 |
|
| If you can't view the Datasheet, Please click here to try to view without PDF Reader . |
|
|
|
| If you can't view the Datasheet, Please click here to try to view without PDF Reader . |
|
|
| Datasheet File OCR Text: |
| to all our customers regarding the change of names ment ioned in the document, such as hitachi electric and hitachi xx, to renesas technology corp. the semiconductor operations of mitsubishi electric and hitachi were transferred to renesas technology corporation on april 1st 2003. these operations include microcomput er, logic, analog and discrete devices, and memory chips other than drams (flash memory, srams etc.) accordingly, although hitachi, hitac hi, ltd., hitachi semiconductors, and other hitachi brand names are m entioned in the document, these names have in fact all been changed to renesas technology corp. thank you for your understanding. ex cept for our corporate trademark, logo and corporate statement, no changes whatsoever have been made to the contents of the document, and these changes do not constitute any alteration to the contents of the document itself. renesas technology home page: www.renesas.com renesas technology corp. customer support dept. april 1, 2003 renesas technology corp.
hitachi superh? risc engine sh7729r hardware manual ade-602-229 rev. 1.0 8/20/01 hitachi, ltd. cautions 1. hitachi neither warrants nor grants licenses of any rights of hitachis or any third partys patent, copyright, trademark, or other intellectual property rights for information contained in this document. hitachi bears no responsibility for problems that may arise with third partys rights, including intellectual property rights, in connection with use of the information contained in this document. 2. products and product specifications may be subject to change without notice. confirm that you have received the latest product standards or specifications before final design, purchase or use. 3. hitachi makes every attempt to ensure that its products are of high quality and reliability. however, contact hitachis sales office before using the product in an application that demands especially high quality and reliability or where its failure or malfunction may directly threaten human life or cause risk of bodily injury, such as aerospace, aeronautics, nuclear power, combustion control, transportation, traffic, safety equipment or medical equipment for life support. 4. design your application so that the product is used within the ranges guaranteed by hitachi particularly for maximum rating, operating supply voltage range, heat radiation characteristics, installation conditions and other characteristics. hitachi bears no responsibility for failure or damage when used beyond the guaranteed ranges. even within the guaranteed ranges, consider normally foreseeable failure rates or failure modes in semiconductor devices and employ systemic measures such as fail-safes, so that the equipment incorporating hitachi product does not cause bodily injury, fire or other consequential damage due to operation of the hitachi product. 5. this product is not designed to be radiation resistant. 6. no one is permitted to reproduce or duplicate, in any form, the whole or part of this document without written approval from hitachi. 7. contact hitachis sales office for any questions regarding this document or hitachi semiconductor products. preface the sh7729r is a microprocessor that integrates peripheral functions necessary for system configuration with a 32-bit internal architecture sh2-dsp cpu as its core. the sh7729r's on-chip peripheral functions include a cache memory, internal x/y memory, an interrupt controller, timers, three serial communication interfaces, a real time clock (rtc), memory management unit (mmu), a user break controller (ubc), a bus state controller (bsc), and i/o ports, making it ideal for use as a microcomputer in electronic devices that require high speed together with low power consumption. intended readership: this manual is intended for users undertaking the design of an application system using the sh7729r. readers using this manual require a basic knowledge of electrical circuits, logic circuits, and microcomputers. purpose: the purpose of this manual is to give users an understanding of the hardware functions and electrical characteristics of the sh7729r.details of execution instructions can be found in the sh-3, sh-3e, sh3-dsp programming manual, which should be read in conjunction with the present manual. using this manual: ? for an overall understanding of the sh7729r's functions follow the table of contents.this manual is broadly divided into sections on the cpu, system control functions, peripheral functions, and electrical characteristics. ? for a detailed understanding of cpu functions refer to the separate publication sh-3, sh-3e, sh3-dsp programming manual. note on bit notation: bits are shown in high-to-low order from left to right. related material: the latest information is available at our web site. please make sure that you have the most up-to-date information available. (http://ww.hitachi.cp.jp/sicd/english/products/micome.htm) user's manuals on the sh7729r: manual title ade no. sh7729r hardware manual this manual sh-3, sh-3e, sh-3dsp programming manual ade-602-096 users manuals for development tools: manual title ade no. c/c++ complier, assembler, optimized linkage editor user's manual ade-702-246 simulator debugger users manual ade-702-186 hitachi embedded workshop users manual ade-702-201 i contents section 1 overview ............................................................................................................ 1 1.1 features ................................................................................................................... ........... 1 1.2 block diagram .............................................................................................................. ..... 7 1.3 pin description............................................................................................................ ....... 8 1.3.1 pin assignment ..................................................................................................... 8 1.3.2 pin function.......................................................................................................... 10 section 2 cpu ..................................................................................................................... 19 2.1 registers.................................................................................................................. ........... 19 2.1.1 general registers.................................................................................................. 23 2.1.2 control registers .................................................................................................. 23 2.1.3 system registers................................................................................................... 30 2.1.4 dsp registers ....................................................................................................... 30 2.2 data formats............................................................................................................... ....... 35 2.2.1 register data format (non-dsp type)................................................................ 35 2.2.2 dsp-type data formats ....................................................................................... 35 2.2.3 memory data formats.......................................................................................... 37 2.3 features of cpu core instructions .................................................................................... 37 2.4 instruction formats ........................................................................................................ .... 41 2.4.1 cpu instruction addressing modes ..................................................................... 41 2.4.2 dsp data addressing ........................................................................................... 45 2.4.3 cpu instruction formats ...................................................................................... 50 2.4.4 dsp instruction formats....................................................................................... 54 2.5 instruction set ............................................................................................................ ........ 60 2.5.1 cpu instruction set .............................................................................................. 60 2.6 dsp extended-function instructions................................................................................. 75 2.6.1 introduction........................................................................................................... 75 2.6.2 added cpu system control instructions ............................................................. 76 2.6.3 single and double data transfer for dsp data instructions ............................... 78 2.6.4 dsp operation instruction set.............................................................................. 82 section 3 memory management unit (mmu) .......................................................... 93 3.1 overview................................................................................................................... ......... 93 3.1.1 features ................................................................................................................. 93 3.1.2 role of mmu........................................................................................................ 93 3.1.3 sh7729r mmu.................................................................................................... 96 3.1.4 register configuration.......................................................................................... 99 3.2 register description....................................................................................................... .... 99 3.3 tlb functions .............................................................................................................. ..... 101 ii 3.3.1 configuration of the tlb ..................................................................................... 101 3.3.2 tlb indexing........................................................................................................ 103 3.3.3 tlb address comparison .................................................................................... 104 3.3.4 page management information............................................................................. 106 3.4 mmu functions .............................................................................................................. ... 107 3.4.1 mmu hardware management.............................................................................. 107 3.4.2 mmu software management ............................................................................... 107 3.4.3 mmu instruction (ldtlb).................................................................................. 108 3.4.4 avoiding synonym problems............................................................................... 109 3.5 mmu exceptions ............................................................................................................. .. 112 3.5.1 tlb miss exception ............................................................................................. 112 3.5.2 tlb protection violation exception .................................................................... 113 3.5.3 tlb invalid exception ......................................................................................... 114 3.5.4 initial page write exception................................................................................. 115 3.5.5 processing flow in event of mmu exception (same processing flow for address error) .......................................................... 117 3.5.6 mmu exception in repeat loop.......................................................................... 119 3.6 memory-mapped tlb ....................................................................................................... 120 3.6.1 address array ....................................................................................................... 120 3.6.2 data array ............................................................................................................ 121 3.6.3 usage examples.................................................................................................... 123 3.7 usage note ................................................................................................................. ........ 123 section 4 exception handling ........................................................................................ 125 4.1 overview................................................................................................................... ......... 125 4.1.1 features ................................................................................................................. 125 4.1.2 register configuration.......................................................................................... 125 4.2 exception handling function ............................................................................................ 125 4.2.1 exception handling flow ..................................................................................... 125 4.2.2 exception vector addresses ................................................................................. 126 4.2.3 acceptance of exceptions..................................................................................... 128 4.2.4 exception codes ................................................................................................... 130 4.2.5 exception request masks ..................................................................................... 131 4.2.6 returning from exception handling .................................................................... 131 4.3 register descriptions ...................................................................................................... ... 132 4.4 exception handling operation .......................................................................................... 133 4.4.1 reset.................................................................................................................... .. 133 4.4.2 interrupts ............................................................................................................... 133 4.4.3 general exceptions ............................................................................................... 134 4.5 individual exception operations ....................................................................................... 134 4.5.1 resets................................................................................................................... . 134 4.5.2 general exceptions ............................................................................................... 135 4.5.3 interrupts ............................................................................................................... 140 iii 4.6 cautions ................................................................................................................... .......... 141 section 5 cache .................................................................................................................. 143 5.1 overview................................................................................................................... ......... 143 5.1.1 features ................................................................................................................. 143 5.1.2 cache structure..................................................................................................... 143 5.1.3 register configuration.......................................................................................... 145 5.2 register descriptions ...................................................................................................... ... 145 5.2.1 cache control register (ccr) ............................................................................. 145 5.2.2 cache control register 2 (ccr2) ........................................................................ 146 5.3 cache operation............................................................................................................ ..... 148 5.3.1 searching the cache.............................................................................................. 148 5.3.2 read access.......................................................................................................... 150 5.3.3 write access ......................................................................................................... 150 5.3.4 write-back buffer ................................................................................................ 150 5.3.5 coherency of cache and external memory.......................................................... 151 5.4 memory-mapped cache .................................................................................................... 151 5.4.1 address array ....................................................................................................... 151 5.4.2 data array ............................................................................................................ 152 5.5 usage examples ............................................................................................................. .... 154 5.5.1 invalidating specific entries................................................................................. 154 5.5.2 reading the data of a specific entry.................................................................... 154 section 6 x/y memory .................................................................................................... 155 6.1 overview................................................................................................................... ......... 155 6.1.1 features ................................................................................................................. 155 6.2 x/y memory access from cpu ........................................................................................ 156 6.3 x/y memory access from dsp ........................................................................................ 158 6.4 x/y memory access from dmac.................................................................................... 158 section 7 interrupt controller (intc) ......................................................................... 159 7.1 overview................................................................................................................... ......... 159 7.1.1 features ................................................................................................................. 159 7.1.2 block diagram...................................................................................................... 160 7.1.3 pin configuration.................................................................................................. 161 7.1.4 register configuration.......................................................................................... 162 7.2 interrupt sources.......................................................................................................... ...... 163 7.2.1 nmi interrupt........................................................................................................ 163 7.2.2 irq interrupts ....................................................................................................... 163 7.2.3 irl interrupts........................................................................................................ 164 7.2.4 pint interrupts ..................................................................................................... 166 7.2.5 on-chip peripheral module interrupts ................................................................. 166 7.2.6 interrupt exception handling and priority ........................................................... 167 iv 7.3 intc registers ............................................................................................................. ..... 173 7.3.1 interrupt priority registers a to e (ipraCipre) ................................................ 173 7.3.2 interrupt control register 0 (icr0)...................................................................... 174 7.3.3 interrupt control register 1 (icr1)...................................................................... 175 7.3.4 interrupt control register 2 (icr2)...................................................................... 178 7.3.5 pint interrupt enable register (pinter) .......................................................... 179 7.3.6 interrupt request register 0 (irr0) ..................................................................... 180 7.3.7 interrupt request register 1 (irr1) ..................................................................... 182 7.3.8 interrupt request register 2 (irr2) ..................................................................... 183 7.4 intc operation ............................................................................................................. .... 185 7.4.1 interrupt sequence ................................................................................................ 185 7.4.2 multiple interrupts ................................................................................................ 187 7.5 interrupt response time.................................................................................................... 187 section 8 user break controller .................................................................................... 191 8.1 overview................................................................................................................... ......... 191 8.1.1 features ................................................................................................................. 191 8.1.2 block diagram...................................................................................................... 193 8.1.3 register configuration.......................................................................................... 194 8.2 register descriptions ...................................................................................................... ... 195 8.2.1 break address register a (bara)...................................................................... 195 8.2.2 break address mask register a (bamra) ........................................................ 196 8.2.3 break bus cycle register a (bbra) .................................................................. 197 8.2.4 break address register b (barb) ...................................................................... 199 8.2.5 break address mask register b (bamrb)......................................................... 200 8.2.6 break data register b (bdrb)............................................................................ 201 8.2.7 break data mask register b (bdmrb) .............................................................. 202 8.2.8 break bus cycle register b (bbrb) ................................................................... 203 8.2.9 break control register (brcr) ........................................................................... 205 8.2.10 break execution times register (betr) ............................................................ 208 8.2.11 branch source register (brsr) .......................................................................... 209 8.2.12 branch destination register (brdr) .................................................................. 211 8.2.13 break asid register a (basra) ....................................................................... 212 8.2.14 break asid register b (basrb)........................................................................ 212 8.3 operation description...................................................................................................... .. 213 8.3.1 flow of the user break operation........................................................................ 213 8.3.2 break on instruction fetch cycle ......................................................................... 213 8.3.3 break by data access cycle................................................................................. 214 8.3.4 break on x/y-memory bus cycle ....................................................................... 215 8.3.5 sequential break ................................................................................................... 215 8.3.6 value of saved program counter ......................................................................... 215 8.3.7 pc trace................................................................................................................ 2 16 8.3.8 examples of use ................................................................................................... 218 v 8.3.9 notes .................................................................................................................... . 223 section 9 power-down modes ...................................................................................... 225 9.1 overview................................................................................................................... ......... 225 9.1.1 power-down modes ............................................................................................. 225 9.1.2 pin configuration.................................................................................................. 227 9.1.3 register configuration.......................................................................................... 227 9.2 register descriptions ...................................................................................................... ... 227 9.2.1 standby control register (stbcr)...................................................................... 227 9.2.2 standby control register 2 (stbcr2) ................................................................ 229 9.3 sleep mode ................................................................................................................. ....... 231 9.3.1 transition to sleep mode...................................................................................... 231 9.3.2 canceling sleep mode.......................................................................................... 231 9.4 standby mode ............................................................................................................... ..... 232 9.4.1 transition to standby mode.................................................................................. 232 9.4.2 canceling standby mode...................................................................................... 233 9.4.3 clock pause function ........................................................................................... 234 9.5 module standby function.................................................................................................. 23 5 9.5.1 transition to module standby function ............................................................... 235 9.5.2 clearing module standby function...................................................................... 236 9.6 timing of status pin changes ...................................................................................... 236 9.6.1 timing for resets.................................................................................................. 236 9.6.2 timing for canceling standby.............................................................................. 238 9.6.3 timing for canceling sleep mode........................................................................ 241 9.7 hardware standby mode ................................................................................................... 243 9.7.1 transition to hardware standby mode................................................................. 243 9.7.2 canceling hardware standby mode ..................................................................... 243 9.7.3 hardware standby mode timing.......................................................................... 244 section 10 on-chip oscillation circuits ....................................................................... 247 10.1 overview.................................................................................................................. .......... 247 10.1.1 features ................................................................................................................ . 247 10.2 overview of cpg........................................................................................................... .... 248 10.2.1 cpg block diagram ............................................................................................. 248 10.2.2 cpg pin configuration ......................................................................................... 250 10.2.3 cpg register configuration ................................................................................. 250 10.3 clock operating modes ..................................................................................................... 251 10.4 register descriptions ..................................................................................................... .... 255 10.4.1 frequency control register (frqcr) ................................................................. 255 10.5 changing the frequency .................................................................................................... 257 10.5.1 changing the multiplication rate......................................................................... 257 10.5.2 changing the division ratio ................................................................................ 257 10.6 overview of wdt ........................................................................................................... .. 258 vi 10.6.1 block diagram of wdt........................................................................................ 258 10.6.2 register configuration.......................................................................................... 258 10.7 wdt registers............................................................................................................. ...... 259 10.7.1 watchdog timer counter (wtcnt).................................................................... 259 10.7.2 watchdog timer control/status register (wtcsr) ........................................... 259 10.7.3 notes on register access...................................................................................... 261 10.8 using the wdt ............................................................................................................. ..... 262 10.8.1 canceling standby ................................................................................................ 262 10.8.2 changing the frequency ....................................................................................... 262 10.8.3 using watchdog timer mode .............................................................................. 263 10.8.4 using interval timer mode .................................................................................. 263 10.9 notes on board design ..................................................................................................... . 264 section 11 bus state controller (bsc) .......................................................................... 267 11.1 overview.................................................................................................................. .......... 267 11.1.1 features ................................................................................................................ . 267 11.1.2 block diagram...................................................................................................... 269 11.1.3 pin configuration.................................................................................................. 270 11.1.4 register configuration.......................................................................................... 272 11.1.5 area overview...................................................................................................... 273 11.1.6 pcmcia support ................................................................................................. 276 11.2 bsc registers ............................................................................................................. ....... 279 11.2.1 bus control register 1 (bcr1)............................................................................ 279 11.2.2 bus control register 2 (bcr2)............................................................................ 283 11.2.3 wait state control register 1 (wcr1) ................................................................ 284 11.2.4 wait state control register 2 (wcr2) ................................................................ 285 11.2.5 individual memory control register (mcr) ....................................................... 289 11.2.6 pcmcia control register (pcr) ........................................................................ 292 11.2.7 synchronous dram mode register (sdmr) ..................................................... 296 11.2.8 refresh timer control/status register (rtcsr) ................................................ 297 11.2.9 refresh timer counter (rtcnt) ........................................................................ 299 11.2.10 refresh time constant register (rtcor).......................................................... 300 11.2.11 refresh count register (rfcr)........................................................................... 300 11.2.12 cautions on accessing refresh control related registers.................................. 301 11.2.13 mcs0 control register (mcscr0) ..................................................................... 302 11.2.14 mcs1 control register (mcscr1) ..................................................................... 303 11.2.15 mcs2 control register (mcscr2) ..................................................................... 303 11.2.16 mcs3 control register (mcscr3) ..................................................................... 303 11.2.17 mcs4 control register (mcscr4) ..................................................................... 303 11.2.18 mcs5 control register (mcscr5) ..................................................................... 303 11.2.19 mcs6 control register (mcscr6) ..................................................................... 303 11.2.20 mcs7 control register (mcscr7) ..................................................................... 303 11.3 bsc operation ............................................................................................................. ...... 304 vii 11.3.1 endian/access size and data alignment ............................................................. 304 11.3.2 description of areas ............................................................................................. 309 11.3.3 basic interface ...................................................................................................... 312 11.3.4 synchronous dram interface ............................................................................. 319 11.3.5 burst rom interface ............................................................................................ 346 11.3.6 pcmcia interface................................................................................................ 349 11.3.7 waits between access cycles .............................................................................. 361 11.3.8 bus arbitration...................................................................................................... 362 11.3.9 bus pull-up .......................................................................................................... 363 11.3.10 mcs[0] to mcs[7] pin control ........................................................................... 365 section 12 direct memory access controller (dmac) .......................................... 369 12.1 overview.................................................................................................................. .......... 369 12.1.1 features ................................................................................................................ . 369 12.1.2 block diagram...................................................................................................... 371 12.1.3 pin configuration.................................................................................................. 372 12.1.4 register configuration.......................................................................................... 373 12.2 register descriptions ..................................................................................................... .... 375 12.2.1 dma source address registers 0C3 (sar0Csar3)........................................... 375 12.2.2 dma destination address registers 0C3 (dar0Cdar3) .................................. 376 12.2.3 dma transfer count registers 0C3 (dmatcr0Cdmatcr3) ......................... 377 12.2.4 dma channel control registers 0C3 (chcr0Cchcr3) ................................... 378 12.2.5 dma operation register (dmaor) ................................................................... 385 12.3 operation................................................................................................................. ........... 387 12.3.1 dma transfer flow ............................................................................................. 387 12.3.2 dma transfer requests ....................................................................................... 389 12.3.3 channel priority.................................................................................................... 391 12.3.4 dma transfer types............................................................................................ 394 12.3.5 number of bus cycle states and dreq pin sampling timing........................... 407 12.3.6 source address reload function.......................................................................... 416 12.3.7 dma transfer ending conditions........................................................................ 418 12.4 compare match timer (cmt)........................................................................................... 420 12.4.1 overview............................................................................................................... 4 20 12.4.2 register descriptions............................................................................................ 421 12.4.3 operation .............................................................................................................. 4 24 12.4.4 compare match .................................................................................................... 425 12.5 examples of use ........................................................................................................... ..... 427 12.5.1 example of dma transfer between on-chip irda and external memory ........ 427 12.5.2 example of dma transfer between a/d converter and external memory (address reload on) ............................................................................................ 428 12.5.3 example of dma transfer between external memory and scif transmitter (indirect address on) ........................................................................................... 429 12.6 usage notes ............................................................................................................... ........ 431 viii section 13 timer (tmu) .................................................................................................... 433 13.1 overview.................................................................................................................. .......... 433 13.1.1 features ................................................................................................................ . 433 13.1.2 block diagram...................................................................................................... 434 13.1.3 pin configuration.................................................................................................. 435 13.1.4 register configuration.......................................................................................... 435 13.2 tmu registers ............................................................................................................. ...... 436 13.2.1 timer output control register (tocr)............................................................... 436 13.2.2 timer start register (tstr) ................................................................................ 436 13.2.3 timer control registers (tcr) ............................................................................ 437 13.2.4 timer constant registers (tcor) ....................................................................... 441 13.2.5 timer counters (tcnt) ....................................................................................... 442 13.2.6 input capture register (tcpr2) .......................................................................... 443 13.3 tmu operation............................................................................................................. ..... 444 13.3.1 overview............................................................................................................... 4 44 13.3.2 basic functions..................................................................................................... 444 13.4 interrupts ................................................................................................................ ............ 448 13.4.1 status flag setting timing.................................................................................... 448 13.4.2 status flag clearing timing ................................................................................. 449 13.4.3 interrupt sources and priorities ............................................................................ 449 13.5 usage notes ............................................................................................................... ........ 450 13.5.1 writing to registers.............................................................................................. 450 13.5.2 reading registers ................................................................................................. 450 section 14 realtime clock (rtc) ................................................................................... 451 14.1 overview.................................................................................................................. .......... 451 14.1.1 features ................................................................................................................ . 451 14.1.2 block diagram...................................................................................................... 452 14.1.3 pin configuration.................................................................................................. 453 14.1.4 rtc register configuration ................................................................................. 454 14.2 rtc registers............................................................................................................. ....... 455 14.2.1 64-hz counter (r64cnt) .................................................................................... 455 14.2.2 second counter (rseccnt) ............................................................................... 455 14.2.3 minute counter (rmincnt) ............................................................................... 456 14.2.4 hour counter (rhrcnt) .................................................................................... 456 14.2.5 day of week counter (rwkcnt) ...................................................................... 457 14.2.6 date counter (rdaycnt).................................................................................. 458 14.2.7 month counter (rmoncnt) .............................................................................. 458 14.2.8 year counter (ryrcnt) ..................................................................................... 459 14.2.9 second alarm register (rsecar)...................................................................... 459 14.2.10 minute alarm register (rminar)...................................................................... 460 14.2.11 hour alarm register (rhrar) ........................................................................... 460 14.2.12 day of week alarm register (rwkar) ............................................................. 461 ix 14.2.13 date alarm register (rdayar) ........................................................................ 462 14.2.14 month alarm register (rmonar) ..................................................................... 462 14.2.15 rtc control register 1 (rcr1) .......................................................................... 463 14.2.16 rtc control register 2 (rcr2) .......................................................................... 464 14.3 rtc operation ............................................................................................................. ...... 466 14.3.1 initial settings of registers after power-on ......................................................... 466 14.3.2 setting the time.................................................................................................... 466 14.3.3 reading the time.................................................................................................. 467 14.3.4 alarm function ..................................................................................................... 468 14.3.5 crystal oscillator circuit...................................................................................... 469 14.4 usage notes ............................................................................................................... ........ 470 14.4.1 register writing during rtc count..................................................................... 470 14.4.2 use of realtime clock (rtc) periodic interrupts................................................ 470 section 15 serial communication interface (sci) ..................................................... 471 15.1 overview.................................................................................................................. .......... 471 15.1.1 features ................................................................................................................ . 471 15.1.2 block diagram...................................................................................................... 472 15.1.3 pin configuration.................................................................................................. 475 15.1.4 register configuration.......................................................................................... 476 15.2 register descriptions ..................................................................................................... .... 476 15.2.1 receive shift register (scrsr) .......................................................................... 476 15.2.2 receive data register (scrdr).......................................................................... 477 15.2.3 transmit shift register (sctsr) ......................................................................... 477 15.2.4 transmit data register (sctdr) ........................................................................ 478 15.2.5 serial mode register (scsmr)............................................................................ 478 15.2.6 serial control register (scscr).......................................................................... 481 15.2.7 serial status register (scssr) ............................................................................ 484 15.2.8 sc port control register (scpcr)/sc port data register (scpdr) ................. 487 15.2.9 bit rate register (scbrr) .................................................................................. 489 15.3 operation................................................................................................................. ........... 497 15.3.1 overview............................................................................................................... 4 97 15.3.2 operation in asynchronous mode........................................................................ 499 15.3.3 multiprocessor communication............................................................................ 509 15.3.4 synchronous operation ........................................................................................ 518 15.4 sci interrupts ............................................................................................................ ......... 528 15.5 usage notes ............................................................................................................... ........ 529 section 16 smart card interface ...................................................................................... 533 16.1 overview.................................................................................................................. .......... 533 16.1.1 features ................................................................................................................ . 533 16.1.2 block diagram...................................................................................................... 534 16.1.3 pin configuration.................................................................................................. 535 x 16.1.4 smart card interface registers ............................................................................. 535 16.2 register descriptions ..................................................................................................... .... 536 16.2.1 smart card mode register (scscmr)................................................................ 536 16.2.2 serial status register (scssr) ............................................................................ 537 16.3 operation................................................................................................................. ........... 538 16.3.1 overview............................................................................................................... 5 38 16.3.2 pin connections .................................................................................................... 539 16.3.3 data format .......................................................................................................... 540 16.3.4 register settings ................................................................................................... 541 16.3.5 clock ................................................................................................................... .. 542 16.3.6 data transmission and reception ........................................................................ 545 16.4 usage notes ............................................................................................................... ........ 551 16.4.1 receive data timing and receive margin in asynchronous mode .................... 551 16.4.2 retransmission (receive and transmit modes) ................................................... 553 section 17 serial communication interface with fifo (scif) ............................. 555 17.1 overview.................................................................................................................. .......... 555 17.1.1 features ................................................................................................................ . 555 17.1.2 block diagram...................................................................................................... 556 17.1.3 pin configuration.................................................................................................. 559 17.1.4 register configuration.......................................................................................... 560 17.2 register descriptions ..................................................................................................... .... 561 17.2.1 receive shift register (scrsr) .......................................................................... 561 17.2.2 receive fifo data register (scfrdr).............................................................. 561 17.2.3 transmit shift register (sctsr) ......................................................................... 561 17.2.4 transmit fifo data register (scftdr)............................................................. 562 17.2.5 serial mode register (scsmr)............................................................................ 562 17.2.6 serial control register (scscr).......................................................................... 564 17.2.7 serial status register (scssr) ............................................................................ 566 17.2.8 bit rate register (scbrr) .................................................................................. 571 17.2.9 fifo control register (scfcr) .......................................................................... 578 17.2.10 fifo data count register (scfdr).................................................................... 580 17.3 operation................................................................................................................. ........... 581 17.3.1 overview............................................................................................................... 5 81 17.3.2 serial operation .................................................................................................... 582 17.4 scif interrupts........................................................................................................... ........ 594 17.5 usage notes ............................................................................................................... ........ 595 section 18 irda .................................................................................................................... 599 18.1 overview.................................................................................................................. .......... 599 18.1.1 features ................................................................................................................ . 599 18.1.2 block diagram...................................................................................................... 600 18.1.3 pin configuration.................................................................................................. 603 xi 18.1.4 register configuration.......................................................................................... 604 18.2 register description...................................................................................................... ..... 605 18.2.1 serial mode register (scsmr)............................................................................ 605 18.3 operation description..................................................................................................... ... 607 18.3.1 overview............................................................................................................... 6 07 18.3.2 transmitting.......................................................................................................... 60 7 18.3.3 receiving .............................................................................................................. 6 08 section 19 pin function controller ................................................................................. 609 19.1 overview.................................................................................................................. .......... 609 19.2 register configuration .................................................................................................... ... 613 19.3 register descriptions ..................................................................................................... .... 614 19.3.1 port a control register (pacr) .......................................................................... 614 19.3.2 port b control register (pbcr)........................................................................... 615 19.3.3 port c control register (pccr)........................................................................... 616 19.3.4 port d control register (pdcr) .......................................................................... 617 19.3.5 port e control register (pecr) ........................................................................... 618 19.3.6 port f control register (pfcr)............................................................................ 619 19.3.7 port g control register (pgcr) .......................................................................... 620 19.3.8 port h control register (phcr) .......................................................................... 621 19.3.9 port j control register (pjcr) ............................................................................. 623 19.3.10 port k control register (pkcr) .......................................................................... 624 19.3.11 port l control register (plcr) ........................................................................... 625 19.3.12 sc port control register (scpcr) ...................................................................... 626 section 20 i/o ports ............................................................................................................. 631 20.1 overview.................................................................................................................. .......... 631 20.2 port a .................................................................................................................... ............. 631 20.2.1 register description.............................................................................................. 631 20.2.2 port a data register (padr)............................................................................... 632 20.3 port b .................................................................................................................... ............. 633 20.3.1 register description.............................................................................................. 633 20.3.2 port b data register (pbdr) ............................................................................... 634 20.4 port c .................................................................................................................... ............. 635 20.4.1 register description.............................................................................................. 635 20.4.2 port c data register (pcdr) ............................................................................... 636 20.5 port d .................................................................................................................... ............. 637 20.5.1 register description.............................................................................................. 637 20.5.2 port d data register (pddr)............................................................................... 638 20.6 port e.................................................................................................................... .............. 639 20.6.1 register description.............................................................................................. 639 20.6.2 port e data register (pedr)................................................................................ 640 20.7 port f.................................................................................................................... .............. 641 xii 20.7.1 register description.............................................................................................. 641 20.7.2 port f data register (pfdr)................................................................................ 642 20.8 port g .................................................................................................................... ............. 643 20.8.1 register description.............................................................................................. 643 20.8.2 port g data register (pgdr)............................................................................... 644 20.9 port h .................................................................................................................... ............. 645 20.9.1 register description.............................................................................................. 645 20.9.2 port h data register (phdr)............................................................................... 646 20.10 port j ................................................................................................................... ............... 647 20.10.1 register description.............................................................................................. 647 20.10.2 port j data register (pjdr) ................................................................................. 648 20.11 port k ................................................................................................................... .............. 649 20.11.1 register description.............................................................................................. 649 20.11.2 port k data register (pkdr)............................................................................... 650 20.12 port l................................................................................................................... ............... 651 20.12.1 register description.............................................................................................. 651 20.12.2 port l data register (pldr)................................................................................ 652 20.13 sc port .................................................................................................................. ............. 653 20.13.1 register description.............................................................................................. 653 20.13.2 port sc data register (scpdr)........................................................................... 654 section 21 a/d converter .................................................................................................. 657 21.1 overview.................................................................................................................. .......... 657 21.1.1 features ................................................................................................................ . 657 21.1.2 block diagram...................................................................................................... 658 21.1.3 input pins .............................................................................................................. 659 21.1.4 register configuration.......................................................................................... 660 21.2 register descriptions ..................................................................................................... .... 661 21.2.1 a/d data registers a to d (addra to addrd).............................................. 661 21.2.2 a/d control/status register (adcsr) ................................................................ 662 21.2.3 a/d control register (adcr) ............................................................................. 664 21.3 bus master interface ...................................................................................................... .... 665 21.4 operation................................................................................................................. ........... 666 21.4.1 single mode (multi = 0) ................................................................................... 666 21.4.2 multi mode (multi = 1)..................................................................................... 668 21.4.3 scan mode (multi = 1, scn = 1)...................................................................... 670 21.4.4 input sampling and a/d conversion time .......................................................... 672 21.4.5 external trigger input timing.............................................................................. 673 21.5 interrupts ................................................................................................................ ............ 674 21.6 definitions of a/d conversion accuracy.......................................................................... 674 21.7 usage notes ............................................................................................................... ........ 675 21.7.1 setting analog input voltage ............................................................................... 675 21.7.2 processing of analog input pins........................................................................... 675 xiii 21.7.3 access size and read data................................................................................... 676 section 22 d/a converter .................................................................................................. 679 22.1 overview.................................................................................................................. .......... 679 22.1.1 features ................................................................................................................ . 679 22.1.2 block diagram...................................................................................................... 679 22.1.3 i/o pins ................................................................................................................ . 680 22.1.4 register configuration.......................................................................................... 680 22.2 register descriptions ..................................................................................................... .... 681 22.2.1 d/a data registers 0 and 1 (dadr0/1) .............................................................. 681 22.2.2 d/a control register (dacr) ............................................................................. 681 22.3 operation................................................................................................................. ........... 683 section 23 hitachi user debugging interface (h-udi) ........................................... 685 23.1 overview.................................................................................................................. .......... 685 23.2 hitachi user debugging interface (h-udi) ...................................................................... 685 23.2.1 pin descriptions.................................................................................................... 685 23.2.2 block diagram...................................................................................................... 686 23.3 register descriptions ..................................................................................................... .... 686 23.3.1 bypass register (sdbpr) .................................................................................... 687 23.3.2 instruction register (sdir).................................................................................. 687 23.3.3 boundary scan register (sdbsr) ....................................................................... 688 23.4 h-udi operation ........................................................................................................... .... 695 23.4.1 tap controller...................................................................................................... 695 23.4.2 reset configuration .............................................................................................. 696 23.4.3 h-udi reset ......................................................................................................... 696 23.4.4 h-udi interrupt .................................................................................................... 697 23.4.5 bypass .................................................................................................................. . 697 23.4.6 using h-udi to recover from sleep mode ......................................................... 697 23.5 boundary scan ............................................................................................................. ...... 697 23.5.1 supported instructions .......................................................................................... 697 23.5.2 points for attention............................................................................................... 698 23.6 usage notes ............................................................................................................... ........ 699 23.7 advanced user debugger (aud)...................................................................................... 699 section 24 electrical characteristics .............................................................................. 701 24.1 absolute maximum ratings .............................................................................................. 701 24.2 dc characteristics ........................................................................................................ ..... 703 24.3 ac characteristics ........................................................................................................ ..... 706 24.3.1 clock timing ........................................................................................................ 707 24.3.2 control signal timing .......................................................................................... 716 24.3.3 ac bus timing ..................................................................................................... 719 24.3.4 basic timing......................................................................................................... 721 xiv 24.3.5 burst rom timing ............................................................................................... 724 24.3.6 synchronous dram timing................................................................................ 727 24.3.7 pcmcia timing .................................................................................................. 745 24.3.8 peripheral module signal timing ........................................................................ 752 24.3.9 h-udi-related pin timing .................................................................................. 756 24.3.10 ac characteristic test conditions ....................................................................... 758 24.3.11 delay time variation due to load capacitance.................................................. 759 24.4 a/d converter characteristics ........................................................................................... 7 60 24.5 d/a converter characteristics ........................................................................................... 76 0 appendix a pin functions .................................................................................................. 761 a.1 pin states................................................................................................................. ........... 761 a.2 pin specifications......................................................................................................... ...... 765 a.3 treatment of unused pins.................................................................................................. 7 69 a.4 pin states in access to each address space...................................................................... 770 appendix b memory-mapped control registers ........................................................ 784 b.1 register address map....................................................................................................... . 784 b.2 register bits .............................................................................................................. ......... 790 appendix c product lineup ............................................................................................... 802 appendix d package dimensions .................................................................................... 803 xv figures figure 1.1 block diagram ........................................................................................................ 7 figure 1.2 pin assignment (fp-208c, fp-208e) ..................................................................... 8 figure 1.3 pin assignment (bp-240a) .................................................................................... 9 figure 2.1 register configuration in each processing mode (1)............................................. 21 figure 2.2 register configuration in each processing mode (2)............................................. 22 figure 2.3 general purpose registers (not in dsp mode)...................................................... 23 figure 2.4 general purpose registers (dsp mode) ................................................................. 24 figure 2.5 control registers.................................................................................................... .27 figure 2.6 system registers ..................................................................................................... 30 figure 2.7 dsp registers........................................................................................................ .. 32 figure 2.8 connections of dsp registers and buses ............................................................... 34 figure 2.9 longword operand ................................................................................................. 35 figure 2.10 data formats ........................................................................................................ ... 36 figure 2.11 byte, word, and longword alignment .................................................................. 37 figure 2.12 x and y data transfer addressing ......................................................................... 46 figure 2.13 single data transfer addressing ............................................................................ 47 figure 2.14 modulo addressing ................................................................................................. 48 figure 2.15 dsp instruction formats ......................................................................................... 54 figure 2.16 sample parallel instruction program ...................................................................... 83 figure 2.17 examples of conditional operations and data transfer instructions..................... 91 figure 3.1 mmu functions ...................................................................................................... 95 figure 3.2 virtual address space mapping ............................................................................. 97 figure 3.3 mmu register contents ......................................................................................... 100 figure 3.4 overall configuration of the tlb .......................................................................... 101 figure 3.5 virtual address and tlb structure ........................................................................ 102 figure 3.6 tlb indexing (ix = 1)............................................................................................ 103 figure 3.7 tlb indexing (ix = 0)............................................................................................ 104 figure 3.8 objects of address comparison.............................................................................. 105 figure 3.9 operation of ldtlb instruction ............................................................................ 109 figure 3.10 synonym problem ................................................................................................... 11 1 figure 3.11 mmu exception generation flowchart.................................................................. 116 figure 3.12 mmu exception signals in instruction fetch ........................................................ 117 figure 3.13 mmu exception signals in data access................................................................ 118 figure 3.14 mmu exception in repeat loop............................................................................ 119 figure 3.15 specifying address and data for memory-mapped tlb access .......................... 122 figure 4.1 vector table......................................................................................................... ... 126 figure 4.2 example of acceptance order of general exceptions............................................ 129 figure 4.3 bit configurations of expevt, intevt, intevt2, and tra registers........... 132 figure 5.1 cache structure ...................................................................................................... . 144 figure 5.2 ccr register configuration ................................................................................... 146 figure 5.3 ccr2 register configuration ................................................................................. 147 xvi figure 5.4 cache search scheme (normal mode) ................................................................... 149 figure 5.5 write-back buffer configuration ........................................................................... 150 figure 5.6 specifying address and data for memory-mapped cache access........................ 153 figure 6.1 x/y memory logical address mapping ................................................................ 157 figure 6.2 x/y memory physical address mapping ............................................................... 158 figure 7.1 block diagram of intc.......................................................................................... 160 figure 7.2 example of irl interrupt connection .................................................................... 164 figure 7.3 interrupt operation flowchart................................................................................. 186 figure 7.4 example of pipeline operations when irl interrupt is accepted.......................... 190 figure 8.1 block diagram of user break controller ............................................................... 193 figure 8.2 when interrupt occurs before branch instruction is executed .............................. 216 figure 9.1 canceling standby mode with stbcr.stby ....................................................... 233 figure 9.2 power-on reset (clock modes 0, 1, 2, and 7) status output ........................... 236 figure 9.3 manual reset status output .............................................................................. 237 figure 9.4 standby to interrupt status output .................................................................... 238 figure 9.5 standby to power-on reset status output ........................................................ 239 figure 9.6 standby to manual reset status output ............................................................ 240 figure 9.7 sleep to interrupt status output ........................................................................ 241 figure 9.8 sleep to power-on reset status output ............................................................ 241 figure 9.9 sleep to manual reset status output ................................................................ 242 figure 9.10 hardware standby mode (when ca goes low in normal operation)................ 244 figure 9.11 hardware standby mode timing (when ca goes low during wdt operation on standby mode cancellation) ... 245 figure 10.1 block diagram of clock pulse generator............................................................... 248 figure 10.2 block diagram of wdt.......................................................................................... 258 figure 10.3 writing to wtcnt and wtcsr............................................................................ 261 figure 10.4 points for attention when using crystal resonator ............................................... 264 figure 10.5 points for attention when using pll oscillator circuit........................................ 265 figure 11.1 block diagram of bus state controller .................................................................. 269 figure 11.2 correspondence between logical address space and physical address space..... 273 figure 11.3 physical space allocation ....................................................................................... 275 figure 11.4 pcmcia space allocation ..................................................................................... 276 figure 11.5 writing to rfcr, rtcsr, rtcnt, and rtcor ................................................. 301 figure 11.6 basic timing of basic interface.............................................................................. 313 figure 11.7 example of 32-bit data-width static ram connection........................................ 314 figure 11.8 example of 16-bit data-width static ram connection........................................ 315 figure 11.9 example of 8-bit data-width static ram connection.......................................... 316 figure 11.10 basic interface wait timing (software wait only) ............................................... 317 figure 11.11 basic interface wait state timing (wait state insertion by wait signal waitsel = 1).......................................... 318 figure 11.12 example of 64-mbit synchronous dram connection (32-bit bus width).......... 320 figure 11.13 example of 64-mbit synchronous dram (16-bit bus width) ............................. 321 figure 11.14 basic timing for synchronous dram burst read................................................ 324 xvii figure 11.15 synchronous dram burst read wait specification timing ................................ 325 figure 11.16 basic timing for synchronous dram single read .............................................. 326 figure 11.17 basic timing for synchronous dram burst write............................................... 328 figure 11.18 basic timing for synchronous dram single write ............................................. 330 figure 11.19 burst read timing (no precharge)......................................................................... 333 figure 11.20 burst read timing (same row address) ............................................................... 334 figure 11.21 burst read timing (different row addresses)...................................................... 335 figure 11.22 burst write timing (no precharge)........................................................................ 336 figure 11.23 burst write timing (same row address) .............................................................. 337 figure 11.24 burst write timing (different row addresses) ..................................................... 338 figure 11.25 auto-refresh operation .......................................................................................... 340 figure 11.26 synchronous dram auto-refresh timing ........................................................... 341 figure 11.27 synchronous dram self-refresh timing ............................................................. 343 figure 11.28 synchronous dram mode write timing.............................................................. 345 figure 11.29 burst rom wait access timing ............................................................................ 347 figure 11.30 burst rom basic access timing ........................................................................... 348 figure 11.31 example of pcmcia interface............................................................................... 350 figure 11.32 basic timing for pcmcia memory card interface .............................................. 352 figure 11.33 wait timing for pcmcia memory card interface................................................ 353 figure 11.34 basic timing for pcmcia memory card interface burst access ........................ 354 figure 11.35 wait timing for pcmcia memory card interface burst access.......................... 355 figure 11.36 pcmcia space allocation ..................................................................................... 356 figure 11.37 basic timing for pcmcia i/o card interface ....................................................... 358 figure 11.38 wait timing for pcmcia i/o card interface........................................................ 359 figure 11.39 dynamic bus sizing timing for pcmcia i/o card interface............................... 360 figure 11.40 waits between access cycles ................................................................................. 362 figure 11.41 pull-up timing for pins a25 to a0........................................................................ 363 figure 11.42 pull-up timing for pins d31 to d0 (read cycle).................................................. 364 figure 11.43 pull-up timing for pins d31 to d0 (write cycle) ................................................. 364 figure 12.1 block diagram of dmac ....................................................................................... 371 figure 12.2 dmac transfer flowchart ..................................................................................... 388 figure 12.3 round-robin mode................................................................................................. 392 figure 12.4 changes in channel priority in round-robin mode .............................................. 393 figure 12.5 operation of direct address mode in dual address mode.................................... 395 figure 12.6 example of dma transfer timing in the direct address mode in dual mode (transfer source: ordinary memory, transfer destination: ordinary memory) ... 396 figure 12.7 example of dma transfer timing in the direct address mode in dual mode (16-byte transfer, transfer source: ordinary memory, transfer destination: ordinary memory) .................................................................................................. 397 figure 12.8 example of dma transfer timing in the direct address mode in dual mode (16-byte transfer, transfer source: synchronous dram, transfer destination: ordinary memory) .................................................................................................. 397 xviii figure 12.9 indirect address operation in dual address mode (when external memory space has a 16-bit width) ............................................. 399 figure 12.10 example of transfer timing in the indirect address mode in dual address mode ............................................................................................ 400 figure 12.11 data flow in single address mode ........................................................................ 401 figure 12.12 example of dma transfer timing in single address mode ................................. 402 figure 12.13 example of dma transfer timing in single address mode (external memory space (ordinary memory) external device with dack) .. 403 figure 12.14 example of transfer in cycle-steal mode.............................................................. 404 figure 12.15 example of transfer in burst mode........................................................................ 404 figure 12.16 bus state when multiple channels are operating ................................................. 406 figure 12.17 cycle-steal mode, level input (cpu access: 2 cycles)........................................ 409 figure 12.18 cycle-steal mode, level input (cpu access: 3 cycles)........................................ 410 figure 12.19 cycle-steal mode, level input (cpu access: 2 cycles, dma rd access: 4 cycles)............................................ 411 figure 12.20 cycle-steal mode, level input (cpu access: 2 cycles, dreq input delayed) ..................................................... 412 figure 12.21 cycle-steal mode, edge input (cpu access: 2 cycles)......................................... 413 figure 12.22 burst mode, level input ......................................................................................... 414 figure 12.23 burst mode, edge input .......................................................................................... 415 figure 12.24 source address reload function diagram ............................................................. 416 figure 12.25 timing chart of source address reload function ................................................. 417 figure 12.26 block diagram of cmt .......................................................................................... 420 figure 12.27 counter operation .................................................................................................. . 424 figure 12.28 count timing ....................................................................................................... ... 425 figure 12.29 cmf setting timing ............................................................................................... 42 6 figure 12.30 timing of cmf clearing by the cpu..................................................................... 426 figure 13.1 block diagram of tmu .......................................................................................... 434 figure 13.2 setting the count operation.................................................................................... 445 figure 13.3 auto-reload count operation ................................................................................ 446 figure 13.4 count timing when operating on internal clock .................................................. 446 figure 13.5 count timing when operating on external clock (both edges detected) ............ 447 figure 13.6 count timing when operating on on-chip rtc clock ........................................ 447 figure 13.7 operation timing when using input capture function (using tclk rising edge) .................................................................................... 448 figure 13.8 unf setting timing................................................................................................ 44 8 figure 13.9 status flag clearing timing ................................................................................... 449 figure 14.1 block diagram of rtc ........................................................................................... 452 figure 14.2 setting the time .................................................................................................... .. 466 figure 14.3 reading the time .................................................................................................... 467 figure 14.4 using the alarm function ....................................................................................... 468 figure 14.5 example of crystal oscillator circuit connection ................................................. 469 figure 14.6 using periodic interrupt function........................................................................... 470 xix figure 15.1 block diagram of sci............................................................................................. 472 figure 15.2 scpt[1]/sck0 pin ................................................................................................. 473 figure 15.3 scpt[0]/txd0 pin.................................................................................................. 47 4 figure 15.4 scpt[0]/rxd0 pin.................................................................................................. 47 5 figure 15.5 example of data format in asynchronous communication (8-bit data with parity and two stop bits)............................................................ 499 figure 15.6 output clock and serial data timing (asynchronous mode)................................ 501 figure 15.7 sample flowchart for sci initialization ................................................................. 502 figure 15.8 sample flowchart for transmitting serial data ..................................................... 503 figure 15.9 example of sci transmit operation in asynchronous mode (8-bit data with parity and one stop bit) .............................................................. 505 figure 15.10 sample flowchart for receiving serial data.......................................................... 506 figure 15.11 example of sci receive operation (8-bit data with parity and one stop bit) .... 509 figure 15.12 communication among processors using multiprocessor format (sending data h'aa to receiving processor a) .................................................... 510 figure 15.13 sample flowchart for transmitting multiprocessor serial data ............................ 511 figure 15.14 example of sci multiprocessor transmit operation (8-bit data with multiprocessor bit and one stop bit).......................................... 513 figure 15.15 sample flowchart for receiving multiprocessor serial data................................. 514 figure 15.16 example of sci receive operation (8-bit data with multiprocessor bit and one stop bit)..........................................516 figure 15.17 data format in synchronous communication........................................................ 518 figure 15.18 sample flowchart for sci initialization ................................................................. 520 figure 15.19 sample flowchart for transmitting serial data ..................................................... 521 figure 15.20 example of sci transmit operation ...................................................................... 522 figure 15.21 sample flowchart for receiving serial data.......................................................... 524 figure 15.22 example of sci receive operation ........................................................................ 526 figure 15.23 sample flowchart for transmitting/receiving serial data .................................... 527 figure 15.24 receive data sampling timing in asynchronous mode........................................ 530 figure 16.1 block diagram of smart card interface ................................................................. 534 figure 16.2 pin connection diagram for smart card interface ................................................. 539 figure 16.3 data format for smart card interface .................................................................... 540 figure 16.4 waveform of start character .................................................................................. 542 figure 16.5 initialization flowchart (example) ......................................................................... 546 figure 16.6 transmission flowchart .......................................................................................... 548 figure 16.7 reception flowchart (example).............................................................................. 550 figure 16.8 receive data sampling timing in smart card mode ............................................ 552 figure 16.9 retransmission in sci receive mode .................................................................... 553 figure 16.10 retransmission in sci transmit mode ................................................................... 554 figure 17.1 block diagram of scif .......................................................................................... 556 figure 17.2 scpt[5]/sck2 pin ................................................................................................. 557 figure 17.3 scpt[4]/txd2 pin.................................................................................................. 55 8 figure 17.4 scpt[4]/rxd2 pin.................................................................................................. 55 9 xx figure 17.5 sample flowchart for scif initialization ............................................................... 584 figure 17.6 sample flowchart for transmitting serial data ..................................................... 586 figure 17.7 example of transmit operation (8-bit data, parity, one stop bit) ....................... 588 figure 17.8 example of operation using modem control (cts) ............................................. 588 figure 17.9 sample flowchart for receiving serial data.......................................................... 590 figure 17.10 sample flowchart for receiving serial data (cont) ............................................... 591 figure 17.11 example of scif receive operation (8-bit data, parity, one stop bit) ............... 593 figure 17.12 example of operation using modem control (rts) ............................................. 593 figure 17.13 receive data sampling timing in asynchronous mode........................................ 596 figure 18.1 block diagram of irda .......................................................................................... 600 figure 18.2 scpt[3]/sck1 pin ................................................................................................. 601 figure 18.3 scpt[2]/txd1 pin.................................................................................................. 60 2 figure 18.4 scpt[2]/rxd1 pin.................................................................................................. 60 3 figure 18.5 transmit/receive operation ................................................................................... 608 figure 20.1 port a .............................................................................................................. ........ 631 figure 20.2 port b .............................................................................................................. ........ 633 figure 20.3 port c .............................................................................................................. ........ 635 figure 20.4 port d .............................................................................................................. ........ 637 figure 20.5 port e.............................................................................................................. ......... 639 figure 20.6 port f .............................................................................................................. ......... 641 figure 20.7 port g .............................................................................................................. ........ 643 figure 20.8 port h .............................................................................................................. ........ 645 figure 20.9 port j.............................................................................................................. .......... 647 figure 20.10 port k ............................................................................................................. ......... 649 figure 20.11 port l............................................................................................................. .......... 651 figure 20.12 sc port ............................................................................................................ ........ 653 figure 21.1 block diagram of a/d converter ........................................................................... 658 figure 21.2 a/d data register access operation (reading h'aa40) ...................................... 665 figure 21.3 example of a/d converter operation (single mode, channel 1 selected) ........... 667 figure 21.4 example of a/d converter operation (multi mode, channels an0 to an2 selected) ..................................................... 669 figure 21.5 example of a/d converter operation (scan mode, channels an0 to an2 selected) ....................................................... 671 figure 21.6 a/d conversion timing.......................................................................................... 672 figure 21.7 external trigger input timing ................................................................................ 673 figure 21.8 definitions of a/d conversion accuracy ............................................................... 675 figure 21.9 example of analog input protection circuit........................................................... 676 figure 21.10 analog input pin equivalent circuit ....................................................................... 676 figure 22.1 block diagram of d/a converter ........................................................................... 679 figure 22.2 example of d/a converter operation .................................................................... 683 figure 23.1 block diagram of h-udi........................................................................................ 686 figure 23.2 tap controller state transitions............................................................................ 695 figure 23.3 h-udi reset......................................................................................................... ... 696 xxi figure 24.1 extal clock input timing ................................................................................... 711 figure 24.2 ckio clock input timing ...................................................................................... 711 figure 24.3 ckio clock output timing.................................................................................... 711 figure 24.4 power-on oscillation settling time........................................................................ 712 figure 24.5 oscillation settling time on return from standby (return by reset) ................... 712 figure 24.6 oscillation settling time on return from standby (return by nmi) .................... 713 figure 24.7 oscillation settling time on return from standby (return by irq4 to irq0 , pint0/1 and irl3 to irl0 )......................................... 713 figure 24.8 pll synchronization settling time in case of reset or nmi ............................... 714 figure 24.9 pll synchronization settling time in case of irq/irl interrupt........................ 714 figure 24.10 pll synchronization settling time in case of irq/irl interrupt........................ 715 figure 24.11 reset input timing................................................................................................. . 717 figure 24.12 interrupt signal input timing ................................................................................. 717 figure 24.13 irqout timing ..................................................................................................... 717 figure 24.14 bus release timing ................................................................................................ 7 18 figure 24.15 pin drive timing at standby .................................................................................. 718 figure 24.16 basic bus cycle (no wait) ..................................................................................... 721 figure 24.17 basic bus cycle (one wait) ................................................................................... 722 figure 24.18 basic bus cycle (external wait, waitsel = 1) .................................................. 723 figure 24.19 burst rom bus cycle (no wait) ........................................................................... 724 figure 24.20 burst rom bus cycle (two waits) ....................................................................... 725 figure 24.21 burst rom bus cycle (external wait, waitsel = 1)......................................... 726 figure 24.22 synchronous dram read bus cycle (rcd = 0, cas latency = 1, tpc = 0)..... 727 figure 24.23 synchronous dram read bus cycle (rcd = 2, cas latency = 2, tpc = 1)..... 728 figure 24.24 synchronous dram read bus cycle (burst read (single read 4), rcd = 0, cas latency = 1, tpc = 1).................. 729 figure 24.25 synchronous dram read bus cycle (burst read (single read 4), rcd = 1, cas latency = 3, tpc = 0).................. 730 figure 24.26 synchronous dram write bus cycle (rcd = 0, tpc = 0, trwl = 0).............. 731 figure 24.27 synchronous dram write bus cycle (rcd = 2, tpc = 1, trwl = 1).............. 732 figure 24.28 synchronous dram write bus cycle (burst write (single write 4), rcd = 0, tpc = 1, trwl = 0).............................................................................. 733 figure 24.29 synchronous dram write bus cycle (burst mode (single write 4), rcd = 1, tpc = 0, trwl = 0).......................... 734 figure 24.30 synchronous dram burst read bus cycle (ras down, same row address, cas latency = 1) ............................................735 figure 24.31 synchronous dram burst read bus cycle (ras down, same row address, cas latency = 2) ............................................736 figure 24.32 synchronous dram burst read bus cycle (ras down, different row address, tpc = 0, rcd = 0, cas latency = 1) ...... 737 figure 24.33 synchronous dram burst read bus cycle (ras down, different row address, tpc = 1, rcd = 0, cas latency = 1) ...... 738 xxii figure 24.34 synchronous dram burst write bus cycle (ras down, same row address) .......................................................................... 739 figure 24.35 synchronous dram burst write bus cycle (ras down, different row address, tpc = 0, rcd = 0) .................................... 740 figure 24.36 synchronous dram burst write bus cycle (ras down, different row address, tpc = 1, rcd = 1) .................................... 741 figure 24.37 synchronous dram auto-refresh timing (tras = 1, tpc = 1)........................ 742 figure 24.38 synchronous dram self-refresh cycle (tras = 1, tpc = 1)............................ 743 figure 24.39 synchronous dram mode register write cycle.................................................. 744 figure 24.40 pcmcia memory bus cycle (ted = 0, teh = 0, no wait)................................ 745 figure 24.41 pcmcia memory bus cycle (ted = 2, teh = 1, one wait, external wait, waitsel = 1) ............................ 746 figure 24.42 pcmcia memory bus cycle (burst read, ted = 0, teh = 0, no wait)............ 747 figure 24.43 pcmcia memory bus cycle (burst read, ted = 1, teh = 1, two waits, burst pitch = 3, waitsel = 1)..... 748 figure 24.44 pcmcia i/o bus cycle (ted = 0, teh = 0, no wait) ........................................ 749 figure 24.45 pcmcia i/o bus cycle (ted = 2, teh = 1, one wait, external wait, waitsel = 1) ............................750 figure 24.46 pcmcia i/o bus cycle (ted = 1, teh = 1, one wait, bus sizing, waitsel = 1) ................................. 751 figure 24.47 tclk input timing ................................................................................................ 75 3 figure 24.48 tclk clock input timing...................................................................................... 753 figure 24.49 rtc crystal oscillator oscillation settling time at power-on ............................. 753 figure 24.50 sck input clock timing ........................................................................................ 753 figure 24.51 sci i/o timing in synchronous mode ................................................................... 754 figure 24.52 i/o port timing .................................................................................................... ... 754 figure 24.53 dreq input timing................................................................................................ 755 figure 24.54 drak output timing ............................................................................................ 755 figure 24.55 tck input timing .................................................................................................. 7 56 figure 24.56 trst input timing (reset hold) ........................................................................... 757 figure 24.57 h-udi data transfer timing.................................................................................. 757 figure 24.58 asemd0 input timing .......................................................................................... 757 figure 24.59 output load circuit ................................................................................................ 758 figure 24.60 load capacitance vs. delay time .......................................................................... 759 figure d.1 package dimensions (fp-208c) ............................................................................. 803 figure d.2 package dimensions (fp-208e).............................................................................. 804 figure d.3 package dimensions (bp-240av).......................................................................... 805 xxiii tables table 1.1 sh7729r features .................................................................................................. 2 table 1.2 characteristics ....................................................................................................... .. 6 table 1.3 sh7729r pin functions.......................................................................................... 10 table 2.1 initial register values ............................................................................................. 22 table 2.2 operation of sr bits in each sh-3 dsp mode...................................................... 29 table 2.3 destination register in dsp instructions................................................................ 31 table 2.4 source register in dsp operations ........................................................................ 32 table 2.5 dsr register bits ................................................................................................... 33 table 2.6 word data sign extension...................................................................................... 38 table 2.7 delayed branch instructions ................................................................................... 38 table 2.8 t bit ................................................................................................................. ....... 39 table 2.9 immediate data referencing................................................................................... 39 table 2.10 absolute address referencing................................................................................ 40 table 2.11 displacement referencing ...................................................................................... 40 table 2.12 addressing modes and effective addresses for cpu instructions ........................ 41 table 2.13 overview of data transfer instructions.................................................................. 45 table 2.14 cpu instruction formats ........................................................................................ 51 table 2.15 double data transfer instruction formats.............................................................. 55 table 2.16 single data transfer instruction formats ............................................................... 56 table 2.17 a-field parallel data transfer instructions ............................................................ 57 table 2.18 b-field alu operation instructions and multiply instructions............................. 58 table 2.19 cpu instruction types............................................................................................ 60 table 2.20 data transfer instructions ....................................................................................... 64 table 2.21 arithmetic operation instructions .......................................................................... 66 table 2.22 logic operation instructions .................................................................................. 69 table 2.23 shift instructions ................................................................................................... .. 7 0 table 2.24 branch instructions ................................................................................................. 7 1 table 2.25 system control instructions.................................................................................... 72 table 2.26 added cpu system control instructions................................................................ 77 table 2.27 double data transfer instructions .......................................................................... 79 table 2.28 single data transfer instructions............................................................................ 80 table 2.29 correspondence between dsp data transfer operands and registers .................. 81 table 2.30 dsp operation instruction formats ........................................................................ 82 table 2.31 correspondence between dsp instruction operands and registers ....................... 83 table 2.32 dsp operation instructions .................................................................................... 84 table 2.33 dc bit update definitions...................................................................................... 90 table 2.34 examples of nopx and nopy instruction codes................................................. 92 table 3.1 register configuration ............................................................................................ 99 table 3.2 access states designated by d, c, and pr bits ..................................................... 106 table 4.1 register configuration ............................................................................................ 125 table 4.2 exception event vectors ......................................................................................... 127 xxiv table 4.3 exception codes...................................................................................................... 1 30 table 4.4 types of reset ........................................................................................................ . 135 table 5.1 cache specifications ............................................................................................... 143 table 5.2 lru and way replacement.................................................................................... 145 table 5.3 register configuration ............................................................................................ 145 table 5.4 lru and way replacement (when w2lock=1) ................................................. 147 table 5.5 lru and way replacement (when w3lock=1) ................................................. 147 table 5.6 lru and way replacement (when w2lock=1 and w3lock=1)..................... 148 table 6.1 x/y memory specifications.................................................................................... 155 table 7.1 intc pins............................................................................................................. ... 161 table 7.2 intc registers........................................................................................................ 162 table 7.3 irl3 C irl0 / irls3 C irls0 pins and interrupt levels............................................ 165 table 7.4 interrupt exception handling sources and priority (irq mode) ........................... 168 table 7.5 interrupt exception handling sources and priority (irl mode)............................ 170 table 7.6 interrupt levels and intevt codes...................................................................... 172 table 7.7 interrupt request sources and ipraCipre............................................................ 173 table 7.8 interrupt response time ......................................................................................... 188 table 8.1 ubc registers......................................................................................................... 194 table 8.2 data access cycle addresses and operand size comparison conditions............. 214 table 8.3 bsa values stored in exception handling before execution of branch destination instruction ........................................................................... 217 table 9.1 power-down modes................................................................................................ 226 table 9.2 pin configuration .................................................................................................... 2 27 table 9.3 register configuration ............................................................................................ 227 table 9.4 register states in standby mode ............................................................................ 232 table 10.1 cpg pins and functions.......................................................................................... 250 table 10.2 cpg register......................................................................................................... .. 250 table 10.3 clock operating modes .......................................................................................... 251 table 10.4 available combinations of clock mode and frqcr values................................ 252 table 10.5 register configuration ............................................................................................ 258 table 11.1 bsc pins ............................................................................................................. .... 270 table 11.2 bsc registers ........................................................................................................ . 272 table 11.3 physical address space map .................................................................................. 274 table 11.4 correspondence between external pins (md4 and md3) and memory size ........ 275 table 11.5 pcmcia interface characteristics.......................................................................... 276 table 11.6 pcmcia support interface .................................................................................... 277 table 11.7 32-bit external device/big-endian access and data alignment .......................... 304 table 11.8 16-bit external device/big-endian access and data alignment .......................... 305 table 11.9 8-bit external device/big-endian access and data alignment ............................ 306 table 11.10 32-bit external device/little-endian access and data alignment........................ 307 table 11.11 16-bit external device/little-endian access and data alignment........................ 307 table 11.12 8-bit external device/little-endian access and data alignment ......................... 308 table 11.13 relationship between bus width, amx bits, and address multiplex output ...... 322 xxv table 11.14 example of correspondence between sh7729r and synchronous dram address pins (amx2C0 = 011 (32-bit bus width)) .............................................. 323 table 11.15 mcscrx settings and mcs[x] assertion conditions (x: 0C7) ............................. 366 table 12.1 dmac pins ............................................................................................................ . 372 table 12.2 dmac registers ..................................................................................................... 37 3 table 12.3 selecting external request modes with rs bits .................................................... 389 table 12.4 selecting on-chip peripheral module request modes with rs bits..................... 390 table 12.5 supported dma transfers...................................................................................... 394 table 12.6 relationship between request modes and bus modes by dma transfer category .................................................................................... 405 table 12.7 register configuration ............................................................................................ 421 table 12.8 transfer conditions and register settings for transfer between on-chip sci and external memory ......................................................... 427 table 12.9 transfer conditions and register settings for transfer between on-chip a/d converter and external memory........................................ 428 table 12.10 values in dmac after end of fourth transfer ...................................................... 429 table 12.11 transfer conditions and register settings for transfer between external memory and scif transmitter .................................................. 430 table 13.1 tmu pin.............................................................................................................. .... 435 table 13.2 tmu registers ........................................................................................................ 435 table 13.3 tmu interrupt sources ........................................................................................... 449 table 14.1 rtc pins ............................................................................................................. .... 453 table 14.2 rtc registers ........................................................................................................ . 454 table 14.3 day-of-week codes (rwkcnt) ........................................................................... 457 table 14.4 day-of-week codes (rwkar) ............................................................................. 461 table 14.5 recommended oscillator circuit constants (recommended values) ................... 469 table 15.1 sci pins............................................................................................................. ...... 475 table 15.2 sci registers........................................................................................................ ... 476 table 15.3 scsmr settings...................................................................................................... 4 90 table 15.4 bit rates and scbrr settings in asynchronous mode......................................... 490 table 15.5 bit rates and scbrr settings in synchronous mode ........................................... 494 table 15.6 maximum bit rates for various frequencies with baud rate generator (asynchronous mode)............................................................................................. 495 table 15.7 maximum bit rates with external clock input (asynchronous mode)................. 496 table 15.8 maximum bit rates with external clock input (synchronous mode)................... 496 table 15.9 serial mode register settings and sci communication formats .......................... 498 table 15.10 scsmr and scscr settings and sci clock source selection............................. 498 table 15.11 serial communication formats (asynchronous mode).......................................... 500 table 15.12 receive error conditions and sci operation ......................................................... 508 table 15.13 sci interrupt sources.............................................................................................. 5 28 table 15.14 scssr status flags and transfer of receive data ................................................ 529 table 16.1 smart card interface pins ....................................................................................... 535 table 16.2 registers............................................................................................................ ...... 535 xxvi table 16.3 register settings for smart card interface ............................................................. 541 table 16.4 relationship of n to cks1 and cks0..................................................................... 543 table 16.5 examples of bit rate b (bits/s) for scbrr settings (n = 0)................................. 543 table 16.6 examples of scbrr settings for bit rate b (bits/s) (n = 0)................................. 543 table 16.7 maximum bit rates for frequencies (smart card interface mode) ....................... 544 table 16.8 register set values and sck pin ........................................................................... 544 table 16.9 smart card mode operating state and interrupt sources....................................... 551 table 17.1 scif pins ............................................................................................................ .... 559 table 17.2 scif registers ....................................................................................................... . 560 table 17.3 scsmr settings...................................................................................................... 5 71 table 17.4 bit rates and scbrr settings................................................................................ 572 table 17.5 maximum bit rates for various frequencies with baud rate generator (asynchronous mode)............................................................................................. 576 table 17.6 maximum bit rates with external clock input (asynchronous mode)................. 577 table 17.7 scsmr settings and scif communication formats ............................................ 581 table 17.8 scscr settings and scif clock source selection................................................ 582 table 17.9 serial communication formats .............................................................................. 582 table 17.10 scif interrupt sources............................................................................................ 59 4 table 18.1 irda pins ............................................................................................................ .... 603 table 18.2 irda registers ....................................................................................................... . 604 table 19.1 list of multiplexed pins.......................................................................................... 609 table 19.2 pin function controller registers ........................................................................... 613 table 20.1 port a register...................................................................................................... .. 631 table 20.2 port a data register (padr) read/write operations ........................................... 632 table 20.3 port b register ...................................................................................................... .. 633 table 20.4 port b data register (pbdr) read/write operations............................................ 634 table 20.5 port c register ...................................................................................................... .. 635 table 20.6 port c data register (pcdr) read/write operations............................................ 636 table 20.7 port d register...................................................................................................... .. 637 table 20.8 port d data register (pddr) read/write operations ........................................... 638 table 20.9 port e register ...................................................................................................... .. 639 table 20.10 port e data register (pedr) read/write operations............................................ 640 table 20.11 port f register ..................................................................................................... ... 641 table 20.12 port f data register (pfdr) read/write operations ............................................ 642 table 20.13 port g register..................................................................................................... ... 643 table 20.14 port g data register (pgdr) read/write operations ........................................... 644 table 20.15 port h register..................................................................................................... ... 645 table 20.16 port h data register (phdr) read/write operations ........................................... 646 table 20.17 port j register ..................................................................................................... .... 647 table 20.18 port j data register (pjdr) read/write operations.............................................. 648 table 20.19 port k register..................................................................................................... ... 649 table 20.20 port k data register (pkdr) read/write operations ........................................... 650 table 20.21 port l register ..................................................................................................... ... 651 xxvii table 20.22 port l data register (pldr) read/write operation.............................................. 652 table 20.23 sc port register.................................................................................................... .. 653 table 20.24 read/write operation of the sc port data register (scpdr) .............................. 654 table 21.1 a/d converter pins ................................................................................................. 65 9 table 21.2 a/d converter registers ......................................................................................... 660 table 21.3 analog input channels and a/d data registers..................................................... 661 table 21.4 a/d conversion time (single mode) ..................................................................... 673 table 21.5 analog input pin ratings ........................................................................................ 677 table 21.6 relationship between access size and read data.................................................. 677 table 22.1 d/a converter pins ................................................................................................. 6 80 table 22.2 d/a converter registers ......................................................................................... 680 table 23.1 h-udi registers..................................................................................................... . 687 table 23.2 h-udi commands .................................................................................................. 688 table 23.3 sh7729r pins and boundary scan register bits ................................................... 689 table 23.4 reset configuration ................................................................................................ 6 96 table 24.1 absolute maximum ratings.................................................................................... 701 table 24.2 dc characteristics .................................................................................................. 703 table 24.3 permissible output current values......................................................................... 706 table 24.4 maximum operating frequencies ........................................................................... 706 table 24.5 clock timing (1) .................................................................................................... . 707 table 24.5 clock timing (2) .................................................................................................... . 709 table 24.5 clock timing (3) .................................................................................................... . 710 table 24.6 control signal timing ............................................................................................ 716 table 24.7 bus timing.......................................................................................................... .... 719 table 24.8 peripheral module signal timing ........................................................................... 752 table 24.9 h-udi-related pin timing..................................................................................... 756 table 24.10 a/d converter characteristics ................................................................................ 760 table 24.11 d/a converter characteristics ................................................................................ 760 table a.1 pin states during resets, power-down states, and bus-released state ................ 761 table a.2 pin specifications................................................................................................... . 765 table a.3 pin states (ordinary memory/little endian).......................................................... 770 table a.4 pin states (ordinary memory/big endian) ............................................................. 772 table a.5 pin states (burst rom/little endian) .................................................................... 774 table a.6 pin states (burst rom/big endian) ....................................................................... 776 table a.7 pin states (synchronous dram/little endian) ..................................................... 778 table a.8 pin states (synchronous dram/big endian) ........................................................ 779 table a.9 pin states (pcmcia/little endian)........................................................................ 780 table a.10 pin states (pcmcia/big endian)........................................................................... 782 table b.1 memory-mapped control registers........................................................................ 784 table b.2 register bits ........................................................................................................ .... 790 table c.1 sh7729r models .................................................................................................... 802 xxviii 1 section 1 overview 1.1 features the sh7729r is a single-chip risc microprocessor that integrates a 32-bit risc-type superh risc engine architecture cpu with a digital signal processing (dsp) extension as its core, together with cache memory, an on-chip x/y memory, and a memory management unit (mmu), as well as peripheral functions required for system configuration such as a timer, a realtime clock, an interrupt controller, and a serial communication interface. the sh7729r includes data protection, virtual memory, and other functions provided by incorporating an mmu, into a superh series microprocessor (sh-1 or sh-2). the provision of on-chip dsp functions enables applications that previously required the use of two chipsa microprocessor and a dspto be implemented with a single chip. the sh7729r chip has the same peripheral modules as the sh7729. high-speed data transfers can be formed by an on-chip direct memory access controller (dmac), and an external memory access support function enables direct connection to different kinds of memory. the sh7729r microprocessor also supports an infrared communication function, an a/d converter, and a d/a converter. a powerful built-in power management function keeps power consumption low, even during high- speed operation. the sh7729r can run at six times the system bus operating speed, making it also ideal for devices such as pdas that require both high speed and low power consumption. the features of the sh7729r are listed in table 1.1. 2 table 1.1 sh7729r features item features cpu ? original hitachi superh architecture ? compatible with sh-1, sh-2 and sh-3 (sh7708) at object code level ? 32-bit internal data bus ? general-registers ? sixteen 32-bit general registers (eight 32-bit shadow registers) ? eight 32-bit control registers ? four 32-bit system registers ? risc-type instruction set ? instruction length: 16-bit fixed length for improved code efficiency ? load/store architecture ? delayed branch instructions ? instruction set based on c language ? instruction execution time: one instruction/cycle for basic instructions ? logical address space: 4 gbytes ? space identifier asid: 8 bits, 256 logical address spaces ? five-stage pipeline dsp ? mixture of 16-bit and 32-bit instructions ? 32-/40-bit internal data paths ? multiplier, alu, barrel shifter, and dsp register ? 16-bit 16-bit 32-bit one cycle multiplier ? large dsp data registers ? six 32-bit data registers ? two 40-bit data registers ? extended harvard architecture for dsp data bus ? two data buses ? one instruction bus ? max. four parallel operations: alu, multiply, and two load or store ? two addressing units to generate addresses for two memory access ? dsp data addressing modes: increment, indexing (with or without modulo addressing) ? zero-overhead repeat loop control ? conditional execution instructions ? user dsp mode and privileged dsp mode 3 table 1.1 sh7729r features (cont) item features clock pulse generator (cpg) ? clock mode: input clock can be selected from external input (extal or ckio) or crystal oscillator ? three types of clocks generated: ? cpu clock: 1C24 times the input clock, maximum 200 mhz ? bus clock: 1C4 times the input clock, maximum 66.67 mhz ? peripheral clock: 1/4C4 times the input clock, maximum 33.34 mhz ? power-down modes: ? sleep mode ? standby mode ? module standby mode ? one-channel watchdog timer memory management unit (mmu) ? 4 gbytes of address space, 256 address spaces (8-bit asid) ? page unit sharing ? supports multiple page sizes: 1 kbyte or 4 kbytes ? 128-entry, 4-way set associative tlb ? supports software selection of replacement method and random-replacement algorithms ? contents of tlb are directly accessible by address mapping cache memory ? 16-kbyte cache, mixed instruction/data ? 256 entries, 4-way set associative, 16-byte block length ? write-back, write-through, lru replacement algorithm ? 1-stage write-back buffer ? maximum 2 ways of the cache can be locked x/y memory ? user-selectable mapping mechanism ? fixed mapping for mission-critical realtime applications ? automatic mapping through tlb for ease of use ? three independent read/write ports ? 8-/16-/32-bit access from the cpu ? maximum two 16-bit accesses from the dsp ? 8-/16-/32-bit access from the dmac ? 8-kbyte ram each for x and y memory interrupt controller (intc) ? 23 external interrupt pins (nmi, irq5Cirq0, pint15Cpint0) ? on-chip peripheral interrupts: priority level set for each module 4 table 1.1 sh7729r features (cont) item features user break controller (ubc) ? two break channels ? addresses, data values, type of access, and data size can all be set as break conditions ? supports a sequential break function bus state controller (bsc) ? physical address space divided into six areas (area 0, areas 2 to 6), each a maximum of 64 mbytes, with the following features settable for each area: ? bus size (8, 16, or 32 bits) ? number of wait cycles (also supports a hardware wait function) ? specifying the memory to be connected to each area enables direct connection to sram, dram, synchronous dram, and burst rom ? supports pcmcia interface (2 channels) ? outputs chip select signal (cs0, cs2Ccs6) for corresponding area ? synchronous dram refresh function ? programmable refresh interval ? supports self-refresh mode ? synchronous dram burst access function ? big or little endian can be set user debugging interface (h-udi) ? e10a emulator support ? jtag-standard pin assignment ? realtime branch trace ? 1-kbyte on-chip ram for fast emulation program execution timer (tmu) ? 3-channel auto-reload-type 32-bit timer ? input capture function ? selection of six counter input clocks ? maximum resolution: 2 mhz realtime clock (rtc) ? built-in clock, calendar functions, and alarm functions ? on-chip 32-khz crystal oscillator circuit with a maximum resolution (cycle interrupt) of 1/256 second serial communi- cation interface 0 (sci0) ? asynchronous mode or clock synchronous mode can be selected ? full-duplex communication ? supports smart card interface serial communi- cation interface 1 (sci1) ? 16-byte fifo for transmission/reception ? dma transfer capability ? irda: interface based on 1.0 5 table 1.1 sh7729r features (cont) item features serial communi- cation interface 2 (sci2) ? 16-byte fifo for transmission/reception ? dma transfer capability ? hardware flow control direct memory access controller (dmac) ? four channels ? burst mode and cycle-steal mode i/o port ? twelve 8-bit ports a/d converter (adc) ? 10 bits 4 lsb, 8 channels ? conversion time: 10 s ? input range: 0Cvcc (max. 3.6 v) d/a converter (dac) ? 8 bits 4 lsb, 2 channels ? conversion time: 16 s ? output range: 0Cvcc (max. 3.6 v) product lineup power supply voltage operating abbr. i/o internal frequency model name package sh7729r 3.3 0.3 v 2.0 0.15 v* 200 mhz HD6417729Rhf200 208-pin plastic hqfp (fp-208e) 1.9 0.15 v 167 mhz HD6417729Rf167 208-pin plastic lqfp (fp-208c) HD6417729Rbp167v 240-pin csp (bp-240a) 1.8 + 0.25 v 1.8 - 0.15v 133 mhz HD6417729Rf133 208-pin plastic lqfp (fp-208c) HD6417729Rbp133v 240-pin csp (bp-240a) 1.7 + 0.25 v 1.7 - 0.15v 100 mhz HD6417729Rf100 208-pin plastic lqfp (fp-208c) HD6417729Rbp100v 240-pin csp (bp-240a) note: * 2.0 + 0.15 v, - 0.1 v when using irl and irls interrupts. 6 table 1.2 characteristics item characteristics power supply voltage ? i/o: 3.3 0.3 v, internal: 2.0 0.15 v (200 mhz)*, 1.9 0.15 v (167 mhz models), 1.8 0.15 v, - 0.15 v (133mhz), 1.7 + 0.25 v, - 0.25 v (100mhz) operating frequency ? internal frequency: 200 mhz (200 mhz models), 167 mhz (167 mhz models), 133.34 mhz (133 mhz models), 100 mhz (100 mhz models) ? external frequency: maximum 66.67 mhz process ? 0.25- m cmos/5-layer metal note: * 2.0 + 0.15 v, - 0.1 v when using irl and irls interrupts. 7 1.2 block diagram xycnt xymem mmu tlb sh3 cpu dsp ubc sci tmu rtc irda scif adc dac aud dmac cmt bridge i/o port external bus interface bsc ccn cache aseram h-udi intc cpg/wdt peripheral bus 1 peripheral bus 2 i bus 1 i bus 2 l bus y bus x bus legend: adc: aseram: aud: bsc: cache: ccn: cmt: cpg/wdt: cpu: dac: dmac: dsp: h-udi: intc: a/d converter ase memory advanced user debugger bus state controller cache memory cache memory controller compare match timer clock pulse generator/watchdog timer central processing unit d/a converter direct memory access controller digital signal processor hitachi user debugging interface interrupt controller irda: mmu: rtc: sci: scif: tlb: tmu: ubc: xycnt: xymem: serial communicatiion interface (with irda) memory management unit realtime clock serial communication interface (with smart card interface) / ),)" translation look-aside buffer timer unit user break controller x/y memory controller x/y memory figure 1.1 block diagram 8 1.3 pin description 1.3.1 pin assignment figure 1.2 shows the pin arrangement of the sh7729r. 156 155 154 153 152 151 150 149 148 147 146 145 144 143 142 141 140 139 138 137 136 135 134 133 132 131 130 129 128 127 126 125 124 123 122 121 120 119 118 117 116 115 114 113 112 111 110 109 108 107 106 105 extal xtal v cc v ss v ss audck/pth[6] v cc -pll2 cap2 v ss -pll2 v ss -pll1 cap1 v cc -pll1 md0 irls0 /ptf[0]/pint[8] irls1 /ptf[1]/pint[9] irls2 /ptf[2]/pint[10] irls3 /ptf[3]/pint[11] 7<17)k3l1,b7kl tdi/ptf[5]/pint[13] tms/ptf[6]/pint[14] trst /ptf[7]/pint[15] 42474k l17&k l v cc audata[1]/ptg[1] v ss audata[2]/ptg[2] audata[3]/ptg[3] ckio2/ptg[4] asebrkak /ptg[5] asemd0 /ptg[6] iois16 /ptg[7] adtrg /pth[5] resetm wait breq back tdo/pte[0] pte[1] ras3u /pte[2] pte[3] pte[6] dack1/ptd[7] dack0/ptd[5] ptj[5] ptj[4] v cc q casu /ptj[3] v ss q casl /ptj[2] ptj[1] ras3l /ptj[0] cke/ptk[5] 104 103 102 101 100 99 98 97 96 95 94 93 92 91 90 89 88 87 86 85 84 83 82 81 80 79 78 77 76 75 74 73 72 71 70 69 68 67 66 65 64 63 62 61 60 59 58 57 56 55 54 53 ce2b /pte[5] ce2a /pte[4] cs6 / ce1b cs5 / ce1a /ptk[3] cs4 /ptk[2] cs3 /ptk[1] cs2 /ptk[0] v cc q cs0 / mcs [0] v ss q audsync /pte[7] rd/ wr we3 /dqmuu/ iciowr /ptk[7] we2 /dqmul/ iciord /ptk[6] we1 /domlu/ we we0 /dqmll rd bs /ptk[4] a25 v cc q a24 v ss q a23 v cc a22 v ss a21 a20 a19 a18 a17 a16 a15 v cc q a14 v ss q a13 a12 a11 a10 a9 a8 a7 a6 a5 v cc q a4 v ss q a3 a2 a1 a0 sh7729r fp-208c fp-208e (top view) index md1 md2 v cc -rtc xtal2 extal2 v ss -rtc nmi irq0/ irl0 /pth[0] irq1/ irl1 /pth[1] irq2/ irl2 /pth[2] irq3/ irl3 /pth[3] irq4/pth[4] d31/ptb[7] d30/ptb[6] d29/ptb[5] d28/ptb[4] d27/ptb[3] d26/ptb[2] v ss q d25/ptb[1] v cc q d24/ptb[0] d23/pta[7] d22/pta[6] d21/pta[5] d20/pta[4] v ss d19/pta[3] v cc d18/pta[2] d17/pta[1] d16/pta[0] v ss q d15 v cc q d14 d13 d12 d11 d10 d9 d8 d7 d6 v ss q d5 v cc q d4 d3 d2 d1 d0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 status0/ptj[6] status1/ptj[7] tclk/pth[7] irqout v ss q ckio v cc q txd0/scpt[0] sck0/scpt[1] txd1/scpt[2] sck1/scpt[3] txd2/scpt[4] sck2/scpt[5] rts2 /scpt[6] rxd0/scpt[0] rxd1/scpt[2] v ss rxd2/scpt[4] v cc cts2 /irq5/scpt[7] mcs[7] /ptc[7]/pint[7] mcs[6] /ptc[6]/pint[6] mcs[5] /ptc[5]/pint[5] mcs[4] /ptc[4]/pint[4] v ss q wakeup /ptd[3] v cc q resetout /ptd[2] mcs[3] /ptc[3]/pint[3] mcs[2] /ptc[2]/pint[2] mcs[1] /ptc[1]/pint[1] mcs[0] /ptc[0]/pint[0] drak0/ptd[1] drak1/ptd[0] dreq0 /ptd[4] dreq1 /ptd[6] resetp ca md3 md4 md5 av ss an[0]/ptl[0] an[1]/ptl[1] an[2]/ptl[2] an[3]/ptl[3] an[4]/ptl[4] an[5]/ptl[5] av cc an[6]/da[1]/ptl[6] an[7]/da[0]/ptl[7] av ss 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 figure 1.2 pin assignment (fp-208c, fp-208e) 9 17 16 15 19 18 14 13 12 11 10 9 8 7 6 5 4 3 2 1 17 16 15 19 18 14 13 12 11 10 9 8 7 6 5 4 3 2 1 sh7729r bp-240a (top view) abcdefghj klmnprtuvw abcdefghj klmnprtuvw note: the area within dotted lines shows a cutaway view of the pins. figure 1.3 pin assignment (bp-240a) 10 1.3.2 pin function table 1.3 shows the pin functions. table 1.3 sh7729r pin functions pin no. fp-208c, fp-208e bp-240a pin name i/o description 1 d2 md1 i clock mode setting 2 c2 md2 i clock mode setting 3 e2 vcc-rtc * 1 rtc power supply v * 4 4 d1 xtal2 o on-chip rtc crystal oscillator pin 5 d3 extal2 i on-chip rtc crystal oscillator pin 6 e1 vss-rtc * 1 rtc power supply (0 v) 7 c3 nmi i nonmaskable interrupt request 8 e3 irq0/ irl0 /pth[0] i external interrupt request/input port h 9 e4 irq1/ irl1 /pth[1] i external interrupt request/input port h 10f1 irq2/ irl2 /pth[2] i external interrupt request/input port h 11 f2 irq3/ irl3 /pth[3] i external interrupt request/input port h 12 f3 irq4/pth[4] i external interrupt request/input port h 13 f4 d31/ptb[7] i/o data bus / i/o port b 14 g1 d30/ptb[6] i/o data bus / i/o port b 15 g2 d29/ptb[5] i/o data bus / i/o port b 16 g3 d28/ptb[4] i/o data bus / i/o port b 17 g4 d27/ptb[3] i/o data bus / i/o port b 18 h1 d26/ptb[2] i/o data bus / i/o port b 19 h2 vssq input/output power supply (0 v) 20h3 d25/ptb[1] i/o data bus / i/o port b 21 h4 vccq input/output power supply (3.3 v) 22 j1 d24/ptb[0] i/o data bus / i/o port b 23 j2 d23/pta[7] i/o data bus / i/o port a 24 j4 d22/pta[6] i/o data bus / i/o port a 11 table 1.3 sh7729r pin functions (cont) pin no. fp-208c, fp-208e bp-240a pin name i/o description 25 j3 d21/pta[5] i/o data bus / i/o port a 26 k2 d20/pta[4] i/o data bus / i/o port a 27 k3 vss power supply (0 v) k4 vss power supply (0 v) 28 k1 d19/pta[3] i/o data bus / i/o port a 29 l3 vcc power supply * 4 l4 vcc power supply * 4 30l2 d18/pta[2] i/o data bus / i/o port a 31 l1 d17/pta[1] i/o data bus / i/o port a 32 m4 d16/pta[0] i/o data bus / i/o port a 33 m3 vssq input/output power supply (0 v) 34 m2 d15 i/o data bus 35 m1 vccq input/output power supply (3.3 v) 36 n4 d14 i/o data bus 37 n3 d13 i/o data bus 38 n2 d12 i/o data bus 39 n1 d11 i/o data bus 40p4 d10 i/o data bus 41 p3 d9 i/o data bus 42 p2 d8 i/o data bus 43 p1 d7 i/o data bus 44 r4 d6 i/o data bus 45 r3 vssq input/output power supply (0 v) 46 t4 d5 i/o data bus 47 r1 vccq input/output power supply (3.3 v) 48 t3 d4 i/o data bus 49 t1 d3 i/o data bus 50r2 d2 i/o data bus 51 u2 d1 i/o data bus 52 t2 d0i/o data bus 53 v4 a0o address bus 12 table 1.3 sh7729r pin functions (cont) pin no. fp-208c, fp-208e bp-240a pin name i/o description 54 v3 a1 o address bus 55 v5 a2 o address bus 56 w4 a3 o address bus 57 u4 vssq input/output power supply (0 v) 58 w5 a4 o address bus 59 u3 vccq input/output power supply (3.3 v) 60u5 a5 o address bus 61 t5 a6 o address bus 62 w6 a7 o address bus 63 v6 a8 o address bus 64 u6 a9 o address bus 65 t6 a10o address bus 66 w7 a11 o address bus 67 v7 a12 o address bus 68 u7 a13 o address bus 69 t7 vssq input/output power supply (0 v) 70w8 a14 o address bus 71 v8 vccq input/output power supply (3.3 v) 72 u8 a15 o address bus 73 t8 a16 o address bus 74 w9 a17 o address bus 75 v9 a18 o address bus 76 t9 a19 o address bus 77 u9 a20o address bus 78 v10a21 o address bus 79 u10vss power supply (0 v) t10vss power supply (0 v) 80w10a22 o address bus 81 u11 vcc power supply * 4 t11 vcc power supply * 4 82 v11 a23 o address bus 13 table 1.3 sh7729r pin functions (cont) pin no. fp-208c, fp-208e bp-240a pin name i/o description 83 w11 vssq input/output power supply (0 v) 84 t12 a24 o address bus 85 u12 vccq input/output power supply (3.3 v) 86 v12 a25 o address bus 87 w12 bs /ptk[4] o / i/o bus cycle start signal / i/o port k 88 t13 rd o read strobe 89 u13 we0 /dqmll o d7Cd0 select signal / dqm (sdram) 90v13 we1 /dqmlu/ we o d15Cd8 select signal / dqm (sdram) 91 w13 we2 /dqmul/ iciord / ptk[6] o / i/o d23Cd16 select signal / dqm (sdram) / pcmcia i/o read / i/o port k 92 t14 we3 /dqmuu/ iciowr / ptk[7] o / i/o d31Cd24 select signal / dqm (sdram) / pcmcia i/o write / i/o port k 93 u14 rd/ wr o read/write 94 v14 audsync /pte[7] o / i/o aud synchronous / i/o port e 95 w14 vssq input/output power supply (0 v) 96 t15 cs0 / mcs[0] o chip select 0/mask rom chip select 0 97 u15 vccq input/output power supply (3.3 v) 98 t16 cs2 /ptk[0] o / i/o chip select 2 / i/o port k 99 w15 cs3 /ptk[1] o / i/o chip select 3 / i/o port k 100 u16 cs4 /ptk[2] o / i/o chip select 4 / i/o port k 101 w16 cs5 / ce1a /ptk[3] o / i/o chip select 5/ce1 (area 5 pcmcia) / i/o port k 102 v15 cs6 / ce1b o chip select 6/ce1 (area 6 pcmcia) 103 v17 ce2a /pte[4] o / i/o area 5 pcmcia card enable / i/o port e 104 v16 ce2b /pte[5] o / i/o area 6 pcmcia card enable / i/o port e 105 t18 cke/ptk[5] o / i/o ck enable (sdram) / i/o port k 14 table 1.3 sh7729r pin functions (cont) pin no. fp-208c, fp-208e bp-240a pin name i/o description 106 u18 ras3l /ptj[0] o / i/o lower 32/64 mb address (sdram) ras / i/o port j 107 u19 ptj[1] i/o i/o port j * 6 108 r18 casl /ptj[2] o / i/o lower 32/64 mb address (sdram) cas / i/o port j 109 t19 vssq input/output power supply (0 v) 110t17 casu /ptj[3] o / i/o upper 32 mb address (sdram) cas / i/o port j 111 r19 vccq input/output power supply (3.3 v) 112 u17 ptj[4] i/o i/o port j 113 r17 ptj[5] i/o i/o port j 114 r16 dack0/ptd[5] o / i/o dma acknowledge 0 / i/o port d 115 p19 dack1/ptd[7] o / i/o dma acknowledge 1 / i/o port d 116 p18 pte[6] i/o i/o port e 117 p17 pte[3] i/o i/o port e 118 p16 ras3u /pte[2] o / i/o upper 32 mb address (sdram) ras / i/o port e 119 n19 pte[1] i/o i/o port e 120n18 tdo/pte[0] o / i/o test data output / i/o port e 121 n17 back o bus acknowledge 122 n16 breq i bus request 123 m19 wait i hardware wait request 124 m18 resetm i manual reset request 125 m17 adtrg /pth[5] i analog trigger / input port h 126 m16 iois16 /ptg[7] i area 6 16-bit input/output / input port g 127 l19 asemd0 /ptg[6] i ase mode* 5 / input port g 128 l18 asebrkak /ptg[5] o/i ase break acknowledge / input port g 129 l16 ptg[4]/ckio2 o/i input port g / clock output 130l17 audata[3]/ptg[3] o/i aud data / input port g 131 k18 audata[2]/ptg[2] o/i aud data / input port g 15 table 1.3 sh7729r pin functions (cont) pin no. fp-208c, fp-208e bp-240a pin name i/o description 132 k17 vss power supply (0 v) k16 vss power supply (0 v) 133 k19 audata[1]/ptg[1] o/i aud data / input port g 134 j17 vcc power supply * 4 j16 vcc power supply * 4 135 j18 audata[0]/ptg[0] o/i aud data / input port g 136 j19 trst /ptf[7]/pint[15] i test reset / input port f / port interrupt 137 h16 tms/ptf[6]/pint[14] i test mode switch / input port f / port interrupt 138 h17 tdi/ptf[5]/pint[13] i test data input / input port f / port interrupt 139 h18 tck/ptf[4]/pint[12] i test clock / input port f / port interrupt 140h19 irls 3/ptf[3]/ pint[11] i external interrupt request / input port f / port interrupt 141 g16 irls 2/ptf[2]/ pint[10] i external interrupt request / input port f / port interrupt 142 g17 irls 1/ptf[1]/pint[9] i external interrupt request / input port f / port interrupt 143 g18 irls 0/ptf[0]/pint[8] i external interrupt request / input port f / port interrupt 144 g19 md0i clock mode setting 145 f16 vcc-pll1 * 2 pll1 power supply * 4 146 f17 cap1 pll1 external capacitance pin 147 f18 vss-pll1 * 2 pll1 power supply (0 v) 148 f19 vss-pll2 * 2 pll2 power supply (0 v) 149 e16 cap2 pll2 external capacitance pin 150e17 vcc-pll2 * 2 pll2 power supply * 4 151 d16 audck/pth[6] i aud clock / input port h 152 e19 vss power supply (0 v) 153 d17 vss power supply (0 v) d19 vss power supply (0 v) 16 table 1.3 sh7729r pin functions (cont) pin no. fp-208c, fp-208e bp-240a pin name i/o description 154 e18 vcc power supply * 4 c19 vcc power supply * 4 155 c18 xtal o clock oscillator pin 156 d18 extal i external clock / crystal oscillator pin 157 b16 status0/ptj[6] o / i/o processor status / i/o port j 158 b17 status1/ptj[7] o / i/o processor status / i/o port j 159 b15 tclk/pth[7] i/o tmu or rtc clock input/output / i/o port h 160a16 irqout o interrupt request notification 161 c16 vssq input/output power supply (0 v) 162 a15 ckio i/o system clock input/output 163 c17 vccq input/output power supply (3.3 v) 164 c15 txd0/scpt[0] o transmit data 0 / sci output port 165 d15 sck0/scpt[1] i/o serial clock 0 / sci i/o port 166 a14 txd1/scpt[2] o transmit data 1 / sci output port 167 b14 sck1/scpt[3] i/o serial clock 1 / sci i/o port 168 c14 txd2/scpt[4] o transmit data 2 / sci output port 169 d14 sck2/scpt[5] i/o serial clock 2 / sci i/o port 170a13 rts2 /scpt[6] o / i/o transmit request 2 / sci i/o port 171 b13 rxd0/scpt[0] i transmit data 0 / sci input port 172 c13 rxd1/scpt[2] i transmit data 1 / sci input port 173 d13 vss power supply (0 v) a12 vss power supply (0 v) 174 b12 rxd2/scpt[4] i transmit data 2 / sci input port 175 c12 vcc power supply * 4 d12 vcc power supply * 4 176 a11 cts2 /irq5/scpt[7] i transmit clear 2 / external interrupt request / sci input port 177 b11 mcs[7] /ptc[7]/pint[7] o / i/o / i mask rom chip select / i/o port c / port interrupt 17 table 1.3 sh7729r pin functions (cont) pin no. fp-208c, fp-208e bp-240a pin name i/o description 178 d11 mcs[6] /ptc[6]/pint[6] o / i/o / i mask rom chip select / i/o port c / port interrupt 179 c11 mcs[5] /ptc[5]/pint[5] o / i/o / i mask rom chip select / i/o port c / port interrupt 180b10 mcs[4] /ptc[4]/pint[4] o / i/o / i mask rom chip select / i/o port c / port interrupt 181 c10vssq input/output power supply (0 v) 182 d10 wakeup /ptd[3] o / i/o standby mode interrupt request notification / i/o port d 183 a10vccq input/output power supply (3.3 v) 184 c9 resetout /ptd[2] o / i/o reset output / i/o port d 185 d9 mcs[3] /ptc[3]/pint[3] o / i/o / i mask rom chip select / i/o port c / port interrupt 186 b9 mcs[2] /ptc[2]/pint[2] o / i/o / i mask rom chip select / i/o port c / port interrupt 187 a9 mcs[1] /ptc[1]/pint[1] o / i/o / i mask rom chip select / i/o port c / port interrupt 188 d8 mcs[0] /ptc[0]/pint[0] o / i/o / i mask rom chip select / i/o port c / port interrupt 189 c8 drak0/ptd[1] o / i/o dma request acceptance / i/o port d 190b8 drak1/ptd[0] o / i/o dma request acceptance / i/o port d 191 a8 dreq0 /ptd[4] i dma request / input port d 192 d7 dreq1 /ptd[6] i dma request / input port d 193 c7 resetp i power-on reset request 194 b7 ca i chip activate / hardware standby request 195 a7 md3 i area 0 bus width setting 196 d6 md4 i area 0 bus width setting 197 c6 md5 i endian setting 198 b6 avss analog power supply (0 v) 199 a6 an[0]/ptl[0] i a/d converter input / input port l 200 d5 an[1]/ptl[1] i a/d converter input / input port l 18 table 1.3 sh7729r pin functions (cont) pin no. fp-208c, fp-208e bp-240a pin name i/o description 201 c5 an[2]/ptl[2] i a/d converter input / input port l 202 d4 an[3]/ptl[3] i a/d converter input / input port l 203 a5 an[4]/ptl[4] i a/d converter input / input port l 204 c4 an[5]/ptl[5] i a/d converter input / input port l 205 a4 avcc analog power supply (3.3 v) 206 b5 an[6]/da[1]/ptl[6] i a/d converter input / input port l 207 b3 an[7]/da[0]/ptl[7] i a/d converter input / input port l 208 b4 avss analog power supply (0 v) notes: *1 must be connected to the power supply even when the rtc is not used. *2 must be connected to the power supply even when the on-chip pll circuits are not used (except in hardware standby mode). *3 except in hardware standby mode, all power supply pins must be connected to the system power supply. (supply power constantly.) in hardware standby mode, power must be supplied at least to v cc Crtc and v ss Crtc. if power is not supplied to power supply pins other than v cc Crtc and v ss Crtc, hold the ca pin low. *4 2.0 v in 200 mhz models, 1.9 v in 167 mhz models, 1.8 v in 133 mhz models, 1.7 v in 100 mhz models *5 drive high when using the user system alone, and not using an emulator or the h-udi. *6 b2, b1, c1, u1, v1, w1, v2, w2, w3, w17, w18, w19, v18, v19, b19, a19, b18, a18, a17, a3, a2, and a1 are nc pins. no connection should be made to these pins. 19 section 2 cpu 2.1 registers the sh7729r has the same registers as the sh-3. in addition, the sh7729r also supports the same dsp-related registers as in the sh2-dsp. the basic software-accessible registers are divided into four distinct groups: ? general registers ? control registers ? system registers ? dsp registers with the exception of a number of dsp registers, all of these registers are 32-bit width. the general registers are accessible from user mode, with r0Cr7 banked to provide each processor mode access to a separate set of r0Cr7 registers (i.e. r0Cr7_bank0, and r0Cr7_bank1). in privileged mode, the register bank (rb) bit in the status register (sr) defines which set of banked registers (r0Cr7_bank0 or r0Cr7_bank1) are accessed as general registers, and which are accessed only by ldc/stc instructions. the control registers can be accessed by ldc/stc instructions. the gbr, rs, re, and mod registers can also be accessed in user mode. control registers are: ? sr: status register ? ssr: saved status register ? spc: saved program counter ? gbr: global base register ? vbr: vector base register ? rs: repeat start register (dsp mode only) ? re: repeat end register (dsp mode only) ? mod: modulo register (dsp mode only) the system registers are accessed by the lds/sts instructions (the pc is software-accessible, but is included here because its contents are saved in, and restored from, spc in exception handling). the system registers are: ? mach: multiply and accumulate high register ? macl: multiply and accumulate low register ? pr: procedure register ? pc: program counter 20 this section explains the usage of these registers in different modes. figures 2.1 and 2.2 show the register configuration in each processing mode. switching between user mode and privileged mode is carried out by means of the processing operation mode bit (md) in the status register. the dsp mode is switched by means of the dsp bit in the status register. 21 31 r0_bank0 * 1 , * 2 r1_bank0 * 2 r2_bank0 * 2 r3_bank0 * 2 r4_bank0 * 2 r5_bank0 * 2 r6_bank0 * 2 r7_bank0 * 2 r8 r9 r10 r11 r12 r13 r14 r15 sr gbr mach macl pr pc 0 31 r0_bank1 * 1 , * 3 r1_bank1 * 3 r2_bank1 * 3 r3_bank1 * 3 r4_bank1 * 3 r5_bank1 * 3 r6_bank1 * 3 r7_bank1 * 3 r0_bank0 * 1 , * 4 r1_bank0 * 4 r2_bank0 * 4 r3_bank0 * 4 r4_bank0 * 4 r5_bank0 * 4 r6_bank0 * 4 r7_bank0 * 4 r8 r9 r10 r11 r12 r13 r14 r15 sr ssr gbr mach macl vbr pr pc spc 0 31 r0_bank0 * 1 , * 4 r1_bank0 * 4 r2_bank0 * 4 r3_bank0 * 4 r4_bank0 * 4 r5_bank0 * 4 r6_bank0 * 4 r7_bank0 * 4 r0_bank1 * 1 , * 3 r1_bank1 * 3 r2_bank1 * 3 r3_bank1 * 3 r4_bank1 * 3 r5_bank1 * 3 r6_bank1 * 3 r7_bank1 * 3 r8 r9 r10 r11 r12 r13 r14 r15 sr ssr gbr mach macl vbr pr pc spc 0 notes: * 1 the r0 register is used as an index register in indexed register indirect addressing mode and indexed gbr indirect addressing mode. * 2 bank register * 3 bank register accessed as a general register when the rb bit is set to 1 in the sr register. accessed only by ldc/stc instructions when the rb bit is cleared to 0. * 4 bank register accessed as a general register when the rb bit is cleared to 0 in the sr register. accessed only by ldc/stc instructions when the rb bit is set to 1. (a) user mode register configuration (b) privileged mode register configuration (rb = 1) (c) privileged mode register configuration (rb = 0) figure 2.1 register configuration in each processing mode (1) 22 39 a0g a1g 32 31 a0 a1 m0 m1 x0 x1 y0 y1 dsr ms me mod 0 (d) dsp mode register configuration (dsp = 1) figure 2.2 register configuration in each processing mode (2) register values after a reset are shown in table 2.1. table 2.1 initial register values type registers initial value * general registers r0 to r15 undefined control registers sr md bit = 1, rb bit = 1, bl bit = 1, i3 to i0 = 1111 (h'f), reserved bits = 0, others undefined gbr, ssr, spc undefined vbr h'00000000 rs, re undefined mod undefined system registers mach, macl, pr undefined pc h'a0000000 dsp registers a0, a0g, a1, a1g, m0, m1, x0, x1, y0, y1 undefined dsr h'00000000 note: * initialized by a power-on or manual reset. 23 2.1.1 general registers there are sixteen 32-bit general registers (rn), designated r0 to r15. the general registers are used for data processing and address calculation. with superh microcomputer type instructions, r0 is used as an index register. with a number of instructions, r0 is the only register that can be used. r15 is used as the stack pointer (sp). in exception handling, r15 is used to reference the stack when saving and restoring the status register (sr) and program counter (pc). with dsp type instructions, eight of the sixteen general registers are used for addressing of x and y data memory and data memory (single data) that uses the l-bus. to access x memory, r4 and r5 are used as the x address register [ax] and r8 is used as the x index register [ix]. to access y memory, r6 and r7 are used as the y address register [ay] and r9 is used as the y index register [iy]. to access single data that uses the l-bus, r2, r3, r4, and r5 are used as the single data address register [as] and r8 is used as the single data index register [is]. figure 2.3 shows the general registers, which are identical to those of the sh3, when dsp extension is disabled. 31 r0 * 1, * 2 r1 * 2 r2 * 2 r3 * 2 r4 * 2 r5 * 2 r6 * 2 r7 * 2 r8 r9 r10 r11 r12 r13 r14 r15 0 general registers (when not in dsp mode) notes: * 1 r0 functions as an index register in the indexed register-indirect addressing mode and indexed gbr-indirect addressing mode. in some instructions, only r0 can be used as the source register or destination register. * 2 r0Cr7 are banked registers. in user mode, bank0 is used. in privileged mode, sr.rb specifies bank. sr.rb = 0; bank0 is used sr.rb = 1; bank1 is used figure 2.3 general purpose registers (not in dsp mode) 24 on the other hand, registers r2Cr9 are also used for dsp data address calculation when dsp extension is enabled (see figure 2.4). other symbols that represent the purpose of the registers in dsp type instructions is shown in [ ]. 31 r0 r1 r2 [as] r3 [as] r4 [as, ax] r5 [as, ax] r6 [ay] r7 [ay] r8 [ix, is] r9 [iy] r10 r11 r12 r13 r14 r15 0 general registers (dsp mode enabled) x or y data transfer operation r4, 5 [ax]: address register set for x data memory. r8 [x]: index register for address register set ax. r6, 7 [ay]: address register set for y data memory. r9 [iy]: index register for address register set ay. single data transfer operation r2C5 [as]: address register set for memory. r8 [is]: index register for address register set as. figure 2.4 general purpose registers (dsp mode) dsp type instructions can access x and y data memory simultaneously. to specify addresses for x and y data memory, two address pointer sets are provided. these are: r8[ix], r4,5[ax] for x memory access, and r9[iy], r6,7[ay] for y memory access. the symbols r2Cr9 are used by the assembler, but users can use other register names (aliases) that indicate the purpose of the register in the dsp instruction. the coding in assembler is as follows. ix: .reg (r8) 25 the name ix is the alias for r8. other aliases are as follows. ax0: .reg (r4) ax1: .reg (r5) ix: .reg (r8) ay0: .reg (r6) ay1: .reg (r7) iy: .reg (r9) as0: .reg (r4) ; this is optional, if another alias is required for single data transfer. as1: .reg (r5) ; this is optional, if another alias is required for single data transfer. as2: .reg (r2) as3: .reg (r3) is: .reg (r8) ; this is optional, if another alias is required for single data transfer. 2.1.2 control registers the sh7729r has 8 control registers: sr, ssr, spc, gbr, vbr, rs, re, and mod (figure 2.5). ssr, spc, gbr and vbr are the same as the sh-3 registers. the dsp mode is activated only when sr.dsp = 1. repeat start register rs, repeat end register re, and repeat counter rc (12-bit part of sr) and repeat control bits rf0 and rf1 are new registers and control bits which are used for repeat control. modulo register mod and modulo control bits dmx and dmy in sr are also new register and control bits. in sr, there are six additional control bits: rc[11:0], rf0, rf1, dmx, dmy and dsp. bits dmx, dmy, rc[11:0], and rf[1:0] can be modified in supervisor mode, supervisor dsp mode, and use dsp mode. dmx and dmy are used for modulo addressing control. if dmx is 1, the modulo addressing mode is effective for the x memory address pointer, ax (r4 or r5). if dmy is 1, the modulo addressing mode is effective for the y memory address pointer, ay (r6 or r7). however, both x and y address pointers cannot be operated in modulo addressing mode even though both dmx and dmy bits are set. the case where dmx = dmy = 1 is reserved for future expansion. if both dmx and dmy are set simultaneously, the hardware will provisionally treat only the y address pointer as the modulo addressing mode pointer. modulo addressing is available for x and y data transfer operations (movx and movy), but not for a single data transfer operation (movs). rf1 and rf0 hold information on the number of repeat steps, and are set when a setrc instruction is executed. when rf[1:0] = 00, the current repeat module consists of one instruction step. rf[1:0] = 01 means two instruction steps, rf[1:0] = 11 means three instruction steps, and rf[1:0] = 10 means the current repeat module consists of four or more instructions. although rc[11:0] and rf[1:0] can be changed by a store/load to sr, use of the dedicated manipulation instruction setrc is recommended. 26 sr also has a 12-bit repeat counter, rc, which is used for efficient loop control. the repeat start register (rs) and repeat end register (re) are also provided for loop control. they hold the start and end addresses of a loop (the contents of the rs and re registers are slightly different from the actual loop start and end addresses). the modulo register, mod, is provided to implement modulo addressing for circular data buffering. mod holds the modulo start address (ms) and modulo end address (me). in order to access rs, re and mod, load/store (control register) instructions for these registers are provided. an example for rs is as follows: ldc rm,rs; rm -> rs ldc.l @rm+,rs; (rm) -> rs, rm+4 -> rm stc rs,rn; rs -> rn stc.l rs,@-rn; rn-4 -> rn, rs -> (rn) address set instructions for rs and re are also provided. ldrs @(disp,pc); disp 2 + pc -> rs ldre @(disp,pc); disp 2 + pc -> re 27 31 0 md rc 0-0 dsp dmy dmx m q i3 i2 i1 i0 rf1 rf0 st rb bl 28 27 16 15 13 12 11 10 9 8 7 6 5 4 3 2 1 0 sr (status register) md bit: processor operation mode md = 1: privileged mode md = 0: user mode rb bit: register bank bit; used to define the general registers in privileged mode. rb = 1: r0_bank1 to r7_bank1 are used as general registers. r0_bank0 to r7_bank0 accessed by ldc/stc instructions. rb = 0: r0_bank0 to r7_bank0 are used as general registers. r0_bank1 to r7_bank1 accessed by ldc/stc instructions. bl bit: block bit; used to mask exception in privileged mode. bl = 1: interrupts are masked (not accepted) bl = 0: interrupts are accepted rc [11:0]: 12-bit repeat counter dsp bit: dsp operation mode dsp = 1: dsp instructions (lds rm, dsr/a0/x0/x1/y0/y1, lds.l @rm+, dsr/a0/x0/x1/y0/y1, sts dsr/a0/x0/x1/y0/y1, rn, sts.l dsr/a0/x0/x1/y0/y1, @Crn, ldc rm, rs/re/mod, ldc.l @rm+, rs/re/mod, stc rs/re/mod,rn, stc.l rs/re/mod, @Crn, ldrs, ldre, setrc, movs, movx, movy, pxxx) are enabled. dsp = 0: all dsp instructions are treated as illegal instructions; only sh3 instructions are supported. dmy bit: modulo addressing enable for y side dmx bit: modulo addressing enable for x side q, m bit: used by div0u/s and div1 instructions. i [3:0]: 4-bit field indicating the interrupt request mask level. rf [1:0]: used for repeat control s bit: used by the mac instructions and dsp data. t bit: the movt, cmp/cond, tas.tst, bt, bf, sett, clrt and dt instructions use the t bit to indicate true (logic one) or false (logic zero). the addv/c, subv/c, div0u/s, div1, negc, shar/l, shlr/l, rotr/l and rotcr/l instructions also use the t bit to indicate a carry, borrow, overflow, or underflow. reserved bits: always read as 0, and should always be written with 0. figure 2.5 control registers 28 ssr 31 0 saved status register (ssr) spc 31 0 saved program counter (spc) gbr 31 0 global base register vbr 31 0 vector base register rs 31 0 repeat start register re 31 0 repeat end register me ms 31 16 15 0 modulo register mod me: modulo end address, ms: modulo start address saved status register (ssr) stores current sr value at time of exception to indicate processor status when returning to instruction stream from exception handler. saved program counter (spc) stores current pc value at time of exception to indicate return address on completion of exception handling. global base register (gbr) stores base address of gbr-indirect addressing mode. the gbr-indirect addressing mode is used for data transfer and logical operations on the on-chip peripheral module register area. vector base register (vbr) stores base address of exception vector area. repeat start register (rs) used in dsp mode only. indicates start address of repeat loop. repeat end register (re) used in dsp mode only. indicates address of repeat loop end. modulo register used in dsp mode only. md[31:16]: me: modulo end address, md[15:0]: modulo start address. in x/y operand address generation, the cpu compares the address with me, and if it is the same, loads ms in either the x or y operand address register (depending on bits dmx and dmy in the sr register). figure 2.5 control registers (cont) 29 details of the status register (sr) when stc/ldc instructions are used are shown below. 1. when the dsp is not operating, operation is the same as for the sh-3. 2. in supervisor dsp mode, operation is the same as in supervisor mode. 3. in user dsp mode, sr can be read with an stc instruction. 4. in user dsp mode, an ldc instruction can be issued for sr, but in this case dsp-related bits are not write-protected. table 2.2 operation of sr bits in each sh-3 dsp mode supervisor mode user mode supervisor dsp mode user dsp mode access to dsp-related bit with field md = 1 & dsp = 0 md = 0 & dsp = 0 md = 1 & dsp = 1 md = 0 & dsp = 1 dedicated instruction initial value after reset md s: ok, l: ok s, l: invalid instruction s: ok, l: ok s: ok, l: ng 1 rb s: ok, l: ok s, l: invalid instruction s: ok, l: ok s: ok, l: ng 1 bl s: ok, l: ok s, l: invalid instruction s: ok, l: ok s: ok, l: ng 1 rc [11:0] s: ok, l: ok s, l: invalid instruction s: ok, l: ok r: ok, l: ok setrc instruction 0b000000000000 dsp s: ok, l: ok s, l: invalid instruction s: ok, l: ok s: ok, l: ng 0 smx s: ok, l: ok s, l: invalid instruction s: ok, l: ok r: ok, l: ok 0 smy s: ok, l: ok s, l: invalid instruction s: ok, l: ok r: ok, l: ok 0 q s: ok, l: ok s, l: invalid instruction s: ok, l: ok s: ok, l: ng x m s: ok, l: ok s, l: invalid instruction s: ok, l: ok s: ok, l: ng x i[3:0] s: ok, l: ok s, l: invalid instruction s: ok, l: ok s: ok, l: ng 1111 rf[1:0] s: ok, l: ok s, l: invalid instruction s: ok, l: ok r: ok, l: ng setrc instruction x s s: ok, l: ok s, l: invalid instruction s: ok, l: ok s: ok, l: ng x ts: ok, l: ok s, l: invalid instruction s: ok, l: ok s: ok, l: ng x legend s (stc): store sr to rn,sr->rn l (ldc): load rn to sr,rn->sr ok: stc/ldc operation is enabled. invalid instruction: exception occurs when an invalid instruction is executed. ng: previous value is retained. no change. 30 2.1.3 system registers the sh7729r has four system registers, macl, mach, pr and pc (figure 2.6). mach macl 31 0 pr 31 0 pc 31 0 multiply and accumulate high and low registers (mach/l) store the results of multiplicationand accumulation operations. procedure register (pr) stores the subroutine procedure return address. program counter (pc) indicates the start address of the current instruction. figure 2.6 system registers the dsr, a0, x0, x1, y0 and y1 registers are also treated as system registers. therefore, instructions for data transfer between general registers and system registers are supported for these registers. 2.1.4 dsp registers the sh7729r has eight data registers and one control register as dsp registers (figure 2.7). the data registers are 32-bit width with the exception of registers a0 and a1. registers a0 and a1 include 8 guard bits (fields a0g and a1g), giving them a total width of 40 bits. three kinds of operation access the dsp data registers. the first is dsp data processing. when a dsp fixed-point data operation uses a0 or a1 as the source register, it uses the guard bits (bits 39C32). when it uses a0 or a1 as the destination register, guard bits 39C32 are valid. when a dsp fixed-point data operation uses a dsp register other than a0 or a1 as the source register, it sign- extends the source value to bits 39C32. when it uses one of these registers as the destination register, bits 39C32 of the result are discarded. the second kind of operation is an x or y data transfer operation, movx.w movy.w. this operation accesses the x and y memories through the 16-bit x and y data buses (figure 2.8). the register to be loaded or stored by this operation always comprises the upper 16 bits (bits 31C16). x0 or x1 can be the destination of an x memory load and y0 or y1 can be the destination of a y memory load, but no other register can be the destination register in this operation. when data is read into the upper 16 bits of a register (bits 31C16), the lower 16 bits of the register (bits 15C0) are automatically cleared. a0 and a1 can be stored in the x or y memory by this operation, but no other registers can be stored. 31 the third kind of operation is a single-data transfer instruction, movs.w or movs.l. these instructions access any memory location through the ldb (figure 2.8). all dsp registers connect to the ldb and can be the source or destination register of the data transfer. these instructions have word and longword access modes. in word mode, registers to be loaded or stored by this instruction comprise the upper 16 bits (bits 31C16) for dsp registers except a0g and a1g. when data is loaded into a register other than a0g and a1g in word mode, the lower half of the register is cleared. when a0 or a1 is used, the data is sign-extended to bits 39C32 and the lower half is cleared. when a0g or a1g is the destination register in word mode, data is loaded into an 8-bit register, but a0 or a1 is not cleared. in longword mode, when the destination register is a0 or a1, it is sign-extended to bits 39C32. tables 2.3 and 2.4 show the data type of registers used in dsp instructions. some instructions cannot use some registers shown in the tables because of instruction code limitations. for example, pmuls can use a1 as the source register, but cannot use a0. these tables ignore details of register selectability. table 2.3 destination register in dsp instructions guard bits register bits registers instructions 39 32 31 16 15 0 a0, a1 dsp fixed-point, psha, pmuls sign-extended 40-bit result integer, pdmsb sign-extended 24-bit result cleared logical, pshl cleared 16-bit result cleared data transfer movs.w sign-extended 16-bit data cleared movs.l sign-extended 32-bit data a0g, a1g data movs.w data no update transfer movs.l data no update x0, x1 y0, y1 dsp fixed-point, psha, pmuls 32-bit result m0, m1 integer, logical, pdmsb, pshl 16-bit result cleared data movx/y.w, movs.w 16-bit result cleared transfer movs.l 32-bit data 32 table 2.4 source register in dsp operations guard bits register bits registers instructions 39 32 31 16 15 0 a0, a1 dsp fixed-point, pdmsb, psha 40-bit data integer 24-bit data logical, pshl, pmuls 16-bit data data movx/y.w, movs.w 16-bit data transfer movs.l 32-bit data a0g, a1g data movs.w data transfer movs.l data x0, x1 y0, y1 dsp fixed-point, pdmsb, psha sign * 32-bit data m0, m1 integer sign * 16-bit data logical, pshl, pmuls 16-bit data data movs.w 16-bit data transfer movs.l 32-bit data note: * the data is sign-extended and input to the alu. 31 32 39 a0 a0g a1g a1 m0 m1 x0 x1 y0 y1 0 1 2 3 4 5 6 7 dc cs [2:0] v n z gt 8 31 0 (a) dsp data registers (b) dsp status register (dsr) reset status dsr: all zeros others: undefined figure 2.7 dsp registers 33 table 2.5 dsr register bits bits name (abbreviation) function 31C8 reserved bits 0: always read as 0; always use 0 as the write value 7 signed greater than bit (gt) indicates that the operation result is positive (except 0), or that operand 1 is greater than operand 2 1: operation result is positive, or operand 1 is greater than operand 2 6 zero bit (z) indicates that the operation result is zero (0), or that operand 1 is equal to operand 2 1: operation result is zero (0), or operands are equal 5 negative bit (n) indicates that the operation result is negative, or that operand 1 is smaller than operand 2 1: operation result is negative, or operand 1 is smaller than operand 2 4 overflow bit (v) indicates that the operation result has overflowed 1: operation result has overflowed 3C1 condition select bits (cs) designate the mode for selecting the operation result status to be set in the dc bit do not set these bits to 110 or 111 000: carry/borrow mode 001: negative value mode 010: zero mode 011: overflow mode 100: signed greater mode 101: signed greater than or equal to mode 0 dsp condition bit (dc) sets the status of the operation result in the mode designated by the cs bits 0: designated mode status has not occurred (false) 1: designated mode status has occurred 34 a0g 32 0 39 31 16 a0 a1 m0 m1 x0 x1 y0 y1 0 7 a1g dsr 16 bit 16 bit 8 bit 32 bit ldb xdb ydb movx.w movs.w, movs.l movs.w, movs.l movy.w figure 2.8 connections of dsp registers and buses the dsp unit has one control register, the dsp status register (dsr). dsr holds the status of dsp data operation results (zero, negative, and so on) and has a dc bit which is similar to the t bit in the cpu. the dc bit indicates one of the status flags. a dsp data processing instruction controls its execution based on the dc bit. this control affects only the operations in the dsp unit; it controls the update of dsp registers only. it cannot control operations in the cpu, such as address register updating and load/store operations. control bits cs[2:0] specify the condition to be reflected in the dc bit. unconditional dsp type data operations, except pmuls, movx, movy and movs, update the condition flags and dc bit, but no cpu instructions, including mac instructions, update the dc bit. conditional dsp type instructions do not update dsr either. dsr is assigned as a system register and the following load/store instructions are provided: sts dsr,rn; sts.l dsr,@-rn; lds rn,dsr; lds.l @rn+,dsr; when dsr is read by an sts instruction, the upper bits (bits 31 to 8) are all 0. 35 2.2 data formats 2.2.1 register data format (non-dsp type) register operands are always longwords (32 bits) (figure 2.9). when the memory operand is only a byte (8 bits) or a word (16 bits), it is sign-extended into a longword when loaded into a register. 31 0 longword figure 2.9 longword operand 2.2.2 dsp-type data formats the sh7729r has several different data formats that depend on the instruction. this section explains the data formats for dsp type instructions. figure 2.10 shows three dsp-type data formats with different binary point positions. a cpu-type data format with the binary point to the right of bit 0 is also shown for reference. the dsp-type fixed point data format has the binary point between bit 31 and bit 30. the dsp- type integer format has the binary point between bit 16 and bit 15. the dsp-type logical format does not have a binary point. the valid data lengths of the data formats depend on the instruction and the dsp register. 36 39 s 31 30 0 C2 8 to +2 8 C 2 C31 39 s 32 31 0 C2 23 to +2 23 C 1 39 s s 31 30 16 15 16 15 0 C1 to +1 C 2 C15 39 31 16 15 0 s 31 0 C2 15 to +2 15 C 1 16 15 31 22 0 C32 to +32 16 15 s 31 21 0 C16 to +16 16 15 s 31 30 0 C1 to +1 C 2 C31 s 31 0 C2 31 to +2 31 C 1 dsp type fixed point with guard bits without guard bits multiplier input dsp type integer dsp type logical with guard bits cpu type integer s: sign bit longword : binary point : does not affect the operations without guard bits shift amount for arithmetic shift (psha) shift amount for logical shift (pshl) figure 2.10 data formats the shift amount for the arithmetic shift (psha) instruction has a 7-bit field that can represent values from C64 to +63, but C32 to +32 are valid numbers for the instruction. also the shift amount for a logical shift operation has a 6-bit field, but C16 to +16 are valid numbers for the instruction. 37 2.2.3 memory data formats memory data formats are classified into byte, word, and longword. byte data can be accessed from any address, but an address error will occur if word data starting from an address other than 2n or longword data starting from an address other than 4n is accessed. in such cases, the data accessed cannot be guaranteed (figure 2.11). 31 0 15 23 7 byte 0 byte 1 byte 2 byte 3 word 1 word 0 address a + 4 address a + 8 longword address a address a address a + 2 address a + 1 address a + 3 big-endian mode 31 0 15 23 7 byte 3 byte 2 byte 1 byte 0 word 0 word 1 address a + 4 address a longword address a + 8 address a + 11 address a + 9 address a + 10 address a + 8 little-endian mode figure 2.11 byte, word, and longword alignment either big-endian or little-endian byte order can be selected for the data format, according to the md5 pin at reset. when md5 is low at reset, the processor operates in big-endian mode. when md5 is high at reset, the processor operates in little-endian mode. 2.3 features of cpu core instructions the cpu core instructions are risc-type instructions with the following features: fixed 16-bit length: all instructions have a fixed length of 16 bits. this improves program code efficiency. one instruction per state: pipelining is used, and basic instructions can be executed in one state. data size: the basic data size for operations is longword. byte, word, or longword can be selected as the memory access size. memory byte or word data is sign-extended and operated on as longword data. immediate data is sign-extended to longword size for arithmetic operations or zero-extended to longword size for logical operations. 38 table 2.6 word data sign extension sh7729r cpu description example of other cpu mov.w @(disp,pc),r1 add r1,r0 ........ .data.w h'1234 sign-extended to 32 bits, r1 becomes h'00001234, and is then operated on by the add instruction. add.w #h'1234,r0 note: immediate data is referenced by @(disp,pc). load/store architecture: basic operations are executed between registers. in operations involving memory, data is first loaded into a register (load/store architecture). however, bit manipulation instructions such as and are executed directly on memory. delayed branching: unconditional branch instructions, etc., are executed as delayed branches. with a delayed branch instruction, the branch is made after execution of the instruction (called the slot instruction) immediately following the delayed branch instruction. this minimizes disruption of the pipeline when a branch is made. with a delayed branch, the actual branch operation occurs after execution of the slot instruction. however, instruction execution for register updating, etc., excluding the branch operation, is performed in delayed branch instruction delay slot instruction order. for example, even though the contents of the register holding the branch destination address are changed in the delay slot, the branch destination address remains as the register contents prior to the change. table 2.7 delayed branch instructions sh7729r cpu description example of other cpu bra trget add r1,r0 add is executed before branch to trget. add.w r1,r0 bra trget multiply/multiply-and-accumulate operations: a 16 16 32 multiply operation is executed in 1 to 3 states, and a 16 16 + 64 64 multiply-and-accumulate operation in 2 to 3 states. a 32 32 64 multiply operation and a 32 32 + 64 64 multiply-and-accumulate operation are each executed in 2 to 4 states. 39 t bit: the result of a comparison is indicated by the t bit in the status register (sr), and a conditional branch is performed according to whether the result is true or false. processing speed has been improved by keeping the number of instructions that modify the t bit to a minimum. table 2.8 t bit sh7729r cpu description example of other cpu cmp/ge r1,r0 bt trget0 bf trget1 if r0 r1, the t bit is set. a branch is made to trget0 if r0 r1, or to trget1 if r0 < r1. cmp.w r1,r0 bge trget0 blt trget1 add #C1,r0 cmp/eq #0,r0 bttrget the t bit is not set by add. if r0 = 0, the t bit is set. a branch is made if r0 = 0. sub.w #1,r0 beq trget immediate data: byte immediate data is placed inside the instruction code. word and longword immediate data is not placed inside the instruction code, but in a table in memory. the table in memory is referenced with an immediate data transfer instruction (mov) using pc-relative addressing mode with displacement. table 2.9 immediate data referencing type sh7729r cpu example of other cpu 8-bit immediate mov #h'12,r0 mov.b #h'12,r0 16-bit immediate mov.w @(disp,pc),r0 ........ .data.w h'1234 mov.w #h'1234,r0 32-bit immediate mov.l @(disp,pc),r0 ........ .data.l h'12345678 mov.l #h'12345678,r0 note: immediate data is referenced by @(disp,pc). 40 absolute addresses: when data is referenced by absolute address, the absolute address value is placed in a table in memory beforehand. using the method whereby immediate data is loaded when an instruction is executed, this value is transferred to a register and the data is referenced using register indirect addressing mode. table 2.10 absolute address referencing type sh7729r cpu example of other cpu absolute address mov.l @(disp,pc),r1 mov.b @r1,r0 ........ .data.l h'12345678 mov.b @h'12345678,r0 16-bit/32-bit displacement: when data is referenced with a 16- or 32-bit displacement, the displacement value is placed in a table in memory beforehand. using the method whereby immediate data is loaded when an instruction is executed, this value is transferred to a register and the data is referenced using indexed register indirect addressing mode. table 2.11 displacement referencing type sh7729r cpu example of other cpu 16-bit displacement mov.w @(disp,pc),r0 mov.w @(r0,r1),r2 ........ .data.w h'1234 mov.w @(h'1234,r1),r2 41 2.4 instruction formats 2.4.1 cpu instruction addressing modes the following table shows addressing modes and effective address calculation methods for instructions executed by the cpu core. table 2.12 addressing modes and effective addresses for cpu instructions addressing mode instruction format effective address calculation method calculation formula register direct rn effective address is register rn. (operand is register rn contents.) register indirect @rn effective address is register rn contents. rn rn rn register indirect with post-increment @rn+ effective address is register rn contents. a constant is added to rn after instruction execution: 1 for a byte operand, 2 for a word operand, 4 for a longword operand. rn rn 1/2/4 + rn + 1/2/4 rn after instruction execution byte: rn + 1 rn word: rn + 2 rn longword: rn + 4 rn register indirect with pre-decrement @Crn effective address is register rn contents, decremented by a constant beforehand: 1 for a byte operand, 2 for a word operand, 4 for a longword operand. rn 1/2/4 rn C 1/2/4 C rn C 1/2/4 byte: rn C 1 rn word: rn C 2 rn longword: rn C 4 rn (instruction executed with rn after calculation) 42 table 2.12 addressing modes and effective addresses for cpu instructions (cont) addressing mode instruction format effective address calculation method calculation formula register indirect with displacement @(disp:4, rn) effective address is register rn contents with 4-bit displacement disp added. after disp is zero-extended, it is multiplied by 1 (byte), 2 (word), or 4 (longword), according to the operand size. rn 1/2/4 + disp (zero-extended) rn + disp 1/2/4 byte: rn + disp word: rn + disp 2 longword: rn + disp 4 indexed register indirect @(r0, rn) effective address is sum of register rn and r0 contents. rn r0 rn + r0 + rn + r0 gbr indirect with displacement @(disp:8, gbr) effective address is register gbr contents with 8-bit displacement disp added. after disp is zero-extended, it is multiplied by 1 (byte), 2 (word), or 4 (longword), according to the operand size. gbr 1/2/4 + disp (zero-extended) gbr + disp 1/2/4 byte: gbr + disp word: gbr + disp 2 longword: gbr + disp 43 table 2.12 addressing modes and effective addresses for cpu instructions (cont) addressing mode instruction format effective address calculation method calculation formula indexed gbr indirect @(r0, gbr) effective address is sum of register gbr and r0 contents. gbr r0 gbr + r0 + gbr + r0 pc-relative with displacement @(disp:8, pc) effective address is pc with 8-bit displacement disp added. after disp is zero-extended, it is multiplied by 2 (word) or 4 (longword), according to the operand size. with a longword operand, the lower 2 bits of pc are masked. pc h'fffffffc + 2/4 x & * * : with longword operand disp (zero-extended) pc + disp 2 or pc&h'fffffffc + disp 4 word: pc + disp 2 longword: pc&h'fffffffc + disp 4 44 table 2.12 addressing modes and effective addresses for cpu instructions (cont) addressing mode instruction format effective address calculation method calculation formula pc-relative disp:8 effective address is pc with 8-bit displacement disp added after being sign-extended and multiplied by 2. pc 2 + disp (sign-extended) pc + disp 2 pc + disp 2 disp:12 effective address is pc with 12-bit displacement disp added after being sign-extended and multiplied by 2 pc 2 + disp (sign-extended) pc + disp 2 pc + disp 2 rn effective address is sum of pc and rn. pc rn + pc + rn pc + rn immediate #imm:8 8-bit immediate data imm of tst, and, or, or xor instruction is zero-extended. #imm:8 8-bit immediate data imm of mov, add, or cmp/eq instruction is sign-extended. #imm:8 8-bit immediate data imm of trapa instruction is zero-extended and multiplied by 4. 45 2.4.2 dsp data addressing two different memory accesses are made with dsp instructions. the two kinds of instructions are x and y data transfer instructions (movx.w, movy.w) and single data transfer instructions (movs.w, movsl). the data addressing is different for these two kinds of instruction. an overview of the data transfer instructions is given in table 2.13. table 2.13 overview of data transfer instructions x/y data transfer processing (movx.w, movy.w) single data transfer processing (movs.w, movs.l) address register ax: r4, r5, ay: r6, r7 as: r2, r3, r4, r5 index register ix: r8, iy: r9 is: r8 addressing nop/inc (+2)/index addition: post-increment nop/inc (+2, +4)/index addition: post-increment dec (C2, C4): pre-decrement modulo addressing possible not possible data bus xdb, ydb ldb data length 16 bits (word) 16/32 bits (word/longword) bus contention no yes memory x/y data memory entire memory space source register dx, dy: a0, a1 ds: a0/a1, m0/m1, x0/x1, y0/y1, a0g, a1g destination register dx: x0/x1, dy: y0/y1 ds: a0/a1, m0/m1, x0/x1, y0/y1, a0g, a1g x/y data addressing: with dsp instructions, the x and y data memory can be accessed simultaneously using the movx.w and movy.w instructions. two address pointers are provided for dsp instructions to enable simultaneous access to x and y data memory. only pointer addressing can be used with dsp instructions; immediate addressing is not available. address registers are divided into two, with register r4 or r5 functioning as the x memory address register (ax), and register r6 or r7 as the y memory address register (ay). the following three kinds of addressing can be used with x and y data transfer instructions. 1. non-update address register addressing: the ax and ay registers are address pointers. they are not updated. 2. addition index register addressing: the ax and ay registers are address pointers. after a data transfer, the value of the ix or iy register is added to each (post-increment). 46 3. increment address register addressing: the ax and ay registers are address pointers. after a data transfer, they are each incremented by 2 (post- increment). there is an index register for each address pointer. the r8 register is the index register (ix) for the x memory address register (ax), and the r9 register is the index register (iy) for the y memory address register (ay). the x and y data transfer instructions perform word-length processing, and use 16-bit access to the x/y data memory. a value of 2 is therefore added to the address register in the increment processing. to perform decrementing, C2 is set in the index register and addition index register addressing is specified. in x/y data addressing, only bits 1 to 15 of the address pointer are valid. when using x/y data addressing, 0 must always be written to bit 0 of the address pointer and index register. x/y data transfer addressing is shown in figure 2.12. when accessing x and y memory using the x and y buses, the upper word of ax (r4 or r5) and ay (r6 or r7) is ignored. the result of @ay+ or @ay+iy is stored in the lower word of ay, while the upper word retains its original value. alu au r8[ix] r4[ax] r5[ax] r9[iy] r6[ay] r7[ay] +2 (inc) +0 (no update) +2 (inc) +0 (no update) note: three address processing methods: 1. increment 2. index register addition (ix/iy) 3. no increment post-updating is used in all cases. the address pointer can be decremented by setting C2/C4 in the index register. au: adder provided for dsp addressing figure 2.12 x and y data transfer addressing single data addressing: dsp instructions include two single data transfer instructions (movs.w, movs.l) that load data into, or store data from, a dsp register. with these instructions, one of registers r2 to r5 is used as the single data transfer address register (as). the following four kinds of addressing can be used with single data transfer instructions. 47 1. non-update address register addressing: the as register is an address pointer. it is not updated. 2. addition index register addressing: the as register is an address pointer. after a data transfer, the value of the is register is added to the as register (post-increment). 3. increment address register addressing: the as register is an address pointer. after a data transfer, the as register is incremented by 2 or 4 (post-increment). 4. decrement address register addressing: the as register is an address pointer. before a data transfer, C2 or C4 is added to the as register (i.e. 2 or 4 is subtracted) (pre-decrement). the r8 register is the index register (is) for the address pointer (as). single data transfer addressing is shown in figure 2.13. alu r8[is] r4[as] r5[as] C2/C4 (dec) +2/+4 (inc) +0 (no update) r3[as] r2[as] 31 0 31 0 mab cab 31 0 note: four address processing methods: 1. no update 2. index register addition (is) 3. increment 4. decrement post-increment pre-decrement figure 2.13 single data transfer addressing modulo addressing: like other dsps, the sh7729r has a modulo addressing mode. address registers are updated in the same way in this mode. when the address pointer value reaches the preset modulo end address, the address pointer value becomes the modulo start address. modulo addressing is only available for the x and y data transfer instructions (movx.w, movy.w). modulo addressing mode is specified for the x address register by setting the dmx bit in the sr register, and for the y address register by setting the dmy bit. modulo addressing is valid for either the x or the y address register, only; it cannot be set for both at the same time. 48 therefore, dmx and dmy cannot both be set simultaneously (if they are, the dmy setting will be valid). the mod register is provided to set the start and end addresses of the modulo address area. the mod register contains ms (modulo start) and me (modulo end). an example of the use of the mod register (ms and me fields) is shown below. mov.l modaddr,rn; rn=modend, modstart ldc rn,mod; me=modend, ms=modstart modaddr: .data.w mend; modend .data.w mstart; modstart modstart: .data : modend: .data the start and end addresses are specified in ms and me, then the dmx or dmy bit is set to 1. the address register contents are compared with me, and if they match, start address ms is stored in the address register. the lower 16 bits of the address register are compared with me. the maximum modulo size is 64 kbytes. this is sufficient to access the x and y data memory. a block diagram of modulo addressing is shown in figure 2.14. alu au r8[ix] r9[iy] r6[ay] r7[ay] +2 +0 31 0 r4[ax] r5[ax] 31 16 15 0 15 1 abx xab +2 +0 31 0 31 16 15 0 15 1 aby yab dmx dmy cont ms cmp me 15 1 15 1 instruction (movx/movy) figure 2.14 modulo addressing an example of modulo addressing is given below. ms = h'7008; me=h'700c; r4=h'a5007008; dmx = 1; dmy = 0: (modulo addressing setting for address register ax (r4, r5)) 49 as a result of the above settings, the r4 register changes as follows. r4: h'a5007008 inc. r4: h'a500700a inc. r4: h'a500700c inc. r4: h'a5007008 (reaches modulo end address, so becomes modulo start address) place the data so that the upper 16 bits of the modulo start and end addresses are the same. this is because the modulo start address overwrites only the lower 15 bits of the address register, excluding bit 0. note: when addition indexing is used for dsp data addressing, the address pointer may exceed the me value without actually reaching it. in this case, the address pointer will not return to the modulo start address. not only with modulo addressing, but when x and y data addressing is used, bit 0 is ignored. 0 must always be written to bit 0 of the address pointer, index register, ms, and me. dsp addressing operations: dsp addressing operations in the pipeline execution stage (ex), including modulo addressing, are shown below. if ( operation is movx.w movy.w ) { abx=ax; aby=ay; /* memory access cycle uses abx and aby. the addresses to be used have not been updated */ /* ax is one of r4,5 */ if ( dmx==0 || dmx==1 && dmy == 1 )} ax=ax+(+2 or r8[ix] or +0); /* inc,index,not-update */ else if (! not-update) ax=modulo( ax, (+2 or r8[ix]) ); /* ay is one of r6,7 */ if ( dmy==0 ) ay=ay+(+2 or r9[iy] or +0); /* inc,index,not-update */ else if (! not-update) ay=modulo( ay, (+2 or r9[iy]) ); } else if ( operation is movs.w or movs.l ) { if ( addressing is nop, inc, add-index-reg ) { mab=as; /* memory access cycle uses mab. the address to be used has not been updated */ /* as is one of r2? */ as=as+(+2 or +4 or r8[is] or +0); /* inc,index,not-update */ 50 else { /* decrement, pre-update */ /* as is one of r2~5 */ as=as+(-2 or -4); mab=as; /* memory access cycle uses mab. the address to be used has been updated */ } /* the value to be added to the address register depends on addressing operations. for example, (+2 or r8[ix] or +0) means that +2 : if operation is increment r8[ix] : if operation is add-index-reg +0 : if operation is not-update */ function modulo ( addrreg, index ) { if ( adrreg[15:0]==me ) adrreg[15:0]==ms; else adrreg=adrreg+index; return addrreg; } 2.4.3 cpu instruction formats table 2.14 shows the instruction formats, and the meaning of the source and destination operands, for instructions executed by the cpu core. the meaning of the operands depends on the instruction code. the following symbols are used in the table. xxxx: instruction code mmmm: source register nnnn: destination register iiii: immediate data dddd: displacement 51 table 2.14 cpu instruction formats instruction format source operand destination operand sample instruction 0 type xxxx xxxx xxxx xxxx 15 0 nop n type xxxx xxxx xxxx nnnn 15 0 nnnn : register direct mov trn control register or system register nnnn : register direct sts mach,rn control register or system register nnnn : pre- decrement register indirect stc.l sr,@-rn m type xxxx mmmm xxxx xxxx 15 0 mmmm : register direct control register or system register ldc rm,sr mmmm : post- increment register indirect control register or system register ldc.l @rm+,sr mmmm : register indirect jmp @rm pc-relative using rm braf rm 52 table 2.14 cpu instruction formats (cont) instruction format source operand destination operand sample instruction nm type nnnn xxxx xxxx 15 0 mmmm mmmm : register direct nnnn : register direct add rm,rn mmmm : register indirect nnnn : register indirect mov.l rm,@rn mmmm : post- increment register indirect (multiply- and-accumulate operation) nnnn : * post- increment register indirect (multiply- and-accumulate operation) mach, macl mac.w @rm+,@rn+ mmmm : post- increment register indirect nnnn : register direct mov.l @rm+,rn mmmm : register direct nnnn : pre- decrement register indirect mov.l rm,@-rn mmmm : register direct nnnn : indexed register indirect mov.l rm,@(r0,rn) md type xxxx dddd 15 0 mmmm xxxx mmmmdddd : register indirect with displacement r0 (register direct) mov.b @(disp,rm),r0 nd4 type dddd nnnn xxxx 15 0 xxxx r0 (register direct) nnnndddd : register indirect with displacement mov.b r0,@(disp,rn) nmd type nnnn xxxx dddd 15 0 mmmm mmmm : register direct nnnndddd : register indirect with displacement mov.l rm,@(disp,rn) mmmmdddd : register indirect with displacement nnnn : register direct mov.l @(disp,rm),rn note: * in multiply-and-accumulate instructions, nnnn is the source register. 53 table 2.14 cpu instruction formats (cont) instruction format source operand destination operand sample instruction d type dddd xxxx 15 0 xxxx dddd dddddddd : gbr indirect with displacement r0 (register direct) mov.l @@(disp,gbr),r0 r0 (register direct) dddddddd : gbr indirect with displacement mov.l @r0,@(disp,gbr) dddddddd : pc-relative with displacement r0 (register direct) mova @(disp,pc),r0 dddddddd : pc-relative bf label d12 type dddd xxxx 15 0 dddd dddd dddddddddddd : pc-relative bra label (label=disp+pc) nd8 type dddd nnnn xxxx 15 0 dddd dddddddd : pc- relative with displacement nnnn : register direct mov.l @(disp,pc),rn i type i i i i xxxx 15 0 xxxx i i i i iiiiiiii : immediate indexed gbr indirect and.b #imm,@(r0,gbr) iiiiiiii : immediate r0 (register direct) and #imm,r0 iiiiiiii : immediate trapa #imm ni type nnnn i i i i xxxx 15 0 i i i i iiiiiiii : immediate nnnn : register direct add #imm,rn 54 2.4.4 dsp instruction formats the sh7729r includes new instructions for digital signal processing. the new instructions are of the following two kinds. 1. memory and dsp register double and single data transfer instructions (16-bit length) 2. parallel processing instructions processed by the dsp unit (32-bit length) the instruction formats are shown in figure 2.15. 15 16 25 26 31 0 0 0 0 15 15 15 9 9 10 10 a field a field a field b field 111110 111101 111100 0000 1110 cpu core instructions double data transfer instructions single data transfer instructions parallel processing instructions . . . figure 2.15 dsp instruction formats 55 double and single data transfer instructions: the format of double data transfer instructions is shown in table 2.15, and that of single data transfer instructions in table 2.16. table 2.15 double data transfer instruction formats type mnemonic 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 x memory nopx 1 1 1 1 0 0 0 0 0 00 data movx.w @ax,dx ax dx 0 01 transfer movx.w @ax+,dx 10 movx.w @ax+ix,dx 11 movx.w da,@ax da 1 01 movx.w da,@ax+ 10 movx.w da,@ax+ix 11 y memory nopy 1 1 1 1 0 0 0 0 0 00 data movy.w @ay,dy ay dy 0 01 transfer movy.w @ay+,dy 10 movy.w @ay+iy,dy 11 movy.w da,@ay da 1 01 movy.w da,@ay+ 10 movy.w da,@ay+iy 11 note: ax: 0 = r4, 1 = r5 ay: 0 = r6, 1 = r7 dx: 0 = x0, 1 = x1 dy: 0 = y0, 1 = y1 da: 0 = a0, 1 = a1 56 table 2.16 single data transfer instruction formats type mnemonic 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 single movs.w @-as,ds 1 1 1 1 0 1 as ds 0:( * ) 0 000 data movs.w @as,ds 0:r4 1:( * )01 transfer movs.w @as+,ds 1:r5 2:( * )10 movs.w @as+ix,ds 2:r2 3:( * )11 movs.w ds,@-as 3:r3 4:( * ) 0 001 movs.w ds,@as 5:a1 0 1 movs.w ds,@as+ 6:( * )10 movs.w ds,@as+ix 7:a0 1 1 movs.l @-as,ds 8:x0 0 0 1 0 movs.l @as,ds 9:x1 0 1 movs.l @as+,ds a:y0 1 0 movs.l @as+ix,ds b:y1 1 1 movs.l ds,@-as c:m0 0 0 1 1 movs.l ds,@as d:a1g 0 1 movs.l ds,@as+ e:m1 1 0 movs.l ds,@as+ix f:a0g 1 1 note: * codes reserved for system use. parallel processing instructions: parallel processing instructions are provided for efficient execution of digital signal processing using the dsp unit. they are 32 bits long and allow four simultaneous processes, an alu operation, multiplication, and two data transfers. parallel processing instructions are divided into an a field and a b field. the a field defines data transfer instructions and the b field an alu operation instruction and multiply instruction. these instructions can be defined independently, and the processing is executed in parallel, independently and simultaneously. a-field parallel data transfer instructions are shown in table 2.17, and b-field alu operation instructions and multiply instructions in table 2.18. 57 table 2.17 a-field parallel data transfer instructions nopx movx.w @ax, dx movx.w @ax+, dx movx.w @ax+ix, dx movx.w da, @ax movx.w da, @ax+ movx.w da, @ax+ix nopy movy.w @ay, dy movy.w @ay+, dy movy.w @ay+iy, dy movy.w da, @ay movy.w da, @ay+ movy.w da, @ay+iy mnemonic x memory data transfer type 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 43 2 1 0 y memory data transfer 111110 note: ax: 0 = r4, 1 = r5 ay: 0 = r6, 1 = r7 dx: 0 = x0, 1 = x1 dy: 0 = y0, 1 = y1 da: 0 = a0, 1 = a1 0 ax 0 ay 0 dx da 0 dy da 0 0 1 0 0 1 0 0 1 1 0 1 1 0 1 0 1 1 0 1 0 0 1 1 0 1 1 0 1 0 1 1 0 1 b field b field 58 table 2.18 b-field alu operation instructions and multiply instructions pshl #imm, dz psha #imm, dz reserved pmuls se, sf, dg reserved psub sx, sy, du pmuls se, sf, dg padd sx, sy, du pmuls se, sf, dg reserved psubc sx, sy, dz paddc sx, sy, dz pcmp sx, sy reserved reserved reserved pabs sx, dz prnd sx, dz pabs sy, dz prnd sy, dz reserved mnemonic imm. shift type 0:( * ) 1:( * ) 2:( * ) 3:( * ) 4:( * ) 5:a1 6:( * ) 7:a0 8:x0 9:x1 a:y0 b:y1 c:m0 d:( * ) e:m1 f:( * ) 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 6-operand parallel instructions 3-operand instructions 111110 0 0 0 0 0 0 0 0 1 0 0 0 0 1 1 1 1 0 0 0 0 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 0 1 C16 < = imm < = +16 C32 < = imm < = +32 se sf sx sy dg du 0:x0 1:x1 2:y0 3:a1 0:y0 1:y1 2:x0 3:a1 0:x0 1:x1 2:a0 3:a1 0:y0 1:y1 2:m0 3:m1 0:m0 1:m1 2:a0 3:a1 0:x0 1:y0 2:a0 3:a1 0 0 1 1 0 1 0 1 00 dz dz note: * codes reserved for system use. a field 59 table 2.18 b-field alu operation instructions and multiply instructions (cont) [if cc] pshl sx, sy, dz [if cc] psha sx, sy, dz [if cc] psub sx, sy, dz [if cc] padd sx, sy, dz reserved [if cc] pand sx, sy, dz [if cc] pxor sx, sy, dz [if cc] por sx, sy, dz [if cc] pdec sx, dz [if cc] pinc sx, dz [if cc] pdec sy, dz [if cc] pinc sy, dz [if cc] pclr dz [if cc] pdmsb sx, dz reserved [if cc] pdmsb sy, dz [if cc] pneg sx, dz [if cc] pcopy sx, dz [if cc] pneg sy, dz [if cc] pcopy sy, dz reserved [if cc] psts mach, dz [if cc] psts macl, dz [if cc] plds dz, mach [if cc] plds dz, macl ( * 2) reserved reserved mnemonic type if cc 01: 10: dct 11: dcf 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 43 2 1 0 conditional 3-operand instructions 111110 11 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 sx 0:x0 1:x1 2:a0 3:a1 0 0 1 1 1 1 0 0 1 0 1 0 1 * 0:( * 1) 1:( * 1) 2:( * 1) 3:( * 1) 4:( * 1) 5:a1 6:( * 1) 7:a0 8:x0 9:x1 a:y0 b:y1 c:m0 d:( * 1) e:m1 f:( * 1) dz 111111 0 0 0 0 sy 0:y0 1:y1 2:m0 3:m1 uncon- ditional if cc notes: * 1 codes reserved for system use. * 2 [if cc]: dct (dc bit true), dcf (dc bit false) or none (unconditional instruction) a field 60 2.5 instruction set 2.5.1 cpu instruction set the sh-1/sh-2/sh-3 compatible instruction set consists of 66 basic instruction types divided into seven functional groups, as shown in table 2.19. tables 2.20 to 2.25 show the instruction notation, machine code, execution time, and function. table 2.19 cpu instruction types type kinds of instruction op code function number of instructions data transfer instructions 5 mov data transfer immediate data transfer peripheral module data transfer structure data transfer 39 mova effective address transfer movtt bit transfer swap upper/lower swap xtrct extraction of middle of linked registers arithmetic 21 add binary addition 33 operation addc binary addition with carry instructions addv binary addition with overflow check cmp/cond comparison div1 division div0s signed division initialization div0u unsigned division initialization dmuls signed double-precision multiplication dmulu unsigned double-precision multiplication dtdecrement and test exts sign extension extu zero extension mac multiply-and-accumulate, double- precision multiply-and-accumulate mul double-precision multiplication (32 32 bits) 61 table 2.19 cpu instruction types (cont) type kinds of instruction op code function number of instructions arithmetic 21 muls signed multiplication (16 16 bits) 34 operation mulu unsigned multiplication (16 16 bits) instructions neg sign inversion negc sign inversion with borrow sub binary subtraction subc binary subtraction with carry subv binary subtraction with underflow logic 6 and logical and 14 operation notbit inversion instructions or logical or tas memory test and bit setting tst logical and and t bit setting xor exclusive logical or shift 12 rotl 1-bit left shift 16 instructions rotr 1-bit right shift rotcl 1-bit left shift with t bit rotcr 1-bit right shift with t bit shal arithmetic 1-bit left shift shar arithmetic 1-bit right shift shll logical 1-bit left shift shlln logical n-bit left shift shlr logical 1-bit right shift shlrn logical n-bit right shift shad arithmetic dynamic shift shld logical dynamic shift 62 table 2.19 cpu instruction types (cont) type kinds of instruction op code function number of instructions branch instructions 9 bf conditional branch, delayed conditional branch (t = 0) 11 btconditional branch, delayed conditional branch (t = 1) bra unconditional branch braf unconditional branch bsr branch to subroutine procedure bsrf branch to subroutine procedure jmp unconditional branch jsr branch to subroutine procedure rts return from subroutine procedure system 15 clrtt bit clear 75 control clrmac mac register clear instructions clrs s bit clear ldc load into control register lds load into system register ldtlb pteh/ptel load into tlb nop no operation pref data prefetch to cache rte return from exception handling sets s bit setting sett t bit setting sleep transition to power-down mode stc store from control register sts store from system register trapa trap exception handling total: 68 189 63 the instruction code, operation, and number of execution states of the cpu instructions are shown in the following tables, classified by instruction type, using the format shown below. instruction instruction code operation privilege execution states t bit indicated by mnemonic. explanation of symbols op.sz src, dest op: operation code sz: size src: source dest: destination rm: source register rn: destination register imm: immediate data disp: displacement indicated in msb ? lsb order. explanation of symbols mmmm: source register nnnn: destination register 0000: r0 0001: r1 ......... 1111: r15 iiii: immediate data dddd: displacement* 2 indicates summary of operation. explanation of symbols , : transfer direction (xx): memory operand m/q/t: flag bits in sr &: logical and of each bit |: logical or of each bit ^: exclusive logical or of each bit ~: logical not of each bit < 64 data transfer instructions table 2.20 data transfer instructions instruction operation code privileged mode cycles t bit mov #imm,rn imm sign extension rn 1110nnnniiiiiiii 1 mov.w @(disp,pc),rn (disp 2 + pc) sign extension rn 1001nnnndddddddd 1 mov.l @(disp,pc),rn (disp 4 + pc) rn 1101nnnndddddddd 1 mov rm,rn rm rn 0110nnnnmmmm0011 1 mov.b rm,@rn rm (rn) 0010nnnnmmmm0000 1 mov.w rm,@rn rm (rn) 0010nnnnmmmm0001 1 mov.l rm,@rn rm (rn) 0010nnnnmmmm0010 1 mov.b @rm,rn (rm) sign extension rn 0110nnnnmmmm0000 1 mov.w @rm,rn (rm) sign extension rn 0110nnnnmmmm0001 1 mov.l @rm,rn (rm) rn 0110nnnnmmmm0010 1 mov.b rm,@ C rn rnC1 rn, rm (rn) 0010nnnnmmmm0100 1 mov.w rm,@ C rn rnC2 rn, rm (rn) 0010nnnnmmmm0101 1 mov.l rm,@ C rn rnC4 rn, rm (rn) 0010nnnnmmmm0110 1 mov.b @rm+,rn (rm) sign extension rn, rm + 1 rm 0110nnnnmmmm0100 1 mov.w @rm+,rn (rm) sign extension rn, rm + 2 rm 0110nnnnmmmm0101 1 mov.l @rm+,rn (rm) rn,rm + 4 rm 0110nnnnmmmm0110 1 mov.b r0,@(disp,rn) r0 (disp + rn) 10000000nnnndddd 1 mov.w r0,@(disp,rn) r0 (disp 2 + rn) 10000001nnnndddd 1 mov.l rm,@(disp,rn) rm (disp 4 + rn) 0001nnnnmmmmdddd 1 mov.b @(disp,rm),r0 (disp + rm) sign extension r0 10000100mmmmdddd 1 mov.w @(disp,rm),r0 (disp 2 + rm) sign extension r0 10000101mmmmdddd 1 mov.l @(disp,rm),rn (disp 4 + rm) rn 0101nnnnmmmmdddd 1 mov.b rm,@(r0,rn) rm (r0 + rn) 0000nnnnmmmm0100 1 65 table 2.20 data transfer instructions (cont) instruction operation code privileged mode cycles t bit mov.w rm,@(r0,rn) rm (r0 + rn) 0000nnnnmmmm0101 1 mov.l rm,@(r0,rn) rm (r0 + rn) 0000nnnnmmmm0110 1 mov.b @(r0,rm),rn (r0 + rm) sign extension rn 0000nnnnmmmm1100 1 mov.w @(r0,rm),rn (r0 + rm) sign extension rn 0000nnnnmmmm1101 1 mov.l @(r0,rm),rn (r0 + rm) rn 0000nnnnmmmm1110 1 mov.b r0,@(disp,gbr) r0 (disp + gbr) 11000000dddddddd 1 mov.w r0,@(disp,gbr) r0 (disp 2 + gbr) 11000001dddddddd 1 mov.l r0,@(disp,gbr) r0 (disp 4 + gbr) 11000010dddddddd 1 mov.b @(disp,gbr),r0 (disp + gbr) sign extension r0 11000100dddddddd 1 mov.w @(disp,gbr),r0 (disp 2 + gbr) sign extension r0 11000101dddddddd 1 mov.l @(disp,gbr),r0 (disp 4 + gbr) r0 11000110dddddddd 1 mova @(disp,pc),r0 disp 4 + pc r0 11000111dddddddd 1 movt rn t rn 0000nnnn00101001 1 swap.b rm,rn rm swap lowest two bytes reg 0110nnnnmmmm1000 1 swap.w rm,rn rm swap two consecutive words rn 0110nnnnmmmm1001 1 xtrct rm,rn rm: middle 32 bits of rn rn 0010nnnnmmmm1101 1 66 arithmetic operation instructions table 2.21 arithmetic operation instructions instruction operation code privileged mode cycles t bit add rm,rn rn + rm rn 0011nnnnmmmm1100 1 add #imm,rn rn + imm rn 0111nnnniiiiiiii 1 addc rm,rn rn + rm + t rn, carry t 0011nnnnmmmm1110 1 carry addv rm,rn rn + rm rn, overflow t 0011nnnnmmmm1111 1 overflow cmp/eq #imm,r0 if r0 = imm, 1 t 10001000iiiiiiii 1 comparison result cmp/eq rm,rn if rn = rm, 1 t 0011nnnnmmmm0000 1 comparison result cmp/hs rm,rn if rn rm with unsigned data, 1 t 0011nnnnmmmm0010 1 comparison result cmp/ge rm,rn if rn rm with signed data, 1 t 0011nnnnmmmm0011 1 comparison result cmp/hi rm,rn if rn > rm with unsigned data, 1 t 0011nnnnmmmm0110 1 comparison result cmp/gt rm,rn if rn > rm with signed data, 1 t 0011nnnnmmmm0111 1 comparison result cmp/pz rn if rn 0, 1 t 0100nnnn00010001 1 comparison result cmp/pl rn if rn > 0, 1 t 0100nnnn00010101 1 comparison result cmp/str rm,rn if rn and rm have an equivalent byte, 1 t 0010nnnnmmmm1100 1 comparison result div1 rm,rn single-step division (rn/rm) 0011nnnnmmmm0100 1 calculation result div0s rm,rn msb of rn q, msb of rm m, m ^ q t 0010nnnnmmmm0111 1 calculation result div0u 0 m/q/t 0000000000011001 10 67 table 2.21 arithmetic operation instructions (cont) instruction operation code privileged mode cycles t bit dmuls.l rm,rn signed operation of rn rm mach, macl 32 32 64 bits 0011nnnnmmmm1101 2(5)* 1 dmulu.l rm,rn unsigned operation of rn rm mach, macl 32 32 64 bits 0011nnnnmmmm0101 2(5)* 1 dt rn rn C 1 rn, if rn = 0, 1 t, else 0 t 0100nnnn00010000 1 comparison result exts.b rm,rn a byte in rm is sign- extended rn 0110nnnnmmmm1110 1 exts.w rm,rn a word in rm is sign- extended rn 0110nnnnmmmm1111 1 extu.b rm,rn a byte in rm is zero- extended rn 0110nnnnmmmm1100 1 extu.w rm,rn a word in rm is zero- extended rn 0110nnnnmmmm1101 1 mac.l @rm+,@rn+ signed operation of (rn) (rm) + mac mac, rn + 4 rn, rm + 4 rm 32 32 + 64 64 bits 0000nnnnmmmm1111 2(5)* 1 mac.w @rm+,@rn+ signed operation of (rn) (rm) + mac mac, rn + 2 rn, rm + 2 rm 16 16 + 64 64 bits 0100nnnnmmmm1111 2(5)* 1 mul.l rm,rn rn rm macl 32 32 32 bits 0000nnnnmmmm0111 2(5)* 1 muls.w rm,rn signed operation of rn rm mac 16 16 32 bits 0010nnnnmmmm1111 1(3)* 2 mulu.w rm,rn unsigned operation of rn rm mac 16 16 32 bits 0010nnnnmmmm1110 1(3)* 2 notes: *1 the normal minimum number of execution cycles is two, but five cycles are required when the operation result is read from the mac register immediately after the instruction. *2 the normal minimum number of execution cycles is one, but three cycles are required when the operation result is read from the mac register immediately after the mul instruction. 68 table 2.21 arithmetic operation instructions (cont) instruction operation code privileged mode cycles t bit neg rm,rn 0Crm rn 0110nnnnmmmm1011 1 negc rm,rn 0CrmCt rn, borrow t 0110nnnnmmmm1010 1 borrow sub rm,rn rnCrm rn 0011nnnnmmmm1000 1 subc rm,rn rnCrmCt rn, borrow t 0011nnnnmmmm1010 1 borrow subv rm,rn rnCrm rn, underflow t 0011nnnnmmmm1011 1 underflow 69 logic operation instructions table 2.22 logic operation instructions instruction operation code privileged mode cycles t bit and rm,rn rn & rm rn 0010nnnnmmmm1001 1 and #imm,r0 r0 & imm r0 11001001iiiiiiii 1 and.b #imm,@(r0,gbr) (r0 + gbr) & imm (r0 + gbr) 11001101iiiiiiii 3 not rm,rn ~rm rn 0110nnnnmmmm0111 1 or rm,rn rn | rm rn 0010nnnnmmmm1011 1 or #imm,r0 r0 | imm r0 11001011iiiiiiii 1 or.b #imm,@(r0,gbr) (r0 + gbr) | imm (r0 + gbr) 11001111iiiiiiii 3 tas.b @rn if (rn) is 0, 1 t; 1 msb of (rn) 0100nnnn00011011 3 test result tst rm,rn rn & rm; if the result is 0, 1 t 0010nnnnmmmm1000 1 test result tst #imm,r0 r0 & imm; if the result is 0, 1 t 11001000iiiiiiii 1 test result tst.b #imm,@(r0,gbr) (r0 + gbr) & imm; if the result is 0, 1 t 11001100iiiiiiii 3 test result xor rm,rn rn ^ rm rn 0010nnnnmmmm1010 1 xor #imm,r0 r0 ^ imm r0 11001010iiiiiiii 1 xor.b #imm,@(r0,gbr) (r0 + gbr) ^ imm (r0 + gbr) 11001110iiiiiiii 3 70 shift instructions table 2.23 shift instructions instruction operation code privileged mode cycles t bit rotl rn t rn msb 0100nnnn00000100 1 msb rotr rn lsb rn t 0100nnnn00000101 1 lsb rotcl rn t rn t 0100nnnn00100100 1 msb rotcr rn t rn t 0100nnnn00100101 1 lsb shad rm,rn rn 0: rn << rm rn rn < 0: rn >> rm [msb rn] 0100nnnnmmmm1100 1 shal rn t rn 0 0100nnnn00100000 1 msb shar rn msb rn t 0100nnnn00100001 1 lsb shld rm,rn rn 0: rn << rm rn rn < 0: rn >> rm [0 rn] 0100nnnnmmmm1101 1 shll rn t rn 0 0100nnnn00000000 1 msb shlr rn 0 rn t 0100nnnn00000001 1 lsb shll2 rn rn << 2 rn 0100nnnn00001000 1 shlr2 rn rn >> 2 rn 0100nnnn00001001 1 shll8 rn rn << 8 rn 0100nnnn00011000 1 shlr8 rn rn >> 8 rn 0100nnnn00011001 1 shll16 rn rn << 16 rn 0100nnnn00101000 1 shlr16 rn rn >> 16 rn 0100nnnn00101001 1 71 branch instructions table 2.24 branch instructions instruction operation code privileged mode cycles t bit bf label if t = 0, disp 2 + pc pc; if t = 1, nop (where label is disp + pc) 10001011dddddddd 3/1* bf/s label delayed branch, if t = 0, disp 2 + pc pc; if t = 1, nop 10001111dddddddd 2/1* bt label delayed branch, if t = 1, disp 2 + pc pc; if t = 0, nop 10001001dddddddd 3/1* bt/s label if t = 1, disp 2 + pc pc; if t = 0, nop 10001101dddddddd 2/1* bra label delayed branch, disp 2 + pc pc 1010dddddddddddd 2 braf rm delayed branch, rm + pc pc 0000mmmm00100011 2 bsr label delayed branch, pc pr, disp 2 + pc pc 1011dddddddddddd 2 bsrf rm delayed branch, pc pr, rm + pc pc 0000mmmm00000011 2 jmp @rm delayed branch, rm pc 0100mmmm00101011 2 jsr @rm delayed branch, pc pr, rm pc 0100mmmm00001011 2 rts delayed branch, pr pc 0000000000001011 2 note: * one state when the branch is not executed. 72 system control instructions table 2.25 system control instructions instruction operation code privileged mode cycles t bit clrmac 0 mach, macl 0000000000101000 1 clrs 0 s 0000000001001000 1 clrt 0 t 0000000000001000 10 ldc rm,sr rm sr 0100mmmm00001110 5 lsb ldc rm,gbr rm gbr 0100mmmm00011110 1 ldc rm,vbr rm vbr 0100mmmm00101110 1 ldc rm,ssr rm ssr 0100mmmm00111110 1 ldc rm,spc rm spc 0100mmmm01001110 1 ldc rm,r0_bank rm r0_bank 0100mmmm10001110 1 ldc rm,r1_bank rm r1_bank 0100mmmm10011110 1 ldc rm,r2_bank rm r2_bank 0100mmmm10101110 1 ldc rm,r3_bank rm r3_bank 0100mmmm10111110 1 ldc rm,r4_bank rm r4_bank 0100mmmm11001110 1 ldc rm,r5_bank rm r5_bank 0100mmmm11011110 1 ldc rm,r6_bank rm r6_bank 0100mmmm11101110 1 ldc rm,r7_bank rm r7_bank 0100mmmm11111110 1 ldc.l @rm+,sr (rm) sr, rm + 4 rm 0100mmmm00000111 7 lsb ldc.l @rm+,gbr (rm) gbr, rm + 4 rm 0100mmmm00010111 1 ldc.l @rm+,vbr (rm) vbr, rm + 4 rm 0100mmmm00100111 1 ldc.l @rm+,ssr (rm) ssr, rm + 4 rm 0100mmmm00110111 1 ldc.l @rm+,spc (rm) spc, rm + 4 rm 0100mmmm01000111 1 ldc.l @rm+, r0_bank (rm) r0_bank, rm + 4 rm 0100mmmm10000111 1 ldc.l @rm+, r1_bank (rm) r1_bank, rm + 4 rm 0100mmmm10010111 1 ldc.l @rm+, r2_bank (rm) r2_bank, rm + 4 rm 0100mmmm10100111 1 ldc.l @rm+, r3_bank (rm) r3_bank, rm + 4 rm 0100mmmm10110111 1 73 table 2.25 system control instructions (cont) instruction operation code privileged mode cycles t bit ldc.l @rm+, r4_bank (rm) r4_bank, rm + 4 rm 0100mmmm11000111 1 ldc.l @rm+, r5_bank (rm) r5_bank, rm + 4 rm 0100mmmm11010111 1 ldc.l @rm+, r6_bank (rm) r6_bank, rm + 4 rm 0100mmmm11100111 1 ldc.l @rm+, r7_bank (rm) r7_bank, rm + 4 rm 0100mmmm11110111 1 lds rm,mach rm mach 0100mmmm00001010 1 lds rm,macl rm macl 0100mmmm00011010 1 lds rm,pr rm pr 0100mmmm00101010 1 lds.l @rm+,mach (rm) mach, rm + 4 rm 0100mmmm00000110 1 lds.l @rm+,macl (rm) macl, rm + 4 rm 0100mmmm00010110 1 lds.l @rm+,pr (rm) pr, rm + 4 rm 0100mmmm00100110 1 ldtlb pteh/ptel tlb 0000000000111000 1 nop no operation 0000000000001001 1 pref @rm (rm) cache 0000mmmm10000011 1 rte delayed branch, ssr/spc sr/pc 0000000000101011 4 sets 1 s 0000000001011000 1 sett 1 t 0000000000011000 11 sleep sleep 0000000000011011 4* stc sr,rn sr rn 0000nnnn00000010 1 stc gbr,rn gbr rn 0000nnnn00010010 1 stc vbr,rn vbr rn 0000nnnn00100010 1 stc ssr,rn ssr rn 0000nnnn00110010 1 stc spc,rn spc rn 0000nnnn01000010 1 stc r0_bank,rn r0_bank rn 0000nnnn10000010 1 stc r1_bank,rn r1_bank rn 0000nnnn10010010 1 stc r2_bank,rn r2_bank rn 0000nnnn10100010 1 stc r3_bank,rn r3_bank rn 0000nnnn10110010 1 74 table 2.25 system control instructions (cont) instruction operation code privileged mode cycles t bit stc r4_bank,rn r4_bank rn 0000nnnn11000010 1 stc r5_bank,rn r5_bank rn 0000nnnn11010010 1 stc r6_bank,rn r6_bank rn 0000nnnn11100010 1 stc r7_bank,rn r7_bank rn 0000nnnn11110010 1 stc.l sr,@ C rn rnC4 rn, sr (rn) 0100nnnn00000011 2 stc.l gbr,@ C rn rnC4 rn, gbr (rn) 0100nnnn00010011 2 stc.l vbr,@ C rn rnC4 rn, vbr (rn) 0100nnnn00100011 2 stc.l ssr,@ C rn rnC4 rn, ssr (rn) 0100nnnn00110011 2 stc.l spc,@ C rn rnC4 rn, spc (rn) 0100nnnn01000011 2 stc.l r0_bank, @ C rn rnC4 rn, r0_bank (rn) 0100nnnn10000011 2 stc.l r1_bank, @ C rn rnC4 rn, r1_bank (rn) 0100nnnn10010011 2 stc.l r2_bank, @ C rn rnC4 rn, r2_bank (rn) 0100nnnn10100011 2 stc.l r3_bank, @ C rn rnC4 rn, r3_bank (rn) 0100nnnn10110011 2 stc.l r4_bank, @ C rn rnC4 rn, r4_bank (rn) 0100nnnn11000011 2 stc.l r5_bank, @ C rn rnC4 rn, r5_bank (rn) 0100nnnn11010011 2 stc.l r6_bank, @ C rn rnC4 rn, r6_bank (rn) 0100nnnn11100011 2 stc.l r7_bank, @ C rn rnC4 rn, r7_bank (rn) 0100nnnn11110011 2 sts mach,rn mach rn 0000nnnn00001010 1 sts macl,rn macl rn 0000nnnn00011010 1 sts pr,rn pr rn 0000nnnn00101010 1 sts.l mach,@ C rn rnC4 rn, mach (rn) 0100nnnn00000010 1 sts.l macl,@ C rn rnC4 rn, macl (rn) 0100nnnn00010010 1 sts.l pr,@ C rn rnC4 rn, pr (rn) 0100nnnn00100010 1 trapa #imm pc spc, sr ssr, imm << 2 tra, vbr + h '0100 pc 11000011iiiiiiii 8 75 note: * number of states before the chip enters the sleep state. the table shows the minimum number of clocks required for execution. in practice, the number of execution cycles will be increased if there is contention between an instruction fetch and a data access, or if the destination register of a load instruction (memory register) is also used by the following instruction. 2.6 dsp extended-function instructions 2.6.1 introduction the newly added instructions are classified into the following three groups: 1. additional system control instructions for the cpu unit 2. dsp unit memory-register single and double data transfer 3. dsp unit parallel processing group 1 instructions are provided to support loop control and data transfer between cpu core registers or memory and new control registers added to the cpu core. dsp operations employ a multi-level nested-loop structure. with a single-level loop, use of the decrement and test, dtrn, and conditional delayed branch bf/s instructions supported by the sh-3 is adequate. however, with nested loops, dsp performance can be improved by means of a zero-overhead loop control function. the rs, re, and mod registers have been added to support loop control and modulo addressing functions. instructions are supported for data transfer between these new control registers and general registers or memory. in addition, the ldrs and ldre address calculation registers have been added to reduce the code size for the initial settings for zero-overhead loop control. an independent control register, dsr, is provided for the dsp engine. this register is treated as a system register such as macl and mach. the a0, x0, x1, y0, and y1 registers are treated as system registers from the cpu side, and lds/sts instructions are supported for the same purpose. table 2.26 shows the instruction code map for the new system control instructions for the cpu core. group 2 instructions are provided to reduce dsp operation program code size. data transfer instructions that perform no data processing are frequently executed by the dsp engine. in this case, a 32-bit instruction code is unnecessarily long, and wastes space in the program memory area. all instructions in this class have a 16-bit code length, the same as conventional sh core instructions. single data transfer instructions have greater flexibility in terms of operands than the double data transfer instruction or parallel instruction class. group 3 instructions are provided for fast execution of digital signal processing operations using the dsp unit. these instructions have a 32-bit instruction code, so that a maximum of four 76 instructionsan alu operation, multiplication, and two data transfer instructionscan be executed in parallel. 2.6.2 added cpu system control instructions the new instructions in this class are treated as part of the cpu core functions, and therefore all the added instructions have a 16-bit code length. all the additional instructions belong to the system control instruction group. table 2.26 summarizes the added system instructions. new control registersrs, re, and modhave been added to the cpu core to support loop control and modulo addressing functions, and ldc and sts type instructions have been provided for these registers. the dsp engines dsr, a0, x0, x1, y0, and y1 registers are treated as system registers such as mach and macl, and therefore sts and lds instructions are supported for these registers. as digital signal processing operations usually employ a multi-level nested-loop structure, dsp performance can be improved by means of a zero-overhead loop control function. setrc type instructions are provided to set the repeat count in the rc field in sr[27:16]. when an immediate operand type setrc instruction is executed, the 8-bit immediate operand data is set in sr[23:16], and 0 is set in the remaining bits, sr[27:24]. when a register operand type setrc instruction is executed, rn[11:0] is set in sr[27:16]. the start address and end address of the repeat loop are set in the rs register and re register. there are two ways of setting the addresses: by using an ldc type instruction, or by using the ldrs and ldre instructions. 77 table 2.26 added cpu system control instructions instruction instruction code operation execu- tion states t bit setrc #imm 10000010iiiiiiii imm rc (of sr) 3 setrc rn 0100nnnn00010100 rn[11:0] r c (of sr) 3 ldrs @(disp,pc) 10001100dddddddd (disp 2 + pc) rs 3 ldre @(disp,pc) 10001110dddddddd (disp 2 + pc) re 3 stc mod,rn 0000nnnn01010010 mod rn 1 stc rs,rn 0000nnnn01100010 rs rn 1 stc re,rn 0000nnnn01110010 re rn 1 sts dsr,rn 0000nnnn01101010 dsr rn 1 sts a0,rn 0000nnnn01111010 a0 rn 1 sts x0,rn 0000nnnn10001010 x0 rn 1 sts x1,rn 0000nnnn10011010 x1 rn 1 sts y0,rn 0000nnnn10101010 y0 rn 1 sts y1,rn 0000nnnn10111010 y1 rn 1 sts.l dsr,@-rn 0100nnnn01100010 rn C 4 rn, dsr (rn) 1 sts.l a0,@-rn 0100nnnn01110010 rn C 4 rn, a0 (rn) 1 sts.l x0,@-rn 0100nnnn10000010 rn C 4 rn, x0 (rn) 1 sts.l x1,@-rn 0100nnnn10010010 rn C 4 rn, x1 (rn) 1 sts.l y0,@-rn 0100nnnn10100010 rn C 4 rn, y0 (rn) 1 sts.l y1,@-rn 0100nnnn10110010 rn C 4 rn, y1 (rn) 1 stc.l mod,@-rn 0100nnnn01010011 rn C 4 rn, mod (rn) 2 stc.l rs,@-rn 0100nnnn01100011 rn C 4 rn, rs (rn) 2 stc.l re,@-rn 0100nnnn01110011 rn C 4 rn, re (rn) 2 lds.l @rn+,dsr 0100nnnn01100110 (rn) dsr, rn + 4 rn 1 lds.l @rn+,a0 0100nnnn01110110 (rn) a0, rn + 4 rn 1 lds.l @rn+,x0 0100nnnn10000110 (rn) x0, rn + 4 rn 1 lds.l @rn+,x1 0100nnnn10010110 (rn) x1, rn + 4 rn 1 lds.l @rn+,y0 0100nnnn10100110 (rn) y0, rn + 4 rn 1 lds.l @rn+,y1 0100nnnn10110110 (rn) y1, rn + 4 rn 1 ldc.l @rn+,mod 0100nnnn01010111 (rn) mod, rn + 4 rn 5 ldc.l @rn+,rs 0100nnnn01100111 (rn) rs, rn + 4 rn 5 78 table 2.26 added cpu system control instructions (cont) instruction instruction code operation execu- tion states t bit ldc.l @rn+,re 0100nnnn01110111 (rn) re, rn + 4 rn 5 lds rn,dsr 0100nnnn01101010 rn dsr 1 lds rn,a0 0100nnnn01111010 rn a0 1 lds rn,x0 0100nnnn10001010 rn x0 1 lds rn,x1 0100nnnn10011010 rn x1 1 lds rn,y0 0100nnnn10101010 rn y0 1 lds rn,y1 0100nnnn10111010 rn y1 1 ldc rn,mod 0100nnnn01011110 rn mod 3 ldc rn,rs 0100nnnn01101110 rn rs 3 ldc rn,re 0100nnnn01111110 rn re 3 2.6.3 single and double data transfer for dsp data instructions the new instructions in this class are provided to reduce the program code size for dsp operations. all the new instructions in this class have a 16-bit code length. instructions in this class are divided into two groups: single data transfer instructions and double data transfer instructions. the operand flexibility of the double data transfer instructions is the same as with the a field in parallel instruction class data transfer instructions described in section 4.3.4. however, conditional load instructions cannot be used with these 16-bit instructions. in single transfer, the ax pointer and two other pointers are used as the as pointer, but the ay pointer is not used. tables 2.26 and 2.27 list the single and double data transfer instructions. with double data transfer group instructions, x memory and y memory can be accessed in parallel. the ax pointer can only be used by x memory access instructions, and the ay pointer only by y memory access instructions. double data transfer instructions can only access the on- chip x and y memory areas. single data transfer instructions use a 16-bit instruction code, and can access any memory address space. rn (n = 2 to 7) registers are normally used as the ax, ay, and as pointers. the pointer names themselves can be changed with the assembler rename function. the following renaming scheme is recommended. r2:as2, r3:as3, r4:ax0 (as0), r5:ax1 (as1), r6:ay0, r7:ay1, r8:ix, r9:iy 79 table 2.27 double data transfer instructions instruction instruction code operation execu- tion states dc x memory nopx 1111000*0*0*00** x memory no access 1 data transfer movx.w @ax,dx 111100a*d*0*01** (ax) msw of dx, 0 lsw of dx 1 movx.w @ax+,dx 111100a*d*0*10** (ax) msw of dx, 0 lsw of dx, ax + 2 ax 1 movx.w @ax+ix,dx 111100a*d*0*11** (ax) msw of dx, 0 lsw of dx, ax + ix ax 1 movx.w da,@ax 111100a*d*1*01** msw of da (ax) 1 movx.w da,@ax+ 111100a*d*1*10** msw of da (ax), ax + 2 ax 1 movx.w da,@ax+ix 111100a*d*1*11** msw of da (ax), ax + ix ax 1 y memory nopy 111100*0*0*0**00 y memory no access 1 data transfer movy.w @ay,dy 111100*a*d*0**01 (ay) msw of dy, 0 lsw of dy 1 movy.w @ay+,dy 111100*a*d*0**10 (ay) msw of dy, 0 lsw of dy, ay + 2 ay 1 movy.w @ay+iy,dy 111100*a*d*0**11 (ay) msw of dy, 0 lsw of dy, ay + iy ay 1 movy.w da,@ay 111100*a*d*1**01 msw of da (ay) 1 movy.w da,@ay+ 111100*a*d*1**10 msw of da (ay), ay + 2 ay 1 movy.w da,@ay+iy 111100*a*d*1**11 msw of da (ay), ay + iy ay 1 80 table 2.28 single data transfer instructions instruction instruction code operation execu- tion states dc movs.w @-as,ds 111101aadddd0000 as C 2 as, (as) msw of ds, 0 lsw of ds 1 movs.w @as,ds 111101aadddd0100 (as) msw of ds, 0 lsw of ds 1 movs.w @as+,ds 111101aadddd1000 (as) msw of ds, 0 lsw of ds, as + 2 as 1 movs.w @as+ix,ds 111101aadddd1100 (asc) msw of ds, 0 lsw of ds, as + ix as 1 movs.w ds,@-as * 111101aadddd0001 as C 2 as, msw of ds (as) 1 movs.w ds,@as * 111101aadddd0101 msw of ds (as) 1 movs.w ds,@as+ * 111101aadddd1001 msw of ds (as), as + 2 as 1 movs.w ds,@as+ix * 111101aadddd1101 msw of ds (as), as + ix as 1 movs.l @-as,ds 111101aadddd0010 as C 4 as, (as) ds 1 movs.l @as,ds 111101aadddd0110 (as) ds 1 movs.l @as+,ds 111101aadddd1010 (as) ds, as + 4 as 1 movs.l @as+ix,ds 111101aadddd1110 (as) ds, as + ix as 1 movs.l ds,@-as 111101aadddd0011 as C 4 as, ds (as) 1 movs.l ds,@as 111101aadddd0111 ds (as) 1 movs.l ds,@as+ 111101aadddd1011 ds (as), as + 4 as 1 movs.l ds,@as+ix 111101aadddd1111 ds (as), as + ix as 1 note: * if guard bit registers a0g and a1g are specified in source operand ds, the data is output to the ldb[7:0] bus and the sign bit is copied into the upper bits, [31:8]. 81 the correspondence between dsp data transfer operands and registers is shown in table 2.29. cpu core registers are used as a pointer address that indicates a memory address. table 2.29 correspondence between dsp data transfer operands and registers register ax ix dx ay iy dy da as ds sh r0 register r1 r2 (as2) yes r3 (as3) yes r4 (ax0) yes yes r5 (ax1) yes yes r6 (ay0) yes r7 (ay1) yes r8 (ix) yes r9 (iy) yes dsp a0 yes yes register a1 yes yes m0 yes m1 yes x0 yes yes x1 yes yes y0 yes yes y1 yes yes a0g yes a1g yes 82 2.6.4 dsp operation instruction set dsp operation instructions are instructions for digital signal processing performed by the dsp unit. these instructions have a 32-bit instruction code, and multiple instructions can be executed in parallel. the instruction code is divided into an a field and b field; a parallel data transfer instruction is specified in the a field, and a single or double data operation instruction in the b field. instructions can be specified independently, and are also executed independently. the parallel data transfer instruction specified in the a field is exactly the same as a double data transfer instruction. the function of the a fieldthat is, the data transfer instruction fieldis basically the same as in the double data transfer instructions described in section 2.6.3, single and double data transfer for dsp data instructions, but has a special function in load instructions. b-field data operation instructions are of three kinds: double data operation instructions, conditional single data operation instructions, and unconditional single data operation instructions. the formats of the dsp operation instructions are shown in table 2.30. the respective operands are selected independently from the dsp registers. the correspondence between dsp operation instruction operands and registers is shown in table 2.31. table 2.30 dsp operation instruction formats type instruction formats double data operation instructions aluop. sx, sy, du mltop. se, df, dg conditional single data operation instructions aluop. sx, sy, dz dct aluop. sx, sy, dz dcf aluop. sx, sy, dz aluop. sx, dz dct aluop. sx, dz dcf aluop. sx, dz aluop. sy, dz dct aluop. sy, dz dcf aluop. sy, dz unconditional single data operation instructions aluop. sx, sy, dz aluop. sx, dz aluop. sy, dz mltop. se, sf, dg 83 table 2.31 correspondence between dsp instruction operands and registers alu/bpu operations multiply operations register sx sy dz du se sf dg a0 yes yes yes yes a1 yes yes yes yes yes yes m0 yes yes yes m1 yes yes yes x0 yes yes yes yes yes x1 yes yes yes y0 yes yes yes yes yes y1 yes yes yes when writing parallel instructions, the b-field instruction is written first, followed by the a-field instruction. a sample parallel processing program is shown in figure 2.16. dcf padd a0, m0, a0 pmuls x0, y0, m0 movx.w @r4+, x0 movy.w @r6+, y0 [;] pinc x1, a1 movx.w a0, @r5+r8 movy.w @r7+, y0 [;] pcmp x1, m0 movx.w @r4 [nopy] [;] figure 2.16 sample parallel instruction program square brackets mean that the contents can be omitted. the no operation instructions nopx and nopy can be omitted. table 2.32 gives an overview of the b field in parallel operation instructions. a semicolon is the instruction line delimiter, but this can also be omitted. if the semicolon delimiter is used, the area to the right of the semicolon can be used as a comment field. this has the same function as with conventional sh tools. the dsr register condition code bit (dc) is always updated on the basis of the result of an unconditional alu or shift operation instruction. conditional instructions do not update the dc bit. multiply instructions, also, do not update the dc bit. dc bit updating is performed by means of bits cs0 to cs2 in the dsr register. the dc bit update rules are shown in table 2.33. 84 table 2.32 dsp operation instructions instruction instruction code operation execu- tion states dc pmuls se,sf,dg 111110********** 0100eeff0000gg00 se * sf dg (signed) 1 padd sx,sy,du pmuls se,sf,dg 111110********** 0111eeffxxyygguu sx + sy du se * sf dg (signed) 1 * psub sx,sy,du pmuls se,sf,dg 111110********** 0110eeffxxyygguu sy C sy du se * sf dg (signed) 1 * padd sx,sy,dz 111110********** 10110001xxyyzzzz sx + sy dz 1 * dct padd sx,sy,dz 111110********** 10110010xxyyzzzz if dc = 1, sx + sy dz if dc = 0, nop 1 * dcf padd sx,sy,dz 111110********** 10110011xxyyzzzz if dc = 0, sx + sy dz if dc = 1, nop 1 * psub sx,sy,dz 111110********** 10100001xxyyzzzz sx C sy dz 1 * dct psub sx,sy,dz 111110********** 10100010xxyyzzzz if dc = 1, sx C sy dz if dc = 0, nop 1 * dcf psub sx,sy,dz 111110********** 10100011xxyyzzzz if dc = 0, sx C sy dz if dc = 1, nop 1 * psha sx,sy,dz 111110********** 1010001xxyyzzzz if sy > = 0, sx << sy dz (arithmetic shift) if sy<0, sx>>sy dz 1 * dct psha sx,sy,dz 111110********** 10010010xxyyzzzz if dc = 1 & sy > = 0, sx << sy dz (arithmetic shift) if dc = 1 & sy < 0, sx >> sy dz if dc = 0, nop 1 * note: * see table 2.33. 85 table 2.32 dsp operation instructions (cont) instruction instruction code operation execu- tion states dc dcf psha sx,sy,dz 111110********** 10010011xxyyzzzz if dc = 0 & sy > = 0, sx << sy dz (arithmetic shift) if dc = 0 & sy < 0, sx >> sy dz if dc = 1, nop 1 * pshl sx,sy,dz 111110********** 10000001xxyyzzzz if sy > = 0, sx << sy dz (logical shift) if sy < 0, sx >> sy dz 1 * dct pshl sx,sy,dz 111110********** 10000010xxyyzzzz if dc = 1 & sy > = 0, sx << sy dz (logical shift) if dc = 1 & sy < 0, sx >> sy dz if dc = 0, nop 1 * dcf pshl sx,sy,dz 111110********** 10000011xxyyzzzz if dc = 0 & sy > = 0, sx << sy dz (logical shift) if dc = 0 & sy < 0, sx >> sy dz if dc = 1, nop 1 * pcopy sx,dz 111110********** 11011001xx00zzzz sx dz 1 * pcopy sy,dz 111110********** 1111100100yyzzzz sy dz 1 * dct pcopy sx,dz 111110********** 11011010xx00zzzz if dc = 1, sx dz if dc = 0, nop 1 * dct pcopy sy,dz 111110********** 1111101000yyzzzz if dc = 1, sy dz if dc = 0, nop 1 * dcf pcopy sx,dz 111110********** 11011011xx00zzzz if dc = 0, sx dz if dc = 1, nop 1 * dcf pcopy sy,dz 111110********** 1111101100yyzzzz if dc = 0, sy dz if dc = 1, nop 1 * note: * see table 2.33. 86 table 2.32 dsp operation instructions (cont) instruction instruction code operation execu- tion states dc pdmsb sx,dz 111110********** 10011101xx00zzzz sx dz normalization count shift value 1 * pdmsb sy,dz 111110********** 1011110100yyzzzz sx dz normalization count shift value 1 * dct pdmsb sx,dz 111110********** 10011110xx00zzzz if dc = 1, normalization count shift value sx dz if dc = 0, nop 1 * dct pdmsb sy,dz 111110********** 1011111000yyzzzz if dc = 1, normalization count shift value sy dz if dc = 0, nop 1 * dcf pdmsb sx,dz 111110********** 10011111xx00zzzz if dc = 0, normalization count shift value sx dz if dc = 1, nop 1 * dcf pdmsb sy,dz 111110********** 1011111100yyzzzz if dc = 0, normalization count shift value sy dz if dc = 1, nop 1 * pinc sx,dz 111110********** 10011001xx00zzzz msw of sx dz 1 * pinc sy,dz 111110********** 1011100100yyzzzz msw of sy dz 1 * dct pinc sx,dz 111110********** 10011010xx00zzzz if dc = 1, msw of sx + 1 dz if dc = 0, nop 1 * dct pinc sy,dz 111110********** 1011101000yyzzzz if dc = 1, msw of sy + 1 dz if dc = 0, nop 1 * dcf pinc sx,dz 111110********** 10011011xx00zzzz if dc = 0, msw of sx + 1 dz if dc = 1, nop 1 * dcf pinc sy,dz 111110********** 1011101100yyzzzz if dc = 0, msw of sy + 1 dz if dc = 1, nop 1 * pneg sx,dz 111110********** 11001001xx00zzzz 0 C sx dz 1 * note: * see table 2.33. 87 table 2.32 dsp operation instructions (cont) instruction instruction code operation execu- tion states dc pneg sy,dz 111110********** 1110100100yyzzzz 0 C sy dz 1 * dct pneg sx,dz 111110********** 11001010xx00zzzz if dc = 1, 0 C sx dz if dc = 0, nop 1 * dct pneg sy,dz 111110********** 1110101000yyzzzz if dc = 1, 0 C sy dz if dc = 0, nop 1 * dcf pneg sx,dz 111110********** 11001011xx00zzzz if dc = 0, 0 C sx dz if dc = 1, nop 1 * dcf pneg sy,dz 111110********** 1110101100yyzzzz if dc = 0, 0 C sy dz if dc = 1, nop 1 * por sx,sy,dz 111110********** 10110101xxyyzzzz sx | sy dz 1 * dct por sx,sy,dz 111110********** 10110110xxyyzzzz if dc = 1, sx | sy dz if dc = 0, nop 1 * dcf por sx,sy,dz 111110********** 10110111xxyyzzzz if dc = 0, sx | sy dz if dc = 1, nop 1 * pand sx,sy,dz 111110********** 10010101xxyyzzzz sx & sy dz 1 * dct pand sx,sy,dz 111110********** 10010110xxyyzzzz if dc = 1, sx & sy dz if dc = 0, nop 1 * dcf pand sx,sy,dz 111110********** 10010111xxyyzzzz if dc = 0, sx & sy dz if dc = 1, nop 1 * pxor sx,sy,dz 111110********** 10100101xxyyzzzz sx ^ sy dz 1 * dct pxor sx,sy,dz 111110********** 10100110xxyyzzzz if dc = 1, sx ^ sy dz if dc = 0, nop 1 * dcf pxor sx,sy,dz 111110********** 10100111xxyyzzzz if dc = 1, sx ^ sy dz if dc = 0, nop 1 * pdec sx,dz 111110********** 10001001xx00zzzz sx [39:16] C 1 dz 1 * note: * see table 2.33. 88 table 2.32 dsp operation instructions (cont) instruction instruction code operation execu- tion states dc pdec sy,dz 111110********** 1010100100yyzzzz sy [31:16] C 1 dz 1 * dct pdec sx,dz 111110********** 10001010xx00zzzz if dc = 1, sx [39:16] C 1 dz if dc = 0, nop 1 * dct pdec sy,dz 111110********** 1010101000yyzzzz if dc = 1, sy [31:16] C 1 dz if dc = 0, nop 1 * dcf pdec sx,dz 111110********** 10001011xx00zzzz if dc = 0, sx [39:16] C 1 dz if dc = 1, nop 1 * dcf pdec sy,dz 111110********** 1010101100yyzzzz if dc = 0, sy [31:16] C 1 dz if dc = 1, nop 1 * pclr dz 111110********** 100011010000zzzz h'00000000 dz 1 * dct pclr dz 111110********** 100011100000zzzz if dc = 1, h'00000000 dz if dc = 0, nop 1 * dcf pclr dz 111110********** 100011110000zzzz if dc = 0, h'00000000 dz if dc = 1, nop 1 * psha #imm,dz 111110********** 00010iiiiiiizzzz if imm > = 0, dz << imm dz (arithmetic shift) if imm<0, dz>>imm dz 1 * pshl #imm,dz 111110********** 00000iiiiiiizzzz if imm > = 0, dz << imm dz (logical shift) if imm < 0, dz >> imm dz 1 * psts mach,dz 111110********** 110011010000zzzz mach dz 1 dct psts mach,dz 111110********** 110011100000zzzz if dc = 1, mach dz 1 dcf psts mach,dz 111110********** 110011110000zzzz if dc = 0, mach dz 1 psts macl,dz 111110********** 110111010000zzzz macl dz 1 note: * see table 2.33. 89 table 2.32 dsp operation instructions (cont) instruction instruction code operation execu- tion states dc dct psts macl,dz 111110********** 110111100000zzzz if dc = 1, macl dz 1 dcf psts macl,dz 111110********** 110111110000zzzz if dc = 0, macl dz 1 plds dz,mach 111110********** 111011010000zzzz dz mach 1 dct plds dz,mach 111110********** 111011100000zzzz if dc = 1, dz mach 1 dcf plds dz,mach 111110********** 111011110000zzzz if dc = 0, dz mach 1 plds dz,macl 111110********** 111111010000zzzz dz macl 1 dct plds dz,macl 111110********** 111111100000zzzz if dc = 1, dz macl 1 dcf plds dz,macl 111110********** 111111110000zzzz if dc = 0, dz macl 1 paddc sx,sy,dz 111110********** 10110000xxyyzzzz sx + sy + dc dz carry dc 1 carry psubc sx,sy,dz 111110********** 10100000xxyyzzzz sx C sy C dc dz borrow dc 1 borrow pcmp sx,sy 111110********** 10000100xxyy0000 sx C sy dc update * 1 * pabs sx,dz 111110********** 10001000xx00zzzz if sx < 0, 0 C sx dz if sx > = 0, nop 1 * pabs sy,dz 111110********** 1010100000yyzzzz if sy < 0, 0 C sy dz if sx > = 0, nop 1 * prnd sx,dz 111110********** 10011000xx00zzzz sx + h'00008000 dz lsw of dz h'0000 1 * prnd sy,dz 111110********** 1011100000yyzzzz sy + h'00008000 dz lsw of dz h'0000 1 * note: * see table 2.33. 90 table 2.33 dc bit update definitions cs [2:0] condition mode description 0 0 0 carry or borrow mode the dc bit is set if an alu arithmetic operation generates a carry or borrow, and is cleared otherwise. when a psha or pshl shift instruction is executed, the last bit data shifted out is copied into the dc bit. when an alu logical operation is executed, the dc bit is always cleared. 0 0 1 negative value mode when an alu or shift (psha) arithmetic operation is executed, the msb of the result, including the guard bits, is copied into the dc bit. when an alu or shift (pshl) logical operation is executed, the msb of the result, excluding the guard bits, is copied into the dc bit. 0 1 0 zero value mode the dc bit is set if the result of an alu or shift operation is all- zeros, and is cleared otherwise. 0 1 1 overflow mode the dc bit is set if the result of an alu or shift (psha) arithmetic operation exceeds the destination register range, excluding the guard bits, and is cleared otherwise. when an alu or shift (pshl) logical operation is executed, the dc bit is always cleared. 1 0 0 signed greater-than mode this mode is similar to signed greater-or-equal mode, but dc is cleared if the result is all-zeros. dc = ~{(negative value ^ over-range) | zero value}; in case of arithmetic operation dc = 0; in case of logical operation 1 0 1 signed greater-or- equal mode if the result of an alu or shift (psha) arithmetic operation exceeds the destination register range, including the guard bits (over- range), the definition is the same as in negative value mode. if the result is not over-range, the definition is the opposite of that in negative value mode. when an alu or shift (pshl) logical operation is executed, the dc bit is always cleared. dc = ~(negative value ^ over-range); in case of arithmetic operation dc = 0 ; in case of logical operation 1 1 0 reserved 1 1 1 reserved 91 conditional operations and data transfer: some instructions belonging to this class can be executed conditionally, as described earlier. the specified condition is valid only for the b field of the instruction, and is not valid for data transfer instructions for which a parallel specification is made. examples are shown in figure 2.17. dct padd x0,y0,a0 movx.w @r4+,x0 movy.w a0,@r6+r9 ; when condition is true before execution: after execution: x0=h'33333333, y0=h'55555555, a0=h'123456789a, r4=h'00008000, r6=h'00008233, r9=h'00000004 (r4)=h'1111, (r6)=h'2222 x0=h'11110000, y0=h'55555555, a0=h'0088888888, r4=h'00008002, r6=h'00008237, r9=h'00000004 (r4)=h'1111, (r6)=h'3456 when condition is false before execution: after execution: x0=h'33333333, y0=h'55555555, a0=h'123456789a, r4=h'00008000, r6=h'00008233, r9=h'00000004 (r4)=h'1111, (r6)=h'2222 x0=h'11110000, y0=h'55555555, a0=h'123456789a, r4=h'00008002, r6=h'00008237, r9=h'00000004 (r4)=h'1111, (r6)=h'3456 figure 2.17 examples of conditional operations and data transfer instructions assignment of nopx and nopy instruction codes: when there is no data transfer instruction to be parallel-processed simultaneously with a dsp operation instruction, an nopx or nopy instruction can be written as the data transfer instruction, or the instruction can be omitted. the instruction code is the same whether an nopx or nopy instruction is written or the instruction is omitted. examples of nopx and nopy instruction codes are shown in table 2.34. 92 table 2.34 examples of nopx and nopy instruction codes instruction code padd x0,y0,a0 movx.w @r4+,x0 movy.w @r6+r9,y0 1111100000001011 1011000100000111 padd x0,y0,a0 nopx movy.w @r6+r9,y0 1111100000000011 1011000100000111 padd x0,y0,a0 nopx nopy 1111100000000000 1011000100000111 padd x0,y0,a0 nopx 1111100000000000 1011000100000111 padd x0,y0,a0 1111100000000000 1011000100000111 movx.w @r4+,x0 movy.w @r6+r9,y0 1111000000001011 movx.w @r4+,x0 nopy 1111000000001000 movs.w @r4+,x0 1111010010001000 nopx movy.w @r6+r9,y0 1111000000000011 movy.w @r6+r9,y0 1111000000000011 nopx nopy 1111000000000000 nop 0000000000001001 93 section 3 memory management unit (mmu) 3.1 overview 3.1.1 features the sh7729r has an on-chip memory management unit (mmu) that implements address translation. the sh7729r features a resident translation look-aside buffer (tlb) that caches information for user-created address translation tables located in external memory. it enables high- speed translation of virtual addresses into physical addresses. address translation uses the paging system and supports two page sizes (1 kbyte and 4 kbytes). the access right to virtual address space can be set for privileged and user modes to provide memory protection. 3.1.2 role of mmu the mmu is a feature designed to make efficient use of physical memory. as shown in figure 3.1, if a process is smaller in size than the physical memory, the entire process can be mapped onto physical memory. however, if the process increases in size to the extent that it no longer fits into physical memory, it becomes necessary to partition the process and to map those parts requiring execution onto memory as occasion demands ((1)). having the process itself consider this mapping onto physical memory would impose a large burden on the process. to lighten this burden, the idea of virtual memory was born as a means of performing en bloc mapping onto physical memory ((2)). in a virtual memory system, substantially more virtual memory than physical memory is provided, and the process is mapped onto this virtual memory. thus a process only has to consider operation in virtual memory. mapping from virtual memory to physical memory is handled by the mmu. the mmu is normally controlled by the operating system, switching physical memory to allow the virtual memory required by a process to be mapped onto physical memory in a smooth fashion. switching of physical memory is carried out via secondary storage, etc. the virtual memory system that came into being in this way is particularly effective in a time- sharing system (tss) in which a number of processes are running simultaneously ((3)). if processes running in a tss had to take mapping onto virtual memory into consideration while running, it would not be possible to increase efficiency. virtual memory is thus used to reduce this load on the individual processes and so improve efficiency ((4)). in the virtual memory system, virtual memory is allocated to each process. the task of the mmu is to perform efficient mapping of these virtual memory areas onto physical memory. it also has a memory protection feature that prevents one process from inadvertently accessing another processs physical memory. when address translation from virtual memory to physical memory is performed using the mmu, it may occur that the relevant translation information is not recorded in the mmu, with the result that one process may inadvertently access the virtual memory allocated to another process. in this 94 case, the mmu will generate an exception, change the physical memory mapping, and record the new address translation information. although the functions of the mmu could also be implemented by software alone, the need for translation to be performed by software each time a process accesses physical memory would result in poor efficiency. for this reason, a buffer for address translation (translation look-aside buffer: tlb) is provided in hardware to hold frequently used address translation information. the tlb can be described as a cache for storing address translation information. unlike cache memory, however, if address translation fails, that is, if an exception is generated, switching of address translation information is normally performed by software. this makes it possible for memory management to be performed flexibly by software. the mmu has two methods of mapping from virtual memory to physical memory: a paging method using fixed-length address translation, and a segment method using variable-length address translation. with the paging method, the unit of translation is a fixed-size address space (usually of 1 to 64 kbytes) called a page. in the following text, the sh7729r address space in virtual memory is referred to as virtual address space, and address space in physical memory as physical memory space. 95 process 1 physical memory mmu (1) (2) (3) (4) process 1 physical memory process 1 virtual memory mmu physical memory process 1 process 2 process 3 physical memory process 1 process 2 process 3 virtual memory physical memory figure 3.1 mmu functions 96 3.1.3 sh7729r mmu virtual address space: the sh7729r uses 32-bit virtual addresses to access a 4-gbyte virtual address space that is divided into several areas. address space mapping is shown in figure 3.2. in privileged mode, there are five areas, p0Cp4. the p0 and p3 areas are mapped onto physical address space in page units, in accordance with address translation table information. write-back or write-through can be selected for write access by means of a cache control register (ccr) setting. mapping of the p1 area is fixed in physical address space (h'00000000 to h'1fffffff). in the p1 area, setting a virtual address msb (bit 31) to 0 generates the corresponding physical address. p1 area accesses can be cached, and the cache control register (ccr) is set to indicate whether to cache or not. write-back or write-through mode can be selected. mapping of the p2 area is fixed in physical address space (h'00000000 to h'1fffffff). in the p2 area, setting the top three virtual address bits (bits 31, 30, and 29) to 0 generates the corresponding physical address. p2 area access cannot be cached. the p1 and p2 areas are not mapped by the address translation table, so the tlb is not used and no exceptions such as tlb misses occur. initialization of mmu-related registers, exception handling, and the like are located in the p1 and p2 areas. because the p1 area is cached, handlers that require high-speed processing are placed there. some peripheral module control registers are located in area 1 of the physical address space. when the physical address space is not used for address translation, these registers should be located in the p2 area. when address translation is to be used, set no caching. the p4 area is used for mapping on-chip control register addresses. in user mode, 2 gbytes of the virtual address space from h'00000000 to h'7fffffff (area u0) can be accessed. u0 is mapped onto physical address space in page units, in accordance with address translation table information. when the dsp bit in cpu status register (sr) is off, 2 gbytes of the virtual address space from h'80000000 to h'ffffffff cannot be accessed in the user mode. attempting to do so creates an address error. write-back or write-through mode can be selected for write accesses by means of a cache control register (ccr) setting. when the dsp bit in cpu status register (sr) is on, a new 16-mbyte address space, uxy, is defined from address h'a5000000 to h'a5ffffff for x/y ram. this uxy space is non-cached, fixed physical address space. any access to address space beyond u0 and uxy creates an address error. for details of the x/y ram space, refer to section 6, x/y memory. 97 h'80000000 h'a0000000 h'c0000000 h'e0000000 h'ffffffff 2-gbyte physical space, cacheable (write-back/write-through) 2-gbyte physical space, cacheable (write-back/write-through) address error h'00000000 h'00000000 h'80000000 h'ffffffff area p0 area p1 area p2 area p3 area p4 area u0 privileged mode user mode 0.5-gbyte fixed physical space, cacheable (write-back/write-through) 0.5-gbyte fixed physical space, non-cacheable 0.5-gbyte physical space, cacheable (write-back/write-through) 0.5-gbyte control space, non-cacheable address error area uxy (exists only * 3 sr.dsp= 1) figure 3.2 virtual address space mapping physical address space: the sh7729r supports a 32-bit physical address space, but the upper 3 bits are actually ignored and treated as a shadow. see section 11, bus state controller (bsc), for details. address translation: when the mmu is enabled, the virtual address space is divided into units called pages. physical addresses are translated in page units. address translation tables in external memory hold information such as the physical address that corresponds to the virtual address and memory protection codes. when an access to an area other than p4 occurs, if the accessed virtual address belongs to area p1 or p2 there is no tlb access and the physical address is uniquely defined. if it belongs to area p0, p3 or u0, the tlb is searched by virtual address and, if that virtual address is registered in the tlb, the access hits the tlb. the corresponding physical address and the page control information are read from the tlb and the physical address is determined. 98 if the virtual address is not registered in the tlb, a tlb miss exception occurs and processing will shift to the tlb miss handler. in the tlb miss handler, the tlb address translation table in external memory is searched and the corresponding physical address and the page control information are registered in the tlb. after returning from the handler, the instruction that caused the tlb miss is re-executed. when the mmu is enabled, address translation information that results in a physical address space of h'80000000Ch'ffffffff should not be registered in the tlb. when the mmu is disabled, the virtual address is used directly as the physical address. as the sh7729r supports a 29-bit address space as the physical address space, the top 3 bits of the physical address are ignored, and constitute a shadow space (see section 11, bus state controller (bsc)). for example, addresses h'00001000 in the p0 area, h'80001000 in the p1 area, h'a0001000 in the p2 area, and h'c0001000 in the p3 area are all mapped onto the same physical address. when access to these addresses is performed with the cache enabled, an address with the top 3 bits of the physical address masked to 0 is stored in the cache address array to ensure data congruity. single virtual memory mode and multiple virtual memory mode: there are two virtual memory modes: single virtual memory mode and multiple virtual memory mode. in single virtual memory mode, multiple processes run in parallel using the virtual address space exclusively and the physical address corresponding to a given virtual address is specified uniquely. in multiple virtual memory mode, multiple processes run in parallel sharing the virtual address space, so a given virtual address may be translated into different physical addresses depending on the process. single or multiple virtual mode is selected by a value set in the mmu control register (mmucr). in terms of operation, the only difference between single virtual memory mode and multiple virtual memory mode is in the tlb address comparison method (see section 3.3.3, tlb address comparison). address space identifier (asid): in multiple virtual memory mode, the address space identifier (asid) is used to differentiate between processes running in parallel and sharing virtual address space. the asid is 8 bits in length and can be set by software setting of the asid of the currently running process in the page table entry register high (pteh) within the mmu. when the process is switched using the asid, the tlb does not have to be purged. in single virtual memory mode, the asid is used to provide memory protection for processes running simultaneously and using the virtual address space exclusively (see section 3.4.2, mmu software management). 99 3.1.4 register configuration table 3.1 shows the configuration of the mmu control registers. table 3.1 register configuration name abbreviation r/w size initial value* 1 address page table entry register high pteh r/w longword undefined h'fffffff0 page table entry register low ptel r/w longword undefined h'fffffff4 translation table base register ttb r/w longword undefined h'fffffff8 tlb exception address register tea r/w longword undefined h'fffffffc mmu control register mmucr r/w longword * 2 h'ffffffe0 notes: *1 initialized by a power-on reset or manual reset. *2 sv bit: undefined other bits: 0 3.2 register description there are five registers for mmu processing. these are all peripheral module registers, so they are located in address space area p4 and can only be accessed from privileged mode by specifying the address. these registers consist of: 1.the page table entry register high (pteh) register residing at address h'fffffff0, which consists of a virtual page number (vpn) and asid. the vpn set is the vpn of the virtual address at which the exception is generated in case of an mmu exception or address error exception. when the page size is 4 kbytes, the vpn is the upper 20 bits of the virtual address, but in this case the upper 22 bits of the virtual address are set. the vpn can also be modified by software. as the asid, software sets the number of the currently executing process. the vpn and asid are recorded in the tlb by the ldtlb instruction. 2.the page table entry register low (ptel) register residing at address h'fffffff4, and used to store the physical page number and page management information to be recorded in the tlb by the ldtlb instruction. the contents of this register are only modified in response to a software command. 3.the translation table base register (ttb) residing at address h'fffffff8, which points to the base address of the current page table. the software does not set any value in ttb automatically. ttb is available to software for general purposes. 100 4.the tlb exception address register (tea) residing at address h'fffffffc, which stores the virtual address corresponding to a tlb or address error exception. this value remains valid until the next exception or interrupt. 5.the mmu control register (mmucr) residing at address h'ffffffe0, which makes the mmu settings described in figure 3.3. any program that modifies mmucr should reside in the p1 or p2 area. the mmu registers are shown in figure 3.3. 31 7 vpn pteh ptel asid 0 ppn 0 10 31 6 4 3 2 1 0 10 31 ttb ttb 0 31 virtual address causing mmu exception or address error exception tea 0 pr sz c d 8 9 7 v 00 sh 0 mmucr 0 31 8 4 65 73210 sv rc 00 0 tf ix at 0: reserved bits. always read as 0. writing is ignored. however, 0 should also be specified in a write to mmucr only. sv: single virtual memory mode bit. set to 1 for the single virtual memory mode, cleared to 0 for the multiple virtual memory mode. rc: a 2-bit random counter, automatically updated by hardware according to the following rules in the event of an mmu exception. when a tlb miss exception occurs, all tlb entry ways corresponding to the virtual address at which the exception occurred are checked, and if all ways are valid, 1 is added to rc; if there is one or more invalid way, they are set by priority from way 0, in the order: way 0, way 1, way 2, way 3. in the event of an mmu exception other than a tlb miss exception, the way which caused the exception is set in rc. tf: tlb flush bit. write 1 to flush the tlb (clear all valid bits of the tlb to 0). always reads 0. ix: index mode bit. when 0, vpn bits 16C12 are used as the tlb index number. when 1, the value obtained by ex-oring asid bits 4C0 in pteh and vpn bits 16C12 is used as the tlb index number. at: address translation bit. enables/disables the mmu. 0: mmu disabled 1: mmu enabled figure 3.3 mmu register contents 101 3.3 tlb functions 3.3.1 configuration of the tlb the tlb caches address translation table information located in external memory. the address translation table stores the physical page number translated from the virtual page number and the control information for the page, which is the unit of address translation. figure 3.4 shows the overall tlb configuration. the tlb is 4-way set associative with 128 entries. there are 32 entries for each way. figure 3.5 shows the configuration of virtual addresses and tlb entries. entry 1 address arra y data arra y entry 0 entry 1 entry 31 way 0C3 way 0C3 vpn(11C10) vpn(31C17) asid(7C0) v entry 0 entry 31 ppn(31C10) pr(1C0) sz c d sh figure 3.4 overall configuration of the tlb 102 31 9 vpn virtual address (1-kbyte page) virtual address (4-kbyte page) tlb entry offset vpn vpn (31C17) vpn (11C10) asid sh sz v pr ppn c d offset 0 10 31 11 0 (15) (2) (2) (22) (8) (1) (1) (1) (1) (1) 12 legend vpn: virtual page number. upper 22 bits of virtual address for a 1-kbyte page, or upper 20 bits of virtual address for a 4-kbyte page. since vpn bits 16-12 are used as the index number, they are not stored in the tlb entry. asid: address space identifier. indicates the process that can access a virtual page. in single virtual memory mode and user mode, or in multiple virtual memory mode, if the sh bit is 0, the address is compared with the asid in pteh when address comparison is performed. sh: share status bit 0 = page not shared between processes 1 = page shared between processes sz: page-size bit 0 = 1-kbyte page 1 = 4-kbyte page v: valid bit. indicates whether entry is valid. 0 = invalid 1 = valid cleared to 0 by a power-on reset. not affected by a manual reset. ppn: physical page number. upper 22 bits of physical address. ppn bits 11-10 are not used in case of a 4-kbyte page. attention must be paid to the synonym problem in case of a 1-kbyte page (see section 3.4.4, avoiding synonym problems). set the most significant bit to 0. pr: protection key field. 2-bit field encoded to define the access rights to the page. 00: reading only is possible in privileged mode. 01: reading/writing is possible in privileged mode. 10: reading only is possible in privileged/user mode. 11: reading/writing is possible in privileged/user mode. c: cacheable bit. indicates whether the page is cacheable. 0: non-cacheable 1: cacheable d: dirty bit. indicates whether the page has been written to. 0 = not written to 1 = written to figure 3.5 virtual address and tlb structure 103 3.3.2 tlb indexing the tlb uses a 4-way set associative scheme, so entries must be selected by index. vpn bits 16 to 12 and asid bits 4 to 0 in pteh are used as the index number. the index number can be generated in two different ways depending on the setting of the ix bit in mmucr. 1.when ix = 0, vpn bits 16C12 alone are used as the index number 2.when ix = 1, vpn bits 16C12 are ex-ored with asid bits 4C0 to generate the index number the second method is used to prevent lowered tlb efficiency that results when multiple processes run simultaneously in the same virtual address space (multiple virtual memory) and a specific entry is selected by indexing of each process. figures 3.6 and 3.7 show the indexing schemes. 31 16 11 12 17 031 0 pteh register virtual address vpn 0asid 7 10 index asid(4C0) exclusive-or way 0 to 3 vpn(31C17) vpn(11C10) asid(7C0) v 0 31 address array data array ppn(31C10) pr(1C0) sz c d sh figure 3.6 tlb indexing (ix = 1) 104 31 16 11 12 17 0 virtual address way 0 to 3 vpn(31C17) vpn(11C10) asid(7C0) v 0 address array data array ppn(31C10) pr(1C0) sz c d sh index 31 figure 3.7 tlb indexing (ix = 0) 3.3.3 tlb address comparison a tlb address comparison is performed when an instruction is fetched from a program in external memory or data in external memory is referenced. the items used in the comparison are vpn and asid. the vpn of the virtual address that accesses external memory is compared to the vpn of the tlb entry selected with the index number. the asid within the pteh is compared to the asid of the indexed tlb entry. all four ways are searched simultaneously. if the compared values match, and the indexed tlb entry is valid (v bit = 1), the hit is registered. it is necessary to have software ensure that tlb hits do not occur simultaneously in more than one way, as hardware operation is not guaranteed if this occurs. for example, if there are two identical tlb entries with the same vpn and a setting is made such that a tlb hit is made only by a process with asid = h'ff when one is in the shared state (sh = 1) and the other in the non-shared state (sh = 0), then if the asid in pteh is set to h'ff, there is a possibility of simultaneous tlb hits in both these ways. it is therefore necessary to ensure that this kind of setting is not made by software. the object compared varies depending on the page management information (sz, sh) in the tlb entry. it also varies depending on whether the system supports multiple virtual memory or single virtual memory. the page-size information determines whether vpn (11C10) is compared. vpn (11C10) is compared for 1-kbyte pages (sz = 0) but not for 4-kbyte pages (sz = 1). 105 the sharing information (sh) determines whether pteh.asid and the asid in the tlb entry are compared. asids are compared when there is no sharing between processes (sh = 0) but not when there is sharing (sh = 1). when single virtual memory is supported (mmucr.sv = 1) and privileged mode is engaged (sr.md = 1), all process resources can be accessed. this means that asids are not compared when single virtual memory is supported and privileged mode is engaged. the objects of address comparison are shown in figure 3.8. bits compared: vpn (31C17) vpn (11C10) sz = 0? yes no yes (1 kbyte) no (4 kbytes) bits compared: vpn (31C17) bits compared: vpn (31C17) vpn (11C10) asid (7C0) sz = 0? yes (1 kbyte) no (4 kbytes) bits compared: vpn (31C17) asid (7C0) sh = 1 or (sr.md = 1 and mmucr.sv = 1)? figure 3.8 objects of address comparison 106 3.3.4 page management information in addition to the sh and sz bits, the page management information of tlb entries also includes d, c, and pr bits. the d bit of a tlb entry indicates whether the page is dirty (i.e., has been written to). if the d bit is 0, an attempt to write to the page results in an initial page write exception. for physical page swapping between secondary memory and main memory, for example, pages are controlled so that a dirty page is paged out of main memory only after that page is written back to secondary memory. to record that there has been a write to a given page in the address translation table in memory, an initial page write exception is used. the c bit in the entry indicates whether the referenced page resides in a cacheable or non- cacheable area of memory. when the control register in area 1 is mapped, set the c bit to 0. the pr field specifies the access rights for the page in privileged and user modes and is used to protect memory. attempts at nonpermitted accesses result in tlb protection violation exceptions. access states designated by the d, c, and pr bits are shown in table 3.2. table 3.2 access states designated by d, c, and pr bits privileged mode user mode reading writing reading writing d bit 0permitted initial page write exception permitted initial page write exception 1 permitted permitted permitted permitted c bit 0permitted (no caching) permitted (no caching) permitted (no caching) permitted (no caching) 1 permitted (with caching) permitted (with caching) permitted (with caching) permitted (with caching) pr bit 00 permitted tlb protection violation exception tlb protection violation exception tlb protection violation exception 01 permitted permitted tlb protection violation exception tlb protection violation exception 10permitted tlb protection violation exception permitted tlb protection violation exception 11 permitted permitted permitted permitted 107 3.4 mmu functions 3.4.1 mmu hardware management there are two kinds of mmu hardware management as follows: 1.the mmu decodes the virtual address accessed by a process and performs address translation by controlling the tlb in accordance with the mmucr settings. 2.in address translation, the mmu receives page management information from the tlb, and determines the mmu exception and whether the cache is to be accessed (using the c bit). for details of the determination method and the hardware processing, see section 3.5, mmu exceptions. 3.4.2 mmu software management there are three kinds of mmu software management, as follows. 1.mmu register setting. mmucr setting, in particular, should be performed in areas p1 and p2 for which address translation is not performed. also, since sv and ix bit changes constitute address translation system changes, in this case, tlb flushing should be performed by simultaneously writing 1 to the tf bit also. since mmu exceptions are not generated in the mmu disabled state with the at bit cleared to 0, use in the disabled state must be avoided with software that does not use the mmu. 2.tlb entry recording, deletion, and reading. tlb entry recording can be done in two ways: by using the ldtlb instruction, or by writing directly to the memory-mapped tlb. for tlb entry deletion and reading, the memory-mapped tlb can be accessed. see section 3.4.3, mmu instruction (ldtlb), for details of the ldtlb instruction, and section 3.6, memory- mapped tlb configuration, for details of the memory-mapped tlb. 3.mmu exception handling. when an mmu exception is generated, it is handled on the basis of information set from the hardware side. see section 3.5, mmu exceptions, for details. when single virtual memory mode is used, it is possible to create a state in which physical memory access is enabled in privileged mode only by clearing the share status bit (sh) to 0 to specify recording of all tlb entries. this strengthens inter-process memory protection, and enables special access levels to be created in privileged mode only. recording a 1-kbyte page tlb entry may result in a synonym problem. see section 3.4.4, avoiding synonym problems. 108 3.4.3 mmu instruction (ldtlb) the load tlb instruction (ldtlb) is used to record tlb entries. when the ix bit in mmucr is 0, the ldtlb instruction changes the tlb entry in the way specified by the rc bit in mmucr to the value specified by pteh and ptel, using vpn bits 16C12 specified in pteh as the index number. when the ix bit in mmucr is 1, the ex-or of vpn bits 16C12 specified in pteh and asid bits 4C0 in pteh is used as the index number. figure 3.9 shows the case where the ix bit in mmucr is 0. when an mmu exception occurs, the virtual page number of the virtual address that caused the exception is set in pteh by hardware. the way is set in the rc bit of mmucr for each exception according to the rules shown in figure 3.9. consequently, if the ldtlb instruction is issued after setting only ptel in the mmu exception handling routine, tlb entry recording is possible. any tlb entry can be updated by software rewriting of pteh and the rc bits in mmucr. as the ldtlb instruction changes address translation information, there is a risk of destroying address translation information if this instruction is issued in the p0, u0, or p3 area. make sure, therefore, that this instruction is issued in the p1 or p2 area. also, an instruction associated with an access to the p0, u0, or p3 area (such as the rte instruction) should be issued at least two instructions after the ldtlb instruction. 109 vpn(31C17) vpn(11C10) asid(7C0) v vpn 0asid vpn 0sv 00rc 0tf ix at ppn 0v 0pr sz c d sh 0 write ppn(31C10) pr(1C0) sz c d sh write data array address array way selection way 0 to 3 31 9 0 mmucr index 31 17 12 108 0 pteh register 31 100 ptel register 0 31 figure 3.9 operation of ldtlb instruction 3.4.4 avoiding synonym problems when a 1-kbyte page is recorded in a tlb entry, a synonym problem may arise. if a number of virtual addresses are mapped onto a single physical address, the same physical address data will be recorded in a number of cache entries, and it will not be possible to guarantee data congruity. the reason why this problem only occurs when using a 1-kbyte page is explained below with reference to figure 3.10. to achieve high-speed operation of the sh7729r cache, an index number is created using virtual address bits 11C4. when a 4-kbyte page is used, virtual address bits 11C4 are included in the offset, and since they are not subject to address translation, they are the same as physical address bits 11C4. in cache-based address comparison and recording in the address array, since the cache tag address is a physical address, physical address bits 31C10 are recorded. when a 1-kbyte page is used, also, a cache index number is created using virtual address bits 11-4. however, in case of a 1-kbyte page, virtual address bits 11 and 10 are subject to address translation and therefore may not be the same as physical address bits 11 and 10. consequently, the physical address is recorded in a different entry from that of the index number indicated by the physical address in the cache address array. 110 note: when multiple address information items use the same physical memory to provide for future expansion of the superh risc engine family, it is recommended that vpn[20:10] be made equal. also, the same physical addresses should not be used with different page size address conversion information. for example, assume that, with 1-kbyte page tlb entries, tlb entries for which the following translation has been performed are recorded in two tlbs: virtual address 1 h'00000000 physical address h'00000c00 virtual address 2 h'00000c00 physical address h'00000c00 virtual address 1 is recorded in cache entry h'00, and virtual address 2 in cache entry h'c0. since two virtual addresses are recorded in different cache entries despite the fact that the physical addresses are the same, memory inconsistency will occur as soon as a write is performed to either virtual address. therefore, when recording a 1-kbyte tlb entry, if the physical address is the same as a physical address already used in another tlb entry, it should be recorded in such a way that physical address bits 11 and 10 are the same. 111 when using a 4-kbyte page virtual address 31 vpn 0 12 11 10 offset physical address 31 ppn 0 offset virtual address (11C4) physical address (31C10) cache address array when using a 1-kbyte page virtual address 31 vpn 0 10 11 offset physical address 31 ppn 0 10 11 offset virtual address (11C4) physical address (31C10) cache address array 9 9 12 11 10 figure 3.10 synonym problem 112 3.5 mmu exceptions there are four mmu exceptions: tlb miss, tlb protection violation, tlb invalid, and initial page write. 3.5.1 tlb miss exception a tlb miss exception occurs when the virtual address and the address array of the selected tlb entry are compared and no match is found. tlb miss exception handling includes both hardware and software operations. hardware operations: in a tlb miss, the sh7729r hardware executes a set of prescribed operations, as follows: 1.the vpn field of the virtual address causing the exception is written to the pteh register. 2.the virtual address causing the exception is written to the tea register. 3.either exception code h'040 for a load access, or h'060 for a store access, is written to the expevt register. 4.the pc value indicating the address of the instruction in which the exception occurred is written to the saved program counter (spc). if the exception occurred in a delay slot, the pc value indicating the address of the related delayed branch instruction is written to spc. 5.the contents of the status register (sr) at the time of the exception are written to the saved status register (ssr). 6.the mode (md) bit in sr is set to 1 to place the sh7729r in privileged mode. 7.the block (bl) bit in sr is set to 1 to mask any further exception requests. 8.the register bank (rb) bit in sr is set to 1. 9.the random counter (rc) field in the mmu control register (mmucr) is incremented by 1 when all ways are checked for the tlb entry corresponding to the virtual address at which the exception occurred, and all ways are valid. if one or more ways are invalid, those ways are set in rc in prioritized order from way 0 through way 1, way 2, and way 3. 10.execution branches to the address obtained by adding the value of the vbr contents and h'00000400 to invoke the user-written tlb miss exception handler. software (tlb miss handler) operations: the software searches the page tables in external memory and allocates the required page table entry. upon retrieving the required page table entry, software must execute the following operations: 1.write the value of the physical page number (ppn) field and the protection key (pr), page size (sz), cacheable (c), dirty (d), share status (sh), and valid (v) bits of the page table entry recorded in the address translation table in external memory into the ptel register in the sh7729r. 113 2.if using software for way selection for entry replacement, write the desired value to the rc field in mmucr. 3.issue an ldtlb instruction to load the contents of pteh and ptel into the tlb. 4.issue an rte (return from exception handler) instruction to terminate the handler and return to the instruction stream. the rte instruction should be issued after two ldtlb instructions. 3.5.2 tlb protection violation exception a tlb protection violation exception occurs when the virtual address and the address array of the selected tlb entry are compared and a valid entry is found to match, but the type of access is not permitted by the access rights specified in the pr field. tlb protection violation exception handling includes both hardware and software operations. hardware operations: in a tlb protection violation exception, the sh7729r hardware executes a set of prescribed operations, as follows: 1.the vpn field of the virtual address causing the exception is written to the pteh register. 2.the virtual address causing the exception is written to the tea register. 3.either exception code h'0a0 for a load access, or h'0c0 for a store access, is written to the expevt register. 4.the pc value indicating the address of the instruction in which the exception occurred is written into spc (if the exception occurred in a delay slot, the pc value indicating the address of the related delayed branch instruction is written into spc). 5.the contents of sr at the time of the exception are written to ssr. 6.the md bit in sr is set to 1 to place the sh7729r in privileged mode. 7.the bl bit in sr is set to 1 to mask any further exception requests. 8.the register bank (rb) bit in sr is set to 1. 9.the way that generated the exception is set in the rc field in mmucr. 10.execution branches to the address obtained by adding the value of the vbr contents and h'00000100 to invoke the tlb protection violation exception handler. software (tlb protection violation handler) operations: software resolves the tlb protection violation and issues an rte (return from exception handler) instruction to terminate the handler and return to the instruction stream. the rte instruction should be issued after two ldtlb instructions. 114 3.5.3 tlb invalid exception a tlb invalid exception occurs when the virtual address is compared to a selected tlb entry address array and a match is found but the entry is not valid (the v bit is 0). tlb invalid exception handling includes both hardware and software operations. hardware operations: in a tlb invalid exception, the sh7729r hardware executes a set of prescribed operations, as follows: 1.the vpn number of the virtual address causing the exception is written to the pteh register. 2.the virtual address causing the exception is written to the tea register. 3.the way number causing the exception is written to rc in mmucr. 4.either exception code h'040 for a load access, or h'060 for a store access, is written to the expevt register. 5.the pc value indicating the address of the instruction in which the exception occurred is written to spc. if the exception occurred in a delay slot, the pc value indicating the address of the delayed branch instruction is written to spc. 6.the contents of sr at the time of the exception are written to ssr. 7.the mode (md) bit in sr is set to 1 to place the sh7729r in privileged mode. 8.the block (bl) bit in sr is set to 1 to mask any further exception requests. 9.the register bank (rb) bit in sr is set to 1. 10.execution branches to the address obtained by adding the value of the vbr contents and h'00000100, and the tlb protection violation exception handler starts. software (tlb invalid exception handler) operations: the software searches the page tables in external memory and assigns the required page table entry. upon retrieving the required page table entry, software must execute the following operations: 1.write the values of the physical page number (ppn) field and the values of the protection key (pr), page size (sz), cacheable (c), dirty (d), share status (sh), and valid (v) bits of the page table entry recorded in the external memory to the ptel register. 2.if using software for way selection for entry replacement, write the desired value to the rc field in mmucr. 3.issue an ldtlb instruction to load the contents of pteh and ptel into the tlb. 4.issue an rte instruction to terminate the handler and return to the instruction stream. the rte instruction should be issued after two ldtlb instructions. 115 3.5.4 initial page write exception an initial page write exception occurs in a write access when the virtual address and the address array of the selected tlb entry are compared and a valid entry with the appropriate access rights is found to match, but the d (dirty) bit of the entry is 0 (the page has not been written to). initial page write exception handling includes both hardware and software operations. hardware operations: in an initial page write exception, the sh7729r hardware executes a set of prescribed operations, as follows: 1.the vpn field of the virtual address causing the exception is written to the pteh register. 2.the virtual address causing the exception is written to the tea register. 3.exception code h'080 is written to the expevt register. 4.the pc value indicating the address of the instruction in which the exception occurred is written to spc. if the exception occurred in a delay slot, the pc value indicating the address of the related delayed branch instruction is written to spc. 5.the contents of sr at the time of the exception are written to ssr. 6.the md bit in sr is set to 1 to place the sh7729r in privileged mode. 7.the bl bit in sr is set to 1 to mask any further exception requests. 8.the register bank (rb) bit in sr is set to 1. 9.the way that caused the exception is set in the rc field in mmucr. 10.execution branches to the address obtained by adding the value of the vbr contents and h'00000100 to invoke the user-written initial page write exception handler. software (initial page write handler) operations: the software must execute the following operations: 1.retrieve the required page table entry from external memory. 2.set the d bit of the page table entry in external memory to 1. 3.write the value of the ppn field and the pr, sz, c, d, sh, and v bits of the page table entry in the external memory to the ptel register. 4.if using software for way selection for entry replacement, write the desired value to the rc field in mmucr. 5.issue an ldtlb instruction to load the contents of pteh and ptel into the tlb. 6.issue an rte instruction to terminate the handler and return to the instruction stream. the rte instruction should be issued after two ldtlb instructions. figure 3.11 shows the flowchart for mmu exceptions. 116 start tlb miss exception initial page write exception pr check pr check yes sh = 0 and (mmucr.sv = 0 or sr.md = 0)? vpns and asids match? vpns match? no yes yes yes yes user or privileged? d = 1? c = 1? v = 1? no no user mode privileged mode no no tlb protection violation exception tlb protection exception cache access w 00/01 10 01/11 00/10 11 ww w rr rr r/w? r/w? r/w? r/w? tlb invalid exception memory access no (noncacheable) yes (cacheable) figure 3.11 mmu exception generation flowchart 117 3.5.5 processing flow in event of mmu exception (same processing flow for address error) mmu exception in instruction fetch mode id ex ma wb id ex ma wb id ex ma wb nop nop if id ex ma wb : exception source stage if id ex ma wb nop mmu exception handler tlb-related exception signals in an instruction fetch handler transition processing = instruction fetch = instruction decode = instruction execution = memory access = write back = no operation if figure 3.12 mmu exception signals in instruction fetch 118 mmu exception in data access mode if id ex if id ex if id id ex ma wb id ex ma wb id ex ma wb nop nop if id ex ma wb : exception source stage : stage cancellation for instruction that has begun execution if id ex ma wb nop = instruction fetch = instruction decode = instruction execution = memory access = write back = no operation mmu exception handler handler transition processing ma wb ma wb ex ma wb tlb-related exception signals in a data access figure 3.13 mmu exception signals in data access 119 3.5.6 mmu exception in repeat loop when an mmu exception or cpu address error occurs immediately before or within a repeat loop, the pc of the instruction that generated the exception cannot be saved in spc correctly and the repeat loop cannot be restarted after returning from the exception handler. expevt is set to h'070 in cases of tlb miss, tlb invalid, and cpu address error. expevt is set to h'0d0 in case of tlb protection violation. figure 3.14 shows where such cases occur. in a repeat loop of 4 or more instructions, only the last 4 instructions are relevant (see figure 3.14 (4)). : exception source stage where spc is not correct and repeat loop can not be restarted (1) 1 instruction repeated (inst1, sr.rc=2) inst-1 inst0 inst1 inst1 inst2 if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb (2) 2 instructions repeated (inst1 and inst2, sr.rc=2) inst-1 inst0 inst1 inst2 inst1 inst2 inst3 if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb (3) 3 instructions repeated (inst1, inst2 and inst3, sr.rc=2) inst-1 inst0 inst1 inst2 inst3 inst1 inst2 inst3 inst4 if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb figure 3.14 mmu exception in repeat loop 120 (4) 4 or more instructions repeated (inst1, inst2, ..., instn, sr.rc=2) inst-1 inst0 inst1 inst2 : instn-3 instn-2 instn-1 instn inst1 inst2 : instn-3 instn-2 instn-1 instn instn+1 if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb : if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb if id ex ma wb : : exception source stage where spc is not correct and repeat loop can not be restarted figure 3.14 mmu exception in repeat loop (cont) 3.6 memory-mapped tlb in order for tlb operations to be managed by software, tlb contents can be read or written to in privileged mode using the mov instruction. the tlb is assigned to the p4 area in the virtual address space. the tlb address array (vpn, v bit, and asid) is assigned to h'f2000000C h'f2ffffff, and the data array (ppn, pr, sz, c, d, and sh bits) to h'f3000000Ch'f3ffffff. the v bit in the address array can also be accessed from the data array. only longword access is possible for both the address array and the data array. 3.6.1 address array the address array is assigned to h'f2000000Ch'f2ffffff. to access an address array, the 32-bit address field (for read/write operations) and 32-bit data field (for write operations) must be specified. the address field specifies information for selecting the entry to be accessed; the data field specifies the vpn, v bit and asid to be written to the address array (figure 3.15 (1)). in the address field, specify vpn (16C12) as the index address for selecting the entry (bits 16C12), the w bits for selecting the way (bits 9C8), and h'f2 to indicate address array access (bits 31C24). the ix bit in mmucr indicates whether the ex-or of vpn (16C12) and asid (4C0) in the pteh register is used as the index address. 121 when writing, the write is performed to the entry selected with the index address and way. when reading, the vpn, v bit, and asid of the entry selected with the index address and way in the format of the data field in figure 3.12 without comparing addresses. 0 is written to data field bits 16C12. to invalidate a specific entry, specify the entry and way, and write 0 to the corresponding v bit. 3.6.2 data array the data array is assigned to h'f3000000Ch'f3ffffff. to access a data array, the 32-bit address field (for read/write operations), and 32-bit data field (for write operations) must be specified. the address section specifies information for selecting the entry to be accessed; the data section specifies the longword data to be written to the data array (figure 3.15 (2)). longword data has the same bit configuration as ptel. in the address field, specify vpn (16C12) as the index address for selecting the entry (bits 16C12), the w bits for selecting the way (bits 9C8), and h'f3 to indicate data array access (bits 31C24). the ix bit in mmucr indicates whether the ex-or of vpn (16C12) and asid (4C0) in the pteh register is used as the index address. both reading and writing use longword data of the data array specified by the entry address and way number. 122 vpn 31 23 11110010 * * 16 (1) tlb address array access ? read access (2) tlb data array access ? read/write access w 0 * vpn * 31 23 24 24 17 17 17 11110010 * ** * 16 ? write access w 60 * * 0 vpn 31 23 24 11110011 * * 16 17 address field w 0 * * 31 data field 10 ppn 8 9 7654 3 2 1 0 x v xx vpn 31 16 data field 12 10 11 8 97 12 10 11 8 97 12 10 11 8 97 12 10 11 8 97 6 * 0 0 asid 0v vpn 0 0 17 vpn 16 12 10 11 vpn 31 asid 8 97 0 * v d c sh pr sz vpn: v: w: virtual page number valid bit way (00: way 0, 01: way 1, 10: way 2, 11: way 3) asid: : address space identifier dont care bit 0 read and written ppn: pr: c: sh: vpn: x: w: physical page number protection key field cacheable bit share status bit virtual page number 0 for read, dont care bit for write way (00: way 0, 01: way 1, 10: way 2, 11: way 3) v: sz: d: : valid bit page-size bit dirty bit dont care bit address field data field address field * * * * * * * figure 3.15 specifying address and data for memory-mapped tlb access 123 3.6.3 usage examples invalidating specific entries: specific tlb entries can be invalidated by writing 0 to the entrys v bit. r0 specifies the write data and r1 specifies the address. ; r0=h'1547 381c r1=h'f201 30 ; mmucr.ix=0 ; vpn(31?7)=b'0001 0101 0100 011 vpn(11?0)=b'10 asid=b'0001 1100 ; corresponding entry association is made from the entry selected by ; the vpn(16?2)=b'1 0011 index, the v bit of the hit way is cleared to ; 0,achieving invalidation. mov.l r0,@r1 reading the data of a specific entry: this example reads the data section of a specific tlb entry. the data is read in the bit order indicated in the data field in figure 3.15 (2) is read. r0 specifies the address and the data section of a selected entry is read to r1. ; r1 = h'f300 4300 vpn(16-12)=b'00100 way 3 ; mov.l @r0,r1 3.7 usage note instructions that manipulate the md or bl bit in register sr (the ldc rm, sr instruction, ldc @rm+, sr instruction, and rte instruction) and the following instruction, or the ldtlb instruction, should be used with the tlb disabled or in a fixed physical address space (the p1 or p2 space). 124 125 section 4 exception handling 4.1 overview 4.1.1 features exception handling is separate from normal program processing, and is performed by a routine separate from the normal program. in response to an exception handling request due to abnormal termination of the executing instruction, control is passed to a user-written exception handler. however, in response to an interrupt request, normal program execution continues until the end of the executing instruction. here, all exceptions other than resets and interrupts will be called general exceptions. there are thus three types of exceptions: resets, general exceptions, and interrupts. 4.1.2 register configuration table 4.1 lists the registers used for exception handling. a register with an undefined initial value should be initialized by software. table 4.1 register configuration register abbr. r/w size initial value address trapa exception register tra r/w longword undefined h'ffffffd0 exception event register expevt r/w longword power-on reset: h'000 manual reset: h'020* 1 h'ffffffd4 interrupt event register intevt r/w longword undefined h'ffffffd8 interrupt event register2 intevt2 r longword undefined h'04000000 (h'a4000000)* 2 notes: *1 h'000 is set in a power-on reset, and h'020 in a manual reset. *2 when address translation by the mmu does not apply, the address in parentheses should be used. 4.2 exception handling function 4.2.1 exception handling flow in exception handling, the contents of the program counter (pc) and status register (sr) are saved in the saved program counter (spc) and saved status register (ssr), respectively, and execution of the exception handler is invoked from a vector address. the return from exception handler (rte) instruction is issued by the exception handler routine on completion of the routine, restoring the 126 contents of pc and sr to return to the processor state at the point of interruption and the address where the exception occurred. a basic exception handling sequence consists of the following operations: 1. the contents of pc and sr are saved in spc and ssr, respectively. 2. the block (bl) bit in sr is set to 1, masking any subsequent exceptions. 3. the mode (md) bit in sr is set to 1 to place the sh7729r in privileged mode. 4. the register bank (rb) bit in sr is set to 1. 5. an exception code identifying the exception event is written to bits 11C0 of the exception event (expevt) or interrupt event (intevt or intevt2) register. 6. instruction execution jumps to the designated exception vector address to invoke the handler routine. 4.2.2 exception vector addresses the reset vector address is fixed at h'a0000000. the other three events are assigned offsets from the vector base address by software. translation look-aside buffer (tlb) miss exceptions have an offset from the vector base address of h'00000400. the vector address offset for general exception events other than tlb miss exceptions is h'00000100. the interrupt vector address offset is h'00000600. the vector base address is loaded into the vector base register (vbr) by software. the vector base address should reside in p1 or p2 fixed physical address space. figure 4.1 shows the relationship between the vector base address, the vector offset, and the vector table. vbr (vector base address) + vector offset h'a000 0000 vector table figure 4.1 vector table in table 4.2, exceptions and their vector addresses are listed by exception type, instruction completion state, relative acceptance priority, relative order of occurrence within an instruction execution sequence and vector address for exceptions and their vector addresses. 127 table 4.2 exception event vectors exception type current instruction exception event priority* 1 exception order vector address vector offset reset aborted power-on 1 h'a00000000 manual reset 1 h'a00000000 h-udi reset 2 h'a00000000 general exception aborted and retried cpu address error (instruction access) 2 1 h'00000100 events tlb miss (instruction access not in repeat loop) 2 2 h'00000400 tlb miss (instruction access in repeat loop)* 4 2 2 h'00000100 tlb invalid (instruction access) 2 3 h'00000100 tlb protection violation (instruction access) 2 4 h'00000100 reserved instruction code exception 2 5 h'00000100 illegal slot instruction exception 2 5 h'00000100 cpu address error (data access) 2 6 h'00000100 tlb miss (data access not in repeat loop) 2 7 h'00000400 tlb miss (data access in repeat loop)* 4 2 7 h'00000100 tlb invalid (data access) 2 8 h'00000100 tlb protection violation (data access) 2 9 h'00000100 initial page write 2 10 h'00000100 completed unconditional trap (trapa instruction) 2 5 h'00000100 128 table 4.2 exception event vectors (cont) exception type current instruction exception event priority* 1 exception order vector address vector offset general exception events completed user breakpoint trap 2 n* 2 h'00000100 dma address error 2 12 h'00000100 general interrupt completed nonmaskable interrupt 3 h'00000600 requests external hardware interrupt 4* 3 h'00000600 h-udi interrupt 4* 3 h'00000600 notes: *1 priorities are indicated from high to low, 1 being the highest and 4 the lowest. *2 the user defines the break point traps. 1 is a break point before instruction execution and 11 is a break point after instruction execution. for an operand break point, use 11. *3 use software to specify relative priorities of external hardware interrupts and peripheral module interrupts (see section 7, interrupt controller (intc)). *4 see section 4.5.2, general exceptions, for details. 4.2.3 acceptance of exceptions processor resets and interrupts are asynchronous events unrelated to the instruction stream. all exception events are prioritized to establish an acceptance order whenever two or more exception events occur simultaneously. the power-on reset and manual reset do not occur simultaneously, so they have the same priority. all general exception events occur in a relative order in the execution sequence of an instruction (i.e. execution order), but are handled at priority level 2 in instruction-stream order (i.e. program order), where an exception detected in a preceding instruction is accepted prior to an exception detected in a subsequent instruction. three general exception events (reserved instruction code exception, unconditional trap, and illegal slot instruction exception) are detected in the decode stage (id stage) of different instructions and are mutually exclusive events in the instruction pipeline. they have the same execution priority. figure 4.2 shows the order of general exception acceptance. 129 if instruction n id ex ma tlb miss (data access) wb if instruction n + 1 instruction n + 2 id ex ma tlb miss (instruction access) wb if id ex ma rie (reserved instruction exception) wb pipeline sequence: tlb miss (instruction n) re-execution of instruction n 1 2 3 tlb miss (instruction n + 1) re-execution of instruction n + 1 rie (instruction n + 2) if id ex ma wb = instruction fetch = instruction decode = instruction execution = memory access = write back handling order: program order: tlb miss (instruction n+1) tlb miss (instruction n) and rie (instruction n + 2) = simultaneous detection detection order: figure 4.2 example of acceptance order of general exceptions all exceptions other than a reset are detected in the pipeline id stage, and accepted at instruction boundaries. however, an exception is not accepted between a delayed branch instruction and the delay slot. a re-execution type exception detected in a delay slot is accepted before execution of the delayed branch instruction. a completion type exception detected in a delayed branch instruction or delay slot is accepted after execution of the delayed branch instruction. the delay 130 slot here refers to the next instruction after a delayed unconditional branch instruction, or the next instruction when a delayed conditional branch instruction is true. 4.2.4 exception codes table 4.3 lists the exception codes written to bits 11C0 of the expevt register (for reset or general exceptions) or the intevt and intevt2 registers (for general interrupt requests) to identify each specific exception event. an additional exception register, the trapa (tra) register, is used to hold the 8-bit immediate data in an unconditional trap (trapa instruction). table 4.3 exception codes exception type exception event exception code reset power-on reset h'000 manual reset h'020 h-udi reset h'000 general exception events tlb miss/invalid (read) h'040 tlb miss/invalid (write) h'060 tlb miss/invalid/cpu address error in repeat loop h'070 initial page write h'080 tlb protection violation (read) h'0a0 tlb protection violation (write) h'0c0 tlb protection violation in repeat loop h'0d0 cpu address error (read) h'0e0 cpu address error (write) h'100 unconditional trap (trapa instruction) h'160 illegal general instruction exception h'180 illegal slot instruction exception h'1a0 user breakpoint trap h'1e0 dma address error h'5c0 general interrupt requests nonmaskable interrupt h'1c0 h-udi interrupt h'5e0 external hardware interrupts: irl3Cirl0 = 0000 h'200 irl3Cirl0 = 0001 h'220 131 table 4.3 exception codes (cont) exception type exception event exception code general interrupt requests external hardware interrupts (cont): (cont) irl3Cirl0 = 0010 h'240 irl3Cirl0 = 0011 h'260 irl3Cirl0 = 0100 h'280 irl3Cirl0 = 0101 h'2a0 irl3Cirl0 = 0110 h'2c0 irl3Cirl0 = 0111 h'2e0 irl3Cirl0 = 1000 h'300 irl3Cirl0 = 1001 h'320 irl3Cirl0 = 1010 h'340 irl3Cirl0 = 1011 h'360 irl3Cirl0 = 1100 h'380 irl3Cirl0 = 1101 h'3a0 irl3Cirl0 = 1110 h'3c0 note: exception codes h'120, h'140, and h'3e0 are reserved. 4.2.5 exception request masks when the bl bit in sr is 0, exceptions and interrupts are accepted. if a general exception event occurs when the bl bit in sr is 1, the cpus internal registers are set to their post-reset state, other module registers retain their contents prior to the general exception, and a branch is made to the same address (h'a0000000) as for a reset. if a general interrupt occurs when bl = 1, the request is masked (held pending) and not accepted until the bl bit is cleared to 0 by software. for reentrant exception handling, spc and ssr must be saved and the bl bit in sr cleared to 0. 4.2.6 returning from exception handling the rte instruction is used to return from exception handling. when rte is executed, the spc value is set in pc, and the ssr value in sr, and the return from exception handling is performed by branching to the spc address. if spc and ssr have been saved in external memory, set the bl bit in sr to 1, then restore spc and ssr, and issue an rte instruction. 132 4.3 register descriptions there are four registers related to exception handling. these are peripheral module registers, and therefore reside in area p4. they can be accessed by specifying the address in privileged mode only. 1. the exception event register (expevt) resides at address h'ffffffd4, and contains a 12-bit exception code. the exception code set in expevt is that for a reset or general exception event. the exception code is set automatically by hardware when an exception occurs. expevt can also be modified by software. 2. interrupt event register 2 (intevt2) resides at address h'04000000, and contains a 12-bit exception code. the exception code set in intevt2 is that for an interrupt request. the exception code is set automatically by hardware when an exception occurs. 3. the interrupt event register (intevt) resides at address h'ffffffd8, and contains a 12-bit interrupt exception code or a code indicating the interrupt priority. which is set when an interrupt occurs depends on the interrupt source (see tables 7.4 and 7.5, interrupt exception sources and priority). the exception code or interrupt priority code is set automatically by hardware when an exception occurs. intevt can also be modified by software. 4. the trapa exception register (tra) resides at address h'ffffffd0, and contains 8-bit immediate data (imm) for the trapa instruction. tra is set automatically by hardware when a trapa instruction is executed. tra can also be modified by software. the bit configurations of the expevt, intevt, intevt2, and tra registers are shown in figure 4.3. 31 0 0 exception code imm 00 11 0 expevt register, intevt and intevt2 registers 8-bit immediate data in trapa instruction imm: 0: reserved bits, always read as 0 31 00 920 tra register figure 4.3 bit configurations of expevt, intevt, intevt2, and tra registers 133 4.4 exception handling operation 4.4.1 reset the reset sequence is used to power up or restart the sh7729r from the initialization state. the resetp and resetm signals are sampled every clock cycle, and in the case of a power-on reset, all processing being executed (excluding the rtc) is suspended, all unfinished events are canceled, and reset processing is executed immediately. in the case of a manual reset, however, processing to retain external memory contents is continued. the reset sequence consists of the following operations: 1. the md bit in sr is set to 1 to place the sh7729r in privileged mode. 2. the bl bit in sr is set to 1, masking any subsequent exceptions (except the nmi interrupt when the blmsk bit is 1). 3. the rb bit in sr is set to 1. 4. an encoded value of h'000 in a power-on reset or h'020 in a manual reset is written to bits 11C 0 of the expevt register to identify the exception event. 5. instruction execution jumps to the user-written exception handler at address h'a0000000. 4.4.2 interrupts an interrupt handling request is accepted on completion of the current instruction. the interrupt acceptance sequence consists of the following operations: 1. the contents of pc and sr are saved to spc and ssr, respectively. 2. the bl bit in sr is set to 1, masking any subsequent exceptions (except the nmi interrupt when the blmsk bit is 1). 3. the md bit in sr is set to 1 to place the sh7729r in privileged mode. 4. the rb bit in sr is set to 1. 5. an encoded value identifying the exception event is written to bits 11C0 of the intevt and intevt2 registers. 6. instruction execution jumps to the vector location designated by the sum of the value of the contents of the vector base register (vbr) and h'00000600 to invoke the exception handler. 134 4.4.3 general exceptions when the sh7729r encounters any exception condition other than a reset or interrupt request, it executes the following operations: 1. the contents of pc and sr are saved to spc and ssr, respectively. 2. the bl bit in sr is set to 1, masking any subsequent exceptions (except the nmi interrupt when the blmsk bit is 1). 3. the md bit in sr is set to 1 to place the sh7729r in privileged mode. 4. the rb bit in sr is set to 1. 5. an encoded value identifying the exception event is written to bits 11C0 of the expevt register. 6. instruction execution jumps to the vector location designated by either the sum of the vector base address and offset h'00000400 in the vector table in a tlb miss trap, or by the sum of the vector base address and offset h'00000100 for exceptions other than tlb miss traps, to invoke the exception handler. 4.5 individual exception operations this section describes the conditions for specific exception handling, and the processor operations. 4.5.1 resets ? power-on reset ? conditions: resetp low ? operations: expevt set to h'000, vbr and sr initialized, branch to pc = h'a0000000. initialization sets the vbr register to h'0000000. in sr, the md, rb and bl bits are set to 1 and the interrupt mask bits (i3 to i0) are set to 1111. the cpu and on-chip peripheral modules are initialized. see the register descriptions in the relevant sections for details. a power-on reset must always be performed when powering on. a low level is output from the resetout pin, and a high level is output from the status0 and status1 pins. ? manual reset ? conditions: resetm low ? operations: expevt set to h'020, vbr and sr initialized, branch to pc = h'a0000000. initialization sets the vbr register to h'0000000. in sr, the md, rb, and bl bits are set to 1 and the interrupt mask bits (i3 to i0) are set to 1111. the cpu and on-chip peripheral modules are initialized. see the register descriptions in the relevant sections for details. a low level is output from the resetout pin, and a high level is output from the status0 and status1 pins. 135 ? h-udi reset ? conditions: h-udi reset command input (see section 23.4.3, h-udi reset) ? operations: expevt set to h'000, vbr and sr initialized, branch to pc = h'a0000000. initialization sets the vbr register to h'0000000. in sr, the md, rb and bl bits are set to 1 and the interrupt mask bits (i3 to i0) are set to 1111. the cpu and on-chip peripheral modules are initialized. see the register descriptions in the relevant sections for details. table 4.4 types of reset internal state type conditions for transition to reset state cpu on-chip peripheral modules power-on reset resetp = low initialized (see register configuration in relevant sections) manual reset resetm = low initialized h-udi reset h-udi reset command input initialized 4.5.2 general exceptions ? tlb miss exception ? conditions: comparison of tlb addresses shows no address match. ? operations: the virtual address (32 bits) that caused the exception is set in tea and the corresponding virtual page number (22 bits) is set in pteh (31C10). the asid of pteh indicates the asid at the time the exception occurred. the rc bit in mmucr is incremented by one for replacement. pc and sr of the instruction that generated the exception are saved to spc and ssr, respectively. if the exception occurred during a read, h'040 is set in expevt; if the exception occurred during a write, h'060 is set in expevt. the bl, md and rb bits in sr are set to 1 and a branch occurs to pc = vbr + h'0400. to speed up tlb miss processing, the offset differs from other exceptions. 136 ? tlb invalid exception ? conditions: comparison of tlb addresses shows address match but v = 0. ? operations: the virtual address (32 bits) that caused the exception is set in tea and the corresponding virtual page number (22 bits) is set in pteh (31C10). the asid of pteh indicates the asid at the time the exception occurred. the way that generated the exception is set in the rc bits in mmucr. pc and sr of the instruction that generated the exception are saved to spc and ssr, respectively. if the exception occurred during a read, h'040 is set in expevt; if the exception occurred during a write, h'060 is set in expevt. the bl, md, and rb bits in sr are set to 1 and a branch occurs to pc = vbr + h'0100. ? tlb exception/cpu address error in repeat loop ? conditions: tlb miss, tlb invalid or cpu address error in the last several instructions of repeat loop (see section 3.5.6, mmu exception in repeat loop) ? operations: tea, pteh and rc bit in mmucr are set in the way of the type of exception. sr of the instruction that generated the exception is saved to ssr, but spc is not the pc of the instruction that generated the exception. a repeat loop cannot be restarted after returning from the exception handler. in order to complete a repeat loop, ensure that a tlb exception or cpu address error does not occur in the last several instructions of the repeat loop (see section 3.5.6, mmu exception in repeat loop). if a tlb exception or cpu address error occurs in the last several instructions of a repeat loop, h'070 is set in expevt. the bl, md, and rb bits in sr are set to 1 and a branch occurs to pc = vbr + h'0100. ? initial page write exception ? conditions: a hit occurred to the tlb for a store access, but d = 0. this occurs for initial writes to the page registered by the load. ? operations: the virtual address (32 bits) that caused the exception is set in tea and the corresponding virtual page number (22 bits) is set in pteh (31C10). the asid of pteh indicates the asid at the time the exception occurred. the way that generated the exception is set in the rc bit in mmucr. pc and sr of the instruction that generated the exception are saved to spc and ssr, respectively. h'080 is set in expevt. the bl, md, and rb bits in sr are set to 1 and a branch occurs to pc = vbr + h'0100. 137 ? tlb protection exception ? conditions: when a hit access violates the tlb protection information (pr bits) shown below: pr privileged mode user mode 00 only read enabled no access 01 read/write enabled no access 10 only read enabled only read enabled 11 read/write enabled read/write enabled ? operations: the virtual address (32 bits) that caused the exception is set in tea and the corresponding virtual page number (22 bits) is set in pteh (31C10). the asid of pteh indicates the asid at the time the exception occurred. the way that generated the exception is set in the rc bits in mmucr. pc and sr of the instruction that generated the exception are saved to spc and ssr, respectively. if the exception occurred during a read, h'0a0 is set in expevt; if the exception occurred during a write, h'0c0 is set in expevt. the bl, md, and rb bits in sr are set to 1 and a branch occurs to pc = vbr + h'0100. ? tlb protection violation in repeat loop ? conditions: tlb protection violation in the last several instruction of a repeat loop (see section 3.5.6, mmu exception in repeat loop) ? operations: tea, pteh, and rc bit in mmucr are set in the way of the type of exception. sr of the instruction that generated the exception is saved to ssr, but spc is not the pc of the instruction that generated the exception. a repeat loop cannot be restarted after returning from the exception handler. in order to complete a repeat loop, ensure that a tlb exception or cpu address error does not occur in the last several instructions of the repeat loop (see section 3.5.6, mmu exception in repeat loop). if a tlb protection violation occurs in an instruction immediately before or during a repeat loop, h'0d0 is set in expevt. the bl, md, and rb bits in sr are set to 1 and a branch occurs to pc = vbr + h'0100. ? cpu address error ? conditions: a. instruction fetch from odd address (4n + 1, 4n + 3) b. word data accessed from addresses other than word boundaries (4n + 1, 4n + 3) c. longword accessed from addresses other than longword boundaries (4n + 1, 4n + 2, 4n + 3) d. virtual space accessed in user mode in the area h'80000000 to h'ffffffff 138 ? operations: the virtual address (32 bits) that caused the exception is set in tea. pc and sr of the instruction that generated the exception are saved to spc and ssr, respectively. if the exception occurred during a read, h'0e0 is set in expevt; if the exception occurred during a write, h'100 is set in expevt. the bl, md, and rb bits in sr are set to 1 and a branch occurs to pc = vbr + h'0100. see section 3.5.5, processing flow in event of mmu exception, for more information. ? unconditional trap ? conditions: trapa instruction executed ? operations: the exception is a processing-completion type, so pc of the instruction after the trapa instruction is saved to spc. sr from the time when the trapa instruction was executing is saved to ssr. the 8-bit immediate value in the trapa instruction is quadrupled and set in tra (9C0). h'160 is set in expevt. the bl, md, and rb bits in sr are set to 1 and a branch occurs to pc = vbr + h'0100. ? illegal general instruction exception ? conditions: a. when undefined code not in a delay slot is decoded delay branch instructions: jmp, jsr, bra, braf, bsr, bsrf, rts, rte, bt/s, bf/s undefined instruction: h'fxxx b. when a privileged instruction not in a delay slot is decoded in user mode privileged instructions: ldc, stc, rte, ldtlb, sleep; instructions that access gbr with ldc/stc are not privileged instructions. c. when a dsp instruction not in a delay slot is decoded without dsp extension (sr.dsp=0) dsp instructions: lds rm, dsr/a0/x0/x1/y0/y1, lds.l @rm+, dsr/a0/x0/x1/y0/y1, sts dsr/a0/x0/x1/y0/y1, rn, sts.l dsr/a0/x0/x1/y0/y1, @-rn, ldc rm, rs/re/mod, ldc.l @rm+, rs/re/mod, stc rs/re/mod, rn, stc.l rs/re/mod, @-rn, ldrs, ldre, setrc, movs, movx, movy, pxxx d. when an instruction that rewrites pc/sr/rs/re in the last three instructions of repeat loop is decoded. instructions that rewrite pc: jmp, jsr, bra, braf, bsr, bsrf, rts, rte, bt, bf, bt/s, bf/s, trapa, ldc rm, sr, ldc.l @rm+, sr instructions that rewrite sr: ldc rm, sr, ldc.l @rm+, sr, setrc instructions that rewrite rs: ldc rm, rs, ldc.l @rm+, rs, ldrs instructions that rewrite re: ldc rm, re, ldc.l @rm+, re, ldre ? operations: pc and sr of the instruction that generated the exception are saved to spc and ssr, respectively. h'180 is set in expevt. the bl, md, and rb bits in sr are set to 1 and a branch occurs to pc = vbr + h'0100. when an undefined instruction other than h'fxxx is decoded, operation cannot be guaranteed. 139 ? illegal slot instruction ? conditions: a. when undefined code in a delay slot is decoded delay branch instructions: jmp, jsr, bra, braf, bsr, bsrf, rts, rte, bt/s, bf/s b. when an instruction that rewrites pc in a delay slot is decoded instructions that rewrite pc: jmp, jsr, bra, braf, bsr, bsrf, rts, rte, bt, bf, bt/s, bf/s, trapa, ldc rm, sr, ldc.l @rm+, sr c. when a privileged instruction in a delay slot is decoded in user mode privileged instructions: ldc, stc, rte, ldtlb, sleep; instructions that access gbr with ldc/stc are not privileged instructions. d. when a dsp instruction in a delay slot is decoded without dsp extension (sr.dsp=0) dsp instructions: lds rm, dsr/a0/x0/x1/y0/y1, lds.l @rm+, dsr/a0/x0/x1/y0/y1, sts dsr/a0/x0/x1/y0/y1, rn, sts.l dsr/a0/x0/x1/y0/y1, @-rn, ldc rm, rs/re/mod, ldc.l @rm+, rs/re/mod, stc rs/re/mod, rn, stc.l rs/re/mod, @-rn, ldrs, ldre, setrc, movs, movx, movy, pxxx ? operations: pc of the previous delay branch instruction is saved to spc. sr of the instruction that generated the exception is saved to ssr. h'1a0 is set in expevt. the bl, md, and rb bits in sr are set to 1 and a branch occurs to pc = vbr + h'0100. when an undefined instruction other than h'fxxx is decoded, operation cannot be guaranteed. ? user break point trap ? conditions: when a break condition set in the user break controller is satisfied ? operations: when a post-execution break occurs, pc of the instruction immediately after the instruction that set the break point is set in spc. if a pre-execution break occurs, pc of the instruction that set the break point is set in spc. sr when the break occurs is set in ssr. h'1e0 is set in expevt. the bl, md, and rb bits in sr are set to 1 and a branch occurs to pc = vbr + h'0100. see section 8, user break controller (ubc), for more information. ? dma address error ? conditions: a. word data accessed from addresses other than word boundaries (4n + 1, 4n + 3) b. longword accessed from addresses other than longword boundaries (4n + 1, 4n + 2, 4n + 3) ? operations: pc of the instruction immediately after the instruction executed before the exception occurs is saved to spc. sr when the exception occurs is saved to ssr. h'5c0 is set in expevt. the bl, md, and rb bits in sr are set to 1 and a branch occurs to pc = vbr + h'0100. 140 4.5.3 interrupts 1. nmi conditions: nmi pin edge detection operations: pc and sr after the instruction that receives the interrupt are saved to spc and ssr, respectively. h'01c0 is set to intevt and intevt2. the bl, md, and rb bits of the sr are set to 1 and a branch occurs to pc = vbr + h'0600. this interrupt is not masked by the interrupt mask bits in sr and is accepted with top priority when the bl bit in sr is 0. when the bl bit is 1, the interrupt is masked. see section 7, interrupt controller (intc), for more information. 2. irl interrupts conditions: the value of the interrupt mask bits in sr is lower than the irl3Cirl0 level and the bl bit in sr is 0. the interrupt is accepted at an instruction boundary. operations: the pc value after the instruction at which the interrupt is accepted is saved to spc. sr at the time the interrupt is accepted is saved to ssr. the code corresponding to the irl3Cirl0 level is set in intevt and intevt2. the corresponding code is given as h'200 + [irl3Cirl0] h'20. see table 7.5, interrupt exception sources and priority, for the corresponding codes. the bl, md, and rb bits in sr are set to 1 and a branch occurs to vbr + h'0600. the received level is not set in the interrupt mask bits in sr. see section 7, interrupt controller (intc), for more information. 3. irq pin interrupts conditions: the irq pin is asserted, the interrupt mask bits in sr are lower than the irq priority level, and the bl bit in sr is 0. the interrupt is accepted at an instruction boundary. operations: the pc value after the instruction at which the interrupt is accepted is saved to spc. sr at the point the interrupt is accepted is saved to ssr. the code corresponding to the interrupt source is set in intevt and intevt2. the bl, md, and rb bits in sr are set to 1 and a branch occurs to vbr + h'0600. the received level is not set in the interrupt mask bits in sr. see section 7, interrupt controller (intc), for more information. 4. pint pin interrupts conditions: the pint pin is asserted, the interrupt mask bits in sr are lower than the pint priority level, and the bl bit in sr is 0. the interrupt is accepted at an instruction boundary. operations: the pc value after the instruction at which the interrupt is accepted is saved to spc. sr at the point the interrupt is accepted is saved to ssr. the code corresponding to the interrupt source is set in intevt and intevt2. the bl, md, and rb bits of sr are set to 1 and a branch occurs to vbr + h'0600. the received level is not set in the interrupt mask bits in sr. see section 7, interrupt controller (intc), for more information. 141 5. on-chip peripheral interrupts conditions: the interrupt mask bits in sr are lower than the on-chip module (tmu, rtc, sci0, sci1, sci2, a/d, dmac, cpg, ref) interrupt level and the bl bit in sr is 0. the interrupt is accepted at an instruction boundary. operations: the pc value after the instruction at which the interrupt is accepted is saved to spc. sr at the point the interrupt is accepted is saved to ssr. the code corresponding to the interrupt source is set in intevt and intevt2. the bl, md, and rb bits in sr are set to 1 and a branch occurs to vbr + h'0600. see section 7, interrupt controller (intc), for more information. 6. h-udi interrupt conditions: an h-udi interrupt command is input (see section 23.4.4, h-udi interrupt), sr.imask is lower than 15, and the bl bit in sr is 0. the interrupt is accepted at an instruction boundary. operations: the pc value after the instruction that accepts the interrupt is saved to spc. sr at the point the interrupt is accepted is saved to ssr. h'5e0 is set to intevt and intevt2. the bl, md, and rb bits in sr are set to 1 and a branch occurs to vbr + h'0600. see section 7, interrupt controller (intc), for more information. 4.6 cautions ? return from exception handling ? check the bl bit in sr with software. when spc and ssr have been saved to external memory, set the bl bit in sr to 1 before restoring them. ? issue an rte instruction, which sets spc in pc and ssr in sr, and causes a branch to the spc address, and return from exception handling. ? operation when exception or interrupt occurs while sr.bl = 1 ? interrupt: acceptance is suppressed until the bl bit in sr is cleared to 0 by software. if there is a request and the reception conditions are satisfied, the interrupt is accepted after the execution of the instruction that clears the bl bit in sr to 0. in sleep or standby mode, however, the interrupt will be accepted even when the bl bit in sr is 1. nmi is accepted when blmsk in icr1 is 1. ? exception: no user break point trap will occur even when the break conditions are met. when one of the other exceptions occurs, a branch is made to the fixed address of the reset (h'a0000000). in this case, the values of the expevt, spc, and ssr registers are undefined. differently from general reset processing, no signal is output from resetout , status0, and status1. 142 ? spc when exception occurs: the pc saved to spc when an exception occurs is as shown below: ? re-executing-type exceptions: pc of the instruction that caused the exception is set in spc and re-executed after return from exception handling. if the exception occurred in a delay slot, however, pc of the immediately prior delayed branch instruction is set in spc. if the condition of the conditional delayed branch instruction is not satisfied, the delay slot pc is set in spc. ? completed-type exceptions and interrupts: pc of the instruction after the one that caused the exception is set in spc. if the exception was caused by a delayed conditional instruction, however, the branch destination pc is set in spc. if the condition of the conditional delayed branch instruction is not satisfied, the delay slot pc is set in spc. ? initial register values after reset ? undefined registers r0_bank0/1Cr7_bank0/1, r8Cr15, gbr, spc, ssr, mach, macl, pr ? initialized registers vbr = h'00000000 sr.md = 1, sr.bl = 1, sr.rb = 1, sr.i3Csr.i0 = h'f. other sr bits are undefined. pc = h'a0000000 ? ensure that an exception is not generated at an rte instruction delay slot, as operation is not guaranteed in this case. ? when the bl bit in the sr register is set to 1, ensure that a tlb-related exception or address error does not occur at an ldc instruction that updates the sr register and the following instruction. this will be identified as the occurrence of multiple exceptions, and may initiate reset processing. 143 section 5 cache 5.1 overview 5.1.1 features the cache specifications are listed in table 5.1. table 5.1 cache specifications parameter specification capacity 16 kbytes structure instructions/data mixed, 4-way set associative lockingway 2 and way 3 are lockable line size 16 bytes number of entries 256 entries/way write system p0, p1, p3, u0: write-back/write-through selectable replacement method least-recently-used (lru) algorithm 5.1.2 cache structure the cache mixes data and instructions and uses a 4-way set associative system. it is composed of four ways (banks), each of which is divided into an address section and a data section. each of the address and data sections is divided into 256 entries. the data section of the entry is called a line. each line consists of 16 bytes (4 bytes 4). the data capacity per way is 4 kbytes (16 bytes 256 entries), with a total of 16 kbytes in the cache as a whole (4 ways). figure 5.1 shows the cache structure. 144 24 (1 + 1 + 22) bits 128 (32 4) bits 6 bits lw0Clw3: longword data 0C3 entry 0 entry 1 entry 255 0 1 255 0 1 255 v u tag address lw0 lw1 lw2 lw3 address array (ways 0C3) data array (ways 0C3) lru . . . . . . . . . . . . . . . . . . figure 5.1 cache structure address array: the v bit indicates whether the entry data is valid. when the v bit is 1, data is valid; when 0, data is not valid. the u bit indicates whether the entry has been written to in write- back mode. when the u bit is 1, the entry has been written to; when 0, it has not. the address tag holds the physical address used in the external memory access. it is composed of 22 bits (address bits 31C10) used for comparison during cache searches. in the sh7729r, the top three of 32 physical address bits are used as shadow bits (see section 11, bus state controller (bsc)), and therefore in a normal replace operation the top three bits of the tag address are cleared to 0. the v and u bits are initialized to 0 by a power-on reset, but are not initialized by a manual reset. the tag address is not initialized by either a power-on or manual reset. data array: holds a 16-byte instruction or data. entries are registered in the cache in line units (16 bytes). the data array is not initialized by a power-on or manual reset. lru: with the 4-way set associative system, up to four instructions or data with the same entry address (address bits 11C4) can be registered in the cache. when an entry is registered, lru shows which of the four ways it is recorded in. there are six lru bits, controlled by hardware. a least-recently-used (lru) algorithm is used to select the way. in normal operation, four ways are used as cache and six lru bits indicate the way to be replaced (table 5.2). if a bit pattern other than those listed in table 5.2 is set in the lru bits by software, the cache will not function correctly. when modifying the lru bits by software, set one of the patterns listed in table 5.2. the lru bits are initialized to 0 by a power-on reset, but are not initialized by a manual reset. 145 table 5.2 lru and way replacement lru (5C0) way to be replaced 000000, 000100, 010100, 100000, 110000, 110100 3 000001, 000011, 001011, 100001, 101001, 101011 2 000110, 000111, 001111, 010110, 011110, 011111 1 111000, 111001, 111011, 111100, 111110, 111111 0 5.1.3 register configuration table 5.3 shows details of the cache control registers. table 5.3 register configuration register abbr. r/w initial value address access size cache control register ccr r/w h'00000000 h'ffffffec 32 cache control register 2 ccr2 w h'00000000 h'040000b0 (h'a40000b0)* 32 note: * when address translation by the mmu does not apply, the address in parentheses should be used. 5.2 register descriptions 5.2.1 cache control register (ccr) the cache is enabled or disabled using the ce bit in the cache control register (ccr). ccr also has a cf bit (which invalidates all cache entries), and wt and cb bits (which select either write- through mode or write-back mode). programs that change the contents of the ccr register should be placed in address space that is not cached. when updating the contents of the ccr register, bit 4 must always be cleared to 0. figure 5.2 shows the configuration of the ccr register. 146 ce wt cf cb 0 0 0 1 2 3 4 5 6 31 bit 5, 4: always set to 0 when setting the register. cf: cache flush bit. writing 1 flushes all cache entries (clears the v, u, and lru bits of all cache entries to 0). always reads 0. write-back to external memory is not performed when the cache is flushed. cb: cache write-back bit. indicates the caches operating mode for area p1. 1 = write-back mode, 0 = write-through mode. wt: write-through bit. indicates the caches operating mode for areas p0, u0 and p3. 1 = write-through mode, 0 = write-back mode. ce: cache enable bit. indicates whether the cache function is used. 1 = cache used, 0 = cache not used. figure 5.2 ccr register configuration 5.2.2 cache control register 2 (ccr2) ccr2 register is used to enable or disable the cache locking mechanism in dsp mode (set by cpu status register bit 12) only. executing a prefetch instruction (pref) in dsp mode will bring the line of data pointed to by rn into the cache, according to the setting of ccr2 [9:8] (w3load, w3lock) and [1:0] (w2load, w2lock). when ccr2[9:8]=11, in dsp mode pref @rn will bring the data into way 3. when ccr2[9:8]=00, 01, or 10 in dsp mode, or any setting in non-dsp mode, pref @rn will place the data into the way pointed to by lru. when ccr2[1:0]=11, in dsp mode pref @rn will bring the data into way 2. when ccr2[1:0]=00, 01, or 10 in dsp mode, or any setting in non-dsp mode, pref @rn will place the data into the way pointed to by lru. ccr2 must be set before the cache is enabled (ccr.ce = 1). when a pref instruction is issued and there is a cache hit, the operation is treated as nop. figure 5.3 shows the configuration of the ccr2 register. ccr2 is a write-only register; if read, an undefined value will be returned. 147 31 9 8 7 2 1 0 w2 load w3 lock w3 load w2 lock w2lock: way 2 lock bit. w2load: way 2 load bit. when w2lock = 1 & w2load = 1 & dsp = 1, the prefetched data will always be loaded into way 2. in all other conditions the prefetched data will be loaded into the way pointed to by lru. w3lock: way 3 lock bit. w3load: way 3 load bit. when w3lock = 1 & w3load = 1 & dsp = 1, the prefetched data will always be loaded into way 3. in all other conditions the prefetched data will be loaded into the way pointed to by lru. note: w2load and w3load should not be set high at the same time. figure 5.3 ccr2 register configuration whenever ccr2 bit 8 (w3lock) or bit 0 (w2lock) is high the cache is locked. the locked data will not be overwritten unless the w3lock bit and w2lock bit are reset or the pref condition in dsp mode matches. in cache locking mode, the lru values in table 5.2 will be replaced by those in tables 5.4 to 5.6. table 5.4 lru and way replacement (when w2lock=1) lru (5C0) way to be replaced 000000, 000001, 000100, 010100, 100000, 100001, 110000, 110100 3 000011, 000110, 000111, 001011, 001111, 010110, 011110, 011111 1 101001, 101011, 111000, 111001, 111011, 111100, 111110, 111111 0 table 5.5 lru and way replacement (when w3lock=1) lru (5C0) way to be replaced 000000, 000001, 000011, 001011, 100000, 100001, 101001, 101011 2 000100, 000110, 000111, 001111, 010100, 010110, 011110, 011111 1 110000, 110100, 111000, 111001, 111011, 111100, 111110, 111111 0 148 table 5.6 lru and way replacement (when w2lock=1 and w3lock=1) lru (5C0) way to be replaced 000000, 000001, 000011, 000100, 000110, 000111, 001011, 001111, 010100, 010110, 011110, 011111 1 100000, 100001, 101001, 101011, 110000, 110100, 111000, 111001, 111011, 111100, 111110, 111111 0 5.3 cache operation 5.3.1 searching the cache if the cache is enabled, whenever instructions or data in memory are accessed the cache will be searched to see if the desired instruction or data is in the cache. figure 5.4 illustrates the method by which the cache is searched. the cache is a physical cache and holds physical addresses in its address section. entries are selected using bits 11C4 of the address (virtual) of the access to memory and the address tag of that entry is read. in parallel to reading of the address tag, the virtual address is translated to a physical address in the mmu. the physical address after translation and the physical address read from the address section are compared. the address comparison uses all four ways. when the comparison shows a match and the selected entry is valid (v = 1), a cache hit occurs. when the comparison does not show a match or the selected entry is not valid (v = 0), a cache miss occurs. figure 5.4 shows a hit on way 1. 149 0 1 255 v u tag address lw0 lw1 lw2 lw3 ways 0C3 ways 0C3 31 12 11 4 3 2 1 0 virtual address cmp0 cmp1 cmp2 cmp3 physical address cmp0: comparison circuit 0 cmp1: comparison circuit 1 cmp2: comparison circuit 2 cmp3: comparison circuit 3 hit signal 1 entry selection longword (lw) selection mmu figure 5.4 cache search scheme (normal mode) 150 5.3.2 read access read hit: in a read access, instructions and data are transferred from the cache to the cpu. the transfer unit is 32 bits. lru is updated. read miss: an external bus cycle starts and the entry is updated. the way replaced is the one least recently used. entries are updated in 16-byte units. when the desired instruction or data that caused the miss is loaded from external memory to the cache, the instruction or data is transferred to the cpu in parallel with being loaded to the cache. when it is loaded in the cache, the u bit is cleared to 0 and the v bit is set to 1. 5.3.3 write access write hit: in a write access in write-back mode, the data is written to the cache and the u bit of the entry written is set to 1. writing occurs only to the cache; no external memory write cycle is issued. in write-through mode, the data is written to the cache and an external memory write cycle is issued. write miss: in write-back mode, an external write cycle starts when a write miss occurs, and the entry is updated. the way to be replaced is the one least recently used. when the u bit of the entry to be replaced is 1, the cache fill cycle starts after the entry is transferred to the write-back buffer. the write-back unit is 16 bytes. data is written to the cache and the u bit is set to 1. after the cache completes its fill cycle, the write-back buffer writes the entry back to the memory. in write- through mode, no write to cache occurs in a write miss; the write is only to the external memory. 5.3.4 write-back buffer when the u bit of the entry to be replaced in the write-back mode is 1, it must be written back to the external memory. to increase performance, the entry to be replaced is first transferred to the write-back buffer and fetching of new entries to the cache takes priority over writing back to the external memory. during the write-back cycles, the cache can be accessed. the write-back buffer can hold one line of cache data (16 bytes) and its physical address. figure 5.5 shows the configuration of the write-back buffer. longword 0 longword 1 longword 2 longword 3 pa (31C4) pa (31C4): longword 0C3: physical address written to external memory the line of cache data to be written to external memory figure 5.5 write-back buffer configuration 151 5.3.5 coherency of cache and external memory use software to ensure coherency between the cache and the external memory. when memory shared by this lsi and another device is accessed, the latest data may be in a write-back mode cache, so invalidate the entry that includes the latest data in the cache, generate a write-back, and update the data in memory before using it. when the caching area is updated by a device other than the sh7729r, invalidate the entry that includes the updated data in the cache. 5.4 memory-mapped cache to allow software management of the cache, cache contents can be read and written by means of mov instructions in privileged mode. the cache is mapped onto the p4 area in virtual address space. the address array is mapped onto addresses h'f0000000 to h'f0ffffff, and the data array onto addresses h'f1000000 to h'f1ffffff. only longword can be used as the access size for the address array and data array, and instruction fetches cannot be performed. 5.4.1 address array the address array is mapped onto h'f0000000 to h'f0ffffff. to access an address array, the 32-bit address field (for read/write accesses) and 32-bit data field (for write accesses) must be specified. the address field specifies information for selecting the entry to be accessed; the data field specifies the address, v bit, u bit, and lru bits to be written to the address array (figure 5.6 (1)). in the address field, specify the entry address selecting the entry (bits 11C4), w for selecting the way (bits 12C11: in normal mode (8-kbyte cache), 00 is way 0, 01 is way 1, 10 is way 2, and 11 is way 3), and h'f0 to indicate address array access (bits 31C24). when writing, specify bit 3 as the a bit. the a bit indicates whether addresses are compared during writing. when the a bit is 1, the addresses of four entries selected by the entry addresses are compared to the addresses to be written into the address array specified in the data field. writing takes place to the way that has a hit. when a miss occurs, nothing is written to the address array and no operation occurs. the way number (w) specified in bits 12C11 is not used. when the a bit is 0, it is written to the entry selected with the entry address and way number without comparing addresses. the address specified by bits 31C10 in the data specification in figure 5.6 (1), address array access, is a virtual address. when the mmu is enabled, the address is translated into a physical address, then the physical address is used in comparing addresses when the a bit is 1. the physical address is written into the address array. when reading, the address tag, v bit, u bit, and lru bits of the entry specified by the entry address and way number (w) are read using the data format shown in figure 5.6 without comparing addresses. to invalidate a specific entry, specify the entry by its entry address and way 152 number, and write 0 to its v bit. to invalidate only an entry for an address to be invalidated, specify 1 for the a bit. when an entry for which 0 is written to the v bit has a u bit set to 1, it will be written back. this allows coherency to be achieved between the external memory and cache by invalidating the entry. however, when 0 is written to the v bit, 0 must also be written to the u bit of that entry. in the sh7729r, the upper 3 bits of the 32-bit physical address are treated as a shadow field (see section 11, bus state controller (bsc)). therefore, when a cache miss occurs, 0 is stored in the upper 3 bits of the address array address tag. when using an mov instruction to modify the address array directly, a nonzero value must not be written to the upper 3 bits of the address tag. 5.4.2 data array the data array is mapped onto h'f1000000 to h'f1ffffff. to access a data array, the 32-bit address field (for read/write accesses) and 32-bit data field (for write accesses) must be specified. the address field specifies information for selecting the entry to be accessed; the data field specifies the longword data to be written to the data array (figure 5.6 (2)). specify the entry address for selecting the entry (bits 11C4), l indicating the longword position within the (16-byte) line (bits 3C2: 00 is longword 0, 01 is longword 1, 10 is longword 2, and 11 is longword 3), w for selecting the way (bits 12C11: in normal mode, 00 is way 0, 01 is way 1, 10 is way 2, and 11 is way 3), and h'f1 to indicate data array access (bits 31C24). both reading and writing use the longword of the data array specified by the entry address, way number and longword address. the access size of the data array is fixed at longword. 153 1. address array access address specification read access write access data specification 2. data array access (both read and write accesses) address specification 31 24 23 14 13 12 11 4 3 0 1111 0000 * * * * * * * * * * * * w entry 31 24 23 14 13 12 11 4 3 0 1111 0000 w entry 2 a 31 30 29 10 4 3 0 lru 2 x 000 x 9 address tag (31C10) uv 1 31 24 23 14 13 12 11 4 3 0 1111 0001 w entry * * 1 2 l data specification 31 0 longword x: 0 for read, dont care for write * : dont care bit 0 2 figure 5.6 specifying address and data for memory-mapped cache access 154 5.5 usage examples 5.5.1 invalidating specific entries specific cache entries can be invalidated by writing 0 to the entrys v bit. when the a bit is 1, the address tag specified by the write data is compared to the address tag within the cache selected by the entry address, and data is written when a match is found. if no match is found, there is no operation. r0 specifies the write data and r1 specifies the address. when the v bit of an entry in the address array is set to 0, the entry is written back if the entrys u bit is 1. ; r0 = h'01100010; vpn = b'0000 0001 0001 0000 0000 00, u = 0, v = 0 ; r1 = h'f0000088; address array access, entry = b'00001000, a = 1 ; mov.l r0,@r1 5.5.2 reading the data of a specific entry this example reads the data section of a specific cache entry. the longword indicated in the data field of the data array in figure 5.6 is read into the register. r0 specifies the address and r1 is read. ; r1 = h'f100 004c; data array access, entry = b'00000100, way = 0, ; longword address = 3 ; mov.l @r0,r1 ; longword 3 is read. 155 section 6 x/y memory 6.1 overview the sh7729r has on-chip x-ram and y-ram. it can be used by the cpu, dsp and dmac to store instructions or data. 6.1.1 features the x/y memory features are listed in table 6.1. table 6.1 x/y memory specifications parameter features addressing method user selectable mapping mechanism ? fixed mapping for mission-critical realtime applications (p2/uxy area) ? automatic mapping through tlb for easy to use (p0/p3/u0 area) ports three independent read/write ports ? 8-/16-/32-bit access from the cpu ? maximum of two simultaneous 16-bit accesses, or 16/32-bit accesses, from the dsp ? 8-/16-/32-bit access from the dmac size 8-kbyte ram each for x and y memory 156 6.2 x/y memory access from cpu the x/y memory can be located in either a mappable area or fixed-mapped area, depending on the mode bit (md) and dsp bit (dsp) setting in the status register (sr). figure 6.1 shows x/y memory logical mapping. 1. privileged mode md = 1, dsp = 0: any physical address in space p0 or p3 can map to x/y memory through tlb translation. addresses ranging from h'a500 0000 to h'a5ff ffff in the p2 space can also fixed-map to x/y memory. since the dsp extension is disabled, the dsp instruction set and registers are not available to the programmer. 2. user mode md = 0, dsp = 0: any physical address in the u0 space can access x/y memory through tlb translation. any access to addresses beyond the u0 space will cause an address error. since the dsp extension is disabled, the dsp instruction set and registers are not available to the programmer. 3. privileged-dsp mode md = 1, dsp = 1: any physical address in space p0 or p3 can map to x/y memory through tlb translation. addresses ranging from h'a500 0000 to h'a5ff ffff in the p2 space can also fixed-map to x/y memory. since the dsp extension is enabled, the dsp instruction set and registers are available to the programmer. 4. user-dsp mode md = 0, dsp = 1: any physical address in space u0 can map to x/y memory through tlb translation. addresses ranging from h'a500 0000 to h'a5ff ffff in the uxy spaces can also fixed-map to x/y memory. any access outside u0 and uxy space will cause an address error. since the dsp extension is enabled, the dsp instruction set and registers are available to the programmer. for the mappable area, the c (cacheable) bit in the tlb entry must be cleared to 0 to guarantee a two-cycle access. mapping through tlb translation provides a flexible x/y memory addressing scheme but takes two cycles even when the c bit in the tlb entry is cleared to 0. fixed mapping provides a one- cycle access for read and two-cycle access for write, which is the appropriate method for mission- critical realtime operations. the x/y memory resides in the second 16 mbytes of physical address space area 1, from h'a500 0000 to h'a5ff ffff. this 16-mbyte address space is shadowed and maps to the same 128-kbyte x/y rom/ram. figures 6.1 and 6.2 show x/y memory physical mapping. 157 p4 p3 p2 p1 p0 md = 1, dsp = 0 privileged mode same as sh-3 when md = 1, cpu can change dsp flag address error u0 md = 0, dsp = 0 user mode same as sh-3 when md = 0, user cannot change dsp flag address error address error u0 md = 0, dsp = 1 user dsp mode uxy: address range from ha500 0000 to ha5ff ffff p4 p3 p2 p1 p0 md = 1, dsp = 1 privileged dsp mode when md = 1, cpu can change dsp flag x y x y address range from h'a500 0000 to h'a5ff ffff figure 6.1 x/y memory logical address mapping 158 128-kbyte x/y memory x-rom/x-ram reserved space x-rom/x-ram reserved space y-rom/x-ram reserved space y-rom/x-ram reserved space x-ram, 8 kbytes y-ram, 8 kbytes 501ffff 5018fff 5017000 5010000 5008fff 5007000 5000000 area 1, 64 mbytes i/o register space 16 mbytes reserved area 16 mbytes reserved area 32 mbytes 7ffffff x/y memory 6000000 5020000 5000000 4000000 figure 6.2 x/y memory physical address mapping 6.3 x/y memory access from dsp the x/y memory can be accessed by the dsp through the x bus and y bus. accesses via the x bus/y bus are always 16-bit, while accesses via the l bus are either 16-bit or 32-bit. accesses via the x bus and y bus cannot be specified simultaneously. 6.4 x/y memory access from dmac the x/y memory also exists on the i bus and can be accessed by the dmac. dmac access uses an 8-/16-/32-bit unit. if the i bus accesses x/y memory simultaneously with an access from the x bus/y bus or l bus, the i bus master has a higher priority. when accessing x/y memory from the dmac, use physical addresses in the range h'05000000 to h'05ffffff. 159 section 7 interrupt controller (intc) 7.1 overview the interrupt controller (intc) ascertains the priority of interrupt sources and controls interrupt requests to the cpu. the intc registers set the order of priority of each interrupt, allowing the user to process interrupt requests according to the user-set priority. 7.1.1 features the intc has the following features: ? 16 levels of interrupt priority can be set: by setting the five interrupt-priority registers, the priorities of on-chip peripheral module, irq, and pint interrupts can be selected from 16 levels for individual request sources. ? nmi noise canceler function: an nmi input-level bit indicates the nmi pin state. by reading this bit in the interrupt exception service routine, the pin state can be checked, enabling it to be used as a noise canceler. ? external devices can be notified that an interrupt has been received ( irqout ): when the sh7729r has released the bus, the external bus master can be notified that an external interrupt, an on-chip peripheral module interrupt, or a memory refresh request has occurred, enabling the bus to be requested. 160 7.1.2 block diagram figure 7.1 shows a block diagram of the intc. ref irda dmac icr input/output control com- parator priority identifier 3 4 6 16 interrupt request sr ipraCipre irl3 C irl0 nmi irq0Cirq5 pint0Cpint15 irqout : timer unit : realtime clock unit : serial communication interface : serial communication interface (with irda) : serial communication interface (with fifo) : watchdog timer : refresh requests in the bus state controller : interrupt control register : interrupt priority registers aCe : status register : direct memory access controller : analog-to-digital converter : hitachi user-debugging interface legend tmu rtc sci irda scif wdt ref icr ipraCipre sr dmac adc h-udi (interrupt request) scif tmu (interrupt request) (interrupt request/ ipr cpu internal bus bus interface 2 1 0 (interrupt request) (interrupt request) rtc wdt (interrupt request) (interrupt request) irls3 C irls0 4 sci (interrupt request) adc (interrupt request) intc h-udi (interrupt request) refresh request) figure 7.1 block diagram of intc 161 7.1.3 pin configuration table 7.1 shows the intc pin configuration. table 7.1 intc pins name abbreviation i/o description nonmaskable interrupt input pin nmi i input of interrupt request signal, not maskable by the interrupt mask bits in sr interrupt input pins irq5Cirq0 irl3 C irl0 irls3 C irls0 i input of interrupt request signals, maskable by the interrupt mask bits in sr port interrupt input pins pint0Cpint15 i input of port interrupt request signals, maskable by the interrupt mask bits in sr interrupt request output pin irqout o output of signal that notifies external devices that an interrupt source or memory refresh has occurred 162 7.1.4 register configuration the intc has the 12 registers listed in table 7.2. table 7.2 intc registers name abbr. r/w initial value * 1 address access size interrupt control register 0 icr0 r/w * 2 h'fffffee0 16 interrupt control register 1 icr1 r/w h'0000 h'04000010 (h'a4000010) * 3 16 interrupt control register 2 icr2 r/w h'0000 h'04000012 (h'a4000012) * 3 16 pint interrupt enable register pinter r/w h'0000 h'04000014 (h'a4000014) * 3 16 interrupt priority register a ipra r/w h'0000 h'fffffee2 16 interrupt priority register b iprb r/w h'0000 h'fffffee4 16 interrupt priority register c iprc r/w h'0000 h'04000016 (h'a4000016) * 3 16 interrupt priority register d iprd r/w h'0000 h'04000018 (h'a4000018) * 3 16 interrupt priority register e ipre r/w h'0000 h'0400001a (h'a400001a) * 3 16 interrupt request register 0 irr0 r/w h'00 h'04000004 (h'a4000004) * 3 8 interrupt request register 1 irr1 r h'00 h'04000006 (h'a4000006) * 3 8 interrupt request register 2 irr2 r h'00 h'04000008 (h'a4000008) * 3 8 notes: * 1 initialized by a power-on or manual reset. * 2 h'8000 when the nmi pin is high, h'0000 when the nmi pin is low. * 3when address translation by the mmu does not apply, the address in parentheses should be used. 163 7.2 interrupt sources there are five types of interrupt sources: nmi, irq, irl,pint, and on-chip peripheral modules. each interrupt has a priority level (0C16), with 0 the lowest and 16 the highest. priority level 0 masks an interrupt. 7.2.1 nmi interrupt the nmi interrupt has the highest priority level of 16. when the blmsk bit in the interrupt control register (icr1) is 1 or the bl bit in the status register (sr) is 0, nmi interrupts are accepted when the mai bit in the icr1 register is 0. nmi interrupts are edge-detected. in sleep or standby mode, the interrupt is accepted regardless of the bl setting. the nmi edge select bit (nmie) in the interrupt control register 0 (icr0) is used to select either rising or falling edge detection. when the nmie bit in the icr0 register is changed, an nmi interrupt is not detected for 20 cycles after changing icr0. nmie to avoid a false detection of nmi. nmi interrupt exception handling does not affect the interrupt mask level bits (i3Ci0) in the status register (sr). when the bl bit is 1 and the blmsk bit in the icr1 register is set to 1 and only nmi interrupts are accepted, the spc register and ssr register are updated by the nmi interrupt handler, making it impossible to return to the original processing from exception handling initiated prior to the nmi interrupt. use should therefore be restricted to cases where return is not necessary. it is possible to wake the chip up from the standby state with an nmi interrupt (except when the mai bit in the icr1 register is set to 1). 7.2.2 irq interrupts irq interrupts are input by level or edge from pins irq0Cirq5. the priority level can be set by interrupt priority registers cCd (iprcCiprd) in a range from 0 to 15. when using edge-sensing for irq interrupts, clear the interrupt source by having software read 1 from the corresponding bit in irr0, then write 0 to the bit. when the icr1 register is rewritten, irq interrupts may be mistakenly detected, depending on the pin states. to prevent this, rewrite the register while interrupts are masked, then release the mask after clearing the illegal interrupt by writing 0 to interrupt request register 0 (irr0). edge input interrupt detection requires input of a pulse width of more than two cycles on a p clock basis. the interrupt mask bits (i3Ci0) in the status register (sr) are not affected by irq interrupt handling. 164 interrupts irq4Cirq0 can wake the chip up from the standby state when the relevant interrupt level is higher than the setting of i3Ci0 in the sr register (but only when the rtc 32-khz oscillator is used). 7.2.3 irl interrupts irl interrupts are input by level at pins irl3 C irl0 and irls3 C irls0 . irls3 C irls0 are enabled when the irqlvl bit and irlsen bit in interrupt control register 1 (icr1) are both 1. the priority level is the higher level indicated by pins irl3 C irl0 and irls3 C irls0 . an irl3 C irl0 / irls3 C irls0 value of 0 (0000) indicates the highest-level interrupt request (interrupt priority level 15). a value of 15 (1111) indicates no interrupt request (interrupt priority level 0). figure 7.2 shows an example of irl interrupt connection. table 7.3 shows irl / irls pins and interrupt levels. interrupt request priority encoder irl3 to irl0 4 sh7729r irl3 to irl0 interrupt request priority encoder irls3 to irls0 4 irls3 to irls0 figure 7.2 example of irl interrupt connection 165 table 7.3 irl3 C irl0 / irls3 C irls0 pins and interrupt levels irl3 / irls3 irl2 / irls2 irl1 / irls1 irl0 / irls0 interrupt priority level interrupt request 0000 15 level 15 interrupt request 0001 14 level 14 interrupt request 0010 13 level 13 interrupt request 0011 12 level 12 interrupt request 0100 11 level 11 interrupt request 0101 10 level 10 interrupt request 0110 9 level 9 interrupt request 0111 8 level 8 interrupt request 1000 7 level 7 interrupt request 1001 6 level 6 interrupt request 1010 5 level 5 interrupt request 1011 4 level 4 interrupt request 1100 3 level 3 interrupt request 1101 2 level 2 interrupt request 1110 1 level 1 interrupt request 1111 0 no interrupt request a noise-cancellation feature is built in, and the irl interrupt is not detected unless the levels sampled at every peripheral module clock cycle remain unchanged for two consecutive cycles, so that no transient level on the irl / irls pin change is detected. in standby mode, as the peripheral clock is stopped, noise cancellation is performed using the 32-khz clock for the rtc instead. therefore when the rtc is not used, interruption by means of irl interrupts cannot be performed in standby mode. the priority level of the irl interrupt must not be lowered until the interrupt is accepted and interrupt handling starts. however, the priority level can be changed to a higher one. the interrupt mask bits (i3Ci0) in the status register (sr) are not affected by irl / irls interrupt handling. if the interrupt level of an irl interrupt is higher than the setting in bits i3-i0 in the sr register, it can be used to recover from the standby state (but only when using the rtc 32-khz oscillator). 166 7.2.4 pint interrupts pint interrupts are input by level from pins pint0Cpint15. the priority level can be set by interrupt priority register d (iprd) in a range from 0 to 15, in groups of pint0Cpint7 and pint8Cpint15. the pint interrupt level should be held until the interrupt is accepted and interrupt handling is started. the interrupt mask bits (i3Ci0) in the status register (sr) are not affected by pint interrupt handling. pint interrupts can wake the chip up from the standby state when the relevant interrupt level is higher than the setting of i3Ci0 in the sr register (but only when the rtc 32-khz oscillator is used). 7.2.5 on-chip peripheral module interrupts on-chip peripheral module interrupts are generated by the following ten modules: ? timer unit (tmu) ? realtime clock (rtc) ? serial communication interfaces (sci, irda, scif) ? bus state controller (bsc) ? watchdog timer (wdt) ? direct memory access controller (dmac) ? analog-to-digital converter (adc) ? hitachi user-debugging interface (h-udi) not every interrupt source is assigned a different interrupt vector. sources are reflected in the interrupt event registers (intevt and intevt2). it is easy to identify sources by using the value of the intevt or intevt2 register as a branch offset. a priority level (from 0 to 15) can be set for each module except h-udi by writing to interrupt priority registers a, b and e (ipra, iprb, and ipre). the priority level of the h-udi interrupt is 15 (fixed). the interrupt mask bits (i3Ci0) in the status register are not affected by on-chip peripheral module interrupt handling. tmu and rtc interrupts can wake the chip up from the standby state when the relevant interrupt level is higher than the setting of i3Ci0 in the sr register (but only when the rtc 32-khz oscillator is used). 167 7.2.6 interrupt exception handling and priority tables 7.4 and 7.5 list the codes for the interrupt event registers (intevt and intevt2), and the order of interrupt priority. each interrupt source is assigned a unique code. the start address of the interrupt service routine is common to each interrupt source. this is why, for instance, the value of intevt or intevt2 is used as offset at the start of the interrupt service routine and branched to in order to identify the interrupt source. the priority of the on-chip peripheral module, irq, and pint interrupts is set within priority levels 0C15 as required by using interrupt priority registers aCe (ipraCipre). the priority of the on-chip peripheral module, irq, and pint interrupts is set to 0 by a reset. when the priorities of multiple interrupt sources are set to the same level and such interrupts are generated simultaneously, they are handled according to the default order shown in tables 7.4 and 7.5. 168 table 7.4 interrupt exception handling sources and priority (irq mode) interrupt source intevt code (intevt2 code) interrupt priority (initial value) ipr (bit numbers) priority within ipr setting unit default priority nmi h'1c0 (h'1c0) 16 high h-udi h'5e0 (h'5e0) 15 irq irq0 h'200C3c0* (h'600) 0C15 (0) iprc (3C0) irq1 h'200C3c0* (h'620) 0C15 (0) iprc (7C4) irq2 h'200C3c0* (h'640) 0C15 (0) iprc (11C8) irq3h'200C3c0* (h'660) 0C15 (0) iprc (15C12) irq4 h'200C3c0* (h'680) 0C15 (0) iprd (3C0) irq5 h'200C3c0* (h'6a0) 0C15 (0) iprd (7C4) pint pint0-7 h'200C3c0* (h'700) 0C15 (0) iprd (15C12) pint8-15 h'200C3c0* (h'720) 0C15 (0) iprd (11C8) dmac dei0 h'200C3c0* (h'800) 0C15 (0) ipre (15C12) high dei1 h'200C3c0* (h'820) dei2 h'200C3c0* (h'840) dei3h'200C3c0* (h'860) low irda eri1 h'200C3c0* (h'880) 0C15 (0) ipre (11C8) high rxi1 h'200C3c0* (h'8a0) bri1 h'200C3c0* (h'8c0) txi1 h'200C3c0* (h'8e0) low scif eri2 h'200C3c0* (h'900) 0C15 (0) ipre (7C4) high rxi2 h'200C3c0* (h'920) bri2 h'200C3c0* (h'940) txi2 h'200C3c0* (h'960) low adc adi h'200C3c0* (h'980) 0C15 (0) ipre (3C0) tmu0 tuni0 h'400 (h'400) 0C15 (0) ipra (15C12) tmu1 tuni1 h'420 (h'420) 0C15 (0) ipra (11C8) tmu2 tuni2 h'440 (h'440) 0C15 (0) ipra (7C4) high ticpi2 h'460 (h'460) low low 169 table 7.4 interrupt exception handling sources and priority (irq mode) (cont) interrupt source intevt code (intevt2 code) interrupt priority (initial value) ipr (bit numbers) priority within ipr setting unit default priority rtc ati h'480 (h'480) 0C15 (0) ipra (3C0) high high pri h'4a0 (h'4a0) cui h'4c0 (h'4c0) low sci0 eri h'4e0 (h'4e0) 0C15 (0) iprb (3C0) high rxi h'500 (h'500) txi h'520 (h'520) tei h'540 (h'540) low wdt iti h'560 (h'560) 0C15 (0) iprb (15C12) ref rcmi h'580 (h'580) 0C15 (0) iprb (11C8) high rovi h'5a0 (h'5a0) low low note: * the code corresponding to an interrupt level shown in table 7.6 is set. 170 table 7.5 interrupt exception handling sources and priority (irl mode) interrupt source intevt code (intevt2 code) interrupt priority (initial value) ipr (bit numbers) priority within ipr setting unit default priority nmi h'1c0 (h'1c0) 16 high h-udi h'5e0 (h'5e0) 15 irl irl(3:0) * 2 = 0000 h'200 (h'200) 15 irl(3:0) * 2 = 0001 h'220 (h'220) 14 irl(3:0) * 2 = 0010 h'240 (h'240) 13 irl(3:0) * 2 = 0011 h'260 (h'260) 12 irl(3:0) * 2 = 0100 h'280 (h'280) 11 irl(3:0) * 2 = 0101 h'2a0 (h'2a0) 10 irl(3:0) * 2 = 0110 h'2c0 (h'2c0) 9 irl(3:0) * 2 = 0111 h'2e0 (h'2e0) 8 irl(3:0) * 2 = 1000 h'300 (h'300) 7 irl(3:0) * 2 = 1001 h'320 (h'320) 6 irl(3:0) * 2 = 1010 h'340 (h'340) 5 irl(3:0) * 2 = 1011 h'360 (h'360) 4 irl(3:0) * 2 = 1100 h'380 (h'380) 3 irl(3:0) * 2 = 1101 h'3a0 (h'3a0) 2 irl(3:0) * 2 = 1110 h'3c0 (h'3c0) 1 irq irq4 h'200C3c0* 1 (h'680) 0C15 (0) iprd (3C0) irq5 h'200C3c0* 1 (h'6a0) 0C15 (0) iprd (7C4) pint pint0C7 h'200C3c0* 1 (h'700) 0C15 (0) iprd (15C12) pint8C15 h'200C3c0* 1 (h'720) 0C15 (0) iprd (11C8) dmac dei0 h'200C3c0* 1 (h'800) 0C15 (0) ipre (15C12) high dei1 h'200C3c0* 1 (h'820) dei2 h'200C3c0* 1 (h'840) dei3h'200C3c0* 1 (h'860) low irda eri1 h'200C3c0* 1 (h'880) 0C15 (0) ipre (11C8) high rxi1 h'200C3c0* 1 (h'8a0) bri1 h'200C3c0* 1 (h'8c0) txi1 h'200C3c0* 1 (h'8e0) low low 171 table 7.5 interrupt exception handling sources and priority (irl mode) (cont) interrupt source intevt code (intevt2 code) interrupt priority (initial value) ipr (bit numbers) priority within ipr setting unit default priority scif eri2 h'200C3c0* 1 (h'900) 0C15 (0) ipre (7C4) high high rxi2 h'200C3c0* 1 (h'920) bri2 h'200C3c0* 1 (h'940) txi2 h'200C3c0* 1 (h'960) low adc adi h'200C3c0* 1 (h'980) 0C15 (0) ipre (3C0) tmu0 tuni0 h'400 (h'400) 0C15 (0) ipra (15C12) tmu1 tuni1 h'420 (h'420) 0C15 (0) ipra (11C8) tmu2 tuni2 h'440 (h'440) 0C15 (0) ipra (7C4) high ticpi2 h'460 (h'460) low rtc ati h'480 (h'480) 0C15 (0) ipra (3C0) high pri h'4a0 (h'4a0) cui h'4c0 (h'4c0) low sci0 eri h'4e0 (h'4e0) 0C15 (0) iprb (7C4) high rxi h'500 (h'500) txi h'520 (h'520) tei h'540 (h'540) low wdt iti h'560 (h'560) 0C15 (0) iprb (15C12) ref rcmi h'580 (h'580) 0C15 (0) iprb (11C8) high rovi h'5a0 (h'5a0) low low notes: *1 the code corresponding to an interrupt level shown in table 7.6 is set. *2 when irls3 C irls0 are enabled, irl is the higher level of irl3 C irl0 and irls3 C irls0 . 172 table 7.6 interrupt levels and intevt codes interrupt level intevt code 15 h'200 14 h'220 13h'240 12 h'260 11 h'280 10 h'2a0 9 h'2c0 8 h'2e0 7 h'300 6 h'320 5 h'340 4 h'360 3h'380 2 h'3a0 1 h'3c0 173 7.3 intc registers 7.3.1 interrupt priority registers a to e (ipraCipre) interrupt priority registers a to e (ipra to ipre) are 16-bit readable/writable registers in which priority levels from 0 to 15 are set for on-chip peripheral module, irq, and pint interrupts. these registers are initialized to h'0000 by a power-on reset or manual reset, but are not initialized in standby mode. bit: 15 14 1312 11 10 9 8 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 32 1 0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w table 7.7 lists the relationship between the interrupt sources and the ipraipre bits. table 7.7 interrupt request sources and ipraCipre register bits 15 to 12 bits 11 to 8 bits 7 to 4 bits 3 to 0 ipra tmu0 tmu1 tmu2 rtc iprb wdt ref sci0 reserved* iprc irq3irq2 irq1 irq0 iprd pint0 to pint7 pint8 to pint15 irq5 irq4 ipre dmac irda scif adc note: * always read as 0. only 0 should be written. as shown in table 7.7, on-chip peripheral module, irq, or pint interrupts are assigned to four 4- bit groups in each register. these 4-bit groups (bits 15 to 12, bits 11 to 8, bits 7 to 4, and bits 3 to 0) are set with values from h'0 (0000) to h'f (1111). setting h'0 means priority level 0 (masking is requested); h'f is priority level 15 (the highest level). a reset initializes ipraCipre to h'0000. 174 7.3.2 interrupt control register 0 (icr0) icr0 is a register that sets the input signal detection mode of external interrupt input pin nmi, and indicates the input signal level at the nmi pin. this register is initialized to h'0000 or h'8000 by a power-on reset or manual reset, but is not initialized in standby mode. bit: 15 14 1312 11 10 9 8 nmil nmie initial value: 0/1* 0 0 0 0 0 0 0 r/w: r r r r r r r r/w bit: 7 6 5 4 32 1 0 initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r r note: * when nmi input is high, 0 when nmi input is low. bit 15nmi input level (nmil): sets the level of the signal input at the nmi pin. this bit can be read to determine the nmi pin level. this bit cannot be modified. bit 15: nmil description 0 nmi input level is low 1 nmi input level is high bit 8nmi edge select (nmie): selects whether the falling or rising edge of the interrupt request signal at the nmi pin is detected. bit 8: nmie description 0 interrupt request is detected on falling edge of nmi input 1 interrupt request is detected on rising edge of nmi input bits 14 to 9 and 7 to 0reserved: these bits are always read as 0. the write value should always be 0. 175 7.3.3 interrupt control register 1 (icr1) icr1 is a 16-bit register that specifies the detection mode for external interrupt input pins irq0 to irq5 individually: rising edge, falling edge, or low level. this register is initialized to h'4000 by a power-on reset or manual reset, but is not initialized in standby mode. bit: 15 14 1312 11 10 9 8 mai irqlvl blmsk irlsen irq51s irq50s irq41s irq40s initial value: 0 1 0 0 0 0 0 0 r/w: r/w r/w r/w rw r/w r/w r/w r/w bit: 7 6 5 4 32 1 0 irq31s irq30s irq21s irq20s irq11s irq10s irq01s irq00s initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit 15mask all interrupts (mai): when set to 1, all interrupt requests are masked while a low level is being input to the nmi pin. masks nmi interrupts in standby mode. bit 15: mai description 0 all interrupt requests are not masked when nmi pin is low level (initial value) 1 all interrupt requests are masked when nmi pin is low level bit 14interrupt request level detect (irqlvl): selects whether the irq3Cirq0 pins are used as four independent interrupt pins or as 15-level interrupt pins encoded as irl3 C irl0 . bit 14: irqlvl description 0 used as four independent interrupt request pins irq3Cirq0 1 used as 15-level interrupt pins encoded as irl3 C irl0 (initial value) bit 13bl bit mask (blmsk): specifies whether nmi interrupts are masked when the bl bit in the sr register is 1. bit 13: blmsk description 0 nmi interrupts are masked when bl bit is 1 (initial value) 1 nmi interrupts are accepted regardless of bl bit setting 176 bit 12 irls enable (irlsen): enables pins irls3 C irls0 . this bit is valid only when the irqlvl bit is 1. bit 12: irlsen description 0 pins irls3 C irls0 disabled (initial value) 1 pins irls3 C irls0 enabled bits 11 and 10irq5 sense select (irq51s, irq50s): select whether the interrupt signal to the irq5 pin is detected at the rising edge, at the falling edge, or at the low level. bit 11: irq51s bit 10: irq50s description 0 0 an interrupt request is detected at irq5 input falling edge (initial value) 1 an interrupt request is detected at irq5 input rising edge 1 0 an interrupt request is detected at irq5 input low level 1 reserved bits 9 and 8irq4 sense select (irq41s, irq40s): select whether the interrupt signal to the irq4 pin is detected at the rising edge, at the falling edge, or at the low level. bit 9: irq41s bit 8: irq40s description 0 0 an interrupt request is detected at irq4 input falling edge (initial value) 1 an interrupt request is detected at irq4 input rising edge 1 0 an interrupt request is detected at irq4 input low level 1 reserved bits 7 and 6irq3 sense select (irq31s, irq30s): select whether the interrupt signal to the irq3 pin is detected at the rising edge, at the falling edge, or at the low level. bit 7: irq31s bit 6: irq30s description 0 0 an interrupt request is detected at irq3 input falling edge (initial value) 1 an interrupt request is detected at irq3 input rising edge 1 0 an interrupt request is detected at irq3 input low level 1 reserved 177 bits 5 and 4irq2 sense select (irq21s, irq20s): select whether the interrupt signal to the irq2 pin is detected at the rising edge, at the falling edge, or at the low level. bit 5: irq21s bit 4: irq20s description 0 0 an interrupt request is detected at irq2 input falling edge (initial value) 1 an interrupt request is detected at irq2 input rising edge 1 0 an interrupt request is detected at irq2 input low level 1 reserved bits 3 and 2irq1 sense select (irq11s, irq10s): select whether the interrupt signal to the irq1 pin is detected at the rising edge, at the falling edge, or at the low level. bit 3: irq11s bit 2: irq10s description 0 0 an interrupt request is detected at irq1 input falling edge (initial value) 1 an interrupt request is detected at irq1 input rising edge 1 0 an interrupt request is detected at irq1 input low level 1 reserved bits 1 and 0irq0 sense select (irq01s, irq00s): select whether the interrupt signal to the irq0 pin is detected at the rising edge, at the falling edge, or at the low level. bit 1: irq01s bit 0: irq00s description 0 0 an interrupt request is detected at irq0 input falling edge (initial value) 1 an interrupt request is detected at irq0 input rising edge 1 0 an interrupt request is detected at irq0 input low level 1 reserved 178 7.3.4 interrupt control register 2 (icr2) icr2 is a 16-bit readable/writable register that sets the detection mode for external interrupt input pins pint0 to pint15. this register is initialized to h'0000 by a power-on reset or manual reset, but is not initialized in standby mode. bit: 15 14 1312 11 10 9 8 pint15s pint14s pint13s pint14s pint11s pint10s pint9s pint8s initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 32 1 0 pint7s pint6s pint5s pint4s pint3s pint2s pint1s pint0s initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bits 15 to 0pint15 to pint0 sense select (pint15s to pint0s): select whether interrupt request signals to pint15 to pint0 are detected at the low level or high level. b it 15C0: pint15s to pint0s description 0 interrupt requests are detected at low level input to the pint pin (initial value) 1 interrupt requests are detected at high level input to the pint pin 179 7.3.5 pint interrupt enable register (pinter) pinter is a 16-bit readable/writable register that enables interrupt requests input to external interrupt input pins pint0 to pint15. this register is initialized to h'0000 by a power-on reset or manual reset, but is not initialized in standby mode. bit: 15 14 1312 11 10 9 8 pint15e pint14e pint13e pint12e pint11e pint10e pint9e pint8e initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 32 1 0 pint7e pint6e pint5e pint4e pint3e pint2e pint1e pint0e initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bits 15 to 0pint15 to pint0 interrupt enable (pint15e to pint0e): enable or diable interrupt request input to pins pint15 to pint0. b it 15C0: pint15e to pint0e description 0 pint input interrupt requests disabled (initial value) 1 pint input interrupt requests enabled when all or some of pins pint0Cpint15 are not used for interrupt input, bits corresponding to pins not used as interrupt request pins should be cleared to 0. 180 7.3.6 interrupt request register 0 (irr0) irr0 is an 8-bit register that indicates interrupt requests from external input pins irq0 to irq5 and pint0 to pint15. this register is initialized to h'00 by a power-on reset or manual reset, but is not initialized in standby mode. bit: 7 6 5 4 32 1 0 pint0r pint1r irq5r irq4r irq3r irq2r irq1r irq0r initial value: 0 0 0 0 0 0 0 0 r/w: r r r/w r/w r/w r/w r/w r/w when clearing an irq5rCirq0r bit to 0, read the bit while bit set to 1, and then write 0. only 0 can be written to bits irq5rCirq0r. bit 7pint0 to pint7 interrupt request (pint0r): indicates whether there is interrupt request input to pins pint0 to pint7. bit 7: pint0r description 0 no interrupt request to pins pint0 to pint7 (initial value) 1 interrupt to pins pint0 to pint7 bit 6pint8 to pint15 interrupt request (pint1r): indicates whether there is interrupt request input to pins pint8 to pint15. bit 6: pint1r description 0 no interrupt request input to pins pint8 to pint15 (initial value) 1 interrupt request input to pins pint8 to pint15 bit 5irq5 interrupt request (irq5r): indicates whether there is interrupt request input to the irq5 pin. when edge detection mode is set for irq5, an interrupt request is cleared by clearing the irq5r bit. bit 5: irq5r description 0 no interrupt request input to irq5 pin (initial value) 1 interrupt request input to irq5 pin 181 bit 4irq4 interrupt request (irq4r): indicates whether there is interrupt request input to the irq4 pin. when edge detection mode is set for irq4, an interrupt request is cleared by clearing the irq4r bit. bit 4: irq4r description 0 no interrupt request input to irq4 pin (initial value) 1 interrupt request input to irq4 pin bit 3irq3 interrupt request (irq3r): indicates whether there is interrupt request input to the irq3 pin. when edge detection mode is set for irq3, an interrupt request is cleared by clearing the irq3r bit. bit 3: irq3r description 0 no interrupt request input to irq3 pin (initial value) 1 interrupt request input to irq3 pin bit 2irq2 interrupt request (irq2r): indicates whether there is interrupt request input to the irq2 pin. when edge detection mode is set for irq2, an interrupt request is cleared by clearing the irq2r bit. bit 2: irq2r description 0 no interrupt request input to irq2 pin (initial value) 1 interrupt request input to irq2 pin bit 1irq1 interrupt request (irq1r): indicates whether there is interrupt request input to the irq1 pin. when edge detection mode is set for irq1, an interrupt request is cleared by clearing the irq1r bit. bit 1: irq1r description 0 no interrupt request input to irq1 pin (initial value) 1 interrupt request input to irq1 pin bit 0irq0 interrupt request (irq0r): indicates whether there is interrupt request input to the irq0 pin. when edge detection mode is set for irq0, an interrupt request is cleared by clearing the irq0r bit. bit 0: irq0r description 0 no interrupt request input to irq0 pin (initial value) 1 interrupt request input to irq0 pin 182 7.3.7 interrupt request register 1 (irr1) irr1 is an 8-bit read-only register that indicates whether dmac or irda interrupt requests have been generated. this register is initialized to h'00 by a power-on reset or manual reset, but is not initialized in standby mode. bit: 7 6 5 4 32 1 0 txi1r bri1r rxi1r eri1r dei3r dei2r dei1r dei0r initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r r bit 7txi1 interrupt request (txi1r): indicates whether a txi1 (irda) interrupt request has been generated. bit 7: txi1 description 0 txi1 interrupt request not generated (initial value) 1 txi1 interrupt request generated bit 6bri1 interrupt request (bri1r): indicates whether a bri1 (irda) interrupt request has been generated. bit 6: bri1r description 0 bri1 interrupt request not generated (initial value) 1 bri1 interrupt request generated bit 5rxi1 interrupt request (rxi1r): indicates whether an rxi1 (irda) interrupt request has been generated. bit 5: rxi1r description 0 rxi1 interrupt request not generated (initial value) 1 rxi1 interrupt request generated bit 4eri1 interrupt request (eri1r): indicates whether an eri1 (irda) interrupt request has been generated. bit 4: eri1r description 0 eri1 interrupt request not generated (initial value) 1 eri1 interrupt request generated 183 bit 3dei3 interrupt request (dei3r): indicates whether a dei3 (dmac) interrupt request has been generated. b it 3: dei3r description 0 dei3 interrupt request not generated (initial value) 1 dei3 interrupt request generated bit 2dei2 interrupt request (dei2r): indicates whether a dei2 (dmac) interrupt request has been generated. b it 2: dei2r description 0 dei2 interrupt request not generated (initial value) 1 dei2 interrupt request generated bit 1dei1 interrupt request (dei1r): indicates whether a dei1 (dmac) interrupt request has been generated. b it 1: dei1r description 0 dei1 interrupt request not generated (initial value) 1 dei1 interrupt request generated bit 0dei0 interrupt request (dei0r): indicates whether a dei0 (dmac) interrupt request has been generated. b it 0: dei0r description 0 dei0 interrupt request not generated (initial value) 1 dei0 interrupt request generated 7.3.8 interrupt request register 2 (irr2) irr2 is an 8-bit read-only register that indicates whether an a/d converter or scif interrupt request has been generated. this register is initialized to h'00 by a power-on reset or manual reset, but is not initialized in standby mode. bit: 7 6 5 4 32 1 0 adir txi2r bri2r rxi2r eri2r initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r r 184 bits 7 to 5reserved: these bits are always read as 0. the write value should always be 0. bit 4adi interrupt request (adir): indicates whether an adi (adc) interrupt request has been generated. b it 4: adir description 0 adi interrupt request not generated (initial value) 1 adi interrupt request generated bit 3txi2 interrupt request (txi2r): indicates whether a txi2 (scif) interrupt request has been generated. b it 3: txi2r description 0 txi2 interrupt request not generated (initial value) 1 txi2 interrupt request generated bit 2bri2 interrupt request (bri2r): indicates whether a bri2 (scif) interrupt request has been generated. b it 2: bri2r description 0 bri2 interrupt request not generated (initial value) 1 bri2 interrupt request generated bit 1rxi2 interrupt request (rxi2r): indicates whether an rxi2 (scif) interrupt request has been generated. b it 1: rxi2r description 0 rxi2 interrupt request not generated (initial value) 1 rxi2 interrupt request generated bit 0eri2 interrupt request (eri2r): indicates whether an eri2 (scif) interrupt request has been generated. b it 0: eri2r description 0 eri2 interrupt request not generated (initial value) 1 eri2 interrupt request generated 185 7.4 intc operation 7.4.1 interrupt sequence the sequence of interrupt operations is described below. figure 7.3 is a flowchart of the operations. 1. the interrupt request sources send interrupt request signals to the interrupt controller. 2. the interrupt controller selects the highest-priority interrupt from the interrupt requests sent, following the priority levels set in interrupt priority registers a to e (ipra to ipre). lower priority interrupts are held pending. if two of these interrupts have the same priority level or if multiple interrupts occur within a single module, the interrupt with the highest default priority or the highest priority within its ipr setting unit (as indicated in tables 7.4 and 7.5) is selected. 3. the priority level of the interrupt selected by the interrupt controller is compared with the interrupt mask bits (i3Ci0) in the status register (sr) of the cpu. if the request priority level is higher than the level in bits i3Ci0, the interrupt controller accepts the interrupt and sends an interrupt request signal to the cpu. when the interrupt controller receives an interrupt, a low level is output from the irqout pin. 4. detection timing: the intc operates, and notifies the cpu of interrupt requests, in synchronization with the peripheral clock (p ). the cpu receives an interrupt at a break in instructions. 5. the interrupt source code is set in the interrupt event registers (intevt and intevt2). 6. the status register (sr) and program counter (pc) are saved to ssr and spc, respectively. 7. the block bit (bl), mode bit (md), and register bank bit (rb) in sr are set to 1. 8. the cpu jumps to the start address of the interrupt handler (the sum of the value set in the vector base register (vbr) and h'00000600). this jump is not a delayed branch. the interrupt handler may branch with the intevt and intevt2 register value as its offset in order to identify the interrupt source. this enables it to branch to the handling routine for the individual interrupt source. notes: 1. the interrupt mask bits (i3Ci0) in the status register (sr) are not changed by acceptance of an interrupt in the sh7729r. 2. irqout outputs a low level until the interrupt request is cleared. however, if the interrupt source is masked by an interrupt mask bit, the irqout pin returns to the high level. the level is output without regard to the bl bit. 3. the interrupt source flag should be cleared in the interrupt handler. to ensure that an interrupt request that should have been cleared is not inadvertently accepted again, read the interrupt source flag after it has been cleared, then wait for the interval shown in table 7.8 (time for priority decision and sr mask bit comparison) before clearing the bl bit or executing an rte instruction. 186 i3Ci0: interrupt mask bits in status register (sr) program execution state icr1.mai = 1? interrupt generated? nmi = low? sr.bl= 0 or sleep mode? yes yes yes yes yes yes yes yes yes yes yes yes yes no no no no no no no no no no no no no icr1.blmsk = 1? nmi? nmi? level 15 interrupt? irqout = low set interrupt cause in intevt, intevt2 save sr to ssr; save pc to spc set bl/md/rb bits in sr to 1 branch to exception handler i3Ci0 level 14 or lower? level 14 interrupt? i3Ci0 level 13 or lower? level 1 interrupt? i3Ci0 level 0? figure 7.3 interrupt operation flowchart 187 7.4.2 multiple interrupts when handling multiple interrupts, an interrupt handler should include the following procedures: 1. branch to a specific interrupt handler corresponding to a code set in intevt and intevt2. the code in intevt and intevt2 can be used as a branch-offset for branching to the specific handler. 2. clear the cause of the interrupt in each specific handler. 3. save ssr and spc to memory. 4. clear the bl bit in sr, and set the accepted interrupt level in the interrupt mask bits in sr. 5. handle the interrupt. 6. execute the rte instruction. when these procedures are followed in order, an interrupt of higher priority than the one being handled can be accepted after clearing bl in step 4. figure 7.3 shows a sample interrupt operation flowchart. 7.5 interrupt response time the time from generation of an interrupt request until interrupt exception handling is performed and fetching of the first instruction of the exception handler is started (the interrupt response time) is shown in table 7.8. figure 7.4 shows an example of pipeline operation when an irl interrupt is accepted. when sr.bl is 1, interrupt exception handling is masked, and is kept waiting until completion of an instruction that clears bl to 0. 188 table 7.8 interrupt response time number of states item nmi irq pint peripheral modules notes time for priority decision and sr mask bit comparison 0.5 icyc + 0.5 bcyc + 0.5 pcyc 0.5 icyc + 1 bcyc + 4.5 pcyc * 4 0.5 icyc + 3.5 pcyc 0.5 icyc + 1.5 pcyc * 6 0.5 icyc + 3 pcyc * 7 wait time until end of sequence being executed by cpu x ( 0) icyc x ( 0) icyc x ( 0) icyc x ( 0) icyc interrupt exception handling is kept waiting until the executing instruc- tion ends. if the number of instruc- tion execution states is s * 1 , the maximum wait time is: x = s C 1. however, if bl is set to 1 by instru- ction execution or by an exception, interrupt exception handling is deferred until completion of an instruction that clears bl to 0. if the following instruction masks interrupt exception handling, the handling may be further deferred. time from interrupt exception handling (save of sr and pc) until fetch of first instruction of exception handler is started 5 icyc 5 icyc 5 icyc 5 icyc 189 table 7.8 interrupt response time (cont) number of states item nmi irq pint peripheral modules notes response time total (5.5 + x) icyc + 0.5 bcyc + 0.5 pcyc (5.5 + x) icyc + 1 bcyc + 4.5 pcyc * 4 (5.5 + x) icyc + 3.5 pcyc * 6 (5.5 + x) icyc + 1.5 pcyc * 6 (5.5 + x) icyc + 3 pcyc * 7 minimum case * 2 7.5 16.5 12.5 8.5 * 6 /11.5 * 7 at 60-mhz (ckio = 30) operation: 0.13C0.28 s maximum case * 3 8.5 + s 26.5 + s 18.5 + s 10.5 + s * 6 16.5 + s * 7 at 60-mhz (ckio = 15) operation: 0.26C0.56 s (in case of operand cache-hit) at 60-mhz (ckio = 15) operation: 0.29C0.59 s (when external memory access is performed with wait = 0) icyc: duration of one cycle of internal clock supplied to cpu. bcyc: duration of one ckio cycle. pcyc: duration of one cycle of peripheral clock supplied to peripheral modules. notes: * 1 s also includes the memory access wait time. the processing requiring the maximum execution time is ldc.l @rm+, sr. when the memory access is a cache-hit, this requires seven instruction execution cycles. when the external access is performed, the corresponding number of cycles must be added. there are also instructions that perform two external memory accesses; if the external memory access is slow, the number of instruction execution cycles will increase accordingly. * 2 the internal clock:ckio:peripheral clock ratio is 2:1:1. * 3the internal clock:ckio:peripheral clock ratio is 4:1:1. * 4 irq mode * 5 irl mode * 6 modules: tmu, rtc, sci, wdt, refc * 7 modules: dmac, adc, irda, scif 190 interrupt acceptance irl 0.5 icyc + 0.5 bcyc + 2 pcyc instruction (instruction replaced by interrupt exception handling) if id ex ex ex ex if if id ex 5 icyc start of interrupt handling if: instruction fetch: instruction is fetched from memory in which program is stored. id: instruction decode: fetched instruction is decoded. ex: instruction execution: data operation and address calculation are performed. overrun fetch first instruction of interrupt handler figure 7.4 example of pipeline operations when irl interrupt is accepted 191 section 8 user break controller 8.1 overview the user break controller (ubc) provides functions that simplify program debugging. break conditions are set in the ubc and a user break is generated according to the conditions of the bus cycle generated by the cpu or on-chip dmac. the breakpoint check function monitors instruction fetches and operand read/writes, generating a variable combination of pre-execution instruction fetch, post-execution instruction fetch, and post-execution operand access breakpoint traps under designated read/write conditions. this function makes it easy to design an effective self-monitoring debugger, enabling the chip to debug programs without using an in-circuit emulator. 8.1.1 features the ubc has the following features: ? the following break comparison conditions can be set. number of break channels: two channels (channels a and b) user break can be requested as either the independent or sequential condition on channels a and b (sequential break setting: when a channel a break condition match is followed by a channel b break condition match, and both matches do not occur in the same bus cycle). ? address (compares 40 bits comprising a 32-bit logical address prefixed with an asid address. comparison bits are maskable in 32-bit units; user can mask addresses at lower 12 bits (4-k page), lower 10 bits (1-k page), or any size of page, etc.) one of four address buses (logic address bus (lab), internal address bus (iab), x-memory address bus (xab), or y-memory address bus (yab)) can be selected. ? data (only on channel b, 32-bit maskable) one of the four data buses (logic data bus (ldb), internal data bus (idb), x-memory data bus (xdb), or y-memory data bus (ydb)) can be selected. ? bus master: cpu or dmac cycle ? bus cycle: instruction fetch or data access ? read/write ? operand size: byte, word, or longword ? user break is generated upon satisfying break conditions. a user-designed user-break condition exception handling routine can be run. ? in an instruction fetch cycle, break setting before or after instruction execution can be set. ? breaks can be specified for on-chip i/o accesses or ldtlb instruction execution in ase mode. 192 ? the number of repetitions can be specified as a break condition. (channel b only) ? maximum repetitions for the break condition: 2 12 C 1 times. ? eight pairs of branch source/destination buffers. 193 8.1.2 block diagram bbra bara bamra basra comparator cpu state signals xab/yab iab lab mdb access comparator address comparator channel a asid comparator comparator address comparator data comparator pc trace control channel b bbrb betr '()* ccn location ldb/idb/ xdb/ydb legend bbra: break bus cycle register a bara: break address register a bamra: break address mask register a basra: break asid register a bbrb: break bus cycle register b barb: break address register b bamrb: break address mask register b basrb: break asid register b bdrb: break data register b bdmrb: break data mask register b betr: break execution times register + branch source register brdr: branch destination register brcr: break control register access control figure 8.1 block diagram of user break controller 194 8.1.3 register configuration table 8.1 ubc registers name abbr. r/w initial value * 1 address access size location break address register a bara r/w h'00000000 h'ffffffb0 32 ubc break address mask register a bamra r/w h'00000000 h'ffffffb4 32 ubc break bus cycle register a bbra r/w h'0000 h'ffffffb8 16 ubc break address register b barb r/w h'00000000 h'ffffffa0 32 ubc break address mask register b bamrb r/w h'00000000 h'ffffffa4 32 ubc break bus cycle register b bbrb r/w h'0000 h'ffffffa8 16 ubc break data register b bdrb r/w h'00000000 h'ffffff90 32 ubc break data mask register b bdmrb r/w h'00000000 h'ffffff94 32 ubc break control register brcr r/w h'00000000 h'ffffff98 32 ubc execution count break register betr r/w h'0000 h'ffffff9c 16 ubc branch source register brsr r undefined* 2 h'ffffffac 32 ubc branch destination register brdr r undefined* 2 h'ffffffbc 32 ubc break asid register a basra r/w undefined h'ffffffe4 8 ccn break asid register b basrb r/w undefined h'ffffffe8 8 ccn notes: *1 initialized by a power-on reset. values held in the standby state and undefined in a manual reset. *2 bit 31 of brsr and brdr (valid flag) is initialized by a power-on reset, but other bits are not. 195 8.2 register descriptions 8.2.1 break address register a (bara) bit: 31 30 29 28 27 26 25 24 baa31 baa30 baa29 baa28 baa27 baa26 baa25 baa24 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 23 22 21 20 19 18 17 16 baa23 baa22 baa21 baa20 baa19 baa18 baa17 baa16 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 15 14 13 12 11 10 9 8 baa15 baa14 baa13 baa12 baa11 baa10 baa9 baa8 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 3 2 1 0 baa7 baa6 baa5 baa4 baa3 baa2 baa1 baa0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bara is a 32-bit readable/writable register that specifies the address used as a break condition in channel a. bara is initialized to h'00000000 by a power-on reset. bits 31 to 0break address a31 to a0 (baa31 to baa0): store the address on the lab or iab specifying break conditions of channel a. 196 8.2.2 break address mask register a (bamra) bit: 31 30 29 28 27 26 25 24 bama31 bama30 bama29 bama28 bama27 bama26 bama25 bama24 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 23 22 21 20 19 18 17 16 bama23 bama22 bama21 bama20 bama19 bama18 bama17 bama16 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 15 14 13 12 11 10 9 8 bama15 bama14 bama13 bama12 bama11 bama10 bama9 bama8 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 3 2 1 0 bama7 bama6 bama5 bama4 bama3 bama2 bama1 bama0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bamra is a 32-bit readable/writable register that specifies bits masked in the break address specified by bara. bamra is initialized to h'00000000 by a power-on reset. bits 31 to 0break address mask register a31 to a0 (bama31 to bama0): specify bits masked in the channel a break address bits specified by bara (baa31Cbaa0). bit 31C0: baman description 0 break address bit baan of channel a is included in the break condition (initial value) 1 break address bit baan of channel a is masked and is not included in the break condition n = 31C0 197 8.2.3 break bus cycle register a (bbra) bit: 15 14 13 12 11 10 9 8 initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r r bit: 7 6 5 4 3 2 1 0 cda1 cda0 ida1 ida0 rwa1 rwa0 sza1 sza0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w break bus cycle register a (bbra) is a 16-bit readable/writable register that specifies (1) cpu cycle or dmac cycle, (2) instruction fetch or data access, (3) read or write, and (4) operand size in the break conditions of channel a. bbra is initialized to h'0000 by a power-on reset. bits 15 to 8reserved: these bits are always read as 0. the write value should always be 0. bits 7 and 6cpu cycle/dmac cycle select a (cda1, cda0): select a cpu cycle or dmac cycle as the bus cycle of the channel a break condition. bit 7: cda1 bit 6: cda0 description 0 0 condition comparison is not performed (initial value) * 1 break condition is cpu cycle 1 0 break condition is dmac cycle note: * dont care bits 5 and 4instruction fetch/data access select a (ida1, ida0): select an instruction fetch cycle or data access cycle as the bus cycle of the channel a break condition. bit 5: ida1 bit 4: ida0 description 0 0 condition comparison is not performed (initial value) 1 break condition is instruction fetch cycle 1 0 break condition is data access cycle 1 break condition is instruction fetch cycle or data access cycle 198 bits 3 and 2read/write select a (rwa1, rwa0): select a read cycle or write cycle as the bus cycle of the channel a break condition. bit 3: rwa1 bit 2: rwa0 description 0 0 condition comparison is not performed (initial value) 1 break condition is read cycle 1 0 break condition is write cycle 1 break condition is read cycle or write cycle bits 1 and 0operand size select a (sza1, sza0): select the operand size of the bus cycle for the channel a break condition. bit 1: sza1 bit 0: sza0 description 0 0 break condition does not include operand size (initial value) 1 break condition is byte access 1 0 break condition is word access 1 break condition is longword access 199 8.2.4 break address register b (barb) bit: 31 30 29 28 27 26 25 24 bab31 bab30 bab29 bab28 bab27 bab26 bab25 bab24 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 23 22 21 20 19 18 17 16 bab23 bab22 bab21 bab20 bab19 bab18 bab17 bab16 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 15 14 13 12 11 10 9 8 bab15 bab14 bab13 bab12 bab11 bab10 bab9 bab8 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 3 2 1 0 bab7 bab6 bab5 bab4 bab3 bab2 bab1 bab0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w barb is a 32-bit readable/writable register that specifies the address used as a break condition in channel b. control bits xye and xys in bbrb select an address bus for break condition b. if xye is 0, then barb specifies the break address on the logic or internal bus, lab or iab. if xye is 1, then bab31C16 specifies the break address on xab (bits 15C1) and bab15C0 specifies the break address on yab (bits 15C1). however, one of two address buses must be chosen for the break. barb is initialized to h'00000000 by a power-on reset. bab31C16 bab15C0 xye = 0 l(i) ab31C16 l(i) ab15C0 xye = 1 xab15C1 (xys = 0) yab15C1 (xys = 1) 200 8.2.5 break address mask register b (bamrb) bit: 31 30 29 28 27 26 25 24 bamb31 bamb30 bamb29 bamb28 bamb27 bamb26 bamb25 bamb24 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 23 22 21 20 19 18 17 16 bamb23 bamb22 bamb21 bamb20 bamb19 bamb18 bamb17 bamb16 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 15 14 13 12 11 10 9 8 bamb15 bamb14 bamb13 bamb12 bamb11 bamb10 bamb9 bamb8 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 3 2 1 0 bamb7 bamb6 bamb5 bamb4 bamb3 bamb2 bamb1 bamb0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bamrb is a 32-bit readable/writable register that specifies bits masked in the break address specified by barb. bamrb is initialized to h'00000000 by a power-on reset. bamb31C16 bamb15C0 xye = 0 mask l(i) ab31C16 mask l(i) ab15C0 xye = 1 mask xab15C1 (xys = 0) mask yab15C1 (xys = 1) bit 31C0: bambn description 0 break address babn of channel b is included in the break condition (initial value) 1 break address babn of channel b is masked and is not included in the break condition n = 31C0 201 8.2.6 break data register b (bdrb) bit: 31 30 29 28 27 26 25 24 bdb31 bdb30 bdb29 bdb28 bdb27 bdb26 bdb25 bdb24 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 23 22 21 20 19 18 17 16 bdb23 bdb22 bdb21 bdb20 bdb19 bdb18 bdb17 bdb16 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 15 14 13 12 11 10 9 8 bdb15 bdb14 bdb13 bdb12 bdb11 bdb10 bdb9 bdb8 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 3 2 1 0 bdb7 bdb6 bdb5 bdb4 bdb3 bdb2 bdb1 bdb0 initial value: 0 0 0 0 0 0 0 0 bdrb is a 32-bit readable/writable register. the control bits xye and xys in bbrb select a data bus for break condition b. if xye is 0, then bdrb specifies the break data on ldb or idb. if xye is 1, then bdb31C16 specifies the break data on xdb (bits 15C0) and bdb15C0 specifies the break data on ydb (bits 15C0). however, one of two data buses must be chosen for the break. bdrb is initialized to h'00000000 by a power-on reset. bdb31C16 bdb15C0 xye = 0 l(i) db31C16 l(i) db15C0 xye = 1 xdb15C0 (xys = 0) ydb15C0 (xys = 1) 202 8.2.7 break data mask register b (bdmrb) bit: 31 30 29 28 27 26 25 24 bdmb31 bdmb30 bdmb29 bdmb28 bdmb27 bdmb26 bdmb25 bdmb24 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 23 22 21 20 19 18 17 16 bdmb23 bdmb22 bdmb21 bdmb20 bdmb19 bdmb18 bdmb17 bdmb16 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 15 14 13 12 11 10 9 8 bdmb15 bdmb14 bdmb13 bdmb12 bdmb11 bdmb10 bdmb9 bdmb8 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 3 2 1 0 bdmb7 bdmb6 bdmb5 bdmb4 bdmb3 bdmb2 bdmb1 bdmb0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bdmrb is a 32-bit readable/writable register that specifies bits masked in the break data specified by bdrb. bdmrb is initialized to h'00000000 by a power-on reset. bdmb31C16 bdmb15C0 xye = 0 mask l(i) db31C16 mask l(i) db15C0 xye = 1 mask xdb15C0 (xys = 0) mask ydb15C0 (xys = 1) bit 31C0: bdmbn description 0 break data bdbn of channel b is included in the break condition (initial value) 1 break data bdbn of channel b is masked and is not included in the break condition n = 31C0 notes: 1. specify an operand size when including the value of the data bus in the break condition. 2. when a byte size is selected as a break condition, the break data must be set in bits 15-8 in bdrb for an even break address and bits 7-0 for an odd break address. other bits have no influence on a break condition. 203 8.2.8 break bus cycle register b (bbrb) bit: 15 14 13 12 11 10 9 8 xyexys initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r/w r/w bit: 7 6 5 4 3 2 1 0 cdb1 cdb0 idb1 idb0 rwb1 rwb0 szb1 szb0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w break bus cycle register b (bbrb) is a 16-bit readable/writable register that specifies (1) logic or internal bus (l or i bus), x bus, or y bus, (2) cpu cycle or dmac cycle, (3) instruction fetch or data access, (4) read/write, and (5) operand size in the break conditions of channel b. bbrb is initialized to h'0000 by a power-on reset. bits 15 to 10reserved: these bits are always read as 0. the write value should always be 0. bit 9x/y memory bus enable (xye): selects the logic or internal bus (l or i bus) or x/y memory bus as the bus of the channel b break condition. bit 9: xye description 0 internal bus (i bus) selected for the channel b break condition 1 x/y memory bus (x/y bus) selected for the channel b break condition bit 8x or y memory bus select (xys): selects the x bus or the y bus as the bus of the channel b break condition. bit 8: xys description 0 x bus selected for the channel b break condition 1 y bus selected for the channel b break condition 204 bits 7 and 6cpu cycle/dmac cycle select b (cdb1, cdb0): select a cpu cycle or dmac cycle as the bus cycle of the channel b break condition. bit 7: cdb1 bit 6: cdb0 description 0 0 condition comparison is not performed (initial value) * 1 break condition is cpu cycle 1 0 break condition is dmac cycle note: * dont care. bits 5 and 4instruction fetch/data access select b (idb1, idb0): select an instruction fetch cycle or data access cycle as the bus cycle of the channel b break condition. bit 5: idb1 bit 4: idb0 description 0 0 condition comparison is not performed (initial value) 1 break condition is instruction fetch cycle 1 0 break condition is data access cycle 1 break condition is instruction fetch cycle or data access cycle bits 3 and 2read/write select b (rwb1, rwb0): select a read cycle or write cycle as the bus cycle of the channel b break condition. bit 3: rwb1 bit 2: rwb0 description 0 0 condition comparison is not performed (initial value) 1 break condition is read cycle 1 0 break condition is write cycle 1 break condition is read cycle or write cycle bits 1 and 0operand size select b (szb1, szb0): select the operand size of the bus cycle for the channel b break condition. bit 1: szb1 bit 0: szb0 description 0 0 break condition does not include operand size (initial value) 1 break condition is byte access 1 0 break condition is word access 1 break condition is longword access 205 8.2.9 break control register (brcr) bit: 31 30 29 28 27 26 25 24 initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r r bit: 23 22 21 20 19 18 17 16 basma basmb initial value: 0 0 0 0 0 0 0 0 r/w: r r r/w r/w r r r r bit: 15 14 13 12 11 10 9 8 scmfca scmfcb scmfda scmfdb pcte pcba initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r r bit: 7 6 5 4 3 2 1 0 dbeb pcbb seq etbe initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r r r/w r r r/w brcr sets the following conditions: 1. use of channels a and b as two independent channel conditions or as a sequential condition 2. break setting before or after instruction execution 3. break setting by the number of execution times 4. determination of whether to include data bus on channel b in comparison conditions 5. enabling of pc trace 6. enabling of asid check 206 the break control register (brcr) is a 32-bit readable/writable register that has break condition match flags and bits for setting a variety of break conditions. brcr is initialized to h'00000000 by a power-on reset. bits 31 to 22reserved: these bits are always read as 0. the write value should always be 0. bit 21break asid mask a (basma): specifies whether or not channel a break bits asid7 to asid0 (basa7 to basa0) set in basra are masked. bit 21: basma description 0 all basra bits are included in break condition, asid is checked (initial value) 1 no basra bits are included in break condition, asid is not checked bit 20break asid mask b (basmb): specifies whether or not channel b break bits asid7 to asid0 (basb7 to basb0) set in basrb are masked. bit 20: basmb description 0 all basrb bits are included in break condition, asid is checked (initial value) 1 no basrb bits are included in break condition, asid is not checked bits 19 to 16reserved: these bits are always read as 0. the write value should always be 0. bit 15cpu condition match flag a (scmfca): when the cpu bus cycle condition in the break conditions set for channel a is satisfied, this flag is set to 1 (not cleared to 0). in order to clear this flag, write 0 to this bit. bit 15: scmfca description 0 cpu cycle condition for channel a is not matched (initial value) 1 cpu cycle condition for channel a is matched bit 14cpu condition match flag b (scmfcb): when the cpu bus cycle condition in the break conditions set for channel b is satisfied, this flag is set to 1 (not cleared to 0). in order to clear this flag, write 0 to this bit. bit 14: scmfcb description 0 cpu cycle condition for channel b is not matched (initial value) 1 cpu cycle condition for channel b is matched 207 bit 13dmac condition match flag a (scmfda): when the on-chip dmac bus cycle condition in the break conditions set for channel a is satisfied, this flag is set to 1 (not cleared to 0). in order to clear this flag, write 0 to this bit. bit 13: scmfda description 0 dmac cycle condition for channel a is not matched (initial value) 1 dmac cycle condition for channel a is matched bit 12dmac condition match flag b (scmfdb): when the on-chip dmac bus cycle condition in the break conditions set for channel b is satisfied, this flag is set to 1 (not cleared to 0). in order to clear this flag, write 0 to this bit. bit 12: scmfdb description 0 dmac cycle condition for channel b is not matched (initial value) 1 dmac cycle condition for channel b is matched bit 11pc trace enable (pcte): enables a pc trace. bit 11: pcte description 0 pc trace disabled (initial value) 1 pc trace enabled bit 10pc break select a (pcba): selects the break timing of the instruction fetch cycle for channel a as before or after instruction execution. bit 10: pcba description 0 pc break of channel a is set before instruction execution (initial value) 1 pc break of channel a is set after instruction execution bits 9 and 8reserved: these bits are always read as 0. the write value should always be 0. bit 7data break enable b (dbeb): selects whether or not the data bus condition is included in the channel b break condition. bit 7: dbeb description 0 data bus condition not included in channel b condition (initial value) 1 data bus condition included in channel b condition 208 bit 6pc break select b (pcbb): selects the break timing of the instruction fetch cycle for channel b as before or after instruction execution. bit 6: pcbb description 0 pc break of channel b is set before instruction execution (initial value) 1 pc break of channel b is set after instruction execution bits 5 and 4reserved: these bits are always read as 0. the write value should always be 0. bit 3sequence condition select (seq): selects two conditions of channels a and b as independent or sequential. bit 3: seq description 0 channels a and b are compared as independent conditions (initial value) 1 channels a and b are compared as a sequential condition bits 2 and 1reserved: these bits are always read as 0. the write value should always be 0. bit 0execution times break enable (etbe): enables the execution-times break condition on channel b only. if this bit is 1 (break enabled), a user break is issued when the number of break conditions matches the number of execution times specified by the betr register. bit 0: etbe description 0 execution-times break condition is masked on channel b (initial value) 1 execution-times break condition is enabled on channel b 8.2.10 break execution times register (betr) bit: 15 14 13 12 11 10 9 8 initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r/w r/w r/w r/w bit: 7 6 5 4 3 2 1 0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w 209 when the execution-times break condition of channel b is enabled, this register specifies the number of execution times to make the break. the maximum number is 2 12 C 1 times. betr is initialized to h'0000 by a power-on reset. when a break condition is satisfied, the betr value is decremented by 1. a break is issued when the break condition is satisfied after the betr value reaches h'0001. bits 15C12 are always read as 0, and 0 should always be written to these bits. 8.2.11 branch source register (brsr) bit: 31 30 29 28 27 26 25 24 svf pid2 pid1 pid0 bsa27 bsa26 bsa25 bsa24 initial value: 0 ** ** * ** r/w: r r r r r r r r bit: 23 22 21 20 19 18 17 16 bsa23 bsa22 bsa21 bsa20 bsa19 bsa18 bsa17 bsa16 initial value: ******** r/w: r r r r r r r r bit: 15 14 13 12 11 10 9 8 bsa15 bsa14 bsa13 bsa12 bsa11 bsa10 bsa9 bsa8 initial value: ******** r/w: r r r r r r r r bit: 7 6 5 4 3 2 1 0 bsa7 bsa6 bsa5 bsa4 bsa3 bsa2 bsa1 bsa0 initial value: ******** r/w: r r r r r r r r note: * undefined brsr is a 32-bit read-only register that stores the last fetched address before a branch and the pointer (3 bits) which indicates the number of cycles from fetch to execution for the last executed instruction. brsr has a flag bit that is set to 1 when a branch occurs. this flag bit is cleared to 0 when brsr is read, and also is initialized by a power-on reset or manual reset. other bits are not initialized by a reset. four brsr registers have a queue structure and the stored register is shifted every branch. 210 bit 31brsr valid flag (svf): indicates whether the address and the pointer that indicates the branch source address can be calculated. when a branch source address is fetched, this flag is set to 1. this flag is cleared to 0 by reading brsr. bit 31: svf description 0 brsr register value is invalid (initial value) 1 brsr register value is valid bits 30 to 28instruction decode pointer (pid2 to pid0): pid is a 3-bit binary pointer (0C7). these bits indicate the instruction buffer number which stores the last instruction executed before a branch. bits 30 to 28: pid description even pid indicates the instruction buffer number odd pid+2 indicates the instruction buffer number bits 27 to 0branch source address (bsa27 to bsa0): these bits store the last address fetched before a branch. 211 8.2.12 branch destination register (brdr) bit: 31 30 29 28 27 26 25 24 dvf bda27 bda26 bda25 bda24 initial value: 0 ** ** * ** r/w: r r r r r r r r bit: 23 22 21 20 19 18 17 16 bda23 bda22 bda21 bda20 bda19 bda18 bda17 bda16 initial value: ******** r/w: r r r r r r r r bit: 15 14 13 12 11 10 9 8 bda15 bda14 bda13 bda12 bda11 bda10 bda9 bda8 initial value: ******** r/w: r r r r r r r r bit: 7 6 5 4 3 2 1 0 bda7 bda6 bda5 bda4 bda3 bda2 bda1 bda0 initial value: ******** r/w: r r r r r r r r note: * undefined brdr is a 32-bit read-only register that stores the branch destination fetch address. brdr has a flag bit that is set to 1 when a branch occurs. this flag bit is cleared to 0 when brdr is read, and is also initialized by a power-on reset or manual reset. other bits are not initialized by a reset. four brdr registers have a queue structure, and the stored register is shifted every branch. bit 31brdr valid flag (dvf): indicates whether a branch destination address is stored. when a branch destination address is fetched, this flag is set to 1. this flag is set to 0 by reading brdr. bit 31: dvf description 0 brdr register value is invalid (initial value) 1 brdr register value is valid bits 30 to 28reserved: these bits are always read as 0. the write value should always be 0. 212 bits 27 to 0branch destination address (bda27 to bda0): these bits store the first address fetched after a branch. 8.2.13 break asid register a (basra) bit: 7 6 5 4 3 2 1 0 basa7 basa6 basa5 basa4 basa3 basa2 basa1 basa0 initial value: ******** r/w: r/w r/w r/w r/w r/w r/w r/w r/w note: * undefined break asid register a (basra) is an 8-bit readable/writable register that specifies the asid that serves as the break condition for channel a. it is not initialized by a reset. it is located in ccn. bits 7 to 0break asid a7 to 0 (basa7 to basa0): these bits store the asid (bits 7 to 0) that is the channel a break condition. 8.2.14 break asid register b (basrb) bit: 7 6 5 4 3 2 1 0 basb7 basb6 basb5 basb4 basb3 basb2 basb1 basb0 initial value: ******** r/w: r/w r/w r/w r/w r/w r/w r/w r/w note: * undefined break asid register b (basrb) is an 8-bit readable/writable register that specifies the asid that serves as the break condition for channel b. it is not initialized by a reset. it is located in ccn. bits 7 to 0: break asid a7 to 0 (basb7 to basb0): these bits store the asid (bits 7 to 0) that is the channel b break condition. 213 8.3 operation description 8.3.1 flow of the user break operation the flow from setting of break conditions to user break exception processing is described below: 1. the break addresses and the corresponding asids are loaded in the break address registers (bara and barb) and break asid registers (basra and basrb in ccn). the masked addresses are set in the break address mask registers (bamra and bamrb). the break data is set in the break data register (bdrb). the masked data is set in the break data mask register (bdmrb). the breaking bus conditions are set in the break bus cycle registers (bbra and bbrb). three groups of the bbra and bbrb (cpu cycle/dmac cycle select, instruction fetch/data access select, and read/write select) are each set. no user break will be generated if even one of these groups is set with 00. the respective conditions are set in the bits of brcr. 2. when the break conditions are satisfied, the ubc sends a user break request to the interrupt controller. the break type will be sent to the cpu indicating instruction fetch, pre/post instruction break, data access break, or on-chip i/o access/ldtlb break. when conditions match, the cpu condition match flags (scmfca and scmfcb) and dmac condition match flags (scmfda and scmfdb) for the respective channels are set. 3. the appropriate condition match flags (scmfca, scmfda, scmfcb, and scmfdb) can be used to check if the set conditions match or not. the matching of the conditions sets flags, but they are not reset. 0 must first be written to them before they can be used again. 4. there is a chance that a data access break and its following instruction fetch break will occur around the same time; there will be only one break request to the cpu, but these two break channel match flags could be both set. 8.3.2 break on instruction fetch cycle 1. when cpu/instruction fetch/read/word or longword is set in the break bus cycle registers (bbra/bbrb), the break condition becomes the cpu instruction fetch cycle. whether it then breaks before or after execution of the instruction can then be selected with the pcba/pcbb bits in the break control register (brcr) for the appropriate channel. 2. an instruction set for a break before execution breaks when it is confirmed that the instruction has been fetched and will be executed. this means this feature cannot be used on instructions fetched by overrun (instructions fetched at a branch or during an interrupt transition, but not to be executed). when this kind of break is set for the delay slot of a delay branch instruction, the break is generated prior to execution of the instruction that then first accepts the break. meanwhile, breaks set for pre-instruction-break on a delay slot instruction and post- instruction-break on a sleep instruction are also prohibited. 214 3. when the condition is specified to occur after execution, the instruction set with the break condition is executed and then the break is generated prior to the execution of the next instruction. as with pre-execution breaks, this cannot be used with overrun fetch instructions. when this kind of break is set for a delay branch instruction, the break is generated at the instruction that then first accepts the break. 4. when an instruction fetch cycle is set for channel b, break data register b (bdrb) is ignored. there is thus no need to set break data for an instruction fetch cycle break. 8.3.3 break by data access cycle 1. the memory cycle in which a cpu data access break occurs depend on the instruction. 2. the relationship between the data access cycle address and the comparison condition for operand size is shown in table 8.2: table 8.2 data access cycle addresses and operand size comparison conditions access size address compared longword compares break address register bits 31C2 to address bus bits 31C2 word compares break address register bits 31C1 to address bus bits 31C1 byte compares break address register bits 31C0 to address bus bits 31C0 this means that when address h'00001003 is set without specifying the size condition, for example, the bus cycle in which the break condition is satisfied is as follows (where other conditions are met). longword access at h'00001000 word access at h'00001002 byte access at h'00001003 3. when the data value is included in the break condition on channel b: when the data value is included in the break condition, longword, word, or byte is specified as the operand size in the break bus cycle registers (bbra and bbrb). when data values are included in break conditions, a break is generated when the address conditions and data conditions both match. to specify byte data for this case, set the same data in two bytes at bits 15C8 and bits 7C0 of the break data register b (bdrb) and break data mask register b (bdmrb). when word or byte is set, bits 31C16 of bdrb and bdmrb are ignored. 4. when the dmac data access is included in the break condition: when the address is included in the break condition on dmac data access, the operand size of the break bus cycle registers (bbra and bbrb) should be byte, word, or no specified operand size. when the data value is included, select either byte or word. 215 8.3.4 break on x/y-memory bus cycle 1. the break condition on an x/y-memory bus cycle is specified only in channel b. if xye in bbrb is set to 1, break address and break data on the x/y-memory bus are selected. at this time, select the x-memory bus or y-memory bus by specifying xys in bbrb. the break condition cannot include both x-memory and y-memory at the same time. the break condition is applied to x/y-memory bus cycles by specifying cpu/data access/read or write/word or no specified operand size in the break bus cycle register b (bbrb). 2. when x-memory address is selected as the break condition, specify the x-memory address in the upper 16 bits of barb and bamrb. when y-memory address is selected, specify the y- memory address in the lower 16 bits. specification of x/y-memory data is the same for bdrb and bdmrb. 8.3.5 sequential break 1. when seq in brcr is set to 1, the sequential break is issued when the channel b break condition matches after the channel a break condition matches. a user break is ignored even if the channel b break condition matches before the channel a break condition matches. when channel a and b conditions match at the same time, a sequential break is not issued. 2. in sequential break specification, the internal/x/y bus can be selected and the execution times break condition can be also specified. for example, when the execution times break condition is specified, the break condition is satisfied by a channel b condition match with betr = h'0001 after a channel a condition match. 8.3.6 value of saved program counter when a break occurs, pc is saved to spc in user breaks but saved to a fixed address (h'fd000000) in the ase space in an ase break. the pc value saved is as follows depending on the type of break. 1. when instruction fetch (before instruction execution) is specified as a break condition: the value of the program counter (pc) saved is the address of the instruction that matches the break condition. the fetched instruction is not executed, and a break occurs before it. 2. when instruction fetch (after instruction execution) is specified as a break condition: the pc value saved is the address of the instruction to be executed following the instruction in which the break condition matches. the fetched instruction is executed, and a break occurs before execution of the next instruction. 3. when data access (address only) is specified as a break condition: the pc value is the address of the instruction to be executed following the instruction that matched the break condition. the instruction that matched the condition is executed and the break occurs before the next instruction is executed. 216 4. when data access (address + data) is specified as a break condition: the pc value is the start address of the instruction that follows the instruction already executed when break processing started. when a data value is added to the break conditions, the place where the break will occur cannot be specified exactly. the break will occur before the execution of an instruction fetched in the vicinity of the data access where the break occurred. 8.3.7 pc trace 1. a pc trace is started by setting the pc trace enable bit (pcte) to 1 in brcr. when a branch (branch instruction, repeat, interrupt) occurs, an address that enables the branch source address to be calculated and the branch destination address are stored in the branch source register (brsr) and branch destination register (brdr). the branch destination instruction fetch address is stored in brdr, while the last instruction fetch address before the branch is stored in brsr. the branch flag register (brfr) holds a pointer that indicates the relationship to the instruction executed immediately before the branch. 2. the address of the instruction executed immediately before the branch can be calculated from the address stored in brsr and the pointer stored in brfr. if the address stored in brsr is bsa, the pointer stored in brfr is pid, and the address prior to the branch is ia, then ia = bsa C 2 pid. with this equation, caution is required in the case where an interrupt (branch) is executed before the branch destination instruction is executed. in the example in figure 8.2, the address of instruction exec executed immediately before the branch is calculated using the equation ia = bsa C 2 pid. however, if branch branch has a delay slot and the branch destination is address 4n + 2, branch destination address dest specified by the branch instruction is stored in brsr. therefore, the equation ia = bsa C 2 pid does not apply in this case, and this pid is invalid. in this case only, bsa is at the 4n + 2 boundary, classified as shown in table 8.3. exec: dest: int: branch dest instr; not executed interrupt interrupt routine figure 8.2 when interrupt occurs before branch instruction is executed 217 table 8.3 bsa values stored in exception handling before execution of branch destination instruction branch branch destination (dest) bsa branch source address calculable by means of brsr and brfr delay 4n 4n exec = ia = bsa C 2 pid 4n + 2 4n + 2 dest = bsa no delay 4n or 4n + 2 4n exec = ia = bsa C 2 pid if pid is an odd number, the value incremented by 2 indicates the instruction buffer, but the equations in the table do not take this into account. therefore, the calculation can be performed using the values of bsa stored in brsr and pid stored in brfr. 3. the location indicated by ia, the address prior to the branch, depends on the type of branch. a. branch instruction: branch instruction address b. repeat loop: second-before-last instruction of the repeat loop repeat_start: inst (1) ;-----> brdr inst (2) ; : inst (n?) ;-----> address calculated from brsr and brfr repeat end: inst (n) ; c. interrupt: instruction executed immediately before the interrupt the start address of the interrupt routine is stored in brdr. in a repeat loop consisting of no more than three instructions, an instruction fetch cycle is not generated. a pc trace is invalid, since the branch destination address is unknown. 4. brsr, brdr, and brfr have a four-queue structure. when reading addresses stored in a pc trace, reads are performed from the head of the queue. brfr, brsr, and brdr are read in that order. after brdr is read, the queue shifts by one. brsr and brdr should be read by longword access. 218 8.3.8 examples of use break condition specified for cpu instruction fetch cycle 1. register specifications bara = h'00000404, bamra = h'00000000, bbra = h'0054, barb = h'00008010, bamrb = h'00000006, bbrb = h'0054, bdrb = h'00000000, bdmrb = h'00000000, brcr = h'00300400 specified conditions: channel a/channel b independent mode ? channel a address: h'00000404, address mask: h'00000000 bus cycle: cpu/instruction fetch (after instruction execution)/read (operand size is not included in the condition) no asid check is included ? channel b address: h'00008010, address mask: h'00000006 data: h'00000000, data mask: h'00000000 bus cycle: cpu/instruction fetch (before instruction execution)/read (operand size is not included in the condition) no asid check is included a user break occurs after the instruction at address h'00000404 is executed or before instructions at adresses h'00008010 to h'00008016 are executed. 2. register specifications bara = h'00037226, bamra = h'00000000, bbra = h'0056, barb = h'0003722e, bamrb = h'00000000, bbrb = h'0056, bdrb = h'00000000, bdmrb = h'00000000, brcr = h'00000008, basra = h'80, basrb = h'70 specified conditions: channel a/channel b sequential mode ? channel a address: h'00037226, address mask: h'00000000, asid = h'80 bus cycle: cpu/instruction fetch (before instruction execution)/read/word ? channel b address: h'0003722e, address mask: h'00000000, asid = h'70 data: h'00000000, data mask: h'00000000 bus cycle: cpu/instruction fetch (before instruction execution)/read/word the instruction with asid = h'80 and address h'00037226 is executed, and a user break occurs before the instruction with asid = h'70 and address h'0003722e is executed. 219 3. register specifications bara = h'00027128, bamra = h'00000000, bbra = h'005a, barb = h'00031415, bamrb = h'00000000, bbrb = h'0054, bdrb = h'00000000, bdmrb = h'00000000, brcr = h'00300000 specified conditions: channel a/channel b independent mode ? channel a address: h'00027128, address mask: h'00000000 bus cycle: cpu/instruction fetch (before instruction execution)/write/word no asid check is included ? channel b address: h'00031415, address mask: h'00000000 data: h'00000000, data mask: h'00000000 bus cycle: cpu/instruction fetch (before instruction execution)/read (operand size is not included in the condition) no asid check is included on channel a, no user break occurs since an instruction fetch is not a write cycle. on channel b, no user break occurs since an instruction fetch is performed for an even address. 4. register specifications bara = h'00037226, bamra = h'00000000, bbra = h'005a, barb = h'0003722e, bamrb = h'00000000, bbrb = h'0056, bdrb = h'00000000, bdmrb = h'00000000, brcr = h'00000008, basra = h'80, basrb = h'70 specified conditions: channel a/channel b sequential mode ? channel a address: h'00037226, address mask: h'00000000, asid: h'80 bus cycle: cpu/instruction fetch (before instruction execution)/write/word ? channel b address: h'0003722e, address mask: h'00000000, asid: h'70 data: h'00000000, data mask: h'00000000 bus cycle: cpu/instruction fetch (before instruction execution)/read/word since the instruction fetch is not a write cycle on channel a, a sequential condition is not matched. therefore, no user break occurs. 220 5. register specifications bara = h'00000500, bamra = h'00000000, bbra = h'0057, barb = h'00001000, bamrb = h'00000000, bbrb = h'0057, bdrb = h'00000000, bdmrb = h'00000000, brcr = h'00300001, betr = h'0005 specified conditions: channel a/channel b independent mode ? channel a address: h'00000500, address mask: h'00000000 bus cycle: cpu/instruction fetch (before instruction execution)/read/longword ? channel b address: h'00001000, address mask: h'00000000 data: h'00000000, data mask: h'00000000 bus cycle: cpu/instruction fetch (before instruction execution)/read/longword execution-times break enabled (5 times) on channel a, a user break occurs before the instruction at address h'00000500 is executed. on channel b, a user break occurs before the fifth instruction execution after the instruction at address h'00001000 has been executed four times. 6. register specifications bara = h'00008404, bamra = h'00000fff, bbra = h'0054, barb = h'00008010, bamrb = h'00000006, bbrb = h'0054, bdrb = h'00000000, bdmrb = h'00000000, brcr = h'00000400, basra = h'80, basrb = h'70 specified conditions: channel a/channel b independent mode ? channel a address: h'00008404, address mask: h'00000fff, asid: h'80 bus cycle: cpu/instruction fetch (after instruction execution)/read (operand size is not included in the condition) ? channel b address: h'00008010, address mask: h'00000006, asid: h'70 data: h'00000000, data mask: h'00000000 bus cycle: cpu/instruction fetch (before instruction execution)/read (operand size is not included in the condition) a user break occurs after an instruction with asid = h'80 and address h'00008000 to h'00008ffe is executed or before instructions with asid = h'70 and addresses h'00008010 to h'00008016 are executed. 221 break condition specified for cpu data access cycle 1. register specifications bara = h'00123456, bamra = h'00000000, bbra = h'0064, barb = h'000abcde, bamrb = h'000000ff, bbrb = h'006a, bdrb = h'0000a512, bdmrb = h'00000000, brcr = h'00000080, basra = h'80, basrb = h'70 specified conditions: channel a/channel b independent mode ? channel a address: h'00123456, address mask: h'00000000, asid: h'80 bus cycle: cpu/data access/read (operand size is not included in the condition) ? channel b address: h'000abcde, address mask: h'000000ff, asid: h'70 data: h'0000a512, data mask: h'00000000 bus cycle: cpu/data access/write/word on channel a, a user break occurs with asid = h'80 during longword read to address h'00123454, word read to address h'00123456, or byte read to address h'00123456. on channel b, a user break occurs with asid = h'70 when word h'a512 is written in addresses h'000abc00 to h'000abcfe. 2. register specifications: bara = h'01000000, bamra = h'00000000, bbra = h'0066, barb = h'0000f000, bamrb = h'ffff0000, bbrb = h'036a, bdrb = h'00004567, bdmrb = h'00000000, brcr = h'00300080 specified conditions: channel a/channel b independent mode ? channel a address: h'01000000, address mask: h'00000000 bus cycle: cpu/data access/read/word no asid check is included ? channel b y address: h'0001f000, address mask: h'ffff0000 data: h'00004567, data mask: h'00000000 bus cycle: cpu/data access/write/word no asid check is included on channel a, a user break occurs during word read to address h'01000000 in the memory space. on channel b, a user break occurs when word h'4567 is written in address h'0001f000 in y memory space. x/y-memory space is changed by a mode specification. 222 break condition specified for dmac data access cycle 1. register specifications: bara = h'00314156, bamra = h'00000000, bbra = h'0094, barb = h'00055555, bamrb = h'00000000, bbrb = h'00a9, bdrb = h'00000078, bdmrb = h'0000000f, brcr = h'00000080, basra = h'80, basrb = h'70 specified conditions: channel a/channel b independent mode ? channel a address: h'00314156, address mask: h'00000000, asid: h'80 bus cycle: dmac/instruction fetch/read (operand size is not included in the condition) ? channel b address: h'00055555, address mask: h'00000000, asid: h'70 data: h'00000078, data mask: h'0000000f bus cycle: dmac/data access/write/byte on channel a, no user break occurs since an instruction fetch is not performed in dmac cycles. on channel b, a user break occurs with asid = h'70 when the dmac writes byte h'7* in address h'00055555. 223 8.3.9 notes 1. only the cpu can read/write to ubc registers. 2. the ubc cannot monitor cpu and dmac access in the same channel. 3. notes on the specification of a sequential break are given below: a. a condition match occurs when a channel b match occurs in a bus cycle after a channel a match occurs in another bus cycle in sequential break setting. therefore, no condition match occurs if a bus cycle in which a channel a match and a channel b match occur simultaneously is set. b. since the cpu has a pipeline configuration, the pipeline determines the order of an instruction fetch cycle and a memory cycle. therefore, when a channel condition matches in the order of bus cycles, a sequential condition is satisfied. c. when the bus cycle condition for channel a is specified as a break before execution (pcba = 0 in brcr) and an instruction fetch cycle (in bbra), the following point must be noted. a break is issued, and condition match flags in brcr are set to 1, when the bus cycle conditions both for channels a and b match simultaneously. 4. the change of a ubc register value is executed in the ma (memory access) stage. therefore, even if the break condition matches in the instruction fetch address following the instruction in which pre-execution break is specified as the break condition, no break occurs. in order to ascertain the timing of a ubc register is change, read the last register written to. instructions after then are valid for the newly written register value. 5. note the following when specifying an instruction in repeat execution, including a repeat instruction, as the break condition: when an instruction in a repeat loop is specified as the break condition, a. a break is not issued during execution of a repeat loop with fewer than three instructions. b. when an execution-times break is set, no instruction fetch from memory occurs during execution of a repeat loop with fewer than three instructions. therefore, the value in the execution times register, betr, is not decremented. 6. the branch instruction should not be executed as soon as pc trace registers brsr and brdr are read. 7. if a pc break and a tlb exception or error occur in the same instruction, the priority is as follows: a. break and instruction fetch exceptions: instruction fetch exception occurs first. b. break before execution and operand exception: break before execution occurs first. c. break after execution and operand exception: operand exception occurs first. 224 225 section 9 power-down modes 9.1 overview in the power-down modes, all cpu and some on-chip peripheral module functions are halted. this lowers power consumption. 9.1.1 power-down modes the sh7729r has the following power-down modes and function: 1. sleep mode 2. standby mode 3. module standby function (tmu, rtc, sci, x/y memory, ubc, dmac, dac, adc, scif, and irda on-chip peripheral modules) 4. hardware standby mode table 9.1 shows the transition conditions for entering the modes from the program execution state, as well as the cpu and peripheral module states in each mode and the procedures for canceling each mode. 226 table 9.1 power-down modes state mode transition conditions cpg cpu cpu reg- ister on-chip memory on-chip peripheral modules pins external memory canceling procedure sleep mode execute sleep instruction with stby bit cleared to 0 in stbcr runs halts (reg- ister: held) held held run held refresh 1. interrupt 2. reset standby mode execute sleep instruction with stby bit set to 1 in stbcr halts halts (reg- ister: held) held held halt* 1 held self- refresh 1. interrupt 2. reset module standby function set mstp bit to 1 in stbcr runs runs or halts held held specified module halts * 2 refresh 1. clear mstp bit to 0 2. reset hardware standby mode drive ca pin low halts halts held held halt* 3 held self- refresh power-on reset notes: *1 the rtc still runs if the start bit in rcr2 is set to 1 (see section 14, realtime clock (rtc)). the tmu still runs when output of the rtc is used as input to its counter (see section 13, timer (tmu)). *2 depends on the on-chip peripheral module. tmu external pin: held sci external pin: reset *3 the rtc still runs if the start bit in rcr2 is set to 1. the tmu does not run. 227 9.1.2 pin configuration table 9.2 lists the pins used for the power-down modes. table 9.2 pin configuration pin name symbol i/o description processing state 1 status1 o operating state of the processor. processing state 0 status0 hh: reset, hl: sleep mode, lh: standby mode, ll: normal operation wakeup from standby mode wakeup o active-low assertion after accepting wakeup interrupt in standby mode until returning to normal operation with wdt overflow note: h: high level; l: low level 9.1.3 register configuration table 9.3 shows the control register configuration for the power-down modes. table 9.3 register configuration name abbreviation r/w initial value access size address standby control register stbcrr/w h'00* byte h'ffffff82 standby control register 2 stbcr2 r/w h'00* byte h'ffffff88 note: * initialized by a power-on reset. this value is not initialized by a manual reset; the current value is retained. 9.2 register descriptions 9.2.1 standby control register (stbcr) the standby control register (stbcr) is an 8-bit readable/writable register that sets the power- down mode. stbcr is initialized to h'00 by a power-on reset. always set bits 6C3 to 0 when writing to the stbcr register. bit: 7 6 5 4 3 2 1 0 stby stbxtl mstp2 mstp1 mstp0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r r r/w r r/w r/w r/w 228 bit 7standby (stby): specifies transition to standby mode. bit 7: stby description 0 executing sleep instruction puts chip into sleep mode (initial value) 1 executing sleep instruction puts chip into standby mode bits 6, 5, and 3reserved: these bits are always read as 0. the write value should always be 0. bit 4standby crystal (stbxtl): specifies halting or operating of the clock pulse generator in standby mode. bit 4: stbxtl description 0 clock pulse generator is halted in standby mode 1 clock pulse generator is operates in standby mode bit 2module standby 2 (mstp2): specifies halting of the clock supply to the timer unit tmu (an on-chip peripheral module). when the mstp2 bit is set to 1, the supply of the clock to the tmu is halted. bit 2: mstp2 description 0 tmu runs (initial value) 1 clock supply to tmu is halted bit 1module standby 1 (mstp1): specifies halting of the clock supply to the realtime clock rtc (an on-chip peripheral module). when the mstp1 bit is set to 1, the supply of the clock to the rtc is halted. when the clock halts, all rtc registers become inaccessible, but the counter keeps running. bit 1: mstp1 description 0 rtc runs (initial value) 1 clock supply to rtc is halted bit 0module standby 0 (mstp0): specifies halting of the clock supply to the serial communication interface sci (an on-chip peripheral module). when the mstp0 bit is set to 1, the supply of the clock to the sci is halted. bit 0: mstp0 description 0 sci operates (initial value) 1 clock supply to sci is halted 229 9.2.2 standby control register 2 (stbcr2) the standby control register 2 (stbcr2) is a readable/writable 8-bit register that sets the power- down mode. stbcr2 is initialized to h'00 by a power-on reset. bit: 7 6 5 4 3 2 1 0 mstp9 mdchg mstp8 mstp7 mstp6 mstp5 mstp4 mstp3 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit 7module stop 9 (mstp9): specifies halting of the clock supply to the x/y memory (an on-chip peripheral module). when the mstp9 bit is set to 1, the supply of the clock to the memory is halted. bit 7: mstp9 description 0 x/y memory runs (initial value) 1 clock supply to x/y memory halted bit 6pin md5 to md0 control (mdchg): specifies whether or not pins md5 to md0 are changed in standby mode. when this bit is set to 1, the md5 to md0 pin values are latched when returning from standby mode by means of a reset or interrupt. bit 6: mdchg description 0 pins md5 to md0 are not changed in standby mode (initial value) 1 pins md5 to md0 are changed in standby mode bit 5 module stop 8 (mstp8): specifies halting of the clock supply to the user break controller ubc (an on-chip peripheral module). when the mstp8 bit is set to 1, the supply of the clock to the ubc is halted. bit 5: mstp8 description 0 ubc runs (initial value) 1 clock supply to ubc is halted 230 bit 4module stop 7 (mstp7): specifies halting of the clock supply to the dmac (an on-chip peripheral module). when the mstp7 bit is set to 1, the supply of the clock to the dmac is halted. bit 4: mstp7 description 0 dmac runs (initial value) 1 clock supply to dmac halted bit 3module stop 6 (mstp6): specifies halting of the clock supply to the dac (an on-chip peripheral module). when the mstp6 bit is set to 1, the supply of the clock to the dac is halted. bit 3: mstp6 description 0 dac runs (initial value) 1 clock supply to dac halted bit 2module stop 5 (mstp5): specifies halting of the clock supply to the adc (an on-chip peripheral module). when the mstp5 bit is set to 1, the supply of the clock to the adc is halted and all registers are initialized. bit 2: mstp5 description 0 adc runs (initial value) 1 clock supply to adc halted and all registers initialized bit 1module stop 4 (mstp4): specifies halting of the clock supply to the sci2 (scif) serial communication interface with fifo (an on-chip peripheral module). when the mstp1 bit is set to 1, the supply of the clock to sci2 (scif) is halted. bit 1: mstp4 description 0 sci2 (scif) runs (initial value) 1 clock supply to sci2 (scif) halted bit 0module stop 3 (mstp3): specifies halting of the clock supply to the sci1 (irda) infrared data association interface with fifo (an on-chip peripheral module). when the mstp1 bit is set to 1, the supply of the clock to sci1 (irda) is halted. bit 0: mstp3 description 0 sci1(irda) runs (initial value) 1 clock supply to sci1(irda) halted 231 9.3 sleep mode 9.3.1 transition to sleep mode executing the sleep instruction when the stby bit in stbcr is 0 causes a transition from the program execution state to sleep mode. although the cpu halts immediately after executing the sleep instruction, the contents of its internal registers remain unchanged. the on-chip peripheral modules continue to run in sleep mode and the clock continues to be output to the ckio and ckio2 pins. in sleep mode, the status1 pin is set high and the status0 pin low. 9.3.2 canceling sleep mode sleep mode is canceled by an interrupt (nmi, irq, irl, on-chip peripheral module, pint) or reset. interrupts are accepted in sleep mode even when the bl bit in the sr register is 1. if necessary, save spc and ssr to the stack before executing the sleep instruction. canceling with an interrupt: when an nmi, irq, irl or on-chip peripheral module interrupt occurs, sleep mode is canceled and interrupt exception handling is executed. a code indicating the interrupt source is set in the intevt and intevt2 registers. canceling with a reset: sleep mode is canceled by a power-on reset or a manual reset. 232 9.4 standby mode 9.4.1 transition to standby mode to enter standby mode, set the stby bit to 1 in stbcr, then execute the sleep instruction. the chip switches from the program execution state to standby mode. in standby mode, power consumption is greatly reduced by halting not only the cpu, but the clock and on-chip peripheral modules as well. the clock output from the ckio and ckio2 pins also halts. cpu and cache register contents are held, but some on-chip peripheral modules are initialized. table 9.4 lists the states of registers in standby mode. table 9.4 register states in standby mode module registers initialized registers retaining data interrupt controller all registers on-chip clock pulse generator all registers user break controller (ubc) all registers bus state controller (bsc) all registers timer unit (tmu) tstr register registers other than tstr realtime clock (rtc) all registers a/d converter (adc) all registers d/a converter (dac) all registers the procedure for moving to standby mode is as follows: 1. clear the tme bit in the wdts timer control register (wtcsr) to 0 to stop the wdt. set the wdts timer counter (wtcnt) and the cks2Ccks0 bits in the wtcsr register to appropriate values to secure the specified oscillation settling time. 2. when pll circuit 1 is running in clock modes 3 and 4, clear the pstby and pllen bits in the frequency control register (frqcr) to 0 to stop pll circuit 1. 3. after the stby bit in the stbcr register is set to 1, a sleep instruction is executed. 4. standby mode is entered and the clocks within the chip are halted. the status1 pin output goes low and the status0 pin output goes high. 233 9.4.2 canceling standby mode standby mode is canceled by an interrupt (nmi, irq, irl, pint, or on-chip peripheral module) or a reset. canceling with an interrupt: the on-chip wdt can be used for hot starts. when the chip detects an nmi, irl, irq, pint* 1 , or on-chip peripheral module (except interval timer)* 2 interrupt, the clock will be supplied to the entire chip and standby mode canceled after the time set in the wdts timer control/status register has elapsed. the status1 and status0 pins both go low. interrupt handling then begins and a code indicating the interrupt source is set in the intevt and intevt2 registers. after the branch to the interrupt handling routine, clear the stby bit in the stbcr register. wtcnt stops automatically. if the stby bit is not cleared, wtcnt continues operation and a transition is made to standby mode* 3 when it reaches h'80. this function prevents the data from being destroyed due to a rise in voltage with an unstable power supply, etc. interrupts are accepted in standby mode even when the bl bit in the sr register is 1. if necessary, save spc and ssr to the stack before executing the sleep instruction. immediately after an interrupt is detected, the phase of the ckio pin clock output may be unstable, until the processor starts interrupt handling. (the canceling condition is that the irl3Cirl0 level is higher than the mask level in the i3Ci0 bits in the sr register.) notes: *1 when the rtc is being used, standby mode can be canceled using irl3Cirl0, irq4C irq0, or pint0/1. *2 standby mode can be canceled with an rtc or tmu (only when running on the rtc clock) interrupt. *3 this standby mode can be canceled only by a power-on reset. wtcnt value h'ff h'80 time interrupt request wdt overflow and branch to interrupt handling routine crystal oscillator settling time and pll synchronization time clear bit stbcr.stby before wtcnt reaches h'80. when stbcr.stby is cleared, wtcnt +) figure 9.1 canceling standby mode with stbcr.stby 234 canceling with a reset: standby mode is canceled by a reset (power-on or manual). keep the reset pin low until the clock oscillation settles. the internal clock will continue to be output to the ckio and ckio2 pins. 9.4.3 clock pause function in standby mode, the clock input from the extal pin or ckio pin can be halted and the frequency can be changed. this function is used as follows: 1. enter standby mode using the appropriate procedures. 2. once standby mode is entered and the clock stopped within the chip, the status1 pin output is low and the status0 pin output is high. 3. once the status1 pin goes low and the status0 pin goes high, the input clock is stopped or the frequency is changed. 4. when the frequency is changed, an nmi, irl, irq, pint or on-chip peripheral module (except interval timer) interrupt is input after the change. when the clock is stopped, the same interrupts are input after the clock is applied. 5. after the time set in the wdt has elapsed, the clock starts being applied internally within the chip, the status1 and status0 pins both go low, and operation resumes from interrupt exception handling. 235 9.5 module standby function 9.5.1 transition to module standby function setting the standby control register mstp9Cmstp0 bits to 1 halts the supply of clocks to the corresponding on-chip peripheral modules. this function can be used to reduce the power consumption in sleep mode. the module standby function holds the state prior to halting the external pins of the on-chip peripheral modules. tmu external pins hold their state prior to the halt. sci external pins go to the reset state. with a few exceptions, all registers hold their values. bit value description mstp9 0 x/y memory runs 1 supply of clock to x/y memory halted mstp8 0 ubc runs 1 supply of clock to ubc halted mstp7 0 dmac runs 1 supply of clock to dmac halted mstp6 0 dac runs 1 supply of clock to dac halted mstp5 0 adc runs 1 supply of clock to adc halted, and all registers initialized mstp4 0 scif runs 1 supply of clock to scif halted mstp3 0 irda runs 1 supply of clock to irda halted mstp2 0 tmu runs 1 supply of clock to tmu halted. registers initialized *1 mstp1 0 rtc runs 1 supply of clock to rtc halted. register access prohibited *2 mstp0 0 sci runs 1 supply of clock to sci halted notes: *1 the registers initialized are the same as in standby mode (see table 9.4). *2 the counter runs. 236 9.5.2 clearing module standby function the module standby function can be cleared by clearing the mstp9Cmstp0 bits to 0, or by a power-on reset or manual reset. 9.6 timing of status pin changes the timing of status1 and status0 pin changes is shown in figures 9.1 through 9.8. 9.6.1 timing for resets power-on reset ckio, ckio2 * 4 resetp status normal * 2 normal * 2 reset * 1 pll settling time 0 to 5 bcyc * 3 0 to 30 bcyc * 3 resetout notes: * 1. reset: hh (status1 high, status0 high) * 2. normal: ll (status1 low, status0 low) * 3. bcyc: bus clock cycle * 4. ckio2 output can only be used in clock modes 0, 1, and 2. figure 9.2 power-on reset (clock modes 0, 1, 2, and 7) status output 237 manual reset ckio, ckio2 * 5 resetm status normal * 3 normal * 3 reset * 2 0 bcyc or more * 1, * 4 0 to 30 bcyc * 4 resetout notes: * 1 in a manual reset, status becomes hh (reset) and the internal reset begins after waiting for the executing bus cycle to end. * 2 reset: hh (status1 high, status0 high) * 3 normal: ll (status1 low, status0 low) * 4 bcyc: bus clock cycle * 5 ckio2 output can only be used in clock modes 0, 1, and 2. figure 9.3 manual reset status output 238 9.6.2 timing for canceling standby standby to interrupt ckio, ckio2 * 3 status normal * 2 normal * 2 wdt count oscillation stops standby * 1 interrupt request wdt overflow wakeup notes: * 1 standby: lh (status1 low, status0 high) * 2 normal: ll (status1 low, status0 low) * 3 ckio2 output can only be used in clock modes 0, 1, and 2. figure 9.4 standby to interrupt status output 239 standby to power-on reset ckio, ckio2 * 7 status normal * 5 normal * 5 oscillation stops standby * 4 0 to 10 bcyc * 6 0 to 30 bcyc * 6 reset reset * 3 resetp * 1 * 2 notes: * 1 when standby mode is cleared with a power-on reset, the wdt does not count. keep resetp low during the plls oscillation settling time. * 2 undefined * 3 reset: hh (status1 high, status0 high) * 4 standby: lh (status1 low, status0 high) * 5 normal: ll (status1 low, status0 low) * 6 bcyc: bus clock cycle * 7 ckio2 output can only be used in clock modes 0, 1, and 2. figure 9.5 standby to power-on reset status output 240 standby to manual reset ckio, ckio2 * 6 status normal * 4 normal * 4 oscillation stops standby * 3 reset * 2 0 to 20 bcyc * 5 reset resetm * 1 notes: * 1 when standby mode is cleared with a manual reset, the wdt does not count. keep resetm low during the plls oscillation settling time. * 2 reset: hh (status1 high, status0 high) * 3 standby: lh (status1 low, status0 high) * 4 normal: ll (status1 low, status0 low) * 5 bcyc: bus clock cycle * 6 ckio2 output can only be used in clock modes 0, 1, and 2. figure 9.6 standby to manual reset status output 241 9.6.3 timing for canceling sleep mode sleep to interrupt ckio, ckio2 * 3 status normal * 2 normal * 2 sleep * 1 interrupt request notes: * 1 sleep: hl (status1 high, status0 low) * 2 normal: ll (status1 low, status0 low) * 3 ckio2 output can only be used in clock modes 0, 1, and 2. figure 9.7 sleep to interrupt status output sleep to power-on reset ckio, ckio2 * 7 status normal * 5 normal * 5 sleep * 4 0 to 10 bcyc * 6 0 to 30 bcyc * 6 reset reset * 3 * 2 resetp * 1 notes: * 1 when the pll1s multiplication ratio is changed by a power-on reset, keep resetp low during the plls oscillation settling time. * 2 undefined * 3 reset: hh (status1 high, status0 high) * 4 sleep: hl (status1 high, status0 low) * 5 normal: ll (status1 low, status0 low) * 6 bcyc: bus clock cycle * 7 ckio2 output can only be used in clock modes 0, 1, and 2. figure 9.8 sleep to power-on reset status output 242 sleep to manual reset ckio, ckio2 * 6 0 to 80 bcyc * 5 0 to 30 bcyc * 5 reset status normal * 4 normal * 4 sleep * 3 reset * 2 resetm * 1 notes: * 1 keep resetm low until status becomes reset. * 2 reset: hh (status1 high, status0 high) * 3 sleep: hl (status1 high, status0 low) * 4 normal: ll (status1 low, status0 low) * 5 bcyc: bus clock cycle * 6 ckio2 output can only be used in clock modes 0, 1, and 2. figure 9.9 sleep to manual reset status output 243 9.7 hardware standby mode 9.7.1 transition to hardware standby mode driving the ca pin low causes a transition to hardware standby mode. in hardware standby mode, all modules except those operating on an rtc clock are halted, as in the standby mode entered on execution of a sleep instruction ((software) standby mode). hardware standby mode differs from (software) standby mode as follows. 1. interrupts and manual resets are not accepted. 2. the tmu does not operate. operation when a low-level signal is input at the ca pin depends on the cpg state, as follows. 1. in standby mode the clock remains stopped and the chip enters the hardware standby state. acceptance of interrupts and manual resets is disabled, tclk output is fixed low, and the tmu halts. 2. during wdt operation when standby mode is canceled by an interrupt the chip enters hardware standby mode after standby mode is canceled and the cpu resumes operation. 3. in sleep mode the chip enters hardware standby mode after sleep mode is canceled and the cpu resumes operation. hold the ca pin low in hardware standby mode. 9.7.2 canceling hardware standby mode hardware standby mode can only be canceled by a power-on reset. when the ca pin is driven high while the resetp pin is low, clock oscillation is started. hold the resetp pin low until clock oscillation stabilizes. when the resetp pin is driven high, the cpu begins power-on reset processing. operation is not guaranteed in the event of an interrupt or manual reset. 244 9.7.3 hardware standby mode timing figures 9.10 and 9.11 show examples of pin timing in hardware standby mode. the ca pin is sampled using extal2 (32.768 khz), and a hardware standby request is only recognized when the pin is low for two consecutive clock cycles. the ca pin must be held low while the chip is in hardware standby mode. clock oscillation starts when the ca pin is driven high after the resetp pin is driven low. normal * 3 status ca ckio, ckio2 * 6 standby * 2 reset * 1 resetp undefined rcyc: extal2 (32.768 khz) cycle 2 rcyc or more * 5 0C10bcyc * 4 notes: * 1 reset: hh (status1 high, status0 high) * 2 standby: lh (status1 low, status0 high) * 3 normal: ll (status1 low, status0 low) * 4 bcyc: bus clock cycle * 5 rcyc: extal2 (32.768 khz) cycle * 6 ckio2 output can only be used in clock modes 0, 1, and 2. figure 9.10 hardware standby mode (when ca goes low in normal operation) 245 normal * 3 status ca ckio, ckio2 * 6 standby * 2 reset * 1 resetp undefined 2 rcyc or more * 5 0C10 bcyc * 4 standby * 2 wdt operation notes: * 1 reset: hh (status1 high, status0 high) * 2 standby: lh (status1 low, status0 high) * 3 normal: ll (status1 low, status0 low) * 4 bcyc: bus clock cycle * 5 rcyc: extal2 (32.768 khz) cycle * 6 ckio2 output can only be used in clock modes 0, 1, and 2. figure 9.11 hardware standby mode timing (when ca goes low during wdt operation on standby mode cancellation) 246 247 section 10 on-chip oscillation circuits 10.1 overview the clock pulse generator (cpg) supplies all clocks to the processor and controls the power-down modes. the watchdog timer (wdt) is a single-channel timer that counts the clock settling time and is used when clearing standby mode and temporary standbys, such as frequency changes. it can also be used as an ordinary watchdog timer or interval timer. 10.1.1 features the cpg has the following features: ? four clock modes: selection of four clock modes for different frequency ranges, power consumption, direct crystal input, and external clock input. ? three clocks generated independently: an internal clock for the cpu, cache, and tlb (i ); a peripheral clock (p ) for the on-chip peripheral modules; and a bus clock (ckio) for the external bus interface. ? frequency change function: internal and peripheral clock frequencies can be changed independently using the pll circuit and divider circuit within the cpg. frequencies are changed by software using frequency control register (frqcr) settings. ? power-down mode control: the clock can be stopped for sleep mode and standby mode and specific modules can be stopped using the module standby function. the wdt has the following features: ? can be used to ensure the clock settling time: use the wdt to cancel standby mode and the temporary standbys which occur when the clock frequency is changed. ? can switch between watchdog timer mode and interval timer mode. ? generates internal resets in watchdog timer mode: internal resets occur after counter overflow. selection of power-on reset or manual reset. ? generates interrupts in interval timer mode: internal timer interrupts occur after counter overflow. ? selection of eight counter input clocks. eight clocks ( 1 to 1/4096) can be obtained by dividing the peripheral clock. 248 10.2 overview of cpg 10.2.1 cpg block diagram a block diagram of the on-chip clock pulse generator is shown in figure 10.1. cap1 ckio cycle = bcyc cap2 xtal extal md2 md1 md0 frqcr internal bus bus interface stbcr pll circuit 1 ( 1, 2, 3, 4, 6) divider 1 internal clock (i ) cycle = icyc peripheral clock (p ) cycle = pcyc standby control divider 2 clock pulse generator pll circuit 2 ( 1, 4) crystal oscillator cpg control unit clock frequency control circuit standby control circuit 1 1/2 1/3 1/4 1/6 1 1/2 1/3 1/4 1/6 legend frqcr: frequency control register stbcr: standby control register figure 10.1 block diagram of clock pulse generator 249 the clock pulse generator blocks function as follows: 1. pll circuit 1: pll circuit 1 doubles, triples, quadruples, sextuples, or leaves unchanged the input clock frequency from the ckio pin. the multiplication rate is set by the frequency control register. when this is done, the phase of the leading edge of the internal clock is controlled so that it will agree with the phase of the leading edge of the ckio pin. 2. pll circuit 2: pll circuit 2 leaves unchanged or quadruples the frequency of the crystal oscillator or the input clock frequency from the extal pin. the multiplication ratio is fixed by the clock operation mode. the clock operation mode is set by pins md0, md1, and md2. see table 10.3 for more information on clock operation modes. 3. crystal oscillator: this oscillator is used when a crystal oscillator element is connected to the xtal and extal pins. it operates according to the clock operating mode setting. 4. divider 1: divider 1 generates a clock at the operating frequency used by the internal clock. the operating frequency can be 1, 1/2, 1/3, 1/4, or 1/6 times the output frequency of pll circuit 1, as long as it is not lower than the ckio pin clock frequency. the division ratio is set in the frequency control register. 5. divider 2: divider 2 generates a clock at the operating frequency used by the peripheral clock. the operating frequency can be 1, 1/2, 1/3, 1/4, or 1/6 times the output frequency of pll circuit 1 or the ckio pin clock frequency, as long as it is not higher than the ckio pin clock frequency. the division ratio is set in the frequency control register. 6. clock frequency control circuit: the clock frequency control circuit controls the clock frequency using the md pins and the frequency control register. 7. standby control circuit: the standby control circuit controls the state of the clock pulse generator and other modules during clock switching and sleep/standby modes. 8. frequency control register: the frequency control register has control bits assigned for the following functions: clock output/non-output from the ckio pin, on/off control of pll circuit 1, pll standby, the frequency multiplication ratio of pll 1, and the frequency division ratio of the internal clock and the peripheral clock. 9. standby control register: the standby control register has bits for controlling the power-down modes. see section 9, power-down modes, for more information. 250 10.2.2 cpg pin configuration table 10.1 lists the cpg pins and their functions. table 10.1 cpg pins and functions pin name symbol i/o description mode control md0 i set the clock operating mode. pins md1 i md2 i crystal i/o pins xtal o connects a crystal oscillator. (clock input pins) extal i connects a crystal oscillator. also used to input an external clock. clock i/o pin ckio i/o inputs or outputs an external clock. level can be fixed during output. capacitor connection pins cap1 i connects capacitor for pll circuit 1 operation (recommended value 470 pf). for pll cap2 i connects capacitor for pll circuit 2 operation (recommended value 470 pf). 10.2.3 cpg register configuration table 10.2 shows the cpg register configuration. table 10.2 cpg register register name abbreviation r/w initial value address access size frequency control register frqcr r/w h'0102 h'ffffff80 16 251 10.3 clock operating modes table 10.3 shows the relationship between the mode control pin (md2Cmd0) combinations and the clock operating modes. table 10.4 shows the usable frequency ranges in the clock operating modes. table 10.3 clock operating modes pin values clock i/o pll2 pll1 divider 1 divider 2 ckio mode md2 md1 md0 source output on/off on/off input input frequency 0 0 0 0 extal ckio on, multi- plication ratio: 1 on pll1 output pll1 (extal) 1 0 0 1 extal ckio on, multi- plication ratio: 4 on pll1 output pll1 (extal) 4 2 0 1 0 crystal oscillator ckio on, multi- plication ratio: 4 on pll1 output pll1 (crystal) 4 7 1 1 1 ckio off on pll1 output pll1 (ckio) values except above reserved (setting prohibited) mode 0: an external clock is input from the extal pin and undergoes waveform shaping by pll circuit 2 before being supplied inside the chip. pll circuit 1 is constantly on, and there are no frequency range restrictions compared to mode 3. an input clock frequency of 25 mhz to 66.67 mhz can be used, and the ckio frequency range is 25 mhz to 66.67 mhz. as pll circuit 1 compensates for fluctuations in the ckio pin load, this mode is suitable for connection of synchronous dram. mode 1: an external clock is input from the extal pin and its frequency is multiplied by 4 by pll circuit 2 before being supplied inside the chip, allowing a low-frequency external clock to be used. an input clock frequency of 6.25 mhz to 16.67 mhz can be used, and the ckio frequency range is 25 mhz to 66.67 mhz. as pll circuit 1 compensates for fluctuations in the ckio pin load, this mode is suitable for connection of synchronous dram. 252 mode 2: the on-chip crystal oscillator operates, with the oscillation frequency being multiplied by 4 by pll circuit 2 before being supplied inside the chip, allowing a low crystal frequency to be used. a crystal oscillation frequency of 6.25 mhz to 16.67 mhz can be used, and the ckio frequency range is 25 mhz to 66.67 mhz. as pll circuit 1 compensates for fluctuations in the ckio pin load, this mode is suitable for connection of synchronous dram. mode 7: in this mode, the ckio pin is an input, an external clock is input to this pin, and undergoes waveform shaping, and also frequency multiplication according to the setting, by pll circuit 1 before being supplied to the chip. in modes 0 to 4, the system clock is generated from the output of the chips ckio pin. consequently, if a large number of ics are operating on the clock cycle, the ckio pin load will be large. this mode, however, assumes a comparatively large-scale system. if a large number of ics are operating on the clock cycle, a clock generator with a number of low-skew clock outputs can be provided, so that the ics can operate synchronously by distributing the clocks to each one. as pll circuit 1 compensates for fluctuations in the ckio pin load, this mode is suitable for connection of synchronous dram. table 10.4 available combinations of clock mode and frqcr values clock mode frqcr pll1 pll2 clock rate* (i:b:p) input frequency range ckio frequency range 0 h'0100 on ( 1) on ( 1) 1:1:1 25 mhz to 33.34 mhz 25 mhz to 33.34 mhz h'0101 on ( 1) on ( 1) 1:1:1/2 25 mhz to 66.67 mhz 25 mhz to 66.67 mhz h'0102 on ( 1) on ( 1) 1:1:1/4 25 mhz to 66.67 mhz 25 mhz to 66.67 mhz h'0111 on ( 2) on ( 1) 2:1:1 25 mhz to 33.34 mhz 25 mhz to 33.34 mhz h'0112 on ( 2) on ( 1) 2:1:1/2 25 mhz to 66.67 mhz 25 mhz to 66.67 mhz h'0115 on ( 2) on ( 1) 1:1:1 25 mhz to 33.34 mhz 25 mhz to 33.34 mhz h'0116 on ( 2) on ( 1) 1:1:1/2 25 mhz to 66.67 mhz 25 mhz to 66.67 mhz h'0122 on ( 4) on ( 1) 4:1:1 25 mhz to 33.34 mhz 25 mhz to 33.34 mhz h'0126 on ( 4) on ( 1) 2:1:1 25 mhz to 33.34 mhz 25 mhz to 33.34 mhz h'012a on ( 4) on ( 1) 1:1:1 25 mhz to 33.34 mhz 25 mhz to 33.34 mhz h'a100 on ( 3) on ( 1) 3:1:1 25 mhz to 33.34 mhz 25 mhz to 33.34 mhz h'a101 on ( 3) on ( 1) 3:1:1/2 25 mhz to 66.67 mhz 25 mhz to 66.67 mhz h'e100 on ( 3) on ( 1) 1:1:1 25 mhz to 33.34 mhz 25 mhz to 33.34 mhz h'e101 on ( 3) on ( 1) 1:1:1/2 25 mhz to 66.67 mhz 25 mhz to 66.67 mhz h'a111 on ( 6) on ( 1) 6:1:1 25 mhz to 33.34 mhz 25 mhz to 33.34 mhz 253 table 10.4 available combinations of clock mode and frqcr values (cont) clock mode frqcr pll1 pll2 clock rate* (i:b:p) input frequency range ckio frequency range 1, 2 h'0100 on ( 1) on ( 4) 4:4:4 6.25 mhz to 8.34 mhz 25 mhz to 33.34 mhz h'0101 on ( 1) on ( 4) 4:4:2 6.25 mhz to 16.67 mhz 25 mhz to 66.67 mhz h'0102 on ( 1) on ( 4) 4:4:1 6.25 mhz to 16.67 mhz 25 mhz to 66.67 mhz h'0111 on ( 2) on ( 4) 8:4:4 6.25 mhz to 8.34 mhz 25 mhz to 33.34 mhz h'0112 on ( 2) on ( 4) 8:4:2 6.25 mhz to 16.67 mhz 25 mhz to 66.67 mhz h'0115 on ( 2) on ( 4) 4:4:4 6.25 mhz to 8.34 mhz 25 mhz to 33.34 mhz h'0116 on ( 2) on ( 4) 4:4:2 6.25 mhz to 16.67 mhz 25 mhz to 66.67 mhz h'0122 on ( 4) on ( 4) 16:4:4 6.25 mhz to 8.34 mhz 25 mhz to 33.34 mhz h'0126 on ( 4) on ( 4) 8:4:4 6.25 mhz to 8.34 mhz 25 mhz to 33.34 mhz h'012a on ( 4) on ( 4) 4:4:4 6.25 mhz to 8.34 mhz 25 mhz to 33.34 mhz h'a100 on ( 3) on ( 4) 12:4:4 6.25 mhz to 8.34 mhz 25 mhz to 33.34 mhz h'a101 on ( 3) on ( 4) 12:4:2 6.25 mhz to 16.67 mhz 25 mhz to 66.67 mhz h'e100 on ( 3) on ( 4) 4:4:4 6.25 mhz to 8.34 mhz 25 mhz to 33.34 mhz h'e101 on ( 3) on ( 4) 4:4:2 6.25 mhz to 16.67 mhz 25 mhz to 66.67 mhz h'a111 on ( 6) on ( 4) 24:4:4 6.25 mhz to 8.34 mhz 25 mhz to 33.34 mhz 7 h'0100 on ( 1) off 1:1:1 25 mhz to 33.34 mhz 25 mhz to 33.34 mhz h'0101 on ( 1) off 1:1:1/2 25 mhz to 66.67 mhz 25 mhz to 66.67 mhz h'0102 on ( 1) off 1:1:1/4 25 mhz to 66.67 mhz 25 mhz to 66.67 mhz h'0111 on ( 2) off 2:1:1 25 mhz to 33.34 mhz 25 mhz to 33.34 mhz h'0112 on ( 2) off 2:1:1/2 25 mhz to 66.67 mhz 25 mhz to 66.67 mhz h'0115 on ( 2) off 1:1:1 25 mhz to 33.34 mhz 25 mhz to 33.34 mhz h'0116 on ( 2) off 1:1:1/2 25 mhz to 66.67 mhz 25 mhz to 66.67 mhz h'0122 on ( 4) off 4:1:1 25 mhz to 33.34 mhz 25 mhz to 33.34 mhz h'0126 on ( 4) off 2:1:1 25 mhz to 33.34 mhz 25 mhz to 33.34 mhz h'012a on ( 4) off 1:1:1 25 mhz to 33.34 mhz 25 mhz to 33.34 mhz h'a100 on ( 3) off 3:1:1 25 mhz to 33.34 mhz 25 mhz to 33.34 mhz h'a101 on ( 3) off 3:1:1/2 25 mhz to 66.67 mhz 25 mhz to 66.67 mhz h'e100 on ( 3) off 1:1:1 25 mhz to 33.34 mhz 25 mhz to 33.34 mhz h'e101 on ( 3) off 1:1:1/2 25 mhz to 66.67 mhz 25 mhz to 66.67 mhz h'a111 on ( 6) off 6:1:1 25 mhz to 33.34 mhz 25 mhz to 33.34 mhz note: * taking input clock as 1 254 cautions: 1. the input to divider 1 becomes the output of pll circuit 1 when pll circuit 1 is on. 2. the input of divider 2 becomes the output of pll circuit 1 when the clock operating mode is 0C2 or 7. 3. the frequency of the internal clock (i ) becomes: ? the product of the frequency of the ckio pin, the frequency multiplication ratio of pll circuit 1, and the division ratio of divider 1 when pll circuit 1 is on. ? do not set the internal clock frequency lower than the ckio pin frequency. 4. the frequency of the peripheral clock (p ) becomes: ? the product of the frequency of the ckio pin, the frequency multiplication ratio of pll circuit 1, and the division ratio of divider 2 when the clock operating mode is 0C2 or 7. ? the peripheral clock frequency should not be set higher than the frequency of the ckio pin, higher than 33 mhz, or lower than 1/8 the internal clock (i ). 5. the output frequency of pll circuit 1 is the product of the ckio frequency and the multiplication ratio of pll circuit 1. 6. 1, 2, 3, 4, or 6 can be used as the multiplication ratio of pll circuit 1. 1, 1/2, 1/ 3, 1/4, and 1/6 can be selected as the division ratios of dividers 1 and 2. set the rate in the frequency control register. the on/off state of pll circuit 2 is determined by the mode. 255 10.4 register descriptions 10.4.1 frequency control register (frqcr) the frequency control register (frqcr) is a 16-bit readable/writable register used to specify the frequency multiplication ratio of pll circuit 1, and the frequency division ratio of the internal clock and the peripheral clock. only word access can be used on the frqcr register. frqcr is initialized to h'0102 by a power-on reset, but retains its value in a manual reset and in standby mode. frqcr: bit: 15 14 13 12 11 10 9 8 stc2 ifc2 pfc2 initial value: 0 0 0 0 0 0 0 1 r/w: r/w r/w r/w r r r r r bit: 7 6 5 4 3 2 1 0 stc1 stc0 ifc1 ifc0 pfc1 pfc0 initial value: 0 0 0 0 0 0 1 0 r/w: r r r/w r/w r/w r/w r/w r/w bits 15, 5, and 4frequency multiplication ratio (stc2, stc1, stc0): these bits specify the frequency multiplication ratio of pll circuit 1. bit 15: stc2 bit 5: stc1 bit 4: stc0 description 000 1 (initial value) 001 2 100 3 010 4 101 6 values except above reserved (setting prohibited) 256 bits 14, 3, and 2internal clock frequency division ratio (ifc2, ifc1, ifc0): these bits specify the frequency division ratio of the internal clock with respect to the output frequency of pll circuit 1. bit 14: ifc2 bit 3: ifc1 bit 2: ifc0 description 000 1 (initial value) 001 1/2 100 1/3 010 1/4 values except above reserved (setting prohibited) note: do not set the internal clock frequency lower than the ckio pin frequency. bits 13, 1, and 0peripheral clock frequency division ratio (pfc2, pfc1, pfc0): these bits specify the division ratio of the peripheral clock frequency with respect to the frequency of the output frequency of pll circuit 1 or the frequency of the ckio pin. bit 13: pfc2 bit 1: pfc1 bit 0: pfc0 description 000 1 001 1/2 100 1/3 010 1/4 (initial value) 101 1/6 values except above reserved (setting prohibited) note: do not set the peripheral clock frequency higher than the ckio pin frequency. bits 12 to 9, 7, and 6reserved: these bits are always read as 0. the write value should always be 0. bit 8reserved: this bit is always read as 1. the write value should always be 1. 257 10.5 changing the frequency the frequency of the internal clock and peripheral clock can be changed either by changing the multiplication ratio of pll circuit 1 or by changing the division ratios of dividers 1 and 2. all of these are controlled by software through the frequency control register. the methods are described below. 10.5.1 changing the multiplication rate a pll settling time is required when the multiplication rate of pll circuit 1 is changed. the on- chip wdt counts the settling time. 1. in the initial state, the multiplication rate of pll circuit 1 is 1. 2. set a value that will become the specified oscillation settling time in the wdt and stop the wdt. the following must be set: wtcsr register tme bit = 0: wdt stops wtcsr register cks2Ccks0 bits: division ratio of wdt count clock wtcnt counter: initial counter value 3. set the desired value in the stc2, stc1, and stc0 bits. the division ratio can also be set in the ifc2Cifc0 bits and pfc2Cpfc0 bits. 4. the processor pauses internally and the wdt starts incrementing. in clock modes 0C2 and 7, the internal and peripheral clocks both stop (except for the peripheral clock supplied to the wdt). 5. supply of the clock that has been set begins at wdt count overflow, and the processor begins operating again. the wdt stops after it overflows. 10.5.2 changing the division ratio the wdt will not count unless the multiplication ratio is changed simultaneously. 1. in the initial state, ifc2Cifc0 = 000 and pfc2Cpfc0 = 010. 2. set the ifc2, ifc1, ifc0, pfc2, pfc1, and pfc0 bits to the new division ratio. the values that can be set are limited by the clock mode and the multiplication ratio of pll circuit 1. note that if the wrong value is set, the processor will malfunction. 3. the clock is immediately supplied at the new division ratio. 258 10.6 overview of wdt 10.6.1 block diagram of wdt figure 10.2 shows a block diagram of the wdt. wtcsr standby control bus interface wtcnt divider clock selector clock internal bus standby mode peripheral clock standby cancellation reset control clock selection wdt overflow internal reset request interrupt control interrupt request wtcsr: wtcnt: legend watchdog timer control/status register watchdog timer counter figure 10.2 block diagram of wdt 10.6.2 register configuration the wdt has two registers that select the clock, switch the timer mode, and perform other functions. table 10.5 shows the wdt registers. table 10.5 register configuration name abbreviation r/w access size initial value address watchdog timer counter wtcnt r/w* r: byte; w: word* h'00 h'ffffff84 watchdog timer control/status register wtcsr r/w* r: byte; w: word* h'00 h'ffffff86 note: * write with word access. write with h'5a and h'a5, respectively, in the upper byte. byte or longword writes are not possible. read with byte access. 259 10.7 wdt registers 10.7.1 watchdog timer counter (wtcnt) the watchdog timer counter (wtcnt) is an 8-bit readable/writable counter that increments on the selected clock. wtcnt differs from other registers in that it is more difficult to write to. see section 10.9.3, notes on register access, for details. when an overflow occurs, it generates a reset in watchdog timer mode and an interrupt in interval time mode. its address is h'ffffff84. the wtcnt counter is initialized to h'00 only by a power-on reset through the resetp pin. use word access to write to the wtcnt counter, with h'5a in the upper byte. use byte access to read wtcnt. bit: 7 6 5 4 3 2 1 0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w 10.7.2 watchdog timer control/status register (wtcsr) the watchdog timer control/status register (wtcsr) is an 8-bit readable/writable register composed of bits to select the clock used for the count, bits to select the timer mode, and overflow flags. wtcsr differs from other registers in that it is more difficult to write to. see section 10.9.3, notes on register access, for details. its address is h'ffffff86. the wtcsr register is initialized to h'00 only by a power-on reset through the resetp pin. when a wdt overflow causes an internal reset, wtcsr retains its value. when used to count the clock settling time for canceling a standby, it retains its value after counter overflow. use word access to write to the wtcsr counter, with h'a5 in the upper byte. use byte access to read wtcsr. bit: 7 6 5 4 3 2 1 0 tme wt/ it rsts wovf iovf cks2 cks1 cks0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit 7timer enable (tme): starts and stops timer operation. clear this bit to 0 when using the wdt in standby mode or when changing the clock frequency. bit 7: tme description 0 timer disabled: count-up stops and wtcnt value is retained (initial value) 1 timer enabled 260 bit 6timer mode select (wt/ it ): selects whether to use the wdt as a watchdog timer or an interval timer. bit 6: wt/ it description 0 used as interval timer (initial value) 1 used as watchdog timer note: if wt/ it is modified when the wdt is running, the up-count may not be performed correctly. bit 5reset select (rsts): selects the type of reset when wtcnt overflows in watchdog timer mode. in interval timer mode, this setting is ignored. bit 5: rsts description 0 power-on reset (initial value) 1 manual reset note: resetout is output. bit 4watchdog timer overflow (wovf): indicates that the wtcnt has overflowed in watchdog timer mode. this bit is not set in interval timer mode. bit 4: wovf description 0 no overflow (initial value) 1 wtcnt has overflowed in watchdog timer mode bit 3interval timer overflow (iovf): indicates that wtcnt has overflowed in interval timer mode. this bit is not set in watchdog timer mode. bit 3: iovf description 0 no overflow (initial value) 1 wtcnt has overflowed in interval timer mode bits 2 to 0clock select 2 to 0 (cks2 to cks0): these bits select the clock to be used for the wtcnt count from the eight types obtainable by dividing the peripheral clock. the overflow period in the table is the value when the peripheral clock (p ) is 15 mhz. 261 bit 2: cks2 bit 1: cks1 bit 0: cks0 clock division ratio overflow period (when p = 15 mhz) 0001 (initial value) 17 s 1 1/4 68 s 1 0 1/16 273 s 1 1/32 546 s 1 0 0 1/64 1.09 ms 1 1/256 4.36 ms 1 0 1/1024 17.48 ms 1 1/4096 69.91 ms note: if bits cks2Ccks0 are modified when the wdt is running, the up-count may not be performed correctly. ensure that these bits are modified only when the wdt is not running. 10.7.3 notes on register access the watchdog timer counter (wtcnt) and watchdog timer control/status register (wtcsr) are more difficult to write to than other registers. the procedure for writing to these registers is given below. writing to wtcnt and wtcsr: these registers must be written to using a word transfer instruction. they cannot be written to with a byte or longword transfer instruction. when writing to wtcnt, set the upper byte to h'5a and transfer the lower byte as the write data, as shown in figure 10.3. when writing to wtcsr, set the upper byte to h'a5 and transfer the lower byte as the write data. this transfer procedure writes the lower byte data to wtcnt or wtcsr. 15 8 7 0 h'5a write data address: h'fffffe84 wtcnt write 15 8 7 0 h'a5 write data address: h'fffffe86 wtcsr write figure 10.3 writing to wtcnt and wtcsr 262 10.8 using the wdt 10.8.1 canceling standby the wdt can be used to cancel standby mode with an nmi or other interrupt. the procedure is described below. (the wdt does not run when a reset is used for canceling, so keep the reset pin low until the clock stabilizes.) 1. before transitioning to standby mode, always clear the tme bit in wtcsr to 0. when the tme bit is 1, an erroneous reset or interval timer interrupt may be generated when the count overflows. 2. set the type of count clock used in the cks2Ccks0 bits in wtcsr and the initial values for the counter in the wtcnt counter. these values should ensure that the time till count overflow is longer than the clock oscillation settling time. 3. switch to standby mode by executing a sleep instruction to stop the clock. 4. the wdt starts counting by detecting the edge change of the nmi signal or detecting interrupts. 5. when the wdt count overflows, the cpg starts supplying the clock and the processor resumes operation. the wovf flag in wtcsr is not set when this happens. 6. since the wdt continues counting from h'00, set the stby bit in the stbcr register to 0 in the interrupt handling routine and this will stop the wdt. when the stby bit remains at 1, the sh7729r again enters standby mode when the wdt has counted up to h'80. this standby mode can be canceled by a power-on reset. 10.8.2 changing the frequency to change the frequency used by the pll, use the wdt. when changing the frequency only by switching the divider, do not use the wdt. 1. before changing the frequency, always clear the tme bit in wtcsr to 0. when the tme bit is 1, an erroneous reset or interval timer interrupt may be generated when the count overflows. 2. set the type of count clock used in the cks2Ccks0 bits of wtcsr and the initial values for the counter in the wtcnt counter. these values should ensure that the time till count overflow is longer than the clock oscillation settling time. 3. when the frequency control register (frqcr) is written to, the clock stops and the processor enters standby mode temporarily. the wdt starts counting. 4. when the wdt count overflows, the cpg resumes supplying the clock and the processor resumes operation. the wovf flag in wtcsr is not set when this happens. 5. the counter stops at a value of h'00 or h'01. the stop value depends on the clock ratio. 263 10.8.3 using watchdog timer mode 1. set the wt/ it bit in the wtcsr register to 1, set the reset type in the rsts bit, set the type of count clock in the cks2Ccks0 bits, and set the initial value of the counter in the wtcnt counter. 2. set the tme bit in wtcsr to 1 to start the count in watchdog timer mode. 3. while operating in watchdog timer mode, rewrite the counter periodically to h'00 to prevent the counter from overflowing. 4. when the counter overflows, the wdt sets the wovf flag in wtcsr to 1 and generates the type of reset specified by the rsts bit. the counter then resumes counting. when a reset is generated, a low level is output at the resetout pin, and a high level at the status0 and status1 pins. the output period is approximately 1 count clock cycle in the case of a power-on reset, and approximately 5 peripheral clock cycles in the case of a manual reset. 10.8.4 using interval timer mode when operating in interval timer mode, interval timer interrupts are generated at every overflow of the counter. this enables interrupts to be generated at set periods. 1. clear the wt/ it bit in the wtcsr register to 0, set the type of count clock in the cks2C cks0 bits, and set the initial value of the counter in the wtcnt counter. 2. set the tme bit in wtcsr to 1 to start the count in interval timer mode. 3. when the counter overflows, the wdt sets the iovf flag in wtcsr to 1 and an interval timer interrupt request is sent to the intc. the counter then resumes counting. 264 10.9 notes on board design when using an external crystal resonator: place the crystal resonator, capacitors cl1 and cl2, and damping resistor r close to the extal and xtal pins. to prevent induction from interfering with correct oscillation, use a common grounding point for the capacitors connected to the resonator, and do not locate a wiring pattern near these components. note: the values for cl1, cl2, and the damping resistance should be determined after consultation with the crystal manufacturer. xtal extal sh7729r r cl2 cl1 avoid crossing signal lines figure 10.4 points for attention when using crystal resonator decoupling capacitors: insert a laminated ceramic capacitor of 0.01 to 0.1 f as a passive capacitor for each v ss /v cc pair. mount the passive capacitors close to the sh7729r power supply pins, and use components with a frequency characteristic suitable for the chips operating frequency, as well as a suitable capacitance value. digital system v ss /v cc pairs: 19-21, 27-29, 33-35, 45-47, 57-59, 69-71, 79-81, 83-85, 95-97, 109- 111, 132-134, 153-154, 161-163, 173-175, 181-183, 205-208 on-chip oscillator v ss /v cc pairs: 3-6, 145-147, 148-150 when using a pll oscillator circuit: keep the wiring from the pll v cc and v ss connection pattern to the power supply pins short, and make the pattern width large, to minimize the inductance component. ground the oscillation stabilization capacitors c1 and c2 to v ss (pll1) and v ss (pll2), respectively. place c1 and c2 close to the cap1 and cap2 pins and do not locate a wiring pattern in the vicinity. in clock mode 7, connect the extal pin to v cc or v ss and leave the xtal pin open. 265 cap2 v cc (pll2) v cc (pll1) v cc c1 = 470 pf c2 = 470 pf v ss cap1 v ss (pll2) v ss (pll1) avoid crossing signal lines power supply reference values c2 c1 figure 10.5 points for attention when using pll oscillator circuit 266 267 section 11 bus state controller (bsc) 11.1 overview the bus state controller (bsc) divides physical address space and output control signals for various types of memory and bus interface specifications. bsc functions enable the chip to link directly with synchronous dram, sram, rom, and other memory storage devices without an external circuit. the bsc also allows direct connection to pcmcia interfaces, simplifying system design and allowing high-speed data transfers in a compact system. 11.1.1 features the bsc has the following features: ? physical address space is divided into six areas ? a maximum 64 mbytes for each of the six areas, 0, 2C6 ? area bus width can be selected by register (area 0 is set by external pin) ? wait states can be inserted using the wait pin ? wait state insertion can be controlled through software. register settings can be used to specify the insertion of 1C10 cycles independently for each area (1C38 cycles for areas 5 and 6 and the pcmcia interface only) ? the type of memory connected can be specified for each area, and control signals are output for direct memory connection ? wait cycles are automatically inserted to avoid data bus conflict for continuous memory accesses to different areas or writes directly following reads in the same area ? direct interface to synchronous dram ? multiplexes row/column addresses according to synchronous dram capacity ? supports burst operation ? supports bank active mode ? has both auto-refresh and self-refresh functions ? controls timing of synchronous dram direct-connection control signals according to register setting ? burst rom interface ? insertion of wait states controllable through software ? register setting control of burst transfers ? pcmcia direct-connection interface ? insertion of wait states controllable through software 268 ? bus sizing function for i/o bus width (only in little-endian mode) ? refresh function ? refresh cycles will be automatically maintained in sleep mode even after the external bus frequency is reduced to 1/4 of its normal operating frequency ? short refresh cycle control ? the overflow interrupt function of the refresh counter enables the refresh function immediately after a self-refresh operation using low power-consumption dram ? the refresh counter can be used as an interval timer ? outputs an interrupt request signal using the compare-match function ? outputs an interrupt request signal when the refresh counter overflows ? automatically disables the output of clock signals to anywhere but the refresh counter, except during execution of external bus cycles 269 11.1.2 block diagram figure 11.1 shows a block diagram of the bus state controller. wcr1 wcr2 bcr1 module bus mcr bsc rfcr rtcnt comparator refresh controller peripheral bus internal bus interrupt controller memory controller area controller wait controller wait cs0 , cs6 C cs2 , ce2a , ce2b mcs0 C mcs7 bs rd rd/ wr we3 C we0 rasxx casx cke iciord , iciowr iois16 wcr: bcr: mcr: pcr: legend bus interface rtcsr rtcor bcr2 pcr mcscrn wait state control register bus control register memory control register pcmcia control register rfcr: rtcnt: rtcor: rtcsr: mcscrn: refresh count register refresh timer count register refresh time constant register refresh timer control/status register mcsn control register (n: 0C7) figure 11.1 block diagram of bus state controller 270 11.1.3 pin configuration table 11.1 shows the bsc pin configuration. table 11.1 bsc pins pin name signal i/o description address bus a25Ca0 o address output data bus d15Cd0 i/o data i/o d31Cd16 i/o data i/o when using 32-bit bus width bus cycle start bs o shows start of bus cycle. during burst transfers, asserted every data cycle. chip select 0, 2C4 cs0 , cs2 C cs4 o chip select signals to indicate area being accessed. chip select 5, 6 cs5 / ce1a , cs6 / ce1b o chip select signals to indicate area being accessed. cs5/ce1a and cs6/ce1b can also be used as ce1a and ce1b of pcmcia. pcmcia card select ce2a , ce2b o ce2a and ce2b signals when pcmcia is used read/write rd/ wr o data bus direction indication signal. pcmcia write indication signal. row address strobe 3l ras3l o when synchronous dram is used in area 3, ras3l for lower 32-mbyte address. row address strobe 3u ras3u o when synchronous dram is used in area 3, ras3u for upper 32-mbyte address. column address strobe casl o when synchronous dram is used, casl signal for lower 32-mbyte address. column address strobe lh casu o when synchronous dram is used, casu signal for upper 32-mbyte address. data enable 0 we0 /dqmll o when memory other than synchronous dram is used, d7Cd0 write strobe signal. when synchronous dram is used, selects d7Cd0. data enable 1 we1 /dqmlu/ we o when memory other than synchronous dram and pcmcia is used, d15Cd8 write strobe signal. when synchronous dram is used, selects d15Cd8. when pcmcia is used, strobe signal indicating write cycle. data enable 2 we2 /dqmul/ iciord o when memory other than synchronous dram and pcmcia is used, d23Cd16 write strobe signal. when synchronous dram is used, selects d23C d16. when pcmcia is used, strobe signal indicating i/o read. 271 table 11.1 bsc pins (cont) pin name signal i/o description data enable 3 we3 /dqmuu/ iciowr o when memory other than synchronous dram and pcmcia is used, d31Cd24 write strobe signal. when synchronous dram is used, selects d31C d24. when pcmcia is used, strobe signal indicating i/o write. read rd o strobe signal indicating read cycle wait wait i wait state request signal clock enable cke o clock enable control signal for synchronous dram iois16 iois16 i signal indicating pcmcia 16-bit i/o. valid only in little-endian mode. bus release request breq i bus release request signal bus release acknowledgment back o bus release acknowledge signal mask rom chip select mcs[0] C mcs[7] o chip select signal for mask rom connected to area 0 or 2. 272 11.1.4 register configuration the bsc has 21 registers (table 11.2). synchronous dram also has a built-in synchronous dram mode register. these registers control direct connection interfaces to memory, wait states, refreshes, and pcmcia devices. table 11.2 bsc registers name abbr. r/w initial value* 1 address bus width bus control register 1 bcr1 r/w h'0000 h'ffffff60 16 bus control register 2 bcr2 r/w h'3ff0 h'ffffff62 16 wait state control register 1 wcr1 r/w h'3ff3 h'ffffff64 16 wait state control register 2 wcr2 r/w h'ffff h'ffffff66 16 individual memory control register mcr r/w h'0000 h'ffffff68 16 pcmcia control register pcr r/w h'0000 h'ffffff6c 16 refresh timer control/status register rtcsr r/w h'0000 h'ffffff6e 16 refresh timer counter rtcnt r/w h'0000 h'ffffff70 16 refresh time constant register rtcor r/w h'0000 h'ffffff72 16 refresh count register rfcr r/w h'0000 h'ffffff74 16 synchronous dram mode register, area 2 sdmr w h'ffffd000C h'ffffdfff 8 synchronous dram mode register, area 3 h'ffffe000C h'ffffefff mcs0 control register mcscr0 r/w h'0000 h'ffffff50 16 mcs1 control register mcscr1 r/w h'0000 h'ffffff52 16 mcs2 control register mcscr2 r/w h'0000 h'ffffff54 16 mcs3 control register mcscr3 r/w h'0000 h'ffffff56 16 mcs4 control register mcscr4 r/w h'0000 h'ffffff58 16 mcs5 control register mcscr5 r/w h'0000 h'ffffff5a 16 mcs6 control register mcscr6 r/w h'0000 h'ffffff5c 16 mcs7 control register mcscr7 r/w h'0000 h'ffffff5e 16 notes: *1 initialized by a power-on reset. *2 for details, see section 11.2.9, synchronous dram mode register. 273 11.1.5 area overview space allocation: in the architecture of the sh7729r, both logical spaces and physical spaces have 32-bit address spaces. the logical space is divided into five areas by the value of the upper bits of the address. the physical space is divided into eight areas. logical space can be allocated to physical space using a memory management unit (mmu). for details, refer to section 3, memory management unit, which describes area allocation for physical space. as shown in table 11.3, the sh7729r can be connected directly to six memory/pcmcia interface areas, and it outputs chip select signals ( cs0 , cs2 C cs6 , ce2a , ce2b ) for each of them. cs0 is asserted during area 0 access; cs6 is asserted during area 6 access. when pcmcia interface is selected in area 5 or 6, in addition to cs5 / cs6 , ce2a / ce2b are asserted for the corresponding bytes accessed. area 0 (cs0) internal i/o area 2 (cs2) area 3 (cs3) area 4 (cs4) area 5 (cs5) area 6 (cs6) h'00000000 h'20000000 h'40000000 h'60000000 h'80000000 h'a0000000 h'c0000000 h'e0000000 h'00000000 h'04000000 h'08000000 h'0c000000 h'10000000 h'14000000 h'18000000 reserved area physical address space logical address space p0, u0 p1 p2 p3 p4 note: for logical address spaces p0 and p3, when the memory management unit (mmu) is on, it can optionally generate a physical address for the logical address. this diagram can be applied when the mmu is off, and when the mmu is on and each physical address corresponding to a logical address is equal except for the upper three bits. figure 11.2 correspondence between logical address space and physical address space 274 table 11.3 physical address space map area connectable memory physical address capacity access size 0 ordinary memory *1 , burst rom h'00000000 to h'03ffffff 64 mbytes 8, 16, 32 *2 h'00000000 + h'20000000 n to h'03ffffff + h'20000000 n shadow n: 1C6 1 internal i/o registers *8 h'04000000 to h'07ffffff 64 mbytes 8, 16, 32 *3 h'04000000 + h'20000000 n to h'07ffffff + h'20000000 n shadow n: 1C6 2 ordinary memory *1 , h'08000000 to h'0bffffff 64 mbytes 8, 16, 32 *3, *4 synchronous dram h'08000000 + h'20000000 n to h'0bffffff + h'20000000 n shadow n: 1C6 3 ordinary memory, h'0c000000 to h'0fffffff 64 mbytes 8, 16, 32 *3, *5 synchronous dram h'0c000000 + h'20000000 n to h'0fffffff + h'20000000 n shadow n: 1C6 4 ordinary memory h'10000000 to h'13ffffff 64 mbytes 8, 16, 32 *3 h'10000000 + h'20000000 n to h'13ffffff + h'20000000 n shadow n: 1C6 5 ordinary memory, pcmcia, burst rom h'14000000 to h'15ffffff 32 mbytes 8, 16, 32 *3, *6 ordinary memory, burst rom h'16000000 to h'17ffffff 32 mbytes h'14000000 + h'20000000 n to h'17ffffff + h'20000000 n shadow n: 1C6 6 ordinary memory, pcmcia, h'18000000 to h'19ffffff 32 mbytes 8, 16, 32 *3, *6 burst rom h'1a000000 to h'1bffffff h'18000000 + h'20000000 n to h'1bffffff + h'20000000 n shadow n: 1C6 7 *7 reserved area h'1c000000 + h'20000000 n to h'1fffffff + h'20000000 n n: 0C7 notes: *1 memory with interface such as sram or rom. *2 use external pin to specify memory bus width. *3 use register to specify memory bus width. *4 with synchronous dram interfaces, bus width must be 16 or 32 bits. *5 with synchronous dram interfaces, bus width must be 16 or 32 bits. *6 with pcmcia interface, bus width must be 8 or 16 bits. *7 do not access the reserved area. if the reserved area is accessed, correct operation cannot be guaranteed. *8 when the control register in area 1 is not used for address translation by the mmu, set the first three bits of the logical address to 101 for allocation to the p2 space. 275 area 0: h'00000000 area 1: h'04000000 area 2: h'08000000 area 3: h'0c000000 area 4: h'10000000 area 5: h'14000000 the pcmcia interface is shared by the memory and i/o card the pcmcia interface is shared by the memory and i/o card area 6: h'18000000 ordinary memory/ burst rom internal i/o ordinary memory/ synchronous dram ordinary memory/ synchronous dram ordinary memory ordinary memory/ burst rom/pcmcia ordinary memory/ burst rom/pcmcia figure 11.3 physical space allocation memory bus width: the memory bus width in the sh7729r can be set for each area. in area 0, external pins can be used to select byte (8 bits), word (16 bits), or longword (32 bits) on power-on reset. the correspondence between the external pins (md4 and md3) and the memory size is shown in table below. table 11.4 correspondence between external pins (md4 and md3) and memory size md4 md3 memory size 0 0 reserved (do not set) 0 1 8 bits 1 0 16 bits 1 1 32 bits for areas 2C6, byte, word, and longword can be chosen for the bus width using bus control register 2 (bcr2) whenever ordinary memory, rom, or burst rom are used. when the synchronous dram interface is used, word or longword can be chosen as the bus width. when the pcmcia interface is used, set the bus width to byte or word. when synchronous dram is connected to both area 2 and area 3, set the same bus width for areas 2 and 3. when using the port function, set each of the bus widths to byte or word for all areas. for more information, see section 11.2.2, bus control register 2 (bcr2). 276 shadow space: areas 0 and 2C6 are decoded by physical addresses a28Ca26, which correspond to areas 000 to 110. address bits 31C29 are ignored. this means that the range of area 0 addresses, for example, is h'00000000 to h'03ffffff, and its corresponding shadow space is the address space obtained by adding to it h'20000000 n (n = 1C6). the address range for area 7, which is on-chip i/o space, is h'1c000000 to h'1fffffff. the address space h'1c000000 + h'20000000 nCh'1fffffff + h'20000000 n (n = 0C7) corresponding to the area 7 shadow space is reserved, and must not be used. 11.1.6 pcmcia support the sh7729r supports pcmcia standard interface specifications in physical space areas 5 and 6. table 11.5 pcmcia interface characteristics item feature access random access data bus 8/16 bits memory type mask rom, otprom, eprom, eeprom, flash memory, sram memory capacity maximum 32 mbytes i/o space capacity maximum 32 mbytes other features dynamic bus sizing of i/o bus width* the pcmcia interface can be accessed from the address translation area or non-address translation area. note: * dynamic bus sizing of the i/o bus width is supported only in little-endian mode. common memory/attribute memory area 5: h'14000000 area 5: h'16000000 common memory/attribute memory area 6: h'18000000 i/o space i/o space area 6: h'1a000000 figure 11.4 pcmcia space allocation 277 table 11.6 pcmcia support interface ic memory card interface i/o card interface pin signal i/o function signal i/o function sh7729r pin 1 gnd ground gnd ground 2 d3 i/o data d3 i/o data d3 3 d4 i/o data d4 i/o data d4 4 d5 i/o data d5 i/o data d5 5 d6 i/o data d6 i/o data d6 6 d7 i/o data d7 i/o data d7 7 ce1 i card enable ce1 i card enable ce1a or ce1b 8 a10 i address a10 i address a10 9 oe i output enable oe i output enable rd 10 a11 i address a11 i address a11 11 a9 i address a9 i address a9 12 a8 i address a8 i address a8 13 a13 i address a13 i address a13 14 a14 i address a14 i address a14 15 we / pgm i write enable we / pgm i write enable we 16 rdy/ bsy o ready/busy ireq o ready/busy 17 v cc operation power v cc operation power 18 vpp1 program power vpp1 program/ peripheral power 19 a16 i address a16 i address a16 20 a15 i address a15 i address a15 21 a12 i address a12 i address a12 22 a7 i address a7 i address a7 23 a6 i address a6 i address a6 24 a5 i address a5 i address a5 25 a4 i address a4 i address a4 26 a3 i address a3 i address a3 27 a2 i address a2 i address a2 28 a1 i address a1 i address a1 29 a0 i address a0 i address a0 30 d0 i/o data d0 i/o data d0 278 table 11.6 pcmcia support interface (cont) ic memory card interface i/o card interface pin signal i/o function signal i/o function sh7729r pin 31 d1 i/o data d1 i/o data d1 32 d2 i/o data d2 i/o data d2 33 wp o write protect iois16 o 16-bit i/o port iois16 34 gnd ground gnd ground 35 gnd ground gnd ground 36 cd1 o card detection cd1 o card detection 37 d11 i/o data d11 i/o data d11 38 d12 i/o data d12 i/o data d12 39 d13 i/o data d13 i/o data d13 40 d14 i/o data d14 i/o data d14 41 d15 i/o data d15 i/o data d15 42 ce2 i card enable ce2 i card enable ce2a or ce2b 43 vs1 i voltage sense 1 vs1 i voltage sense 1 44 rfu reserved iord i i/o read iciord 45 rfu reserved iowr i i/o write iciowr 46 a17 i address a17 i address a17 47 a18 i address a18 i address a18 48 a19 i address a19 i address a19 49 a20 i address a20 i address a20 50 a21 i address a21 i address a21 51 v cc power supply v cc power supply 52 vpp2 program power vpp2 program/ peripheral power 53 a22 i address a22 i address a22 54 a23 i address a23 i address a23 55 a24 i address a24 i address a24 56 a25 i address a25 i address a25 57 vs2 i voltage sense 2 vs2 i voltage sense 2 58 reset i reset reset i reset 59 wait o wait request wait o wait request 60 rfu reserved inpack o input acknowledge 279 table 11.6 pcmcia support interface (cont) ic memory card interface i/o card interface pin signal i/o function signal i/o function sh7729r pin 61 reg i attribute memory space select reg i attribute memory space select 62 bvd2 o battery voltage detection spkr o digital voice signal 63 bvd1 o battery voltage detection stschg o card state change 64 d8 i/o data d8 i/o data d8 65 d9 i/o data d9 i/o data d9 66 d10 i/o data d10 i/o data d10 67 cd2 o card detection cd2 o card detection 68 gnd ground gnd ground 11.2 bsc registers 11.2.1 bus control register 1 (bcr1) bus control register 1 (bcr1) is a 16-bit readable/writable register that sets the functions and bus cycle state for each area. it is initialized to h'0000 by a power-on reset, but is not initialized by a manual reset or in standby mode. do not access external memory outside area 0 until bcr1 register initialization is complete. bit: 15 14 13 12 11 10 9 8 pula puld hizmem hizcnt endian a0bst1 a0bst0 a5bst1 initial value: 0 0 0 0 0/1* 0 0 0 r/w: r/w r/w r/w r/w r r/w r/w r/w bit: 7 6 5 4 3 2 1 0 a5bst0 a6bst1 a6bst0 dram tp2 dram tp1 dram tp0 a5 pcm a6 pcm initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w note: * samples the value of the external pin (md5) designating the endian in a power-on reset. 280 bit 15pin a25 to a0 pull-up (pula): specifies whether or not pins a25 to a0 are pulled up for 4 cycles immediately after back is asserted. bit 15: pula description 0 not pulled up (initial value) 1 pulled up bit 14pin d31 to d0 pull-up (puld): specifies whether or not pins d31 to d0 are pulled up when not in use. bit 14: puld description 0 not pulled up (initial value) 1 pulled up bit 13hi-z memory control (hizmem): specifies the state of a25Ca0, bs , cs , rd/ wr , we /dqm, rd , ce2a , ce2b and drak0/1 in standby mode. bit 13: hizmem description 0 a25Ca0, bs, cs, rd/wr, we/dqm, rd, ce2a, ce2b and drak0/1 are hi-z in standby mode (initial value) 1 a25Ca0, bs, cs, rd/wr, we/dqm, rd, ce2a, ce2b and drak0/1 are high in standby mode bit 12high-z control (hizcnt): specifies the state of the ras and cas signals in standby mode and when the bus is released. bit 12: hizcnt description 0 ras and cas signals are high-impedance (high-z) in standby mode and when bus is released (initial value) 1 ras and cas signals are driven in standby mode and when bus is released bit 11endian flag (endian): samples the value of the external pin designating the endian in a power-on reset. the endian for all physical spaces is decided by this bit, which is read-only. bit 11: endian description 0 (on reset) endian setting external pin (md5) is low. indicates the sh7729r is set as big-endian 1 (on reset) endian setting external pin (md5) is high. indicates the sh7729r is set as little-endian 281 bits 10 and 9area 0 burst rom control (a0bst1, a0bst0): specify whether to use burst rom in physical space area 0. when burst rom is used, these bits set the number of burst transfers. bit 10: a0bst1 bit 9: a0bst0 description 0 0 access area 0 accessed as ordinary memory (initial value) 1 access area 0 accessed as burst rom (4 consecutive accesses). can be used when bus width is 8, 16, or 32 1 0 access area 0 accessed as burst rom (8 consecutive accesses). can be used when bus width is 8 or 16 1 access area 0 accessed as burst rom (16 consecutive accesses). can be used only when bus width is 8 bits 8 and 7area 5 burst enable (a5bst1, a5bst0): specify whether to use burst rom and pcmcia burst mode in physical space area 5. when burst rom and pcmcia burst mode are used, these bits set the number of burst transfers. bit 8: a5bst1 bit 7: a5bst0 description 0 0 access area 5 accessed as ordinary memory (initial value) 1 burst access of area 5 (4 consecutive accesses). can be used when bus width is 8, 16, or 32 1 0 burst access of area 5 (8 consecutive accesses). can be used when bus width is 8 or 16 1 burst access of area 5 (16 consecutive accesses). can be used only when bus width is 8 bits 6 and 5area 6 burst enable (a6bst1, a6bst0): specify whether to use burst rom and pcmcia burst mode in physical space area 6. when burst rom and pcmcia burst mode are used, these bits set the number of burst transfers. bit 6: a6bst1 bit 5: a6bst0 description 0 0 access area 6 accessed as ordinary memory (initial value) 1 burst access of area 6 (4 consecutive accesses). can be used when bus width is 8, 16, or 32 1 0 burst access of area 6 (8 consecutive accesses). can be used when bus width is 8 or 16 1 burst access of area 6 (16 consecutive accesses). can be used only when bus width is 8 282 bits 4 to 2area 2, area 3 memory type (dramtp2, dramtp1, dramtp0): designate the types of memory connected to physical space areas 2 and 3. ordinary memory, such as rom, sram, or flash rom, can be directly connected. synchronous dram can also be directly connected. bit 4: dramtp2 bit 3: dramtp1 bit 2: dramtp0 description 0 0 0 areas 2 and 3 are ordinary memory (initial value) 1 reserved (setting prohibited) 1 0 area 2: ordinary memory; area 3: synchronous dram* 2 1 areas 2 and 3 are synchronous dram* 1, * 2 1 0 0 reserved (setting prohibited) 1 reserved (setting prohibited) 1 0 reserved (setting prohibited) 1 reserved (setting prohibited) notes: *1 when selecting this mode, set the same bus width for area 2 and area 3. *2 do not access synchronous dram when clock ratio i :b = 1:1. bit 1area 5 bus type (a5pcm): designates whether to access physical space area 5 as pcmcia space. bit 1: a5pcm description 0 physical space area 5 accessed as ordinary memory (initial value) 1 physical space area 5 accessed as pcmcia space bit 0area 6 bus type (a6pcm): designates whether to access physical space area 6 as pcmcia space. bit 0: a6pcm description 0 physical space area 6 accessed as ordinary memory (initial value) 1 physical space area 6 accessed as pcmcia space 283 11.2.2 bus control register 2 (bcr2) bus control register 2 (bcr2) is a 16-bit readable/writable register that selects the bus size of each area. it is initialized to h'3ff0 by a power-on reset, but is not initialized by a manual reset or in standby mode. do not access external memory outside area 0 until bcr2 register initialization is complete. bit: 15 14 13 12 11 10 9 8 a6sz1 a6sz0 a5sz1 a5sz0 a4sz1 a4sz0 initial value: 0 0 1 1 1 1 1 1 r/w: r r r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 3 2 1 0 a3sz1 a3sz0 a2sz1 a2sz0 initial value: 1 1 1 1 0 0 0 0 r/w: r/w r/w r/w r/w r r r r bits 15, 14, 3, 2, 1, and 0reserved: these bits are always read as 0. the write value should always be 0. bits 2n + 1, 2narea n (2C6) bus size specification (ansz1, ansz0): specify the bus size of physical space area n (n = 2 to 6). bit 2n + 1: ansz1 bit 2n: ansz0 port a / b description 0 0 not used reserved (setting prohibited) 1 byte (8-bit) size 1 0 word (16-bit) size 1 longword (32-bit) size 0 0 used reserved (setting prohibited) 1 byte (8-bit) size 1 0 word (16-bit) size 1 reserved (setting prohibited) 284 11.2.3 wait state control register 1 (wcr1) wait state control register 1 (wcr1) is a 16-bit readable/writable register that specifies the number of idle (wait) state cycles inserted for each area. for some memories, data bus drive may not be turned off quickly even when the read signal from the external device is turned off. this can result in conflicts between data buses when consecutive memory accesses are to different memories or when a write immediately follows a memory read. the sh7729r automatically inserts the number of idle states set in wcr1 in those cases. wcr1 is initialized to h'3ff3 by a power-on reset. it is not initialized by a manual reset or in standby mode. bit: 15 14 13 12 11 10 9 8 waitsel a6iw1 a6iw0 a5iw1 a5iw0 a4iw1 a4iw0 initial value: 0 0 1 1 1 1 1 1 r/w: r/w r r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 3 2 1 0 a3iw1 a3iw0 a2iw1 a2iw0 a0iw1 a0iw0 initial value: 1 1 1 1 0 0 1 1 r/w: r/w r/w r/w r/w r r r/w r/w bit 15wait sampling timing select (waitsel): specifies the wait signal sampling timing. bit 15: waitsel description 0 setting to 1 when using the wait signal * (initial value) 1 sampled wait signal at fall of ckio note: * operation is not guaranteed if wait is asserted while waitsel = 0. bits 14, 3, and 2 reserved: these bits are always read as 0. the write value should always be 0. 285 bits 2n + 1, 2narea n (6C2, 0) intercycle idle specification (aniw1, aniw0): specify the number of idles inserted between bus cycles when switching between physical space area n (6C2, 0) and another space or between a read access and a write access in the same physical space. bit 2n + 1: aniw1 bit 2n: aniw0 description 0 0 1 idle cycle inserted 1 1 idle cycle inserted 1 0 2 idle cycles inserted 1 3 idle cycles inserted (initial value) 11.2.4 wait state control register 2 (wcr2) wait state control register 2 (wcr2) is a 16-bit readable/writable register that specifies the number of wait state cycles inserted for each area. it also specifies the data access pitch for burst memory accesses. this allows direct connection of even low-speed memories without an external circuit. wcr2 is initialized to h'ffff by a power-on reset. it is not initialized by a manual reset or in standby mode. bit: 15 14 13 12 11 10 9 8 a6 w2 a6 w1 a6 w0 a5 w2 a5 w1 a5 w0 a4 w2 a4 w1 initial value: 1 1 1 1 1 1 1 1 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 3 2 1 0 a4 w0 a3 w1 a3 w0 a2 w1 a2 w0 a0 w2 a0 w1 a0 w0 initial value: 1 1 1 1 1 1 1 1 r/w: r/w r/w r/w r/w r/w r/w r/w r/w 286 bits 15 to 13area 6 wait control (a6w2, a6w1, a6w0): specify the number of wait states inserted in physical space area 6. also specify the burst pitch for burst transfer. description first cycle burst cycle (excluding first cycle) bit 15: a6w2 bit 14: a6w1 bit 13: a6w0 inserted wait states wait pin number of states per data transfer wait pin 0 0 0 0 disabled 2 enabled 1 1 enabled 2 enabled 1 0 2 enabled 3 enabled 1 3 enabled 4 enabled 1 0 0 4 enabled 4 enabled 1 6 enabled 6 enabled 1 0 8 enabled 8 enabled 110 (initial value) enabled 10 enabled bits 12 to 10area 5 wait control (a5w2, a5w1, a5w0): specify the number of wait states inserted in physical space area 5. also specify the burst pitch for burst transfer. description first cycle burst cycle (excluding first cycle) bit 12: a5w2 bit 11: a5w1 bit 10: a5w0 inserted wait states wait pin number of states per data transfer wait pin 0 0 0 0 disabled 2 enabled 1 1 enabled 2 enabled 1 0 2 enabled 3 enabled 1 3 enabled 4 enabled 1 0 0 4 enabled 4 enabled 1 6 enabled 6 enabled 1 0 8 enabled 8 enabled 110 (initial value) enabled 10 enabled 287 bits 9 to 7area 4 wait control (a4w2, a4w1, a4w0): specify the number of wait states inserted in physical space area 4. description bit 9: a4w2 bit 8: a4w1 bit 7: a4w0 inserted wait state wait pin 0000 ign ored 1 1 enabled 1 0 2 enabled 1 3 enabled 1004 en abled 1 6 enabled 1 0 8 enabled 1 10 enabled (initial value) bits 6 and 5area 3 wait control (a3w1, a3w0): specify the number of wait states inserted in physical space area 3. ? for ordinary memory description bit 6: a3w1 bit 5: a3w0 inserted wait states wait pin 0 0 0 ignored 1 1 enabled 1 0 2 enabled 1 3 enabled (initial value) ? for synchronous dram description bit 6: a3w1 bit 5: a3w0 synchronous dram: cas latency 00 1 11 10 2 1 3 (initial value) 288 bits 4 and 3area 2 wait control (a2w1, a2w0): specify the number of wait states inserted in physical space area 2. ? for ordinary memory description bit 4: a2w0 bit 3: a2w0 inserted wait states wait pin 0 0 0 ignored 1 1 enabled 1 0 2 enabled 1 3 enabled (initial value) ? for synchronous dram description bit 4: a2w1 bit 3: a2w0 synchronous dram: cas latency 001 11 102 1 3 (initial value) bits 2 to 0area 0 wait control (a0w2, a0w1, a0w0): specify the number of wait states inserted in physical space area 0. also specify the burst pitch for burst transfer. description first cycle burst cycle (excluding first cycle) bit 2: a0w2 bit 1: a0w1 bit 0: a0w0 inserted wait states wait pin number of states per data transfer wait pin 0 0 0 0 ignored 2 enabled 1 1 enabled 2 enabled 1 0 2 enabled 3 enabled 1 3 enabled 4 enabled 1 0 0 4 enabled 4 enabled 1 6 enabled 6 enabled 1 0 8 enabled 8 enabled 110 (initial value) enabled 10 enabled 289 11.2.5 individual memory control register (mcr) the individual memory control register (mcr) is a 16-bit readable/writable register that specifies ras and cas timing and burst control for synchronous dram (areas 2 and 3), specifies address multiplexing, and controls refresh. this enables direct connection of synchronous dram without external circuits. mcr is initialized to h'0000 by a power-on reset, but is not initialized by a manual reset or in standby mode. bits tpc1Ctpc0, rcd1Crcd0, trwl1Ctrwl0, tras1Ctras0, rasd, be, amx2Camx0, and edomode are written to in the initialization after a power-on reset and should not then be modified again. when rfsh and rmode are written to, write the same values to the other bits. when using synchronous dram, do not access areas 2 and 3 until this register is initialized. bit: 15 14 13 12 11 10 9 8 tpc1 tpc0 rcd1 rcd0 trwl1 trwl0 tras1 tras0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 3 2 1 0 rasd amx3 amx2 amx1 amx0 rfsh rmode initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r bits 15 and 14ras precharge time (tpc1, tpc0): when synchronous dram interface is selected as connected memory, they set the minimum number of cycles until output of the next bank-active command after precharge. bit 15: tpc1 bit 14: tpc0 description 0 0 1 cycle (initial value) 1 2 cycles 1 0 3 cycles 1 4 cycles 290 bits 13 and 12rasCcas delay (rcd1, rcd0): when synchronous dram interface is selected as connected memory, these bits set the bank active read/write command delay time. bit 13: rcd1 bit 12: rcd0 description 0 0 1 cycle (initial value) 1 2 cycles 1 0 3 cycles 1 4 cycles bits 11 and 10write-precharge delay (trwl1, trwl0): set the synchronous dram write-precharge delay time. this designates the time between the end of a write cycle and the next bank-active command. this setting is valid only when synchronous dram is connected. after the write cycle, the next bank-active command is not issued for the period tpc + trwl. bit 11: trwl1 bit 10: trwl0 description 0 0 1 cycle (initial value) 1 2 cycles 1 0 3 cycles 1 reserved (setting prohibited) bits 9 and 8 cas -before- ras refresh ras assert time (tras1, tras0): when synchronous dram interface is selected as a connected memory, no bank-active command is issued during the period tpc + tras after an auto-refresh command. bit 9: tras1 bit 8: tras0 description 0 0 2 cycles (initial value) 1 3 cycles 1 0 4 cycles 1 5 cycles bit 7synchronous dram bank active (rasd): specifies whether synchronous dram is used in bank active mode or auto-precharge mode. set auto-precharge mode when areas 2 and 3 are both designated as synchronous dram space, or when 16 bits is set as the bus width. bit 7: rasd description 0 auto-precharge mode (initial value) 1 bank active mode 291 bits 6 to 3address multiplex (amx3, amx2, amx1, amx0): specify address multiplexing for synchronous dram. for synchronous dram interface: bit6: amx3 bit5: amx2 bit 4: amx1 bit 3: amx0 description 1101 the row address begins with a10 . (the a10 value is output at a1 when the row address is output. 4m 16-bit 4-bank products) 1 0 the row address begins with a11 . (the a11 value is output at a1 when the row address is output. 8m 16-bit 4-bank products) * 0100 the row address begins with a9 . (the a9 value is output at a1 when the row address is output. 1m 16-bit 4-bank products) (initial value) 1 the row address begins with a10 . (the a10 value is output at a1 when the row address is output. 2m 8-bit 4-bank products) 1 1 the row address begins with a9 . (the a9 value is output at a1 when the row address is output. 512k 32-bit 4-bank products) 0000 begin synchronous dram access after setting amx3 to 0 = *1**. values except above reserved (setting prohibited) note: * can only be set when using a 16-bit bus width. bit 2refresh control (rfsh): the rfsh bit determines whether or not synchronous dram refresh operations are is performed. if the refresh function is not used, the timer for generation of periodic refresh requests can also be used as an interval timer. bit 2: rfsh description 0 no refresh (initial value) 1 refresh 292 bit 1refresh mode (rmode): selects whether to perform an ordinary refresh or a self- refresh when the rfsh bit is 1. when the rfsh bit is 1 and this bit is 0, an auto-refresh is performed on synchronous dram at the period set by refresh-related registers rtcnt, rtcor, and rtcsr. when a refresh request occurs during an external bus cycle, the refresh cycle is performed after the bus cycle ends. when the rfsh bit is 1 and this bit is also 1, the synchronous dram will wait for the end of any executing external bus cycle before going into a self-refresh. all refresh requests to memory that is in the self-refresh state are ignored. bit 1: rmode description 0 auto refresh (rfsh must be 1) (initial value) 1 self-refresh (rfsh must be 1) bit 0reserved: this bit is always read as 0. the write value should always be 0. 11.2.6 pcmcia control register (pcr) the pcmcia control register (pcr) is a 16-bit readable/writable register that specifies the assertion and negation timing of the oe and we signals for the pcmcia interface connected to areas 5 and 6. the oe and we signal assertion width is set by the wait control bits in the wcr2 register. pcr is initialized to h'0000 by a power-on reset, but is not initialized, and retains its contents, in a manual reset and in standby mode. bit: 15 14 13 12 11 10 9 8 a6w3 a5w3 a5ted2 a6ted2 a5teh2 a6teh2 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 3 2 1 0 a5ted1 a5ted0 a6ted1 a6ted0 a5teh1 a5teh0 a6teh1 a6teh0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w 293 bit 15area 6 wait control (a6w3): specifies the number of inserted wait states for area 6 combined with bits a6w2Ca6w0 in wcr2. also specifies the number of transfer states in burst transfer. clear this bit to 0 when area 6 is not set to pcmcia. first cycle burst cycle a6w3 a6w2 a6w1 a6w0 inserted wait states wait pin number of states per one-data transfer wait pin 00000 ign ored 2 enabled 00011 en abled 2 enabled 00102 en abled 3 enabled 00113 en abled 4 enabled 01004 en abled 5 enabled 01016 en abled 7 enabled 01108 en abled 9 enabled 011110 (initial value) enabled 11 enabled 100012 en abled 13 enabled 100114 en abled 15 enabled 101018 en abled 19 enabled 101122 en abled 23 enabled 110026 en abled 27 enabled 110130 en abled 31 enabled 111034 en abled 35 enabled 111138 en abled 39 enabled bit 14area 5 wait control (a5w3): specifies the number of inserted wait states for area 5 combined with bits a5w2Ca5w0 in wcr2. also specifies the number of transfer states in burst transfer. clear this bit to 0 when area 5 is not set to pcmcia. the relationship between the set value and the number of waits is the same as for a6w3. bits 13 and 12reserved: these bits are always read as 0. the write value should always be 0. 294 bits 11, 7, and 6area 5 address oe / we assert delay (a5ted2, a5ted1, a5ted0): specify the delay time from address output to oe / we assertion for the pcmcia interface connected to area 5. bit 11: a5ted2 bit 7: a5ted1 bit 6: a5ted0 description 0 0 0 0.5-cycle delay (initial value) 1 1.5-cycle delay 1 0 2.5-cycle delay 1 3.5-cycle delay 1 0 0 4.5-cycle delay 1 5.5-cycle delay 1 0 6.5-cycle delay 1 7.5-cycle delay bits 10, 5, and 4area 6 address oe / we assert delay (a6ted2, a6ted1, a6ted0): the a6ted bits specify the delay time from address output to oe / we assertion for the pcmcia interface connected to area 6. bit 10: a6ted2 bit 5: a6ted1 bit 4: a6ted0 description 0 0 0 0.5-cycle delay (initial value) 1 1.5-cycle delay 1 0 2.5-cycle delay 1 3.5-cycle delay 1 0 0 4.5-cycle delay 1 5.5-cycle delay 1 0 6.5-cycle delay 1 7.5-cycle delay 295 bits 9, 3, and 2area 5 oe / we negate address delay (a5teh2, a5teh1, a5teh0): specify the address hold delay time from oe / we negation for the pcmcia interface connected to area 5. bit 9: a5teh2 bit 3: a5teh1 bit 2: a5teh0 description 0 0 0 0.5-cycle delay (initial value) 1 1.5-cycle delay 1 0 2.5-cycle delay 1 3.5-cycle delay 1 0 0 4.5-cycle delay 1 5.5-cycle delay 1 0 6.5-cycle delay 1 7.5-cycle delay bits 8, 1, and 0area 6 oe / we negate address delay (a6teh2, a6teh1, a6teh0): specify the address hold delay time from oe / we negation for the pcmcia interface connected to area 6. bit 8: a6teh2 bit 1: a6teh1 bit 0: a6teh0 description 0 0 0 0.5-cycle delay (initial value) 1 1.5-cycle delay 1 0 2.5-cycle delay 1 3.5-cycle delay 1 0 0 4.5-cycle delay 1 5.5-cycle delay 1 0 6.5-cycle delay 1 7.5-cycle delay 296 11.2.7 synchronous dram mode register (sdmr) the synchronous dram mode register (sdmr) is an 8-bit write-only register that is written to via the synchronous dram address bus. it sets synchronous dram mode for areas 2 and 3. sdmr is undefined after a power-on reset. the register contents are not initialized by a manual reset or in standby mode; values remain unchanged. writes to the synchronous dram mode register use the address bus rather than the data bus. if the value to be set is x and the sdmr address is y, the value x is written in the synchronous dram mode register by writing in address x + y. since, with a 32-bit bus width, a0 of the synchronous dram is connected to a2 of the chip and a1 of the synchronous dram is connected to a3 of the chip, the value actually written to the synchronous dram is the x value shifted two bits right. with a 16-bit bus width, the value written is the x value shifted one bit right. for example, with a 32-bit bus width, when h'0230 is written to the sdmr register of area 2, random data is written to the address h'ffffd000 (address y) + h'08c0 (value x), or h'ffffd8c0. as a result, h'0230 is written to the sdmr register. the range for value x is h'0000 to h'0ffc. when h'0230 is written to the sdmr register of area 3, random data is written to the address h'ffffe000 (address y) + h'08c0 (value x), or h'ffffe8c0. as a result, h'0230 is written to the sdmr register. the range for value x is h'0000 to h'0ffc. bit: 31 12 11 10 9 8 sdmr address initial value: ...................... r/w: ...................... w* w* w w bit: 7 6 5 4 3 2 1 0 initial value: r/w: w w w w w w note: * depending on the type of synchronous dram. 297 11.2.8 refresh timer control/status register (rtcsr) the refresh timer control/status register (rtcsr) is a 16-bit readable/writable register that specifies the refresh cycle, whether to generate an interrupt, and the cycle of that interrupt. it is initialized to h'0000 by a power-on reset, but is not initialized by a manual reset or in standby mode. make the rtcor setting before setting bits cks2 to cks0 in rtcsr. note: the method of writing to rtcsr differs from that for general registers to ensure that rtcsr is not rewritten incorrectly. use a word transfer instruction to set the upper byte as b'10100101 and the lower byte as the write data. for details, see section 11.2.12, cautions on accessing refresh control related registers. bit: 15 14 13 12 11 10 9 8 initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r r bit: 7 6 5 4 3 2 1 0 cmf cmie cks2 cks1 cks0 ovf ovie lmts initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bits 15 to 8reserved: these bits are always read as 0. the write value should always be 0. bit 7compare match flag (cmf): indicates that the values of rtcnt and rtcor match. bit 7: cmf description 0 the values of rtcnt and rtcor do not match (initial value) clearing condition: when a refresh is performed after 0 has been written to cmf and rfsh = 1 and rmode = 0 (to perform a cbr refresh) 1 the values of rtcnt and rtcor match setting condition: rtcnt = rtcor* note: * contents do not change when 1 is written to cmf. bit 6compare match interrupt enable (cmie): enables or disables an interrupt request caused when cmf in rtcsr is set to 1. do not set this bit to 1 when using auto-refresh. bit 6: cmie description 0 interrupt request by cmf is disabled (initial value) 1 interrupt request by cmf is enabled 298 bits 5 to 3clock select bits (cks2 to cks0): select the clock input to rtcnt. the source clock is the external bus clock (bclk). the rtcnt count clock is ckio divided by the specified ratio. rtcor must be set before setting cks2Ccks0. description bit 5: cks2 bit 4: cks1 bit 3: cks0 normal external bus clock 0 0 0 clock input disabled 1 bus clock (ckio)/4 1 0 ckio/16 1 ckio/64 1 0 0 ckio/256 1 ckio/1024 1 0 ckio/2048 1 ckio/4096 bit 2refresh count overflow flag (ovf): indicates when the number of refresh requests indicated in the refresh count register (rfcr) exceeds the limit set in the lmts bit in rtcsr. bit 2: ovf description 0 rfcr has not exceeded the count limit value set in lmts (initial value) clearing condition: when 0 is written to ovf 1 rfcr has exceeded the count limit value set in lmts setting condition: when the rfcr value has exceeded the count limit value set in lmts* note: * contents do not change when 1 is written to ovf. bit 1refresh count overflow interrupt enable (ovie): selects whether to suppress generation of interrupt requests by the ovf bit in rtcsr when ovf is set to 1. bit 1: ovie description 0 interrupt request by ovf is disabled (initial value) 1 interrupt request by ovf is enabled 299 bit 0refresh count overflow limit select (lmts): indicates the count limit value to be compared to the number of refreshes indicated in the refresh count register (rfcr). when the value in rfcr overflows the value specified by lmts, the ovf flag is set. bit 0: lmts description 0 count limit value is 1024 (initial value) 1 count limit value is 512 11.2.9 refresh timer counter (rtcnt) rtcnt is a 16-bit register containing a readable/writable 8-bit counter that counts up on an input clock. the clock select bits (cks2Ccks0) in rtcsr select the input clock. when rtcnt matches rtcor, the cmf bit in rtcsr is set and rtcnt is cleared. rtcnt is initialized to h'00 by a power-on reset, but continues incrementing after a manual reset. it is not initialized in standby mode, but holds its contents. note: the method of writing to rtcnt differs from that for general registers to ensure that rtcnt is not rewritten incorrectly. use a word transfer instruction to set the upper byte as b'10100101 and the lower byte as the write data. for details, see section 11.2.12, cautions on accessing refresh control related registers. bit: 15 14 13 12 11 10 9 8 initial value: 0 0 0 0 0 0 0 0 r/w: bit: 7 6 5 4 3 2 1 0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w 300 11.2.10 refresh time constant register (rtcor) the refresh time constant register (rtcor) is a 16-bit register with a readable/writable lower 8 bits. the values of rtcor and rtcnt (lower 8 bits) are constantly compared. when the values match, the compare match flag (cmf) in rtcsr is set and rtcnt is cleared to 0. when the refresh bit (rfsh) in the individual memory control register (mcr) is set to 1 and the refresh mode is set to auto refresh, a memory refresh cycle occurs when the cmf bit is set. rtcor is initialized to h'00 by a power-on reset. it is not initialized by a manual reset or in standby mode, but holds its contents. make the rtcor setting before setting bits cks2 to cks0 in rtcsr. note: the method of writing to rtcor differs from that for general registers to ensure that rtcor is not rewritten incorrectly. use a word transfer instruction to set the upper byte as b'10100101 and the lower byte as the write data. for details, see section 11.2.12, cautions on accessing refresh control related registers. bit: 15 14 13 12 11 10 9 8 initial value: 0 0 0 0 0 0 0 0 r/w: bit: 7 6 5 4 3 2 1 0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w 11.2.11 refresh count register (rfcr) the refresh count register (rfcr) is a 16-bit register containing a readable/writable 10-bit counter that increments every time rtcor and rtcnt match. when rfcr exceeds the count limit value set in the lmts bit in rtcsr, the ovf bit in rtcsr is set and rfcr is cleared. rfcr is initialized to h'0000 by a power-on reset. it is not initialized by a manual reset or in standby mode, but holds its contents. note: the method of writing to rfcr differs from that for general registers to ensure that rfcr is not rewritten incorrectly. use a word transfer instruction to set the six bits starting from the msb in the upper byte as b'101001, and the remaining bits as the write data. for details, see section 11.2.12, cautions on accessing refresh control related registers. 301 bit: 15 14 13 12 11 10 9 8 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w bit: 7 6 5 4 3 2 1 0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w 11.2.12 cautions on accessing refresh control related registers rfcr, rtcsr, rtcnt, and rtcor require that a specific code be appended to the data when it is written to prevent data from being mistakenly overwritten by program overruns or other write operations (figure 11.5). perform reads and writes using the following methods: 1. when writing to rfcr, rtcsr, rtcnt, and rtcor, use only word transfer instructions. byte transfer instructions cannot be used. when writing to rtcnt, rtcsr, or rtcor, place b'10100101 in the upper byte and the write data in the lower byte. when writing to rfcr, place b'101001 in the upper 6 bits and the write data in the remaining bits, as shown in figure 11.5. 2. when reading from rfcr, rtcsr, rtcnt, and rtcor, carry out reads with a 16-bit width. 0 is read from undefined bits. 15 10 8 rtcsr, rtcnt, rtcor 0 rfcr 7 100101 15 10 10 0 9 1 0 0 1 write data write data figure 11.5 writing to rfcr, rtcsr, rtcnt, and rtcor 302 11.2.13 mcs0 control register (mcscr0) the mcs0 control register (mcscr0) is a 16-bit readable/writable register that specifies the mcs[0] pin output conditions. mcscr0 is initialized to h'0000 by a power-on reset, but is not initialized by a manual reset or in standby mode. as the mcs[0] pin is multiplexed as the ptc0 pin, when using the pin as mcs[0] , bits pc0md[1:0] in the pccr register should be set to 00 (other function). bit: 15 14 13 12 11 10 9 8 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 3 2 1 0 cs2/0 cap1 cap0 a25 a24 a23 a22 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bits 15 to 7reserved: these bits are always read as 0. the write value should always be 0. bit 6cs2/cs0 select (cs2/0): selects whether an area 2 or area 0 address is to be decoded. bit 6: cs2/0 description 0 area 0 is selected 1 area 2 is selected bits 5 and 4connected memory size specification (cap1, cap0) bit 5: cap1 bit 4: cap0 description 0 0 32-mbit memory is connected 0 1 64-mbit memory is connected 1 0 128-mbit memory is connected 1 1 256-mbit memory is connected bits 3 to 0start address specification (a25, a24, a23, a22): these bits specify the start address of the memory area for which mcs[0] is asserted. 303 11.2.14 mcs1 control register (mcscr1) the mcs1 control register (mcscr1) specifies the mcs[1] pin output conditions. the bit configuration and functions are the same as those of mcscr0. 11.2.15 mcs2 control register (mcscr2) the mcs2 control register (mcscr2) specifies the mcs[2] pin output conditions. the bit configuration and functions are the same as those of mcscr0. 11.2.16 mcs3 control register (mcscr3) the mcs3 control register (mcscr3) specifies the mcs[3] pin output conditions. the bit configuration and functions are the same as those of mcscr0. 11.2.17 mcs4 control register (mcscr4) the mcs4 control register (mcscr4) specifies the mcs[4] pin output conditions. the bit configuration and functions are the same as those of mcscr0. 11.2.18 mcs5 control register (mcscr5) the mcs5 control register (mcscr5) specifies the mcs[5] pin output conditions. the bit configuration and functions are the same as those of mcscr0. 11.2.19 mcs6 control register (mcscr6) the mcs6 control register (mcscr6) specifies the mcs[6] pin output conditions. the bit configuration and functions are the same as those of mcscr0. 11.2.20 mcs7 control register (mcscr7) the mcs7 control register (mcscr7) specifies the mcs[7] pin output conditions. the bit configuration and functions are the same as those of mcscr0. 304 11.3 bsc operation 11.3.1 endian/access size and data alignment the sh7729r supports both big endian, in which the 0 address is the most significant byte in the byte data, and little endian, in which the 0 address is the least significant byte. switching between the two is designated by an external pin (md5 pin) at the time of a power-on reset. after a power- on reset, big endian is engaged when md5 is low; little endian is engaged when md5 is high. three data bus widths are available for ordinary memory (byte, word, longword) and two data bus widths (word and longword) for synchronous dram. for the pcmcia interface, choose from byte and word. this means data alignment is done by matching the devices data width and endian. the access unit must also be matched to the devices bus width. this also means that when longword data is read from a byte-width device, four read operations must be performed. in the sh7729r, data alignment and conversion of data length is performed automatically between the respective interfaces. tables 11.7 through 11.12 show the relationship between endian, device data width, and access unit. table 11.7 32-bit external device/big-endian access and data alignment data bus strobe signals operation d31Cd24 d23Cd16 d15Cd8 d7Cd0 we3 , dqmuu we2 , dqmul we1 , dqmlu we0 , dqmll byte access at 0 data 7C0 asserted byte access at 1 data 7C0 asserted byte access at 2 data 7C0 asserted byte access at 3 data 7C0 asserted word access at 0 data 15C8 data 7C0 asserted asserted word access at 2 data 15C8 data 7C0 asserted asserted longword access at 0 data 31C24 data 23C16 data 15C8 data 7C0 asserted asserted asserted asserted 305 table 11.8 16-bit external device/big-endian access and data alignment data bus strobe signals operation d31C d24 d23C d16 d15Cd8 d7Cd0 we3 , dqmuu we2 , dqmul we1 , dqmlu we0 , dqmll byte access at 0 data 7C0 asserted byte access at 1 data 7C0 asserted byte access at 2 data 7C0 asserted byte access at 3 data 7C0 asserted word access at 0 data 15C8 data 7C0 asserted asserted word access at 2 data 15C8 data 7C0 asserted asserted longword access 1st time at 0 data 31C24 data 23C16 asserted asserted at 0 2nd time at 2 data 15C8 data 7C0 asserted asserted 306 table 11.9 8-bit external device/big-endian access and data alignment data bus strobe signals operation d31C d24 d23C d16 d15C d8 d7Cd0 we3 , dqmuu we2 , dqmul we1 , dqmlu we0 , dqmll byte access at 0 data 7C0 asserted byte access at 1 data 7C0 asserted byte access at 2 data 7C0 asserted byte access at 3 data 7C0 asserted word access at 0 1st time at 0 data 15C8 asserted 2nd time at 1 data 7C0 asserted word access at 2 1st time at 2 data 15C8 asserted 2nd time at 3 data 7C0 asserted longword access at 0 1st time at 0 data 31C24 asserted 2nd time at 1 data 23C16 asserted 3rd time at 2 data 15C8 asserted 4th time at 3 data 7C0 asserted 307 table 11.10 32-bit external device/little-endian access and data alignment data bus strobe signals operation d31Cd24 d23Cd16 d15Cd8 d7Cd0 we3 , dqmuu we2 , dqmul we1 , dqmlu we0 , dqmll byte access at 0 data 7C0 asserted byte access at 1 data 7C0 asserted byte access at 2 data 7C0 asserted byte access at 3 data 7C0 asserted word access at 0 data 15C8 data 7C0 asserted asserted word access at 2 data 15C8 data 7C0 asserted asserted longword access at 0 data 31C24 data 23C16 data 15C8 data 7C0 asserted asserted asserted asserted table 11.11 16-bit external device/little-endian access and data alignment data bus strobe signals operation d31C d24 d23C d16 d15Cd8 d7Cd0 we3 , dqmuu we2 , dqmul we1 , dqmlu we0 , dqmll byte access at 0 data 7C0 asserted byte access at 1 data 7C0 asserted byte access at 2 data 7C0 asserted byte access at 3 data 7C0 asserted word access at 0 data 15C8 data 7C0 asserted asserted word access at 2 data 15C8 data 7C0 asserted asserted longword access 1st time at 0 data 15C8 data 7C0 asserted asserted at 0 2nd time at 2 data 31C24 data 23C16 asserted asserted 308 table 11.12 8-bit external device/little-endian access and data alignment data bus strobe signals operation d31C d24 d23C d16 d15C d8 d7Cd0 we3 , dqmuu we2 , dqmul we1 , dqmlu we0 , dqmll byte access at 0 data 7C0 asserted byte access at 1 data 7C0 asserted byte access at 2 data 7C0 asserted byte access at 3 data 7C0 asserted word access at 0 1st time at 0 data 7C0 asserted 2nd time at 1 data 15C8 asserted word access at 2 1st time at 2 data 7C0 asserted 2nd time at 3 data 15C8 asserted longword access at 0 1st time at 0 data 7C0 asserted 2nd time at 1 data 15C8 asserted 3rd time at 2 data 23C16 asserted 4th time at 3 data 31C24 asserted 309 11.3.2 description of areas area 0: area 0 physical address bits a28Ca26 are 000. address bits a31Ca29 are ignored and the address range is h'00000000 + h'20000000 n C h'03ffffff + h'20000000 n (n = 0C6 and n = 1C6 are the shadow spaces). ordinary memories such as sram, rom, and burst rom can be connected to this space. byte, word, or longword can be selected as the bus width using external pins md3 and md4. when the area 0 space is accessed, the cs0 signal is asserted. the rd signal that can be used as oe and the we0 C we3 signals for write control are also asserted. the number of bus cycles is selected between 0 and 10 wait cycles using the a0w2Ca0w0 bits in wcr2. when the burst function is used, the bus cycle pitch of the burst cycle is determined within a range of 2C10 according to the number of waits. area 1: area 1 physical address bits a28Ca26 are 001. address bits a31Ca29 are ignored and the address range is h'04000000 + h'20000000 n C h'07ffffff + h'20000000 n (n = 0C6 and n = 1C6 are the shadow spaces). area 1 is the area specifically for internal peripheral modules. external memories cannot be connected. control registers of the peripheral modules shown below are mapped to this area 1. their addresses are physical addresses, to which logical addresses can be mapped when the mmu is enabled: dmac, port, irda, scif, adc, dac, intc (except intevt, ipra, iprb) these registers must be set not to be cached. area 2: area 2 physical address bits a28Ca26 are 010. address bits a31Ca29 are ignored and the address range is h'08000000 + h'20000000 n C h'0bffffff + h'20000000 n (n = 0C6 and n = 1C6 are the shadow spaces). ordinary memories such as sram and rom, as well as synchronous dram, can be connected to this space. byte, word, or longword can be selected as the bus width using bits a2sz1 and a2sz0 in bcr2 for ordinary memory. when the area 2 space is accessed, the cs2 signal is asserted. when ordinary memories are connected, the rd signal that can be used as oe and the we0 C we3 signals for write control are also asserted and the number of bus cycles is selected between 0 and 3 wait cycles using bits a2w1 and a2w0 bits in wcr2. when synchronous dram is connected, the ras3u and ras3l signals, casu and casl signals, rd/ wr signal, and byte control signals dqmhh, dqmhl, dqmlh, and dqmll are all asserted and addresses multiplexed. control of ras3u , ras3l , casu , casl , data timing, and address multiplexing is set with mcr. 310 area 3: area 3 physical address bits a28Ca26 are 011. address bits a31Ca29 are ignored and the address range is h'0c000000 + h'20000000 n C h'0fffffff + h'20000000 n (n = 0C6 and n = 1C6 are the shadow spaces). ordinary memories such as sram and rom, as well as synchronous dram, can be connected to this space. byte, word or longword can be selected as the bus width using bits a3sz1 and a3sz0 bits in bcr2 for ordinary memory. when area 3 space is accessed, cs3 is asserted. when ordinary memories are connected, the rd signal that can be used as oe and the we0 C we3 signals for write control are asserted and the number of bus cycles is selected between 0 and 3 wait cycles using the a3w1 and a3w0 bits in wcr2. when synchronous dram is connected, the ras3u and ras3l signals, casu and casl signals, rd/ wr signal, and byte control signals dqmhh, dqmhl, dqmlh, and dqmll are all asserted and addresses multiplexed. the ras3u and ras3l signals, cashh signal, cashl signal, caslh signal, casll signal, and rd/ wr signal are all asserted and addresses multiplexed. area 4: area 4 physical address bits a28Ca26 are 100. address bits a31Ca29 are ignored and the address range is h'10000000 + h'20000000 n C h'13ffffff + h'20000000 n (n = 0C6 and n = 1C6 are the shadow spaces). only ordinary memories such as sram and rom can be connected to this space. byte, word, or longword can be selected as the bus width using bits a4sz1 and a4sz0 in bcr2. when the area 4 space is accessed, the cs4 signal is asserted. the rd signal that can be used as oe and the we0 C we3 signals for write control are also asserted. the number of bus cycles is selected between 0 and 10 wait cycles using the a4w2Ca4w0 bits in wcr2. area 5: area 5 physical address bits a28Ca26 are 101. address bits a31Ca29 are ignored and the address range is the 64 mbytes at h'14000000 + h'20000000 n C h'17ffffff + h'20000000 n (n = 0C6 and n = 1C6 are the shadow spaces). ordinary memories such as sram and rom as well as burst rom and pcmcia interfaces can be connected to this space. when the pcmcia interface is used, the ic memory card interface address range comprises the 32 mbytes at h'14000000 + h'20000000 n to h'15ffffff + h'20000000 n (n = 0C6 and n = 1C6 are the shadow spaces), and the i/o card interface address range comprises the 32 mbytes at h'16000000 + h'20000000 n to h'17ffffff + h'20000000 n (n = 0C6 and n = 1C6 are the shadow spaces). for ordinary memory and burst rom, byte, word, or longword can be selected as the bus width using bits a5sz1 and a5sz0 in bcr2. for the pcmcia interface, byte or word can be selected as the bus width using bits a5sz1 and a5sz0 bits in bcr2. 311 when the area 5 space is accessed and ordinary memory is connected, the cs5 signal is asserted. the rd signal that can be used as oe and the we0 C we3 signals for write control are also asserted. when the pcmcia interface is used, the ce1a signal, ce2a signal, rd signal as oe signal, and we1 signal are asserted. the number of bus cycles is selected between 0 and 10 wait cycles using the a5w2Ca5w0 bits in wcr2. with the pcmcia interface, from 0 to 38 wait cycles can be selected using the a5w2C a5w0 bits in wcr2 and the a5w3 bit in pcr. in addition, any number of waits can be inserted in each bus cycle by means of the external wait pin ( wait ). when a burst function is used, the bus cycle pitch of the burst cycle is determined within a range of 2C11 (2C39 for the pcmcia interface) according to the number of waits. the setup and hold times of address/ cs5 for the read/write strobe signals can be set in the range 0.5C7.5 using bits a5ted2Ca5ted0 and a5teh2Ca5teh0 in the pcr register. area 6: area 6 physical address bits a28Ca26 are 110. address bits a31Ca29 are ignored and the address range is the 64 mbytes at h'18000000 + h'20000000 n C h'1bffffff + h'20000000 n (n = 0C6 and n = 1C6 are the shadow spaces). ordinary memories such as sram and rom as well as burst rom and pcmcia interfaces can be connected to this space. when the pcmcia interface is used, the ic memory card interface address range is 32 mbytes at h'18000000 + h'20000000 n C h'19ffffff + h'20000000 n and the i/o card interface address range is 32 mbytes at h'1a000000 + h'20000000 n C h'1bffffff + h'20000000 n (n = 0C6 and n = 1C6 are the shadow spaces). for ordinary memory and burst rom, byte, word, or longword can be selected as the bus width using bits a6sz1 and a6sz0 in bcr2. for the pcmcia interface, byte or word can be selected as the bus width using bits a6sz1 and a6sz0 in bcr2. when the area 6 space is accessed and ordinary memory is connected, the cs6 signal is asserted. the rd signal that can be used as oe and the we0 C we3 signals for write control are also asserted. when the pcmcia interface is used, the ce1b signal, ce2b signal, rd signal as oe signal, and we , iciord , and iciowr signals are asserted. the number of bus cycles is selected between 0 and 10 wait cycles using the a6w2Ca6w0 bits in wcr2. with the pcmcia interface, from 0 to 38 wait cycles can be selected using the a6w2C a6w0 bits in wcr2 and the a6w3 bit in pcr. in addition, any number of waits can be inserted in each bus cycle by means of the external wait pin ( wait ). the bus cycle pitch of the burst cycle is determined within a range of 2C11 (2C39 for the pcmcia interface) according to the number of waits. the address/ cs6 setup and hold times for the read/write strobe signals can be set in the range 0.5C7.5 using bits a6ted2Ca6ted0 and a6teh2Ca6teh0 in the pcr register. 312 11.3.3 basic interface basic timing: the basic interface of the sh7729r uses strobe signal output in consideration of the fact that mainly static ram will be directly connected. figure 11.6 shows the basic timing of normal space accesses. a no-wait normal access is completed in two cycles. the bs signal is asserted for one cycle to indicate the start of a bus cycle. the csn signal is negated on the t2 clock falling edge to secure the negation period. therefore, in case of access at minimum pitch, there is a half-cycle negation period. there is no access size specification when reading. the correct access start address is output in the least significant bit of the address, but since there is no access size specification, 32 bits are always read in case of a 32-bit device, and 16 bits in case of a 16-bit device. when writing, only the we signal for the byte to be written is asserted. for details, see section 11.3.1, endian/access size and data alignment. read/write for cache fill or write-back follows the set bus width and transfers a total of 16 bytes continuously. the bus is not released during this transfer. for cache misses that occur during byte or word operand accesses or branching to odd word boundaries, the fill is always performed by longword accesses on the chip-external interface. write-through-area write access and non- cacheable read/write access are based on the actual address size. 313 t1 ckio a25 to a0 csn rd/wr rd d31 to d0 wen d31 to d0 bs t2 read write figure 11.6 basic timing of basic interface 314 figures 11.7, 11.8, and 11.9 show examples of connection to 32, 16, and 8-bit data-width static ram, respectively. ???? ???? ???? ???? ???? a16 a0 cs oe i/o7 i/o0 we ???? ???? ???? ???? a18 a2 csn rd d31 d24 we3 d23 d16 we2 d15 d8 we1 d7 d0 we0 sh7729r 128k 8-bit sram ???? a16 a0 cs oe i/o7 i/o0 we ???? ???? ???? ???? ???? a16 a0 cs oe i/o7 i/o0 we ???? ???? ???? ???? a16 a0 cs oe i/o7 i/o0 we ???? ???? ???? ???? ???? figure 11.7 example of 32-bit data-width static ram connection 315 a16 a0 cs oe i/o7 i/o0 we ???? ???? ???? ???? a17 a1 csn rd d15 d8 we1 d7 d0 we0 sh7729r 128k 8-bit sram ???? a16 a0 cs oe i/o7 i/o0 we ???? ???? ???? ???? ???? ???? ???? ???? figure 11.8 example of 16-bit data-width static ram connection 316 a16 a0 csn rd d7 d0 we0 sh7729r 128k 8-bit sram ???? a16 a0 cs oe i/o7 i/o0 we ???? ???? ???? ???? ???? ???? ???? figure 11.9 example of 8-bit data-width static ram connection 317 wait state control: wait state insertion on the basic interface can be controlled by the wcr2 settings. if the wcr2 wait specification bits corresponding to a particular area are not zero, a software wait is inserted in accordance with that specification. for details, see section 11.2.5, wait control register 2 (wcr2). the specified number of tw cycles are inserted as wait cycles using the basic interface wait timing shown in figure 11.10. t1 ckio a25 to a0 csn rd/wr rd d31 to d0 wen d31 to d0 bs tw t2 read write figure 11.10 basic interface wait timing (software wait only) when software wait insertion is specified by wcr2, the external wait input wait signal is also sampled. wait pin sampling is shown in figure 11.11. a 2-cycle wait is specified as a software wait. sampling is performed at the transition from the tw state to the t2 state; therefore, if the wait signal has no effect if asserted in the t1 cycle or the first tw cycle. 318 when the waitsel bit in the wcr1 register is set to 1, the wait signal is sampled at the falling edge of the clock. if the setup time and hold times with respect to the falling edge of the clock are not satisfied, the value sampled at the next falling edge is used. however, the wait signal is ignored in the following three cases: ? a write to external address space in dual address mode with 16-byte dma transfer ? transfer from an external device with dack to external address space in single address mode with 16-byte dma transfer ? cache write-back access t1 ckio a25 to a0 csn rd/wr rd d31 to d0 wen d31 to d0 wait tw tw tw t2 read write bs wait states inserted by wait signal figure 11.11 basic interface wait state timing (wait state insertion by wait signal waitsel = 1) 319 11.3.4 synchronous dram interface synchronous dram direct connection: since synchronous dram can be selected by the cs signal, physical space areas 2 and 3 can be connected using ras and other control signals in common. if the memory type bits (dramtp2C0) in bcr1 are set to 010, area 2 is ordinary memory space and area 3 is synchronous dram space; if set to 011, areas 2 and 3 are both synchronous dram space. note, however, that synchronous dram must not be accessed when clock ratio i :b = 1:1. with the sh7729r, burst length 1 burst read/single write mode is supported as the synchronous dram operating mode. a data bus width of 16 or 32 bits can be selected. the burst enable bit (be) in mcr is ignored, a 16-bit burst transfer is performed in a cache fill/write-back cycle, and only one access is performed in a write-through area write or a non-cacheable area read/write. the control signals for direct connection of synchronous dram are ras3l , ras3u , casl , casu , rd/ wr , cs2 or cs3 , dqmuu, dqmul, dqmlu, dqmll, and cke. all the signals other than cs2 and cs3 are common to all areas, and signals other than cke are valid and fetched to the synchronous dram only when cs2 or cs3 is asserted. synchronous dram can therefore be connected in parallel to a number of areas. cke is negated (low) only when self-refreshing is performed, and is always asserted (high) at other times. in the refresh cycle and mode-register write cycle, ras3u and ras3l or casu and casl are output. commands for synchronous dram are specified by ras3l , ras3u , casl , casu , rd/ wr , and special address signals. the commands are nop, auto-refresh (ref), self-refresh (self), precharge all banks (pall), row address strobe bank active (actv), read (read), read with precharge (reada), write (writ), write with precharge (writa), and mode register write (mrs). byte specification is performed by dqmuu, dqmul, dqmlu, and dqmll. a read/write is performed for the byte for which the corresponding dqm is low. in big-endian mode, dqmuu specifies an access to address 4n, and dqmll specifies an access to address 4n + 3. in little- endian mode, dqmuu specifies an access to address 4n + 3, and dqmll specifies an access to address 4n. figures 11.12 and 11.13 show examples of the connection of two 1m 16-bit 4-bank synchronous drams and one 1m 16-bit 4-bank synchronous dram, respectively. 320 a15 a2 cki0 cke csn ras3x casx rd/ wr d31 d16 dqmuu dqmul d15 d0 dqmlu dqmll sh7729r 1 m 16-bit 4-bank synchronous dram ???? a13 a0 clk cke cs ras cas we i/o15 i/o0 dqmu dqml ???? ???? ???? a13 a0 clk cke cs ras cas we i/o15 i/o0 dqmu dqml ???? ???? ???? ???? ???? ???? ???? ???? ???? ???? note : "x" is u or l figure 11.12 example of 64-mbit synchronous dram connection (32-bit bus width) 321 sh7729r 64m synchronous dram (1m 16 bit 4 bank) a14 a13 a12 a1 ckio cke csn ras3x casx rd/ wr d15 d0 dqmlu dqmll a13 a12 a11 a0 clk cke cs ras cas we dq15 dq0 dqmu dqml ??? ??? ??? ??? ??? ??? ??? ??? figure 11.13 example of 64-mbit synchronous dram (16-bit bus width) address multiplexing: synchronous dram can be connected without external multiplexing circuitry in accordance with the address multiplex specification bits amx2-amx0 in mcr. table 11.14 shows the relationship between the address multiplex specification bits and the bits output at the address pins. a25Ca16 and a0 are not multiplexed; the original values are always output at these pins. when a0, the lsb of the synchronous dram address, is connected to the sh7729r, it performs longword address specification. connection should therefore be made in the following order: with a 32-bit bus width, connect pin a0 of the synchronous dram to pin a2 of the sh7729r, then connect pin a1 to pin a3; with a 16-bit bus width, connect pin a0 of the synchronous dram to pin a1 of the sh7729r, then connect pin a1 to pin a2. 322 table 11.13 relationship between bus width, amx bits, and address multiplex output setting external address pins bus width memory type amx3 amx2 amx1 amx0 output timing a1 to a8 a9 a10 a11 a12 a13 a14 a15 a16 32 bits 4m 16 bits 1 1 0 1 column address a1 to a8 a9 a10 a11 l/h* 3 a13 a23 a24* 4 a25* 4 4 banks* 1 row address a10 to a17 a18 a19 a20 a21 a22 a23 a24* 4 a25* 4 2m 16 bits 0 1 0 1 column address a1 to a8 a9 a10 a11 l/h* 3 a13 a23* 4 a24* 4 4 banks* 2 row address a10 to a17 a18 a19 a20 a21 a22 a23* 4 a24* 4 1m 16 bits 0 1 0 0 column address a1 to a8 a9 a10 a11 l/h* 3 a13 a22* 4 a23* 4 4 banks* 2 row address a9 to a16 a17 a18 a19 a20 a21 a22* 4 a23* 4 2m 8 bits 0 1 0 1 column address a1 to a8 a9 a10 a11 l/h* 3 a13 a23* 4 a24* 4 4 banks* 2 row address a10 to a17 a18 a19 a20 a21 a22 a23* 4 a24* 4 512k 32 bits 0 1 1 1 column address a1 to a8 a9 a10 a11 l/h* 3 a21* 4 a22* 4 a15 4 banks* 2 row address a9 to a16 a17 a18 a19 a20 a21* 4 a22* 4 a23 16 bits 8m 16 bits 1 1 1 0 column address a1 to a8 a9 a10 l/h* 3 a12 a23 a24* 4 a25* 4 4 banks* 1 row address a11 to a18 a19 a20 a21 a22 a23 a24* 4 a25* 4 4m 16 bits 1 1 0 1 column address a1 to a8 a9 a10 l/h* 3 a12 a22 a23* 4 a24* 4 4 banks* 2 row address a10 to a17 a18 a19 a20 a21 a22 a23* 4 a24* 4 2m 16 bits 0 1 0 1 column address a1 to a8 a9 a10 l/h* 3 a12 a22* 4 a23* 4 a24 4 banks* 2 row address a10 to a17 a18 a19 a20 a21 a22* 4 a23* 4 a24 1m 16 bits 0 1 0 0 column address a1 to a8 a9 a10 l/h* 3 a12 a21* 4 a22* 4 a15 4 banks* 2 row address a9 to a16 a17 a18 a19 a20 a21* 4 a22* 4 a23 2m 8 bits 0 1 0 1 column address a1 to a8 a9 a10 l/h* 3 a12 a22* 4 a23* 4 a24 4 banks* 2 row address a10 to a17 a18 a19 a20 a21 a22* 4 a23* 4 a24 323 table 11.14 example of correspondence between sh7729r and synchronous dram address pins (amx2C0 = 011 (32-bit bus width)) sh7729r address pin synchronous dram address pin ras cycle cas cycle function a11 a19 a19 a9 bank select bank address a10 a18 l/h a8 address precharge setting a9 a17 a9 a7 address a8 a16 a8 a6 a7 a15 a7 a5 a6 a14 a6 a4 a5 a13 a5 a3 a4 a12 a4 a2 a3 a11 a3 a1 a2 a10 a2 a0 a1 a9 a1 not used a0 a0 a0 not used burst read: in the example in figure 11.14 it is assumed that four 2m 8-bit synchronous drams are connected and a 32-bit data width is used, and the burst length is 1. following the tr cycle in which actv command output is performed, a read command is issued in the tc1, tc2, and tc3 cycles, and a reada command in the tc4 cycle, and the read data is accepted at the rising edge of the external command clock (ckio) from cycle td1 to cycle td4. the tpc cycle is used to wait for completion of auto-precharge based on the reada command inside the synchronous dram; no new access command can be issued to the same bank during this cycle, but access to synchronous dram for another area is possible. in the sh7729r, the number of tpc cycles is determined by the tpc bit specification in mcr, and commands cannot be issued for the same synchronous dram during this interval. the example in figure 11.14 shows the basic timing. to connect low-speed synchronous dram, the cycle can be extended by setting wcr2 and mcr bits. the number of cycles from the actv command output cycle, tr, to the read command output cycle, tc1, can be specified by the rcd bits in mcr, with values of 0 to 3 specifying 1 to 4 cycles, respectively. in case of 2 or more cycles, a trw cycle, in which an nop command is issued for the synchronous dram, is inserted between the tr cycle and the tc cycle. the number of cycles from read and reada command output cycles tc1-tc4 to the first read data latch cycle, td1, can be specified as 1 to 3 cycles independently for areas 2 and 3 by means of bits a2w1 and a2w0 or a3w1 and a3w0 in wcr2. this number of cycles corresponds to the number of synchronous dram cas latency cycles. 324 ckio a25 to a16, a13 a12 a15, a14, a11 to a0 cs2 or cs3 ras3x casx rd/ wr dqmxx d31 to d0 bs tr tc1 tc2/td1 tc3/td2 tc4/td3 td4 tpc figure 11.14 basic timing for synchronous dram burst read 325 figure 11.15 shows the burst read timing when rcd is set to 1, a3w1 and a3w0 are set to 10, and tpc is set to 1. the bs cycle, which is asserted for one cycle at the start of a bus cycle for normal access space, is asserted in each of cycles td1Ctd4 in a synchronous dram cycle. when a burst read is performed, the address is updated each time cas is asserted. as the unit of burst transfer is 16 bytes, address updating is performed for a3 and a2 only (a3, a2, and a1 in the case of a 16-bit bus width). the order of access is as follows: in a fill operation in the event of a cache miss, the missed data is read first, then 16-byte boundary data including the missed data is read in wraparound mode. ckio a25 to a16, a13 a12 a15, a14, a11 to a0 cs2 or cs3 ras3x casx rd/wr dqmxx d31 to d0 bs tr tc1 tc2 tc3/td1 tc4/td2 td3 tpc trw td4 figure 11.15 synchronous dram burst read wait specification timing 326 single read: figure 11.16 shows the timing when a single address read is performed. as the burst length is set to 1 in synchronous dram burst read/single write mode, only the required data is output. consequently, no unnecessary bus cycles are generated even when a cache-through area is accessed. ckio a25 to a16, a13 a12 a15, a14, a11 to a0 cs2 or cs3 ras3x casx rd/ wr dqmxx d31 to d0 bs tr tc1 td1 tpc figure 11.16 basic timing for synchronous dram single read 327 burst write: the timing chart for a burst write is shown in figure 11.17. in the sh7729r, a burst write occurs in the event of cache write-back or 16-byte dmac transfer. in a burst write operation, following the tr cycle in which actv command output is performed, a writ command is issued in the tc1, tc2, and tc3 cycles, and a writa command that performs auto- precharge is issued in the tc4 cycle. in the write cycle, the write data is output at the same time as the write command. in case of the write with auto-precharge command, precharging of the relevant bank is performed in the synchronous dram after completion of the write command, and therefore no command can be issued for the same bank until precharging is completed. consequently, in addition to the precharge wait cycle, tpc, used in a read access, cycle trwl is also added as a wait interval until precharging is started following the write command. issuance of a new command for the same bank is deferred during this interval. the number of trwl cycles can be specified by the trwl bits in mcr. 328 ckio csn rd/ wr ras3x casx dqmxx d31 to d0 (read) bs tr tc1 tc2 tc3 tc4 (tpc) (tpc) address upper bits a12, a11, a10 or a9 address lower bits figure 11.17 basic timing for synchronous dram burst write 329 single write: the basic timing chart for write access is shown in figure 11.18. in a single write operation, following the tr cycle in which actv command output is performed, a writa command that performs auto-precharge is issued in the tc1 cycle. in the write cycle, the write data is output at the same time as the write command. in case of the write with auto-precharge command, precharging of the relevant bank is performed in the synchronous dram after completion of the write command, and therefore no command can be issued for the same bank until precharging is completed. consequently, in addition to the precharge wait cycle, tpc, used in a read access, cycle trwl is also added as a wait interval until precharging is started following the write command. issuance of a new command for the same bank is deferred during this interval. the number of trwl cycles can be specified by the trwl bits in mcr. 330 ckio csn rd/ wr ras3x casx dqmxx d31 to d0 bs address upper bits a12 or a10 address lower bits cke tr tc1 (trwl) (tpc) figure 11.18 basic timing for synchronous dram single write 331 bank active: the synchronous dram bank function is used to support high-speed accesses to the same row address. when the rasd bit in mcr is 1, read/write command accesses are performed using commands without auto-precharge (read, writ). in this case, precharging is not performed when the access ends. when accessing the same row address in the same bank, it is possible to issue the read or writ command immediately, without issuing an actv command, in the same way as in the ras down state in dram fast page mode. as synchronous dram is internally divided into two or four banks, it is possible to activate one row address in each bank. if the next access is to a different row address, a pre command is first issued to precharge the relevant bank, then when precharging is completed, the access is performed by issuing an actv command followed by a read or writ command. if this is followed by an access to a different row address, the access time will be longer because of the precharging performed after the access request is issued. in a write, when auto-precharge is performed, a command cannot be issued for a period of trwl + tpc cycles after issuance of the writa command. when bank active mode is used, read or writ commands can be issued successively if the row address is the same. the number of cycles can thus be reduced by trwl + tpc cycles for each write. the number of cycles between issuance of the precharge command and the row address strobe command is determined by the tpc bits in mcr. whether faster execution speed is achieved by use of bank active mode or by use of basic access is determined by the probability of accessing the same row address (p1), and the average number of cycles from completion of one access to the next access (ta). if ta is greater than tpc, the delay due to the precharge wait when writing is imperceptible. in this case, the access speed for bank active mode and basic access is determined by the number of cycles from the start of access to issuance of the read/write command: (tpc + trcd) (1 C p1) and trcd, respectively. there is a limit on tras, the time for placing each bank in the active state. if there is no guarantee that there will not be a cache hit and another row address will be accessed within the period in which this value is maintained by program execution, it is necessary to set auto-refresh and set the refresh cycle to no more than the maximum value of tras. in this way, it is possible to observe the restrictions on the maximum active state time for each bank. if auto-refresh is not used, measures must be taken in the program to ensure that the banks do not remain active for longer than the prescribed time. a burst read cycle without auto-precharge is shown in figure 11.19, a burst read cycle for the same row address in figure 11.20, and a burst read cycle for different row addresses in figure 11.21. similarly, a burst write cycle without auto-precharge is shown in figure 11.22, a burst write cycle for the same row address in figure 11.23, and a burst write cycle for different row addresses in figure 11.24. 332 a tnop cycle, in which no operation is performed, is inserted before the tc cycle in which the read command is issued in figure 11.20, but when synchronous dram is read, there is a two- cycle latency for the dqmxx signal that performs the byte specification. if the tc cycle were performed immediately, without inserting a tnop cycle, it would not be possible to perform the dqmxx signal specification for td1 cycle data output. this is the reason for inserting the tnop cycle. if the cas latency is two cycles or longer, tnop cycle insertion is not performed, since the timing requirements will be met even if the dqmxx signal is set after the tc cycle. when bank active mode is set, if only accesses to the respective banks in the area 3 space are considered, as long as accesses to the same row address continue, the operation starts with the cycle in figure 11.19 or 11.22, followed by repetition of the cycle in figure 11.20 or 11.23. an access to a different area 3 space during this time has no effect. if there is an access to a different row address in the bank active state, after this is detected the bus cycle in figure 11.21 or 11.24 is executed instead of that in figure 11.20 or 11.23. in bank active mode, too, all banks become inactive after a refresh cycle or after the bus is released as the result of bus arbitration. 333 ckio a25Ca16, a13 (a25C a16, a11) a12 (a10) a15, a14, a11Ca0 (a15Ca12, a9Ca0) cs2 or cs3 ras3x casx rd/ wr dqmxx d31Cd0 bs tr tc1 tc2/td1 tc3/td2 tc4/td3 td4 figure 11.19 burst read timing (no precharge) 334 ckio a25Ca16, a13 (a25C a16, a11) a12 (a10) a15, a14, a11Ca0 (a15Ca12, a9Ca0) cs2 or cs3 ras3x casx rd/ wr dqmxx d31Cd0 bs tnop tc1 tc2/td1 tc3/td2 tc4/td3 td4 figure 11.20 burst read timing (same row address) 335 ckio a25Ca16, a13 (a25C a16, a11) a12 (a10) a15, a14, a11Ca0 (a15Ca12, a9Ca0) cs2 or cs3 ras3x casx rd/ wr dqmxx d31Cd0 bs tp tr tc1 tc2/td1 tc3/td2 tc4/td3 td4 figure 11.21 burst read timing (different row addresses) 336 ckio a25Ca16, a13 (a25C a16, a11) a12 (a10) a15, a14, a11Ca0 (a15Ca12, a9Ca0) cs2 or cs3 ras3x casx rd/ wr dqmxx d31Cd0 bs tr tc1 tc2 tc3 tc4 figure 11.22 burst write timing (no precharge) 337 ckio a25Ca16, a13 (a25C a16, a11) a12 (a10) a15, a14, a11Ca0 (a15Ca12, a9Ca0) cs2 or cs3 ras3x casx rd/ wr dqmxx d31Cd0 bs tc1 tc2 tc3 tc4 figure 11.23 burst write timing (same row address) 338 ckio a25Ca16, a13 (a25C a16, a11) a12 (a10) a15, a14, a11Ca0 (a15Ca12, a9Ca0) cs2 or cs3 ras3x casx rd/ wr dqmxx d31Cd0 bs tp tr tc1 tc2 tc3 td4 figure 11.24 burst write timing (different row addresses) 339 refreshing: the bus state controller is provided with a function for controlling synchronous dram refreshing. auto-refreshing can be performed by clearing the rmode bit to 0 and setting the rfsh bit to 1 in mcr. if synchronous dram is not accessed for a long period, self-refresh mode, in which the power consumption for data retention is low, can be activated by setting both the rmode bit and the rfsh bit to 1. ? auto-refreshing refreshing is performed at intervals determined by the input clock selected by bits cks2-0 in rtcsr, and the value set in rtcor. the value of bits cks2-0 in rtcor should be set so as to satisfy the refresh interval stipulation for the synchronous dram used. first make the settings for rtcor, rtcnt, and the rmode and rfsh bits in mcr, then make the cks2- cks0 setting. when the clock is selected by cks2-cks0, rtcnt starts counting up from the value at that time. the rtcnt value is constantly compared with the rtcor value, and if the two values are the same, a refresh request is generated and an auto-refresh is performed. at the same time, rtcnt is cleared to zero and the count-up is restarted. figure 11.38 shows the auto-refresh cycle timing. all-bank precharging is performed in the tp cycle, then an ref command is issued in the trr cycle following the interval specified by the tpc bits in mcr. after the trr cycle, new command output cannot be performed for the duration of the number of cycles specified by the tras bits in mcr plus the number of cycles specified by the tpc bits in mcr. the tras and tpc bits must be set so as to satisfy the synchronous dram refresh cycle time stipulation (active/active command delay time). auto-refreshing is performed in normal operation, in sleep mode, and in case of a manual reset. 340 rtcor value rtcnt h'00000000 rtcsr.cks(2C0) cmf external bus cmf flag cleared by start of refresh cycle = 000 000 rtcnt cleared to 0 when rtcnt = rtcor auto-refresh cycle time figure 11.25 auto-refresh operation 341 tp trr trrw trrw ckio cke csn ras3u , ras3l casu , casl rd/ wr figure 11.26 synchronous dram auto-refresh timing 342 ? self-refreshing self-refresh mode is a kind of standby mode in which the refresh timing and refresh addresses are generated within the synchronous dram. self-refreshing is activated by setting both the rmode bit and the rfsh bit to 1. the self-refresh state is maintained while the cke signal is low. synchronous dram cannot be accessed while in the self-refresh state. self-refresh mode is cleared by clearing the rmode bit to 0. after self-refresh mode has been cleared, command issuance is disabled for the number of cycles specified by the tpc bits in mcr. self-refresh timing is shown in figure 11.27. settings must be made so that self-refresh clearing and data retention are performed correctly, and auto-refreshing is performed at the correct intervals. when self-refreshing is activated from the state in which auto-refreshing is set, or when exiting standby mode other than through a power-on reset, auto-refreshing is restarted if rfsh is set to 1 and rmode is cleared to 0 when self-refresh mode is cleared. if the transition from clearing of self-refresh mode to the start of auto-refreshing takes time, this time should be taken into consideration when setting the initial value of rtcnt. making the rtcnt value 1 less than the rtcor value will enable refreshing to be started immediately. after self-refreshing has been set, the self-refresh state continues even if the chip standby state is entered using the sh7729rs standby function, and is maintained even after recovery from standby mode other than through a power-on reset. in case of a power-on reset, the bus state controllers registers are initialized, and therefore the self-refresh state is cleared. self-refreshing is performed in normal operation, in sleep mode, in standby mode, and in case of a manual reset. when using synchronous dram, use the following procedure to initiate self-refreshing. 1. clear the refresh control bit to 0. 2. write h'00 to the rtcnt register. 3. set the refresh control bit and refresh mode bit to 1. 343 trs1 ckio rd/wr csn ras3u, ras3l casu, casl cke (trs2) (trs2) trs3 (tpc) (tpc) tp figure 11.27 synchronous dram self-refresh timing ? relationship between refresh requests and bus cycle requests if a refresh request is generated during execution of a bus cycle, execution of the refresh is deferred until the bus cycle is completed. if a refresh request occurs when the bus has been released by the bus arbiter, refresh execution is deferred until the bus is acquired. if a match between rtcnt and rtcor occurs while a refresh is waiting to be executed, so that a new refresh request is generated, the previous refresh request is eliminated. in order for refreshing to be performed normally, care must be taken to ensure that no bus cycle or bus mastership occurs that is longer than the refresh interval. when a refresh request is generated, the irqout pin is asserted (driven low). therefore, normal refreshing can be performed by having the irqout pin monitored by a bus master other than the sh7729r requesting the bus, or the bus arbiter, and returning the bus to the sh7729r. when refreshing is started, and if no other interrupt request has been generated, the irqout pin is negated (driven high). 344 power-on sequence: in order to use synchronous dram, mode setting must first be performed after powering on. to perform synchronous dram initialization correctly, the bus state controller registers must first be set, followed by a write to the synchronous dram mode register. in synchronous dram mode register setting, the address signal value at that time is latched by a combination of the ras , cas , and rd/ wr signals. if the value to be set is x, the bus state controller provides for value x to be written to the synchronous dram mode register by performing a write to address h'ffffd000 + x for area 2 synchronous dram, and to address h'ffffe000 + x for area 3 synchronous dram. in this operation the data is ignored, but the mode write is performed as a byte-size access. to set burst read/single write, cas latency 1 to 3, wrap type = sequential, and burst length 1 supported by the sh7729r, arbitrary data is written in a byte-size access to the following addresses. with 32-bit bus width: area 2 area 3 cas latency 1 ffffd840 ffffe840 cas latency 2 ffffd880 ffffe880 cas latency 3 ffffd8c0 ffffe8c0 mode register setting timing is shown in figure 11.41. as a result of the write to address h'ffffd000 + x or h'ffffe000 + x, a precharge all banks (pall) command is first issued in the trp1 cycle, then a mode register write command is issued in the tmw1 cycle. address signals, when the mode-register write command is issued, are as follows: a15Ca9 = 0000100 (burst read and single write) a8Ca6 = cas latency a5 = 0 (burst type = sequential) a4Ca2 = 000 (burst length 1) before mode register setting, a 100 s idle time (depending on the memory manufacturer) must be guaranteed after powering on requested by the synchronous dram. if the reset signal pulse width is greater than this idle time, there is no problem in performing mode register setting immediately. the number of dummy auto-refresh cycles specified by the manufacturer (usually 8) or more must be executed. this is usually achieved automatically while various kinds of initialization are being performed after auto-refresh setting, but a way of carrying this out more dependably is to set a short refresh request generation interval just while these dummy cycles are being executed. with simple read or write access, the address counter in the synchronous dram used for auto- refreshing is not initialized, and so the cycle must always be an auto-refresh cycle. 345 ckio a11 a12 or a10 a9 to a2 csn rd/wr ras3u or ras3l casu or casl d31 to d0 cke trp1 trp2 trp3 trp4 tmw1 tmw2 tmw3 tmw4 (high) a15 to a13 or a15 to a12 figure 11.28 synchronous dram mode write timing 346 11.3.5 burst rom interface setting bits a0bst1C0, a5bst1C0, and a6bst1C0 in bcr1 to a non-zero value allows burst rom to be connected to areas 0, 5, and 6. the burst rom interface provides high-speed access to rom that has a nibble access function. the timing for nibble access to burst rom is shown in figure 11.29. two wait cycles are set. basically, access is performed in the same way as for normal space, but when the first cycle ends the cs0 signal is not negated, and only the address is changed before the next access is executed. when 8-bit rom is connected, the number of consecutive accesses can be set as 4, 8, or 16 by bits a0bst1C0, a5bst1C0, or a6bst1C0. when 16-bit rom is connected, 4 or 8 can be set in the same way. when 32-bit rom is connected, only 4 can be set. wait pin sampling is performed in the first access if one or more wait states are set, and is always performed in the second and subsequent accesses. the second and subsequent access cycles also comprise two cycles when a burst rom setting is made and the wait specification is 0. the timing in this case is shown in figure 11.30. 347 t1 tw tw tb2 tb1 tw tb2 ckio a25 to a4 a3 to a0 csn rd/wr rd d31 to d0 bs wait t2 note: for a write cycle, a basic bus cycle (write cycle) is performed. tb1 figure 11.29 burst rom wait access timing 348 t1 tb2 tb1 tb2 tb1 tb2 tb1 t2 ckio a25 to a4 a3 to a0 csn rd/wr rd d31 to d0 bs wait note: for a write cycle, a basic bus cycle (write cycle) is performed. figure 11.30 burst rom basic access timing 349 11.3.6 pcmcia interface in the sh7729r, setting the a5pcm bit in bcr1 to 1 makes the bus interface for physical space area 5 an ic memory card and i/o card interface as stipulated in jeida version 4.2 (pcmcia2.1). setting the a6pcm bit to 1 makes the bus interface for physical space area 6 an ic memory card and i/o card interface as stipulated in jeida version 4.2. when the pcmcia interface is used, a bus size of 8 or 16 bits can be set by bits a5sz1 and a5sz0, or a6sz1 and a6sz0, in bcr2. figure 11.31 shows an example of pcmcia card connection to the sh7729r. to enable active insertion of the pcmcia cards (i.e. insertion or removal while system power is being supplied), a 3-state buffer must be connected between the sh7729rs bus interface and the pcmcia cards. as operation in big-endian mode is not explicitly stipulated in the jeida/pcmcia specifications, the pcmcia interface for the sh7729r in big-endian mode is stipulated independently. however, the wait signal is ignored in the following three cases: ? a write to external address space in dual address mode with 16-byte dma transfer ? transfer from an external device with dack to external address space in single address mode with 16-byte dma transfer ? cache write-back access 350 a24 to a0 d15 to d0 rd/wr ce1b/(cs6) ce1a/(cs5) rd we1 iciord iciowr wait iois16 sh7729r a25 to a0 d15 to d0 ce2 oe we/pgm (iord) (iowr) wait (iois16) cd1, cd2 ce1 pc card (memory/io) g g g dir dir g d7 to d0 d15 to d8 a25 to a0 d15 to d0 ce2 oe we/pgm wait cd1, cd2 ce1 pc card (memory/io) g g g dir dir g d7 to d0 d15 to d8 ce2b ce2a output port card detection circuit card detection circuit figure 11.31 example of pcmcia interface 351 memory card interface basic timing: figure 11.32 shows the basic timing for the pcmcia ic memory card interface. when physical space areas 5 and 6 are designated as pcmcia interface areas, bus accesses are automatically performed as ic memory card interface accesses. with a high external bus frequency (ckio), the setup and hold times for the address (a24Ca0), card enable ( cs5 , ce2a , cs6 , ce2b ), and write data (d15Cd0) in a write cycle, become insufficient with respect to rd and wr (the we pin in the sh7729r). the sh7729r provides for this by enabling setup and hold times to be set for physical space areas 5 and 6 in the pcr register. also, software waits by means of a wcr2 register setting and hardware waits by means of the wait pin can be inserted in the same way as for the basic interface. figure 11.33 shows the pcmcia memory bus wait timing. 352 ckio tpcm1 tpcm2 a25 to a0 cexx rd/ wr d15 to d0 (read) d15 to d0 (read) rd (read) we (write) bs figure 11.32 basic timing for pcmcia memory card interface 353 ckio tpcm0 a25 to a0 rd/wr cexx rd (read) d15 to d0 (read) d15 to d0 (write) we (write) bs wait tpcm0w tpcm1 tpcm1w tpcm1w tpcm2 tpcm2w figure 11.33 wait timing for pcmcia memory card interface 354 memory card interface burst timing: in the sh7729r, when the ic memory card interface is selected, page mode burst access mode can be used, for read access only, by setting bits a5bst1 and a5bst0 in bcr1 for physical space area 5, or bits a6bst1 and a6bst0 in bcr1 for area 6. this burst access mode is not stipulated in jeida version 4.2 (pcmcia2.1), but allows high- speed data access using rom provided with a burst mode, etc. burst access mode timing is shown in figures 11.34 and 11.35. ckio tpcm1 a25 to a4 cexx a3 to a0 rd/ wr rd (read) d15 to d0 (read) bs tpcm2 tpcm1 tpcm2 tpcm1 tpcm2 tpcm1 tpcm2 figure 11.34 basic timing for pcmcia memory card interface burst access 355 ckio tpcm0 a25 to a4 cexx a3 to a0 rd/wr rd (read) d15 to d0 (read) bs wait tpcm1 tpcm1w tpcm1w tpcm1w tpcm2 tpcm1 tpcm1w tpcm2 tpcm2w figure 11.35 wait timing for pcmcia memory card interface burst access 356 when the entire 32-mbyte memory space is used as ic memory card interface space, the common memory/attribute memory switching signal reg is generated using a port, etc. if 16 mbytes or less of memory space is sufficient, using 16 mbytes of memory space as common memory space and 16 mbytes as attribute memory space enables the a24 pin to be used for the reg signal. i/o space i/o space i/o space area 5: h'14000000 area 5: h'16000000 area 6: h'18000000 area 6: h'1a000000 area 5: h'14000000 area 5: h'15000000 area 5: h'16000000 h'17000000 area 6: h'18000000 area 6: h'19000000 area 6: h'1a000000 h'1b000000 attribute memory common memory attribute memory common memory i/o space up to 16-mbyte capacity (reg = a24) 32-mbyte capacity (reg = i/o port) common memory/ attribute memory common memory/ attribute memory figure 11.36 pcmcia space allocation 357 i/o card interface timing: figures 11.37 and 11.38 show the timing for the pcmcia i/o card interface. switching between the i/o card interface and the ic memory card interface is performed according to the accessed address. when pcmcia is designed for physical space area 5, the bus access is automatically performed as an i/o card interface access when a physical address from h'16000000 to h'17ffffff is accessed. when pcmcia is designated for physical space area 6, the bus access is automatically performed as an i/o card interface access when a physical address from h'1a000000 to h'1bffffff is accessed. when accessing a pcmcia i/o card, the access should be performed using a non-cacheable area in virtual space (p2 or p3 space) or an area specified as non-cacheable by the mmu. when an i/o card interface access is made to a pcmcia card in little-endian mode, dynamic sizing of the i/o bus width is possible using the iois16 pin. when a 16-bit bus width is set for area 6, if the iois16 signal is high during a word-size i/o bus cycle, the i/o port is recognized as being 8 bits in width. in this case, a data access for only 8 bits is performed in the i/o bus cycle being executed, followed automatically by a data access for the remaining 8 bits. figure 11.39 shows the basic timing for dynamic bus sizing. in big-endian mode, the iois16 signal is not supported, and should be fixed low. 358 ckio tpci1 tpci2 a25 to a0 rd/wr cexx iciord (read) d15 to d0 (read) iciowr (write) d15 to d0 (write) bs figure 11.37 basic timing for pcmcia i/o card interface 359 ckio a25 to a0 rd/wr cexx iciord (read) iciowr (write) d15 to d0 (read) d15 to d0 (write) bs wait iois16 tpci0 tpci0w tpci1 tpci1w tpci1w tpci2 tpci2w figure 11.38 wait timing for pcmcia i/o card interface 360 ckio tpci0 a25 to a1 cexx a0 rd/wr iciord (read) d15 to d0 (read) iciowr (write) d15 to d0 (write) bs wait iois16 tpci1 tpci1w tpci2 tpci1 tpci1w tpci2 tpci2w figure 11.39 dynamic bus sizing timing for pcmcia i/o card interface 361 11.3.7 waits between access cycles a problem associated with higher external memory bus operating frequencies is that data buffer turn-off on completion of a read from a low-speed device may be too slow, causing a collision with data in the next access. this results in lower reliability or incorrect operation. to avoid this problem, a data collision prevention feature has been provided. this memorizes the preceding access area and the kind of read/write. if there is a possibility of a bus collision when the next access is started, a wait cycle is inserted before the access cycle thus preventing a data collision. there are two cases in which a wait cycle is inserted: when an access is followed by an access to a different area, and when a read access is followed by a write access from the sh7729r. when the sh7729r performs consecutive write cycles, the data transfer direction is fixed (from the sh7729r to other memory) and there is no problem. with read accesses to the same area, in principle, data is output from the same data buffer, and wait cycle insertion is not performed. bits aniw1 and aniw0 (n = 0, 2C6) in wcr1 specify the number of idle cycles to be inserted between access cycles when a physical space area access is followed by an access to another area, or when the sh7729r performs a write access after a read access to physical space area n. if there is originally space between accesses, the number of idle cycles inserted is the specified number of idle cycles minus the number of empty cycles. waits are not inserted between accesses when bus arbitration is performed, since empty cycles are inserted for arbitration purposes. 362 t1 ckio csm csn a25 to a0 bs rd/wr rd d31 to d0 t2 twait t1 t2 twait t1 t2 area m read area m inter-access wait specification area n inter-access wait specification area n space read area n space write figure 11.40 waits between access cycles 11.3.8 bus arbitration when a bus release request ( breq ) is received from an external device, buses are released after the bus cycle being executed is completed and a bus grant signal ( back ) is output . the bus is not released during burst transfers for cache fills or tas instruction execution between the read cycle and write cycle. bus arbitration is not executed in multiple bus cycles that are generated when the data bus width is shorter than the access size; i.e. in the bus cycles when longword access is executed for the 8-bit memory. at the negation of breq , back is negated and bus use is restarted . see appendix a.1, pin states, for the pin states when the bus is released. the sh7729r sometimes needs to retrieve a bus it has released. for example, when memory generates a refresh request or an interrupt request internally, the sh7729r must perform the appropriate processing. the sh7729r has a bus request signal ( irqout ) for this purpose. when it must retrieve the bus, it asserts the irqout signal. devices asserting an external bus release request receive the assertion of the irqout signal and negate the breq signal to release the bus. the sh7729r retrieves the bus and carries out the processing. 363 irqout pin assertion conditions: ? when a memory refresh request has been generated but the refresh cycle has not yet begun ? when an interrupt is generated with an interrupt request level higher than the setting of the interrupt mask bits (i3Ci0) in the status register (sr). (this does not depend on the sr.bl bit.) 11.3.9 bus pull-up with the sh7729r, address pin pull-up can be performed when the bus is released by setting the pula bit in bcr1 to 1. the address pins are pulled up for a 4-clock period after back is asserted. figure 11.41 shows the address pin pull-up timing. similarly, data pin pull-up can be performed by setting the puld bit in bcr1 to 1. the data pins should be pulled up when the data bus is not in use. the data pin pull-up timing for a read cycle is shown in figure 11.42, and the timing for a write cycle in figure 11.43. hi-z pull-up ckio a25Ca0 back figure 11.41 pull-up timing for pins a25 to a0 364 pull-up ckio d31Cd0 rd csn pull-up figure 11.42 pull-up timing for pins d31 to d0 (read cycle) pull-up ckio d31Cd0 wen csn pull-up figure 11.43 pull-up timing for pins d31 to d0 (write cycle) 365 11.3.10 mcs[0] to mcs[7] pin control the sh7729r is provided with pins mcs[0] C mcs[7] as dedicated cs pins for the rom connected to area 0 or 2. assertion of mcs[0] C mcs[7] is controlled by settings in mcscr0C mcscr7. this enables 32-, 64-, 128-, or 256-mbit memory to be connected to area 0 or area 2. table 11.15 shows mcscr0Cmcscr7 settings and mcs[0] C mcs[7] assertion conditions. as the mcs[0] C mcs[7] pins are multiplexed as the ptc0Cptc7 pins, when using these pins as mcs[0] C mcs[7] , the corresponding bits in the pccr register should be set to other function. when cs2/0 = 0 in the mcscr0 and when the ptc0 pin is switched to mcs[0] (when pcomd1Cpcomd0 are set to other function), the cs0 pin is also switched to mcs[0] . as port register writes operate on the peripheral clock, they take time compared with instruction execution by the cpu operating on the high-speed internal clock. therefore, if an instruction that accesses mcs[1] to mcs[7] is located several instructions after an instruction that switches port c to mcs , the switch from ptc[n] to mcsn and from cs0 to mcs[0] may not be performed correctly. to prevent this problem, the following switching procedure should be used. ? when the program runs with cache on (1) to switch port c to mcs , set the corresponding bits in the pccr register to 00 ("other function"). (2) read the pccr register and check whether the set value is read. repeat until the set value is read. (3) perform a dummy read from non-cacheable cs0 space (e.g. address h'a0000000). this will result in an access to the cs0 space, and immediately afterward, cs0 will be switched to mcs[0] , and port c[n] will be switched to mcs[n] . (4) access can now be made to the mcs[1] to mcs[7] spaces. ? when the program runs in mcs[0] space with cache off (1) set the pccr register as in (1) above. (2) place at least three nop instructions after the instruction in (1). as a result, when the pccr register is rewritten, an access to the cs0 space will be generated, and immediately afterward, cs0 will be switched to mcs[0] , and port c[n] will be switched to mcs[n] . (3) access can now be made to the mcs[1] to mcs[7] spaces. 366 table 11.15 mcscrx settings and mcs[x] assertion conditions (x: 0C7) mcscrx settings mcs[x] assertion conditions cs2/0 cap1 cap0 a25 a24 a23 a22 cs0 cs2 address bus a [25:0] notes 0110lhh' 0000000 to h'1ffffff 256-mbit rom 1 l h h'2000000 to h'3ffffff 1 0 0 0 l h h'0000000 to h'0ffffff 128-mbit rom 0 1 l h h'1000000 to h'1ffffff 1 0 l h h'2000000 to h'2ffffff 1 1 l h h'3000000 to h'3ffffff 0 1 000 l hh' 0000000 to h'07fffff 64-mbit rom 001 l hh' 0800000 to h'0ffffff 010 l hh' 1000000 to h'17fffff 011 l hh' 1800000 to h'1ffffff 100 l hh' 2000000 to h'27fffff 101 l hh' 2800000 to h'2ffffff 110 l hh' 3000000 to h'37fffff 111 l hh' 3800000 to h'3ffffff 0 0 0000 l hh' 0000000 to h'03fffff 32-mbit rom 0001 l hh' 0400000 to h'07fffff 0010 l hh' 0800000 to h'0bfffff 0011 l hh' 0c00000 to h'0ffffff 0100 l hh' 1000000 to h'13fffff 0101 l hh' 1400000 to h'17fffff 0110 l hh' 1800000 to h'1bfffff 0111 l hh' 1c00000 to h'1ffffff 1000 l hh' 2000000 to h'23fffff 1001 l hh' 2400000 to h'27fffff 1010 l hh' 2800000 to h'2bfffff 1011 l hh' 2c00000 to h'2ffffff 1100 l hh' 3000000 to h'33fffff 1101 l hh' 3400000 to h'37fffff 1110 l hh' 3800000 to h'3bfffff 1111 l hh' 3c00000 to h'3ffffff 367 table 11.15 mcscrx settings and mcs[x] assertion conditions (x: 0C7) (cont) mcscrx settings mcs[x] assertion conditions cs2/0 cap1 cap0 a25 a24 a23 a22 cs0 cs2 address bus a[25:0] notes 1110hlh' 0000000 to h'1ffffff 256-mbit rom 1 h l h'2000000 to h'3ffffff 1 0 0 0 h l h'0000000 to h'0ffffff 128-mbit rom 0 1 h l h'1000000 to h'1ffffff 1 0 h l h'2000000 to h'2ffffff 1 1 h l h'3000000 to h'3ffffff 0 1 000 hl h' 0000000 to h'07fffff 64-mbit rom 001 hl h' 0800000 to h'0ffffff 010 hl h' 1000000 to h'17fffff 011 hl h' 1800000 to h'1ffffff 100 hl h' 2000000 to h'27fffff 101 hl h' 2800000 to h'2ffffff 110 hl h' 3000000 to h'37fffff 111 hl h' 3800000 to h'3ffffff 0 0 0000 hl h' 0000000 to h'03fffff 32-mbit rom 0001 hl h' 0400000 to h'07fffff 0010 hl h' 0800000 to h'0bfffff 0011 hl h' 0c00000 to h'0ffffff 0100 hl h' 1000000 to h'13fffff 0101 hl h' 1400000 to h'17fffff 0110 hl h' 1800000 to h'1bfffff 0111 hl h' 1c00000 to h'1ffffff 1000 hl h' 2000000 to h'23fffff 1001 hl h' 2400000 to h'27fffff 1010 hl h' 2800000 to h'2bfffff 1011 hl h' 2c00000 to h'2ffffff 1100 hl h' 3000000 to h'33fffff 1101 hl h' 3400000 to h'37fffff 1110 hl h' 3800000 to h'3bfffff 1111 hl h' 3c00000 to h'3ffffff 368 369 section 12 direct memory access controller (dmac) 12.1 overview the sh7729r includes a four-channel direct memory access controller (dmac). the dmac can be used in place of the cpu to perform high-speed transfers between external devices that have dack (transfer request acknowledge signal), external memory, memory-mapped external devices, and on-chip peripheral modules (irda, scif, a/d converter, and d/a converter). using the dmac reduces the burden on the cpu and increases overall operating efficiency. 12.1.1 features the dmac has the following features. ? four channels ? 4-gb physical address space ? 8-bit, 16-bit, 32-bit, or 16-byte transfer (in 16-byte transfer, four 32-bit reads are executed, followed by four 32-bit writes.) ? 16 mbytes (16,777,216 transfers) ? address mode: dual address mode and single address mode are supported. in addition, direct address transfer mode or indirect address transfer mode can be selected. ? dual address mode transfer: both the transfer source and transfer destination are accessed by address. dual address mode has direct address transfer mode and indirect address transfer mode. direct address transfer mode: the values specified in the dmac registers indicates the transfer source and transfer destination. two bus cycles are required for one data transfer. indirect address transfer mode: data is transferred with the address stored prior to the address specified in the transfer source address in the dmac. other operations are the same as those of direct address transfer mode. this function is only available in channel 3. four bus cycles are required for one data transfer. ? single address mode transfer: either the transfer source or transfer destination peripheral device is accessed (selected) by means of the dack signal, and the other device is accessed by address. one transfer unit of data is transferred in one bus cycle. ? channel functions: the transfer mode that can be specified depends on the channel: ? channel 0: external request can be accepted. ? channel 1: external request can be accepted. ? channel 2: this channel has a source address reload function, which reloads a source address every four transfers. ? channel 3: in this channel, direct address mode or indirect address transfer mode can be specified. 370 ? reload function: the value that was specified in the source address register can be automatically reloaded every four dma transfers. this function is only available in channel 2. ? transfer requests ? external request (from two dreq pins (channels 0 and 1 only). dreq can be detected either by edge or by level.) ? on-chip module request (requests from on-chip peripheral modules such as serial communications interface (irda and scif), a/d converter (a/d) and a timer (cmt). this request can be accepted in all the channels.) ? auto request (the transfer request is generated automatically within the dmac.) ? selectable bus modes: cycle-steal mode or burst mode ? selectable channel priority levels: fixed mode: the channel priority is fixed. round-robin mode: the priority of the channel in which the execution request was accepted is made the lowest. ? interrupt request: an interrupt request to the cpu can be generated after the specified number of transfers. 371 12.1.2 block diagram figure 12.1 shows a block diagram of the dmac. peripheral bus internal bus dreq0 , dreq1 iteration control sarn dmac module register control start-up control request priority control bus interface bus state controller on-chip peripheral module darn dmatcrn chcrn dmaor irda, scif a/d converter cmt dein x/y memory external ram external rom external i/o (memory mapped) external i/o (with acknowledge) dack0, dack1 drak0, drak1 legend dmaor: sarn: darn: dmatcrn: chcrn: dein: n: dmac operation register dmac source address register dmac destination address register dmac transfer count register dmac channel control register dma transfer-end interrupt request to cpu 0 to 3 figure 12.1 block diagram of dmac 372 12.1.3 pin configuration table 12.1 shows the dmac pins. table 12.1 dmac pins channel name symbol i/o function 0 dma transfer request dreq0 i dma transfer request input from external device to channel 0 dreq acknowledge dack0 o strobe output to an external i/o upon dma transfer request from external device to channel 0 dma request acknowledge drak0 o output showing that dreq0 has been accepted 1 dma transfer request dreq1 i dma transfer request input from external device to channel 1 dreq acknowledge dack1 o strobe output to an external i/o upon dma transfer request from external device to channel 1 dma request acknowledge drak1 o output showing that dreq1 has been accepted 373 12.1.4 register configuration table 12.2 summarizes the dmac registers. the dmac has a total of 17 registers: four control registers for each other control register shared by all channels. table 12.2 dmac registers channel name abbrevi- ation r/w initial value address access size 0 dma source address register 0 sar0 r/w undefined h'04000020 (h'a4000020)* 4 16, 32* 2 dma destination address register 0 dar0 r/w undefined h'04000024 (h'a4000024)* 4 16, 32* 2 dma transfer count register 0 dmatcr0 r/w undefined h'04000028 (h'a4000028)* 4 16, 32* 3 dma channel control register 0 chcr0 r/w* 1 h'00000000 h'0400002c (h'a400002c)* 4 8, 16, 32* 2 1 dma source address register 1 sar1 r/w undefined h'04000030 (h'a4000030)* 4 16, 32* 2 dma destination address register 1 dar1 r/w undefined h'04000034 (h'a4000034)* 4 16, 32* 2 dma transfer count register 1 dmatcr1 r/w undefined h'04000038 (h'a4000038)* 4 16, 32* 3 dma channel control register 1 chcr1 r/w* 1 h'00000000 h'0400003c (h'a400003c)* 4 8, 16, 32* 2 2 dma source address register 2 sar2 r/w undefined h'04000040 (h'a4000040)* 4 16, 32* 2 dma destination address register 2 dar2 r/w undefined h'04000044 (h'a4000044)* 4 16, 32* 2 dma transfer count register 2 dmatcr2 r/w undefined h'04000048 (h'a4000048)* 4 16, 32* 3 dma channel control register 2 chcr2 r/w* 1 h'00000000 h'0400004c (h'a400004c)* 4 8, 16, 32* 2 374 table 12.2 dmac registers (cont) channel name abbrevi- ation r/w initial value address access size 3 dma source address register 3 sar3 r/w undefined h'04000050 (h'a4000050)* 4 16, 32* 2 dma destination address register 3 dar3 r/w undefined h'04000054 (h'a4000054)* 4 16, 32* 2 dma transfer count register 3 dmatcr3 r/w undefined h'04000058 (h'a4000058)* 4 16, 32* 3 dma channel control register 3 chcr3 r/w* 1 h'00000000 h'0400005c (h'a400005c)* 4 8, 16, 32* 2 shareddma operation register dmaor r/w* 1 h'0000 h'04000060 (h'a4000060)* 4 8, 16* 2 notes: these registers are located in area 1 of physical space. therefore, when the cache is on, either access these registers from the p2 area of logical space or else make an appropriate setting using the mmu so that these registers are not cached. *1 only 0 can be written to bit 1 of chcr0 to chcr3, and bits 1 and 2 of dmaor to clear the flag after 1 is read. *2 if 16-bit access is used on sar0 to sar3, dar0 to dar3, and chcr0 to chcr3, the value in the 16 bits that were not accessed is retained. *3 dmatcr comprises the 24 bits from bit 0 to bit 23. the upper 8 bits, bits 24 to 31, cannot be written with 1 and are always read as 0. *4 when address translation by the mmu does not apply, the address in parentheses should be used. 375 12.2 register descriptions 12.2.1 dma source address registers 0C3 (sar0Csar3) dma source address registers 0C3 (sar0Csar3) are 32-bit readable/writable registers that specify the source address of a dma transfer. during a dma transfer, these registers indicate the next source address. to transfer data in 16 bits or in 32 bits, specify a 16-bit or 32-bit address boundary address. when transferring data in 16-byte units, a 16-byte boundary (address 16n) must be set for the source address value. operation is not guaranteed if other addresses are specified. the initial value is undefined in a reset. the previous value is retained in standby mode. bit: 31 30 29 28 27 26 25 24 initial value: r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 23 22 21 20 0 initial value: r/w: r/w r/w r/w r/w r/w 376 12.2.2 dma destination address registers 0C3 (dar0Cdar3) dma destination address registers 0C3 (dar0Cdar3) are 32-bit readable/writable registers that specify the destination address of a dma transfer. these registers include a count function, and during a dma transfer, these registers indicate the next destination address. to transfer data in 16 bits or in 32 bits, specify a 16-bit or 32-bit address boundary address. operation is not guaranteed if other addresses are specified. the initial value is undefined in a reset. the previous value is retained in standby mode. bit: 31 30 29 28 27 26 25 24 initial value: r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 23 22 21 20 0 initial value: r/w: r/w r/w r/w r/w r/w 377 12.2.3 dma transfer count registers 0C3 (dmatcr0Cdmatcr3) dma transfer count registers 0C3 (dmatcr0Cdmatcr3) are 24-bit readable/writable registers that specify the dma transfer count (bytes, words, or longwords). the number of transfers is 1 when the setting is h'000001, and 16,777,216 (the maximum) when h'000000 is set. during a dma transfer, these registers indicate the remaining number of transfers. in 16-byte transfer, one 16-byte transfer (128 bits) is counted as one. writing to upper eight bits in dmatcr is invalid; 0s are read if these bits are read. when using 16-byte transfer, an integral multiple of 4 (4n) must be set for the number of transfers to ensure normal operation. the initial value is undefined in a reset. the previous value is retained in standby mode. bit: 31 30 29 28 27 26 25 24 initial value: r/w: r r r r r r r r bit: 23 22 21 20 ... 0 ... initial value: ... r/w: r/w r/w r/w r/w ... r/w 378 12.2.4 dma channel control registers 0C3 (chcr0Cchcr3) dma channel control registers 0C3 (chcr0Cchcr3) are 32-bit readable/writable registers that specify the operation mode, transfer method, etc., for each channel. writing to bits 31 to 21 and 7 in this register is invalid; 0s are read if these bits are read. bit 20 is only used in chcr3; it is not used in chcr0 to chcr2. consequently, writing to this bit is invalid in chcr0 to chcr2; 0 is read if this bit is read. bit 19 is only used in chcr2; it is not used in chcr0, chcr1, and chcr3. consequently, writing to this bit is invalid in chcr0, chcr1, and chcr3; 0 is read if this bit is read. bits 6 and 16 to 18 are only used in chcr0 and chcr1; they are not used in chcr2 and chcr3. consequently, writing to these bits is invalid in chcr2 and chcr3; 0s are read if these bits are read. these register values are initialized to zero in a power-on reset. the previous value is retained in standby mode. bit: 31 ... 21 20 19 18 17 16 ... di ro rl am al initial value: 0 ... 0 0 0 0 0 0 r/w: r ... r (r/w)* 2 (r/w)* 2 (r/w)* 2 (r/w)* 2 (r/w)* 2 bit: 15 14 13 12 11 10 9 8 dm1 dm0 sm1 sm0 rs3 rs2 rs1 rs0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 3 2 1 0 ds tm ts1 ts0 ie te de initial value: 0 0 0 0 0 0 0 0 r/w: r (r/w)* 2 r/w r/w r/w r/w r/(w)* 1 r/w notes: *1 only 0 can be written to the te bit after 1 is read. *2 the di, ro, rl, am, al, and ds bits are not included in some channels. 379 bits 31 to 21, 7reserved: these bits are always read as 0. the write value should always be 0. bit 20direct/indirect selection (di): selects direct address mode or indirect address mode in channel 3. this bit is only valid in chcr3. writing to this bit is invalid in chcr0 to chcr2; 0 is read if this bit is read. when using 16-byte transfer, direct address mode must be specified. operation is not guaranteed if indirect address mode is specified. bit 20: di description 0 direct address mode (initial value) 1 indirect address mode bit 19source address reload bit (ro): selects whether the source address initial value is reloaded in channel 2. this bit is only valid in chcr2. writing to this bit is invalid in chcr0, chcr1, and chcr3; 0 is read if this bit is read. when using 16-byte transfer, this bit must be cleared to 0, specifying non- reloading. operation is not guaranteed if reloading is specified. bit 19: ro description 0 source address is not reloaded (initial value) 1 source address is reloaded bit 18request check level bit (rl): specifies whether drak ( dreq acknowledge) signal output is active-high or active-low. this bit is only valid in chcr0 and chcr1. writing to this bit is invalid in chcr2 and chcr3; 0 is read if this bit is read. bit 18: rl description 0 active-low drak output (initial value) 1 active-high drak output 380 bit 17acknowledge mode bit (am): specifies whether dack is output in the data read cycle or in the data write cycle in dual address mode. this bit is only valid in chcr0 and chcr1. writing to this bit is invalid in chcr2 and chcr3; 0 is read if this bit is read. bit 17: am description 0 dack output in read cycle (initial value) 1 dack output in write cycle bit 16acknowledge level (al): specifies whether dack (acknowledge) signal output is active-high or active-low. this bit is only valid in chcr0 and chcr1. writing to this bit is invalid in chcr2 and chcr3; 0 is read if this bit is read. bit 16: al description 0 active-low dack output (initial value) 1 active-high dack output bits 15 and 14destination address mode bits 1 and 0 (dm1, dm0): select whether the dma destination address is incremented, decremented, or left fixed. bit 15: dm1 bit 14: dm0 description 0 0 fixed destination address * (initial value) 0 1 destination address is incremented (+1 in 8-bit transfer, +2 in 16-bit transfer, +4 in 32-bit transfer, +16 in 16-byte transfer) 1 0 destination address is decremented (C1 in 8-bit transfer, C2 in 16-bit transfer, C4 in 32-bit transfer; illegal setting in 16-byte transfer) 1 1 setting prohibited note: * this setting cannot be used when the transfer destination is x/y memory in 16-byte transfer. 381 bits 13 and 12source address mode bits 1 and 0 (sm1, sm0): select whether the dma source address is incremented, decremented, or left fixed. bit 13: sm1 bit 12: sm0 description 0 0 fixed source address * (initial value) 0 1 source address is incremented (+1 in 8-bit transfer, +2 in 16- bit transfer, +4 in 32-bit transfer, +16 in 16-byte transfer) 1 0 source address is decremented (C1 in 8-bit transfer, C2 in 16- bit transfer, C4 in 32-bit transfer; illegal setting in 16-byte transfer) 1 1 setting prohibited note: * this setting cannot be used when the transfer destination is x/y memory in 16-byte transfer. if the transfer source is specified by indirect address, specify the address holding the value of the address in which the data to be transferred is stored (i.e. the indirect address) in source address register 3 (sar3). specification of sar3 incrementing or decrementing in indirect address mode depends on the sm1 and sm0 settings. in this case, however, the sar3 increment or decrement value is +4, C4, or fixed at 0, regardless of the transfer data size specified in ts1 and ts0. 382 bits 11 to 8resource select bits 3 to 0 (rs3 to rs0): specify which transfer requests will be sent to the dmac. bit 11: rs3 bit 10: rs2 bit 9: rs1 bit 8: rs0 description 0000 external request * , dual address mode (initial value) 0001 setting prohibited 0010 external request / single address mode external address space external device with dack 0011 external request / single address mode external device with dack external address space 0100 auto request 0101 setting prohibited 0110 setting prohibited 0111 setting prohibited 1000 setting prohibited 1001 setting prohibited 1010 irda transmission 1011 irda reception 1100 scif transmission 1101 scif reception 1110 internal a/d 1111cmt notes: when using 16-byte transfer, the following settings must not be made: 1010 irda transmission 1011 irda reception 1100 scif transmission 1101 scif reception 1110 a/d converter 1111 cmt operation is not guaranteed if these settings are made. * external request specification is valid only in channels 0 and 1. none of the request sources can be selected in channels 2 and 3. 383 bit 6 dreq select bit (ds): selects low-level or falling-edge detection as the sampling method for the dreq pin used in external request mode. this bit is only valid in chcr0 and chcr1. writing to this bit is invalid in chcr2 and chcr3; 0 is read if this bit is read. in channels 0 and 1, if an on-chip peripheral module is specified as a transfer request source or an auto-request is specified, the specification of this bit is ignored and falling-edge detection is fixed except in an auto-request. bit 6: ds description 0 dreq detected by low level (initial value) 1 dreq detected at falling edge bit 5transmit mode (tm): specifies the bus mode when transferring data. bit 5: tm description 0 cycle-steal mode (initial value) 1 burst mode bits 4 and 3transmit size bits 1 and 0 (ts1, ts0): specify the size of data to be transferred. bit 4: ts1 bit 3: ts0 description 0 0 byte size (8 bits) (initial value) 0 1 word size (16 bits) 1 0 longword size (32 bits) 1 1 16-byte unit (4 longword transfers) bit 2interrupt enable bit (ie): if this bit is set to 1, an interrupt is requested on completion of the number of data transfers specified in dmatcr (i.e. when te = 1). bit 2: ie description 0 interrupt request is not generated on completion of data transfers specified in dmatcr (initial value) 1 interrupt request is generated on completion of data transfers specified in dmatcr 384 bit 1transfer end bit (te): set to 1 on completion of the number of data transfers specified in dmatcr. at this time, if the ie bit is set to 1, an interrupt request is generated. if data transfer ends due to an nmi interrupt, a dmac address error, or clearing of the de bit or the dme bit in dmaor before this bit is set to 1, this bit will not be set to 1. even if the de bit is set to 1 while this bit is set to 1, transfer is not enabled. bit 1: te description 0 data transfers specified in dmatcr not completed (initial value) clearing conditions: writing 0 to te after reading te = 1 power-on reset, manual reset 1 data transfers specified in dmatcr completed bit 0dmac enable bit (de): enables operation of the corresponding channel. bit 0: de description 0 channel operation disabled (initial value) 1 channel operation enabled if an auto-request is specified (rs3 to rs0), transfer starts when this bit is set to 1. in an external request or an internal module request, transfer starts when a transfer request is generated after this bit is set to 1. clearing this bit during transfer terminates the transfer. even if the de bit is set, transfer is not enabled if the te bit is 1, the dme bit in dmaor is 0, or the nmif bit in dmaor is 1. 385 12.2.5 dma operation register (dmaor) the dma operation register (dmaor) is a 16-bit readable/writable register that controls the dmac transfer mode. writing to bits 15 to 10 and bits 7 to 3 is invalid in this register; 0 is always read if these bits are read. dmaor is initialized to 0 by a power-on reset, and in hardware standby mode or software standby mode. bit: 15 14 13 12 11 10 9 8 pr1pr0 initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r/w r/w bit: 7 6 5 4 3 2 1 0 ae nmif dme initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r/(w)* r/(w)* r/w note: * only 0 can be written to the ae and nmif bits after 1 is read. bits 15 to 10reserved: these bits are always read as 0. the write value should always be 0. bits 9 and 8priority mode bits 1 and 0 (pr1, pr0): select the priority level between channels when there are simultaneous transfer requests for multiple channels. bit 9: pr1 bit 8: pr0 description 0 0 ch0 > ch1 > ch2 > ch3 (initial value) 0 1 ch0 > ch2 > ch3 > ch1 1 0 ch2 > ch0 > ch1 > ch3 1 1 round-robin bits 7 to 3reserved: these bits are always read as 0. the write value should always be 0. 386 bit 2address error flag bit (ae): indicates that an address error occurred during dma transfer. if this bit is set during data transfer, transfers on all channels are suspended. the cpu cannot write 1 to this bit. this bit can only be cleared by writing 0 after reading 1. bit 2: ae description 0 no dmac address error; dma transfer is enabled (initial value) clearing conditions: writing 0 to ae after reading ae = 1 power-on reset, manual reset 1 dmac address error; dma transfer is disabled this bit is set by occurrence of a dmac address error bit 1nmi flag bit (nmif): indicates that an nmi interrupt occurred. this bit is set regardless of whether the dmac is in the operating or halted state. the cpu cannot write 1 to this bit. only 0 can be written to clear this bit after 1 is read. bit 1: nmif description 0 no nmi input; dma transfer is enabled(initial value) clearing conditions: writing 0 to nmif after reading nmif = 1 power-on reset, manual reset 1 nmi input; dma transfer is disabled this bit is set by occurrence of an nmi interrupt bit 0dma master enable bit (dme): enables or disables dma transfers on all channels. if the dme bit and the de bit corresponding to each channel in chcr are set to 1, transfer is enabled on the corresponding channel. if this bit is cleared during transfer, transfer on all the channels will be terminated. even if the dme bit is set, transfer is not enabled if the te bit is 1 or the de bit is 0 in chcr, or the nmif bit is 1 in dmaor. bit 0: dme description 0 dma transfer disabled on all channels (initial value) 1 dma transfer enabled on all channels 387 12.3 operation when there is a dma transfer request, the dmac starts the transfer according to the predetermined channel priority order; when the transfer end conditions are satisfied, it ends the transfer. transfers can be requested in three modes: auto-request, external request, and on-chip module request. the dual address mode has direct address transfer mode and indirect address transfer mode. burst mode or cycle-steal mode can be selected as the bus mode. 12.3.1 dma transfer flow after the dma source address register (sar), dma destination address register (dar), dma transfer count register (dmatcr), dma channel control register (chcr), and dma operation register (dmaor) are set, the dmac transfers data according to the following procedure: 1. checks to see if transfer is enabled (de = 1, dme = 1, te = 0, nmif = 0) 2. when a transfer request comes and transfer is enabled, the dmac transfers 1 transfer unit of data (according to the ts0 and ts1 settings). for an auto-request, the transfer begins automatically when the de bit and dme bit are set to 1. the dmatcr value will be decremented for each transfer. the actual transfer flows vary by address mode and bus mode. 3. when the specified number of transfers have been completed (when dmatcr reaches 0), the transfer ends normally. if the ie bit in chcr is set to 1 at this time, a dei interrupt is sent to the cpu. 4. when an nmi interrupt is generated, the transfer is aborted. transfers are also aborted when the de bit in chcr or the dme bit in dmaor are changed to 0. figure 12.2 is a flowchart of this procedure. 388 normal end nmif = 1 or de = 0 or dme = 0? bus mode, transfer request mode, dreq detection selection system initial settings (sar, dar, dmatcr, chcr, dmaor) transfer (1 transfer unit); dmatcr C 1 dmatcr, sar and dar updated dei interrupt request (when ie = 1) no yes no yes no yes yes no yes no * 3 * 2 start transfer aborted dmatcr = 0? transfer request?* 1 de, dme = 1 and nmif, te = 0? does nmif = 1 or de = 0 or dme = 0? transfer end notes: *1 in auto-request mode, transfer begins when nmif and te are both 0 and the de and dme bits are set to 1. *2 dreq = level detection in burst mode (external request) or cycle-steal mode. *3 dreq = edge detection in burst mode (external request), or auto-request mode in burst mode. figure 12.2 dmac transfer flowchart 389 12.3.2 dma transfer requests dma transfer requests are basically generated in either the data transfer source or destination, but they can also be generated by devices and on-chip peripheral modules that are neither the source nor the destination. transfers can be requested in three modes: auto-request, external request, and on-chip module request. the request mode is selected in the rs3Crs0 bits of dma channel control registers 0C3 (chcr0Cchcr3). auto-request mode: when there is no transfer request signal from an external source, as in a memory-to-memory transfer or a transfer between memory and an on-chip peripheral module unable to request a transfer, the auto-request mode allows the dmac to automatically generate a transfer request signal internally. when the de bit of chcr0Cchcr3 and the dme bit of dmaor are set to 1, the transfer begins so long as the te bit of chcr0Cchcr3 and the nmif bit of dmaor are 0. external request mode: in this mode a transfer is performed in response to the request signal ( dreq ) of an external device. choose one of the modes shown in table 12.3 according to the application system. when this mode is selected, if dma transfer is enabled (de = 1, dme = 1, te = 0, nmif = 0), a transfer is performed upon a request at the dreq input. choose dreq detection by either a falling edge or low level of the signal input with the ds bit in chcr0 and chcr1 (ds = 0 for level detection, ds = 1 for edge detection). the source of the transfer request does not have to be the data transfer source or destination. table 12.3 selecting external request modes with rs bits rs3 rs2 rs1 rs0 address mode source destination 0000 dual address mode any* any* 1 0 single address mode external memory, memory-mapped external device external device with dack 1 external device with dack external memory, memory-mapped external device note: * external memory, memory-mapped external device, on-chip memory, on-chip peripheral module (excluding dmac, ubc, and bsc) on-chip module request: in this mode a transfer is performed in response to a transfer request signal (interrupt request signal) of an on-chip module. this mode cannot be set in case of 16-byte transfer. these are six transfer request signals: the receive-data-full interrupts (rxi) and the transmit-data-empty interrupts (txi) from two serial communication interfaces (irda, scif), the a/d conversion end interrupt (adi) of the a/d converter, and the compare match timer interrupt (cmi) of the cmt (table 12.4). when this mode is selected, if dma transfer is enabled (de = 1, 390 dme = 1, te = 0, nmif = 0), a transfer is performed upon input of a transfer request signal. the source of the transfer request does not have to be the data transfer source or destination. when rxi is set as the transfer request, however, the transfer source must be the sci's receive data register (rdr). likewise, when txi is set as the transfer request, the transfer source must be the sci's transmit data register (tdr). if the transfer requester is the a/d converter, the data transfer source must be the a/d data register. table 12.4 selecting on-chip peripheral module request modes with rs bits rs3 rs2 rs1 rs0 dma transfer request source dma transfer request signal source desti- nation bus mode 1010 irda transmitter txi1 (irda transmit-data-empty interrupt transfer request) any* tdr1 cycle-steal 1011 irda receiver rxi1 (irda receive-data-full interrupt transfer request) rdr1 any* cycle-steal 1100 scif transmitter txi2 (scif transmit-data-empty interrupt transfer request) any* tdr2 cycle-steal 1101 scif receiver rxi2 (scif receive-data-full interrupt transfer request) rdr1 any* cycle-steal 1110a/d converter adi (a/d conversion end interrupt) addr any* cycle-steal 1111cmt cmi (compare match timer interrupt) any* any* burst/ cycle-steal addr: a/d data register of a/d converter note: * external memory, memory-mapped external device, on-chip peripheral module (excluding dmac, bsc, ubc) when outputting transfer requests from on-chip peripheral modules, the appropriate interrupt enable bits must be set to output the interrupt signals. if the interrupt request signal of the on-chip peripheral module is used as a dma transfer request signal, an interrupt is not sent to the cpu. the dma transfer request signals in table 12.4 are automatically discontinued when the corresponding dma transfer is performed. if cycle-steal mode is being employed, they are withdrawn at the first transfer; if burst mode is being used, they are discontinued at the last transfer. 391 12.3.3 channel priority when the dmac receives simultaneous transfer requests on two or more channels, it selects a channel according to a predetermined priority order. two modes (fixed mode and round-robin mode) are selected by priority bits pr1 and pr0 in the dma operation register (dmaor). fixed mode: in these modes, the priority order of the channels remain fixed. there are three kinds of fixed modes as follows: ch0 > ch1 > ch2 > ch3 ch0 > ch2 > ch3 > ch1 ch2 > ch0 > ch1 > ch3 these are selected by the pr1 and pr0 bits in dmaor. round-robin mode: each time one word, byte, or longword is transferred on one channel, the priority order is rotated. the channel on which the transfer was just finished rotates to the bottom of the priority order. the round-robin mode operation is shown in figure 12.3. the priority of the round-robin mode is ch0 > ch1 > ch2 > ch3 immediately after reset. 392 ch1 > ch2 > ch3 > ch0 ch0 > ch1 > ch2 > ch3 ch2 > ch3 > ch0 > ch1 ch0 > ch1 > ch2 > ch3 ch2 > ch3 > ch0 > ch1 ch0 > ch1 > ch2 > ch3 ch0 > ch1 > ch2 > ch3 ch3 > ch0 > ch1 > ch2 ch0 > ch1 > ch2 > ch3 (1) when channel 0 transfers initial priority order initial priority order initial priority order priority order after transfer priority order after transfer priority order does not change. channel 2 becomes lowest- priority. the priority of channels 0 and 1, which were higher than channel 2, are also shifted. if immediately after there is a request to transfer channel 1 only, channel 1 becomes lowest-priority and the priority of channels 0 and 3, which were higher than channel 1, is also shifted. channel 0 becomes lowest- priority. the priority of channel 0, which was higher than channel 3, is also shifted. channel 0 becomes lowest- priority. priority order after transfer priority order after transfer priority order after transfer post-transfer priority order when there is an immediate transfer request to channel 1 only (2) when channel 1 transfers (3) when channel 2 transfers (4) when channel 3 transfers figure 12.3 round-robin mode 393 figure 12.4 shows how the priority order changes when channel 0 and channel 3 transfers are requested simultaneously and a channel 1 transfer is requested during the channel 0 transfer. the dmac operates as follows: 1. transfer requests are generated simultaneously for channels 0 and 3. 2. channel 0 has a higher priority, so the channel 0 transfer begins first (channel 3 waits for transfer). 3. a channel 1 transfer request occurs during the channel 0 transfer (channels 1 and 3 are both waiting) 4. when the channel 0 transfer ends, channel 0 becomes lowest-priority. 5. at this point, channel 1 has a higher priority than channel 3, so the channel 1 transfer begins (channel 3 waits for transfer). 6. when the channel 1 transfer ends, channel 1 becomes lowest-priority. 7. the channel 3 transfer begins. 8. when the channel 3 transfer ends, channels 3 and 2 shift downward in priority so that channel 3 becomes the lowest-priority. transfer request waiting channel(s) dmac operation channel priority (1) channels 0 and 3 (3) channel 1 0 > 1 > 2 > 3 (2) channel 0 transfer start (4) channel 0 transfer ends (5) channel 1 transfer starts (6) channel 1 transfer ends (7) channel 3 transfer starts (8) channel 3 transfer ends 1 > 2 > 3 > 0 2 > 3 > 0 > 1 0 > 1 > 2 > 3 priority order changes priority order changes priority order changes none 3 3 1,3 figure 12.4 changes in channel priority in round-robin mode 394 12.3.4 dma transfer types the dmac supports the transfers shown in table 12.5. in dual address mode, both the transfer source address and the transfer destination address are output. dual address mode has a direct address mode and indirect address mode. in direct address mode, an output address value is the data transfer target address; in indirect address mode, the value stored in the output address, not the output address value itself, is the data transfer target address. data transfer timing depends on the bus mode, which may be cycle-steal mode or burst mode. table 12.5 supported dma transfers destination source external device with dack external memory memory- mapped external device on-chip peripheral module xy memory external device with dack not available dual, single dual, single not available not available external memory dual, single dual dual dual dual memory-mapped external device dual, single dual dual dual dual on-chip peripheral module not available dual dual dual dual x/y memory not available dual dual dual dual notes: 1. dual: dual address mode 2. single: single address mode 3. dual address mode includes direct address mode and indirect address mode. 4. 16-byte transfer is not available for on-chip peripheral modules. address modes: ? dual address mode in dual address mode, both the transfer source and destination are accessed (selectable) by an address. the source and destination can be located externally or internally. dual address mode has (1) a direct address transfer mode and (2) an indirect address transfer mode. 395 (1) in direct address transfer mode, dma transfer requires two bus cycles because data is read from the transfer source in a data read cycle and written to the transfer destination in a data write cycle. at this time, transfer data is temporarily stored in the dmac. in the transfer between external memories as shown in figure 12.5, data is read to the dmac from one external memory in a data read cycle, and then that data is written to the other external memory in a write cycle. figures 12.6 to 12.8 show examples of the timing at this time. data buffer address bus data bus address bus data bus memory transfer source module transfer destination module memory transfer source module transfer destination module sar dar data buffer sar dar the sar value is an address, data is read from the transfer source module, and the data is temporarily stored in the dmac. first bus cycle second bus cycle the dar value is an address, and the value stored in the data buffer in the dmac is written to the transfer destination module. dmac dmac figure 12.5 operation of direct address mode in dual address mode 396 (1st cycle) (2nd cycle) data read cycle data write cycle transfer source address transfer destination address ckio a25 to a0 csn d31 to d0 rd wen dackn note: in transfer between external memories, with dack output in the read cycle, dack output timing is the same as that of csn . figure 12.6 example of dma transfer timing in the direct address mode in dual mode (transfer source: ordinary memory, transfer destination: ordinary memory) 397 transfer source address +4 +8 +12 transfer destination address +4 +8 +12 data read cycle (1st cycle) (2nd cycle) a25-a0 ckio csn rd wem dackn d31-d0 note: in transfer between external memories, with dack output in the read cycle, dack output timing is the same as that of csn . figure 12.7 example of dma transfer timing in the direct address mode in dual mode (16-byte transfer, transfer source: ordinary memory, transfer destination: ordinary memory) transfer source address transfer destination address +4 +8 +12 a25-a0 ckio csn ras cas rd/ wr dackn d31-d0 data read cycle (1st cycle) (2nd cycle) data write cycle note: in transfer between external memories, with dack output in the read cycle, dack output timing is the same as that of csn . figure 12.8 example of dma transfer timing in the direct address mode in dual mode (16-byte transfer, transfer source: synchronous dram, transfer destination: ordinary memory) 398 (2) in indirect address transfer mode, the address of memory in which data to be transferred is stored is specified in the transfer source address register (sar3) in the dmac. 16-byte transfer is not possible. consequently, in this mode, the address value specified in the transfer source address register in the dmac is read first. this value is temporarily stored in the dmac. next, the read value is output as an address, and the value stored in that address is stored in the dmac again. then, the value read afterwards is written to the address specified in the transfer destination address; this completes one dma transfer. figure 12.9 shows an example. in this example, the transfer destination, the transfer source, and the storage destination of the indirect address are external memories, and transfer data is 16 or 8 bits. figure 12.10 shows an example of the transfer timing. in this mode, one nop cycle (ck1 cycle shown in figure 12.10) is required to output data read as an indirect address to an address bus. if transfer data is 32 bits, the third and fourth bus cycles shown in figure 12.10 are required twice for each; a total of six bus cycles and one nop cycle are required. 399 memory transfer source module transfer destination module sar3 dar3 data buffer temporary buffer d m a c when the value in sar3 is an address, the memory data is read and the value is stored in the temporary buffer. the value to be read must be 32 bits since it is used for the address. memory transfer source module data bus address bus transfer destination module sar3 dar3 data buffer temporary buffer d m a c memory transfer source module data bus address bus transfer destination module sar3 dar3 data buffer temporary buffer d m a c first and second bus cycles when the value in the temporary buffer is an address, the data is read from the transfer source module to the data buffer. third bus cycle fourth bus cycle when the value in sar3 is an address, the value in the data buffer is written to the transfer source module. note: this example shows memory, the transfer source module, and the transfer destination module; in practice, any module can be connected in the addressing space. data bus address bus figure 12.9 indirect address operation in dual address mode (when external memory space has a 16-bit width) 400 transfer source address (h) transfer source address (l) indirect address nop transfer destination address indirect address (h) indirect address (l) transfer data transfer data transfer data transfer data transfer data transfer source address ? 1 transfer source address ? 2 indirect address nop indirect address address read cycle (1st) (2nd) (3rd) nop cycle data read cycle (4th) data write cycle ckio a25 to a0 csn d31 to d0 internal address bus internal data bus dmac indirect address buffer dmac data buffer rd wen notes: * 1 * 2 the internal address bus value does not change, and is controlled by the port. the dmac does not fetch the value until 32-bit data is output to the internal data bus. transfer between external memories figure 12.10 example of transfer timing in the indirect address mode in dual address mode 401 ? single address mode in single address mode, either the transfer source or transfer destination peripheral device is accessed (selected) by means of the dack signal, and the other device is accessed by address. in this mode, the dmac performs one dma transfer in one bus cycle, accessing one of the external devices by outputting the dack transfer request acknowledge signal to it, and at the same time outputting an address to the other device involved in the transfer. for example, in the case of transfer between external memory and an external device with dack shown in figure 12.11, when the external device outputs data to the data bus, that data is written to the external memory in the same bus cycle. dmac sh7729r dack dreq external address bus external data bus external memory external device with dack data flow figure 12.11 data flow in single address mode two kinds of transfer are possible in single address mode: (1) transfer between an external device with dack and a memory-mapped external device, and (2) transfer between an external device with dack and external memory. in both cases, only the external request signal ( dreq ) is used for transfer requests. figures 12.12 and 12.13 show examples of dma transfer timing in single address mode. 402 address output to external memory space data output from external device with dack dack signal (active-low) to external device with dack write strobe signal to external memory space address output to external memory space data output from external memory space dack signal (active-low) to external device with dack read strobe signal to external memory space (a) external device with dack external memory space (ordinary memory) (b) external memory space external device with dack (active-low) ckio a25 to a0 d31 to d0 dackn csn we bs ckio a25 to a0 d31 to d0 dackn csn rd bs figure 12.12 example of dma transfer timing in single address mode 403 ckio a25 to a0 d31 to d0 rd wen dackn csn transfer source address +4 +8 +12 figure 12.13 example of dma transfer timing in single address mode (external memory space (ordinary memory) external device with dack) bus modes: there are two bus modes: cycle-steal and burst. select the mode in the tm bits of chcr0Cchcr3. ? cycle-steal mode in cycle-steal mode, the bus is given to another bus master after a one-transfer-unit (8-, 16-, or 32-bit unit) dma transfer. when another transfer request occurs, the bus is obtained from the other bus master and transfer is performed for one transfer unit. when that transfer ends, the bus is passed to the other bus master. this is repeated until the transfer end conditions are satisfied. in the cycle-steal mode, transfer areas are not affected regardless of the transfer request source, transfer source, and transfer destination settings. figure 12.14 shows an example of dma transfer timing in cycle-steal mode. transfer conditions shown in the figure are: dual address mode dreq level detection 404 cpu cpu cpu dmac dmac cpu dmac dmac cpu cpu dreq bus cycle bus returned to cpu readwrite write read figure 12.14 example of transfer in cycle-steal mode ? burst mode once the bus is obtained, the transfer is performed continuously until the transfer end condition is satisfied. in external request mode with low level detection of the dreq pin, however, when the dreq pin is driven high, the bus passes to the other bus master after the dmac transfer request that has already been accepted ends, even if the transfer end conditions have not been satisfied. burst mode cannot be used when a serial communication interface (irda or sci) is the transfer request source. figure 12.15 shows an example of burst mode timing. cpu cpu cpu dmac dmac dmac dmac dmac dmac cpu dreq bus cycle readread read write write write figure 12.15 example of transfer in burst mode 405 relationship between request modes and bus modes by dma transfer category: table 12.6 shows the relationship between request modes and bus modes by dma transfer category. table 12.6 relationship between request modes and bus modes by dma transfer category address mode transfer category request mode bus mode transfer size (bits) usable channels dual external device with dack and external memory external b/c 8/16/32/128 0,1 external device with dack and memory-mapped external device external b/c 8/16/32/128 0, 1 external memory and external memory all* 1 b/c 8/16/32/128 0C3* 5 external memory and memory- mapped external device all * 1 b/c 8/16/32/128 0C3* 5 memory-mapped external device and memory-mapped external device all * 1 b/c 8/16/32/128 0C3* 5 external memory and on-chip peripheral module all * 2 b/c* 3 8/16/32* 4 0C3* 5 memory-mapped external device and on-chip peripheral module all * 2 b/c* 3 8/16/32* 4 0C3* 5 on-chip peripheral module and on- chip peripheral module all * 2 b/c* 3 8/16/32* 4 0C3* 5 x/y memory and x/y memory all b/c 8/16/32/128 0C3 x/y memory and memory-mapped external device all * 1 b/c 8/16/32/128 0C3 x/y memory and on-chip peripheral module all * 2 b/c* 3 8/16/32 0C3 x/y memory and external memory all b/c 8/16/32/128 0C3 single external device with dack and external memory external b/c 8/16/32/128 0, 1 external device with dack and memory-mapped external device external b/c 8/16/32/128 0, 1 b: burst, c: cycle-steal notes: *1 external requests, auto requests and on-chip peripheral module (cmt) requests are all available. *2 external requests, auto requests and on-chip peripheral module requests are all available. when the irda, sci, or a/d converter is also the transfer request source, however, the transfer destination or transfer source must be the irda, sci, or a/d converter, respectively. 406 3. if the transfer request source is the irda, sci, or a/d converter only cycle-steal mode is available. 4. the access size permitted when the transfer destination or source is an on-chip peripheral module register. 5. if the transfer request is an external request, only channels 0 and 1 are available. 6. if the transfer request source is dram or sdram, the transfer size should be set smaller than the bus width. bus mode and channel priority order: when, for example, channel 1 is transferring in burst mode and there is a transfer request to channel 0, which has higher priority, the channel 0 transfer will begin immediately. at this time, if the priority is set in the fixed mode (ch0 > ch1), the channel 1 transfer will continue when the channel 0 transfer has completely finished, even if channel 0 is operating in cycle-steal mode or burst mode. if the priority is set in round-robin mode, channel 1 will begin operating again after channel 0 completes the transfer of one transfer unit, even if channel 0 is in cycle-steal mode or burst mode. the bus will then switch between the two in the order channel 1, channel 0, channel 1, channel 0. even if the priority is set in fixed mode or in round-robin mode, the bus will not be given to the cpu since channel 1 is in burst mode. this example is illustrated in figure 12.16. cpu dmac ch1 dmac ch1 dmac ch0 dmac ch1 dmac ch0 dmac ch1 dmac ch1 cpu ch0 ch1 ch0 round-robin mode in dmac ch0 and ch1 dmac ch1 burst mode cpu cpu priority: round-robin mode ch0: cycle-steal mode ch1: burst mode dmac ch1 burst mode figure 12.16 bus state when multiple channels are operating 407 12.3.5 number of bus cycle states and dreq pin sampling timing number of bus cycle states: when the dmac is the bus master, the number of bus cycle states is controlled by the bus state controller (bsc) in the same way as when the cpu is the bus master. for details, see section 11, bus state controller (bsc). dreq pin sampling timing: in external request mode, the dreq pin is sampled by clock pulse (ckio) falling edge or low level detection. when dreq input is detected, a dmac bus cycle is generated and dma transfer performed, at the earliest, three states later. the second and subsequent dreq sampling operations are started two cycles after the first sample. operation ? cycle-steal mode in cycle-steal mode, the dreq sampling timing is the same regardless of whether level or edge detection is used. for example, in figure 12.17 (cycle-steal mode, level detection), dmac transfer begins, at the earliest, three cycles after the first sampling is performed. the second sampling is started two cycles after the first. if dreq is not detected at this time, sampling is performed in each subsequent cycle. thus, dreq sampling is performed one step in advance. the third sampling operation is not performed until the idle cycle following the end of the first dma transfer. the above conditions are the same whatever the number of cpu transfer cycles, as shown in figure 12.18. the above conditions are also the same whatever the number of dma transfer cycles, as shown in figure 12.19. dack is output in a read in the example in figure 12.17, and in a write in the example in figure 12.18. in both cases, dack is output for the same duration as csn . figure 12.20 shows an example in which sampling is executed in all subsequent cycles when dreq cannot be detected. ? burst mode, level detection in the case of burst mode with level detection, the dreq sampling timing is the same as in cycle-steal mode. 408 for example, in figure 12.22, dmac transfer begins, at the earliest, three cycles after the first sampling is performed. the second sampling is started two cycles after the first. subsequent sampling operations are performed in the idle cycle following the end of the dma transfer cycle. in burst mode, also, the dack output period is the same as in cycle-steal mode. ? burst mode, edge detection in the case of burst mode with edge detection, dreq sampling is only performed once. for example, in figure 12.23, dmac transfer begins, at the earliest, three cycles after the first sampling is performed. after this, dmac transfer is executed continuously until the number of data transfers set in the dmatcr register have been completed. dreq is not sampled during this time. to restart dma transfer after it has been suspended by an nmi, first clear nmif, then input an edge request again. in burst mode, also, the dack output period is the same as in cycle-steal mode. 409 ckio drak dreq dack bus cycle dmac(r) cpu dmac(w) dmac(r) cpu dmac(w) 1st sampling 2nd sampling 3rd sampling figure 12.17 cycle-steal mode, level input (cpu access: 2 cycles) 410 cpu cpu ckio drak dreq dack dmac(r) dmac(w) dmac(r) 1st sampling 2nd sampling 3rd sampling bus cycle figure 12.18 cycle-steal mode, level input (cpu access: 3 cycles) 411 ckio drak (high output) bus cycle dreq dack (rd output) dmac(w) cpu dmac(w) dmac(r) cpu 1st sampling 2nd sampling 3rd sampling figure 12.19 cycle-steal mode, level input (cpu access: 2 cycles, dma rd access: 4 cycles) 412 ckio drak bus cycle dreq dack (rd output) cpu cpu dmac(w) dmac(r) dmac(w) dmac(r) cpu 3rd sampling is performed, but since dreq is high, per-cycle sampling starts 2nd sampling is performed, but since dreq is high, per-cycle sampling starts 1st sampling 2nd sampling 3rd sampling figure 12.20 cycle-steal mode, level input (cpu access: 2 cycles, dreq input delayed) 413 ckio drak bus cycle dreq dack (rd output) cpu cpu dmac(w) dmac(r) dmac(w) dmac(r) cpu high high high high 2nd sampling is performed, but since there is no dreq falling edge, per-cycle sampling starts 3rd sampling is performed, but since there is no dreq falling edge, per-cycle sampling starts 1st sampling 2nd sampling 3rd sampling note: when a dreq falling edge is detected, dreq must be high for at least one cycle before the sampling point. figure 12.21 cycle-steal mode, edge input (cpu access: 2 cycles) 414 ckio drak dreq dack bus cycle dmac(r) dmac(w) dmac(r) dmac(w) dmac(r) cpu 1st sampling 2nd sampling 3rd sampling figure 12.22 burst mode, level input 415 ckio drak dreq dack bus cycle cpu dmac(r) dmac(w) dmac(r) dmac(w) dmac(r) 1st sampling figure 12.23 burst mode, edge input 416 12.3.6 source address reload function channel 2 includes a reload function, in which the value is returned to the value set in the source address register (sar2) every four transfers by setting the ro bit in chcr2. 16-byte transfer cannot be used. figure 12.24 shows this operation. figure 12.25 shows a timing chart for the source address reload function under the following conditions: burst mode, auto-request, 16-bit transfer data size, sar2 incremented, dar2 fixed, reload function on, and use of channel 2 only. sar2 (initial value) dmac transfer request dmac control reload control 4 time count chcr2 dmatcr2 sar2 ro bit = 1 count signal reload signal reload signal address bus figure 12.24 source address reload function diagram 417 ck internal address bus internal data bus sar2 dar2 dar2 dar2 dar2 sar2+2 sar2+4 sar2+6 sar2 sar2 data sar2+2 data sar2+4 data sar2+6 data first transfer on channel 2 second transfer third transfer fourth transfer fifth transfer sar2 output dar2 output sar2+2 output dar2 output sar2+4 output dar2 output sar2+6 output dar2 output sar2 reload sar2 output dar2 output figure 12.25 timing chart of source address reload function the reload function can be executed with a transfer data size of 8, 16, or 32 bits. dmatcr2, which specifies the transfer count, increments 1 each time a transfer ends regardless of whether the reload function is on or off. consequently, a multiple of four must be specified in dmatcr2 when the reload function is on. operation is not guaranteed if other values are specified. the counter that counts the execution of four transfers for the reload function is reset by clearing the dme bit in dmaor or the de bit in chcr2, by setting the transfer end flag (te bit in chcr2), and by nmi input, as well as by a reset or standby transition, but the sar2, dar2, and dmatcr2 registers are not reset. therefore, if these sources are generated, there will be a mix of an initialized counter and uninitialized registers in the dmac, and a malfunction will be caused by restarting the dmac in that state. consequently, if one of these sources other than setting of the te bit occurs during use of the reload function, set sar2, dar2, and dmatcr2 again. 418 12.3.7 dma transfer ending conditions the dma transfer ending conditions are different for ending on an individual channel and ending on all channels together. at the end of transfer, the following conditions are applied except in the case where the value set in the dma transfer count register (dmatcr) reaches 0. (a) cycle-steal mode (external request, internal request, and auto-request) when the transfer ending conditions are satisfied, dmac transfer request acceptance is suspended. the dmac stops operating after completing the number of transfers that it has accepted until the ending conditions are satisfied. in cycle-steal mode, the operation is the same regardless of whether the transfer request is detected by level or edge. (b) burst mode, edge detection (external request, internal request, and auto-request) the timing from the point where the ending conditions are satisfied to the point where the dmac stops operating is the same as in cycle-steal mode. with edge detection in burst mode, though only one transfer request is generated to start the dmac, stop request sampling is performed at the same timing as transfer request sampling in cycle-steal mode. as a result, the period when a stop request is not sampled is regarded as the period when a transfer request is generated, and after performing the dma transfer for this period, the dmac stops operating. (c) burst mode, level detection (external request) same as in (a). (d) bus timing when transfer is suspended transfer is suspended when one transfer ends. even if transfer ending conditions are satisfied during a read in direct address transfer in dual address mode, the subsequent write process is executed, and after the transfer in (a) to (c) above has been executed, dmac operation is suspended. 419 individual channel ending conditions: there are two ending conditions. a transfer ends when the value of the channels dma transfer count register (dmatcr) is 0, or when the de bit in the channels chcr register is cleared to 0. ? when dmatcr is 0: when the dmatcr value becomes 0 and the corresponding channel's dma transfer ends, the transfer end flag bit (te) is set in chcr. if the ie (interrupt enable) bit has been set, a dmac interrupt (dei) request is sent to the cpu. this transfer ending does not apply to (a) through (d) described above. ? when de in chcr is 0: software can halt a dma transfer by clearing the de bit in the channels chcr register. the te bit is not set when this happens. this transfer ending applies to (a) through (d) described above. conditions for ending on all channels simultaneously: transfers on all channels end when 1) the nmif (nmi flag) bit is set to 1 in dmaor, or 2) when the dme bit in dmaor is cleared to 0. ? transfer ending when the nmif bit is set to 1 in dmaor: when an nmi interrupt occurs, the nmif bit is set to 1 in dmaor and all channels stop their transfers according to the conditions in (a) to (d) described above, and pass the bus to an other bus master. consequently, even if the nmi bit is set to 1 during transfer, sar, dar, dmatcr are updated. the te bit is not set. to resume transfer after nmi interrupt exception handling, clear the nmif bit to 0. at this time, if there are channels that should not be restarted, clear the corresponding de bit in chcr. ? transfer ending when dme is cleared to 0 in dmaor: clearing the dme bit to 0 in dmaor forcibly aborts transfer on all channels. the te bit is not set. all channels abort their transfer according to the conditions in (a) to (d) in 12.3.7, dma transfer ending conditions, as in nmi interrupt generation. in this case, the values in sar, dar, and dmatcr are also updated. 420 12.4 compare match timer (cmt) 12.4.1 overview the dmac has an on-chip compare match timer (cmt) to generate dma transfer requests. the cmt has a 16-bit counter. features the cmt has the following features: ? four types of counter input clock can be selected ? one of four internal clocks (p /4, p /8, p /16, p /64) can be selected. ? generates a dma transfer request when compare match occurs. block diagram figure 12.26 shows a block diagram of the cmt. internal bus bus interface control circuit clock selection cmstr cmcsr0 cmcor0 comparator cmcnt0 module bus cmt p /4 p /8 p /16 p /64 cmstr: cmcsr0: cmcor0: cmcnt0: compare match timer start register compare match timer control/status register 0 compare match timer constant register 0 com p are match timer counter 0 figure 12.26 block diagram of cmt 421 register configuration table 12.7 summarizes the cmt register configuration. table 12.7 register configuration name abbreviation r/w initial value address access size (bits) compare match timer start register cmstr r/(w) h'0000 h'04000070 (h'a4000070)* 2 8, 16, 32 compare match timer control/status register 0 cmcsr0 r/(w)* 1 h'0000 h'04000072 (h'a4000072)* 2 8, 16, 32 compare match counter 0 cmcnt0 r/w h'0000 h'04000074 (h'a4000074)* 2 8, 16, 32 compare match constant register 0 cmcor0 r/w h'ffff h'04000076 (h'a4000076)* 2 8, 16, 32 notes: *1 the only value that can be written to the cmf bit in cmcsr0 is 0 to clear the flag. *2 when address translation by the mmu does not apply, the address in parentheses should be used. 12.4.2 register descriptions compare match timer start register (cmstr) the compare match timer start register (cmstr) is a 16-bit register that selects whether compare match counter 0 (cmcnt0) is operated or halted. it is initialized to h'0000 by a reset, but retains its previous value in standby mode. bit: 15 14 13 12 11 10 9 8 initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r r bit: 7 6 5 4 3 2 1 0 str0 initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r/w r/w bits 15 to 2reserved: these bits are always read as 0. the write value should alway be 0. bit 1reserved: this bit can be read or written. the write value should always be 0. 422 bit 0count start 0 (str0): selects whether to operate or halt compare match timer counter 0. bit 0: str0 description 0 cmcnt0 count operation halted (initial value) 1 cmcnt0 count operation compare match timer control/status register 0 (cmcsr0) the compare match timer control/status register 0 (cmcsr0) is a 16-bit register that indicates the occurrence of compare matches, sets the enable/disable status of interrupts, and establishes the clock used for incrementation. it is initialized to h'0000 by a reset, but retains its previous value in standby mode. bit: 15 14 13 12 11 10 9 8 initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r r bit: 7 6 5 4 3 2 1 0 cmf cks1 cks0 initial value: 0 0 0 0 0 0 0 0 r/w: r/(w)* r/w r r r r r/w r/w note: * the only value that can be written is 0 to clear the flag. bits 15 to 8 and 5 to 2reserved: these bits are always read as 0. the write value should always be 0. bit 7compare match flag (cmf): indicates whether or not the compare match timer counter 0 cmcnt0 and compare match timer constant 0 (cmcor0) values match. bit 7: cmf description 0 cmcnt0 and cmcor0 values do not match (initial value) clearing condition: write 0 to cmf after reading cmf = 1 1 cmcnt0 and cmcor0 values match 423 bit 6reserved: this bit can be read or written. the wite value should always be 0. bits 1 and 0clock select 1 and 0 (cks1, cks0): select the clock input to cmcnt from among the four internal clocks obtained by dividing the system clock (p ). when the str bit in cmstr is set to 1, cmcnt0 begins incrementing on the clock selected by cks1 and cks0. bit 1: cks1 bit 0: cks0 description 00p /4 (initial value) 1p /8 1 0 p /16 1p /64 compare match counter 0 (cmcnt0) compare match counter 0 (cmcnt0) is a 16-bit register used as an up-counter. when an internal clock is selected with the cks1 and cks0 bits in the cmcsr0 register and the str bit in cmstr is set to 1, cmcnt0 begins incrementing on that clock. when the cmcnt0 value matches that of compare match constant register 0 (cmcor0), cmcnt0 is cleared to h'0000 and the cmf flag in cmcsr0 is set to 1. cmcnt0 is initialized to h'0000 by a reset, but retains its previous value in standby mode. bit: 15 14 13 12 11 10 9 8 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 3 2 1 0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w 424 compare match constant register 0 (cmcor0) compare match constant register 0 (cmcor0) is a 16-bit register that sets the cmcnt0 compare match period. cmcor0 is initialized to h'ffff by a reset, but retains its previous value in standby mode. bit: 15 14 13 12 11 10 9 8 initial value: 1 1 1 1 1 1 1 1 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 3 2 1 0 initial value: 1 1 1 1 1 1 1 1 r/w: r/w r/w r/w r/w r/w r/w r/w r/w 12.4.3 operation period count operation when an internal clock is selected with the cks1 and cks0 bits in the cmcsr0 register and the str bit in cmstr is set to 1, cmcnt0 begins incrementing on the selected clock. when the cmcnt counter value matches that of cmcor0, the cmcnt0 counter is cleared to h'0000 and the cmf flag in the cmcsr0 register is set to 1. the cmcnt0 counter begins counting up again from h'0000. figure 12.27 shows the compare match counter operation. counter cleared by cmcor0 compare match cmcnt0 value cmcor0 h'0000 time figure 12.27 counter operation 425 cmcnt0 count timing one of four clocks (p /4, p /8, p /16, p /64) obtained by dividing the p clock can be selected with the cks1 and cks0 bits in cmcsr0. figure 12.28 shows the timing. n+1 ck internal clock cmcnt0 input clock cmcnt0 n-1 n figure 12.28 count timing 12.4.4 compare match compare match flag setting timing the cmf bit in the cmcsr0 register is set to 1 by the compare match signal generated when the cmcor0 register and the cmcnt0 counter match. the compare match signal is generated in the final state of the match (timing at which the cmcnt0 counter matching count value is updated). consequently, after the cmcor0 register and the cmcnt0 counter match, a compare match signal will not be generated until a cmcnt0 counter input clock occurs. figure 12.29 shows the cmf bit setting timing. 426 ck cmcor0 cmcnt0 input clock compare match signal cmf cmi cmcnt0 n n 0 figure 12.29 cmf setting timing compare match flag clearing timing the cmf bit in the cmcsr0 register is cleared by writing 0 to it after reading 1. figure 12.30 shows the timing when the cmf bit is cleared by the cpu. ck cmf cmcsr0 write cycle t 1 t 2 figure 12.30 timing of cmf clearing by the cpu 427 12.5 examples of use 12.5.1 example of dma transfer between on-chip irda and external memory in this example, receive data of the on-chip irda is transferred to external memory using dmac channel 3. table 12.8 shows the transfer conditions and register settings. in addition, it is recommended that the trigger for the number of receive fifo data bytes in irda be set to 1 (rtrg1 = rtrg0 = 0 in scfcr). table 12.8 transfer conditions and register settings for transfer between on-chip sci and external memory transfer conditions register setting transfer source: rdr1 of on-chip irda sar3 h'0400014a transfer destination: external memory dar3 h'00400000 number of transfers: 64 dmatcr3 h'00000040 transfer source address: fixed chcr3 h'00004b05 transfer destination address: incremented transfer request source: irda (rxi1) bus mode: cycle-steal transfer unit: byte interrupt request generated at end of transfer channel priority order: 0 > 2 > 3 > 1 dmaor h'0101 428 12.5.2 example of dma transfer between a/d converter and external memory (address reload on) in this example, dma transfer is performed between the on-chip a/d converter (transfer source) and the external memory (transfer destination) with the address reload function on. table 12.9 shows the transfer conditions and register settings. table 12.9 transfer conditions and register settings for transfer between on-chip a/d converter and external memory transfer conditions register setting transfer source: on-chip a/d converter sar2 h'04000080 transfer destination: external memory dar2 h'00400000 number of transfers: 128 (reloading 32 times) dmatcr2 h'00000080 transfer source address: incremented chcr2 h'00089e35 transfer destination address: decremented transfer request source: a/d converter bus mode: burst transfer unit: longword interrupt request generated at end of transfer channel priority order: 0 > 2 > 3 > 1 dmaor h'0101 when the address reload function is on, the value set in sar returns to the initially set value every four transfers. in this example, when an interrupt request is generated from the a/d converter, byte data is read from the register at address h'04000080 in the a/d converter, and is written to external memory address h'00400000. since longword data has been transferred, the values in sar and dar are h'04000084 and h'003ffffc, respectively. the bus is kept and data transfers are performed successively because this transfer is in burst mode. after four transfers end, fifth and sixth transfers are performed if the address reload function is off, and the value in sar is incremented from h'0400008c to h'04000090, h'04000094 if the address reload function is on, dma transfer stops after the fourth transfer ends, and the bus request signal to the cpu is cleared. at this time, the value stored in sar is not incremented from h'0400008c to h'04000090, but returns to the initially set value, h'04000080. the value in dar continues to be incremented regardless of whether the address reload function is on or off. 429 as a result, the values in the dmac are as shown in table 12.10 when the fourth transfer ends, depending on whether the address reload function is on or off. table 12.10 values in dmac after end of fourth transfer items address reload on address reload off sar h'04000080 h'04000090 dar h'003ffffc h'003ffffc dmatcr h'0000007c h'0000007c bus releasedheld dmac operation stops keeps operating interrupt not generatednot generated transfer request source flag clearing executednot executed notes: 1. an interrupt is generated regardless of whether the address reload function is on or off, if transfers are executed until the value in dmatcr reaches 0 and the ie bit in chcr has been set to 1. 2. the transfer request source flag is cleared regardless of whether the address reload function is on or off, if transfers are executed until the value in dmatcr reaches 0. 3. specify burst mode when using the address reload function. this function may not be correctly executed in cycle-steal mode. 4. set a multiple of four in dmatcr when using the address reload function. this function may not be correctly executed if other values are specified. 12.5.3 example of dma transfer between external memory and scif transmitter (indirect address on) in this example, dma transfer is performed between the external memory specified by indirect address (transfer source) and the scif transmitter (transfer destination) using dmac channel 3. table 12.11 shows the transfer conditions and register settings. in addition, the trigger for the number of transmit fifo data bytes is set to 1 (ttrg1 = ttrg0 = 1 in scfcr). 430 table 12.11 transfer conditions and register settings for transfer between external memory and scif transmitter transfer conditions register setting transfer source: external memory sar3 h'00400000 value stored in address h'00400000 h'00450000 value stored in address h'04500000 h'55 transfer destination: on-chip scif tdr2 dar3 h'04000156 number of transfers: 10 dmatcr3 h'0000000a transfer source address: incremented chcr3 h'00011c01 transfer destination address: fixed transfer request source: scif (txi2) bus mode: cycle-steal transfer unit: byte no interrupt request generated at end of transfer channel priority order: 0 > 1 > 2 > 3 dmaor h'0001 if the indirect address is on, data stored in the address set in sar is not used as transfer source data. in the indirect address, after the value stored in the address set in sar is read, that read value is used as an address again, and the value stored in that address is read and stored in the address set in dar. in the example shown in table 12.11, when an scif transfer request is generated, the dmac reads the value in address h'00400000 set in sar3. since the value h'00450000 is stored in that address, the dmac reads the value h'00450000. next, the dmac uses that read value as an address again, and reads the value h'55 stored in that address. then, the dmac writes the value h'55 to address h'04000156 set in dar3; this completes one indirect address transfer. in the indirect address, when data is read first from the address set in sar3, the data transfer size is always longword regardless of the settings of the ts0 and ts1 bits that specify the transfer data size. however, whether the transfer source address is fixed, incremented, or decremented is specified by the sm0 and sm1 bits. therefore, in this example, though the transfer data size is specified as byte, the value in sar3 is h'00400004 when one transfer ends. write operations are the same as in normal dual address transfer. 431 12.6 usage notes 1. the dma channel control registers (chcr0Cchcr3) can be accessed with any data size. the dma operation register (dmaor) must be accessed by byte (eight bits) or word (16 bits); other registers must be accessed by word (16 bits) or longword (32 bits). 2. before rewriting the rs0Crs3 bits in chcr0Cchcr3, first clear the de bit to 0 (when rewriting chcr with a byte address, be sure to set the de bit to 0 in advance). 3. even if an nmi interrupt is input when the dmac is not operating, the nmif bit in dmaor will be set. 4. before entering standby mode, the dme bit in dmaor must be cleared to 0 and the transfers accepted by the dmac completed. 5. the on-chip peripherals which the dmac can access are the irda, scif, a/d converter, d/a converter, and i/o ports. do not access other peripherals with the dmac. 6. when starting up the dmac, set chcr or dmaor last. normal operation is not guaranteed if settings for another register are made last. 7. even if the maximum number of transfers are performed in the same channel after the dmatcr count reaches 0 and dma transfer ends normally, write 0 to dmatcr. otherwise, normal dma transfer may not be performed. 8. when using the address reload function, specify burst mode as the transfer mode. in cycle-steal mode, normal dma transfer may not be performed. 9. when using the address reload function, set a multiple of four in dmatcr. normal operation is not guaranteed if other values are specified. 10. when detecting an external request at the falling edge, keep the external request pin high when setting the dmac. 11. do not access the space from h'4000062 to h'400006f, which is not used in the dmac. accessing this space may cause malfunctions. 12. the wait signal is ignored in the case of a write to external address space in dual address mode with 16-byte transfer, or transfer from an external device with dack to external address space in single address mode with 16-byte transfer. 432 433 section 13 timer (tmu) 13.1 overview the sh7729r has a three-channel (channel 0 to 2) 32-bit timer unit (tmu). 13.1.1 features the tmu has the following features: ? each channel is provided with an auto-reload 32-bit down counter ? channel 2 is provided with an input capture function ? all channels are provided with 32-bit constant registers and 32-bit down counters that can be read or written to at any time. ? all channels generate interrupt requests when the 32-bit down counter underflows (h'00000000 h'ffffffff). ? allows selection between 6 counter input clocks: external clock (tclk), on-chip rtc output clock (16 khz), p /4, p /16, p /64, p /256. (p is the internal clock for peripheral modules.) see section 10, on-chip oscillation circuits, for more information on the clock pulse generator. ? all channels can operate when the sh7729r is in standby mode: when the rtc output clock is being used as the counter input clock, the sh7729r is still able to count in standby mode. ? synchronized read: tcnt is a sequentially changing 32-bit register. since the peripheral module used has an internal bus width of 16 bits, a time lag can occur between the time when the upper 16 bits and lower 16 bits are read. to correct the discrepancy in the counter read value caused by this time lag, a synchronization circuit is built into the tcnt so that the entire 32-bit data in the tcnt can be read at once. ? the maximum operating frequency of the 32-bit counter is 2 mhz on all channels: operate the sh7729r so that the clock input to the timer counters of each channel (obtained by dividing the external clock and internal clock with the prescaler) does not exceed the maximum operating frequency. 434 13.1.2 block diagram figure 13.1 shows a block diagram of the tmu. tocr prescaler tstr tcr0 tcnt0 module bus internal bus tcor0 tcr1 tcnt1 tcor1 counter controller tclk p rtcclk tuni0 bus interface ch. 0 interrupt controller interrupt controller interrupt controller counter controller counter controller tuni1 tuni2 ticpi2 tcr2 tcpr2 tcnt2 tcor2 tmu ch. 1 ch. 2 clock controller tocr: tstr: tcr: legend timer output control register timer start register tcnt: tcor: tcpr2: 32-bit timer counter 32-bit timer constant register 32-bit input capture register timer control register figure 13.1 block diagram of tmu 435 13.1.3 pin configuration table 13.1 shows the pin configuration of the tmu. table 13.1 tmu pin channel pin i/o description clock input/clock output tclk i/o external clock input pin/input capture control input pin/realtime clock (rtc) output pin 13.1.4 register configuration table 13.2 shows the tmu register configuration. table 13.2 tmu registers channel register abbreviation r/w initial value * address access size common timer output control register tocr r/w h'00 h'fffffe90 8 timer start register tstr r/w h'00 h'fffffe92 8 0 timer constant register 0 tcor0 r/w h'ffffffff h'fffffe94 32 timer counter 0 tcnt0 r/w h'ffffffff h'fffffe98 32 timer control register 0 tcr0 r/w h'0000 h'fffffe9c 16 1 timer constant register 1 tcor1 r/w h'ffffffff h'fffffea0 32 timer counter 1 tcnt1 r/w h'ffffffff h'fffffea4 32 timer control register 1 tcr1 r/w h'0000 h'fffffea8 16 2 timer constant register 2 tcor2 r/w h'ffffffff h'fffffeac 32 timer counter 2 tcnt2 r/w h'ffffffff h'fffffeb0 32 timer control register 2 tcr2 r/w h'0000 h'fffffeb4 16 input capture register 2 tcpr2 r undefined h'fffffeb8 32 note: * initialized by power-on resets or manual resets. 436 13.2 tmu registers 13.2.1 timer output control register (tocr) tocr is an 8-bit readable/writable register that selects whether to use the external tclk pin as an external clock or an input capture control usage input pin, or an output pin for the on-chip rtc output clock. tocr is initialized to h'00 by a power-on reset or manual reset, but is not initialized in standby mode. bit: 7 6 5 4 3 2 1 0 tcoe initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r r/w bits 7 to 1reserved: these bits are always read as 0. the write value should always be 0. bit 0timer clock pin control (tcoe): selects use of the timer clock pin (tclk) as an external clock output pin or input pin for input capture control for the on-chip timer, or as an output pin for the on-chip rtc output clock. since the tclk pin is multiplexed as the pth7 pin, when the pin is used as tclk, bits ph7md1 and ph7md0 in the phcr register should be set to 00 (the "other function" setting). bit 0: tcoe description 0 timer clock pin (tclk) used as external clock input or input capture control input pin for the on-chip timer (initial value) 1 timer clock pin (tclk) used as output pin for on-chip rtc output clock 13.2.2 timer start register (tstr) tstr is an 8-bit readable/writable register that selects whether to run or halt the timer counters (tcnt) for channels 0C2. tstr is initialized to h'00 by a power-on reset or manual reset, but is not initialized in standby mode when the input clock selected for the channel is the on-chip rtc clock (rtcclk). only when an external clock (tclk) or the peripheral clock (p ) is used as the input clock, it is initialized in standby mode when the multiplication ratio of pll circuit 1 is changed or when the mstp2 bit in stbcr is set to 1. bit: 7 6 5 4 3 2 1 0 str2 str1 str0 initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r/w r/w r/w 437 bits 7 to 3reserved: these bits are always read as 0. the write value should always be 0. bit 2counter start 2 (str2): selects whether to run or halt timer counter 2 (tcnt2). bit 2: str2 description 0 tcnt2 count halted (initial value) 1 tcnt2 counts bit 1counter start 1 (str1): selects whether to run or halt timer counter 1 (tcnt1). bit 1: str1 description 0 tcnt1 count halted (initial value) 1 tcnt1 counts bit 0counter start 0 (str0): selects whether to run or halt timer counter 0 (tcnt0). bit 0: str0 description 0 tcnt0 count halted (initial value) 1 tcnt0 counts 13.2.3 timer control registers (tcr) the timer control registers (tcr) control the timer counters (tcnt) and interrupts. the tmu has three tcr registers, one for each channel. the tcr registers are 16-bit readable/writable registers that control the issuance of interrupts when the flag indicating timer counter (tcnt) underflow has been set to 1, and also carry out counter clock selection. when the external clock has been selected, they also select its edge. additionally, tcr2 controls the channel 2 input capture function and the issuance of interrupts during input capture. the tcr registers are initialized to h'0000 by a power-on reset and manual reset, but are not initialized in standby mode. 438 channel 0 and 1 tcr bit configuration: bit: 15 14 13 12 11 10 9 8 unf initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r r/w bit: 7 6 5 4 3 2 1 0 unie ckeg1 ckeg0 tpsc2 tpsc1 tpsc0 initial value: 0 0 0 0 0 0 0 0 r/w: r r r/w r/w r/w r/w r/w r/w channel 2 tcr bit configuration: bit: 15 14 13 12 11 10 9 8 icpf unf initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r/w r/w bit: 7 6 5 4 3 2 1 0 icpe1 icpe0 unie ckeg1 ckeg0 tpsc2 tpsc1 tpsc0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bits 15 to 10, 9 (except tcr2), 7, and 6 (except tcr2)reserved: these bits are always read as 0. the write value should always be 0. bit 9input capture interrupt flag (icpf): a function of channel 2 only: the flag is set when input capture is requested via the tclk pin. bit 9: icpf description 0 no input capture request has been issued clearing condition: when 0 is written to icpf (initial value) 1 input capture has been requested via the tclk pin setting condition: when input capture is requested via the tclk pin* note: * contents do not change when 1 is written to icpf. 439 bit 8underflow flag (unf): status flag that indicates occurrence of a tcnt underflow. bit 8: unf description 0 tcnt has not underflowed clearing condition: when 0 is written to unf (initial value) 1 tcnt has underflowed (h'00000000 h'ffffffff) setting condition: when tcnt underflows* note: * contents do not change when 1 is written to unf. bits 7 and 6input capture control (icpe1, icpe0): a function of channel 2 only: determines whether the input capture function can be used, and when used, whether or not to enable interrupts. when using this input capture function it is necessary to set the tclk pin to input mode with the tcoe bit in the tocr register. additionally, use the ckeg bit to designate use of either the rising or falling edge of the tclk pin to set the value in tcnt2 in the input capture register (tcpr2). bit 7: icpe1 bit 6: icpe0description 0 0 input capture function is not used (initial value) 1 reserved (setting prohibited) 1 0 input capture function is used. interrupt due to icpf (ticpi2) is not enabled 1 input capture function is used. interrupt due to icpf (ticpi2) is enabled bit 5underflow interrupt control (unie): controls enabling of interrupt generation when the status flag (unf) indicating tcnt underflow has been set to 1. bit 5: unie description 0 interrupt due to unf (tuni) is not enabled (initial value) 1 interrupt due to unf (tuni) is enabled 440 bits 4 and 3clock edge 1 and 0 (ckeg1, ckeg0): select the external clock edge when the external clock is selected, or when the input capture function is used. bit 4: ckeg1 bit 3: ckeg0description 0 0 count/capture register set on rising edge (initial value) 1 count/capture register set on falling edge 1 x count/capture register set on both rising and falling edge note: x means 0, 1, or dont care. bits 2 to 0timer prescaler 2 to 0 (tpsc2 to tpsc0): select the tcnt count clock. bit 2: tpsc2 bit 1: tpsc1 bit 0: tpsc0 description 0 0 0 internal clock: count on p /4 (initial value) 1 internal clock: count on p / 16 1 0 internal clock: count on p /64 1 internal clock: count on p /256 1 0 0 internal clock: count on clock output of on-chip rtc (rtcclk) 1 external clock: count on tclk pin input 1 0 reserved (setting prohibited) 1 reserved (setting prohibited) 441 13.2.4 timer constant registers (tcor) the timer constant registers are 32-bit registers. the tmu has three tcor registers, one for each channel. tcor is a 32-bit readable/writable register. when a tcnt count-down results in an underflow, the tcor value is set in tcnt and the count-down continues from that value. tcor is initialized to h'ffffffff by a power-on reset or manual reset, but is not initialized, and retains its contents, in standby mode. bit: 31 30 29 28 27 26 25 24 initial value: 1 1 1 1 1 1 1 1 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 23 22 21 20 19 18 17 16 initial value: 1 1 1 1 1 1 1 1 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 15 14 13 12 11 10 9 8 initial value: 1 1 1 1 1 1 1 1 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 3 2 1 0 initial value: 1 1 1 1 1 1 1 1 r/w: r/w r/w r/w r/w r/w r/w r/w r/w 442 13.2.5 timer counters (tcnt) the timer counters are 32-bit readable/writable registers. the tmu has three timer counters, one for each channel. tcnt counts down upon input of a clock. the clock input is selected using the tpsc2 C tpsc0 bits in the timer control register (tcr). when a tcnt count-down results in an underflow (h'00000000 h'ffffffff), the underflow flag (unf) in the timer control register (tcr) of the relevant channel is set. the tcor value is simultaneously set in tcnt itself and the count-down continues from that value. because the internal bus for the sh7729r on-chip peripheral modules is 16 bits wide, a time lag can occur between the time when the upper 16 bits and lower 16 bits are read. since tcnt counts sequentially, this time lag can create discrepancies between the data in the upper and lower halves. to correct the discrepancy, a buffer register is connected to tcnt so that the upper and lower halves are not read separately. the entire 32-bit data in tcnt can thus be read at once. tcnt is initialized to h'ffffffff by a power-on reset or manual reset, but is not initialized, and retains its contents, in standby mode. bit: 31 30 29 28 27 26 25 24 initial value: 1 1 1 1 1 1 1 1 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 23 22 21 20 19 18 17 16 initial value: 1 1 1 1 1 1 1 1 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 15 14 13 12 11 10 9 8 initial value: 1 1 1 1 1 1 1 1 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit: 7 6 5 4 3 2 1 0 initial value: 1 1 1 1 1 1 1 1 r/w: r/w r/w r/w r/w r/w r/w r/w r/w 443 13.2.6 input capture register (tcpr2) input capture register 2 (tcpr2) is a read-only 32-bit register provided only in timer 2. control of tcpr2 setting conditions due to the tclk pin is affected by the input capture function bits (icpe1/icpe2 and ckeg1/ckeg0) in tcr2. when a tcpr2 setting indication due to the tclk pin occurs, the value of tcnt2 is copied into tcpr2. tcnt2 is not initialized by a power-on reset or manual reset, or in standby mode. bit: 31 30 29 28 27 26 25 24 initial value: r/w: r r r r r r r r bit: 23 22 21 20 19 18 17 16 initial value: r/w: r r r r r r r r bit: 15 14 13 12 11 10 9 8 initial value: r/w: r r r r r r r r bit: 7 6 5 4 3 2 1 0 initial value: r/w: r r r r r r r r 444 13.3 tmu operation 13.3.1 overview each of three channels has a 32-bit timer counter (tcnt) and a 32-bit timer constant register (tcor). tcnt counts down. the auto-reload function enables synchronized counting and counting by external events. channel 2 has an input capture function. 13.3.2 basic functions counter operation: when the str0 C str2 bits in the timer start register (tstr) are set, the corresponding timer counter (tcnt) starts counting. when a tcnt underflows, the unf flag of the corresponding timer control register (tcr) is set. at this time, if the unie bit in tcr is 1, an interrupt request is sent to the cpu. also at this time, the value is copied from tcor to tcnt and the down-count operation is continued. the count operation is set as follows (figure 13.2): 1. select the counter clock with the tpsc2Ctpsc0 bits in the timer control register (tcr). if the external clock is selected, set the tclk pin to input mode with the toce bit in tocr, and select its edge with the ckeg1 and ckeg0 bits in tcr. 2. use the unie bit in tcr to set whether to generate an interrupt when tcnt underflows. 3. when using the input capture function, set the icpe bits in tcr, including the choice of whether or not to use the interrupt function (channel 2 only). 4. set a value in the timer constant register (tcor) (the cycle is the set value plus 1). 5. set the initial value in the timer counter (tcnt). 6. set the str bit in the timer start register (tstr) to 1 to start operation. 445 select operation select counter clock set underflow interrupt generation set timer constant register initialize timer counter start counting (1) (2) (4) (5) (6) set interrupt generation when using input capture function (3) note: when an interrupt has been generated, clear the flag in the interrupt handler that caused it. if interrupts are enabled without clearing the flag, another interrupt will be generated. figure 13.2 setting the count operation 446 auto-reload count operation: figure 13.3 shows the tcnt auto-reload operation. tcnt value tcor h'00000000 str0Cstr2 unf tcor value set to tcnt during underflow time figure 13.3 auto-reload count operation tcnt count timing: ? internal clock operation: set the tpsc2Ctpsc0 bits in tcr to select whether peripheral module clock p or one of the four internal clocks created by dividing it is used (p /4, p /16, p /64, p /256). figure 13.4 shows the timing. p internal clock tcnt input clock tcnt n + 1 n n C 1 figure 13.4 count timing when operating on internal clock 447 ? external clock operation: set the tpsc2Ctpsc0 bits in tcr to select the external clock (tclk) as the timer clock. use the ckeg1 and ckeg0 bits in tcr to select the detection edge. rising, falling, or both edges may be selected. the pulse width of the external clock must be at least 1.5 peripheral module clock cycles for single edges or 2.5 peripheral module clock cycles for both edges. a shorter pulse width will result in accurate operation. figure 13.5 shows the timing for both-edge detection. p external clock input pin tcnt input clock tcnt n + 1 n n C 1 figure 13.5 count timing when operating on external clock (both edges detected) ? on-chip rtc clock operation: set the tpsc2Ctpsc0 bits in tcr to select the on-chip rtc clock as the timer clock. figure 13.6 shows the timing. rtc output clock tcnt tcnt input clock n + 1 n n C 1 figure 13.6 count timing when operating on on-chip rtc clock input capture function: channel 2 has an input capture function (figure 13.7). when using the input capture function, set the tclk pin to input mode with the tcoe bit in the timer output control register (tocr) and set the timer operation clock to internal clock or on-chip rtc clock with the tpcs2Ctpcs0 bits in the timer control register (tcr2). also, designate use of the input capture function and whether to generate interrupts on input capture with the ipce1Cipce0 bits in tcr2, and designate the use of either the rising or falling edge of the tclk pin to set the timer counter (tcnt2) value into the input capture register (tcpr2) with the ckeg1Cckeg0 bits in tcr2. the input capture function cannot be used in standby mode. 448 tcnt value tcor h'00000000 tclk tcpr2 set tcnt value icpi tcor value set to tcnt during underflow time figure 13.7 operation timing when using input capture function (using tclk rising edge) 13.4 interrupts there are two sources of tmu interrupts: underflow interrupts (tuni) and interrupts when using the input capture function (ticpi2). 13.4.1 status flag setting timing unf is set to 1 when the tcnt underflows. figure 13.8 shows the timing. p tcnt underflow signal unf tuni tcor value h'00000000 figure 13.8 unf setting timing 449 13.4.2 status flag clearing timing the status flag can be cleared by writing 0 from the cpu. figure 13.9 shows the timing. p peripheral address bus unf tcr address t1 t2 tcr write cycle t3 figure 13.9 status flag clearing timing 13.4.3 interrupt sources and priorities the tmu produces underflow interrupts for each channel. when the interrupt request flag and interrupt enable bit are both set to 1, an interrupt is requested. codes are set in the interrupt event registers (intevt, intevt2) for these interrupts and interrupt handling occurs according to the codes. the relative priorities of channels can be changed using the interrupt controller (see section 4, exception handling, and section 7, interrupt controller (intc)). table 13.3 lists tmu interrupt sources. table 13.3 tmu interrupt sources channel interrupt source description priority 0 tuni0 underflow interrupt 0 high 1 tuni1 underflow interrupt 1 2 tuni2 underflow interrupt 2 2 ticpi2 input capture interrupt 2 low 450 13.5 usage notes 13.5.1 writing to registers synchronization processing is not performed for timer counting during register writes. when writing to registers, always clear the appropriate start bits for the channel (str2Cstr0) in the timer start register (tstr) to halt timer counting. 13.5.2 reading registers synchronization processing is performed for timer counting during register reads. when timer counting and register read processing are performed simultaneously, the register value before tcnt counting down (with synchronization processing) is read. 451 section 14 realtime clock (rtc) 14.1 overview the sh7729r has a realtime clock (rtc) with its own 32.768-khz crystal oscillator. 14.1.1 features ? clock and calendar functions (bcd display): seconds, minutes, hours, date, day of the week, month, and year ? 1-hz to 64-hz timer (binary display) ? start/stop function ? 30-second adjust function ? alarm interrupt: frame comparison of seconds, minutes, hours, date, day of the week, and month can be used as conditions for the alarm interrupt ? cyclic interrupts: the interrupt cycle may be 1/256 second, 1/64 second, 1/16 second, 1/4 second, 1/2 second, 1 second, or 2 seconds ? carry interrupt: a carry interrupt indicates when a carry occurs during a counter read ? automatic leap year correction 452 14.1.2 block diagram figure 14.1 shows a block diagram of the rtc. module bus rtc internal bus interrupt control circuit prescaler ( 2) rtcclk bus interface carry detection circuit ati pri cui legend r64cnt: 64 hz counter rseccnt: second counter rmincnt: minute counter rhrcnt: hour counter rwkcnt: day-of-week counter rdaycnt: day counter rmoncnt: month counter ryrcnt: year counter rsecar: second alarm register rminar: minute alarm register rhrar: hour alarm register rwkar: day-of-week alarm register rdayar: day alarm register rmonar: month alarm register rcr1: rtc control register 1 rcr2: rtc control register 2 r64cnt reset rseccnt rmincnt rhrcnt rwkcnt 16.384 khz rdaycnt rmoncnt ryrcnt comparator rsecar rminar rhrar rwkar rdayar rcr1 rcr2 30- second adj extal2 32.768 khz 128 hz xtal2 externally connected circuit oscillator circuit prescaler ( 128) rmonar figure 14.1 block diagram of rtc 453 14.1.3 pin configuration table 14.1 shows the rtc pin configuration. table 14.1 rtc pins pin abbreviation i/o description rtc oscillator crystal pin extal2 i connects crystal to rtc oscillator* 2 rtc oscillator crystal pin xtal2 o connects crystal to rtc oscillator* 2 clock input/clock output tclk i/o external clock input pin/input capture control input pin/realtime clock (rtc) output pin (shared by tmu) dedicated power-supply pin for rtc vccCrtc dedicated power-supply pin for rtc* 1 dedicated gnd pin for rtc vssCrtc dedicated gnd pin for rtc* 1 notes: *1 except in hardware standby mode, power must be supplied to all power supply pins, including these, even when only the rtc is used (including standby mode). *2 when the rtc is not used, pull extal2 up (to vcc) and make no connection for xtal2. 454 14.1.4 rtc register configuration table 14.2 shows the rtc register configuration. table 14.2 rtc registers name abbreviation r/w initial value address access size 64-hz counter r64cnt r undefined h'fffffec0 8 second counter rseccnt r/w undefined h'fffffec2 8 minute counter rmincnt r/w undefined h'fffffec4 8 hour counter rhrcnt r/w undefined h'fffffec6 8 day of week counter rwkcnt r/w undefined h'fffffec8 8 date counter rdaycnt r/w undefined h'fffffeca 8 month counter rmoncnt r/w undefined h'fffffecc 8 year counter ryrcnt r/w undefined h'fffffece 8 second alarm register rsecar r/w undefined* h'fffffed0 8 minute alarm register rminar r/w undefined* h'fffffed2 8 hour alarm register rhrar r/w undefined* h'fffffed4 8 day of week alarm register rwkar r/w undefined* h'fffffed6 8 date alarm register rdayar r/w undefined* h'fffffed8 8 month alarm register rmonar r/w undefined* h'fffffeda 8 rtc control register 1 rcr1 r/w h'00 h'fffffedc 8 rtc control register 2 rcr2 r/w h'09 h'fffffede 8 note: * only the enb bits of each register are initialized. 455 14.2 rtc registers 14.2.1 64-hz counter (r64cnt) the 64-hz counter (r64cnt) is an 8-bit read-only register that indicates the state of the rtc divider circuit between 64 hz and 1 hz. r64cnt is reset to h'00 by setting the reset bit in rtc control register 2 (rcr2) or the adj bit in rcr2 to 1. r64cnt is not initialized by a power-on reset or manual reset, or in standby mode. bit 7 is always read as 0. bit: 7 6 5 4 3 2 1 0 1hz 2hz 4hz 8hz 16hz 32hz 64hz initial value: 0 r/w: r r r r r r r r 14.2.2 second counter (rseccnt) the second counter (rseccnt) is an 8-bit readable/writable register used for setting/counting in the bcd-coded second section of the rtc. the count operation is performed by a carry for each second of the 64 - hz counter. the range that can be set is 00 C 59 (decimal). errant operation will result if any other value is set. carry out write processing after halting the count operation with the start bit in rcr2. rseccnt is not initialized by a power-on reset or manual reset, or in standby mode. bit: 7 6 5 4 3 2 1 0 10 seconds 1 second initial value: 0 r/w: r r/w r/w r/w r/w r/w r/w r/w 456 14.2.3 minute counter (rmincnt) the minute counter (rmincnt) is an 8-bit readable/writable register used for setting/counting in the bcd-coded minute section of the rtc. the count operation is performed by a carry for each minute of the second counter. the range that can be set is 00C59 (decimal). errant operation will result if any other value is set. carry out write processing after halting the count operation with the start bit in rcr2. rmincnt is not initialized by a power-on reset or manual reset, or in standby mode. bit: 7 6 5 4 3 2 1 0 10 minutes 1 minute initial value: 0 r/w: r r/w r/w r/w r/w r/w r/w r/w 14.2.4 hour counter (rhrcnt) the hour counter (rhrcnt) is an 8-bit readable/writable register used for setting/counting in the bcd-coded hour section of the rtc. the count operation is performed by a carry for each 1 hour of the minute counter. the range that can be set is 00 C 23 (decimal). errant operation will result if any other value is set. carry out write processing after halting the count operation with the start bit in rcr2 or using a carry flag as shown in figure 14.2. rhrcnt is not initialized by a power-on reset or manual reset, or in standby mode. bit: 7 6 5 4 3 2 1 0 10 hours 1 hour initial value: 0 0 r/w: r r r/w r/w r/w r/w r/w r/w 457 14.2.5 day of week counter (rwkcnt) the day of week counter (rwkcnt) is an 8-bit readable/writable register used for setting/counting in the bcd-coded day of week section of the rtc. the count operation is performed by a carry for each day of the date counter. the range that can be set is 0 C 6 (decimal). errant operation will result if any other value is set. carry out write processing after halting the count operation with the start bit in rcr2. rwkcnt is not initialized by a power-on reset or manual reset, or in standby mode. bit: 7 6 5 4 3 2 1 0 day of week initial value: 0 0 0 0 0 r/w: r r r r r r/w r/w r/w days of the week are coded as shown in table 14.3. table 14.3 day-of-week codes (rwkcnt) day of weekcode sunday0 monday1 tuesday2 wednesday3 thursday4 friday5 saturday6 458 14.2.6 date counter (rdaycnt) the date counter (rdaycnt) is an 8-bit readable/writable register used for setting/counting in the bcd-coded date section of the rtc. the count operation is performed by a carry for each day of the hour counter. the range that can be set is 01 C 31 (decimal). errant operation will result if any other value is set. carry out write processing after halting the count operation with the start bit in rcr2. rdaycnt is not initialized by a power-on reset or manual reset, or in standby mode. the rdaycnt range that can be set changes with each month and in leap years. please confirm the correct setting. bit: 7 6 5 4 3 2 1 0 10 days 1 day initial value: 0 0 r/w: r r r/w r/w r/w r/w r/w r/w 14.2.7 month counter (rmoncnt) the month counter (rmoncnt) is an 8-bit readable/writable register used for setting/counting in the bcd-coded month section of the rtc. the count operation is performed by a carry for each month of the date counter. the range that can be set is 00 C 12 (decimal). errant operation will result if any other value is set. carry out write processing after halting the count operation with the start bit in rcr2. rmoncnt is not initialized by a power-on reset or manual reset, or in standby mode. bit: 7 6 5 4 3 2 1 0 10 months 1 month initial value: 0 0 0 r/w: r r r r/w r/w r/w r/w r/w 459 14.2.8 year counter (ryrcnt) the year counter (ryrcnt) is an 8-bit readable/writable register used for setting/counting in the bcd-coded year section of the rtc. the least significant 2 digits of the western calendar year are displayed. the count operation is performed by a carry for each year of the month counter. the range that can be set is 00 C 99 (decimal). errant operation will result if any other value is set. carry out write processing after halting the count operation with the start bit in rcr2. ryrcnt is not initialized by a power-on reset or manual reset, or in standby mode. leap years are recognized by dividing the year counter value by 4 and obtaining a fractional result of 0. bit: 7 6 5 4 3 2 1 0 10 years 1 year initial value: r/w: r/w r/w r/w r/w r/w r/w r/w r/w 14.2.9 second alarm register (rsecar) the second alarm register (rsecar) is an 8-bit readable/writable register, and an alarm register corresponding to the bcd-coded second section counter rseccnt of the rtc. when the enb bit is set to 1, a comparison with the rseccnt value is performed. from among the rsecar/rminar/rhrar/rwkar/rdayar/rmonar registers, the counter and alarm register comparison is performed only on those with enb bits set to 1, and if each of those coincide, an rtc alarm interrupt is generated. the range that can be set is 00 C 59 (decimal) + enb bit. errant operation will result if any other value is set. the enb bit in rsecar is initialized to 0 by a power-on reset. the remaining rsecar fields are not initialized by a power-on reset or manual reset, or in standby mode. bit: 7 6 5 4 3 2 1 0 enb 10 seconds 1 second initial value: 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w 460 14.2.10 minute alarm register (rminar) the minute alarm register (rminar) is an 8-bit readable/writable register, and an alarm register corresponding to the bcd-coded minute section counter rmincnt of the rtc. when the enb bit is set to 1, a comparison with the rmincnt value is performed. from among the rsecar/rminar/rhrar/rwkar/rdayar/rmonar registers, the counter and alarm register comparison is performed only on those with enb bits set to 1, and if each of those coincide, an rtc alarm interrupt is generated. the range that can be set is 00 C 59 (decimal) + enb bit. errant operation will result if any other value is set. the enb bit in rminar is initialized by a power-on reset. the remaining rminar fields are not initialized by a power-on reset or manual reset, or in standby mode. bit: 7 6 5 4 3 2 1 0 enb 10 minutes 1 minute initial value: 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w 14.2.11 hour alarm register (rhrar) the hour alarm register (rhrar) is an 8-bit readable/writable register, and an alarm register corresponding to the bcd-coded hour section counter rhrcnt of the rtc. when the enb bit is set to 1, a comparison with the rhrcnt value is performed. from among the rsecar/rminar/rhrar/rwkar/rdayar/rmonar registers, the counter and alarm register comparison is performed only on those with enb bits set to 1, and if each of those coincide, an rtc alarm interrupt is generated. the range that can be set is 00 C 23 (decimal) + enb bit. errant operation will result if any other value is set. the enb bit in rhrar is initialized by a power-on reset. the remaining rhrar fields are not initialized by a power-on reset or manual reset, or in standby mode. bit: 7 6 5 4 3 2 1 0 enb 10 hours 1 hour initial value: 0 0 r/w: r/w r r/w r/w r/w r/w r/w r/w 461 14.2.12 day of week alarm register (rwkar) the day of week alarm register (rwkar) is an 8-bit readable/writable register, and an alarm register corresponding to the bcd-coded day of week section counter rwkcnt of the rtc. when the enb bit is set to 1, a comparison with the rwkcnt value is performed. from among the rsecar/rminar/rhrar/rwkar/rdayar/rmonar registers, the counter and alarm register comparison is performed only on those with enb bits set to 1, and if each of those coincide, an rtc alarm interrupt is generated. the range that can be set is 0 C 6 (decimal) + enb bit. errant operation will result if any other value is set. the enb bit in rwkar is initialized by a power-on reset. the remaining rwkar fields are not initialized by a power-on reset or manual reset, or in standby mode. bit: 7 6 5 4 3 2 1 0 enb day of week initial value: 0 0 0 0 0 r/w: r/w r r r r r/w r/w r/w days of the week are coded as shown in table 14.4. table 14.4 day-of-week codes (rwkar) day of weekcode sunday0 monday1 tuesday2 wednesday3 thursday4 friday5 saturday6 462 14.2.13 date alarm register (rdayar) the date alarm register (rdayar) is an 8-bit readable/writable register, and an alarm register corresponding to the bcd-coded date section counter rdaycnt of the rtc. when the enb bit is set to 1, a comparison with the rdaycnt value is performed. from among the registers rsecar, rminar, rhrar, rwkar, rdayar, rmonar, the counter and alarm register comparison is performed only on those with enb bits set to 1, and if each of those coincide, an rtc alarm interrupt is generated. the range that can be set is 01 C 31 (decimal) + enb bit. errant operation will result if any other value is set. the rdaycnt range that can be set changes with some months and in leap years. please confirm the correct setting. the enb bit in rdayar is initialized by a power-on reset. the remaining rdayar fields are not initialized by a power-on reset or manual reset, or in standby mode. bit: 7 6 5 4 3 2 1 0 enb 10 days 1 day initial value: 0 0 r/w: r/w r r/w r/w r/w r/w r/w r/w 14.2.14 month alarm register (rmonar) the month alarm register (rmonar) is an 8-bit readable/writable register, and an alarm register corresponding to the bcd-coded month section counter rmoncnt of the rtc. when the enb bit is set to 1, a comparison with the rmoncnt value is performed. from among the registers rsecar, rminar, rhrar, rwkar, rdayar, rmonar, the counter and alarm register comparison is performed only on those with enb bits set to 1, and if each of those coincide, an rtc alarm interrupt is generated. the range that can be set is 01 C 12 (decimal) + enb bit. errant operation will result if any other value is set. the enb bit in rmonar is initialized by a power-on reset. the remaining rmonar fields are not initialized by a power-on reset or manual reset, or in standby mode. bit: 7 6 5 4 3 2 1 0 enb 10 months 1 month initial value: 0 0 0 r/w: r/w r r r/w r/w r/w r/w r/w 463 14.2.15 rtc control register 1 (rcr1) the rtc control register 1 (rcr1) is an 8-bit readable/writable register that affects carry flags and alarm flags. it also selects whether to generate interrupts for each flag. because flags are sometimes set after an operand read, do not use this register in read-modify-write processing. rcr1 is initialized to h'00 by a power-on reset. in a manual reset, all bits are initialized to 0 except for the cf flag, which is undefined. when using the cf flag, it must be initialized beforehand. this register is not initialized in standby mode. bit: 7 6 5 4 3 2 1 0 cf cie aie af initial value: 0 0 0 0 0 0 0 0 r/w: r/w r r r/w r/w r r r/w bit 7 carry flag (cf): status flag that indicates that a carry has occurred. cf is set to 1 when a count-up to r64cnt or rseccnt occurs. a count register value read at this time cannot be guaranteed; another read is required. bit 7: cf description 0 no count up of r64cnt or rseccnt clearing condition: when 0 is written to cf (initial value) 1 count up of r64cnt or rseccnt setting condition: when 1 is written to cf bits 6, 5, 2, and 1 reserved: these bits are always read as 0. the write value should always be 0. bit 4 carry interrupt enable flag (cie): when the carry flag (cf) is set to 1, the cie bit enables interrupts. bit 4: cie description 0 a carry interrupt is not generated when the cf flag is set to 1 (initial value) 1 a carry interrupt is generated when the cf flag is set to 1 bit 3 alarm interrupt enable flag (aie): when the alarm flag (af) is set to 1, the aie bit allows interrupts. 464 bit 3: aie description 0 an alarm interrupt is not generated when the af flag is set to 1 (initial value) 1 an alarm interrupt is generated when the af flag is set to 1 bit 0 alarm flag (af): the af flag is set to 1 when the alarm time set in an alarm register (only registers with enb bit set to 1) matches the clock and calendar time. this flag is cleared to 0 when 0 is written, but holds its previous value when 1 is written. bit 0: af description 0 clock/calendar and alarm register have not matched since last reset to 0 clearing condition: when 0 is written to af (initial value) 1 setting condition: clock/calendar and alarm register have matched (only registers with enb set)* note: * contents do not change when 1 is written to af. 14.2.16 rtc control register 2 (rcr2) the rtc control register 2 (rcr2) is an 8-bit readable/writable register for periodic interrupt control, 30-second adjustment adj, divider circuit reset, and rtc count start/stop control. it is initialized to h'09 by a power-on reset. it is initialized except for rtcen and start by a manual reset. it is not initialized, and retains its contents, in standby mode. bit: 7 6 5 4 3 2 1 0 pef pes2 pes1 pes0 rtcen adj reset start initial value: 0 0 0 0 1 0 0 1 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit 7 periodic interrupt flag (pef): indicates interrupt generation with the period designated by the pes bits. when set to 1, pef generates periodic interrupts. bit 7: pef description 0 interrupts not generated with the period designated by the pes bits clearing condition: when 0 is written to pef (initial value) 1 interrupts generated with the period designated by the pes bits setting condition: when 1 is written to pef 465 bits 6 to 4periodic interrupt flags (pes2-pes0): specify the periodic interrupt. bit 6: pes2 bit 5: pes1 bit 4: pes0 description 0 0 0 no periodic interrupts generated (initial value) 1 periodic interrupt generated every 1/256 second 1 0 periodic interrupt generated every 1/64 second 1 periodic interrupt generated every 1/16 second 1 0 0 periodic interrupt generated every 1/4 second 1 periodic interrupt generated every 1/2 second 1 0 periodic interrupt generated every 1 second 1 periodic interrupt generated every 2 seconds bit 3 rtcen: controls the operation of the crystal oscillator for the rtc. bit 3: rtcen description 0 crystal oscillator for rtc is halted 1 crystal oscillator for rtc runs (initial value) bit 2 30 second adjustment (adj): when 1 is written to the adj bit, times of 29 seconds or less will be rounded to 00 seconds and 30 seconds or more to 1 minute. the divider circuit will be simultaneously reset. this bit is always read as 0. bit 2: adj description 0 runs normally(initial value) 1 (write) 30-second adjustment bit 1 reset (reset): when 1 is written, initializes the divider circuit (rtc prescaler and r64cnt). this bit is always read as 0. bit 1: reset description 0 runs normally (initial value) 1 (write) divider circuit is reset 466 bit 0 start bit (start): halts and restarts the counter (clock). bit 0: start description 0 second/minute/hour/day/week/month/year counter halts 1 second/minute/hour/day/week/month/year counter runs normally (initial value) note: the 64-hz counter always runs unless stopped with the rtcen bit. 14.3 rtc operation 14.3.1 initial settings of registers after power-on all the registers should be set after the power is turned on. 14.3.2 setting the time figure 14.2 shows how to set the time when the clock is stopped. this works when the entire calendar or clock is to be set. write 1 to reset and 0 to start in the rcr2 register order is irrelevant write 1 to start in the rcr2 register set seconds, minutes, hour, day, day of the week, month and year stop clock, reset divider circuit to reset the divider circuit (rtc prescaler and r64cnt) and set the counter start clock figure 14.2 setting the time 467 14.3.3 reading the time figure 14.3 shows how to read the time. if a carry occurs while reading the time, the correct time will not be obtained, so it must be read again. part (a) in figure 14.3 shows the method of reading the time without using interrupts; part (b) in figure 14.3 shows the method using carry interrupts. to keep programming simple, method (a) should normally be used. write 0 to cie in rcr1 note: set af to 1 so that alarm flag is not cleared. read rcr1 and check cf write 0 to cie in rcr1 carry flag = 1? no yes clear the carry flag disable the carry interrupt read counter register write 1 to cie in rcr1, and write 0 to cf in rcr1 note: set af in rcr1 to 1 so that alarm flag is not cleared. interrupt generated? no yes enable the carry interrupt clear the carry flag disable the carry interrupt read counter register to read the time without using interrupts b. to use interrupts a. figure 14.3 reading the time 468 14.3.4 alarm function figure 14.4 shows how to use the alarm function. alarms can be generated using seconds, minutes, hours, day of the week, date, month, or any combination of these. set the enb bit (bit 7) to 1 in the register to which the alarm applies, and then set the alarm time in the lower bits. clear the enb bit to 0 in registers to which the alarm does not apply. when the clock and alarm times match, 1 is set in the af bit (bit 0) in rcr1. alarm detection can be checked by reading this bit, but normally it is done by interrupt. if 1 is placed in the aie bit (bit 3) in rcr1, an interrupt is generated when an alarm occurs. when using interrupts, the interrupt enable bit (bit 3 of rcr1) is 1. clock running set alarm time set whether to use alarm interrupt always set, since the flag may have been set while the alarm time was being set. write 0 to bit 0 of rcr1 to clear it. clear alarm flag monitor alarm time (wait for interrupt or check alarm flag) figure 14.4 using the alarm function 469 14.3.5 crystal oscillator circuit crystal oscillator circuit constants (recommended values) are shown in table 14.5, and the rtc crystal oscillator circuit in figure 14.5. table 14.5 recommended oscillator circuit constants (recommended values) fosc cin cout 32.768 khz 10 to 22 pf 10 to 22 pf sh7729r extal2 xtal2 xtal c in c out r f r d notes: 1. select either the c in or c out side for the frequency adjustment variable capacitor according to requirements such as frequency range, degree of stability, etc. 2. built-in resistance value r f (typ value) = 10 m ? , r d (typ value) = 400 k ? 3. c in and c out values include floating capacitance due to the wiring. take care when using a ground plane. 4. the crystal oscillation settling time depends on the mounted circuit constants, floating capacitance, etc., and should be decided after consultation with the crystal resonator manufacturer. 5. place the crystal resonator and load capacitors c in and c out as close as possible to the chip. (correct oscillation may not be possible if there is externally induced noise in the extal2 and xtal2 pins.) 6. ensure that the crystal resonator connection pin (extal2, xtal2) wiring is routed as far away as possible from other power lines (except gnd) and signal lines. figure 14.5 example of crystal oscillator circuit connection 470 14.4 usage notes 14.4.1 register writing during rtc count the following rtc registers cannot be written to during an rtc count (while bit 0 = 1 in rcr2). rseccnt, rmincnt, rhrcnt, rdaycnt, rwkcnt, rmoncnt, ryrcnt the rtc count must be halted before writing to any of the above registers. 14.4.2 use of realtime clock (rtc) periodic interrupts the method of using the periodic interrupt function is shown in figure 14.6. a periodic interrupt can be generated periodically at the interval set by the periodic interrupt enable flag (pes) in rtc control register 2 (rcr2). when the time set by the periodic interrupt enable flag (pes) has elapsed, the periodic interrupt flag (pef) is set to 1. the periodic interrupt flag (pef) is cleared to 0 upon periodic interrupt generation when the periodic interrupt enable flag (pes) is set. periodic interrupt generation can be confirmed by reading this bit, but normally the interrupt function is used. set pes, and clear pef to 0, in rcr2 clear pef to 0 set pes, clear pef elapse of time set by pes clear pef figure 14.6 using periodic interrupt function 471 section 15 serial communication interface (sci) 15.1 overview the sh7729r has an on-chip serial communication interface (sci) that supports both asynchronous and clock synchronous serial communication. it also has a multiprocessor communication function for serial communication among two or more processors. the sci supports a smart card interface, which is a serial communication feature for ic card interfaces that conforms to the iso/iec standard 7816-3 for identification cards. see section 16, smart card interface, for more information. 15.1.1 features selection of asynchronous or synchronous as the serial communication mode. ? asynchronous mode: ? serial data communication is synchronized by start-stop in character units. the sci can communicate with a universal asynchronous receiver/transmitter (uart), an asynchronous communication interface adapter (acia), or any other communications chip that employs a standard asynchronous serial system. it can also communicate with two or more other processors using the multiprocessor communication function. there are 12 selectable serial data communication formats. ? data length: 7 or 8 bits ? stop bit length: 1 or 2 bits ? parity: even, odd, or none ? multiprocessor bit: 1 or 0 ? receive error detection: parity, overrun, and framing errors ? break detection: by reading the rxd level directly from the port sc data register (scsptr) when a framing error occurs ? synchronous mode: ? serial data communication is synchronized with a clock signal. the sci can communicate with other chips having a synchronous communication function. there is one serial data communication format. ? data length: 8 bits ? receive error detection: overrun errors ? full duplex communication: the transmitting and receiving sections are independent, so the sci can transmit and receive simultaneously. both sections use double buffering, so continuous data transfer is possible in both the transmit and receive directions. ? on-chip baud rate generator with selectable bit rates 472 ? internal or external transmit/receive clock source: from either baud rate generator (internal) or sck pin (external) ? four types of interrupts: transmit-data-empty, transmit-end, receive-data-full, and receive- error interrupts are requested independently. ? when the sci is not in use, it can be stopped by halting the clock supplied to it, saving power. 15.1.2 block diagram figure 15.1 shows a block diagram of the sci. rxd txd sck sci scbrr scssr scscr sctsr scrdr scrsr scsmr scpcr scpdr parity generation parity check clock external clock module data bus internal data bus p p /4 p /16 p /64 txi tei rxi eri bus interface baud rate generator transmit/ receive control scrsr: scrdr: sctsr: sctdr: scsmr: legend receive shift register receive data register transmit shift register transmit data register serial mode register scscr: scssr: scbrr: scpdr: scpcr: serial control register serial status register bit rate register sc port data register sc port control register sctdr figure 15.1 block diagram of sci 473 figures 15.2, 15.3, and 15.4 show block diagrams of the sci i/o port pins. scif pin i/o and data control is performed by bits 11 to 8 of scpcr and bits 5 and 4 of scpdr. for details, see section 15.2.8, sc port control register (scpcr)/sc port data register (scpdr). internal data bus output enable clock input enable sci serial clock output serial clock input r scp1md0 pcrw reset c q q d r scp1md1 pcrw reset c qd r scp1dt1 pdrw reset scpt[1]/sck0 c d pdrw: scpdr write pdrr: pcrw: scpdr read scpcr write pdrr * note: * legend when reading the sck0 pin, clear the c/ a bit in scsmr and the cke1 and cke0 bits in scscr to 0, and set the scp1md1 bit in scpcr to 1 (see section 15.2.8, sc port control register (scpcr)/sc port data register (scpdr)). figure 15.2 scpt[1]/sck0 pin 474 internal data bus output enable sci serial transmission output r scp0md0 pcrw reset c q q d r scp0md1 pcrw reset c qd r scp0dt1 pdrw reset scpt[0]/txd0 c d pcrw: pdrw: scpcr write scpdr write legend figure 15.3 scpt[0]/txd0 pin 475 sci serial receive data internal data bus pdrr * scpt[0]/rxd0 pdrr: pdr read note: * when reading the rxd0 pin, set the re bit in scscr to 1. legend figure 15.4 scpt[0]/rxd0 pin 15.1.3 pin configuration the sci has the serial pins summarized in table 15.1. table 15.1 sci pins pin name abbreviation i/o function serial clock pin sck0 i/o clock i/o receive data pin rxd0 input receive data input transmit data pin txd0 output transmit data output note: these pins are made to function as serial pins by performing sci operation settings with the te, re, ckei, and ckeo bits in scscr and the c/ a bit in scsmr. break state transmission and detection can be performed by means of the scis scsptr register. 476 15.1.4 register configuration table 15.2 summarizes the sci internal registers. these registers select the communication mode (asynchronous or synchronous), specify the data format and bit rate, and control the transmitter and receiver sections. table 15.2 sci registers name abbreviation r/w initial value address access size serial mode register scsmr r/w h'00 h'fffffe80 8 bit rate register scbrr r/w h'ff h'fffffe82 8 serial control register scscr r/w h'00 h'fffffe84 8 transmit data register sctdr r/w h'ff h'fffffe86 8 serial status register scssr r/(w)* 1 h'84 h'fffffe88 8 receive data register scrdr r h'00 h'fffffe8a 8 sc port data register scpdr r/w h'00 h'04000136 (h'a4000136)* 2 8 sc port control register scpcr r/w h'a888 h'04000116 (h'a4000116)* 2 16 notes: these registers are located in area 1 of physical space. therefore, when the cache is on, either access these registers from the p2 area of logical space or else make an appropriate setting using the mmu so that these registers are not cached. *1 the only value that can be written is 0 to clear the flags. *2 when address translation by the mmu does not apply, the address in parentheses should be used. 15.2 register descriptions 15.2.1 receive shift register (scrsr) the receive shift register (scrsr) receives serial data. data input at the rxd pin is loaded into scrsr in the order received, lsb (bit 0) first, converting the data to parallel form. when one byte has been received, it is automatically transferred to scrdr. the cpu cannot read or write to scrsr directly. bit: 7 6 5 4 3 2 1 0 r/w: 477 15.2.2 receive data register (scrdr) the receive data register (scrdr) stores serial receive data. the sci completes the reception of one byte of serial data by moving the received data from the receive shift register (scrsr) into scrdr for storage. scrsr is then ready to receive the next data. this double buffering allows the sci to receive data continuously. the cpu can read but not write to scrdr. scrdr is initialized to h'00 by a reset and in standby or module standby mode. bit: 7 6 5 4 3 2 1 0 initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r r 15.2.3 transmit shift register (sctsr) the transmit shift register (sctsr) transmits serial data. the sci loads transmit data from the transmit data register (sctdr) into sctsr, then transmits the data serially from the txd pin, lsb (bit 0) first. after transmitting one-byte data, the sci automatically loads the next transmit data from sctdr into sctsr and starts transmitting again. if the tdre bit in scssr is 1, however, the sci does not load the sctdr contents into sctsr. the cpu cannot read or write to sctsr directly. bit: 7 6 5 4 3 2 1 0 r/w: 478 15.2.4 transmit data register (sctdr) the transmit data register (sctdr) is an 8-bit register that stores data for serial transmission. when the sci detects that the transmit shift register (sctsr) is empty, it moves transmit data written in sctdr into sctsr and starts serial transmission. continuous serial transmission is possible by writing the next transmit data in sctdr during serial transmission from sctsr. the cpu can always read and write to sctdr. sctdr is initialized to h'ff by a reset and in standby or module standby mode. bit: 7 6 5 4 3 2 1 0 initial value: 1 1 1 1 1 1 1 1 r/w: r/w r/w r/w r/w r/w r/w r/w r/w 15.2.5 serial mode register (scsmr) the serial mode register (scsmr) is an 8-bit register that specifies the sci serial communication format and selects the clock source for the baud rate generator. the cpu can always read and write to scsmr. scsmr is initialized to h'00 by a reset and in standby or module standby mode. bit: 7 6 5 4 3 2 1 0 c/ a chr pe o/ e stop mp cks1 cks0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit 7communication mode (c/ a ): selects whether the sci operates in asynchronous or synchronous mode. bit 7: c/ a description 0 asynchronous mode (initial value) 1 synchronous mode 479 bit 6character length (chr): selects 7-bit or 8-bit data in asynchronous mode. in the synchronous mode, the data length is always eight bits, regardless of the chr setting. bit 6: chr description 0 8-bit data (initial value) 1 7-bit data when 7-bit data is selected, the msb (bit 7) of the transmit data register is not transmitted. bit 5parity enable (pe): selects whether to add a parity bit to transmit data and to check the parity of receive data, in asynchronous mode. in synchronous mode, a parity bit is neither added nor checked, regardless of the pe setting. bit 5: pe description 0 parity bit not added or checked (initial value) 1 parity bit added and checked when pe is set to 1, an even or odd parity bit is added to transmit data, depending on the parity mode (o/ e ) setting. receive data parity is checked according to the even/odd (o/ e ) mode setting. bit 4parity mode (o/ e ): selects even or odd parity when parity bits are added and checked. the o/ e setting is used only in asynchronous mode and only when the parity enable bit (pe) is set to 1 to enable parity addition and checking. the o/ e setting is ignored in synchronous mode, or in asynchronous mode when parity addition and checking is disabled. bit 4: o/ e description 0 even parity (initial value) if even parity is selected, the parity bit is added to transmit data to make an even number of 1s in the transmitted character and parity bit combined. receive data is checked to see if it has an even number of 1s in the received character and parity bit combined. 1 odd parity if odd parity is selected, the parity bit is added to transmit data to make an odd number of 1s in the transmitted character and parity bit combined. receive data is checked to see if it has an odd number of 1s in the received character and parity bit combined. 480 bit 3stop bit length (stop): selects one or two bits as the stop bit length in asynchronous mode. this setting is used only in asynchronous mode. it is ignored in synchronous mode because no stop bits are added. when receiving, only the first stop bit is checked, regardless of the stop bit setting. if the second stop bit is 1, it is treated as a stop bit, but if the second stop bit is 0, it is treated as the start bit of the next incoming character. bit 3: stop description 0 one stop bit (initial value) when transmitting, a single 1-bit is added at the end of each transmitted character. 1 two stop bits when transmitting, two 1-bits are added at the end of each transmitted character. bit 2multiprocessor mode (mp): selects multiprocessor format. when multiprocessor format is selected, settings of the parity enable (pe) and parity mode (o/ e ) bits are ignored. the mp bit setting is used only in asynchronous mode; it is ignored in synchronous mode. for the multiprocessor communication function, see section 15.3.3, multiprocessor communication. bit 2: mp description 0 multiprocessor function disabled (initial value) 1 multiprocessor format selected bits 1 and 0clock select 1 and 0 (cks1, cks0): select the internal clock source of the on- chip baud rate generator. four clock sources are available. p , p /4, p /16 and p /64. for further information on the clock source, bit rate register settings, and baud rate, see section 15.2.9, bit rate register (scbrr). bit 1: cks1 bit 0: cks0 description 00p (initial value) 1p /4 10p /16 1p /64 note: p : peripheral clock 481 15.2.6 serial control register (scscr) the serial control register (scscr) operates the sci transmitter/receiver, selects the serial clock output in asynchronous mode, enables/disables interrupt requests, and selects the transmit/receive clock source. the cpu can always read and write to scscr. scscr is initialized to h'00 by a reset and in standby or module standby mode. bit: 7 6 5 4 3 2 1 0 tie rie te re mpie teie cke1 cke0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w bit 7transmit interrupt enable (tie): enables or disables the transmit-data-empty interrupt (txi) requested when the transmit data register empty bit (tdre) in the serial status register (scssr) is set to 1 due to transfer of serial transmit data from sctdr to sctsr. bit 7: tie description 0 transmit-data-empty interrupt request (txi) is disabled (initial value) the txi interrupt request can be cleared by reading tdre after it has been set to 1, then clearing tdre to 0, or by clearing tie to 0. 1 transmit-data-empty interrupt request (txi) is enabled bit 6receive interrupt enable (rie): enables or disables the receive-data-full interrupt (rxi) requested when the receive data register full bit (rdrf) in the serial status register (scssr) is set to 1 due to transfer of serial receive data from scrsr to scrdr. it also enables or disables receive-error interrupt (eri) requests. bit 6: rie description 0 receive-data-full interrupt (rxi) and receive-error interrupt (eri) requests are disabled (initial value) rxi and eri interrupt requests can be cleared by reading the rdrf flag or error flag (fer, per, or orer) after it has been set to 1, then clearing the flag to 0, or by clearing rie to 0. 1 receive-data-full interrupt (rxi) and receive-error interrupt (eri) requests are enabled 482 bit 5transmit enable (te): enables or disables the sci serial transmitter. bit 5: te description 0 transmitter disabled (initial value) the transmit data register empty bit (tdre) in the serial status register (scssr) is fixed at 1. 1 transmitter enabled serial transmission starts when the transmit data register empty (tdre) bit in the serial status register (scssr) is cleared to 0 after writing of transmit data into the sctdr. select the transmit format in scsmr before setting te to 1. bit 4receive enable (re): enables or disables the sci serial receiver. bit 4: re description 0 receiver disabled (initial value) clearing re to 0 does not affect the receive flags (rdrf, fer, per, orer). these flags retain their previous values. 1 receiver enabled serial reception starts when a start bit is detected in asynchronous mode, or synchronous clock input is detected in synchronous mode. select the receive format in scsmr before setting re to 1. bit 3multiprocessor interrupt enable (mpie): enables or disables multiprocessor interrupts. the mpie setting is used only in asynchronous mode, and only if the multiprocessor mode bit (mp) in the serial mode register (scsmr) is set to 1 during reception. the mpie setting is ignored in synchronous mode or when the mp bit is cleared to 0. bit 3: mpie description 0 multiprocessor interrupts are disabled (normal receive operation) (initial value) mpe is cleared to 0 when mpie is cleared to 0, or the multiprocessor bit (mpb) is set to 1 in receive data. 1 multiprocessor interrupts are enabled receive-data-full interrupt requests (rxi), receive-error interrupt requests (eri), and setting of the rdrf, fer, and orer status flags in the serial status register (scssr) are disabled until data with a multiprocessor bit of 1 is received. the sci does not transfer receive data from scrsr to scrdr, does not detect receive errors, and does not set the rdrf, fer, and orer flags in the serial status register (scssr). when it receives data that includes mpb = 1, the scssrs mpb flag is set to 1, and the sci automatically clears mpie to 0, generates rxi and eri interrupts (if the tie and rie bits in the scscr are set to 1), and allows the fer and orer bits to be set. 483 bit 2transmit-end interrupt enable (teie): enables or disables the transmit-end interrupt (tei) requested if sctdr does not contain new transmit data when the msb is transmitted. bit 2: teie description 0 transmit-end interrupt (tei) requests are disabled* (initial value) 1 transmit-end interrupt (tei) requests are enabled* note: * the tei request can be cleared by reading the tdre bit in the serial status register (scssr) after it has been set to 1, then clearing tdre to 0 and clearing the transmit end (tend) bit to 0, or by clearing the teie bit to 0. bits 1 and 0clock enable 1 and 0 (cke1, cke0): select the sci clock source and enable or disable clock output from the sck pin. depending on the combination of cke1 and cke0, the sck pin can be used for serial clock output or serial clock input. the cke0 setting is valid only in asynchronous mode, and only when the sci is internally clocked (cke1 = 0). the cke0 setting is ignored in synchronous mode, or when an external clock source is selected (cke1 = 1). before selecting the sci operating mode in the serial mode register (scsmr), set cke1 and cke0. for further details on selection of the sci clock source, see table 15.10 in section 15.3, operation. bit 1: cke1 bit 0: cke0 description 0 0 asynchronous mode internal clock, sck pin used for input pin (input signal is ignored) (initial value) synchronous mode internal clock, sck pin used for serial clock output (initial value) 1 asynchronous mode internal clock, sck pin used for clock output* 1 synchronous mode internal clock, sck pin used for serial clock output 1 0 asynchronous mode external clock, sck pin used for clock input* 2 synchronous mode external clock, sck pin used for serial clock input 1 asynchronous mode external clock, sck pin used for clock input* 2 synchronous mode external clock, sck pin used for serial clock input notes: *1 the output clock frequency is the same as the bit rate. *2 the input clock frequency is 16 times the bit rate. 484 15.2.7 serial status register (scssr) the serial status register (scssr) is an 8-bit register containing multiprocessor bit values, and status flags that indicate the sci operating state. the cpu can always read and write to scssr, but cannot write 1 to the status flags (tdre, rdrf, orer, per, and fer). these flags can be cleared to 0 only if they have first been read (after being set to 1). bits 2 (tend) and 1 (mpb) are read-only bits that cannot be written. scssr is initialized to h'84 by a reset and in standby or module standby mode. bit: 7 6 5 4 3 2 1 0 tdre rdrf orer fer per tend mpb mpbt initial value: 1 0 0 0 0 1 0 0 r/w: r/(w)* r/(w)* r/(w)* r/(w)* r/(w)* r r r/w note: * the only value that can be written is 0 to clear the flag. bit 7transmit data register empty (tdre): indicates that the sci has loaded transmit data from sctdr into sctsr and new serial transmit data can be written in sctdr. bit 7: tdre description 0 sctdr contains valid transmit data tdre is cleared to 0 when software reads tdre after it has been set to 1, then writes 0 in tdre or data is written in sctdr. 1 sctdr does not contain valid transmit data (initial value) tdre is set to 1 when the chip is reset or enters standby mode, the te bit in the serial control register (scscr) is cleared to 0, or sctdr contents are loaded into sctsr, so new data can be written in sctdr. bit 6receive data register full (rdrf): indicates that scrdr contains received data. bit 6: rdrf description 0 scrdr does not contain valid receive data (initial value) rdrf is cleared to 0 when the chip is reset or enters standby mode, or software reads rdrf after it has been set to 1, then writes 0 in rdrf. 1 scrdr contains valid receive data rdrf is set to 1 when serial data is received normally and transferred from scrsr to scrdr. note: scrdr and rdrf are not affected by detection of receive errors or by clearing of the re bit to 0 in the serial control register. they retain their previous contents. if rdrf is still set to 1 when reception of the next data ends, an overrun error (orer) occurs and the receive data is lost. 485 bit 5overrun error (orer): indicates that data reception aborted due to an overrun error. bit 5: orer description 0 receiving is in progress or has ended normally* 1 (initial value) orer is cleared to 0 when the chip is reset or enters standby mode, or when software reads orer after it has been set to 1, then writes 0 to orer. 1 a receive overrun error occurred* 2 orer is set to 1 if reception of the next serial data ends when rdrf is set to 1. notes: *1 clearing the re bit to 0 in the serial control register does not affect the orer bit, which retains its previous value. *2 scrdr continues to hold the data received before the overrun error, so subsequent receive data is lost. serial receiving cannot continue while orer is set to 1. in synchronous mode, serial transmitting is also disabled. bit 4framing error (fer): indicates that data reception aborted due to a framing error in asynchronous mode. bit 4: fer description 0 receiving is in progress or has ended normally (initial value) clearing the re bit to 0 in the serial control register does not affect the fer bit, which retains its previous value. fer is cleared to 0 when the chip is reset or enters standby mode, or when software reads fer after it has been set to 1, then writes 0 to fer. 1 a receive framing error occurred when the stop bit length is two bits, only the first bit is checked. the second stop bit is not checked. when a framing error occurs, the sci transfers the receive data into scrdr but does not set rdrf. serial receiving cannot continue while fer is set to 1. in synchronous mode, serial transmitting is also disabled. fer is set to 1 if the stop bit at the end of receive data is checked and found to be 0. 486 bit 3parity error (per): indicates that data reception (with parity) aborted due to a parity error in asynchronous mode. bit 3: per description 0 receiving is in progress or has ended normally (initial value) clearing the re bit to 0 in the serial control register does not affect the per bit, which retains its previous value. per is cleared to 0 when the chip is reset or enters standby mode, or when software reads per after it has been set to 1, then writes 0 to per. 1 a receive parity error occurred when a parity error occurs, the sci transfers the receive data into scrdr but does not set rdrf. serial receiving cannot continue while per is set to 1. in synchronous mode, serial transmitting is also disabled. per is set to 1 if the number of 1s in receive data, including the parity bit, does not match the even or odd parity setting of the parity mode bit (o/ e ) in the serial mode register (scsmr). bit 2transmit end (tend): indicates that when the last bit of a serial character was transmitted, sctdr did not contain valid data, so transmission has ended. tend is a read-only bit and cannot be written to. bit 2: tenddescription 0 transmission is in progress tend is cleared to 0 when software reads tdre after it has been set to 1, then writes 0 to tdre. 1 end of transmission (initial value) tend is set to 1 when the chip is reset or enters standby mode, when te is cleared to 0 in the serial control register (scscr), or if tdre is 1 when the last bit of a one-byte serial character is transmitted. bit 1multiprocessor bit (mpb): stores the value of the multiprocessor bit in receive data when a multiprocessor format is selected for receiving in asynchronous mode. mpb is a read-only bit and cannot be written to. bit 1: mpb description 0 multiprocessor bit value in receive data is 0 (initial value) if re is cleared to 0 when a multiprocessor format is selected, mpb retains its previous value. 1 multiprocessor bit value in receive data is 1 487 bit 0multiprocessor bit transfer (mpbt): stores the value of the multiprocessor bit added to transmit data when a multiprocessor format is selected for transmitting in asynchronous mode. the mpbt setting is ignored in synchronous mode, when a multiprocessor format is not selected, or when the sci is not transmitting. bit 0: mpbt description 0 multiprocessor bit value in transmit data is 0 (initial value) 1 multiprocessor bit value in transmit data is 1 15.2.8 sc port control register (scpcr)/sc port data register (scpdr) the sc port control register (scpcr) and sc port data register (scpdr) control i/o and data for the port pins multiplexed with the serial communication interface (sci) pins. scpcr settings are used to perform i/o control, to enable data written in scpdr to be output to the txd pin, and input data to be read from the rxd pin, and to control serial transmission/reception breaks. it is also possible to read data on the sck pin, and write output data. scpcr bit: 1514131211109876543210 scp7 md1 scp7 md0 scp6 md1 scp6 md0 scp5 md1 scp5 md0 scp4 md1 scp4 md0 scp3 md1 scp3 md0 scp2 md1 scp2 md0 scp1 md1 scp1 md0 scp0 md1 scp0 md0 initial value: 1010100010001000 r/w: r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w scpdr bit: 7 6 5 4 3 2 1 0 scp7dt scp6dt scp5dt scp4dt scp3dt scp2dt scp1dt scp0dt initial value: 0 0 0 0 0 0 0 0 r/w: r r/w r/w r/w r/w r/w r/w r/w sci pin i/o and data control are performed by bits 3C0 of scpcr and bits 1 and 0 of scpdr. 488 scpcr bits 3 and 2serial clock port i/o (scp1md1, scp1md0): specify serial port sck pin i/o. when the sck pin is actually used as a port i/o pin, clear the c/ a bit in scsmr and bits cke1 and cke0 in scscr to 0. bit 3: scp1md1 bit 2: scp1md0 description 0 0 scp1dt bit value is not output to sck pin 0 1 scp1dt bit value is output to sck pin 1 0 sck pin value is read from scp1dt bit 1 1 (initial values: 1 and 0) scpdr bit 1serial clock port data (scp1dt): specifies the serial port sck pin i/o data. input or output is specified by the scp1md0 and scp1md1 bits. in output mode, the value of the scp1dt bit is output to the sck pin. in output mode, the sck pin value is read from the scp1dt bit. bit 1: scp1dt description 0 i/o data is low (initial value) 1 i/o data is high scpcr bits 1 and 0serial port break i/o (scp0md1, scp0md0): specify the serial port txd pin output condition. when the txd pin is actually used as a port output pin and outputs the value set with the scp0dt bit, clear the te bit in scscr to 0. bit 1: scp0md1 bit 0: scp0md0 description 0 0 scp0dt bit value is not output to txd pin (initial value) 0 1 scp0dt bit value is output to txd pin 489 scpdr bit 0serial port break data (scp0dt): specifies the serial port rxd pin input data and txd pin output data. the txd pin output condition is specified by the scp0md0 and scp0md1 bits. when the txd pin is set to output mode, the value of the scp0dt bit is output to the txd pin. the rxd pin value is read from the scp0dt bit regardless of the values of the scp0md0 and scp0md1 bits, if re in scscr is set to 1. the initial value of this bit after a power-on reset is undefined. bit 0: scp0dt description 0 i/o data is low (initial value) 1 i/o data is high block diagrams of the sci i/o port pins are shown in figures 15.2, 15.3, and 15.4. 15.2.9 bit rate register (scbrr) the bit rate register (scbrr) is an 8-bit register that, together with the baud rate generator clock source selected by the cks1 and cks0 bits in the serial mode register (scsmr), determines the serial transmit/receive bit rate. the cpu can always read and write to scbrr. scbrr is initialized to h'ff by a reset, and in module standby or standby mode. each channel has independent baud rate generator control, so different values can be set in two channels. bit: 7 6 5 4 3 2 1 0 initial value: 1 1 1 1 1 1 1 1 r/w: r/w r/w r/w r/w r/w r/w r/w r/w the scbrr setting is calculated as follows: asynchronous mode: n = p 64 2 2n C 1 b 10 6 C 1 synchronous mode: n = p 8 2 2n C 1 b 10 6 C 1 b: bit rate (bits/s) n: scbrr setting for baud rate generator (0 n 255) p : operating frequency for peripheral modules (mhz) n: baud rate generator clock source (n = 0, 1, 2, 3) (for the clock sources and values of n, see table 15.3.) 490 table 15.3 scsmr settings scsmr settings n clock source cks1 cks0 0p 00 1p /4 0 1 2p /16 1 0 3p /64 1 1 note: the bit rate error in asynchronous is given by the following formula: error ( % ) = ( p 10 6 (n + 1) b 64 2 2n C 1 ) 100 table 15.4 lists examples of scbrr settings in asynchronous mode, and table 15.5 lists examples of scbrr settings in synchronous mode. table 15.4 bit rates and scbrr settings in asynchronous mode p (mhz) 2 2.097152 2.4576 bit rate (bits/s) n n error ( % ) n n error ( % ) n n error ( % ) 110 1 141 0.03 1 148 C0.04 1 174 C0.26 150 1 103 0.16 1 108 0.21 1 127 0.00 300 0 207 0.16 0 217 0.21 0 255 0.00 600 0 103 0.16 0 108 0.21 0 127 0.00 1200 0 51 0.16 0 54 C0.70 0 63 0.00 2400 0 25 0.16 0 26 1.14 0 31 0.00 4800 0 12 0.16 0 13 C2.48 0 15 0.00 9600 0 6 C6.99 0 6 C2.48 0 7 0.00 19200 0 2 8.51 0 2 13.78 0 3 0.00 31250 0 1 0.00 0 1 4.86 0 1 22.88 38400 0 1 C18.62 0 1 C14.67 0 1 0.00 491 table 15.4 bit rates and scbrr settings in asynchronous mode (cont) p (mhz) 3 3.6864 4 bit rate (bits/s) n n error ( % ) n n error ( % ) n n error ( % ) 110 1 212 0.03 2 64 0.70 2 70 0.03 150 1 155 0.16 1 191 0.00 1 207 0.16 300 1 77 0.16 1 95 0.00 1 103 0.16 600 0 155 0.16 0 191 0.00 0 207 0.16 1200 0 77 0.16 0 95 0.00 0 103 0.16 2400 0 38 0.16 0 47 0.00 0 51 0.16 4800 0 19 C2.34 0 23 0.00 0 25 0.16 9600 0 9 C2.34 0 11 0.00 0 12 0.16 19200 0 4 C2.34 0 5 0.00 0 6 C6.99 31250 0 2 0.00 0 3 0.00 38400 0 2 0.00 0 2 8.51 p (mhz) 4.9152 5 6 bit rate (bits/s) n n error ( % ) n n error ( % ) n n error ( % ) 110 2 86 0.31 2 88 C0.25 2 106 C0.44 150 1 255 0.00 2 64 0.16 2 77 0.16 300 1 127 0.00 1 129 0.16 1 155 0.16 600 0 255 0.00 1 64 0.16 1 77 0.16 1200 0 127 0.00 0 129 0.16 0 155 0.16 2400 0 63 0.00 0 64 0.16 0 77 0.16 4800 0 31 0.00 0 32 C1.36 0 38 0.16 9600 0 15 0.00 0 15 1.73 0 19 C2.34 19200 0 7 0.00 0 7 1.73 0 9 C2.34 31250 0 4 C1.70 0 4 0.00 0 5 0.00 38400 0 3 0.00 0 3 1.73 0 4 C2.34 492 table 15.4 bit rates and scbrr settings in asynchronous mode (cont) p (mhz) 6.144 7.3728 8 bit rate (bits/s) n n error ( % ) n n error ( % ) n n error ( % ) 110 2 108 0.08 2 130 C0.07 2 141 0.03 150 2 79 0.00 2 95 0.00 2 103 0.16 300 1 159 0.00 1 191 0.00 1 207 0.16 600 1 79 0.00 1 95 0.00 1 103 0.16 1200 0 159 0.00 0 191 0.00 0 207 0.16 2400 0 79 0.00 0 95 0.00 0 103 0.16 4800 0 39 0.00 0 47 0.00 0 51 0.16 9600 0 19 0.00 0 23 0.00 0 25 0.16 19200 0 9 0.00 0 11 0.00 0 12 0.16 31250 0 5 2.40 0 6 5.33 0 7 0.00 38400 0 4 0.00 0 5 0.00 0 6 C6.99 p (mhz) 9.8304 10 12 12.288 bit rate (bits/s) n n error ( % )n n error ( % )n n error ( % )n n error ( % ) 110 2 174 C0.26 2 177 C0.25 2 212 0.03 2 217 0.08 150 2 127 0.00 2 129 0.16 2 155 0.16 2 159 0.00 300 1 255 0.00 2 64 0.16 2 77 0.16 2 79 0.00 600 1 127 0.00 1 129 0.16 1 155 0.16 1 159 0.00 1200 0 255 0.00 1 64 0.16 1 77 0.16 1 79 0.00 2400 0 127 0.00 0 129 0.16 0 155 0.16 0 159 0.00 4800 0 63 0.00 0 64 0.16 0 77 0.16 0 79 0.00 9600 0 31 0.00 0 32 C1.36 0 38 0.16 0 39 0.00 19200 0 15 0.00 0 15 1.73 0 19 0.16 0 19 0.00 31250 0 9 C1.70 0 9 0.00 0 11 0.00 0 11 2.40 38400 0 7 0.00 0 7 1.73 0 9 C2.34 0 9 0.00 493 table 15.4 bit rates and scbrr settings in asynchronous mode (cont) p (mhz) 14.7456 16 19.6608 20 bit rate (bits/s) n n error ( % )n n error ( % )n n error ( % )n n error ( % ) 110 3 64 0.70 3 70 0.03 3 86 0.31 3 88 C0.25 150 2 191 0.00 2 207 0.16 2 255 0.00 3 64 0.16 300 2 95 0.00 2 103 0.16 2 127 0.00 2 129 0.16 600 1 191 0.00 1 207 0.16 1 255 0.00 2 64 0.16 1200 1 95 0.00 1 103 0.16 1 127 0.00 1 129 0.16 2400 0 191 0.00 0 207 0.16 0 255 0.00 1 64 0.16 4800 0 95 0.00 0 103 0.16 0 127 0.00 0 129 0.16 9600 0 47 0.00 0 51 0.16 0 63 0.00 0 64 0.16 19200 0 23 0.00 0 25 0.16 0 31 0.00 0 32 C1.36 31250 0 14 C1.70 0 15 0.00 0 19 C1.70 0 19 0.00 38400 0 11 0.00 0 12 0.16 0 15 0.00 0 15 1.73 p (mhz) 24 24.576 28.7 30 bit rate (bits/s) n n error ( % )n n error ( % )n n error ( % )n n error ( % ) 110 3 106 C0.44 3 108 0.08 3 126 0.31 3 132 0.13 150 3 77 0.16 3 79 0.00 3 92 0.46 3 97 C0.35 300 2 155 0.16 2 159 0.00 2 186 C0.08 2 194 0.16 600 2 77 0.16 2 79 0.00 2 92 0.46 2 97 C0.35 1200 1 155 0.16 1 159 0.00 1 186 C0.08 1 194 0.16 2400 1 77 0.16 1 79 0.00 1 92 0.46 1 97 C0.35 4800 0 155 0.16 0 159 0.00 0 186 C0.08 0 194 C1.36 9600 0 77 0.16 0 79 0.00 0 92 0.46 0 97 C0.35 19200 0 38 0.16 0 39 0.00 0 46 C0.61 0 48 C0.35 31250 0 23 0.00 0 24 C1.70 0 28 C1.03 0 29 0.00 38400 0 19 C2.34 0 19 0.00 0 22 1.55 0 23 1.73 494 table 15.5 bit rates and scbrr settings in synchronous mode p (mhz) bit rate 4 8 16 28.7 30 (bits/s) n n n n n n n n n n 110 250 2 249 3 124 3 249 500 2 124 2 249 3 124 3 223 3 233 1k 1 249 2 124 2 249 3 111 3 116 2.5k 1 99 1 199 2 99 2 178 2 187 5k 0 199 1 99 1 199 2 89 2 93 10k 0 99 0 199 1 99 1 178 1 187 25k 0 39 0 79 0 159 1 71 1 74 50k 0 19 0 39 0 79 0 143 0 149 100k 09019039071074 250k 0307015029 500k 010307014 1m 00*0103 2m 0 0* 0 1 note: settings with an error of 1 % or less are recommended. blank: no setting possible : setting possible, but error occurs *: continuous transmit/receive operation not possible 495 table 15.6 indicates the maximum bit rates in asynchronous mode when the baud rate generator is used. tables 15.7 and 15.8 list the maximum rates for external clock input. table 15.6 maximum bit rates for various frequencies with baud rate generator (asynchronous mode) settings p (mhz) maximum bit rate (bits/s) n n 2 62500 0 0 2.097152 65536 0 0 2.4576 76800 0 0 3 93750 0 0 3.6864 115200 0 0 4 125000 0 0 4.9152 153600 0 0 8 250000 0 0 9.8304 307200 0 0 12 375000 0 0 14.7456 460800 0 0 16 500000 0 0 19.6608 614400 0 0 20 625000 0 0 24 750000 0 0 24.576 768000 0 0 28.7 896875 0 0 30 937500 0 0 496 table 15.7 maximum bit rates with external clock input (asynchronous mode) p (mhz) external input clock (mhz) maximum bit rate (bits/s) 2 0.5000 31250 2.097152 0.5243 32768 2.4576 0.6144 38400 3 0.7500 46875 3.6864 0.9216 57600 4 1.0000 62500 4.9152 1.2288 76800 8 2.0000 125000 9.8304 2.4576 153600 12 3.0000 187500 14.7456 3.6864 230400 16 4.0000 250000 19.6608 4.9152 307200 20 5.0000 312500 24 6.0000 375000 24.576 6.1440 384000 28.7 7.1750 448436 30 7.5000 468750 table 15.8 maximum bit rates with external clock input (synchronous mode) p (mhz) external input clock (mhz) maximum bit rate (bits/s) 8 1.3333 1333333.3 16 2.6667 2666666.7 24 4.0000 4000000.0 28.7 4.7833 4783333.3 30 5.0000 5000000.0 497 15.3 operation 15.3.1 overview for serial communication, the sci has an asynchronous mode in which characters are synchronized individually, and a synchronous mode in which communication is synchronized with clock pulses. asynchronous/synchronous mode and the transmission format are selected in the serial mode register (scsmr), as shown in table 15.9. the sci clock source is selected by the combination of the c/ a bit in the serial mode register (scsmr) and the cke1 and cke0 bits in the serial control register (scscr), as shown in table 15.10. asynchronous mode: ? data length is selectable: 7 or 8 bits. ? parity and multiprocessor bits are selectable. so is the stop bit length (1 or 2 bits). the combination of the preceding selections constitutes the communication format and character length. ? in receiving, it is possible to detect framing errors (fer), parity errors (per), overrun errors (orer) and breaks. ? an internal or external clock can be selected as the sci clock source. ? when an internal clock is selected, the sci operates using the on-chip baud rate generator, and can output a serial clock signal with a frequency matching the bit rate. ? when an external clock is selected, the external clock input must have a frequency 16 times the bit rate. (the on-chip baud rate generator is not used.) synchronous mode: ? the transmission/reception format has a fixed 8-bit data length. ? in receiving, it is possible to detect overrun errors (orer). ? an internal or external clock can be selected as the sci clock source. ? when an internal clock is selected, the sci operates using the on-chip baud rate generator, and outputs a serial clock signal to external devices. ? when an external clock is selected, the sci operates on the input serial clock. the on-chip baud rate generator is not used. 498 table 15.9 serial mode register settings and sci communication formats scsmr settings sci communication format bit 7 c/ a bit 6 chr bit 5 pe bit 2 mp bit 3 stop mode data length parity bit multipro- cessor bit stop bit length 00000 asynchronous 8-bit not set not set 1 bit 1 2 bits 1 0 set 1 bit 1 2 bits 1 0 0 7-bit not set 1 bit 1 2 bits 1 0 set 1 bit 1 2 bits 0 * 1 0 asynchronous 8-bit not set set 1 bit *1 (multiprocessor 2 bits 1* 0 format) 7-bit 1 bit * 1 2 bits 1**** synchronous 8-bit not set none note: asterisks (*) indicate dont-care bits. table 15.10 scsmr and scscr settings and sci clock source selection scsmr scscr settings sci transmit/receive clock bit 7 c/ a bit 1 cke1 bit 0 cke0 mode clock source sck pin function 0 0 0 asynchronous internal sci does not use the sck pin 1 mode outputs a clock with frequency matching the bit rate 1 0 external inputs a clock with frequency 16 1 times the bit rate 1 0 0 synchronous internal outputs the serial clock 1 mode 1 0 external inputs the serial clock 1 499 15.3.2 operation in asynchronous mode in asynchronous mode, each transmitted or received character begins with a start bit and ends with a stop bit. serial communication is synchronized one character at a time. the transmitting and receiving sections of the sci are independent, so full duplex communication is possible. the transmitter and receiver are both double buffered, so data can be written and read while transmitting and receiving are in progress, enabling continuous transmitting and receiving. figure 15.5 shows the general format of asynchronous serial communication. in asynchronous serial communication, the communication line is normally held in the mark (high) state. the sci monitors the line and starts serial communication when the line goes to the space (low) state, indicating a start bit. one serial character consists of a start bit (low), data (lsb first), parity bit (high or low), and stop bit (high), in that order. when receiving in asynchronous mode, the sci synchronizes at the falling edge of the start bit. the sci samples each data bit on the eighth pulse of a clock with a frequency 16 times the bit rate. receive data is latched at the center of each bit. 0d 0 d 1 d 2 d 3 d 4 d 5 d 6 d 7 1 1 0/1 1 1 (lsb) (msb) serial data start bit 1 bit transmit/receive data 7 or 8 bits one unit of communication data (character or frame) idle (mark) state parity bit stop bit 1 or no bit 1 or 2 bits figure 15.5 example of data format in asynchronous communication (8-bit data with parity and two stop bits) 500 transmit/receive formats: table 15.11 lists the 12 communication formats that can be selected in asynchronous mode. the format is selected by settings in the serial mode register (scsmr). table 15.11 serial communication formats (asynchronous mode) scsmr bits serial transmit/receive format and frame length chr pe mp stop 1 23456789 10 11 12 0 0 0 0 start 8-bit data stop 0 0 0 1 start 8-bit data stop stop 0 1 0 0 start 8-bit data p stop 0 1 0 1 start 8-bit data p stop stop 1 0 0 0 start 7-bit data stop 1 0 0 1 start 7-bit data stop stop 1 1 0 0 start 7-bit data p stop 1 1 0 1 start 7-bit data p stop stop 0 1 0 start 8-bit data mpb stop 0 1 1 start 8-bit data mpb stop stop 1 1 0 start 7-bit data mpb stop 1 1 1 start 7-bit data mpb stop stop : dont care bits start: start bit stop: stop bit p: parity bit mpb: multiprocessor bit clock: an internal clock generated by the on-chip baud rate generator or an external clock input from the sck pin can be selected as the sci transmit/receive clock. the clock source is selected by the c/ a bit in the serial mode register (scsmr) and bits cke1 and cke0 in the serial control register (scscr) (table 15.10). when an external clock is input at the sck pin, it must have a frequency equal to 16 times the desired bit rate. 501 when the sci operates on an internal clock, it can output a clock signal at the sck pin. the frequency of this output clock is equal to the bit rate. the phase is aligned as in figure 15.6 so that the rising edge of the clock occurs at the center of each transmit data bit. 0 d0d1d2d3d4d5d6d70/1 1 1 1 frame figure 15.6 output clock and serial data timing (asynchronous mode) transmitting and receiving data (sci initialization (asynchronous mode)): before transmitting or receiving, clear the te and re bits to 0 in the serial control register (scscr), then initialize the sci as follows. when changing the operation mode or communication format, always clear the te and re bits to 0 before following the procedure given below. clearing te to 0 sets tdre to 1 and initializes the transmit shift register (sctsr). clearing re to 0, however, does not initialize the rdrf, per, fer, and orer flags or receive data register (scrdr), which retain their previous contents. when an external clock is used, the clock should not be stopped during initialization or subsequent operation. sci operation becomes unreliable if the clock is stopped. figure 15.7 shows a sample flowchart for initializing the sci. the procedure for initializing the sci is: 1. select the clock source in the serial control register (scscr). leave rie, tie, teie, mpie, te, and re cleared to 0. if clock output is selected in asynchronous mode, clock output starts immediately after the setting is made in scscr. 2. select the communication format in the serial mode register (scsmr). 3. write the value corresponding to the bit rate in the bit rate register (scbrr) (not necessary if an external clock is used). 4. wait for at least the interval required to transmit or receive one bit, then set te or re in the serial control register (scscr) to 1. also set rie, tie, teie, and mpie as necessary. setting te or re enables the sci to use the txd or rxd pin. the initial state is the mark state when transmitting, or the idle state (waiting for a start bit) when receiving. 502 initialization clear te and re bits in scscr to 0 select communication format in scsmr set value in scbrr set cke1 and cke0 bits in scscr (te and re bits are 0) wait set te and re bits in scscr to 1 and set rie, teie, and mpie bits has a 1-bit interval elapsed? end no yes (1) (2) (3) (4) note: numbers in parentheses refer to steps in the preceding procedure description. figure 15.7 sample flowchart for sci initialization transmitting serial data (asynchronous mode): figure 15.8 shows a sample flowchart for transmitting serial data. the procedure for transmitting serial data is: 1. sci status check and transmit data write: read the serial status register (scssr), check that the tdre bit is 1, then write transmit data in the transmit data register (sctdr) and clear tdre to 0. 2. to continue transmitting serial data: read the tdre bit to check whether it is safe to write (if it reads 1); if so, write data in sctdr, then clear tdre to 0. 3. to output a break at the end of serial transmission: set the port sc data register (scpdr) and port sc control register (scpcr), then clear the te bit to 0 in the serial control register (scscr). for scpcr and scpdr settings, see section 15.2.8, port sc control register (scpcr)/port sc data register (scpdr). 503 tdre = 1? write transmit data to sctdr and clear tdre bit in scssr to 0 all data transmitted? yes tend = 1? read tend bit in scssr break output? yes clear te bit in scscr to 0 end of transmission yes read tdre bit in scssr no no yes no no (1) (2) (3) start of transmission set scpdr and scpcr note: numbers in parentheses refer to steps in the preceding procedure description. figure 15.8 sample flowchart for transmitting serial data 504 in transmitting serial data, the sci operates as follows: 1. the sci monitors the tdre bit in scssr. when tdre is cleared to 0, the sci recognizes that the transmit data register (sctdr) contains new data, and loads this data from sctdr into the transmit shift register (sctsr). 2. after loading the data from sctdr into sctsr, the sci sets the tdre bit to 1 and starts transmitting. if the transmit-data-empty interrupt enable bit (tie) is set to 1 in scscr, the sci requests a transmit-data-empty interrupt (txi) at this time. serial transmit data is transmitted in the following order from the txd pin: a. start bit: one 0 bit is output. b. transmit data: seven or eight bits of data are output, lsb first. c. parity bit or multiprocessor bit: one parity bit (even or odd parity) or one multiprocessor bit is output. formats in which neither a parity bit nor a multiprocessor bit is output can also be selected. d. stop bit: one or two 1-bits (stop bits) are output. e. marking: output of 1-bits continues until the start bit of the next transmit data. 3. the sci checks the tdre bit when it outputs the stop bit. if tdre is 0, the sci loads new data from sctdr into sctsr, outputs the stop bit, then begins serial transmission of the next frame. if tdre is 1, the sci sets the tend bit to 1 in scssr, outputs the stop bit, then continues output of 1-bits (marking). if the transmit-end interrupt enable bit (teie) in scscr is set to 1, a transmit-end interrupt (tei) is requested. 505 figure 15.9 shows an example of sci transmit operation in asynchronous mode. 01 1 1 0/1 0 1 tdre tend parity bit parity bit serial data start bit data stop bit start bit data stop bit idle (mark) state txi interrupt request generated tei interrupt request generated txi interrupt handler writes data to sctdr and clears tdre bit to 0 1 frame d 0 d 1 d 7 d 0 d 1 d 7 0/1 txi interrupt request generated figure 15.9 example of sci transmit operation in asynchronous mode (8-bit data with parity and one stop bit) receiving serial data (asynchronous mode): figure 15.10 shows a sample flowchart for receiving serial data. the procedure for receiving serial data after enabling the sci for reception is: 1. receive error handling and break detection: if a receive error occurs, read the orer, per and fer bits in scssr to identify the error. after executing the necessary error handling, clear orer, per and fer to 0. receiving cannot resume if orer, per or fer remains set to 1. when a framing error occurs, the rxd pin can be read to detect the break state. 2. sci status check and receive-data read: read the serial status register (scssr), check that rdrf is set to 1, then read receive data from the receive data register (scrdr) and clear rdrf to 0. the rxi interrupt can also be used to determine if the rdrf bit has changed from 0 to 1. 3. to continue receiving serial data: read the rdrf and scrdr bits and clear rdrf to 0 before the stop bit of the current frame is received. 506 start of reception read orer, per, and fer bits in scssr all data received? end of reception no yes per fer orer = 1? rdrf = 1? yes yes clear re bit in scscr to 0 no no read rdrf bit in scssr error handling read receive data from scrdr and clear rdrf bit in scssr to 0 (1) (2) (3) note: numbers in parentheses refer to steps in the preceding procedure description. figure 15.10 sample flowchart for receiving serial data 507 error handling orer = 1? overrun error handling fer = 1? yes break? no framing error handling per = 1? yes parity error handling clear orer, per, and fer bits in scssr to 0 end no no no yes yes clear re bit in scscr to 0 figure 15.10 sample flowchart for receiving serial data (cont) 508 in receiving, the sci operates as follows: 1. the sci monitors the communication line. when it detects a start bit (0), the sci synchronizes internally and starts receiving. 2. receive data is shifted into scrsr in order from the lsb to the msb. 3. the parity bit and stop bit are received. after receiving these bits, the sci makes the following checks: a. parity check: the number of 1s in the receive data must match the even or odd parity setting of the o/ e bit in scsmr. b. stop bit check: the stop bit value must be 1. if there are two stop bits, only the first stop bit is checked. c. status check: rdrf must be 0 so that receive data can be loaded from scrsr into scrdr. if these checks all pass, the sci sets rdrf to 1 and stores the received data in scrdr. if one of the checks fails (receive error), the sci operates as indicated in table 15.12. note: when a receive error flag is set, further receiving is disabled. the rdrf bit is not set to 1. be sure to clear the error flags. 4. after setting rdrf to 1, if the receive-data-full interrupt enable bit (rie) is set to 1 in scscr, the sci requests a receive-data-full interrupt (rxi). if one of the error flags (orer, per, or fer) is set to 1 and the receive-data-full interrupt enable bit (rie) in scscr is also set to 1, the sci requests a receive-error interrupt (eri). table 15.12 receive error conditions and sci operation receive error abbreviation condition data transfer overrun error orer receiving of next data ends while rdrf is still set to 1 in scssr receive data not loaded from scrsr into scrdr framing error fer stop bit is 0 receive data loaded from scrsr into scrdr parity error per parity of receive data differs from even/odd parity setting in scsmr receive data loaded from scrsr into scrdr 509 figure 15.11 shows an example of sci receive operation in asynchronous mode. rdrf fer eri interrupt request generated by framing error 1 frame rxi interrupt handler reads data and clears rdrf bit to 0 rxi interrupt request generated 01 1 1 0/1 0 1 parity bit parity bit serial data start bit data stop bit start bit data stop bit idle (mark) state d 0 d 1 d 7 d 0 d 1 d 7 0/1 figure 15.11 example of sci receive operation (8-bit data with parity and one stop bit) 15.3.3 multiprocessor communication the multiprocessor communication function enables several processors to share a single serial communication line. the processors communicate in asynchronous mode using a format with an additional multiprocessor bit (multiprocessor format). in multiprocessor communication, each receiving processor is addressed by a unique id. a serial communication cycle consists of an id-sending cycle that identifies the receiving processor, and a data-sending cycle. the multiprocessor bit distinguishes id-sending cycles from data-sending cycles. the transmitting processor starts by sending the id of the receiving processor with which it wants to communicate as data with the multiprocessor bit set to 1. next the transmitting processor sends transmit data with the multiprocessor bit cleared to 0. receiving processors skip incoming data until they receive data with the multiprocessor bit set to 1. when they receive data with the multiprocessor bit set to 1, receiving processors compare the data with their ids. the receiving processor with a matching id continues to receive further incoming data. processors with ids not matching the received data skip further incoming data until they again receive data with the multiprocessor bit set to 1. multiple processors can send and receive data in this way. figure 15.12 shows an example of communication among processors using the multiprocessor format. 510 receiving station a (id = 01) (id = 02) (id = 03) (id = 04) receiving station b receiving station c serial communication line h'01 h'aa (mpb = 0) (mpb = 1) id transmit cycle: specifies receiving station serial data transmitting station receiving station d data transmit cycle: data transmission to receiving station specified by id mpb: multiprocessor bit figure 15.12 communication among processors using multiprocessor format (sending data h'aa to receiving processor a) communication formats: four formats are available. parity-bit settings are ignored when the multiprocessor format is selected. for details see table 15.11. clock: see the description in the asynchronous mode section. transmitting multiprocessor serial data: figure 15.13 shows a sample flowchart for transmitting multiprocessor serial data. the procedure for transmitting multiprocessor serial data is: 1. sci status check and transmit data write: read the serial status register (scssr), check that the tdre bit is 1, then write transmit data in the transmit data register (sctdr). also set mpbt (multiprocessor bit transfer) to 0 or 1 in scssr. finally, clear tdre to 0. 2. to continue transmitting serial data: read the tdre bit to check whether it is safe to write (if it reads 1); if so, write data in sctdr, then clear tdre to 0. 3. to output a break at the end of serial transmission: set the port sc data register (scpdr) and port sc control register (scpcr), then clear the te bit to 0 in the serial control register (scscr). for scpcr and scpdr settings, see section 15.2.8, port sc control register (scpcr)/port sc data register (scpdr). 511 tdre = 1? write transmit data to sctdr and set mpbt bit in scssr transmission ended? yes tend = 1? read tend bit in scssr clear tdre bit to 0 break output? yes clear te bit scscr to 0 end of transmission yes read tdre bit in scssr no no yes no no (1) (2) (3) start of transmission set scpdr and scpcr note: numbers in parentheses refer to steps in the preceding procedure description. figure 15.13 sample flowchart for transmitting multiprocessor serial data 512 in transmitting serial data, the sci operates as follows: 1. the sci monitors the tdre bit in scssr. when tdre is cleared to 0 the sci recognizes that the transmit data register (sctdr) contains new data, and loads this data from sctdr into the transmit shift register (sctsr). 2. after loading the data from sctdr into sctsr, the sci sets the tdre bit to 1 and starts transmitting. if the transmit-data-empty interrupt enable bit (tie) in scscr is set to 1, the sci requests a transmit-data-empty interrupt (txi) at this time. serial transmit data is transmitted in the following order from the txd pin: a. start bit: one 0-bit is output. b. transmit data: seven or eight bits are output, lsb first. c. multiprocessor bit: one multiprocessor bit (mpbt value) is output. d. stop bit: one or two 1-bits (stop bits) are output. e. marking: output of 1-bits continues until the start bit of the next transmit data. 3. the sci checks the tdre bit when it outputs the stop bit. if tdre is 0, the sci loads data from sctdr into sctsr, outputs the stop bit, then begins serial transmission of the next frame. if tdre is 1, the sci sets the tend bit in scssr to 1, outputs the stop bit, then continues output of 1 bits in the mark state. if the transmit-end interrupt enable bit (teie) in scscr is set to 1, a transmit-end interrupt (tei) is requested at this time. figure 15.14 shows sci transmission with a multiprocessor format. 513 tdre tend txi interrupt request generated tei interrupt request generated writes data to tdr with the txi interrupt processing routine and clears tdre bit to 0 1 frame 01 1 1 0/1 0 1 multi- processor bit serial data start bit data stop bit start bit data stop bit idle (mark) state d 0 d 1 d 7 d 0 d 1 d 7 0/1 multi- processor bit txi interrupt request generated figure 15.14 example of sci multiprocessor transmit operation (8-bit data with multiprocessor bit and one stop bit) receiving multiprocessor serial data: figure 15.15 shows a sample flowchart for receiving multiprocessor serial data. the procedure for receiving multiprocessor serial data is: 1. id receive cycle: set the mpie bit in the serial control register (scscr) to 1. 2. sci status check and compare to id reception: read the serial status register (scssr), check that rdrf is set to 1, then read data from the receive data register (scrdr) and compare with the processors own id. if the id does not match the receive data, set mpie to 1 again and clear rdrf to 0. if the id matches the receive data, clear rdrf to 0. 3. sci status check and data receiving: read scssr, check that rdrf is set to 1, then read data from the receive data register (scrdr). 4. receive error handling and break detection: if a receive error occurs, read the orer and fer bits in scssr to identify the error. after executing the necessary error handling, clear both orer and fer to 0. receiving cannot resume if orer or fer remain set to 1. when a framing error occurs, the rxd pin can be read to detect the break state. 514 rdrf = 1? fer = 1 or orer = 1? rdrf = 1? all data received? no end of reception yes set mpie bit in scscr to 1 read rdrf bit in scssr clear re bit in scscr to 0 no no read orer and fer bits in scssr fer = 1 or orer = 1? read rdrf bit in scssr read receive data from scrdr is id the stations id? yes read orer and fer bits in sscsr no error handling yes yes yes no start of reception no yes read receive data from scrdr (1) (2) (4) (3) note: numbers in parentheses refer to steps in the preceding procedure description. figure 15.15 sample flowchart for receiving multiprocessor serial data 515 orer = 1? break? yes framing error handling yes error handling overrun error handling yes fer = 1? clear orer and fer bits in scssr to 0 end no no no clear re bit in scscr to 0 figure 15.15 sample flowchart for receiving multiprocessor serial data (cont) 516 figure 15.16 shows an example of sci receive operation using a multiprocessor format. rdrf mpie rdr value id1 rxi interrupt request (multiprocessor interrupt) generated, mpie = 0 example: own id does not match data rxi interrupt handler reads rdr data and clears rdrf bit to 0 id is not stations id, so mpie bit is set to 1 again no rxi interrupt generated; rdr state is maintained 01 1 1 10 1 stop bit mpb serial data start bit data (id1) data (data 1) start bit mpb stop bit idle (mark) state d 0 d 1 d 7 d 0 d 1 d 7 0 figure 15.16 example of sci receive operation (8-bit data with multiprocessor bit and one stop bit) 517 rdrf mpie rdr value example: own id matches data id1 id2 data2 01 1 1 10 1 mpb mpb serial data start bit data (id2) data (data 2) stop bit start bit stop bit idle (mark) state d 0 d 1 d 7 d 0 d 1 d 7 0 rxi interrupt request (multiprocessor interrupt) generated, mpie = 0 rxi interrupt handler reads rdr data and clears rdrf bit to 0 id is that of station, so reception continues unchanged and data is received by rxi interrupt handler mpie bit set to 1 again figure 15.16 example of sci receive operation (cont) (8-bit data with multiprocessor bit and one stop bit) 518 15.3.4 synchronous operation in synchronous mode, the sci transmits and receives data in synchronization with clock pulses. this mode is suitable for high-speed serial communication. the sci transmitter and receiver are independent, so full-duplex communication is possible while sharing the same clock. the transmitter and receiver are also double buffered, so continuous transmitting or receiving is possible by reading or writing data while transmitting or receiving is in progress. figure 15.17 shows the general format in synchronous serial communication. bit 0 bit 1 bit 2 bit 3 bit 4 bit 5 bit 6 bit 7 lsb msb serial clock serial data ** one unit of communication data (character or frame) note: * high except in continuous transmitting or receiving figure 15.17 data format in synchronous communication in synchronous serial communication, each data bit is output on the communication line from one falling edge of the serial clock to the next. data is guaranteed valid at the rising edge of the serial clock. in each character, the serial data bits are transmitted in order from the lsb (first) to the msb (last). after output of the msb, the communication line remains in the state of the msb. in synchronous mode, the sci transmits or receives data by synchronizing with the falling edge of the serial clock. communication format: the data length is fixed at eight bits. no parity bit or multiprocessor bit can be added. 519 clock: an internal clock generated by the on-chip baud rate generator or an external clock input from the sck pin can be selected as the sci transmit/receive clock. the clock source is selected by the c/ a bit in the serial mode register (scsmr) and bits cke1 and cke0 in the serial control register (scscr). see table 15.10. when the sci operates on an internal clock, it outputs the clock signal at the sck pin. eight clock pulses are output per transmitted or received character. when the sci is not transmitting or receiving, the clock signal remains in the high state. when only receiving, the sci receives in 2- character units, so a 16-pulse serial clock is output. to receive in 1-character units, select an external clock source. transmitting and receiving data sci initialization (synchronous mode): before transmitting, receiving, or changing the mode or communication format, the software must clear the te and re bits to 0 in the serial control register (scscr), then initialize the sci. clearing te to 0 sets tdre to 1 and initializes the transmit shift register (sctsr). clearing re to 0, however, does not initialize the rdrf, per, fer, and orer flags and receive data register (scrdr), which retain their previous contents. figure 15.18 shows a sample flowchart for initializing the sci. the procedure for initializing the sci is: 1. select the clock source in the serial control register (scscr). leave rie, tie, teie, mpie, te and re cleared to 0. 2. select the communication format in the serial mode register (scsmr). 3. write the value corresponding to the bit rate in the bit rate register (scbrr) (not necessary if an external clock is used). 4. wait for at least the interval required to transmit or receive one bit, then set te or re in the serial control register (scscr) to 1. also set rie, tie, teie and mpie. setting te and re allows use of the txd and rxd pins. 520 initialization clear te and re bits in scscr to 0 (1) has a 1-bit period elapsed? set te and re bits in scscr to 1 and set rie, tie, teie, and mpie bits set transmit/receive format in scsmr yes no set value in scbrr set rie, tie, teie, mpie, cke1, and cke0 bits in scscr (te and re are 0) end wait (2) (3) (4) note: numbers in parentheses refer to steps in the preceding procedure description. figure 15.18 sample flowchart for sci initialization transmitting serial data (synchronous mode): figure 15.19 shows a sample flowchart for transmitting serial data. the procedure for transmitting serial data is: 1. sci status check and transmit data write: read the serial status register (scssr), check that the tdre bit is 1, then write transmit data in the transmit data register (sctdr) and clear tdre to 0. 2. to continue transmitting serial data: read the tdre bit to check whether it is safe to write (if it reads 1); if so, write data in sctdr, then clear tdre to 0. 521 start of transmission read tdre bit in scssr all data transmitted? yes no end of transmission (1) (2) tdre = 1? write transmit data to sctdr and clear tdre bit in scssr to 0 yes no read tend bit in scssr tend = 1? yes no clear te bit in scscr to 0 note: numbers in parentheses refer to steps in the preceding procedure description. figure 15.19 sample flowchart for transmitting serial data 522 in transmitting serial data, the sci operates as follows: 1. the sci monitors the tdre bit in scssr. when tdre is cleared to 0 the sci recognizes that the transmit data register (sctdr) contains new data and loads this data from sctdr into the transmit shift register (sctsr). 2. after loading the data from sctdr into sctsr, the sci sets the tdre bit to 1 and starts transmitting. if the transmit-data-empty interrupt enable bit (tie) in scscr is set to 1, the sci requests a transmit-data-empty interrupt (txi) at this time. if clock output mode is selected, the sci outputs eight synchronous clock pulses. if an external clock source is selected, the sci outputs data in synchronization with the input clock. data is output from the txd pin in order from the lsb (bit 0) to the msb (bit 7). 3. the sci checks the tdre bit when it outputs the msb (bit 7). if tdre is 0, the sci loads data from sctdr into sctsr, then begins serial transmission of the next frame. if tdre is 1, the sci sets the tend bit in scssr to 1, transmits the msb, then holds the transmit data pin (txd) in the msb state. if the transmit-end interrupt enable bit (teie) in scscr is set to 1, a transmit-end interrupt (tei) is requested at this time. 4. after the end of serial transmission, the sck pin is held in the high state. figure 15.20 shows an example of sci transmit operation. bit 0 bit 1 bit 7 bit 0 bit 1 bit 6 serial clock serial data transfer direction bit 7 txi interrupt handler writes data to tdr and clears tdre bit to 0 1 frame tdre tend lsb msb txi interrupt request generated txi interrupt request generated tei interrupt request generated figure 15.20 example of sci transmit operation 523 receiving serial data (synchronous mode): figure 15.21 shows a sample flowchart for receiving serial data. when switching from asynchronous mode to synchronous mode, make sure that orer, per, and fer are cleared to 0. if per or fer is set to 1, the rdrf bit will not be set and both transmitting and receiving will be disabled. the procedure for receiving serial data is: 1. receive error handling: if a receive error occurs, read the orer bit in scssr to identify the error. after executing the necessary error handling, clear orer to 0. transmitting/receiving cannot resume if orer remains set to 1. 2. sci status check and receive data read: read the serial status register (scssr), check that rdrf is set to 1, then read receive data from the receive data register (scrdr) and clear rdrf to 0. the rxi interrupt can also be used to determine if the rdrf bit has changed from 0 to 1. 3. to continue receiving serial data: read scrdr, and clear rdrf to 0 before the msb (bit 7) of the current frame is received. 524 read orer bit in scssr all data received? end of reception no yes orer = 1? rdrf = 1? yes clear re bit in scscr to 0 no no read rdrf bit in scssr (3) (2) yes error handling (1) read receive data from scrdr and clear rdrf bit in scssr to 0 start of reception note: numbers in parentheses refer to steps in the preceding procedure description. figure 15.21 sample flowchart for receiving serial data 525 error handling end orer = 1? no clear orer bit in scssr to 0 yes overrun error handling figure 15.21 sample flowchart for receiving serial data (cont) in receiving, the sci operates as follows: 1. the sci synchronizes with serial clock input or output and initializes internally. 2. receive data is shifted into scrsr in order from the lsb to the msb. after receiving the data, the sci checks that rdrf is 0 so that receive data can be loaded from scrsr into scrdr. if this check is passed, the sci sets rdrf to 1 and stores the received data in scrdr. if the check is not passed (receive error), the sci operates as indicated in table 15.12. this state prevents further transmission or reception. while receiving, the rdrf bit is not set to 1. be sure to clear the error flag. 3. after setting rdrf to 1, if the receive-data-full interrupt enable bit (rie) is set to 1 in scscr, the sci requests a receive-data-full interrupt (rxi). if the orer bit is set to 1 and the receive- data-full interrupt enable bit (rie) in scscr is also set to 1, the sci requests a receive-error interrupt (eri). figure 15.22 shows an example of sci receive operation. 526 bit 7 bit 0 bit 7 bit 0 bit 1 bit 6 serial clock serial data transfer direction bit 7 rxi interrupt handler reads data and clears rdrf bit to 0 1 frame rxi interrupt request generated rxi interrupt request generated eri interrupt request generated by overrun error rdrf orer figure 15.22 example of sci receive operation transmitting and receiving serial data simultaneously (synchronous mode): figure 15.23 shows a sample flowchart for transmitting and receiving serial data simultaneously. the procedure for setting the sci to transmit and receive serial data simultaneously is: 1. sci status check and transmit data write: read the serial status register (scssr), check that the tdre bit is 1, then write transmit data in the transmit data register (sctdr) and clear tdre to 0. the txi interrupt can also be used to determine if the tdre bit has changed from 0 to 1. 2. receive error handling: if a receive error occurs, read the orer bit in scssr to identify the error. after executing the necessary error handling, clear orer to 0. transmitting/receiving cannot resume if orer remains set to 1. 3. sci status check and receive data read: read the serial status register (scssr), check that rdrf is set to 1, then read receive data from the receive data register (scrdr) and clear rdrf to 0. the rxi interrupt can also be used to determine if the rdrf bit has changed from 0 to 1. 4. to continue transmitting and receiving serial data: read the rdrf bit and scrdr, and clear rdrf to 0 before the msb (bit 7) of the current frame is received. also read the tdre bit to check whether it is safe to write (if it reads 1); if so, write data in sctdr, then clear tdre to 0 before the msb (bit 7) of the current frame is transmitted. 527 start of transmission/reception read tdre bit in scssr all data transmitted/received? end of transmission/reception (1) no yes tdre = 1? write transmit data to sctdr and clear tdre bit in scssr to 0 rdrf = 1? no yes yes no read orer bit in scssr error processing (2) orer = 1? no read rdrf bit in scssr (4) yes (3) read receive data from scrdr and clear rdrf bit in scssr to 0 clear te and re bits in scscr to 0 notes: 1. 2. numbers in parentheses refer to steps in the preceding procedure description. in switching from transmitting or receiving to simultaneous transmitting and receiving, clear both te and re to 0, then set both te and re to 1 simultaneously. figure 15.23 sample flowchart for transmitting/receiving serial data 528 15.4 sci interrupts the sci has four interrupt sources transmit-end (tei), receive-error (eri), receive-data-full (rxi), and transmit-data-empty (txi). table 15.13 lists the interrupt sources and indicates their priority. these interrupts can be enabled and disabled by the tie, rie, and teie bits in the serial control register (scscr). each interrupt request is sent separately to the interrupt controller. txi is requested when the tdre bit in scssr is set to 1. tdre is automatically cleared to 0 when data is written in the transmit data register (sctdr). rxi is requested when the rdrf bit in scssr is set to 1. rdrf is automatically cleared to 0 when the receive data register (scrdr) is read. eri is requested when the orer, per, or fer bit in scssr is set to 1. tei is requested when the tend bit in scssr is set to 1. where the txi interrupt indicates that transmit data writing is enabled, the tei interrupt indicates that the transmit operation is complete. table 15.13 sci interrupt sources interrupt source description priority when reset is cleared eri receive error (orer, per, or fer) high rxi receive data full (rdrf) txi transmit data empty (tdre) tei transmit end (tend) low see section 4, exception handling, for priorities and the relationship to non-sci interrupts. 529 15.5 usage notes note the following points when using the sci. sctdr writing and tdre flag: the tdre bit in the serial status register (scssr) is a status flag indicating loading of transmit data from sctdr into sctsr. the sci sets tdre to 1 when it transfers data from sctdr to sctsr. data can be written to sctdr regardless of the tdre bit state. if new data is written in sctdr when tdre is 0, however, the old data stored in sctdr will be lost because the data has not yet been transferred to sctsr. before writing transmit data to sctdr, be sure to check that tdre is set to 1. simultaneous multiple receive errors: table 15.14 indicates the state of scssr status flags when multiple receive errors occur simultaneously. when an overrun error occurs, the scrsr contents cannot be transferred to scrdr, so receive data is lost. table 15.14 scssr status flags and transfer of receive data scssr status flags receive data transfer receive error status rdrf orer fer per scrsr scrdr overrun error 1 1 0 0 x framing error 0 0 1 0 o parity error 0 0 0 1 o overrun error + framing error 1 1 1 0 x overrun error + parity error 1 1 0 1 x framing error + parity error 0 0 1 1 o overrun error + framing error + parity error 1 1 1 1 x x: receive data is not transferred from scrsr to scrdr. o: receive data is transferred from scrsr to scrdr. break detection and processing: break signals can be detected by reading the rxd pin directly when a framing error (fer) is detected. in the break state, the input from the rxd pin consists of all 0s, so fer is set and the parity error flag (per) may also be set. in the break state, the sci receiver continues to operate, so if the fer bit is cleared to 0, it will be set to 1 again. sending a break signal: the txd pin i/o condition and level can be determined by means of the scp0dt bit in the port sc data register (scpdr) and bits scp0md0 and scp0md1 in the port sc control register (scpcr). this feature can be used to send breaks. to send a break during serial transmission, clear the scp0dt bit to 0 (designating low level), then clear the te bit to 0 (halting transmission). when the te bit is cleared to 0, the transmitter is initialized regardless of the current transmission state, and 0 is output from the txd pin. 530 tend flag and te bit processing: the tend flag is set to 1 during transmission of the stop bit of the last data. consequently, if the te bit is cleared to 0 immediately after setting of the tend flag has been confirmed, the stop bit will be in the process of transmission and will not be transmitted normally. therefore, the te bit should not be cleared to 0 for at least 0.5 serial clock cycles (or 1.5 cycles if two stop bits are used) after setting of the tend flag is confirmed. receive error flags and transmitter operation (synchronous mode only): when a receive error flag (orer, per, or fer) is set to 1, the sci will not start transmitting even if tdre is set to 1. be sure to clear the receive error flags to 0 before starting to transmit. note that clearing re to 0 does not clear the receive error flags. receive data sampling timing and receive margin in asynchronous mode: in asynchronous mode, the sci operates on a base clock of 16 times the transfer rate frequency. in receiving, the sci synchronizes internally with the falling edge of the start bit, which it samples on the base clock. receive data is latched at the rising edge of the eighth base clock pulse (figure 15.24). 0 1 2 3 4 5 6 7 8 9 10111213 1415 0 1 2 3 4 5 6 7 8 9 10111213 1415 0 1 2 3 4 5 base clock receive data (rxd) synchro- nization sampling timing data sampling timing 8 clocks 16 clocks start bit C7.5 clocks +7.5 clocks d0 d1 figure 15.24 receive data sampling timing in asynchronous mode 531 the receive margin in asynchronous mode can therefore be expressed as in equation 1. equation 1: m = 0.5 C 1 2n d C 0.5 n C (l C 0.5)f C (1 + f) 100% where: m = receive margin ( % ) n = ratio of clock frequency to bit rate (n = 16) d = clock duty cycle (d = 0 to 1.0) l = frame length (l = 9 to 12) f = absolute deviation of clock frequency from equation 1, if f = 0 and d = 0.5, the receive margin is 46.875%, as in equation 2. equation 2: m = (0.5 C 1/(2 16)) 100 % = 46.875 % this is a theoretical value. a reasonable margin to allow in system designs is 20% to 30%. notes on synchronous external clock mode: ? set te = re = 1 only when external clock sck is 1. ? do not set te = re = 1 until at least four clocks after external clock sck has changed from 0 to 1. ? when receiving, rdrf is set to 1 when re is set to zero 2.5C3.5 clocks after the rising edge of the sck input of the d7 bit in rxd, but data cannot be copied to scrdr. note on synchronous internal clock mode: when receiving, rdrf is set to 1 when re is cleared to zero 1.5 clocks after the rising edge of the sck output of the d7 bit in rxd, but data cannot be copied to scrdr. 532 533 section 16 smart card interface 16.1 overview as an added serial communications interface function, the sci supports an ic card (smart card) interface that conforms to the iso/iec7816-3 (identification card) data transmission protocol format t = 0 (asynchronous full-duplex character transmission protocol). register settings are used to switch between the normal serial communication interface and the smart card interface. 16.1.1 features the smart card interface has the following features: ? asynchronous mode ? data length: 8 bits ? parity bit generation and check ? receive mode error signal detection (parity error) ? transmit mode error signal detection and automatic re-transmission of data ? supports both direct convention and inverse convention ? bit rate can be selected using on-chip baud rate generator. ? three types of interrupts: transmit-data-empty, receive-data-full, and communication-error interrupts are requested independently. 534 16.1.2 block diagram figure 16.1 shows a block diagram of the smart card interface. rxd txd sck sci scbrr scscr scsmr sctdr sctsr scrdr scrsr scscmr scssr parity generation parity check clock external clock module data bus internal data bus p p /4 p /16 p /64 txi rxi eri bus interface baud rate generator transmit/ receive control scscmr: scrsr: scrdr: sctsr: sctdr: scsmr: scscr: scssr: scbrr: legend smart card mode register receive shift register receive data register transmit shift register transmit data register serial mode register serial control register serial status register bit rate register figure 16.1 block diagram of smart card interface 535 16.1.3 pin configuration table 16.1 summarizes the smart card interface pins. table 16.1 smart card interface pins pin name abbreviation i/o function serial clock pin sck0 output clock output receive data pin rxd0 input receive data input transmit data pin txd0 output transmit data output 16.1.4 smart card interface registers table 16.2 summarizes the registers used by the smart card interface. the scsmr, scbrr, scscr, sctdr, and scrdr registers are the same as for the normal sci function. they are described in section 15, serial communication interface. table 16.2 registers name abbreviation r/w initial value* 3 address access size serial mode register scsmr r/w h'00 h'fffffe80 8 bit rate register scbrr r/w h'ff h'fffffe82 8 serial control register scscr r/w h'00 h'fffffe84 8 transmit data register sctdr r/w h'ff h'fffffe86 8 serial status register scssr r/(w)* 1 h'84 h'fffffe88 8 receive data register scrdr r h'00 h'fffffe8a 8 smart card mode register scscmr r/w h'00* 2 h'fffffe8c 8 notes: *1 only 0 can be written, to clear the flags. *2 bits 0, 2, and 3 are cleared. the value of the other bits is undefined. *3 initialized by a power-on or manual reset. 536 16.2 register descriptions this section describes the registers added for the smart card interface and the bits whose functions are changed. 16.2.1 smart card mode register (scscmr) the smart card mode register (scscmr) is an 8-bit readable/writable register that selects smart card interface functions. scscmr bits 0, 2, and 3 are initialized to h'00 by a reset and in standby mode. bit: 7 6 5 4 3 2 1 0 sdir sinv smif initial value: 0 0 0 r/w: r r r r r/w r/w r r/w bits 7 to 4 and 1reserved: these bits are always read as 0. the write value should always be 0. bit 3smart card data transfer direction (sdir): selects the serial/parallel conversion format. bit 3: sdir description 0 contents of sctdr are transferred lsb-first, and receive data is stored in scrdr lsb-first (initial value) 1 contents of sctdr are transferred msb-first, and receive data is stored in scrdr msb-first bit 2smart card data inversion (sinv): specifies whether to invert the logic level of the data. this function is used in combination with bit 3 for transmitting and receiving with an inverse convention card. sinv does not affect the logic level of the parity bit. see section 16.3.4, register settings, for information on how parity is set. bit 2: sinv description 0 contents of sctdr are transferred unchanged, and receive data is stored in scrdr unchanged (initial value) 1 contents of sctdr are inverted before transfer, and receive data is inverted before storage in scrdr 537 bit 0smart card interface mode select (smif): enables the smart card interface function. bit 0 : smif description 0 smart card interface function disabled (initial value) 1 smart card interface function enabled 16.2.2 serial status register (scssr) in smart card interface mode, the function of scssr bit 4 is changed. the setting conditions for bit 2, the tend bit, are also changed. bit: 7 6 5 4 3 2 1 0 tdre rdrf orer fer/ers per tend mpb mpbt initial value: 1 0 0 0 0 1 0 0 r/w: r/(w)* r/(w)* r/(w)* r/(w)* r/(w)* r r r/w note: * only 0 can be written, to clear the flag. bits 7 to 5: these bits have the same function as in the ordinary sci. see section 15, serial communication interface (sci), for more information. bit 4error signal status (ers): in the smart card interface mode, bit 4 indicates the state of the error signal returned from the receiving side during transmission. the smart card interface cannot detect framing errors. bit 4: ers description 0 receiving ended normally with no error signal (initial value) ers is cleared to 0 when the chip is reset or enters standby mode, or when software reads ers after it has been set to 1, then writes 0 to ers. 1 an error signal indicating a parity error was transmitted from the receiving side ers is set to 1 if the error signal sampled is low. note: the ers flag maintains its state even when the te bit in scscr is cleared to 0. 538 bits 3 to 0: these bits have the same function as in the ordinary sci. see section 15, serial communication interface (sci), for more information. the setting conditions for bit 2, the transmit end bit (tend), are changed as follows. bit 2: tend description 0 transmission is in progress tend is cleared to 0 when software reads tdre after it has been set to 1, then writes 0 to tdre. 1 end of transmission (initial value) tend is set to 1 when: ? the chip is reset or enters standby mode, ? the te bit in scscr is 0 and the fer/ers bit is also 0, ? the c/ a bit in scsmr is 0, and tdre = 1 and fer/ers = 0 (normal transmission) 2.5 etu after a one-byte serial character is transmitted, or ? the c/ a bit in scsmr is 1, and tdre = 1 and fer/ers = 0 (normal transmission) 1.0 etu after a one-byte serial character is transmitted. note: etu is an abbreviation of elementary time unit, which is the period for the transfer of 1 bit. 16.3 operation 16.3.1 overview the primary functions of the smart card interface are described below. 1. each frame consists of 8-bit data and 1 parity bit. 2. during transmission, the card leaves a guard time of at least 2 etu (elementary time units: the period for 1 bit to transfer) from the end of the parity bit to the start of the next frame. 3. during reception, the card outputs an error signal low level for 1 etu after 10.5 etu has elapsed from the start bit if a parity error was detected. 4. during transmission, it automatically transmits the same data after allowing at least 2 etu from the time the error signal is sampled. 5. only start-stop type asynchronous communication functions are supported; no synchronous communication functions are available. 539 16.3.2 pin connections figure 16.2 shows the pin connection diagram for the smart card interface. during communication with an ic card, transmission and reception are both carried out over the same data transfer line, so connect the txd and rxd pins on the chip. pull up the data transfer line to the power supply v cc side with a resistor. when using the clock generated by the smart card interface on an ic card, input the sck pin output to the ic cards clk pin. this connection is not necessary when the internal clock is used on the ic card. use the chips port output as the reset signal. apart from these pins, power and ground pin connections are usually also required. note: when the ic card is not connected and both re and te are set to 1, closed communication is possible and auto-diagnosis can be performed. lsi txd io clk rst rxd sck px (port) clock line data line reset line ic card connected device v cc figure 16.2 pin connection diagram for smart card interface 540 16.3.3 data format figure 16.3 shows the data format for the smart card interface. in this mode, parity is checked every frame while receiving and error signals sent to the transmitting side whenever an error is detected so that data can be re-transmitted. during transmission, error signals are sampled and data re-transmitted whenever an error signal is detected. ds d0 d1 d2 d3 d4 d5 d6 d7 dp with no parity error transmitting station output ds d0 d1 d2 d3 d4 d5 d6 d7 dp de with parity error transmitting station output receiving station output ds: d0Cd7: dp: de: start bit data bits parity bit error signal figure 16.3 data format for smart card interface the operating sequence is: 1. the data line is high-impedance when not in use and is fixed high with a pull-up resistor. 2. the transmitting side starts one frame of data transmission. the data frame starts with a start bit (ds, low level). the start bit is followed by eight data bits (d0Cd7) and a parity bit (dp). 3. on the smart card interface, the data line returns to high-impedance after this. the data line is pulled high with a pull-up resistor. 4. the receiving side checks parity. when the data is received normally with no parity errors, the receiving side then waits to receive the next data. when a parity error occurs, the receiving side outputs an error signal (de, low level) and requests re-transfer of data. the receiving station returns the signal line to high-impedance after outputting the error signal for a specified period. the signal line is pulled high with a pull-up resistor. 541 5. the transmitting side transmits the next frame of data unless it receives an error signal. if it does receive an error signal, it returns to step 2 to re-transmit the erroneous data. 16.3.4 register settings table 16.3 shows the bit map of the registers that the smart card interface uses. bits shown as 1 or 0 must be set to the indicated value. the settings for the other bits are described below. table 16.3 register settings for smart card interface register address bit 7 bit 6 bit 5 bit 4 bit 3 bit 2 bit 1 bit 0 scsmr h'fffffe80 c/ a 01o/ e 1 0 cks1 cks0 scbrr h'fffffe82 brr7 brr6 brr5 brr4 brr3 brr2 brr1 brr0 scscr h'fffffe84 tie rie te re 0 0 cke1 cke0 sctdr h'fffffe86 tdr7 tdr6 tdr5 tdr4 tdr3 tdr2 tdr1 tdr0 scssr h'fffffe88 tdre rdrf orer fer/ ers per tend 0 0 scrdr h'fffffe8a rdr7 rdr6 rdr5 rdr4 rdr3 rdr2 rdr1 rdr0 scscmr h'fffffe8c sdir sinv smif note: dashes indicate unused bits. 1. setting the serial mode register (scsmr): the c/ a bit selects the setting timing of the tend flag, and selects the clock output state in combination with bits cke1 and cke0 in the serial control register (scscr). clear the o/ e bit to 0 if the ic card uses the direct convention, and set it to 1 if the card uses the inverse convention. select the on-chip baud rate generator clock source with the cks1 and cks0 bits (see section 16.3.5, clock). 2. setting the bit rate register (scbrr): set the bit rate. see section 16.3.5, clock, to see how to calculate the set value. 3. setting the serial control register (scscr): the tie, rie, te and re bits function as they do for the ordinary sci. see section 15, serial communication interface (sci), for more information. the cke0 bit specifies the clock output. when no clock is output, clear cke0 to 0; when a clock is output, set cke0 to 1. 4. setting the smart card mode register (scscmr): the sdir and sinv bits are both cleared to 0 for ic cards that use the direct convention, and both set to 1 when the inverse convention is used. the smif bit is set to 1 for the smart card interface. figure 16.4 shows sample waveforms for register settings of the two types of ic cards (direct convention and inverse convention) and their start characters. in the direct convention type, the logical 1 level is state z, the logical 0 level is state a, and communication is lsb-first. the start character data is h'3b. parity is even (from the smart card standard), and so the parity bit is 1. 542 in the inverse convention type, the logical 1 level is state a, the logical 0 level is state z, and communication is msb first. the start character data is h'3f. parity is even (from the smart card standard), and so the parity bit is 0, which corresponds to state z. only data bits d7Cd0 are inverted by the sinv bit. to invert the parity bit, set the o/ e bit in scsmr to odd parity mode. this applies to both transmission and reception. ds d0 d1 d2 d3 d4 d5 d6 d7 dp a (z) z z a z z z a a z (z) state a. direct convention (sdir, sinv, and o/ e are all 0) ds d7 d6 d5 d4 d3 d2 d1 d0 dp a (z) z z a a a a a a z (z) state b. inverse convention (sdir, sinv, and o/ e are all 1) figure 16.4 waveform of start character 16.3.5 clock only the internal clock generated by the on-chip baud rate generator can be used as the communication clock in the smart card interface. the bit rate for the clock is set by the bit rate register (scbrr) and the cks1 and cks0 bits in the serial mode register (scsmr), and is calculated using the equation below. table 16.5 shows sample bit rates. if clock output is then selected by setting cke0 to 1, a clock with a frequency 372 times the bit rate is output from the sck0 pin. b = 10 6 1488 2 2nC1 (n + 1) p where: n = value set in scbrr (0 n 255) b = bit rate (bits/s) p = peripheral module operating frequency (mhz) n = 0 to 3 (table 16.4) 543 table 16.4 relationship of n to cks1 and cks0 n cks1 cks0 000 101 210 311 table 16.5 examples of bit rate b (bits/s) for scbrr settings (n = 0) p (mhz) n 7.1424 10.00 10.7136 13.00 14.2848 16.00 18.00 0 9600.0 13440.9 14400.0 17473.1 19200.0 21505.4 24193.5 1 4800.0 6720.4 7200.0 8736.6 9600.0 10752.7 12096.8 2 3200.0 4480.3 4800.0 5824.4 6400.0 7168.5 8064.5 note: the bit rate is rounded to one decimal place. calculate the value to be set in the bit rate register (scbrr) from the operating frequency and the bit rate. n is an integer in the range 0 n 255, specifying a smallish error. n = 10 6 C 1 1488 2 2nC1 b p table 16.6 examples of scbrr settings for bit rate b (bits/s) (n = 0) (mhz) (9600 bits/s) 7.1424 10.00 10.7136 13.00 14.2848 16.00 18.00 n error n error n error n error n error n error n error 0 0.00 1 30.00 1 25.00 1 8.99 1 0.00 1 12.01 2 15.99 544 table 16.7 maximum bit rates for frequencies (smart card interface mode) p (mhz) maximum bit rate (bits/s) n n 7.1424 9600 0 0 10.00 13441 0 0 10.7136 14400 0 0 13.00 17473 0 0 14.2848 19200 0 0 16.00 21505 0 0 18.00 24194 0 0 the bit rate error is found as follows: error (%) = ( 10 6 C 1) 100 1488 2 2nC1 b (n + 1) p table 16.8 shows the relationship between transmit/receive clock register set values and output states on the smart card interface. table 16.8 register set values and sck pin register value sck pin setting smif c/ a cke1 cke0 output state 1* 1 1000 port determined by setting of port register scp1md1 and scp1md0 bits 1001 sck (serial clock) output state 2* 2 1100 low output low output state 1101 sck (serial clock) output state 3* 2 1110 high output high output state 1111 sck (serial clock) output state notes: *1 the sck output state changes as soon as the cke0 bit is modified. the cke1 bit should be cleared to 0. *2 the clock duty remains constant despite stopping and starting of the clock by modification of the cke0 bit. 545 16.3.6 data transmission and reception initialization: initialize the sci using the following procedure before sending or receiving data. initialization is also required for switching from transmit mode to receive mode or from receive mode to transmit mode. figure 16.5 shows a flowchart of the initialization process. 1. clear te and re in the serial control register (scscr) to 0. 2. clear error flags fer/ers, per, and orer to 0 in the serial status register (scssr). 3. set the c/ a bit, parity bit (o/ e bit), and baud rate generator select bits (cks1 and cks0 bits) in the serial mode register (scsmr). at this time also clear the chr and mp bits to 0 and set the stop and pe bits to 1. 4. set the smif, sdir, and sinv bits in the smart card mode register (scscmr). when the smif bit is set to 1, the txd and rxd pins both switch from ports to sci pins and become high-impedance. 5. set the value corresponding to the bit rate in the bit rate register (scbrr). 6. set the clock source select bits (cke1 and cke0 bits) in the serial control register (scscr). clear the tie, rie, te, re, mpie, and teie bits to 0. when the cke0 bit is set to 1, a clock is output from the sck pin. 7. after waiting at least 1 bit, set the tie, rie, te, and re bits in scscr. do not set the te and re bits simultaneously unless performing auto-diagnosis. 546 initialization clear te and re bits in scscr to 0 set value in scbrr clear fer/ers, per and orer flags in scssr to 0 wait set tie, rie, te, and re bits in scscr has a 1-bit interval elapsed? end (2) set parity in o/ e bit, set clock in cks1 and cks0 bits, and set c/ a , in scsmr (3) set clock in cke1 and cke0 bits, and clear tie, rie, te, re, mpie, and teie bits to 0, in scscr (6) (5) (4) (1) (7) no yes set smif, sdir, and sinv bits in scsmr note: numbers in parentheses refer to steps in the preceding procedure description. figure 16.5 initialization flowchart (example) 547 serial data transmission: the processing procedures in the smart card mode differ from ordinary sci processing because data is retransmitted when an error signal is sampled during a data transmission. this results in the transmission processing flowchart shown in figure 16.6. 1. initialize the smart card interface mode as described in initialization above. 2. check that the fer/ers bit in scssr is cleared to 0. 3. repeat steps 2 and 3 until the tend flag in scssr is set to 1. 4. write the transmit data into sctdr, clear the tdre flag to 0 and start transmitting. the tend flag will be cleared to 0. 5. to transmit more data, return to step 2. 6. to end transmission, clear the te bit to 0. this processing can be interrupted. when the tie bit is set to 1 and interrupt requests are enabled, a transmit-data-empty interrupt (txi) will be requested when the tend flag is set to 1 at the end of transmission. when the rie bit is set to 1 and interrupt requests are enabled, a communication error interrupt (eri) will be requested when the ers flag is set to 1 when an error occurs in transmission. see interrupt operation below for more information. 548 start end of transmission start of transmission initialize write transmit data in sctdr and clear tdre flag in scssr to 0 (1) clear te bit in scscr to 0 (6) error handling (2) fer/ers = 0? tend = 1? yes yes yes yes no no all data transmitted? no tend = 1? no error handling fer/ers = 0? yes no (4) (5) (3) note: numbers in parentheses refer to steps in the preceding procedure description. figure 16.6 transmission flowchart 549 serial data reception: the processing procedures in smart card mode are the same as in ordinary sci processing. the reception processing flowchart is shown in figure 16.7. 1. initialize the smart card interface mode as described above in initialization and in figure 16.5. 2. check that the orer and per flags in scssr are cleared to 0. if either flag is set, clear both to 0 after performing the appropriate error handling procedures. 3. repeat steps 2 and 3 until the rdrf flag is set to 1. 4. read the receive data from scrdr. 5. to receive more data, clear the rdrf flag to 0 and return to step 2. 6. to end reception, clear the re bit to 0. this processing can be interrupted. when the rie bit is set to 1 and interrupt requests are enabled, a receive-data-full interrupt (rxi) will be requested when the rdrf flag is set to 1 at the end of reception. when an error occurs during reception and either the orer or per flag is set to 1, a communication error interrupt (eri) will be requested. see interrupt operation below for more information. the received data will be transferred to scrdr even when a parity error occurs during reception and per is set to 1, so this data can still be read. 550 start end of reception start of reception initialize write receive data from scrdr and clear rdrf flag in scssr to 0 (1) clear re bit in scscr to 0 (6) error handling (2) orer = 0 or per = 0? rdrf = 1? yes yes yes no no all data received? no (4) (5) (3) note: numbers in parentheses refer to steps in the preceding procedure description. figure 16.7 reception flowchart (example) 551 switching modes: when switching from receive mode to transmit mode, check that the receive operation is completed before starting initialization, clearing re to 0, and setting te to 1. the rdrf, per, and orer flags can be used to check if reception is completed. when switching from transmit mode to receive mode, check that the transmit operation is completed before starting initialization, clearing te to 0, and setting re to 1. the tend flag can be used to check if transmission is completed. interrupt operation: in the smart card interface mode, there are three types of interrupts: transmit-data-empty (txi), communication error (eri) and receive-data-full (rxi). in this mode, the transmit-end interrupt (tei) cannot be requested. set the tend flag in scssr to 1 to request a txi interrupt. set the rdrf flag in scssr to 1 to request an rxi interrupt. set the orer, per, or fer/ers flag in scssr to 1 to request an eri interrupt (table 16.9). table 16.9 smart card mode operating state and interrupt sources mode state flag mask bit interrupt source transmit mode normal tend tie txi error fer/ers rie eri receive mode normal rdrf rie rxi error per, orer rie eri 16.4 usage notes when the sci is used as a smart card interface, be sure that all criteria in sections 16.4.1 and 16.4.2 are applied. 16.4.1 receive data timing and receive margin in asynchronous mode in asynchronous mode, the sci runs on a base clock with a frequency of 372 times the transfer rate. during reception, the sci samples the falling of the start bit using the base clock to achieve internal synchronization. receive data is latched internally at the rising edge of the 186th base clock cycle (figure 16.8). 552 0 185 371 0 185 371 0 base clock receive data (rxd) synchro- nization sampling timing data sampling timing 186 clock cycles 372 clock cycles start bit d0 d1 figure 16.8 receive data sampling timing in smart card mode the receive margin is found from the following equation: for smart card mode: m = (0.5 C ) 1 2n d C 0.5 n C (l C 0.5)f C (1 + f) 100% where: m = receive margin (%) n = ratio of bit rate to clock (n = 372) d = clock duty (d = 0 to 1.0) l = frame length (l = 10) f = absolute value of clock frequency deviation using this equation, the receive margin when f = 0 and d = 0.5 is as follows: m = (0.5 C 1/2 372) 100 % = 49.866 % 553 16.4.2 retransmission (receive and transmit modes) retransmission when sci is in receive mode: figure 16.9 shows the retransmission operation in the sci receive mode. 1. when the received parity bit is checked and an error is found, the per bit in scssr is automatically set to 1. if the rie bit in scscr is enabled at this time, an eri interrupt is requested. be sure to clear the per bit before the next parity bit is sampled. 2. the rdrf bit in scssr is not set in the frame that caused the error. 3. when the received parity bit is checked and no error is found, the per bit in scssr is not set. 4. when the received parity bit is checked and no error is found, reception is considered to have been completed normally and the rdrf bit in scssr is automatically set to 1. if the rie bit in scscr is enabled at this time, an rxi interrupt is requested. 5. when a normal frame is received, the pin maintains a three-state state when it transmits the error signal. d0 ds d2 d1 d4 d3 d6 d5 dp de d7 d0 ds d2 d1 d4 d3 d6 d5 dp (de) d7 d0 ds d2 d1 d4 d3 nth transfer frame rdrf per 2 1 4 5 3 retransmitted frame transfer frame n + 1 figure 16.9 retransmission in sci receive mode 554 retransmission when sci is in transmit mode: figure 16.10 shows the retransmission operation in the sci transmit mode. 1. after transmission of one frame is completed, the fer/ers bit in scssr is set to 1 when a error signal is returned from the receiving side. if the rie bit in scscr is enabled at this time, an eri interrupt is requested. be sure to clear the fer/ers bit before the next parity bit is sampled. 2. the tend bit in scssr is not set in the frame that received the error signal indicating the error. 3. the fer/ers bit in scssr is not set when no error signal is returned from the receiving side. 4. when no error signal is returned from the receiving side, the tend bit in scssr is set to 1 when the transmission of the frame that includes the retransmission is considered completed. if the tie bit in scscr is enabled at this time, a txi interrupt will be requested. d0 ds d2 d1 d4 d3 d6 d5 dp de d7 d0 ds d2 d1 d4 d3 d6 d5 dp (de) d7 d0 ds d2 d1 d4 d3 nth transfer frame tend fer/ers transfer from sctdr to sctsr transfer from sctdr to sctsr transfer from sctdr to sctsr 1 2 4 3 retransmitted frame transfer frame n + 1 tdre figure 16.10 retransmission in sci transmit mode 555 section 17 serial communication interface with fifo (scif) 17.1 overview the sh7729r has a two-channel serial communication interface with fifo (scif) that supports asynchronous serial communication. it also has 16-stage fifo registers for both transmission and reception that enable the sh7729r to perform efficient high-speed continuous communication. 17.1.1 features ? asynchronous serial communication: ? serial data communication is performed by start-stop in character units. the sci can communicate with a universal asynchronous receiver/transmitter (uart), an asynchronous communication interface adapter (acia), or any other communications chip that employs a standard asynchronous serial system. there are eight selectable serial data communication formats. ? data length: 7 or 8 bits ? stop bit length: 1 or 2 bits ? parity: even, odd, or none ? receive error detection: parity and framing errors ? break detection: break is detected when a framing error is followed by at least one frame at the space 0 level (low level). it is also detected by reading the rxd level directly from the port sc data register (scpdr) when a framing error occurs. ? full duplex communication: the transmitting and receiving sections are independent, so the sci can transmit and receive simultaneously. both sections use 16-stage fifo buffering, so high-speed continuous data transfer is possible in both the transmit and receive directions. ? on-chip baud rate generator with selectable bit rates ? internal or external transmit/receive clock source: from either baud rate generator (internal) or sck pin (external) ? four types of interrupts: transmit-fifo-data-empty, break, receive-fifo-data-full, and receive-error interrupts are requested independently. the direct memory access controller (dmac) can be activated to execute a data transfer by a transmit-fifo-data-empty or receive- fifo-data-full interrupt. ? when the scif is not in use, it can be stopped by halting the clock supplied to it, saving power. ? on-chip modem control functions (rts and cts) ? the quantity of data in the transmit and receive fifo registers and the number of receive errors of the receive data in the receive fifo register can be ascertained. ? a time-out error (dr) can be detected when receiving. 556 17.1.2 block diagram figure 17.1 shows a block diagram of the scif. rxd txd sck scif scbrr scssr2 scscr2 scftdr2 sctsr scfrdr2 (16 (16 stages) stages) scrsr scsmr2 scfdr2 scfcr2 scpcr scfdr parity generation parity check clock external clock module data bus internal data bus p p /4 p /16 p /64 txi tei rxi bri bus interface baud rate generator transmit/ receive control scrsr: scfrdr: sctsr: scftdr2: scsmr2: scscr2: legend receive shift register receive fifo data register transmit shift register transmit fifo data register serial mode register serial control register scssr2: scbrr2: scfcr2: scfdr2: scpdr: scpcr: serial status register bit rate register fifo control register fifo data count register port sc data register port sc control register figure 17.1 block diagram of scif 557 figures 17.2 to 17.4 show the scif i/o port pins. scif pin i/o and data control is performed by bits 11 to 8 of scpcr and bits 5 and 4 of scpdr. for details, see section 15.2.8, sc port control register (scpcr)/sc port data register (scpdr). internal data bus output enable clock input enable scif serial clock output serial clock input r scp5md0 pcrw reset c q q d r scp5md1 pcrw reset c qd r scp5dt1 pdrw reset scpt[5]/sck2 c d pdrw: legend scpdr write pdrr: pcrw: scpdr read scpcr write pdrr * note: * when reading the sck2 pin, clear the c/ a bit in scsmr and the cke1 and cke0 bits in scscr to 0, and set the scp5md1 bit in scspr to 1 (see section 15.2.8, sc port control register (scpcr)/sc port data register (scpdr)). figure 17.2 scpt[5]/sck2 pin 558 internal data bus output enable scif serial transmission output r scp4md0 pcrw reset c q q d r scp4md1 pcrw reset c qd r scp4dt1 pdrw reset scpt[4]/txd2 c d pcrw: pdrw: scpcr write scpdr write legend figure 17.3 scpt[4]/txd2 pin 559 scif internal data bus pdrr * serial receive data scpt[4]/rxd2 pdrr: scpdr read legend note: * when reading the rxd2 pin, set the re bit in scscr to 1. figure 17.4 scpt[4]/rxd2 pin 17.1.3 pin configuration the scif has the serial pins summarized in table 17.1. table 17.1 scif pins pin name abbreviation i/o function serial clock pin sck2 i/o clock i/o receive data pin rxd2 input receive data input transmit data pin txd2 output transmit data output request to send pin rts2 output request to send clear to send pin cts2 input clear to send 560 17.1.4 register configuration table 17.2 summarizes the scif internal registers. these registers specify the data format and bit rate, and control the transmitter and receiver sections. table 17.2 scif registers register name abbreviation r/w initial value address access size serial mode register 2 scsmr2 r/w h'00 h'04000150 (h'a4000150)* 2 8 bits bit rate register 2 scbrr2 r/w h'ff h'04000152 (h'a4000152)* 2 8 bits serial control register 2 scscr2 r/w h'00 h'04000154 (h'a4000154)* 2 8 bits transmit fifo data register 2 scftdr2 w h'04000156 (h'a4000156)* 2 8 bits serial status register 2 scssr2 r/(w)* 1 h'0060 h'04000158 (h'a4000158)* 2 16 bits receive fifo data register 2 scfrdr2 r undefined h'0400015a (h'a400015a)* 2 8 bits fifo control register 2 scfcr2 r/w h'00 h'0400015c (h'a400015c)* 2 8 bits fifo data count register 2 scfdr2 r h'0000 h'0400015e (h'a400015e)* 2 16 bits notes: these registers are located in area 1 of physical space. therefore, when the cache is on, either access these registers from the p2 area of logical space or else make an appropriate setting using the mmu so that these registers are not cached. *1 only 0 can be written to clear the flag. *2 when address translation by the mmu does not apply, the address in parentheses should be used. 561 17.2 register descriptions 17.2.1 receive shift register (scrsr) the receive shift register (scrsr) receives serial data. data input at the rxd pin is loaded into scrsr in the order received, lsb (bit 0) first, converting the data to parallel form. when one byte has been received, it is automatically transferred to scfrdr, the receive fifo data register. the cpu cannot read or write to scrsr directly. bit: 7 6 5 4 3 2 1 0 r/w: 17.2.2 receive fifo data register (scfrdr) the 16-byte receive fifo data register (scfrdr) stores serial receive data. the scif completes the reception of one byte of serial data by moving the received data from the receive shift register (scrsr) into scfrdr for storage. continuous reception is possible until 16 bytes are stored. the cpu can read but not write to scfrdr. if data is read when there is no receive data in the scfrdr, the value is undefined. when this register is full of receive data, subsequent serial data is lost. bit: 7 6 5 4 3 2 1 0 r/w: r r r r r r r r 17.2.3 transmit shift register (sctsr) the transmit shift register (sctsr) transmits serial data. the sci loads transmit data from the transmit fifo data register (scftdr) into sctsr, then transmits the data serially from the txd pin, lsb (bit 0) first. after transmitting one data byte, the sci automatically loads the next transmit data from scftdr into sctsr and starts transmitting again. the cpu cannot read or write to sctsr directly. bit: 7 6 5 4 3 2 1 0 r/w: 562 17.2.4 transmit fifo data register (scftdr) the transmit fifo data register (scftdr) is a 16-byte fifo register that stores data for serial transmission. when the scif detects that the transmit shift register (sctsr) is empty, it moves transmit data written in the scftdr into sctsr and starts serial transmission. continuous serial transmission is performed until there is no transmit data left in scftdr. the cpu can always write to scftdr. when scftdr is full of transmit data (16 bytes), no more data can be written. if writing of new data is attempted, the data is ignored. bit: 7 6 5 4 3 2 1 0 r/w: w w w w w w w w 17.2.5serial mode register (scsmr) the serial mode register (scsmr) is an 8-bit register that specifies the scif serial communication format and selects the clock source for the baud rate generator. the cpu can always read and write to scsmr. scsmr is initialized to h'00 by a reset and in standby or module standby mode. bit: 7 6 5 4 3 2 1 0 chr pe o/ e stop cks1 cks0 initial value: 0 0 0 0 0 0 0 0 r/w: r r/w r/w r/w r/w r r/w r/w bit 7reserved: this bit is always read as 0. the write value should always be 0. bit 6character length (chr): selects 7-bit or 8-bit data in asynchronous mode. bit 6: chr description 0 8-bit data (initial value) 1 7-bit data when 7-bit data is selected, the msb (bit 7) of the transmit fifo data register is not transmitted. 563 bit 5parity enable (pe): selects whether to add a parity bit to transmit data and to check the parity of receive data. bit 5: pe description 0 parity bit not added or checked (initial value) 1 parity bit added and checked when pe is set to 1, an even or odd parity bit is added to transmit data, depending on the parity mode (o/ e ) setting. receive data parity is checked according to the even/odd (o/ e ) mode setting. bit 4parity mode (o/ e ): selects even or odd parity when parity bits are added and checked. the o/ e setting is used only when the parity enable bit (pe) is set to 1 to enable parity addition and checking. the o/ e setting is ignored when parity addition and checking is disabled. bit 4: o/ e description 0 even parity (initial value) if even parity is selected, the parity bit is added to transmit data to make an even number of 1s in the transmitted character and parity bit combined. receive data is checked to see if it has an even number of 1s in the received character and parity bit combined. 1 odd parity if odd parity is selected, the parity bit is added to transmit data to make an odd number of 1s in the transmitted character and parity bit combined. receive data is checked to see if it has an odd number of 1s in the received character and parity bit combined. bit 3stop bit length (stop): selects one or two bits as the stop bit length. when receiving, only the first stop bit is checked, regardless of the stop bit setting. if the second stop bit is 1, it is treated as a stop bit, but if the second stop bit is 0, it is treated as the start bit of the next incoming character. bit 3: stop description 0 one stop bit (initial value) when transmitting, a single 1-bit is added at the end of each transmitted character. 1 two stop bits when transmitting, two 1-bits are added at the end of each transmitted character. bit 2reserved: this bit is always read as 0. the write value should always be 0. 564 bits 1 and 0clock select 1 and 0 (cks1, cks0): select the internal clock source of the on- chip baud rate generator. four clock sources are available. p , p /4, p /16 and p /64. for further information on the clock source, bit rate register settings, and baud rate, see section 17.2.8, bit rate register (scbrr). bit 1: cks1 bit 0: cks0 description 00p (initial value) 1p /4 10p /16 1p /64 note: p : peripheral clock 17.2.6 serial control register (scscr) the serial control register (scscr) operates the scif transmitter/receiver, selects the serial clock output in asynchronous mode, enables/disables interrupt requests, and selects the transmit/receive clock source. the cpu can always read and write to scscr. scscr is initialized to h'00 by a reset and in standby or module standby mode. bit: 7 6 5 4 3 2 1 0 tie rie te re cke1 cke0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r r r/w r/w bit 7transmit interrupt enable (tie): enables or disables the transmit-fifo-data-empty interrupt (txi) requested when the serial transmit data is transferred from the transmit fifo data register (scftdr) to the transmit shift register (sctsr), when the quantity of data in the transmit fifo register becomes less than the specified number of transmission triggers, and when the tdfe flag in the serial fifo status register (scfsr) is set to1. bit 7: tie description 0 transmit-fifo-data-empty interrupt request (txi) is disabled (initial value) the txi interrupt request can be cleared by writing a greater quantity of transmit data than the specified transmission trigger number to scftdr and by clearing tdfe to 0 after reading 1 from tdfe, or can be cleared by clearing tie to 0. 1 transmit-fifo-data-empty interrupt request (txi) is enabled 565 bit 6receive interrupt enable (rie): enables or disables the receive-data-full (rxi) and receive-error (eri) interrupts requested when serial receive data is transferred from the receive shift register (scrsr) to the receive fifo data register (scfrdr), when the quantity of data in the receive fifo register becomes more than the specified receive trigger number, and when the rdrf flag in scssr is set to1. bit 6: rie description 0 receive-data-full interrupt (rxi), receive-error interrupt (eri), and receive break interrupt (bri) requests are disabled (initial value) rxi and eri interrupt requests can be cleared by reading the dr, er, or rdf flag after it has been set to 1, then clearing the flag to 0, or by clearing rie to 0. with the rdf flag, read 1 from the rdf flag and clear it to 0, after reading receive data from scrdr until the quantity of receive data becomes less than the specified receive trigger number. 1 receive-data-full interrupt (rxi) and receive-error interrupt (eri) requests are enabled bit 5transmit enable (te): enables or disables the scif serial transmitter. bit 5: te description 0 transmitter disabled (initial value) 1 transmitter enabled serial transmission starts after writing of transmit data into scftdr2. select the transmit format in scsmr2 and scfcr2 and reset the tfifo before setting te to 1. bit 4receive enable (re): enables or disables the scif serial receiver. bit 4: re description 0 receiver disabled (initial value) clearing re to 0 does not affect the receive flags (dr, er, brk, fer, per, and orer). these flags retain their previous values. 1 receiver enabled serial reception starts when a start bit is detected. select the receive format in scsmr2 before setting re to 1. bits 3 and 2reserved: these bits are always read as 0. the write value should always be 0. 566 bits 1 and 0clock enable 1 and 0 (cke1, cke0): select the scif clock source and enable or disable clock output from the sck pin. depending on the combination of cke1 and cke0, the sck pin can be used for serial clock output or serial clock input. the cke0 setting is valid only when the scif is operating on the internal clock (cke1 = 0). the cke0 setting is ignored when an external clock source is selected (cke1 = 1). before selecting the scif operating mode in the serial mode register (scsmr), set cke1 and cke0. for further details on selection of the scif clock source, see table 17.7 in section 17.3, operation. bit 1: cke1 bit 0: cke0 description 0 0 internal clock, sck pin used for input pin (input signal is ignored) (initial value) 1 internal clock, sck pin used for clock output* 1 1 0 external clock, sck pin used for clock input* 2 1 external clock, sck pin used for clock input* 2 notes: *1 the output clock frequency is 16 times the bit rate. *2 the input clock frequency is 16 times the bit rate. 17.2.7 serial status register (scssr) the serial status register (scssr) is a 16-bit register. the upper 8 bits indicate the number of receive errors in the scfrdr data, and the lower 8 bits indicate the scif operating state. the cpu can always read and write to scssr, but cannot write 1 to the status flags (er, tend, tdfe, brk, oper, and dr). these flags can be cleared to 0 only if they have first been read (after being set to 1). bits 3 (fer) and 2 (per) are read-only bits that cannot be written. scssr is initialized to h'0060 by a reset and in standby or module standby mode. lower 8 bits: 7 6 5 4 3 2 1 0 er tend tdfe brk fer per rdf dr initial value: 0 1 1 0 0 0 0 0 r/w: r/(w)* r/(w)* r/(w)* r/(w)* r r r/(w)* r/(w)* note: * the only value that can be written is 0 to clear the flag. 567 bit 7receive error (er): indicates the occurrence of a framing error, or of a parity error when receiving data that includes parity. bit 7: er description 0 receiving is in progress or has ended normally* 1 (initial value) er is cleared to 0 when the chip is reset or enters standby mode, or when 0 is written after 1 is read from er. 1 a framing error or parity error has occurred er is set to 1 when the stop bit is 0 after checking whether or not the last stop bit of the received data is 1 at the end of one data receive operation* 2 , or when the total number of 1s in the receive data plus parity bit does not match the even/odd parity specified by the o/ e bit in scsmr. notes: *1 clearing the re bit to 0 in scscr does not affect the er bit, which retains its previous value. even if a receive error occurs, the receive data is transferred to scfrdr and the receive operation is continued. whether or not the data read from scrdr includes a receive error can be detected by the fer and per bits in scssr. *2 in stop mode, only the first stop bit is checked; the second stop bit is not checked. bit 6transmit end (tend): indicates that when the last bit of a serial character was transmitted, scftdr did not contain valid data, so transmission has ended. bit 6: tend description 0 transmission is in progress tend is cleared to 0 when data is written in scftdr. 1 end of transmission (initial value) tend is set to 1 when the chip is reset or enters standby mode, when te is cleared to 0 in the serial control register (scscr), or when scftdr does not contain receive data when the last bit of a one-byte serial character is transmitted. 568 bit 5transmit fifo data empty (tdfe): indicates that data has been transferred from the transmit fifo data register (scftdr) to the transmit shift register (sctsr), the quantity of data in scftdr has become less than the transmission trigger number specified by the ttrg1 and ttrg0 bits in the fifo control register (scfcr), and writing of transmit data to scftdr is enabled. bit 5: tdfe description 0 the quantity of transmit data written to scftdr is greater than the specified transmission trigger number (initial value) tdfe is cleared to 0 when data exceeding the specified transmission trigger number is written to scftdr, or when software reads tdfe after it has been set to 1, then writes 0 to tdfe. 1 the quantity of transmit data in scftdr is less than the specified transmission trigger number * tdfe is set to 1 by a reset or in standby mode, or when the quantity of transmit data in scftdr becomes less than the specified transmission trigger number as a result of transmission. note: * since scftdr is a 16-byte fifo register, the maximum quantity of data that can be written when tdfe is 1 is 16 minus the specified transmission trigger number. if an attempt is made to write additional data, the data is ignored. the quantity of data in scftdr is indicated by the upper 8 bits of scftdr. bit 4break detection (brk): indicates that a break signal has been detected in receive data. bit 4: brk description 0 no break signal received (initial value) brk is cleared to 0 when the chip is reset or enters standby mode, or when software reads brk after it has been set to 1, then writes 0 to brk. 1 break signal received * brk is set to 1 when data including a framing error is received, and a framing error occurs with space 0 in the subsequent receive data. note: * when a break is detected, transfer of the receive data (h'00) to scfrdr stops after detection. when the break ends and the receive signal becomes mark 1, the transfer of receive data resumes. the receive data of a frame in which a break signal is detected is transferred to scfrdr. after this, however, no receive data is transferred until a break ends with the received signal being mark 1, and the next data is received. 569 bit 3framing error (fer): indicates a framing error in the data read from the receive fifo data register (scfrdr). bit 3: fer description 0 no receive framing error occurred in the data read from scfrdr (initial value) fer is cleared to 0 when the chip undergoes a power-on reset or enters standby mode, or when no framing error is present in the data read from scfrdr. 1 a receive framing error occurred in the data read from scfrdr. fer is set to 1 when a framing error is present in the data read from scfrdr. bit 2parity error (per): indicates a parity error in the data read from the receive fifo data register (scfrdr). bit 2: per description 0 no receive parity error occurred in the data read from scfrdr (initial value) per is cleared to 0 when the chip undergoes a power-on reset or enters standby mode, or when no parity error is present in the data read from scfrdr. 1 a receive framing error occurred in the data read from scfrdr per is set to 1 when a parity error is present in the data read from scfrdr. bit 1receive fifo data full (rdf): indicates that receive data has been transferred to the receive fifo data register (scfrdr), and the quantity of data in scfrdr has become equal or greater than the receive trigger number specified by the rtrg1 and rtrg0 bits in the fifo control register (scfcr). bit 1: rdf description 0 the quantity of transmit data written to scfrdr is less than the specified receive trigger number (initial value) rdf is cleared to 0 by a power-on reset or in standby mode, or when the scfrdr is read until the quantity of receive data in scfrdr becomes less than the specified receive trigger number, or when 1 is read from rdf, and 0 is then written. 1 the quantity of receive data in scfrdr is equal or greater than the specified receive trigger number rdf is set to 1 when a quantity of receive data equal or greater than the specified receive trigger number is stored in scfrdr.* note: * since scftdr is a 16-byte fifo register, the maximum quantity of data that can be read when rdf is 1 is the specified receive trigger number. if an attempt is made to read after all the data in scfrdr has been read, the data is undefined. the quantity of receive data in scfrdr is indicated by the lower 8 bits of scftdr. 570 bit 0receive data ready (dr): indicates that the quantity of data in the receive fifo data register (scfrdr) is less than the specified receive trigger number, and that the next data has not yet been received after the elapse of 15 etu from the last stop bit. bit 0: dr description 0 receiving is in progress, or no receive data remains in scfrdr after receiving ended normally (initial value) dr is cleared to 0 when the chip undergoes a power-on reset or enters standby mode, or when software reads dr after it has been set to 1, then writes 0 to dr. 1 next receive data has not been received dr is set to 1 when scfrdr contains less data than the specified receive trigger number, and the next data has not yet been received after the elapse of 15 etu from the last stop bit.* note: * this is equivalent to 1.5 frames with the 8-bit, 1-stop-bit format. (etu: elementary time unit) upper 8 bits: 15 14 13 12 11 10 9 8 per3 per2 per1 per0 fer3 fer2 fer1 fer0 initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r r bits 15 to 12number of parity errors 3 to 0 (per3 to per0): indicate the quantity of data including a parity error in the receive data stored in the receive fifo data register (scfrdr). the value indicated by bits 15 to 12 represents the number of parity errors in scfrdr. bits 11 to 8number of framing errors 3 to 0 (fer3 to fer0): indicate the quantity of data including a framing error in the receive data stored in scfrdr. the value indicated by bits 11 to 8 represents the number of framing errors in scfrdr. 571 17.2.8 bit rate register (scbrr) the bit rate register (scbrr) is an 8-bit register that, together with the baud rate generator clock source selected by the cks1 and cks0 bits in the serial mode register (scsmr), determines the serial transmit/receive bit rate. the cpu can always read and write to scbrr. scbrr is initialized to h'ff by a reset and in module standby or standby mode. each channel has independent baud rate generator control, so different values can be set in two channels. bit: 7 6 5 4 3 2 1 0 initial value: 1 1 1 1 1 1 1 1 r/w: r/w r/w r/w r/w r/w r/w r/w r/w the scbrr setting is calculated as follows: asynchronous mode: n = p 64 2 2n C 1 b 10 6 C 1 b: bit rate (bits/s) n: scbrr setting for baud rate generator (0 n 255) p : operating frequency for peripheral modules (mhz) n: baud rate generator clock source (n = 0, 1, 2, 3) (for the clock sources and values of n, see table 17.3.) table 17.3 scsmr settings scsmr settings n clock source cks1 cks0 0p 00 1p /4 0 1 2p /16 1 0 3p /64 1 1 note: the bit rate error is given by the following formula: error (%) = p? 10 6 C 1 100 (n+1) 64 2 2nC1 b 572 table 17.4 lists examples of scbrr settings. table 17.4 bit rates and scbrr settings p (mhz) 2 2.097152 2.4576 bit rate (bits/s) n n error ( % ) n n error ( % ) n n error ( % ) 110 1 141 0.03 1 148 C0.04 1 174 C0.26 150 1 103 0.16 1 108 0.21 1 127 0.00 300 0 207 0.16 0 217 0.21 0 255 0.00 600 0 103 0.16 0 108 0.21 0 127 0.00 1200 0 51 0.16 0 54 C0.70 0 63 0.00 2400 0 25 0.16 0 26 1.14 0 31 0.00 4800 0 12 0.16 0 13 C2.48 0 15 0.00 9600 0 6 C6.99 0 6 C2.48 0 7 0.00 19200 0 2 8.51 0 2 13.78 0 3 0.00 31250 0 1 0.00 0 1 4.86 0 1 22.88 38400 0 1 C18.62 0 0 C14.67 0 1 0.00 p (mhz) 3 3.6864 4 bit rate (bits/s) n n error ( % ) n n error ( % ) n n error ( % ) 110 1 212 0.03 2 64 0.70 2 70 0.03 150 1 155 0.16 1 191 0.00 1 207 0.16 300 1 77 0.16 1 95 0.00 1 103 0.16 600 0 155 0.16 0 191 0.00 0 207 0.16 1200 0 77 0.16 0 95 0.00 0 103 0.16 2400 0 38 0.16 0 47 0.00 0 51 0.16 4800 0 19 C2.34 0 23 0.00 0 25 0.16 9600 0 9 C2.34 0 11 0.00 0 12 0.16 19200 0 4 C2.34 0 5 0.00 0 6 C6.99 31250 0 2 0.00 0 3 C7.84 0 3 0.00 38400 0 2 0.00 0 2 8.51 573 table 17.4 bit rates and scbrr settings (cont) p (mhz) 4.9152 5 6 bit rate (bits/s) n n error ( % ) n n error ( % ) n n error ( % ) 110 2 86 0.31 2 88 C0.25 2 106 C0.44 150 1 255 0.00 2 64 0.16 2 77 0.16 300 1 127 0.00 1 129 0.16 1 155 0.16 600 0 255 0.00 1 64 0.16 1 77 0.16 1200 0 127 0.00 0 129 0.16 0 155 0.16 2400 0 63 0.00 0 64 0.16 0 77 0.16 4800 0 31 0.00 0 32 C1.36 0 38 0.16 9600 0 15 0.00 0 15 1.73 0 19 C2.34 19200 0 7 0.00 0 7 1.73 0 9 C2.34 31250 0 4 C1.70 0 4 0.00 0 5 0.00 38400 0 3 0.00 0 3 1.73 0 4 C2.34 p (mhz) 6.144 7.3728 8 bit rate (bits/s) n n error ( % ) n n error ( % ) n n error ( % ) 110 2 108 0.08 2 130 C0.07 2 141 0.03 150 2 79 0.00 2 95 0.00 2 103 0.16 300 1 159 0.00 1 191 0.00 1 207 0.16 600 1 79 0.00 1 95 0.00 1 103 0.16 1200 0 159 0.00 0 191 0.00 0 207 0.16 2400 0 79 0.00 0 95 0.00 0 103 0.16 4800 0 39 0.00 0 47 0.00 0 51 0.16 9600 0 19 0.00 0 23 0.00 0 25 0.16 19200 0 9 0.00 0 11 0.00 0 12 0.16 31250 0 5 2.40 0 6 5.33 0 7 0.00 38400 0 4 0.00 0 5 0.00 0 6 C6.99 574 table 17.4 bit rates and scbrr settings (cont) p (mhz) 9.8304 10 12 12.288 bit rate (bits/s) n n error ( % )n n error ( % )n n error ( % )n n error ( % ) 110 1 174 C0.26 2 177 C0.25 1 212 0.03 2 217 0.08 150 1 127 0.00 2 129 0.16 1 155 0.16 2 159 0.00 300 0 255 0.00 2 64 0.16 1 77 0.16 2 79 0.00 600 0 127 0.00 1 129 0.16 0 155 0.16 1 159 0.00 1200 0 255 0.00 1 64 0.16 0 77 0.16 1 79 0.00 2400 0 127 0.00 0 129 0.16 0 38 0.16 0 159 0.00 4800 0 63 0.00 0 64 0.16 0 19 0.16 0 79 0.00 9600 0 31 0.00 0 32 C1.36 0 9 0.16 0 39 0.00 19200 0 15 0.00 0 15 1.73 0 4 0.16 0 19 0.00 31250 0 9 C1.70 0 9 0.00 0 2 0.00 0 11 2.40 38400 0 1 0.00 0 7 1.73 0 9 C2.34 0 9 0.00 p (mhz) 14.7456 16 19.6608 20 bit rate (bits/s) n n error ( % )n n error ( % )n n error ( % )n n error ( % ) 110 3 64 0.70 3 70 0.03 3 86 0.31 3 88 C0.25 150 2 191 0.00 2 207 0.16 2 255 0.00 2 64 0.16 300 2 95 0.00 2 103 0.16 2 127 0.00 2 129 0.16 600 1 191 0.00 1 207 0.16 1 255 0.00 1 64 0.16 1200 1 95 0.00 1 103 0.16 1 127 0.00 1 129 0.16 2400 0 191 0.00 0 207 0.16 0 255 0.00 0 64 0.16 4800 0 95 0.00 0 103 0.16 0 127 0.00 0 129 0.16 9600 0 47 0.00 0 51 0.16 0 63 0.00 0 64 0.16 19200 0 23 0.00 0 25 0.16 0 31 0.00 0 32 C1.36 31250 0 14 C1.70 0 15 0.00 0 19 C1.70 0 19 0.00 38400 0 11 0.00 0 12 0.16 0 15 0.00 0 15 1.73 115200 0 3 0.00 0 3 8.51 0 4 6.67 0 4 8.51 500000 0 0 C7.84 0 0 0.00 0 0 22.9 0 0 25.0 575 table 17.4 bit rates and scbrr settings (cont) p (mhz) 24 24.576 28.7 30 bit rate (bits/s) n n error ( % )n n error ( % )n n error ( % )n n error ( % ) 110 3 106 C0.44 3 108 0.08 3 126 0.31 3 132 0.13 150 3 77 0.16 3 79 0.00 3 92 0.46 3 97 C0.35 300 2 155 0.16 2 159 0.00 2 186 C0.08 2 194 0.16 600 2 77 0.16 2 79 0.00 2 92 0.46 2 97 C0.35 1200 1 155 0.16 1 159 0.00 1 186 C0.08 1 194 0.16 2400 1 77 0.16 1 79 0.00 1 92 0.46 1 97 C0.35 4800 0 155 0.16 0 159 0.00 0 186 C0.08 0 194 C1.36 9600 0 77 0.16 0 79 0.00 0 92 0.46 0 97 C0.35 19200 0 38 0.16 0 39 0.00 0 46 C0.61 0 48 C0.35 31250 0 23 0.00 0 24 C1.70 0 28 C1.03 0 29 0.00 38400 0 19 C2.34 0 19 0.00 0 22 1.55 0 23 1.73 115200 0 6 C6.99 0 6 C4.76 0 7 C2.68 0 7 1.73 500000 0 1 C25.0 0 1 C23.2 0 1 C10.3 0 1 C6.25 576 table 17.5 indicates the maximum bit rates in asynchronous mode when the baud rate generator is used. table 17.6 list the maximum rates for external clock input. table 17.5maximum bit rates for various frequencies with baud rate generator (asynchronous mode) settings p (mhz) maximum bit rate (bits/s) n n 2 62500 0 0 2.097152 65536 0 0 2.4576 76800 0 0 3 93750 0 0 3.6864 115200 0 0 4 125000 0 0 4.9152 153600 0 0 8 250000 0 0 9.8304 307200 0 0 12 375000 0 0 14.7456 460800 0 0 16 500000 0 0 19.6608 614400 0 0 20 625000 0 0 24 750000 0 0 24.576 768000 0 0 28.7 896875 0 0 30 937500 0 0 577 table 17.6 maximum bit rates with external clock input (asynchronous mode) p (mhz) external input clock (mhz) maximum bit rate (bits/s) 2 0.5000 31250 2.097152 0.5243 32768 2.4576 0.6144 38400 3 0.7500 46875 3.6864 0.9216 57600 4 1.0000 62500 4.9152 1.2288 76800 8 2.0000 125000 9.8304 2.4576 153600 12 3.0000 187500 14.7456 3.6864 230400 16 4.0000 250000 19.6608 4.9152 307200 20 5.0000 312500 24 6.0000 375000 24.576 6.1440 384000 28.7 7.1750 448436 30 7.5000 468750 578 17.2.9 fifo control register (scfcr) bit: 7 6 5 4 3 2 1 0 rtrg1 rtrg0 ttrg1 ttrg0 mce tfrst rfrst loop initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w the fifo control register (scfcr) resets the quantity of data in the transmit and receive fifo registers, sets the trigger data quantity, and contains an enable bit for loop-back testing. scfcr can always be read and written to by the cpu. it is initialized to h'00 by a reset, by the module standby function, and in standby mode. bits 7 and 6receive fifo data trigger (rtrg1, rtrg0): set the quantity of receive data which sets the receive data full (rdf) flag in the serial status register (scssr). the rdf flag is set when the quantity of receive data stored in the receive fifo register (scfrdr) becomes equal or greater than the set trigger number shown below. bit 7: rtrg1 bit 6: rtrg0 receive trigger number 0 0 1 (initial value) 014 108 1114 bits 5 and 4transmit fifo data trigger (ttrg1, ttrg0): set the quantity of remaining transmit data which sets the transmit fifo data register empty (tdfe) flag in the serial status register (scssr). the tdfe flag is set when the quantity of transmit data in the transmit fifo data register (scftdr) becomes less than the set trigger number shown below. bit 5: ttrg1 bit 4: ttrg0 transmit trigger number 0 0 8 (8)* 0 1 4 (12) 1 0 2 (14) 1 1 1 (15) note: * initial value. values in parentheses mean the number of empty bits in scftdr when the tdfe flag is set to 1. 579 bit 3modem control enable (mce): enables modem control signals cts and rts. bit 3: mce description 0 modem signal disabled* (initial value) 1 modem signal enabled note: * cts is fixed at active 0 regardless of the input value, and rts is also fixed at 0. bit 2transmit fifo data register reset (tfrst): disables the transmit data in the transmit fifo data register and resets the data to the empty state. bit 2: tfrst description 0 reset operation disabled* (initial value) 1 reset operation enabled note: * reset is executed in a reset or in standby mode. bit 1receive fifo data register reset (rfrst): disables the receive data in the receive fifo data register and resets the data to the empty state. bit 1: rfrst description 0 reset operation disabled* (initial value) 1 reset operation enabled note: * reset is executed in a reset or in standby mode. bit 0loop-back test (loop): internally connects the transmit output pin (txd) and receive input pin (rxd) and enables loop-back testing. bit 0: loop description 0 loop back test disabled (initial value) 1 loop back test enabled 580 17.2.10 fifo data count register (scfdr) scfdr is a 16-bit register which indicates the quantity of data stored in the transmit fifo data register (scftdr) and the receive fifo data register (scfrdr). it indicates the quantity of transmit data in scftdr with the upper 8 bits, and the quantity of receive data in scfrdr with the lower 8 bits. scfdr can always be read by the cpu. upper 8 bits: 15 14 13 12 11 10 9 8 t4t3t2t1t0 initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r r the upper 8 bits of scfdr indicate the quantity of non-transmitted data stored in scftdr. h'00 means no transmit data, and h'10 means that scftdr is full of transmit data. lower 8 bits: 7 6 5 4 3 2 1 0 r4r3r2r1r0 initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r r the lower 8 bits of scfdr indicate the quantity of receive data stored in scfrdr. h'00 means no receive data, and h'10 means that scfrdr full of receive data. 581 17.3 operation 17.3.1 overview for serial communication, the scif has an asynchronous mode in which characters are synchronized individually. refer to section 15.3.2, operation in asynchronous mode. the scif has a 16-byte fifo buffer for both transmit and receive operations, reducing the overhead of the cpu, and enabling continuous high-speed communication. moreover, it has rts and cts signals as modem control signals. the transmission format is selected in the serial mode register (scsmr), as shown in table 17.7. the sci clock source is selected by the combination of the cke1 and cke0 bits in the serial control register (scscr), as shown in table 17.8. ? data length is selectable: 7 or 8 bits. ? parity and multiprocessor bits are selectable, as is the stop bit length (1 or 2 bits). the combination of the preceding selections constitutes the communication format and character length. ? in receiving, it is possible to detect framing errors (fer), parity errors (per), receive fifo data full, receive data ready, and breaks. ? in transmitting, it is possible to detect transmit fifo data empty. ? the number of stored data bytes is indicated for both the transmit and receive fifo registers. ? an internal or external clock can be selected as the scif clock source. ? when an internal clock is selected, the scif operates using the on-chip baud rate generator, and can output a serial clock signal with a frequency 16 times the bit rate. ? when an external clock is selected, the external clock input must have a frequency 16 times the bit rate. (the on-chip baud rate generator is not used.) table 17.7 scsmr settings and scif communication formats scsmr settings scif communication format mode bit 6 chr bit 5 pe bit 3 stop data length parity bit stop bit length asynchronous 0 0 0 8-bit not set 1 bit 1 2 bits 1 0 set 1 bit 1 2 bits 1 0 0 7-bit not set 1 bit 1 2 bits 1 0 set 1 bit 1 2 bits 582 table 17.8 scscr settings and scif clock source selection scscr settings scif transmit/receive clock mode bit 1 cke1 bit 0 cke0 clock source sck pin function asynchronous 0 0 internal scif does not use the sck pin mode 1 outputs a clock with a frequency 16 times the bit rate 1 0 external inputs a clock with frequency 16 times the 1 bit rate 17.3.2 serial operation transmit/receive formats: table 17.8 lists the eight communication formats that can be selected. the format is selected by settings in the serial mode register (scsmr). table 17.9 serial communication formats scsmr bits serial transmit/receive format and frame length chr pe stop 1 2345678 9 10 11 12 0 0 0 start 8-bit data stop 0 0 1 start 8-bit data stop stop 0 1 0 start 8-bit data p stop 0 1 1 start 8-bit data p stop stop 1 0 0 start 7-bit data stop 1 0 1 start 7-bit data stop stop 1 1 0 start 7-bit data p stop 1 1 1 start 7-bit data p stop stop start: start bit stop: stop bit p: parity bit 583 clock: an internal clock generated by the on-chip baud rate generator or an external clock input from the sck pin can be selected as the scif transmit/receive clock. the clock source is selected by bits cke1 and cke0 in the serial control register (scscr) (table 17.7). when an external clock is input at the sck pin, it must have a frequency equal to 16 times the desired bit rate. when the scif operates on an internal clock, it can output a clock signal at the sck pin. the frequency of this output clock is 16 times the bit rate. transmitting and receiving data (scif initialization): before transmitting or receiving, clear the te and re bits to 0 in the serial control register (scscr), then initialize the scif as follows. when changing the communication format, always clear the te and re bits to 0 before following the procedure given below. clearing te to 0 initializes the transmit shift register (sctsr). clearing te and re to 0, however, does not initialize the serial status register (scssr), transmit fifo data register (scftdr), or receive fifo data register (scfrdr), which retain their previous contents. clear te to 0 after all transmit data has been transmitted and the tend flag in the scssr is set. the te bit can be cleared to 0 during transmission, but the transmit data goes to the high impedance state after the bit is cleared to 0. set the tfrst bit in scfcr to 1 and reset scftdr before te is set again to start transmission. when an external clock is used, the clock should not be stopped during initialization or subsequent operation. scif operation becomes unreliable if the clock is stopped. figure 17.5 shows a sample flowchart for initializing the scif. the procedure for initializing the scif is: 1. set the clock selection in scscr. be sure to clear bits rie tie, te, and re to 0. when clock output is selected, the clock is output immediately after scscr settings are made. 2. set the communication format in scsmr. 3. write a value corresponding to the bit rate into the bit rate register (scbrr). (not necessary if an external clock is used.) 4. wait at least one bit interval, then set the te bit or re bit in scscr to 1. also set the rie and tie bits. setting the te and re bits enables the txd and rxd pins to be used. when transmitting, the scif will go to the mark state; when receiving, it will go to the idle state, waiting for a start bit. 584 initialization clear te and re bits in scscr to 0 set tfrst and rfrst bits in scfcr to 1 1-bit interval elapsed? set rtrg1-0, ttrg1-0, and mce in scfcr clear tfrst and rfrst bits to 0 set te and re bits in scscr to 1,and set rie, tie, teie, and mpie bits set communication format in scsmr yes no set value in scbrr set cke1 and cke0 bits in scscr (leaving te and re bits cleared to 0) end (1) (2) (3) (4) wait note: numbers in parentheses refer to steps in the preceding procedure description. figure 17.5 sample flowchart for scif initialization 585 ? serial data transmission figure 17.6 shows a sample flowchart for serial transmission. use the following procedure for serial data transmission after enabling the scif for transmission. 1. scif status check and transmit data write: read serial status register (scssr) and check that the tdfe flag is set to 1, then write transmit data to the transmit fifo data register (scftdr), read 1 from the tdfe and tend flags, then clear these flags to 0. the number of transmit data bytes that can be written is 16 - (transmit trigger set number). 2. serial transmission continuation procedure: to continue serial transmission, read 1 from the tdfe flag to confirm that writing is possible, then write data to scftdr, and then clear the tdfe flag to 0. 3. break output at the end of serial transmission: to output a break in serial transmission, set the port sc data register (scpdr) and port sc control register (scpcr), then clear the te bit to 0 in the serial control register (scscr). for information on scpdr and scpcr, see section 17.2.8. in steps 1 and 2, it is possible to ascertain the number of data bytes that can be written from the number of transmit data bytes in scftdr indicated by the upper 8 bits of the fifo data count register (scfdr). 586 start of transmission read tdfe bit in scssr tend= 1? read tend bit in scssr clear te bit in scscr to 0 set scpdr and scpcr yes no tdfe= 1? no all data transmitted? no yes yes break output? yes no write transmit data (16 - transmit trigger set number) to scftdr, read 1 from tdfe bit and tend flag in scssr, then clear to 0 end of transmission (1) (2) (3) note: numbers in parentheses refer to steps in the preceding procedure description. figure 17.6 sample flowchart for transmitting serial data 587 in serial transmission, the scif operates as described below. 1. when data is written into the transmit fifo data register (scftdr), the scif transfers the data from scftdr to the transmit shift register (sctsr) and starts transmitting. confirm that the tdfe flag in the serial status register (scssr) is set to 1 before writing transmit data to scftdr. the number of data bytes that can be written is (16 C transmit trigger setting). 2. when data is transferred from scftdr to sctsr and transmission is started, consecutive transmit operations are performed until there is no transmit data left in scftdr. when the number of transmit data bytes in scftdr falls below the transmit trigger number set in the fifo control register (scfcr), the tdfe flag is set. if the tie bit in the serial control register (scsr) is set to 1 at this time, a transmit-fifo-data-empty interrupt (txi) request is generated. the serial transmit data is sent from the txd pin in the following order. a. start bit: one-bit 0 is output. b. transmit data: 8-bit or 7-bit data is output in lsb-first order. c. parity bit: one parity bit (even or odd parity) is output. (a format in which a parity bit is not output can also be selected.) d. stop bit(s): one or two 1-bits (stop bits) are output. e. mark state: 1 is output continuously until the start bit that starts the next transmission is sent. 3. the scif checks the scftdr transmit data at the timing for sending the stop bit. if data is present, the data is transferred from scftdr to sctsr, the stop bit is sent, and then serial transmission of the next frame is started. if there is no transmit data, the tend flag in scssr is set to 1, the stop bit is sent, and then the line goes to the mark state in which 1 is output continuously. 588 figure 17.7 shows an example of the operation for transmission. 01 1 1 0/1 0 1 tdfe tend parity bit parity bit serial data start bit data stop bit start bit data stop bit idle (mark) state txi interrupt request data written to scftdr and tdfe flag read as 1 then cleared to 0 by txi interrupt handler one frame d 0 d 1 d 7 d 0 d 1 d 7 0/1 txi interrupt request figure 17.7 example of transmit operation (8-bit data, parity, one stop bit) 4. when modem control is enabled, transmission can be stopped and restarted in accordance with the cts input value. when cts is set to 1, if transmission is in progress, the line goes to the mark state after transmission of one frame. when cts is set to 0, the next transmit data is output starting from the start bit. figure 17.8 shows an example of the operation when modem control is used. 0 0/1 0 cts parity bit serial data txd start bit stop bit start bit d 0 d 1 d 7 d 0 d 1 d 7 0/1 rise at this point before stop bit figure 17.8 example of operation using modem control ( cts ) 589 ? serial data reception figures 17.9 and 17.10 show a sample flowchart for serial reception. use the following procedure for serial data reception after enabling the scif for reception. 1. receive error handling and break detection: read the dr, er, and brk flags in scssr to identify any error, perform the appropriate error handling, then clear the dr, er, and brk flags to 0. in the case of a framing error, a break can also be detected by reading the value of the rxd pin. 2. scif status check and receive data read : read the serial status register (scssr) and check that rdf = 1, then read the receive data in the receive fifo data register (scfrdr), read 1 from the rdf flag, and then clear the rdf flag to 0. the transition of the rdf flag from 0 to 1 can be identified by an rxi interrupt. 3. serial reception continuation procedure: to continue serial reception, read at least the receive trigger set number of receive data bytes from scfrdr, read 1 from the rdf flag, then clear the rdf flag to 0. the number of receive data bytes in scfrdr can be ascertained by reading the lower bits of scfdr. 590 start of reception read orer, per, fer flags in scssr all data received? end of reception no yes brk v er v dr = 1? rdf = 1? yes yes clear re bit in scscr to 0 no no read rdf flag in scssr error handling read receive data from scfrdr, and clear rdf flag in scssr to 0 (1) (2) (3) note: numbers in parentheses refer to steps in the preceding procedure description. figure 17.9 sample flowchart for receiving serial data 591 1. whether a framing error or parity error has occurred in the receive data read from scfrdr can be ascertained from the fer and per bits in scssr. 2. when a break signal is received, receive data is not transferred to scfrdr while the brk flag is set. however, note that the last data in scfrdr is h'00 and the break data in which a framing error occurred is stored. error handling end brk= 1? dr= 1? er = 1? yes yes clear dr, er, brk flags in scssr to 0 no no no receive error handling break processing read receive data from scfrdr figure 17.10 sample flowchart for receiving serial data (cont) 592 in serial reception, the scif operates as described below. 1. the scif monitors the transmission line, and if a 0 start bit is detected, performs internal synchronization and starts reception. 2. the received data is stored in scrsr in lsb-to-msb order. 3. the parity bit and stop bit are received. after receiving these bits, the scif carries out the following checks. a. stop bit check: the scif checks whether the stop bit is 1. if there are two stop bits, only the first is checked. b. the scif checks whether receive data can be transferred from the receive shift register (scrsr) to scfrdr. c. break check: the scif checks that the brk flag is 0, indicating that the break state is not set. if all the above checks are passed, the receive data is stored in scfrdr. note: reception is not suspended when a receive error occurs. 4. if the rie bit in scsr is set to 1 when the rdf or dr flag changes to 1, a receive-fifo-data- full interrupt (rxi) request is generated. if the rie bit in scsr is set to 1 when the er flag changes to 1, a receive-error interrupt (eri) request is generated. if the rie bit in scsr is set to 1 when the brk flag changes to 1, a break reception interrupt (bri) request is generated. figure 17.11 shows an example of the operation for reception. 593 rdf fer eri interrupt request generated by receive error one frame data read and rdf flag read as 1 then cleared to 0 by rxi interrupt handler rxi interrupt request 01 1 1 0/1 0 1 parity bit parity bit serial data start bit data stop bit start bit data stop bit idle (mark) state d 0 d 1 d 7 d 0 d 1 d 7 0/1 figure 17.11 example of scif receive operation (8-bit data, parity, one stop bit) 5. when modem control is enabled, the rts signal is output when scfrdr is empty. when rts is 0, reception is possible. when rts is 1, this indicates that scfrdr is full and reception is not possible. figure 17.12 shows an example of the operation when modem control is used. 0 0/1 0 1 rts parity bit serial data rxd start bit start d 0 d 1 d 2 d 7 figure 17.12 example of operation using modem control ( rts ) 594 17.4 scif interrupts the scif has four interrupt sources: transmit-fifo-data-empty (txi), receive-error (eri), receive-data-full (rxi), and break (bri). table 17.10 shows the interrupt sources and their order of priority. the interrupt sources are enabled or disabled by means of the tie and rie bits in scscr. a separate interrupt request is sent to the interrupt controller for each of these interrupt sources. when the tdfe flag in the serial status register (scssr) is set to 1, a txi interrupt request is generated. the dmac can be activated and data transfer performed when this interrupt is generated. the tdfe flag is automatically cleared to 0 when data is written to the transmit data register (scftdr) by the dmac. when the rdf flag in scssr is set to 1, an rxi interrupt request is generated. the dmac can be activated and data transfer performed when the rdf flag in scssr is set to 1. the rdf flag is automatically cleared to 0 when data is read from the receive data register (scfrdr) by the dmac. when the er flag in scssr is set to 1, an eri interrupt request is generated. when the brk flag in scssr is set to 1, a bri interrupt request is generated. the txi interrupt indicates that transmit data can be written, and the rxi interrupt indicates that there is receive data in scfrdr. table 17.10 scif interrupt sources interrupt source description dmac activation priority on reset release eri interrupt initiated by receive error flag (er) not possible high rxi interrupt initiated by receive data fifo full flag (rdf) or data ready flag (dr) possible (rdf only) bri interrupt initiated by break flag (brk) not possible txi interrupt initiated by transmit fifo data empty flag (tdfe) possible low see section 4, exception handling, for priorities and the relationship to non-scif interrupts. 595 17.5usage notes note the following when using the scif. 1. scftdr writing and tdfe flag: the tdfe flag in the serial status register (scssr) is set when the number of transmit data bytes written in the transmit fifo data register (scftdr) has fallen below the transmit trigger number set by bits ttrg1 and ttrg0 in the fifo control register (scfcr). after tdfe is set, transmit data up to the number of empty bytes in scftdr can be written, allowing efficient continuous transmission. however, if the number of data bytes written in scftdr is equal to or less than the transmit trigger number, the tdfe flag will be set to 1 again after being read as 1 and cleared to 0. tdfe clearing should therefore be carried out when scftdr contains more than the transmit trigger number of transmit data bytes. the number of transmit data bytes in scftdr can be found from the upper 8 bits of the fifo data count register (scfdr). 2. scfrdr reading and rdf flag: the rdf flag in the serial status register (scssr) is set when the number of receive data bytes in the receive fifo data register (scfrdr) has become equal to or greater than the receive trigger number set by bits rtrg1 and rtrg0 in the fifo control register (scfcr). after rdf is set, receive data equivalent to the trigger number can be read from scfrdr, allowing efficient continuous reception. however, if the number of data bytes in scfrdr is equal to or greater than the trigger number, the rdf flag will be set to 1 again if it is cleared to 0. rdf should therefore be cleared to 0 after being read as 1 after all the receive data has been read. the number of receive data bytes in scfrdr can be found from the lower 8 bits of the fifo data count register (scfdr). 3. break detection and processing: break signals can be detected by reading the rxd pin directly when a framing error (fer) is detected. in the break state the input from the rxd pin consists of all 0s, so the fer flag is set and the parity error flag (per) may also be set. note that, although transfer of receive data to scfrdr is halted in the break state, the scif receiver continues to operate, so if the brk flag is cleared to 0 it will be set to 1 again. 4. sending a break signal: the i/o condition and level of the txd pin are determined by the scp4dt bit in the port sc data register (scpdr) and bits scp4md0 and scp4md1 in the port sc control register (scpcr). this feature can be used to send a break signal. to send a break signal during serial transmission, clear the cp4dt bit to 0 (designating low level), then set the scp4md0 and scp4md1 bits to 0 and 1, respectively, and finally clear the te bit to 0 (halting transmission). when the te bit is cleared to 0, the transmitter is initialized regardless of the current transmission state, and 0 is output from the txd pin. 596 5. tend flag and te bit processing: the tend flag is set to 1 during transmission of the stop bit of the last data. consequently, if the te bit is cleared to 0 immediately after setting of the tend flag has been confirmed, the stop bit will be in the process of transmission and will not be transmitted normally. therefore, the te bit should not be cleared to 0 for at least 0.5 serial clock cycles (or 1.5 cycles if two stop bits are used) after setting of the tend flag is confirmed. 6. receive data sampling timing and receive margin: the scif operates on a base clock with a frequency of 16 times the transfer rate. in reception, the scif synchronizes internally with the fall of the start bit, which it samples on the base clock. receive data is latched at the rising edge of the eighth base clock pulse. the timing is shown in figure 17.13. 0 1 2 3 4 5 6 7 8 9 10111213 1415 0 1 2 3 4 5 6 7 8 9 10111213 1415 0 1 2 3 4 5 base clock receive data (rxd) synchro- nization sampling timing data sampling timing 8 clocks 16 clocks start bit C7.5 clocks +7.5 clocks d0 d1 figure 17.13 receive data sampling timing in asynchronous mode the receive margin in asynchronous mode can therefore be expressed as shown in equation 1. equation 1: m = 0.5 C 1 2n d C 0.5 n C (l C 0.5) f C (1 + f) 100% where: m: receive margin (%) n: ratio of clock frequency to bit rate (n = 16) d: clock duty cycle (d = 0 to 1.0) l: frame length (l = 9 to 12) f: absolute deviation of clock frequency from equation 1, if f = 0 and d = 0.5, the receive margin is 46.875%, as given by equation 2. 597 equation 2: when d = 0.5 and f = 0: m = (0.5 C 1/(2 16)) 100 % = 46.875 % this is a theoretical value. a reasonable margin to allow in system designs is 20% to 30%. 7. note on use of modem signals: the following bug may occur when modem signals are used by the scif (serial communication interface with fifo) (i.e. when the mce bit in scfcr is set to 1). ? conditions when, in transmission, the cts signal goes high when the last byte of data in the fifo has been transmitted. when the last byte of data in the fifo has been transmitted means the state in which there is data in sctsr only (and no data in scftdr), and that data, including the start bit, has been output to txd. it does not cover the case where the next data is written to scftdr before transmission of the last data is started. ? problem transmission may halt midway through a frame when the entire frame should be transmitted (from start bit to stop bit). in this case, a high level is output from txd. ? preventive measures this bug can be avoided by taking the following measures. ? do not perform hard flow control with the cts signal. ? use an irq signal instead of the cts signal, and control starting and stopping of transmission by means of interrupt handling. ? perform transmission/reception with dummy data as the last data (in systems where this is permissible). 598 599 section 18 irda 18.1 overview the sh7729r has an on-chip infrared data association (irda) interface which is based on the irda 1.0 system and can perform infrared communication. it also can be used as the scif by making register settings. 18.1.1 features ? conforms to the irda 1.0 system ? asynchronous serial communication ? data length: 8 bits ? stop bit length: 1 bit ? parity bit: none ? on-chip 16-stage fifo buffers for both transmit and receive operations ? on-chip baud rate generator with selectable bit rates ? guard functions to protect the receiver during transmission ? clock supply halted to reduce power consumption when not using the irda interface 600 18.1.2 block diagram figure 18.1 shows a block diagram of the irda. scif clock input txd transfer clock rxd switching irda/scif irda sck txd1 rxd1 modulation unit demodulation unit legend scif: serial communication interface with fifo figure 18.1 block diagram of irda 601 figures 18.2 to 18.4 show the irda i/o port pins. scif pin i/o and data control is performed by bits 7 to 4 of scpcr and bits 3 and 2 of scpdr. for details, see section 15.2.8, sc port control register (scpcr)/sc port data register (scpdr). internal data bus output enable clock input enable irda serial clock output serial clock input r scp3md0 pcrw reset c q q d r scp3md1 pcrw reset c qd r scp3dt1 pdrw reset scpt[3]/sck1 c d pdrw: legend scpdr write pdrr: pcrw: scpdr read scpcr write pdrr * note: * when reading the sck1 pin, the cke1 and cke0 bits in scscr to 0, and set the scp3md1 bit in scspr to 1 (see section 15.2.8, sc port control register (scpcr)/sc port data register (scpdr)). figure 18.2 scpt[3]/sck1 pin 602 internal data bus output enable irda serial transfer output q r scp2dt1 pdrw reset scpt[2]/txd1 c d pcrw: pdrw: legend scpcr write scpdr write r scp2md0 pcrw reset c q d r scp2md1 pcrw reset c qd figure 18.3 scpt[2]/txd1 pin 603 irda serial receive data internal data bus pdrr * scpt[2]/rxd1 legend pdrr: scpdr read note: * when reading the rxd1 pin, set the re bit in scscr to 1. figure 18.4 scpt[2]/rxd1 pin 18.1.3 pin configuration the irda has the serial pins summarized in table 18.1. table 18.1 irda pins pin name signal name i/o function serial clock pin sck1 i/o clock i/o receive data pin rxd1 input receive data input transmit data pin txd1 output transmit data output note: clock input from the serial clock pin cannot be set in irda mode. 604 18.1.4 register configuration the irda has the internal registers shown in table 18.2. these registers select irda or scif mode, specify the data format and a bit rate, and control the transmit and receive units. table 18.2 irda registers register name abbreviation r/w initial value address access size serial mode register 1 scsmr1 r/w h'00 h'04000140 (h'a4000140)* 2 8 bits bit rate register 1 scbrr1 r/w h'ff h'04000142 (h'a4000142)* 2 8 bits serial control register 1 scscr1 r/w h'00 h'04000144 (h'a4000144)* 2 8 bits transmit fifo data register 1 scftdr1 w h'04000146 (h'a4000146)* 2 8 bits serial status register 1 scssr1 r/(w)* 1 h'0060 h'04000148 (h'a4000148)* 2 16 bits receive fifo data register 1 scfrdr1 r undefined h'0400014a (h'a400014a)* 2 8 bits fifo control register 1 scfcr1 r/w h'00 h'0400014c (h'a400014c)* 2 8 bits fifo data count register 1 scfdr1 r h'0000 h'0400014e (h'a400014e)* 2 16 bits notes: these registers are located in area 1 of physical space. therefore, when the cache is on, either access these registers from the p2 area of logical space or else make an appropriate setting using the mmu so that these registers are not cached. *1 only 0 can be written to clear the flag. *2 when address translation by the mmu does not apply, the address in parentheses should be used. 605 18.2 register description specifications of the registers in the irda are the same as those in the scif except for the serial mode register described below. therefore, refer to section 17, serial communication interface with fifo (scif), for details of these registers. 18.2.1 serial mode register (scsmr) bit: 7 6 5 4 3 2 1 0 irmod ick3 ick2 ick1 ick0 psel cks1 cks0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w scsmr is an 8-bit register that selects irda or scif mode, specifies the scif serial communication format, selects the irda output pulse width, and selects the baud rate generator clock source. this module operates as irda when the irmod bit is set to 1. at this time, bits 3 to 6 are fixed at 0. this register functions in the same way as the scsmr register in the scif when the irmod bit is cleared to 0; therefore, this module can also operate as an scif. scsmr is initialized to h'00 by a power-on reset or manual reset, when the module is stopped by the module standby function, and in standby mode. bit 7irda mode (irmod): selects whether this module operates as an irda serial communication interface or as an scif. bit 7: irmod description 0 operates as an scif (initial value) 1 operates as an irda 606 bits 6 to 3ir clock select bits (ick3 to ick0) bit 2output pulse width select (psel): psel selects an irda output pulse width that is 3/16 of the bit length for 115 kbps or 3/16 of the bit length for the selected baud rate. the ir clock select bits should be set properly to fix the output pulse width at 3/16 of the bit length for 115 kbps by setting the psel bit to 1. bit 6: ick3 bit 5: ick2 bit 4: ick1 bit 3: ick0 bit 2: psel description ick3 ick2 ick1 ick0 1 pulse width: 3/16 of 115 kbps bit length dont care dont care dont care dont care 0 pulse width: 3/16 of bit length it is necessary to generate a fixed clock pulse, irclk, by dividing the p clock by 1/2n + 2 (with the value of n determined by the setting of ick3Cick0). example: p clock: 14.7456 mhz irclk: 921.6 khz (fixed) n: setting of ick3Cick0 (0 n 15) C 1 7 n p 2xirclk accordingly, n is 7. bits 1 and 0clock select 1 and 0 (cks1, cks0): select the internal baud rate generator clock source. p , p /4, p /16, or p /64 can be selected by setting the cks1 and cks0 bits. refer to section 15.2.9, bit rate register (scbrr), for the relationship between the clock source, the bit rate register set value, and the baud rate. bit 1: cks1 bit 0: cks0 description 00p clock (initial value) 01p /4 clock 10p /16 11p /64 note: p : peripheral clock 607 18.3 operation description the irda module can perform infrared communication conforming to irda 1.0 by connecting infrared transmit/receive units. the serial communication interface unit includes a 16-stage fifo buffer in the transmit unit and the receive unit, allowing cpu overhead to be reduced and continuous high-speed communication to be performed. this module also supports dmac data transfer. the irda module differs from the scif described in section 17 in that it does not include modem control signals rts and cts. refer to section 17.3, operation, for scif mode operation. 18.3.1 overview the irda module modifies txd/rxd transmit/receive data waveforms to satisfy the irda 1.0 specification for infrared communication. in the irda 1.0 specification, communication is first performed at a speed of 9600 bps, and the communication speed is changed. however, the communication rate cannot be automatically changed in this module, so the communication speed should be confirmed, and the appropriate speed set for this module by software. note: in irda mode, reception cannot be performed when the te bit in the serial control register (scscr) is set to 1 (enabling transmission). when performing reception, clear the te bit in scscr to 0. as the sh7729r's rxd1 pin is active-high in irda mode, a (schmidt) inverter must be inserted when connecting an active-low irda module. the rxd1 pin is active-low in scif mode. 18.3.2 transmitting in the case of a serial output signal (uart frame) from the scif, its waveforms are modified and the signal is converted into the ir frame serial output signal by the irda module, as shown in figure 18.5. when serial data is 0, a pulse of 3/16 the ir frame bit width is generated and output. when serial data is 1, no pulse is output. an infrared led is driven by this signal demodulated to 3/16 width. 608 18.3.3 receiving received 3/16 ir frame bit-width pulses are demodulated and converted to a uart frame, as shown in figure 18.5. demodulation to 0 is performed for pulse output, and demodulation to 1 is performed for no pulse output. uart frame data ir frame data receive transmit stop bit stop bit start bit start bit bit cycle 3/16-bit cycle pulse width 01 01 0 0 11 0 1 01 01 0 01 1 01 figure 18.5 transmit/receive operation 609 section 19 pin function controller 19.1 overview the pin function controller (pfc) is composed of registers for selecting the function of multiplexed pins and the input/output direction. the pin function and input/output direction can be selected for each pin individually without regard to the operating mode of the chip. table 19.1 lists the multiplexed pins. table 19.1 list of multiplexed pins port port function (related module) other function (related module) a pta7 input/output (port) d23 input/output (data bus) a pta6 input/output (port) d22 input/output (data bus) a pta5 input/output (port) d21 input/output (data bus) a pta4 input/output (port) d20 input/output (data bus) a pta3 input/output (port) d19 input/output (data bus) a pta2 input/output (port) d18 input/output (data bus) a pta1 input/output (port) d17 input/output (data bus) a pta0 input/output (port) d16 input/output (data bus) b ptb7 input/output (port) d31 input/output (data bus) b ptb6 input/output (port) d30 input/output (data bus) b ptb5 input/output (port) d29 input/output (data bus) b ptb4 input/output (port) d28 input/output (data bus) b ptb3 input/output (port) d27 input/output (data bus) b ptb2 input/output (port) d26 input/output (data bus) b ptb1 input/output (port) d25 input/output (data bus) b ptb0 input/output (port) d24 input/output (data bus) c ptc7 input/output (port)/pint7 input (intc) mcs7 output (bsc) c ptc6 input/output (port)/pint6 input (intc) mcs6 output (bsc) c ptc5 input/output (port)/pint5 input (intc) mcs5 output (bsc) c ptc4 input/output (port)/pint4 input (intc) mcs4 output (bsc) c ptc3 input/output (port)/pint3 input (intc) mcs3 output (bsc) c ptc2 input/output (port)/pint2 input (intc) mcs2 output (bsc) c ptc1 input/output (port)/pint1 input (intc) mcs1 output (bsc) 610 table 19.1 list of multiplexed pins (cont) port port function (related module) other function (related module) c ptc0 input/output (port)/pint0 input (intc) mcs0 output (bsc) d ptd7 input/output (port) dack1 output (dmac) d ptd6 input (port) dreq1 input (dmac) d ptd5 input/output (port) dack0 output (dmac) d ptd4 input (port) dreq0 input (dmac) d ptd3 input/output (port) wakeup output (wtc) d ptd2 input/output (port) resetout output d ptd1 input/output (port) drak0 output (dmac) d ptd0 input/output (port) drak1 output (dmac) e pte7 input/output (port) audsync output (aud) * 1 e pte6 input/output (port) e pte5 input/output (port) ce2b output (pcmcia) e pte4 input/output (port) ce2a output (pcmcia) e pte3 input/output (port) e pte2 input/output (port) ras3u output (bsc) e pte1 input/output (port) e pte0 input/output (port) tdo output (h-udi) * 2 f ptf7 input (port)/pint15 input (intc) trst input (aud, h-udi) * 2 f ptf6 input (port)/pint14 input (intc) tms input (h-udi) * 2 f ptf5 input (port)/pint13 input (intc) td1 input (h-udi) * 2 f ptf4 input (port)/pint12 input (intc) tck input (h-udi) * 2 f ptf3 input (port)/pint11 input (intc) irls3 input (intc) f ptf2 input (port)/pint10 input (intc) irls2 input (intc) f ptf1 input (port)/pint9 input (intc) irls1 input (intc) f ptf0 input (port)/pint8 input (intc) irls0 input (intc) g ptg7 input (port) iois16 input (pcmcia) g ptg6 input (port) asemd0 input (aud, h-udi) * 2 g ptg5 input (port) asebrkak output (aud) * 1 g ptg4 input (port) ckio2 output (cpg) g ptg3 input (port) audata3 output (aud) * 1 g ptg2 input (port) audata2 output (aud) * 1 611 table 19.1 list of multiplexed pins (cont) port port function (related module) other function (related module) g ptg1 input (port) audata1 output (aud) * 1 g ptg0 input (port) audata0 output (aud) * 1 h pth7 input/output (port) tclk input/output (tmu) h pth6 input (port) audck input (aud) * 1 h pth5 input (port) adtrg input (adc) h pth4 input (port)/irq4 input (intc) irq4 input (intc) h pth3 input (port)/irq3 input (intc) irq3 input (intc) h pth2 input (port)/irq2 input (intc) irq2 input (intc) h pth1 input (port)/irq1 input (intc) irq1 input (intc) h pth0 input (port)/irq0 input (intc) irq0 input (intc) j ptj7 input/output (port) status1 output (cpg) j ptj6 input/output (port) status0 output (cpg) j ptj5 input/output (port) j ptj4 input/output (port) j ptj3 input/output (port) casu output (bsc) j ptj2 input/output (port) casl output (bsc) j ptj1 input/output (port) j ptj0 input/output (port) ras3l output (bsc) k ptk7 input/output (port) we3 output (bsc)/dqmuu output (bsc)/ iciowr output (bsc) k ptk6 input/output (port) we2 output (bsc)/dqmul output (bsc)/ iciord output (bsc) k ptk5 input/output (port) cke output (bsc) k ptk4 input/output (port) bs output (bsc) k ptk3 input/output (port) cs5 output (bsc)/ ce1a output (bsc) k ptk2 input/output (port) cs4 output (bsc) k ptk1 input/output (port) cs3 output (bsc) k ptk0 input/output (port) cs2 output (bsc) l ptl7 input (port) an7 input (adc)/da0 output (dac) l ptl6 input (port) an6 input (adc)/da1 output (dac) l ptl5 input (port) an5 input (adc) 612 table 19.1 list of multiplexed pins (cont) port port function (related module) other function (related module) l ptl4 input (port) an4 input (adc) l ptl3 input (port) an3 input (adc) l ptl2 input (port) an2 input (adc) l ptl1 input (port) an1 input (adc) l ptl0 input (port) an0 input (adc) scpt scpt7 input (port)/irq5 input (intc) cts2 input (uart ch 3)/irq5 input (intc) scpt scpt6 input/output (port) rts2 output (uart ch 3) scpt scpt5 input/output (port) sck2 input/output (uart ch 3) scpt scpt4 input (port) rxd2 input (uart ch 3) scpt4 output (port) txd2 output (uart ch 3) scpt scpt3 input/output (port) sck1 input/output (uart ch 2) scpt scpt2 input (port) rxd1 input (uart ch 2) scpt2 output (port) txd1 output (uart ch 2) scpt scpt1 input/output (port) sck0 input/output (uart ch 1) scpt scpt0 input (port) rxd0 input (uart ch 1) scpt0 output (port) txd0 output (uart ch 1) notes: scpt0, scpt2, and scpt4 have the same data register to be accessed although they have different input pins and output pins. * 1 for use of emulator only * 2 for use of emulator or boundary scan only 613 19.2 register configuration table 19.2 summarizes the registers of the pin function controller. table 19.2 pin function controller registers name abbreviation r/w initial value address access size port a control register pacr r/w h'0000 h'04000100 (h'a4000100)* 16 port b control register pbcr r/w h'0000 h'04000102 (h'a4000102)* 16 port c control register pccr r/w h'aaaa h'04000104 (h'a4000104)* 16 port d control register pdcr r/w h'aa8a h'04000106 (h'a4000106)* 16 port e control register pecr r/w h'aaaa/h'2aa8 h'04000108 (h'a4000108)* 16 port f control register pfcr r/w h'aaaa/h'00aa h'0400010a (h'a400010a)* 16 port g control register pgcr r/w h'aaaa/h'8200 h'0400010c (h'a400010c)* 16 port h control register phcr r/w h'aaaa/h'8aaa h'0400010e (h'a400010e)* 16 port j control register pjcr r/w h'0000 h'04000110 (h'a4000110)* 16 port k control register pkcr r/w h'0000 h'04000112 (h'a4000112)* 16 port l control register plcr r/w h'0000 h'04000114 (h'a4000114)* 16 sc port control register scpcr r/w h'a888 h'04000116 (h'a4000116)* 16 notes: these registers are located in area 1 of physical space. therefore, when the cache is on, either access these registers from the p2 area of logical space or else make an appropriate setting using the mmu so that these registers are not cached. 1. the initial value of the port e, f, g, and h control registers depends on the state of the asemd0 pin. if a low level is input at the asemd0 pin while the resetp pin is asserted, ase mode is entered; if a high level is input, normal mode is entered. see section 23, hitachi user debugging interface (h-udi), for more information on the h-udi. 2. when address translation by the mmu does not apply, the address in parentheses should be used. 614 19.3 register descriptions 19.3.1 port a control register (pacr) bit:1514131211109876543210 pa7 md1 pa7 md0 pa6 md1 pa6 md0 pa5 md1 pa5 md0 pa4 md1 pa4 md0 pa3 md1 pa3 md0 pa2 md1 pa2 md0 pa1 md1 pa1 md0 pa0 md1 pa0 md0 initial value: 0000000000000000 r/w: r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w the port a control register (pacr) is a 16-bit readable/writable register that selects the pin functions. pacr is initialized to h'0000 by a power-on reset, but is not initialized by a manual reset, in standby mode, or in sleep mode. bits 15 and 14pa7 mode 1 and 0 (pa7md1, pa7md0) bits 13 and 12pa6 mode 1 and 0 (pa6md1, pa6md0) bits 11 and 10pa5 mode 1 and 0 (pa5md1, pa5md0) bits 9 and 8pa4 mode 1 and 0 (pa4md1, pa4md0) bits 7 and 6pa3 mode 1 and 0 (pa3md1, pa3md0) bits 5 and 4pa2 mode 1 and 0 (pa2md1, pa2md0) bits 3 and 2pa1 mode 1 and 0 (pa1md1, pa1md0) bits 1 and 0pa0 mode 1 and 0 (pa0md1, pa0md0) these bits select the pin functions and perform input pull-up mos control. bit (2n + 1) bit 2n panmd1 panmd0 pin function 0 0 other function (see table 19.1) (initial value) 0 1 port output 1 0 port input (pull-up mos: on) 1 1 port input (pull-up mos: off) (n = 0 to 7) 615 19.3.2 port b control register (pbcr) bit:1514131211109876543210 pb7 md1 pb7 md0 pb6 md1 pb6 md0 pb5 md1 pb5 md0 pb4 md1 pb4 md0 pb3 md1 pb3 md0 pb2 md1 pb2 md0 pb1 md1 pb1 md0 pb0 md1 pb0 md0 initial value: 0000000000000000 r/w: r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w the port b control register (pbcr) is a 16-bit readable/writable register that selects the pin functions. pbcr is initialized to h'0000 by a power-on reset, but is not initialized by a manual reset, in standby mode, or in sleep mode. bits 15 and 14pb7 mode 1 and 0 (pb7md1, pb7md0) bits 13 and 12pb6 mode 1 and 0 (pb6md1, pb6md0) bits 11 and 10pb5 mode 1 and 0 (pb5md1, pb5md0) bits 9 and 8pb4 mode 1 and 0 (pb4md1, pb4md0) bits 7 and 6pb3 mode 1 and 0 (pb3md1, pb3md0) bits 5 and 4pb2 mode 1 and 0 (pb2md1, pb2md0) bits 3 and 2pb1 mode 1 and 0 (pb1md1, pb1md0) bits 1 and 0pb0 mode 1 and 0 (pb0md1, pb0md0) these bits select the pin functions and perform input pull-up mos control. bit (2n + 1) bit 2n pbnmd1 pbnmd0 pin function 0 0 other function (see table 19.1) (initial value) 0 1 port output 1 0 port input (pull-up mos: on) 1 1 port input (pull-up mos: off) (n = 0 to 7) 616 19.3.3 port c control register (pccr) bit:1514131211109876543210 pc7 md1 pc7 md0 pc6 md1 pc6 md0 pc5 md1 pc5 md0 pc4 md1 pc4 md0 pc3 md1 pc3 md0 pc2 md1 pc2 md0 pc1 md1 pc1 md0 pc0 md1 pc0 md0 initial value: 1010101010101010 r/w: r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w the port c control register (pccr) is a 16-bit readable/writable register that selects the pin functions. pccr is initialized to h'aaaa by a power-on reset, but is not initialized by a manual reset, in standby mode, or in sleep mode. bits 15 and 14pc7 mode 1 and 0 (pc7md1, pc7md0) bits 13 and 12pb6 mode 1 and 0 (pc6md1, pc6md0) bits 11 and 10pc5 mode 1 and 0 (pc5md1, pc5md0) bits 9 and 8pc4 mode 1 and 0 (pc4md1, pc4md0) bits 7 and 6pc3 mode 1 and 0 (pc3md1, pc3md0) bits 5 and 4pc2 mode 1 and 0 (pc2md1, pc2md0) bits 3 and 2pc1 mode 1 and 0 (pc1md1, pc1md0) bits 1 and 0pc0 mode 1 and 0 (pc0md1, pc0md0) these bits select the pin functions and perform input pull-up mos control. bit (2n + 1) bit 2n pcnmd1 pcnmd0 pin function 0 0 other function (see table 19.1) 0 1 port output 1 0 port input (pull-up mos: on) (initial value) 1 1 port input (pull-up mos: off) (n = 0 to 7) 617 19.3.4 port d control register (pdcr) bit:1514131211109876543210 pd7 md1 pd7 md0 pd6 md1 pd6 md0 pd5 md1 pd5 md0 pd4 md1 pd4 md0 pd3 md1 pd3 md0 pd2 md1 pd2 md0 pd1 md1 pd1 md0 pd0 md1 pd0 md0 initial value: 1010101010001010 r/w: r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w the port d control register (pdcr) is a 16-bit readable/writable register that selects the pin functions. pdcr is initialized to h'aa8a by a power-on reset, but is not initialized by a manual reset, in standby mode, or in sleep mode. bits 15 and 14pd7 mode 1 and 0 (pd7md1, pd7md0) bits 11 and 10pd5 mode 1 and 0 (pd5md1, pd5md0) bits 7 and 6pd3 mode 1 and 0 (pd3md1, pd3md0) bits 5 and 4pd2 mode 1 and 0 (pd2md1, pd2md0) bits 3 and 2pd1 mode 1 and 0 (pd1md1, pd1md0) bits 1 and 0pd0 mode 1 and 0 (pd0md1, pd0md0) these bits select the pin functions and perform input pull-up mos control. bit (2n + 1) bit 2n pdnmd1 pdnmd0 pin function 0 0 other function (see table 19.1) (initial value) (n = 2) 0 1 port output 1 0 port input (pull-up mos: on) (initial value) (n = 0, 1, 3, 5, 7) 1 1 port input (pull-up mos: off) (n = 0 to 3, 5, 7) bits 13 and 12pd6 mode 1 and 0 (pd6md1, pd6md0) bits 9 and 8pd4 mode 1 and 0 (pd4md1, pd4md0) these bits select the pin functions and perform input pull-up mos control. bit (2n + 1) bit 2n pdnmd1 pdnmd0 pin function 0 0 other function (see table 19.1) 0 1 reserved 1 0 port input (pull-up mos: on) (initial value) 1 1 port input (pull-up mos: off) (n = 4, 6) 618 19.3.5 port e control register (pecr) bit:1514131211109876543210 pe7 md1 pe7 md0 pe6 md1 pe6 md0 pe5 md1 pe5 md0 pe4 md1 pe4 md0 pe3 md1 pe3 md0 pe2 md1 pe2 md0 pe1 md1 pe1 md0 pe0 md1 pe0 md0 initial value: 1/0 01010101010101/00 r/w: r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w the port e control register (pecr) is a 16-bit readable/writable register that selects the pin functions. pecr is initialized to h'aaaa ( asemd0 = 1) or h'2aa8 ( asemd0 = 0) by a power-on reset, but is not initialized by a manual reset, in software standby mode, or in sleep mode. bits 15 and 14pe7 mode 1 and 0 (pe7md1, pe7md0) bits 13 and 12pe6 mode 1 and 0 (pe6md1, pe6md0) bits 11 and 10pe5 mode 1 and 0 (pe5md1, pe5md0) bits 9 and 8pe4 mode 1 and 0 (pe4md1, pe4md0) bits 7 and 6pe3 mode 1 and 0 (pe3md1, pe3md0) bits 5 and 4pe2 mode 1 and 0 (pe2md1, pe2md0) bits 3 and 2pe1 mode 1 and 0 (pe1md1, pe1md0) bits 1 and 0pe0 mode 1 and 0 (pe0md1, pe0md0) these bits select the pin functions and perform input pull-up mos control. bit (2n + 1) bit 2n penmd1 penmd0 pin function 0 0 reserved (n = 0, 7) (see table 19.1) (initial value) ( asemd0 = 0) 0 1 port output 1 0 port input (pull-up mos: on) (initial value) ( asemd0 = 1) 1 1 port input (pull-up mos: off) (n = 0, 7) bit (2n + 1) bit 2n penmd1 penmd0 pin function 0 0 other function (n = 2, 4, 5) (see table 19.1), reserved (n = 1, 3, 6) 0 1 port output 1 0 port input (pull-up mos: on) (initial value) 1 1 port input (pull-up mos: off) (n = 1 to 6) 619 19.3.6 port f control register (pfcr) bit:1514131211109876543210 pf7 md1 pf7 md0 pf6 md1 pf6 md0 pf5 md1 pf5 md0 pf4 md1 pf4 md0 pf3 md1 pf3 md0 pf2 md1 pf2 md0 pf1 md1 pf1 md0 pf0 md1 pf0 md0 initial value: 1/0 0 1/0 0 1/0 0 1/0 010101010 r/w: r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w the port f control register (pfcr) is a 16-bit readable/writable register that selects the pin functions. pfcr is initialized to h'aaaa (asemd0 = 1) or h'00aa (asemd0 = 0) by a power- on reset, but is not initialized by a manual reset, in standby mode, or in sleep mode. bits 15 and 14pf7 mode 1 and 0 (pf7md1, pf7md0) bits 13 and 12pf6 mode 1 and 0 (pf6md1, pf6md0) bits 11 and 10pf5 mode 1 and 0 (pf5md1, pf5md0) bits 9 and 8pf4 mode 1 and 0 (pf4md1, pf4md0) bits 7 and 6pf3 mode 1 and 0 (pf3md1, pf3md0) bits 5 and 4pf2 mode 1 and 0 (pf2md1, pf2md0) bits 3 and 2pf1 mode 1 and 0 (pf1md1, pf1md0) bits 1 and 0pf0 mode 1 and 0 (pf0md1, pf0md0) these bits select the pin functions and perform input pull-up mos control. bit (2n + 1) bit 2n pfnmd1 pfnmd0 pin function 0 0 other function (see table 19.1) (initial value) ( asemd0 = 0) 0 1 reserved 1 0 port input (pull-up mos: on) (initial value) ( asemd0 = 1) 1 1 port input (pull-up mos: off) (n = 4 to 7) bit (2n + 1) bit 2n pfnmd1 pfnmd0 pin function 0 0 other function (see table 19.1) 0 1 reserved 1 0 port input (pull-up mos: on) (initial value) 1 1 port input (pull-up mos: off) (n = 0 to 3) 620 19.3.7 port g control register (pgcr) bit:1514131211109876543210 pg7 md1 pg7 md0 pg6 md1 pg6 md0 pg5 md1 pg5 md0 pg4 md1 pg4 md0 pg3 md1 pg3 md0 pg2 md1 pg2 md0 pg1 md1 pg1 md0 pg0 md1 pg0 md0 initial value: 1 0 1/0 0 1/0 0 1 0 1/0 0 1/0 0 1/0 0 1/0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w the port g control register (pgcr) is a 16-bit readable/writable register that selects the pin functions. pgcr is initialized to h'aaaa ( asemd0 = 1) or h'8200 ( asemd0 = 0) by a power- on reset, but is not initialized by a manual reset, in standby mode, or in sleep mode. bits 15 and 14pg7 mode 1 and 0 (pg7md1, pg7md0) bits 13 and 12pg6 mode 1 and 0 (pg6md1, pg6md0) bits 11 and 10pg5 mode 1 and 0 (pg5md1, pg5md0) bits 9 and 8pg4 mode 1 and 0 (pg4md1, pg4md0) bits 7 and 6pg3 mode 1 and 0 (pg3md1, pg3md0) bits 5 and 4pg2 mode 1 and 0 (pg2md1, pg2md0) bits 3 and 2pg1 mode 1 and 0 (pg1md1, pg1md0) bits 1 and 0pg0 mode 1 and 0 (pg0md1, pg0md0) these bits select the pin functions and perform input pull-up mos control. bit (2n + 1) bit 2n pgnmd1 pgnmd0 pin function 0 0 other function (n = 0C3, 5) (see table 19.1) (initial value) ( asemd0 = 0) 0 1 reserved 1 0 port input (pull-up mos: on) (initial value) ( asemd0 = 1) 1 1 port input (pull-up mos: off) (n = 0 to 3, 5) bit (2n + 1) bit 2n pgnmd1 pgnmd0 pin function 0 0 other function (n = 4, 6, 7) (see table 19.1) 0 1 reserved 1 0 port input (pull-up mos: on) (initial value) * 1 1 port input (pull-up mos: off) (n = 4, 6, 7) note: * when n = 6, asemd0 /ptg6 functions as asemd0 input while the reset signal is asserted, and as ptg6 input immediately after the reset signal is negated. 621 19.3.8 port h control register (phcr) bit:1514131211109876543210 ph7 md1 ph7 md0 ph6 md1 ph6 md0 ph5 md1 ph5 md0 ph4 md1 ph4 md0 ph3 md1 ph3 md0 ph2 md1 ph2 md0 ph1 md1 ph1 md0 ph0 md1 ph0 md0 initial value: 1 0 1/0 0101010101010 r/w: r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w the port h control register (phcr) is a 16-bit readable/writable register that selects the pin functions. phcr is initialized to h'aaaa ( asemd0 = 1) or h'8aaa ( asemd0 = 0) by a power-on reset, but is not initialized by a manual reset, in standby mode, or in sleep mode. bits 15, 14ph7 mode 1, 0 (ph7md1, ph7md0): these bits select the pin functions and perform input pull-up mos control. bit 15 bit 14 ph7md1 ph7md0 pin function 0 0 other function (see table 19.1) 0 1 port output 1 0 port input (pull-up mos: on) (initial value) 1 1 port input (pull-up mos: off) bits 13 and 12ph6 mode 1 and 0 (ph6md1, ph6md0) bits 11 and 10ph5 mode 1 and 0 (ph5md1, ph5md0) bits 9 and 8ph4 mode 1 and 0 (ph4md1, ph4md0) bits 7 and 6ph3 mode 1 and 0 (ph3md1, ph3md0) bits 5 and 4ph2 mode 1 and 0 (ph2md1, ph2md0) bits 3 and 2ph1 mode 1 and 0 (ph1md1, ph1md0) bits 1 and 0ph0 mode 1 and 0 (ph0md1, ph0md0) these bits select the pin functions and perform input pull-up mos control. 622 bit 13 bit 12 ph6md1 ph6md0 pin function 0 0 other function (see table 19.1) (initial value) ( asemd0 = 0) 0 1 reserved 1 0 port input (pull-up mos: on) (initial value) ( asemd0 = 1) 1 1 port input (pull-up mos: off) bit (2n + 1) bit 2n phnmd1 phnmd0 pin function 0 0 other function (see table 19.1) 0 1 reserved 1 0 port input (pull-up mos: on) (initial value) 1 1 port input (pull-up mos: off) (n = 0 to 5) 623 19.3.9 port j control register (pjcr) bit:1514131211109876543210 pj7 md1 pj7 md0 pj6 md1 pj6 md0 pj5 md1 pj5 md0 pj4 md1 pj4 md0 pj3 md1 pj3 md0 pj2 md1 pj2 md0 pj1 md1 pj1 md0 pj0 md1 pj0 md0 initial value: 0000000000000000 r/w: r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w the port j control register (pjcr) is a 16-bit readable/writable register that selects the pin functions. pjcr is initialized to h'0000 by a power-on reset, but is not initialized by a manual reset, in standby mode, or in sleep mode. bits 15 and 14pj7 mode 1 and 0 (pj7md1, pj7md0) bits 13 and 12pj6 mode 1 and 0 (pj6md1, pj6md0) bits 11 and 10pj5 mode 1 and 0 (pj5md1, pj5md0) bits 9 and 8pj4 mode 1 and 0 (pj4md1, pj4md0) bits 7 and 6pj3 mode 1 and 0 (pj3md1, pj3md0) bits 5 and 4pj2 mode 1 and 0 (pj2md1, pj2md0) bits 3 and 2pj1 mode 1 and 0 (pj1md1, pj1md0) bits 1 and 0pj0 mode 1 and 0 (pj0md1, pj0md0) these bits select the pin functions and perform input pull-up mos control. bit (2n + 1) bit 2n pjnmd1 pjnmd0 pin function 0 0 other function (n = 0, 2, 3, 6, 7) (see table 19.1), reserved (n = 1, 4, 5) (initial value) 0 1 port output 1 0 port input (pull-up mos: on) 1 1 port input (pull-up mos: off) (n = 0 to 7) 624 19.3.10 port k control register (pkcr) bit:1514131211109876543210 pk7 md1 pk7 md0 pk6 md1 pk6 md0 pk5 md1 pk5 md0 pk4 md1 pk4 md0 pk3 md1 pk3 md0 pk2 md1 pk2 md0 pk1 md1 pk1 md0 pk0 md1 pk0 md0 initial value: 0000000000000000 r/w: r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w the port k control register (pkcr) is a 16-bit readable/writable register that selects the pin functions. pkcr is initialized to h'0000 by a power-on reset, but is not initialized by a manual reset, in standby mode, or in sleep mode. bits 15 and 14pk7 mode 1 and 0 (pk7md1, pk7md0) bits 13 and 12pk6 mode 1 and 0 (pk6md1, pk6md0) bits 11 and 10pk5 mode 1 and 0 (pk5md1, pk5md0) bits 9 and 8pk4 mode 1 and 0 (pk4md1, pk4md0) bits 7 and 6pk3 mode 1 and 0 (pk3md1, pk3md0) bits 5 and 4pk2 mode 1 and 0 (pk2md1, pk2md0) bits 3 and 2pk1 mode 1 and 0 (pk1md1, pk1md0) bits 1 and 0pk0 mode 1 and 0 (pk0md1, pk0md0) these bits select the pin functions and perform input pull-up mos control. bit (2n + 1) bit 2n pknmd1 pknmd0 pin function 0 0 other function (see table 19.1) (initial value) 0 1 port output 1 0 port input (pull-up mos: on) 1 1 port input (pull-up mos: off) (n = 0 to 7) 625 19.3.11 port l control register (plcr) bit:1514131211109876543210 pl7 md1 pl7 md0 pl6 md1 pl6 md0 pl5 md1 pl5 md0 pl4 md1 pl4 md0 pl3 md1 pl3 md0 pl2 md1 pl2 md0 pl1 md1 pl1 md0 pl0 md1 pl0 md0 initial value: 0000000000000000 r/w: r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w the port l control register (plcr) is a 16-bit readable/writable register that selects the pin functions. plcr is initialized to h'0000 by a power-on reset, but is not initialized by a manual reset, in standby mode, or in sleep mode. bits 15 and 14pl7 mode 1 and 0 (pl7md1, pl7md0) bits 13 and 12pl6 mode 1 and 0 (pl6md1, pl6md0) bits 11 and 10pl5 mode 1 and 0 (pl5md1, pl5md0) bits 9 and 8pl4 mode 1 and 0 (pl4md1, pl4md0) bits 7 and 6pl3 mode 1 and 0 (pl3md1, pl3md0) bits 5 and 4pl2 mode 1 and 0 (pl2md1, pl2md0) bits 3 and 2pl1 mode 1 and 0 (pl1md1, pl1md0) bits 1 and 0pl0 mode 1 and 0 (pl0md1, pl0md0) these bits select the pin functions and perform input pull-up mos control. bit (2n + 1) bit 2n plnmd1 plnmd0 pin function 0 0 other function (see table 19.1) (initial value) 0 1 reserved 1 0 port input 1 1 port input (n = 0 to 7) when the da0 and da1 pins are used as the d/a converter outputs or when ptl7 and ptl6 are used in the other function state, plcr should by kept at its initial value. 626 19.3.12 sc port control register (scpcr) bit:1514131211109876543210 scp7 md1 scp7 md0 scp6 md1 scp6 md0 scp5 md1 scp5 md0 scp4 md1 scp4 md0 scp3 md1 scp3 md0 scp2 md1 scp2 md0 scp1 md1 scp1 md0 scp0 md1 scp0 md0 initial value: 1010100010001000 r/w: r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w r/w the sc port control register (scpcr) is a 16-bit readable/writable register that selects the pin functions. the setting of scpcr is valid only when transmit/receive operations are disabled in the scscr register. scpcr is initialized to h'a888 by a power-on reset, but is not initialized by a manual reset, in standby mode, or in sleep mode. when the te bit in scscr is set to 1, the other function output state has a higher priority than the scpcr setting for the txd[2:0] pins. when the re bit in scscr is set to 1, the input state has a higher priority than the scpcr setting for the rxd[2:0] pins. bits 15 and 14scp7 mode 1 and 0 (scp7md1, scp7md0): these bits select the pin function and perform input pull-up mos control. bit 15 bit 14 scp7md1 scp7md0 pin function 0 0 other function (see table 19.1) 0 1 reserved 1 0 port input (pull-up mos: on) (initial value) 1 1 port input (pull-up mos: off) bits 13, 12scp6 mode 1, 0 (scp6md1, scp6md0): these bits select the pin function and perform input pull-up mos control. bit 13 bit 12 scp6md1 scp6md0 pin function 0 0 other function (see table 19.1) 0 1 port output 1 0 port input (pull-up mos: on) (initial value) 1 1 port input (pull-up mos: off) 627 bits 11 and 10scp5 mode 1 and 0 (scp5md1, scp5md0): these bits select the pin functions and perform input pull-up mos control. bit 11 bit 10 scp5md1 scp5md0 pin function 0 0 other function (see table 19.1) 0 1 port output 1 0 port input (pull-up mos: on) (initial value) 1 1 port input (pull-up mos: off) bits 9 and 8scp4 mode 1 and 0 (scp4md1, scp4md0): these bits select the pin function and perform input pull-up mos control. bit 9 bit 8 scp4md1 scp4md0 pin function 0 0 transmit data output 2 (txd2) receive data input 2 (rxd2) (initial value) 0 1 general output (scpt[4] output pin) receive data input 2 (rxd2) 1 0 scpt[4] input pin pull-up (input pin) transmit data output 2 (txd2) 1 1 general input (scpt[4] input pin) transmit data output 2 (txd2) note: there is no scpt[4] simultaneous i/o combination because one bit (scp4dt) is accessed using two pins, txd2 and rxd2. when port input is set (bit scpnmd1 is set to 1) and when the te bit in scscr is set to 1, the txd2 pin is in the output state. when the te bit is cleared to 0, the txd2 pin goes to the high- impedance state. bits 7 and 6scp3 mode 1 and 0 (scp3md1, scp3md0): these bits select the pin function and perform input pull-up mos control. bit 7 bit 6 scp3md1 scp3md0 pin function 0 0 other function (see table 19.1) 0 1 port output 1 0 port input (pull-up mos: on) (initial value) 1 1 port input (pull-up mos: off) 628 bits 5 and 4scp2 mode 1 and 0 (scp2md1, scp2md0): these bits select the pin function and perform input pull-up mos control. bit 5 bit 4 scp2md1 scp2md0 pin function 0 0 transmit data output 1 (txd1) receive data input 1 (rxd1) (initial value) 0 1 general output (scpt[2] output pin) receive data input 1 (rxd1) 1 0 scpt[2] input pin pull-up (input pin) transmit data output 1 (txd1) 1 1 general input (scpt[2] input pin) transmit data output 1 (txd1) note: there is no scpt[2] simultaneous i/o combination because one bit (scp2dt) is accessed using two pins, txd1 and rxd1. when port input is set (bit scpnmd1 is set to 1) and when the te bit in scscr is set to 1, the txd1 pin is in the output state. when the te bit is cleared to 0, the txd1 pin goes to the high- impedance state. bits 3 and 2scp1 mode 1 and 0 (scp1md1, scp1md0): these bits select the pin function and perform input pull-up mos control. bit 3 bit 2 scp1md1 scp1md0 pin function 0 0 other function (see table 19.1) 0 1 port output 1 0 port input (pull-up mos: on) (initial value) 1 1 port input (pull-up mos: off) 629 bits 1 and 0scp0 mode 1 and 0 (scp0md1, scp0md0): these bits select the pin function and perform input pull-up mos control. bit 1 bit 0 scp0md1 scp0md0 pin function 0 0 transmit data output 0 (txd0) receive data input 0 (rxd0) (initial value) 0 1 general output (scpt[0] output pin) receive data input 0 (rxd0) 1 0 scpt[0] input pin pull-up (input pin) transmit data output 0 (txd0) 1 1 general input (scpt[0] input pin) transmit data output 0 (txd0) note: there is no scpt[0] simultaneous i/o combination because one bit (scp0dt) is accessed using two pins, txd0 and rxd0. when port input is set (bit scpnmd1 is set to 1) and when the te bit in scscr is set to 1, the txd0 pin is in the output state. when the te bit is cleared to 0, the txd0 pin goes to the high- impedance state. 630 631 section 20 i/o ports 20.1 overview the sh7729r has twelve 8-bit ports (ports a to l and sc). all port pins are multiplexed with other pin functions (the pin function controller (pfc) handles the selection of pin functions and pull-up mos control). each port has a data register which stores data for the pins. 20.2 port a port a is an 8-bit input/output port with the pin configuration shown in figure 20.1. each pin has an input pull-up mos, which is controlled by the port a control register (pacr) in the pfc. pta7 (input/output) / d23 (input/output) pta6 (input/output) / d22 (input/output) pta5 (input/output) / d21 (input/output) pta4 (input/output) / d20 (input/output) pta3 (input/output) / d19 (input/output) pta2 (input/output) / d18 (input/output) pta1 (input/output) / d17 (input/output) pta0 (input/output) / d16 (input/output) port a figure 20.1 port a 20.2.1 register description table 20.1 summarizes the port a register. table 20.1 port a register name abbreviation r/w initial value address access size port a data register padr r/wh'00 h'04000120 (h'a4000120)* 8 notes: this register is located in area 1 of physical space. therefore, when the cache is on, either access this register from the p2 area of logical space or else make an appropriate setting using the mmu so that this register is not cached. * when address translation by the mmu does not apply, the address in parentheses should be used. 632 20.2.2 port a data register (padr) bit: 7 6 5 4 3 2 1 0 pa7dt pa6dt pa5dt pa4dt pa3dt pa2dt pa1dt pa0dt initial value: 0 0 0 0 0 0 0 0 r/w: r/wr/wr/wr/wr/wr/wr/wr/w the port a data register (padr) is an 8-bit readable/writable register that stores data for pins pta7 to pta0. bits pa7dt to pa0dt correspond to pins pta7 to pta0. when the pin function is general output port, if the port is read the value of the corresponding padr bit is returned directly. when the function is general input port, if the port is read the corresponding pin level is read. table 20.2 shows the function of padr. padr is initialized to h'00 by a power-on reset. it retains its previous value in standby mode and sleep mode, and in a manual reset. table 20.2 port a data register (padr) read/write operations panmd1 panmd0 pin state read write 0 0 other function (see table 19.1) padr value value is written to padr, but does not affect pin state. 1 output padr value write value is output from pin. 1 0 input (pull-up mos on) pin state value is written to padr, but does not affect pin state. 1 input (pull-up mos off) pin state value is written to padr, but does not affect pin state. (n = 7 to 0) 633 20.3 port b port b is an 8-bit input/output port with the pin configuration shown in figure 20.2. each pin has an input pull-up mos, which is controlled by the port b control register (pbcr) in the pfc. ptb7 (input/output) / d31 (input/output) ptb6 (input/output) / d30 (input/output) ptb5 (input/output) / d29 (input/output) ptb4 (input/output) / d28 (input/output) ptb3 (input/output) / d27 (input/output) ptb2 (input/output) / d26 (input/output) ptb1 (input/output) / d25 (input/output) ptb0 (input/output) / d24 (input/output) port b figure 20.2 port b 20.3.1 register description table 20.3 summarizes the port b register. table 20.3 port b register name abbreviation r/w initial value address access size port b data register pbdr r/wh'00 h'04000122 (h'a4000122)* 8 notes: this register is located in area 1 of physical space. therefore, when the cache is on, either access this register from the p2 area of logical space or else make an appropriate setting using the mmu so that this register is not cached. * when address translation by the mmu does not apply, the address in parentheses should be used. 634 20.3.2 port b data register (pbdr) bit: 7 6 5 4 3 2 1 0 pb7dt pb6dt pb5dt pb4dt pb3dt pb2dt pb1dt pb0dt initial value: 0 0 0 0 0 0 0 0 r/w: r/wr/wr/wr/wr/wr/wr/wr/w the port b data register (pbdr) is an 8-bit readable/writable register that stores data for pins ptb7 to ptb0. bits pb7dt to pb0dt correspond to pins ptb7 to ptb0. when the pin function is general output port, if the port is read the value of the corresponding pbdr bit is returned directly. when the function is general input port, if the port is read the corresponding pin level is read. table 20.4 shows the function of pbdr. pbdr is initialized to h'00 by a power-on reset. it retains its previous value in standby mode and sleep mode, and in a manual reset. table 20.4 port b data register (pbdr) read/write operations pbnmd1 pbnmd0 pin state read write 0 0 other function (see table 19.1) pbdr value value is written to pbdr, but does not affect pin state. 1 output pbdr value write value is output from pin. 1 0 input (pull-up mos on) pin state value is written to pbdr, but does not affect pin state. 1 input (pull-up mos off) pin state value is written to pbdr, but does not affect pin state. (n = 7 to 0) 635 20.4 port c port c is an 8-bit input/output port with the pin configuration shown in figure 20.3. each pin has an input pull-up mos, which is controlled by the port c control register (pccr) in the pfc. ptc7 (input/output) / pint7 (input) / msc7 (output) ptc6 (input/output) / pint6 (input) / msc6 (output) ptc5 (input/output) / pint5 (input) / msc5 (output) ptc4 (input/output) / pint4 (input) / msc4 (output) ptc3 (input/output) / pint3 (input) / msc3 (output) ptc2 (input/output) / pint2 (input) / msc2 (output) ptc1 (input/output) / pint1 (input) / msc1 (output) ptc0 (input/output) / pint0 (input) / msc0 (output) port c figure 20.3 port c 20.4.1 register description table 20.5 summarizes the port c register. table 20.5 port c register name abbreviation r/w initial value address access size port c data register pcdr r/wh'00 h'04000124 (h'a4000124)* 8 notes: this register is located in area 1 of physical space. therefore, when the cache is on, either access this register from the p2 area of logical space or else make an appropriate setting using the mmu so that this register is not cached. * when address translation by the mmu does not apply, the address in parentheses should be used. 636 20.4.2 port c data register (pcdr) bit: 7 6 5 4 3 2 1 0 pc7dt pc6dt pc5dt pc4dt pc3dt pc2dt pc1dt pc0dt initial value: 0 0 0 0 0 0 0 0 r/w: r/wr/wr/wr/wr/wr/wr/wr/w the port c data register (pcdr) is an 8-bit readable/writable register that stores data for pins ptc7 to ptc0. bits pc7dt to pc0dt correspond to pins ptc7 to ptc0. when the pin function is general output port, if the port is read, the value of the corresponding pcdr bit is returned directly. when the function is general input port, if the port is read, the corresponding pin level is read. table 20.6 shows the function of pcdr. pcdr is initialized to h'00 by a power-on reset, after which the general input port function (pull- up mos on) is set as the initial pin function, and the corresponding pin levels are read. table 20.6 port c data register (pcdr) read/write operations pcnmd1 pcnmd0 pin state read write 0 0 other function (see table 19.1) pcdr value value is written to pcdr, but does not affect pin state. 1 output pcdr value write value is output from pin. 1 0 input (pull-up mos on) pin state value is written to pcdr, but does not affect pin state. 1 input (pull-up mos off) pin state value is written to pcdr, but does not affect pin state. (n = 7 to 0) 637 20.5 port d port d comprises a 6-bit input/output port and 2-bit input port with the pin configuration shown in figure 20.4. each pin has an input pull-up mos, which is controlled by the port d control register (pdcr) in the pfc. ptd7 (input/output) / dack1 (output) ptd6 (input) / dreq1 (input) ptd5 (input/output) / dack0 (output) ptd4 (input) / dreq0 (input) ptd3 (input/output) / wakeup (output) ptd2 (input/output) / resetout (output) ptd1 (input/output) / drak0 (output) ptd0 (input/output) / drak1 (output) port d figure 20.4 port d 20.5.1 register description table 20.7 summarizes the port d register. table 20.7 port d register name abbreviation r/w initial value address access size port d data register pddr r/w or r b'0 * 0 * 0000 h'04000126 (h'a4000126)* 1 8 notes: this register is located in area 1 of physical space. therefore, when the cache is on, either access this register from the p2 area of logical space or else make an appropriate setting using the mmu so that this register is not cached. * means no value. *1 when address translation by the mmu does not apply, the address in parentheses should be used. 638 20.5.2 port d data register (pddr) bit: 7 6 5 4 3 2 1 0 pd7dt pd6dt pd5dt pd4dt pd3dt pd2dt pd1dt pd0dt initial value: 0 * 0 * 0 0 0 0 r/w: r/wr r/wr r/wr/wr/wr/w note: * undefined the port d data register (pddr) is a 6-bit readable/writable and 2-bit read-only register that stores data for pins ptd7 to ptd0. bits pd7dt to pd0dt correspond to pins ptd7 to ptd0. when the pin function is general output port, if the port is read, the value of the corresponding pddr bit is returned directly. when the function is general input port, if the port is read, the corresponding pin level is read. table 20.8 shows the function of pddr. pddr is initialized to b'0*0*0000 by a power-on reset. after initialization, the general input port function (pull-up mos on) is set as the initial pin function, and the corresponding pin levels are read from bits pd7dtpd3dt, pd1dt, and pd0dt. pddr retains its previous value in standby mode and sleep mode, and in a manual reset. note that the low level is read if bits 6 and 4 are read except in general-purpose input. table 20.8 port d data register (pddr) read/write operations pdnmd1 pdnmd0 pin state read write 0 0 other function (see table 19.1) pddr value value is written to pddr, but does not affect pin state. 1 output pddr value write value is output from pin. 1 0 input (pull-up mos on) pin state value is written to pddr, but does not affect pin state. 1 input (pull-up mos off) pin state value is written to pddr, but does not affect pin state. (n = 0, 1, 2, 3, 5, 7) pdnmd1 pdnmd0 pin state read write 0 0 other function (see table 19.1) low level ignored (no effect on pin state) 1 reserved low level ignored (no effect on pin state) 1 0 input (pull-up mos on) pin state ignored (no effect on pin state) 1 input (pull-up mos off) pin state ignored (no effect on pin state) (n = 4, 6) 639 20.6 port e port e is an 8-bit input/output port with the pin configuration shown in figure 20.5. each pin has an input pull-up mos, which is controlled by the port e control register (pecr) in the pfc. pte7 (input/output) / audsync (output) pte6 (input/output) pte5 (input/output) / ce2b (output) pte4 (input/output) / ce2a (output) pte3 (input/output) pte2 (input/output) / ras3u (output) pte1 (input/output) pte0 (input/output) / tdo (output) port e figure 20.5 port e 20.6.1 register description table 20.9 summarizes the port e register. table 20.9 port e register name abbreviation r/w initial value address access size port e data register pedr r/wh'00 h'04000128 (h'a4000128)* 8 notes: this register is located in area 1 of physical space. therefore, when the cache is on, either access this register from the p2 area of logical space or else make an appropriate setting using the mmu so that this register is not cached. * when address translation by the mmu does not apply, the address in parentheses should be used. 640 20.6.2 port e data register (pedr) bit: 7 6 5 4 3 2 1 0 pe7dt pe6dt pe5dt pe4dt pe3dt pe2dt pe1dt pe0dt initial value: 0 0 0 0 0 0 0 0 r/w: r/wr/wr/wr/wr/wr/wr/wr/w the port e data register (pedr) is an 8-bit readable/writable register that stores data for pins pte7 to pte0. bits pe7dt to pe0dt correspond to pins pte7 to pte0. when the pin function is general output port, if the port is read the value of the corresponding pedr bit is returned directly. when the function is general input port, if the port is read the corresponding pin level is read. table 20.10 shows the function of pedr. pedr is initialized to h'00 by a power-on reset, after which the general input port function (pull- up mos on) is set as the initial pin function, and the corresponding pin levels are read. it retains its previous value in standby mode and sleep mode, and in a manual reset. table 20.10 port e data register (pedr) read/write operations penmd1 penmd0 pin state read write 0 0 other function (see table 19.1) pedr value value is written to pedr, but does not affect pin state. 1 output pedr value write value is output from pin. 1 0 input (pull-up mos on) pin state value is written to pedr, but does not affect pin state. 1 input (pull-up mos off) pin state value is written to pedr, but does not affect pin state. (n = 0 to 7) 641 20.7 port f port f is an 8-bit input port with the pin configuration shown in figure 20.6. each pin has an input pull-up mos, which is controlled by the port f control register (pfcr) in the pfc. ptf7 (input) / pint15 (input) / trst (input) ptf6 (input) / pint14 (input) / tms (input) ptf5 (input) / pint13 (input) / tdi (input) ptf4 (input) / pint12 (input) / tck (input) ptf3 (input) / pint11 (input) / irls3 (input) ptf2 (input) / pint10 (input) / irsl2 (input) ptf1 (input) / pint9 (input) / irls1 (input) ptf0 (input) / pint8 (input) / irls0 (input) port f figure 20.6 port f 20.7.1 register description table 20.11 summarizes the port f register. table 20.11 port f register name abbreviation r/w initial value address access size port f data register pfdr r h' ** h'0400012a (h'a400012a)* 1 8 notes: this register is located in area 1 of physical space. therefore, when the cache is on, either access this register from the p2 area of logical space or else make an appropriate setting using the mmu so that this register is not cached. * means no value. *1 when address translation by the mmu does not apply, the address in parentheses should be used. 642 20.7.2 port f data register (pfdr) bit: 7 6 5 4 3 2 1 0 pf7dt pf6dt pf5dt pf4dt pf3dt pf2dt pf1dt pf0dt initial value: * * * * * * * * r/w: r r r r r r r r note: * undefined the port f data register (pfdr) is an 8-bit read-only register that stores data for pins ptf7 to ptf0. bits pf7dt to pf0dt correspond to pins ptf7 to ptf0. when the function is general input port, if the port is read the corresponding pin level is read. table 20.12 shows the function of pfdr. pfdr is initialized by a power-on reset, after which the general input port function (pull-up mos on) is set as the initial pin function, and the corresponding pin levels are read. table 20.12 port f data register (pfdr) read/write operations pfnmd1 pfnmd0 pin state read write 0 0 other function (see table 19.1) h'00 ignored (no effect on pin state) 1 reserved h'00 ignored (no effect on pin state) 1 0 input (pull-up mos on) pin state ignored (no effect on pin state) 1 input (pull-up mos off) pin state ignored (no effect on pin state) (n = 0 to 7) 643 20.8 port g port g comprises a 5-bit input/output port and 3-bit input port with the pin configuration shown in figure 20.7. each pin has an input pull-up mos, which is controlled by the port g control register (pgcr) in the pfc. ptg7 (input) / iois16 (input) ptg6 (input) / asemd0 (input) ptg5 (input) / asebrkak (output) ptg4 (input) / ckio2 (output) ptg3 (input) / audata3 (input/output) ptg2 (input) / audata2 (input/output) ptg1 (input) / audata1 (input/output) ptg0 (input) / audata0 (input/output) port g figure 20.7 port g 20.8.1 register description table 20.13 summarizes the port g register. table 20.13 port g register name abbreviation r/w initial value address access size port g data register pgdr r/wh' ** h'0400012c (h'a400012c)* 1 8 notes: this register is located in area 1 of physical space. therefore, when the cache is on, either access this register from the p2 area of logical space or else make an appropriate setting using the mmu so that this register is not cached. * means no value. *1 when address translation by the mmu does not apply, the address in parentheses should be used. 644 20.8.2 port g data register (pgdr) bit: 7 6 5 4 3 2 1 0 pg7dt pg6dt pg5dt pg4dt pg3dt pg2dt pg1dt pg0dt initial value: * * * * * * * * r/w: r r r r r r r r note: * undefined the port g data register (pgdr) is an 8-bit read-only register that stores data for pins ptg7 to ptg0. bits pg7dt to pg0dt correspond to pins ptg7 to ptg0. when the function is general input port, if the port is read the corresponding pin level is read. table 20.14 shows the function of pgdr. pgdr is initialized by a power-on reset, after which the general input port function (pull-up mos on) is set as the initial pin function, and the corresponding pin levels are read. table 20.14 port g data register (pgdr) read/write operations pgnmd1 pgnmd0 pin state read write 0 0 other function (see table 19.1) h'00 ignored (no effect on pin state) 1 reserved h'00 ignored (no effect on pin state) 1 0 input (pull-up mos on) pin state ignored (no effect on pin state) 1 input (pull-up mos off) pin state ignored (no effect on pin state) (n = 0 to 7) 645 20.9 port h port h comprises a 1-bit input/output port and 5-bit input port with the pin configuration shown in figure 20.8. each pin has an input pull-up mos, which is controlled by the port h control register (phcr) in the pfc. pth7 (input/output) / tclk (output) pth4 (input) / irq4 (input) pth3 (input) / irq3 (input) pth2 (input) / irq2 (input) pth6 (input) / audck (input) pth5 (input) / adtrg (input) pth1 (input) / irq1 (input) pth0 (input) / irq0 (input) port h figure 20.8 port h 20.9.1 register description table 20.15 summarizes the port h register. table 20.15 port h register name abbreviation r/w initial value address access size port h data register phdr r/w or r b'0 ******* h'0400012e (h'a400012e)* 1 8 notes: this register is located in area 1 of physical space. therefore, when the cache is on, either access this register from the p2 area of logical space or else make an appropriate setting using the mmu so that this register is not cached. * means no value. *1 when address translation by the mmu does not apply, the address in parentheses should be used. 646 20.9.2 port h data register (phdr) bit: 7 6 5 4 3 2 1 0 ph7dt ph6dt ph5dt ph4dt ph3dt ph2dt ph1dt ph0dt initial value: 0 * * * * * * * r/w: r/w r r r r r r r note: * undefined the port h data register (phdr) is a 1-bit readable/writable and 7-bit read-only register that stores data for pins pth7 to pth0. bits ph7dt to ph0dt correspond to pins pth7 to pth0. when the pin function is general output port, if the port is read, the value of the corresponding phdr bit is returned directly. when the function is general input port, if the port is read, the corresponding pin level is read. table 20.16 shows the function of phdr. phdr is initialized to b'0******* by a power-on reset, after which the general input port function (pull-up mos on) is set as the initial pin function, and the corresponding pin levels are read. it retains its previous value in standby mode and sleep mode, and in a manual reset. note that the low level is read if bits 6 to 0 are read except in general-purpose input. table 20.16 port h data register (phdr) read/write operations phnmd1 phnmd0 pin state read write 0 0 other function (see table 19.1) phdr value value is written to phdr, but does not affect pin state. 1 output phdr value write value is output from pin. 1 0 input (pull-up mos on) pin state value is written to phdr, but does not affect pin state. 1 input (pull-up mos off) pin state value is written to phdr, but does not affect pin state. (n = 7) phnmd1 phnmd0 pin state read write 0 0 other function (see table 19.1) low level ignored (no effect on pin state) 1 reserved low level ignored (no effect on pin state) 1 0 input (pull-up mos on) pin state ignored (no effect on pin state) 1 input (pull-up mos off) pin state ignored (no effect on pin state) (n = 0 to 6) 647 20.10 port j port j is an 8-bit input/output port with the pin configuration shown in figure 20.9. each pin has an input pull-up mos, which is controlled by the port j control register (pjcr) in the pfc. ptj7 (input/output) / status1 (output) ptj6 (input/output) / status0 (output) ptj5 (input/output) ptj4 (input/output) ptj3 (input/output) / casu (output) ptj2 (input/output) / casl (output) ptj1 (input/output) ptj0 (input/output) / ras3l (output) port j figure 20.9 port j 20.10.1 register description table 20.17 summarizes the port j register. table 20.17 port j register name abbreviation r/w initial value address access size port j data register pjdr r/wh'00 h'04000130 (h'a4000130)* 8 notes: this register is located in area 1 of physical space. therefore, when the cache is on, either access this register from the p2 area of logical space or else make an appropriate setting using the mmu so that this register is not cached. * when address translation by the mmu does not apply, the address in parentheses should be used. 648 20.10.2 port j data register (pjdr) bit: 7 6 5 4 3 2 1 0 pj7dt pj6dt pj5dt pj4dt pj3dt pj2dt pj1dt pj0dt initial value: 0 0 0 0 0 0 0 0 r/w: r/wr/wr/wr/wr/wr/wr/wr/w the port j data register (pjdr) is an 8-bit readable/writable register that stores data for pins ptj7 to ptj0. bits pj7dt to pj0dt correspond to pins ptj7 to ptj0. when the pin function is general output port, if the port is read the value of the corresponding pjdr bit is returned directly. when the function is general input port, if the port is read, the corresponding pin level is read. table 20.18 shows the function of pjdr. pjdr is initialized to h'00 by a power-on reset. it retains its previous value in software standby mode and sleep mode, and in a manual reset. table 20.18 port j data register (pjdr) read/write operations pjnmd1 pjnmd0 pin state read write 0 0 other function (see table 19.1) pjdr value value is written to pjdr, but does not affect pin state. 1 output pjdr value write value is output from pin. 1 0 input (pull-up mos on) pin state value is written to pjdr, but does not affect pin state. 1 input (pull-up mos off) pin state value is written to pjdr, but does not affect pin state. (n = 0 to 7) 649 20.11 port k port k is an 8-bit input/output port with the pin configuration shown in figure 20.10. each pin has an input pull-up mos, which is controlled by the port k control register (pkcr) in the pfc. ptk7 (input/output) / we3 (output) / dqmuu (output) / iciowr (output) ptk6 (input/output) / we2 (output) / dqmul (output) / iciord (output) ptk5 (input/output) / cke (output) ptk4 (input/output) / bs (output) ptk3 (input/output) / cs5 (output) / ce1a (output) ptk2 (input/output) / cs4 (output) ptk1 (input/output) / cs3 (output) ptk0 (input/output) / cs2 (output) port k figure 20.10 port k 20.11.1 register description table 20.19 summarizes the port k register. table 20.19 port k register name abbreviation r/w initial value address access size port k data register pkdr r/wh'00 h'04000132 (h'a4000132)* 8 notes: this register is located in area 1 of physical space. therefore, when the cache is on, either access this register from the p2 area of logical space or else make an appropriate setting using the mmu so that this register is not cached. * when address translation by the mmu does not apply, the address in parentheses should be used. 650 20.11.2 port k data register (pkdr) bit: 7 6 5 4 3 2 1 0 pk7dt pk6dt pk5dt pk4dt pk3dt pk2dt pk1dt pk0dt initial value: 0 0 0 0 0 0 0 0 r/w: r/wr/wr/wr/wr/wr/wr/wr/w the port k data register (pkdr) is an 8-bit readable/writable register that stores data for pins ptk7 to ptk0. bits pk7dt to pk0dt correspond to pins ptk7 to ptk0. when the pin function is general output port, if the port is read, the value of the corresponding pkdr bit is returned directly. when the function is general input port, if the port is read, the corresponding pin level is read. table 20.20 shows the function of pkdr. pkdr is initialized to h'00 by a power-on reset. it retains its previous value in standby mode and sleep mode, and in a manual reset. table 20.20 port k data register (pkdr) read/write operations pknmd1 pknmd0 pin state read write 0 0 other function (see table 19.1) pkdr value value is written to pkdr, but does not affect pin state. 1 output pkdr value write value is output from pin. 1 0 input (pull-up mos on) pin state value is written to pkdr, but does not affect pin state. 1 input (pull-up mos off) pin state value is written to pkdr, but does not affect pin state. (n = 0 to 7) 651 20.12 port l port l is an 8-bit input port with the pin configuration shown in figure 20.11. ptl7 (input) / an7 (input) / da0 (output) ptl6 (input) / an6 (input) / da1 (output) ptl5 (input) / an5 (input) ptl4 (input) / an4 (input) ptl3 (input) / an3 (input) ptl2 (input) / an2 (input) ptl1 (input) / an1 (input) ptl0 (input) / an0 (input) port l figure 20.11 port l 20.12.1 register description table 20.21 summarizes the port l register. table 20.21 port l register name abbreviation r/w initial value address access size port l data register pldr r h'00 h'04000134 (h'a4000134)* 8 notes: this register is located in area 1 of physical space. therefore, when the cache is on, either access this register from the p2 area of logical space or else make an appropriate setting using the mmu so that this register is not cached. * when address translation by the mmu does not apply, the address in parentheses should be used. 652 20.12.2 port l data register (pldr) bit: 7 6 5 4 3 2 1 0 pl7dt pl6dt pl5dt pl4dt pl3dt pl2dt pl1dt pl0dt initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r r the port l data register (pldr) is an 8-bit read-only register that stores data for pins ptl7 to ptl0. bits pl7dt to pl0dt correspond to pins ptl7 to ptl0. when the function is general input port, if the port is read, the corresponding pin level is read. table 20.22 shows the function of pldr. pldr is initialized to h'00 by power-on reset. it retains its previous value in software standby mode and sleep mode, and in a manual reset. as port l also has analog pin functions, it has no pull-up mos. table 20.22 port l data register (pldr) read/write operation plnmd1 plnmd0 pin state read write 0 0 other function (see table 19.1) h'00 ignored (no effect on pin state) 1 reserved h'00 ignored (no effect on pin state) 1 0 input pin state ignored (no effect on pin state) 1 input pin state ignored (no effect on pin state) (n = 0 to 7) 653 20.13 sc port the sc port comprises a 4-bit input/output port, 3-bit output port, and 4-bit input port with the pin configuration shown in figure 20.12. each pin has an input pull-up mos, which is controlled by the sc port control register (scpcr) in the pfc. scpt7 (input) / cts2 (input) / irq5 (input) scpt6 (input/output) / rts2 (output) scpt5 (input/output) / sck2 (input/output) scpt4 (input) / rxd2 (input) scpt4 (output) / txd2 (output) scpt3 (input/output) / sck1 (input/output) scpt2 (input) / rxd1 (input) scpt2 (output) / txd1 (output) scpt1 (input/output) / sck0 (input/output) scpt0 (input) / rxd0 (input) scpt0 (output) / txd0 (output) sc port figure 20.12 sc port 20.13.1 register description table 20.23 summarizes the sc port register. table 20.23 sc port register name abbreviation r/w initial value address access size sc port data register scpdr r/w or r b' * 0000000 h'04000136 (h'a4000136)* 1 8 notes: this register is located in area 1 of physical space. therefore, when the cache is on, either access this register from the p2 area of logical space or else make an appropriate setting using the mmu so that this register is not cached. * means no value. *1 when address translation by the mmu does not apply, the address in parentheses should be used. 654 20.13.2 port sc data register (scpdr) bit: 7 6 5 4 3 2 1 0 scp7dt scp6dt scp5dt scp4dt scp3dt scp2dt scp1dt scp0dt initial value: * 0 0 0 0 0 0 0 r/w: r r/wr/wr/wr/wr/wr/wr/w note: * undefined the port sc data register (scpdr) is a 7-bit readable/writable and 1-bit read-only register that stores data for pins scpt7 to scpt0. bits scp7dt to scp0dt correspond to pins scpt7 to scpt0. when the pin function is general output port, if the port is read, the value of the corresponding scpdr bit is returned directly. when the function is general input port, if the port is read, the corresponding pin level is read. table 20.24 shows the function of scpdr. scpdr is initialized to b'*0000000 by a power-on reset. after initialization, the general input port function (pull-up mos on) is set as the initial pin function, and the corresponding pin levels are read from bits scp7dtscp5dt, scp3dt, and scp1dt. scpdr retains its previous value in standby mode and sleep mode, and in a manual reset. note that the low level is read if bit 7 is read except in general-purpose input. table 20.24 read/write operation of the sc port data register (scpdr) scpnmd1 scpnmd0 pin state read write 0 0 other function (see table 19.1) scpdr value value is written to scpdr, but does not affect pin state. 1 output scpdr value write value is output from pin. 1 0 input (pull-up mos on) pin state value is written to scpdr, but does not affect pin state. 1 input (pull-up mos off) pin state value is written to scpdr, but does not affect pin state. (n = 0 to 6) 655 scpnmd1 scpnmd0 pin state read write 0 0 other function (see table 19.1) low level ignored (no effect on pin state) 1 output low level ignored (no effect on pin state) 1 0 input (pull-up mos on) pin state ignored (no effect on pin state) 1 input (pull-up mos off) pin state ignored (no effect on pin state) (n = 7) 656 657 section 21 a/d converter 21.1 overview the sh7729r includes a 10-bit successive-approximation a/d converter allowing selection of up to eight analog input channels. 21.1.1 features a/d converter features are listed below. ? 10-bit resolution ? eight input channels ? high-speed conversion ? conversion time: maximum 8.1 s per channel (p? = 33 mhz operation) ? three conversion modes ? single mode: a/d conversion on one channel ? multi mode: a/d conversion on one to four channels ? scan mode: continuous a/d conversion on one to four channels ? four 16-bit data registers ? a/d conversion results are transferred for storage into data registers corresponding to the channels. ? sample-and-hold function ? a/d conversion can be externally triggered ? a/d interrupt requested at the end of conversion ? at the end of a/d conversion, an a/d end interrupt (adi) can be requested. 658 21.1.2 block diagram figure 21.1 shows a block diagram of the a/d converter. 10-bit d/a addra addrb addrd bus interface peripheral data bus analog multi- plexer control circuit successive approxi- mation register + C comparator sample-and- hold circuit adi interrupt ( ## an0 an1 an2 an3 an4 an5 an6 an7 /8 /16 adcsr adcr av cc a/d converter adcr: a/d control register adcsr: a/d control/status register addra: a/d data register a addrb: a/d data register b addrc: a/d data register c addrd: a/d data register d legend internal data bus adtrg addrc figure 21.1 block diagram of a/d converter 659 21.1.3 input pins table 21.1 summarizes the a/d converters input pins. the eight analog input pins are divided into two groups: group 0 (an0 to an3), and group 1 (an4 to an7). avcc and avss are the power supply inputs for the analog circuits in the a/d converter. avcc also functions as the a/d converter reference voltage pin. table 21.1 a/d converter pins pin name abbreviation i/o function analog power supply pin avcc input analog power supply analog ground pin avss input analog ground and reference voltage analog input pin 0 an0 input group 0 analog inputs analog input pin 1 an1 input analog input pin 2 an2 input analog input pin3 an3 input analog input pin 4 an4 input group1 analog inputs analog input pin 5 an5 input analog input pin6 an6 input analog input pin7 an7 input a/d external trigger input pin adtrg input external trigger input for starting a/d conversion 660 21.1.4 register configuration table 21.2 summarizes the a/d converters registers. table 21.2 a/d converter registers name abbreviation r/w initial value address access size a/d data register a (high) addrah r h'00 h'04000080 (h'a4000080)* 2 16, 8 a/d data register a (low) addral r h'00 h'04000082 (h'a4000082)* 2 8 a/d data register b (high) addrbh r h'00 h'04000084 (h'a4000084)* 2 16, 8 a/d data register b (low) addrbl r h'00 h'04000086 (h'a4000086)* 2 8 a/d data register c (high) addrch r h'00 h'04000088 (h'a4000088)* 2 16, 8 a/d data register c (low) addrcl r h'00 h'0400008a (h'a400008a)* 2 8 a/d data register d (high) addrdh r h'00 h'0400008c (h'a400008c)* 2 16, 8 a/d data register d (low) addrdl r h'00 h'0400008e (h'a400008e)* 2 8 a/d control/status register adcsr r/(w)* 1 h'00 h'04000090 (h'a4000090)* 2 8 a/d control register adcr r/w h'3f h'04000092 (h'a4000092)* 2 8 notes: these registers are located in area 1 of physical space. therefore, when the cache is on, either access these registers from the p2 area of logical space or else make an appropriate setting using the mmu so that these registers are not cached. *1 only 0 can be written to bit 7, to clear the flag. *2 when address translation by the mmu does not apply, the address in parentheses should be used. 661 21.2 register descriptions 21.2.1 a/d data registers a to d (addra to addrd) bit: 15 14 13 12 11 10 9 8 ad9 ad8 ad7 ad6 ad5 ad4 ad3 ad2 initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r r bit: 7 6 5 4 3 2 1 0 ad1 ad0 initial value: 0 0 0 0 0 0 0 0 r/w: r r r r r r r r n = a to d the four a/d data registers (addra to addrd) are 16-bit read-only registers that store the results of a/d conversion. an a/d conversion produces 10-bit data, which is transferred for storage into the a/d data register corresponding to the selected channel. the upper 8 bits of the result are stored in the upper byte (bits 15 to 8) of the a/d data register. the lower 2 bits are stored in the lower byte (bits 7 and 6). bits 5 to 0 of an a/d data register are reserved bits that are always read as 0. table 21.3 indicates the pairings of analog input channels and a/d data registers. the a/d data registers are initialized to h'0000 by a reset and in standby mode. table 21.3 analog input channels and a/d data registers analog input channel group 0 group 1 a/d data register an0 an4 addra an1 an5 addrb an2 an6 addrc an3 an7 addrd 662 21.2.2 a/d control/status register (adcsr) bit: 7 6 5 4 3 2 1 0 adf adie adst multi cks ch2 ch1 ch0 initial value: 0 0 0 0 0 0 0 0 r/w: r/(w)* r/w r/w r/w r/w r/w r/w r/w note: * write 0 to clear the flag. adcsr is an 8-bit readable/writable register that selects the mode and controls the a/d converter. adcsr is initialized to h'00 by a reset and in standby mode. bit 7a/d end flag (adf): indicates the end of a/d conversion. bit 7: adf description 0 [clearing conditions] (initial value) (1) cleared by reading adf while adf = 1, then writing 0 to adf (2) cleared when dmac is activated by adi interrupt and addr is read 1 [setting conditions] single mode: a/d conversion ends multi mode: a/d conversion ends on all selected channels bit 6a/d interrupt enable (adie): enables or disables the interrupt (adi) requested at the end of a/d conversion. bit 6: adie description 0 a/d end interrupt request (adi) is disabled (initial value) 1 a/d end interrupt request (adi) is enabled bit 5a/d start (adst): starts or stops a/d conversion. the adst bit remains set to 1 during a/d conversion. it can also be set to 1 by external trigger input at the adtrg pin. bit 5: adst description 0 a/d conversion is stopped (initial value) 1 single mode: a/d conversion starts; adst is automatically cleared to 0 when conversion ends scan mode: a/d conversion starts and continues, cycling through the selected channels, until adst is cleared to 0 by software, by a reset, or by a transition to standby mode 663 bit 4multi mode (multi): selects single mode, multi mode or scan mode. for further information on operation in these modes, see section 21.4, operation. bit 4: multi adcr: bit5: scn description 0 single mode (initial value) 1 0 multi mode 1 scan mode bit 3clock select (cks): selects the a/d conversion time. clear the adst bit to 0 before changing the conversion time. bit 3:cks description 0 conversion time = 536 states (maximum) (initial value) 1 conversion time = 266 states (maximum) bits 2 to 0channel select 2 to 0 (ch2 to ch0): these bits and the multi bit select the analog input channels. clear the adst bit to 0 before changing the channel selection. channel selection description ch2 ch1 ch0 single mode (multi = 0) multi mode (multi = 1) 0 0 0 an0 (initial value) an0 1 an1 an0, an1 1 0 an2 an0 to an2 1 an3 an0 to an3 1 0 0 an4 an4 1 an5 an4, an5 1 0 an6 an4 to an6 1 an7 an4 to an7 664 21.2.3 a/d control register (adcr) bit: 76543210 trge1 trge0 scn resvd1 resvd2 initial value: 00000111 r/w: r/w r/w r/w r/w r/w r r r adcr is an 8-bit readable/writable register that enables or disables external triggering of a/d conversion. adcr is initialized to h'07 by a reset and in standby mode. bit 7 and 6trigger enable (trge1, trge0): enables or disables external triggering of a/d conversion. the trge1 and trge0 bits should only be set when conversion is not in progress. bit 7: trge1 bit 6: trge0 description 0 0 a/d conversion does not start when an external trigger is input 01 (initial value) 10 1 1 a/d conversion starts at the falling edge of an input signal from the external trigger pin ( adtrg ) bit 5scan mode (scn): selects multi mode or scan mode when the multi bit is set to 1. see the description of bit 4 in section 21.2.2, a/d control/status register (adcsr). bits 4 and 3reserved (resvd1, resvd2): these bits are always read as 0. the write value should always be 0. bits 2 to 0reserved: these bits are always read as 1. the write value should always be 1. 665 21.3 bus master interface addra to addrd are 16-bit registers, but they are connected to the bus master by the upper 8 bits of the 16-bit peripheral data bus. therefore, although the upper byte can be accessed directly by the bus master, the lower byte is read through an 8-bit temporary register (temp). an a/d data register is read as follows. when the upper byte is read, the upper-byte value is transferred directly to the bus master and the lower-byte value is transferred into temp. next, when the lower byte is read, the temp contents are transferred to the bus master. when reading an a/d data register, always read the upper byte before the lower byte. it is possible to read only the upper byte, but if only the lower byte is read, the read value is not guaranteed. figure 21.2 shows the data flow for access to an a/d data register. see section 21.7.3, access size and read data. bus interface temp [h'40] addrn l [h'40] addrn h [h'aa] n = a to d cpu receives data h'aa upper byte read module internal data bus bus interface temp [h'40] addrn l [h'40] addrn h [h'aa] n = a to d cpu receives data h'40 lower byte read module internal data bus figure 21.2 a/d data register access operation (reading h'aa40) 666 21.4 operation the a/d converter operates by successive approximations with 10-bit resolution. it has three operating modes: single mode, multi mode, and scan mode. 21.4.1 single mode (multi = 0) single mode should be selected when only one a/d conversion on one channel is required. a/d conversion starts when the adst bit is set to 1 by software, or by external trigger input. the adst bit remains set to 1 during a/d conversion and is automatically cleared to 0 when conversion ends. when conversion ends the adf bit is set to 1. if the adie bit is also set to 1, an adi interrupt is requested at this time. to clear the adf flag to 0, first read adcsr, then write 0 to adf. when the mode or analog input channel must be switched during a/d conversion, to prevent incorrect operation, first clear the adst bit to 0 in adcsr to halt a/d conversion. after making the necessary changes, set the adst bit to 1 to start a/d conversion again. the adst bit can be set at the same time as the mode or channel is changed. typical operations when channel 1 (an1) is selected in single mode are described next. figure 21.3 shows a timing diagram for this example. 1. single mode is selected (multi = 0), input channel an1 is selected (ch2 = ch1 = 0, ch0 = 1), the a/d interrupt is enabled (adie = 1), and a/d conversion is started (adst = 1). 2. when a/d conversion is completed, the result is transferred into addrb. at the same time the adf flag is set to 1, the adst bit is cleared to 0, and the a/d converter becomes idle. 3. since adf = 1 and adie = 1, an adi interrupt is requested. 4. the a/d interrupt handling routine starts. 5. the routine reads adcsr, then writes 0 to the adf flag. 6. the routine reads and processes the conversion result (addrb = 0). 7. execution of the a/d interrupt handling routine ends. then, when the adst bit is set to 1, a/d conversion starts and steps 2 to 7 are executed. 667 channel 0 (an0) operating adie adst adf channel 1 (an1) operating channel 2 (an2) operating channel 3 (an3) operating addra addrb addrc addrd waiting waiting waiting waiting waiting waiting a/d conversion starts set set set clear* clear a/d conversion result 1 a/d conversion result 2 read result read result a/d conversion 1 a/d conversion result 2 note: vertical arrows ( ) indicate instruction execution by software. figure 21.3 example of a/d converter operation (single mode, channel 1 selected) 668 21.4.2 multi mode (multi = 1) multi mode should be selected when performing a/d conversions on one or more channels. when the adst bit is set to 1 by software or external trigger input, a/d conversion starts on the first channel in the group (an0 when ch2 = 0, an4 when ch2 = 1). when two or more channels are selected, after conversion of the first channel ends, conversion of the second channel (an1 or an5) starts immediately. when a/d conversions end on the selected channels, the adst bit is cleared to 0. the conversion results are transferred for storage into the a/d data registers corresponding to the channels. when the mode or analog input channel selection must be changed during a/d conversion, to prevent incorrect operation, first clear the adst bit to 0 in adcsr to halt a/d conversion. after making the necessary changes, set the adst bit to 1. a/d conversion will start again from the first channel in the group. the adst bit can be set at the same time as the mode or channel selection is changed. typical operations when three channels in group 0 (an0 to an2) are selected in scan mode are described next. figure 21.4 shows a timing diagram for this example. 1. multi mode is selected (multi = 1), channel group 0 is selected (ch2 = 0), analog input channels an0 to an2 are selected (ch1 = 1, ch0 = 0), and a/d conversion is started (adst = 1). 2. when a/d conversion of the first channel (an0) is completed, the result is transferred into addra. next, conversion of the second channel (an1) starts automatically. 3. conversion proceeds in the same way through the third channel (an2). 4. when conversion of all selected channels (an0 to an2) is completed, the adf flag is set to 1 and adst bit is cleared to 0. if the adie bit is set to 1, an adi interrupt is requested at this time. when the adst bit is cleared to 0, a/d conversion stops. after that, if the adst bit is set to 1, a/d conversion starts again from the first channel (an0). 669 channel 0 (an0) operating adst adf channel 1 (an1) operating channel 2 (an2) operating channel 3 (an3) operating addra addrb addrc addrd waiting waiting waiting waiting set clear clear a/d conversion result 2 waiting waiting a/d conversion result 3 a/d conversion 1 waiting a/d conversion result 1 transfer a/d conversion 3 a/d conversion a/d conversion 2 note: vertical arrows ( ) indicate instruction execution by software. figure 21.4 example of a/d converter operation (multi mode, channels an0 to an2 selected) 670 21.4.3 scan mode (multi = 1, scn = 1) scan mode is useful for monitoring analog inputs in a group of one or more channels. when the adst bit in the a/d control/status register (adcsr) is set to 1 by software or external trigger input, a/d conversion starts on the first channel in the group (an0 when ch2 = 0, an4 when ch2 = 1)). when two or more channels are selected, after conversion of the first channel ends, conversion of the second channel (an1 or an5) starts immediately. a/d conversion continues cyclically on the selected channels until the adst bit is cleared to 0. the conversion results are transferred for storage into the a/d data registers corresponding to the channels. when the mode or analog input channel must be changed during analog conversion, to prevent incorrect operation, first clear the adst bit to 0 to halt a/d conversion. after making the necessary changes, set the adst bit to 1. a/d conversion will start again from the first channel in the group. the adst bit can be set at the same time as the mode or channel selection is changed. typical operations when three channels (an0 to an2) are selected in scan mode are described next. figure 21.5 shows a timing diagram for this example. 1. scan mode is selected (multi = 1, scn = 1), channel group 2 is selected (ch2 = 0), analog input channels an0 to an2 are selected (ch1 = 1, ch0 = 0), and a/d conversion is started (adst = 1). 2. when a/d conversion of the first channel (an0) is completed, the result is transferred into addra. next, conversion of the second channel (an1) starts automatically. 3. conversion proceeds in the same way through the third channel (an2). 4. when conversion of all the selected channels (an0 to an2) is completed, the adf flag is set to 1 and conversion of the first channel (an0) starts again. if the adie bit is set to 1, an adi interrupt is requested at this time. 5. steps 2 to 4 are repeated as long as the adst bit remains set to 1. when the adst bit is cleared to 0, a/d conversion stops. after that, if the adst bit is set to 1, a/d conversion starts again from the first channel (an0). 671 adst adf channel 0 (an 0 ) operating channel 1 (an 1 ) operating channel 2 (an 2 ) operating channel 3 (an 3 ) operating addra addrb addrc addrd waiting waiting waiting waiting waiting waiting waiting waiting waiting transfer a/d conversion 1 a/d conversion 4 a/d conversion 2 a/d conversion 3 a/d conversion result 1 a/d conversion result 4 a/d conversion result 2 a/d conversion result 3 clear * clear * set * continuous a/d conversion a/d conversion 5 note: * vertical arrows ( ) indicate instruction execution by software. figure 21.5 example of a/d converter operation (scan mode, channels an0 to an2 selected) 672 21.4.4 input sampling and a/d conversion time the a/d converter has a built-in sample-and-hold circuit. the a/d converter samples the analog input at a time t d after the adst bit is set to 1, then starts conversion. figure 21.6 shows the a/d conversion timing. table 21.4 indicates the a/d conversion time. as indicated in figure 21.6, the a/d conversion time includes t d and the input sampling time. the length of t d varies depending on the timing of the write access to adcsr. the total conversion time therefore varies within the ranges indicated in table 21.4. in multi mode and scan mode, the values given in table 21.4 apply to the first conversion. in the second and subsequent conversions the conversion time is fixed at 256 states when cks = 0 or 128 states when cks = 1. p? write signal adf * 1 input sampling timing t d a/d conversion start delay t spl input sampling time t conv a/d conversion time notes: 1. adcsr write cycle 2. adcsr address address * 2 t d t spl t conv figure 21.6 a/d conversion timing 673 table 21.4 a/d conversion time (single mode) cks = 0 cks = 1 symbol min typ max min typ max a/d conversion start delay t d 172810 17 input sampling time t spl 12965 a/d conversion time t conv 514 525 259 266 note: values in the table are numbers of states (t cyc ). 21.4.5 external trigger input timing a/d conversion can be externally triggered. when the trge1 and trge0 bits are set to 1 in adcr, external trigger input is enabled at the adtrg pin. a high-to-low transition at the adtrg pin sets the adst bit to 1 in adcsr, starting a/d conversion. other operations, regardless of the conversion mode, are the same as if the adst bit had been set to 1 by software. figure 21.7 shows the timing. a/d conversion p? adtrg external trigger signal adst figure 21.7 external trigger input timing 674 21.5 interrupts the a/d converter generates an interrupt (adi) at the end of a/d conversion. the adi interrupt request can be enabled or disabled by the adie bit in adcsr. 21.6 definitions of a/d conversion accuracy the a/d converter compares an analog value input from an analog input channel with its analog reference value and converts it to 10-bit digital data. the absolute accuracy of this a/d conversion is the deviation between the input analog value and the output digital value. it includes the following errors: ? offset error ? full-scale error ? quantization error ? nonlinearity error these four error quantities are explained below with reference to figure 21.8. in the figure, the 10 bits of the a/d converter have been simplified to 3 bits. offset error is the deviation between actual and ideal a/d conversion characteristics when the digital output value changes from the minimum (zero voltage) 0000000000 (000 in the figure) to 000000001 (001 in the figure)(figure 21.8, item (1)). full-scale error is the deviation between actual and ideal a/d conversion characteristics when the digital output value changes from the 1111111110 (110 in the figure) to the maximum 1111111111 (111 in the figure)(figure 21.8, item (2)). quantization error is the intrinsic error of the a/d converter and is expressed as 1/2 lsb (figure 21.8, item (3)). nonlinearity error is the deviation between actual and ideal a/d conversion characteristics between zero voltage and full-scale voltage (figure 21.8, item (4)). note that it does not include offset, full-scale, or quantization error. 675 111 110 101 100 011 010 001 000 0 1/8 2/8 3/8 4/8 5/8 6/8 7/8 fs analog input voltage fs: full-scale voltage (3) quantization error ideal a/d conversion characteristic (4) nonlinearity error ideal a/d conversion characteristic actual a/d convertion characteristic (2) full-scale error digital output analog input voltage (1) offset error fs digital output figure 21.8 definitions of a/d conversion accuracy 21.7 usage notes when using the a/d converter, note the following points. 21.7.1 setting analog input voltage ? analog input voltage range: during a/d conversion, the voltages input to the analog input pins ann should be in the range av ss ann av cc (n = 0 to 7). ? relationships of av cc and av ss to v cc and v ss : av cc , av ss , v cc and v ss should be related as follows: av cc = v cc 0.3 v and av ss = v ss . 21.7.2 processing of analog input pins to prevent damage from voltage surges at the analog input pins (an0 to an7), connect an input protection circuit like the one shown in figure 21.9. the circuit shown also includes an rc filter to suppress noise. this circuit is shown as an example; the circuit constants should be selected according to actual application conditions. table 21.5 lists the analog input pin specifications and figure 21.10 shows an equivalent circuit diagram of the analog input ports. 676 21.7.3 access size and read data table 21.6 shows the relationship between access size and read data. note the read data obtained with different access sizes, bus widths, and endian modes. the case is shown here in which h'3ff is obtained when av cc is input as an analog input. ff is the data containing the upper 8 bits of the conversion result, and c0 is the data containing the lower 2 bits. 0.01 f 10 f av cc an0 to an7 av ss sh7729r * 1 100 ? 0.1 f note: figure 21.9 example of analog input protection circuit 1.0 k ? an0 to an7 20 pf 1 m ? figure 21.10 analog input pin equivalent circuit 677 table 21.5 analog input pin ratings item min max unit analog input capacitance 20 pf allowable signal-source impedance 5 k ? table 21.6 relationship between access size and read data access bus width 32 bits (d31Cd0) 16 bits (d15Cd0) 8 bits (d7Cd0) size command endian big little big little big little byte access mov.l mov.b mov.l mov.b #addrah,r9 @r9,r8 #addral,r9 @r9,r8 ffffffff c0c0c0c0 ffffffff c0c0c0c0 ffff c0c0 ffff c0c0 ff c0 ff c0 word access mov.l mov.w mov.l mov.w #addrah,r9 @r9,r8 #addral,r9 @r9,r8 ffxxffxx c0xxc0xx ffxxffxx c0xxc0xx ffxx c0xx ffxx c0xx ffxx c0xx xxff xxc0 longword access mov.l mov.l #addrah,r9 @r9,r8 ffxxc0xx ffxxc0xx ffxxc0xx c0xxffxx ffxxc0xx xxc0xxff in this table: #addrah .equ h'04000080 #addral .equ h'04000082 values are shown in hexadecimal for the case where read data is output to an external device via r8. 678 679 section 22 d/a converter 22.1 overview the sh7729r includes a d/a converter with two channels. 22.1.1 features d/a converter features are listed below. ? eight-bit resolution ? two output channels ? conversion time: maximum 10 s (with 20-pf capacitive load) ? output voltage: 0 v to avcc 22.1.2 block diagram figure 22.1 shows a block diagram of the d/a converter. av cc da 0 da 1 dacr dadr0 dadr1 module data bus bus interface on-chip data bus control circuit legend dacr: d/a control register dadr0: d/a data register 0 dadr1: d/a data register 1 8-bit d/a av ss figure 22.1 block diagram of d/a converter 680 22.1.3 i/o pins table 22.1 summarizes the d/a converters input and output pins. table 22.1 d/a converter pins pin name abbreviation i/o function analog power supply pin avcc input analog power supply analog ground pin avss input analog ground and reference voltage analog output pin 0 da0 output analog output, channel 0 analog output pin 1 da1 output analog output, channel 1 22.1.4 register configuration table 22.2 summarizes the d/a converters registers. table 22.2 d/a converter registers name abbreviation r/w initial value address* 1 d/a data register 0 dadr0 r/w h'00 h'040000a0 (h'a40000a0)* 2 d/a data register 1 dadr1 r/w h'00 h'040000a2 (h'a40000a2)* 2 d/a control register dacr r/w h'1f h'040000a4 (h'a40000a4)* 2 notes: these registers are located in area 1 of physical space. therefore, when the cache is on, either access these registers from the p2 area of logical space or else make an appropriate setting using the mmu so that these registers are not cached. *1 lower 16 bits of the address *2 when address translation by the mmu does not apply, the address in parentheses should be used. 681 22.2 register descriptions 22.2.1 d/a data registers 0 and 1 (dadr0/1) bit: 7 6 5 4 3 2 1 0 initial value: 0 0 0 0 0 0 0 0 r/w: r/w r/w r/w r/w r/w r/w r/w r/w the d/a data registers (dadr0 and dadr1) are 8-bit readable/writable registers that store the data to be converted. when analog output is enabled, the d/a data register values are constantly converted and output at the analog output pins. the d/a data registers are initialized to h'00 by a reset. 22.2.2 d/a control register (dacr) bit: 7 6 5 4 3 2 1 0 daoe1 daoe0 dae initial value: 0 0 0 1 1 1 1 1 r/w: r/w r/w r/w r r r r r dacr is an 8-bit readable/writable register that controls the operation of the d/a converter. dacr is initialized to h'1f by a reset. bit 7d/a output enable 1 (daoe1): controls d/a conversion and analog output. bit 7: daoe1 description 0 da1 analog output is disabled (initial value) 1 channel-1 d/a conversion and da1 analog output are enabled bit 6d/a output enable 0 (daoe0): controls d/a conversion and analog output. bit 6: daoe0 description 0 da0 analog output is disabled (initial value) 1 channel-0 d/a conversion and da0 analog output are enabled 682 bit 5d/a enable (dae): controls d/a conversion, together with bits daoe0 and daoe1. when the dae bit is cleared to 0, d/a conversion is controlled independently in channels 0 and 1. when the chip enters standby mode while d/a conversion is enabled, the d/a output is held and the analog power-supply current is equivalent to that during d/a conversion. to reduce the analog power-supply current in standby mode, clear the daoe0 and daoe1 bits and disable the d/a output. bit 7: daoe1 bit 6: daoe0 bit 5: dae description 0 0 d/a conversion is disabled in channels 0 and 1 (initial value) 0 1 0 d/a conversion is enabled in channel 0 d/a conversion is disabled in channel 1 0 1 1 d/a conversion is enabled in channels 0 and 1 1 0 0 d/a conversion is disabled in channel 0 d/a conversion is enabled in channel 1 1 0 1 d/a conversion is enabled in channels 0 and 1 1 1 d/a conversion is enabled in channels 0 and 1 when the dae bit is set to 1, even if bits daoe0 and daoe1 in dacr and the adst bit in adcsr are cleared to 0, the same current is drawn from the analog power supply as during a/d and d/a conversion. bits 4 to 0reserved: read-only bits, always read as 1. 683 22.3 operation the d/a converter has two built-in d/a conversion circuits that can perform conversion independently. d/a conversion is performed constantly while enabled in dacr. if the dadr0 or dadr1 value is modified, conversion of the new data begins immediately. the conversion results are output when bits daoe0 and daoe1 are set to 1. an example of d/a conversion on channel 0 is given next. timing is indicated in figure 22.2. 1. data to be converted is written in dadr0. 2. bit daoe0 is set to 1 in dacr. d/a conversion starts and da0 becomes an output pin. the converted result is output after the conversion time. the output value is (dadr0 contents/256) avcc. output of this conversion result continues until the value in dadr0 is modified or the daoe0 bit is cleared to 0. 3. if the dadr0 value is modified, conversion starts immediately, and the result is output after the conversion time. 4. when the daoe0 bit is cleared to 0, da0 becomes an input pin. dadr0 write cycle t dconv high-impedance state conversion result 1 conversion data 1 conversion data 2 conversion result 2 t dconv ? address bus dadr0 daoe0 da0 t dconv : d/a conversion time legend dacr write cycle dadr0 write cycle dacr write cycle figure 22.2 example of d/a converter operation 684 685 section 23 hitachi user debugging interface (h-udi) 23.1 overview the sh7729r incorporates a hitachi user debugging interface (h-udi) and advanced user debugger (aud) for program debugging. 23.2 hitachi user debugging interface (h-udi) the h-udi (hitachi user debugging interface) performs on-chip debugging which is supported by the sh7729r. the h-udi described here is a serial interface which is compatible with jtag (joint test action group, ieee standard 1149.1 and ieee standard test access port and boundary-scan architecture) specifications. the h-udi in the sh7729r supports a boundary scan mode, and is also used for emulator connection. when using an emulator, h-udi functions should not be used. refer to the emulator manual for the method of connecting the emulator. 23.2.1 pin descriptions tck: h-udi serial data input/output clock pin. data is serially supplied to the h-udi from the data input pin (tdi), and output from the data output pin (tdo), in synchronization with this clock. tms: mode select input pin. the state of the tap control circuit is determined by changing this signal in synchronization with tck. the protocol conforms to the jtag standard (ieee std. 1140.1). trst : h-udi reset input pin. input is accepted asynchronously with respect to tck, and when low, the h-udi is reset. see section 23.2, reset configuration, for more information. tdi: h-udi serial data input pin. data transfer to the h-udi is executed by changing this signal in synchronization with tck. tdo: h-udi serial data output pin. data output from the h-udi is executed by reading this signal in synchronization with tck. asemd0 : ase mode select pin. if a low level is input at the asemd0 pin while the resetp pin is asserted, ase mode is entered; if a high level is input, normal mode is entered. in ase mode, boundary scan and emulator functions can be used. the input level at the asemd0 pin should be held for at least one cycle after resetp negation. 686 asebrkak : dedicated emulator pin 23.2.2 block diagram figure 23.1 shows a block diagram of the h-udi. sdir tck tdo tdi tms trst sdbpr mux sdbsr shift register tap controller decoder local bus figure 23.1 block diagram of h-udi 23.3 register descriptions the h-udi has the following registers. ? sdbpr: bypass register ? sdir: instruction register ? sdbsr: boundary scan register 687 table 23.1 shows the h-udi register configuration. table 23.1 h-udi registers cpu side h-udi side initial name abbreviation r/w size address r/w size value * bypass register sdbpr r/w 1 undefined instruction register sdir r 16 h'04000200 r/w 16 h'ffff boundary register sdbsr r/w undefined note: * initialized when trst pin is low or when tap is in the test-logic-reset state. 23.3.1 bypass register (sdbpr) the bypass register is a 1-bit register that cannot be accessed by the cpu. when sdir is set to the bypass mode, sdbpr is connected between h-udi pins tdi and tdo. 23.3.2 instruction register (sdir) the instruction register (sdir) is a 16-bit read-only register. the register is in bypass mode in its initial state. it is initialized by trst or in the tap test-logic-reset state, and can be written to by the h-udi irrespective of the cpu mode. operation is not guaranteed if a reserved command is set in this register bit: 15 14 13 12 11 10 9 8 ti3 ti2 ti1 ti0 initial value: 1 1 1 1 1 1 1 1 bit: 7 6 5 4 3 2 1 0 initial value: 1 1 1 1 1 1 1 1 bits 15 to 12test instruction bits (ti3 to ti0): cannot be written by the cpu. 688 table 23.2 h-udi commands ti3ti2 ti1 ti0 description 0000 extest 0100 sample/preload 0101 reserved 0110 h-udi reset negate 0111 h-udi reset assert 100 reserved 101 h-udi interrupt 110 reserved 1110 reserved 1111 bypass mode (initial value) 0001 recovery from sleep bits 11 to 0reserved: these bits are always read as 1. 23.3.3 boundary scan register (sdbsr) the boundary scan register (sdbsr) is a shift register, located on the pad, for controlling the input/output pins of the sh7729r. using the extest and sample/preload commands, a boundary scan test conforming to the jtag standard can be carried out. table 23.3 shows the correspondence between sh7729r pins and boundary scan register bits. 689 table 23.3 sh7729r pins and boundary scan register bits bit pin name i/o bit pin name i/o from tdi 308 d1 in 338 d31/ptb7 in 307 d0 in 337 d30/ptb6 in 306 md1 in 336 d29/ptb5 in 305 md2 in 335 d28/ptb4 in 304 nmi in 334 d27/ptb3 in 303 irq0/ irl0 /pth0 in 333 d26/ptb2 in 302 irq1/ irl1 /pth1 in 332 d25/ptb1 in 301 irq2/ irl2 /pth2 in 331 d24/ptb0 in 300 irq3/ irl3 /pth3 in 330 d23/pta7 in 299 irq4/pth4 in 329 d22/pta6 in 298 d31/ptb7 out 328 d21/pta5 in 297 d30/ptb6 out 327 d20/pta4 in 296 d29/ptb5 out 326 d19/pta3 in 295 d28/ptb4 out 325 d18/pta2 in 294 d27/ptb3 out 324 d17/pta1 in 293 d26/ptb2 out 323 d16/pta0 in 292 d25/ptb1 out 322 d15 in 291 d24/ptb0 out 321 d14 in 290 d23/pta7 out 320 d13 in 289 d22/pta6 out 319 d12 in 288 d21/pta5 out 318 d11 in 287 d20/pta4 out 317 d10 in 286 d19/pta3 out 316 d9 in 285 d18/pta2 out 315 d8 in 284 d17/pta1 out 314 d7 in 283 d16/pta0 out 313 d6 in 282 d15 out 312 d5 in 281 d14 out 311 d4 in 280 d13 out 310 d3 in 279 d12 out 309 d2 in 278 d11 out 690 table 23.3 sh7729r pins and boundary scan register bits (cont) bit pin name i/o bit pin name i/o 277 d10 out 247 d12 control 276 d9 out 246 d11 control 275 d8 out 245 d10 control 274 d7 out 244 d9 control 273 d6 out 243 d8 control 272 d5 out 242 d7 control 271 d4 out 241 d6 control 270 d3 out 240 d5 control 269 d2 out 239 d4 control 268 d1 out 238 d3 control 267 d0 out 237 d2 control 266 d31/ptb7 control 236 d1 control 265 d30/ptb6 control 235 d0 control 264 d29/ptb5 control 234 bs /ptk4 in 263 d28/ptb4 control 233 we2 /dqmul/ iciord / ptk6 in 262 d27/ptb3 control 232 we3 /dqmuu/ iciord / ptk7 in 261 d26/ptb2 control 231 audsync /pte7 in 260 d25/ptb1 control 230 cs2 /ptk0 in 259 d24/ptb0 control 229 cs3 /ptk1 in 258 d23/pta7 control 228 cs4 /ptk2 in 257 d22/pta6 control 227 cs5 / ce1a /ptk3 in 256 d21/pta5 control 226 ce2a /pte4 in 255 d20/pta4 control 225 ce2b /pte5 in 254 d19/pta3 control 224 a0 out 253 d18/pta2 control 223 a1 out 252 d17/pta1 control 222 a2 out 251 d16/pta0 control 221 a3 out 250 d15 control 220 a4 out 249 d14 control 219 a5 out 248 d13 control 218 a6 out 691 table 23.3 sh7729r pins and boundary scan register bits (cont) bit pin name i/o bit pin name i/o 217 a7 out 187 cs4 /ptk2 out 216 a8 out 186 cs5 /ce1a/ptk3 out 215 a9 out 185 cs6 / ce1b out 214 a10 out 184 ce2a /pte4 out 213 a11 out 183 ce2b /pte5 out 212 a12 out 182 a0 control 211 a13 out 181 a1 control 210 a14 out 180 a2 control 209 a15 out 179 a3 control 208 a16 out 178 a4 control 207 a17 out 177 a5 control 206 a18 out 176 a6 control 205 a19 out 175 a7 control 204 a20 out 174 a8 control 203 a21 out 173 a9 control 202 a22 out 172 a10 control 201 a23 out 171 a11 control 200 a24 out 170 a12 control 199 a25 out 169 a13 control 198 bs/ptk4 out 168 a14 control 197 rd out 167 a15 control 196 we0 /dqmll out 166 a16 control 195 we1 /dqmlu/ we out 165 a17 control 194 we2 /dqmul/ iciord / ptk6 out 164 a18 control 193 we3 /dqmuu/ iciowr / ptk7 out 163 a19 control 192 rd/ wr out 162 a20 control 191 audsync /pte7 out 161 a21 control 190 cs0 /mcs0 out 160 a22 control 189 cs2 /ptk0 out 159 a23 control 188 cs3 /ptk1 out 158 a24 control 692 table 23.3 sh7729r pins and boundary scan register bits (cont) bit pin name i/o bit pin name i/o 157 a25 control 127 breq in 156 bs/ptk4 control 126 wait in 155 rd control 125 audck/pth6 in 154 we0 /dqmll control 124 iois16 /ptg7 in 153 we1 /dqmlu/ we control 123 asebrkak /ptg5 in 152 we2 /dqmul/ iciord / ptk6 control 122 ptg4 in 151 we3 /dqmuu/ iciowr / ptk7 control 121 audata3/ptg3 in 150 rd/ wr control 120 audata2/ptg2 in 149 audsync /pte7 control 119 audata1/ptg1 in 148 cs0 /mcs0 control 118 audata0/ptg0 in 147 cs2 /ptk0 control 117 adtrg /pth5 in 146 cs3 /ptk1 control 116 irls3 /ptf3/pint11 in 145 cs4 /ptk2 control 115 irls2 /ptf2/pint10 in 144 cs5 / ce1a /ptk3 control 114 irls1 /ptf1/pint9 in 143 cs6 / ce1b control 113 irls0 /ptf0/pint8 in 142 ce2a /pte4 control 112 md0 in 141 ce2b /pte5 control 111 cke/ptk5 out 140 cke/ptk5 in 110 ras3l /ptj0 out 139 ras3l /ptj0 in 109 ras2l /ptj1 out 138 ras2l /ptj1 in 108 casll / casl /ptj2 out 137 casll / casl /ptj2 in 107 caslh / casu /ptj3 out 136 caslh / casu /ptj3 in 106 cashl /ptj4 out 135 cashl /ptj4 in 105 cashh /ptj5 out 134 cashh /ptj5 in 104 dack0/ptd5 out 133 dack0/ptd5 in 103 dack1/ptd7 out 132 dack1/ptd7 in 102 cas2l /pte6 out 131 cas2l /pte6 in 101 cas2h /pte3 out 130 cas2h /pte3 in 100 ras3u /pte2 out 129 ras3u /pte2 in 99 ras2u /pte1 out 128 ras2u /pte1 in 98 back out 693 table 23.3 sh7729r pins and boundary scan register bits (cont) bit pin name i/o bit pin name i/o 97 asebrkak/ptg5 out 65 rxd2/scpt4 in 96 audata3/ptg3 out 64 wakeup /ptd3 in 95 audata2/ptg2 out 63 resetout /ptd2 in 94 audata1/ptg1 out 62 drak0/ptd1 in 93 audata0/ptg0 out 61 drak1/ptd0 in 92 cke/ptk5 control 60 dreq0 /ptd4 in 91 ras3l /ptj0 control 59 dreq1 /ptd6 in 90 ras2l /ptj1 control 58 rxd1/scpt2 in 89 casll / casl /ptj2 control 57 cts2 /irq5/scpt7 in 88 caslh / casu /ptj3 control 56 mcs7 /ptc7/pint7 in 87 cashl /ptj4 control 55 mcs6 /ptc6/pint6 in 86 cashh /ptj5 control 54 mcs5 /ptc5/pint5 in 85 dack0/ptd5 control 53 mcs4 /ptc4/pint4 in 84 dack1/ptd7 control 52 mcs3 /ptc3/pint3 in 83 cas2l /pte6 control 51 mcs2 /ptc2/pint2 in 82 cas2h /pte3 control 50 mcs1 /ptc1/pint1 in 81 ras3u /pte2 control 49 mcs0 /ptc0/pint0 in 80 ras2u /pte1 control 48 md3 in 79 back control 47 md4 in 78 asebrkak /ptg5 control 46 md5 in 77 audata3/ptg3 control 45 status0/ptj6 out 76 audata2/ptg2 control 44 status1/ptj7 out 75 audata1/ptg1 control 43 tclk/pth7 out 74 audata0/ptg0 control 42 irqout out 73 status0/ptj6 in 41 txd0/scpt0 out 72 status1/ptj7 in 40 sck0/scpt1 out 71 tclk/pth7 in 39 txd1/scpt2 out 70 sck0/scpt1 in 38 sck1/scpt3 out 69 sck1/scpt3 in 37 txd2/scpt4 out 68 sck2/scpt5 in 36 sck2/scpt5 out 67 rts2 /scpt6 in 35 rts2/scpt6 out 66 rxd0/scpt0 in 34 mcs7 /ptc7/pint7 out 694 table 23.3 sh7729r pins and boundary scan register bits (cont) bit pin name i/o bit pin name i/o 33 mcs6 /ptc6/pint6 out 15 sck1/scpt3 control 32 mcs5 /ptc5/pint5 out 14 txd2/scpt4 control 31 mcs4 /ptc4/pint4 out 13 sck2/scpt5 control 30 wakeup /ptd3 out 12 rts2/scpt6 control 29 resetout /ptd2 out 11 mcs7 /ptc7/pint7 control 28 mcs3 /ptc3/pint3 out 10 mcs6 /ptc6/pint6 control 27 mcs2 /ptc2/pint2 out 9 mcs5 /ptc5/pint5 control 26 mcs1 /ptc1/pint1 out 8 mcs4 /ptc4/pint4 control 25 mcs0 /ptc0/pint0 out 7 wakeup /ptd3 control 24 drak0/ptd1 out 6 resetout /ptd2 control 23 drak1/ptd0 out 5 mcs3 /ptc3/pint3 control 22 status0/ptj6 control 4 mcs2 /ptc2/pint2 control 21 status1/ptj7 control 3 mcs1 /ptc1/pint1 control 20 tclk/pth7 control 2 mcs0 /ptc0/pint0 control 19 irqout control 1 drak0/ptd1 control 18 txd0/scpt0 control 0 drak1/ptd0 control 17 sck0/scpt1 control to tdo 16 txd1/scpt2 control note: control is an active-low signal. when control is driven low, the corresponding pin is driven by the value of out. 695 23.4 h-udi operation 23.4.1 tap controller figure 23.2 shows the internal states of the tap controller. state transitions basically conform with the jtag standard. test-logic-reset capture-dr shift-dr exit1-dr pause-dr exit2-dr update-dr select-dr-scan run-test/idle 1 0 0 0 0 11 1 1 0 0 1 0 1 1 10 capture-ir shift-ir exit1-ir pause-ir exit2-ir update-ir select-ir-scan 0 0 1 0 0 1 0 1 1 10 0 figure 23.2 tap controller state transitions note: the transition condition is the tms value at the rising edge of tck. the tdi value is sampled at the rising edge of tck; shifting occurs at the falling edge of tck. the tdo value changes at the tck falling edge. the tdo is at high impedance, except with shift- dr (shift-sr) and shift-ir states. during the change to trstn = 0, there is a transition to test-logic-reset asynchronously with tck. 696 23.4.2 reset configuration table 23.4 reset configuration asdmd0 * 1 resetp trst chip state h llnormal reset and h-udi reset h normal reset h lh-udi reset only h normal operation lllreset hold * 2 h normal reset h lh-udi reset only h normal operation notes: * 1 performs main chip mode and ase mode settings asemd0 = h, main chip mode asemd0 = l, ase mode * 2 in ase mode, reset hold is enabled by driving the resetp and trst pins low for a constant cycle. in this state, the cpu does not start up, even if resetp is driven high. when trst is driven high, h-udi operation is enabled, but the cpu does not start up. the reset hold state is cancelled by the following: ? boot request from h-udi (boot sequence) ? another resetp assert (power-on reset) 23.4.3 h-udi reset an h-udi reset is executed by setting an h-udi reset assert command in sdir. an h-udi reset is of the same kind as a power-on reset. an h-udi reset is released by inputting an h-udi reset negate command. h-udi reset assert h-udi reset negate sdir chip internal reset cpu state branch to h'a0000000 figure 23.3 h-udi reset 697 23.4.4 h-udi interrupt the h-udi interrupt function generates an interrupt by setting a command from the h-udi in the sdir. an h-udi interrupt is a general exception/interrupt operation, resulting in a branch to an address based on the vbr value plus offset, and with return by the rte instruction. this interrupt request has a fixed priority level of 15. h-udi interrupts are not accepted in sleep mode or standby mode. 23.4.5 bypass the jtag-based bypass mode for the h-udi pins can be selected by setting a command from the h-udi in sdir. 23.4.6 using h-udi to recover from sleep mode it is possible to recover from sleep mode by setting a command (0001) from the h-udi in sdir. 23.5 boundary scan a command can be set in sdir by the h-udi to place the h-udi pins in the boundary scan mode stipulated by jtag. 23.5.1 supported instructions the sh7729r supports the three essential instructions defined in the jtag standard (bypass, sample/preload, and extest). bypass: the bypass instruction is an essential standard instruction that operates the bypass register. this instruction shortens the shift path to speed up serial data transfer involving other chips on the printed circuit board. while this instruction is executing, the test circuit has no effect on the system circuits. the instruction code is 1111. sample/preload: the sample/preload instruction inputs values from the sh7729r's internal circuitry to the boundary scan register, outputs values from the scan path, and loads data onto the scan path. when this instruction is executing, the sh7729r's input pin signals are transmitted directly to the internal circuitry, and internal circuit values are directly output externally from the output pins. the sh7729r's system circuits are not affected by execution of this instruction. the instruction code is 0100. 698 in a sample operation, a snapshot of a value to be transferred from an input pin to the internal circuitry, or a value to be transferred from the internal circuitry to an output pin, is latched into the boundary scan register and read from the scan path. snapshot latching is performed in synchronization with the rise of tck in the capture-dr state. snapshot latching does not affect normal operation of the sh7729r. in a preload operation, an initial value is set in the parallel output latch of the boundary scan register from the scan path prior to the extest instruction. without a preload operation, when the extest instruction was executed an undefined value would be output from the output pin until completion of the initial scan sequence (transfer to the output latch) (with the extest instruction, the parallel output latch value is constantly output to the output pin). extest: this instruction is provided to test external circuitry when the sh7729r is mounted on a printed circuit board. when this instruction is executed, output pins are used to output test data (previously set by the sample/preload instruction) from the boundary scan register to the printed circuit board, and input pins are used to latch test results into the boundary scan register from the printed circuit board. if testing is carried out by using the extest instruction n times, the nth test data is scanned-in when test data (n-1) is scanned out. data loaded into the output pin boundary scan register in the capture-dr state is not used for external circuit testing (it is replaced by a shift operation). the instruction code is 0000. 23.5.2 points for attention 1. boundary scan mode covers clock-related signals (extal, extal2, xtal, xtal2, ckio). 2. boundary scan mode does not cover reset-related signals ( resetp , resetm , ca). 3. boundary scan mode does not cover h-udi-related signals (tck, tdi, tdo, tms, trst). 4. when a boundary scan test is carried out, ensure that the ckio clock operates constantly. the ckio frequency range is as follows: minimum: 1 mhz maximum: maximum frequency for respective clock mode specified in the cpg section set pins md[2:0] to the clock mode to be used. after powering on, wait for the ckio clock to stabilize before performing a boundary scan test. 5. fix the resetp pin low. 6. fix the ca pin high, and the asemd0 pin low. 699 23.6 usage notes 1. an h-udi command other than an h-udi interrupt, once set, will not be modified as long as another command is not re-issued from the h-udi. an h-udi interrupt command, however, will be changed to a bypass command once set. 2. because chip operations are suspended in standby mode, h-udi commands are not accepted. however, the tap controller remains in operation at this time. 3. the h-udi is used for emulator connection. therefore, h-udi functions cannot be used when using an emulator. 23.7 advanced user debugger (aud) the aud is a function exclusively for use by an emulator. refer to the user's manual for the relevant emulator for details of the aud. 700 701 section 24 electrical characteristics 24.1 absolute maximum ratings table 24.1 shows the absolute maximum ratings. table 24.1 absolute maximum ratings item symbol rating unit power supply voltage (i/o) vccq C0.3 to 4.2 v power supply voltage (internal) vcc vcc C pll1 vcc C pll2 vcc C rtc C0.3 to 2.5 v input voltage (except port l) vin C0.3 to vccq + 0.3 v input voltage (port l) vin C0.3 to avcc + 0.3 v analog power supply voltage avcc C0.3 to 4.6 v analog input voltage v an C0.3 to avcc + 0.3 v operating temperature topr C20 to 75 c storage temperature tstr C55 to 125 c caution: operating the chip in excess of the absolute maximum rating may result in permanent damage. ? order of turning on 1.7 v/1.8 v/1.9 v/2.0 v power (vcc, vcc-pll1, vcc-pll2, vcc-rtc) and 3.3 v power (vccq, avcc): 1. first turn on the 3.3 v power, then turn on the 1.7 v/1.8 v/1.9 v/2.0 v power within 100 s. this interval should be as short as possible. 2. until voltage is applied to all power supplies and a low level is input at the resetp pin, internal circuits remain unsettled, and so pin states are also undefined. the system design must ensure that these undefined states do not cause erroneous system operation. waveforms at power-on are shown in the following figure. 702 pin states undefined (max. 100 s) 3.3 v 1.7 v/1.8 v/ 1.9 v/2.0 v power 3.3 v power 1.7 v/1.8 v/ 1.9 v/2.0 v power resetp all other pins * pin states undefined power-on reset state +6 * except power/gnd, clock related, and analog pins power-on sequence ? power-off order 1. in the reverse order of powering-on, first turn off the 1.7 v/1.8 v/1.9 v/2.0 v power, then turn off the 3.3 v power within 100 s. this interval should be as short as possible. 2. pin states are undefined while only the 1.7 v/1.8 v/1.9 v/2.0 v power is off. the system design must ensure that these undefined states do not cause erroneous system operation. 703 24.2 dc characteristics tables 24.2 and 24.3 list the dc characteristics. table 24.2 dc characteristics (vccq = 3.3 0.3 v, vcc = 1.55 to 2.15 v, avcc = 3.3 0.3 v, ta = C20 to 75c) item symbol min typ max unit test conditions power supply voltage vccq 3.0 3.3 3.6 v vcc, 1.85 2.00 2.15 200 mhz models vcc-pll1, 1.75 1.90 2.05 167 mhz models vcc-pll2, 1.65 1.80 2.05 100/133 mhz models vcc-rtc 1.55 1.70 1.95 100 mhz models current dissipation normal operation icc 510 820 ma vcc = 2.0 v i = 200 mhz 400 650 vcc = 1.9 v i = 167 mhz 310 500 vcc = 1.8 v i = 133 mhz 230 380 vcc = 1.7 v i = 100 mhz iccq 20 40 vccq = 3.3 v b = 33 mhz in sleep mode* 1 icc 15 30 *1: no external bus cycles except refresh cycles iccq 10 20 b = 33 mhz in standby mode icc 40 120 a ta = 25c (rtc on) vccq = 3.3 v iccq 10 30 vcc = 1.55 v to 2.15 v icc 290 900 ta = 25c (rtc on) vccq = 3.3 v iccq 10 30 vcc = 1.55 v to 2.15 v 704 table 24.2 dc characteristics (cont) (vccq = 3.3 0.3 v, vcc = 1.55 to 2.15 v, avcc = 3.3 0.3 v, ta = C20 to 75c) item symbol min typ max unit test conditions input high voltage resetp , resetm , nmi, irq5 C irq0 , md5Cmd0, irl3 C irl0 , irls3 C irls0 , pint15C pint0, asemd0 , adtrg , trst , extal, ckio, rxd1, ca v ih vccq 0.9 vccq + 0.3 v extal2 if a crystal resonator is not connected, connect to vcc. port l 2.0 avcc + 0.3 other input pins 2.0 vccq + 0.3 input low voltage resetp , resetm , nmi, irq5 C irq0 , md5Cmd0, irl3 C irl0 , irls3 C irls0 , pint15C pint0, asemd0 , adtrg , trst , extal, ckio, rxd1, ca v il C0.3 vccq 0.1 v extal2 if a crystal resonator is not connected, connect to vcc. port l C0.3 avcc 0.2 other input pins C0.3 vccq 0.2 705 table 24.2 dc characteristics (cont) (vccq = 3.3 0.3 v, vcc = 1.55 to 2.15 v, avcc = 3.3 0.3 v, ta = C20 to 75c) item symbol min typ max unit test conditions input leak current all input pins i iin i 1.0 a vin = 0.5 to vccqC0.5 v three-state leak current i/o, all output pins (off condition) i isti i 1.0 a vin = 0.5 to vccqC0.5 v output high voltage all output pins v oh 2.4 v vccq = 3.0 v, i oh = C200 a 2.0 vccq = 3.0 v, i oh = C2 ma v ol 0.55 vccq = 3.6 v, i ol = 1.6 ma pull-up resistance port pin ppull 30 60 120 k ? pin capacity all pins c 10 pf analog power supply voltage avcc 3.0 3.3 3.6 v analog power supply during a/d conversion aicc 0.8 2 ma current during a/d and d/a conversion 2.4 6 ma idle 1.0 20 a ta = 25c notes: 1. when the pll circuits are not used, connect vcc-pll1 and vcc-pll2 to vcc, and vss- pll1 and vss-pll2 to vss. 2. power must be supplied between vcc-rtc and vss-rtc even if the rtc is not used. 3. avcc conditions must be: vccq C 0.3 v avcc vccq + 0.3 v. if the a/d and d/a converters are not used, do not leave the avcc and avss pins open. connect avcc to vccq, and connect avss to vssq. 4. current dissipation values shown are for v ih min = vccq C 0.5 v and v il max = 0.5 v with all output pins unloaded. 706 table 24.3 permissible output current values (vccq = 3.3 0.3 v, vcc = 1.55 v to 2.15 v, avcc = 3.3 0.3 v, ta = C20 to 75c) item symbol min typ max unit permissible output low current (per pin) i ol 2.0 ma permissible output low current (total) i ol 120 ma permissible output low current (per pin) Ci oh 2.0 ma permissible output low current (total) (Ci oh )40ma note: to ensure chip reliability, do not exceed the output current values given in table 24.3. 24.3 ac characteristics in general, sh7729r input should be synchronous. observe the setup and hold times for each input signal unless otherwise specified. table 24.4 maximum operating frequencies (vccq = 3.3 0.3 v, vccq = 1.55 v to 2.15 v, avcc = 3.3 0.3 v, ta = C20 to 75c) item symbol min * typ max unit remarks operating frequency cpu, cache, tlb f 25 200 mhz 200 mhz models 167 167 mhz models 133 133 mhz models 100 100 mhz models external bus 30 66.67 200 mhz models 25 167 mhz, 133 mhz, 100 mhz models peripheral module 7.5 33.34 200 mhz models 6.25 167 mhz, 133 mhz, 100 mhz models note: * the min value depends on the clock mode used. see table 10.4, available combinations of clock mode and frqcr values. 707 24.3.1 clock timing table 24.5 clock timing (1) (vccq = 3.3 0.3 v, vccq = 1.55 v to 2.15 v, avcc = 3.3 0.3 v, ta = C20 to 75c, maximum external bus operating frequency: 25 mhz) item symbol min max unit figure extal clock input frequency f ex 6.25 25 mhz 24.1 extal clock input cycle time t excyc 40 160 ns extal clock input low pulse width t exl 8ns extal clock input high pulse width t exh 8ns extal clock input rise time t exr 4ns extal clock input fall time t exf 4ns ckio clock input frequency f cki 25 25 mhz 24.2 ckio clock input cycle time t ckicyc 40 40 ns ckio clock input low pulse width t ckil 8ns ckio clock input high pulse width t ckih 8ns ckio clock input rise time t ckir 4ns ckio clock input fall time t ckif 4ns ckio clock output frequency f op 25 25 mhz 24.3 ckio clock output cycle time t cyc 40 40 ns ckio clock output low pulse width t ckol 20 ns ckio clock output high pulse width t ckoh 10.6 ns ckio clock output rise time t ckor 7ns ckio clock output fall time t ckof 7ns ckio2 clock output delay time t ck2d -3 3 ns ckio2 clock output rise time t ck2or 7ns ckio2 clock output fall time t ck2of 7ns power-on oscillation settling time t osc1 10 ms 24.4 resetp setup time t resps 20 ns 24.4, 24.5 resetm setup time t resms 6ns resetp assert time t respw 20 tcyc resetm assert time t resmw 20 tcyc standby return oscillation settling time 1 t osc2 10 ms 24.5 standby return oscillation settling time 2 t osc3 10 ms 24.6 standby return oscillation settling time 3 t osc4 11 ms 24.7 708 table 24.5 clock timing (1) (cont) (vccq = 3.3 0.3 v, vcc = 1.55 v to 2.15 v, avcc = 3.3 0.3 v, ta = C20 to 75c, maximum external bus operating frequency: 25 mhz) item symbol min max unit figure pll synchronization settling time 1 (standby release) t pll1 100 s 24.8, 24.9 pll synchronization settling time 2 (multiplication change) t pll2 100 s 24.10 irq/irl interrupt determination time (rtc used and standby mode) t irqstb 100 s 24.9 709 table 24.5 clock timing (2) (vccq = 3.3 0.3 v, vcc = 1.55 v to 2.15 v, avcc = 3.3 0.3 v, ta = C20 to 75c, maximum external bus operating frequency: 33 mhz) item symbol min max unit figure extal clock input frequency f ex 6.25 33 mhz 24.1 extal clock input cycle time t excyc 30.3 160 ns extal clock input low pulse width t exl 7ns extal clock input high pulse width t exh 7ns extal clock input rise time t exr 4ns extal clock input fall time t exf 4ns ckio clock input frequency f cki 25 33 mhz 24.2 ckio clock input cycle time t ckicyc 30.3 40 ns ckio clock input low pulse width t ckil 7ns ckio clock input high pulse width t ckih 7ns ckio clock input rise time t ckir 3ns ckio clock input fall time t ckif 3ns ckio clock output frequency f op 25 33 mhz 24.3 ckio clock output cycle time t cyc 30.3 40 ns ckio clock output low pulse width t ckol 8ns ckio clock output high pulse width t ckoh 8ns ckio clock output rise time t ckor 6ns ckio clock output fall time t ckof 6ns ckio2 clock output delay time t ck2d -3 3 ns ckio2 clock output rise time t ck2or 7ns ckio2 clock output fall time t ck2of 7ns power-on oscillation settling time t osc1 10 ms 24.4 resetp setup time t resps 20 ns 24.4, 24.5 resetm setup time t resms 6ns resetp assert time t respw 20 tcyc resetm assert time t resmw 20 tcyc standby return oscillation settling time 1 t osc2 10 ms 24.5 standby return oscillation settling time 2 t osc3 10 ms 24.6 standby return oscillation settling time 3 t osc4 11 ms 24.7 pll synchronization settling time 1 (standby release) t pll1 100 s 24.8, 24.9 pll synchronization settling time 2 (multiplication change) t pll2 100 s 24.10 irq/irl interrupt determination time (rtc used and standby mode) t irqstb 100 s 24.9 710 table 24.5 clock timing (3) (vccq = 3.3 0.3 v, vcc = 1.55 v to 2.15 v, avcc = 3.3 0.3 v, ta = C20 to 75c, maximum external bus operating frequency: 66 mhz) item symbol min max unit figure extal clock input frequency f ex 6.25 66 mhz 24.1 extal clock input cycle time t excyc 15.2 160 ns extal clock input low pulse width t exl 4ns extal clock input high pulse width t exh 4ns extal clock input rise time t exr 2ns extal clock input fall time t exf 2ns ckio clock input frequency f cki 25 66 mhz 24.2 ckio clock input cycle time t ckicyc 15.2 40 ns ckio clock input low pulse width t ckil 4ns ckio clock input high pulse width t ckih 4ns ckio clock input rise time t ckir 2ns ckio clock input fall time t ckif 2ns ckio clock output frequency f op 25 66 mhz 24.3 ckio clock output cycle time t cyc 15.2 40 ns ckio clock output low pulse width t ckol 3ns ckio clock output high pulse width t ckoh 3ns ckio clock output rise time t ckor 5ns ckio clock output fall time t ckof 5ns ckio2 clock output delay time t ck2d -3 3 ns ckio2 clock output rise time t ck2or 7ns ckio2 clock output fall time t ck2of 7ns power-on oscillation settling time t osc1 10 ms 24.4 resetp setup time t resps 20 ns 24.4, 24.5 resetm setup time t resms 6ns resetp assert time t respw 20 tcyc resetm assert time t resmw 20 tcyc standby return oscillation settling time 1 t osc2 10 ms 24.5 standby return oscillation settling time 2 t osc3 10 ms 24.6 standby return oscillation settling time 3 t osc4 11 ms 24.7 pll synchronization settling time 1 (standby release) t pll1 100 s 24.8, 24.9 pll synchronization settling time 2 (multiplication change) t pll2 100 s 24.10 irq/irl interrupt determination time (rtc used and standby mode) t irqstb 100 s 24.9 711 t exh t exf t exr t exl t excyc v ih v ih v ih 1/2 v cc q 1/2 v cc q v il v il extal * (input) note: * when clock is input from extal pin figure 24.1 extal clock input timing t ckih t ckif t ckir t ckil t ckicyc v ih 1/2 v cc q 1/2 v cc q v ih v il v ih v il ckio (input) figure 24.2 ckio clock input timing t cyc t ckol t ckoh v oh v oh 1/2v cc q ckio (output) ckio2 (output) 1/2v cc q t ckor t ckof v oh v ol v ol v oh t ck2d t ck2d t ck2or t ck2of figure 24.3 ckio clock output timing 712 v cc min t resp/mw t resp/ms t osc1 v cc resetp resetm ckio, internal clock stable oscillation note: oscillation settling time when on-chip oscillator is used figure 24.4 power-on oscillation settling time ckio, internal clock stable oscillation standby t osc2 t resp/mw resetp resetm note: oscillation settling time when on-chip oscillator is used figure 24.5 oscillation settling time on return from standby (return by reset) 713 ckio, internal clock stable oscillation standby t osc3 nmi note: oscillation settling time when using on-chip oscillator in oscillation stop mode wakeup figure 24.6 oscillation settling time on return from standby (return by nmi) ckio, internal clock stable oscillation standby t osc4 irl3 to irl0 irq4 to irq0 pint0/1 note: oscillation settling time when using on-chip oscillator in oscillation stop mode (only when using rtc) wakeup figure 24.7 oscillation settling time on return from standby (return by irq4 to irq0 , pint0/1 and irl3 to irl0 ) 714 extal input or ckio input stable input clock reset or nmi interrupt request stable input clock normal normal standby pll output, ckio output internal clock status 0 status 1 pll synchronization note: pll oscillation settling time when clock is input from extal pin or ckio pin in oscillation continuation mode t pll1 pll synchronization figure 24.8 pll synchronization settling time in case of reset or nmi extal input or ckio input stable input clock irq4 C irq0/ irl3 C irl0 interrupt request stable input clock normal normal pll output, ckio output internal clock status 0 status 1 note: pll oscillation settling time when clock is input from extal pin or ckio pin in oscillation continuation mode t pll1 pll synchronization standby pll synchronization t irlstb figure 24.9 pll synchronization settling time in case of irq/irl interrupt 715 extal input * 1 multiplication ratio change ckio output * 2 , pll output internal clock notes: *1 cklo input in clock mode 7 *2 pll output in clock mode 7 t pll2 figure 24.10 pll synchronization settling time in case of irq/irl interrupt 716 24.3.2 control signal timing table 24.6 control signal timing (vccq = 3.3 0.3 v, vcc = 1.55 v to 2.15 v, avcc = 3.3 0.3 v, ta = C20 to 75c) C66* 2 item symbol min max unit figure resetp pulse width t respw 20 * 3 tcyc 24.11, resetp setup time *1 t resps 20 ns 24.12 resetp hold time t resph 4ns resetm pulse width t resmw 20 * 4 tcyc resetm setup time t resms 6ns resetm hold time t resmh 34 ns breq setup time t breqs 6 ns 24.14 breq hold time t breqh 4ns nmi setup time * 1 t nmis 10 ns 24.12 nmi hold time t nmih 4ns irq5Cirq0 setup time * 1 t irqs 10 ns irq5Cirq0 hold time t irqh 4ns irqout delay time t irqod 10 ns 24.13 back delay time t backd 10 ns 24.14, status1, status0 delay time t std 10 ns 24.15 bus tri-state delay time 1 t boff1 015ns bus tri-state delay time 2 t boff2 015ns bus buffer-on time 1 t bon1 015ns bus buffer-on time 2 t bon2 015ns notes: 1. resetp , nmi, and irq5 to irq0 are asynchronous. changes are detected at the clock fall when the setup time shown is used. if the setup time cannot be used, detection can be delayed until the next clock falls. 2. the upper limit of the external bus clock is 66 mhz. 3. t respw = t osc1 (100 s) when xtal oscillation is continued in standby mode, and t respw = t osc2 (10 ms) when oscillation is stopped. in sleep mode, t respw = t pll1 (100 s). when the clock multiplication ratio is changed, t respw = t pll1 (100 s). 4. in standby mode, t resmw = t osc2 (10 ms). in sleep mode, resetm must be kept low until status (0-1) changes to reset (hh). when the clock multiplication ratio is changed, resetm must be kept low until status (0-1) changes to reset (hh). 717 ckio t resps/ms t resps/ms resetp resetm t respw/mw figure 24.11 reset input timing ckio resetp resetm t resph/mh t resps/ms v ih v il nmi t nmih t nmis v ih v il irq5 to irq0 t irqh t irqs v ih v il figure 24.12 interrupt signal input timing ckio t irqod t irqod irqout figure 24.13 irqout timing 718 ckio breq back rd , rd/ wr , ras , cas , csn , wen , bs , a25 to a0, d31 to d0 t backd t boff2 t boff1 t bon1 t backd t bon2 t breqh t breqh t breqs t breqs mcsn figure 24.14 bus release timing ckio t std t boff2 t boff1 t std t bon2 t bon1 normal mode standby mode normal mode status 0 status 1 rd , rd/ wr , ras , cas , csn , wen , bs , mcsn a25 to a0, d31 to d0 figure 24.15 pin drive timing at standby 719 24.3.3 ac bus timing table 24.7 bus timing (clock modes 0/1/2/7, vccq = 3.3 0.3 v, vcc = 1.55 v to 2.15 v, avcc = 3.3 0.3 v, ta = C20 to 75c) C66* item symbol min max unit figure address delay time t ad 1.5 12 ns 24.16C24.36, 24.39C24.46 address setup time t as 0 ns 24.16C24.18 address hold time t ah 4 ns 24.16C24.21 bs delay time t bsd 10 ns 24.16C24.36, 24.39C24.46 cs delay time 1 t csd1 10 ns 24.16C24.36, 24.39C24.46 cs delay time 2 t csd3 10 ns 24.16C24.21 cs delay time 3 (sdram access) t csd3 1.5 10 ns 24.22C24.39 read/write delay time t rwd 1.5 10 ns 24.16C24.36, 24.39C24.46 read/write hold time t rwh 0 ns 24.16C24.21 read strobe delay time t rsd 10 ns 24.16C24.21, 24.40C24.43 read data setup time 1 t rds1 6 ns 24.16C24.21, 24.40C24.46 read data setup time 2 t rds2 5 ns 24.22C24.25, 24.30C24.33 read data hold time 1 t rdh1 0 ns 24.16C24.25, 24.40C24.46 read data hold time 2 t rdh2 1 ns 24.22C24.25, 24.30C24.33 write enable delay time t wed 10 ns 24.16C24.18, 24.40, 24.41 write data delay time 1 t wdd1 14 ns 24.16C24.18, 24.40, 24.41, 24.44C24.46 write data delay time 2 t wdd2 1.5 12 ns 24.26C24.29, 24.34C24.36 720 table 24.7 bus timing (cont) (clock modes 0/ 1/ 2/ 7, vccq = 3.3 0.3 v, vcc = 1.55 v to 2.15 v, avcc = 3.3 0.3 v, ta = C20 to 75c) C66* item symbol min max unit figure write data hold time 1 t wdh1 1.5 ns 24.16C24.18, 25.40, 25.41, 24.44C24.46 write data hold time 2 t wdh2 1.5 ns 24.26C24.29, 24.34C24.36 write data hold time 3 t wdh3 2 ns 24.16C24.18 write data hold time 4 t wdh4 2 ns 24.40, 24.41, 24.44C24.46 wait setup time t wts 5 ns 24.17C24.21, 24.41, 24.43, 24.45, 24.46 wait hold time t wth 0 ns 24.17C24.21, 24.41, 24.43, 24.45, 24.46 ras delay time 2 t rasd2 1.5 10 ns 24.22C24.39 cas delay time 2 t casd2 1.5 10 ns 24.22C24.39 dqm delay time t dqmd 1.5 10 ns 24.22C24.36 cke delay time t cked 1.5 10 ns 24.38 iciord delay time t icrsd 10 ns 24.44C24.46 iciowr delay time t icwsd 10 ns 24.44C24.46 iois16 setup time t io16s 6 ns 24.45, 24.46 iois16 hold time t io16h 4 ns 24.45, 24.46 dack delay time 1 (based on cklo rise) t dakd1 10 ns 24.16C24.36, 24.39C24.46 dack delay time 2 (based on cklo fall) t dakd2 10 ns 24.16C24.21 note: * the upper limit of the external bus clock is 66 mhz. 721 24.3.4 basic timing t 1 ckio a25 to a0 csn rd/ wr rd d31 to d0 (read) wen d31 to d0 (write) bs t 2 t ad t as t ah t ad t csd1 t rwd t rsd t csd2 t wed t wdd1 t rds1 t bsd t bsd t dakd1 t dakd1 t rdh1 t rdh1 t wed t rsd t ah t rwh t rwd t wdh1 t rwh t rwh t ah t wdh3 dackn (read) (write) figure 24.16 basic bus cycle (no wait) 722 t 1 t w t 2 ckio a25 to a0 csn rd/ wr rd d31 to d0 (read) wen d31 to d0 (write) bs wait t ad t ad t rwd t rwh t ah t ah t rsd t csd1 t wed t wdd1 t bsd t wts t wth t bsd t rds1 t csd2 t wed t rsd t rdh1 t rdh1 t rwd t ah t rwh t wdh3 t wdh1 t rwh t dakd1 t dakd1 dackn (read) (write) t as figure 24.17 basic bus cycle (one wait) 723 t 1 t w t w t 2 ckio a25 to a0 csn rd/ wr rd d31 to d0 (read) wen d31 to d0 (write) bs wait t ad t ad t rwd t rsd t wed t wts t wth t bsd t bsd t rds1 t wts t wth t csd1 t csd2 t rsd t wed t rdh1 t ah t rwh t rdh1 t ah t rwh t rwd t rwh t ah t wdh3 t wdh1 t dakd1 t dakd1 dackn t wdd1 (read) (write) t as figure 24.18 basic bus cycle (external wait, waitsel = 1) 724 24.3.5 burst rom timing ckio a25 to a4 a3 to a0 csn rd/ wr rd d31 to d0 bs dackn wait t ad t ad t ad t ad t csd1 t rwd t bsd t bsd t ah t bsd t dakd1 t dakd1 t csd2 t rsd t rds1 t wts t wth t rds t rsd t 1 t b2 t b1 t b2 t b1 t b2 t b1 t 2 t rsd t rdh1 t rsd t ah t rdh1 t rwh t ah t rwh t rwd t rdh1 t bsd note: in the write cycle, the basic bus cycle is performed. figure 24.19 burst rom bus cycle (no wait) 725 ckio a25 to a4 a3 to a0 csn rd/ wr rd d31 to d0 bs wait dackn t ad t ad t ad t csd1 t rwh t rwd t rsd t rsd t rdh1 t rdh1 t rds1 t bsd t dakd1 t dakd1 t bsd t bsd t bsd t wts t wth t wts t wth t 1 t w t w t b2 t b1 t b2 t w t 2 t 2 t csd2 t rds1 t rsd t rsd t ah t ah t rdh1 t ah t rsd t rwd t rwh t rdh1 note: in the write cycle, the basic bus cycle is performed. figure 24.20 burst rom bus cycle (two waits) 726 ckio a25 to a4 a3 to a0 csn rd/ wr rd d31 to d0 bs dackn wait t 1 t w t w t b2 t b1 t 2 t bw t ad t ad t csd1 t csd2 t rwd t rwh t rdh1 t ah t ah t rwd t rsd t rsd1 t ah t ad t bsd t bsd t wts t wth t wts t wth t wts t wth t wts t wth t bsd t bsd t rds1 t rdh1 t rsd t dakd1 t dakd1 t rdh1 t rwh t rsd1 t rds note: in the write cycle, the basic bus cycle is performed. figure 24.21 burst rom bus cycle (external wait, waitsel = 1) 727 24.3.6 synchronous dram timing ckio a12 or a10 rd/ wr csn ras cas bs dqmxx cke a25 to a16 a15 to a0 tr tad row address row address read a command row address column address tc1 tc2 (tpc) d31 to d0 tad tad tad tad tcsd3 trwd tcsd3 trwd trasd2 tdqmd tdqmd trdh2 tbsd tbsd (high) trds2 trasd2 tcasd2 tcasd2 tad tad tad dackn t dakd1 t dakd1 figure 24.22 synchronous dram read bus cycle (rcd = 0, cas latency = 1, tpc = 0) 728 ckio a12 or a10 rd/ wr csn ras cas bs dqmxx cke (high) a25 to a16 a15 to a0 tr trw trw tc1 tcw td1 (tpc) (tpc) 9 9 # // # // #/ ) command row address column address tad tad tad tad tcsd3 trwd tdqmd trdh2 tbsd tbsd trds2 tcsd3 trwd trasd2 tdqmd trasd2 tcasd2 tcasd2 tad tad tad tad dackn t dakd1 t dakd1 figure 24.23 synchronous dram read bus cycle (rcd = 2, cas latency = 2, tpc = 1) 729 ckio a12 or a10 rd/ wr csn ras cas bs dqmxx cke (high) a25 to a16 a15 to a0 tr tc1 tc2/td1 tc3/td2 tc4/td3 td4 (tpc) (tpc) 9 9 # // # address read a command read command row address column address (1-4) tad tad tad tad tad tad tad tad tad tcsd3 tcsd3 trwd trwd trasd2 tdqmd tbsd tbsd trds2 trdh2 trds2 trdh2 tdqmd trasd2 tcasd2 tcasd2 t dakd1 t dakd1 dackn figure 24.24 synchronous dram read bus cycle (burst read (single read 4), rcd = 0, cas latency = 1, tpc = 1) 730 %d ) ) #9 wr csn ras cas bs dqmxx cke a25 to a16 a15 to a0 tr trw tc1 tc2 tc3 tc4/td1 td2 td3 td4 (tpc) 9 9 (read) )9 )9 )9 )9 )9 )9 )9 )9 )9 )9 %.9 #f9 919 #9. @.9 @.9 #9< #9. #9< %.9 #f9 #).9 #).9 %).9 919 %).9 # // # address row address read command (high) column address (1-4) t dakd1 t dakd1 dackn figure 24.25 synchronous dram read bus cycle (burst read (single read 4), rcd = 1, cas latency = 3, tpc = 0) 731 ckio a12 or a10 rd/ wr csn ras cas bs dqmxx cke a25 to a16 a15 to a0 tr tc1 (trwl) (tpc) (high) d31 to d0 )9 # // # // f ) command row address column address tad tad tcsd3 trwd trasd2 tad tad tad tad tad tcsd3 trwd trwd trasd2 tcasd2 tdqmd twdd2 tbsd tdqmd twdh2 tbsd tcasd2 t dakd1 t dakd1 dackn figure 24.26 synchronous dram write bus cycle (rcd = 0, tpc = 0, trwl = 0) 732 ckio a12 or a10 rd/ wr csn ras cas bs dqmxx cke a25 to a16 a15 to a0 tr trw trw tc1 (trwl) (trwl) (tpc) (tpc) (high) d31 to d0 )9 # // # address write a command row address column address tad tad tad tad tad tad tcsd3 trwd trwd tad tad tad tcsd3 trwd trasd2 trasd2 tdqmd twdd2 tbsd tcasd2 tdqmd twdh2 tbsd tcasd2 t dakd1 t dakd1 dackn figure 24.27 synchronous dram write bus cycle (rcd = 2, tpc = 1, trwl = 1) 733 ckio a12 or a10 rd/ wr csn ras cas bs dqmxx cke a25 to a16 a15 to a0 tr tc1 tc2 tc3 tc4 (trwl) (tpc) (tpc) (high) d31 to d0 )9 # // # address write a command write command row address tad tad tad tad tcsd3 trwd trwd tad tad tad tad tcsd3 trwd trasd2 trasd2 tdqmd twdd2 twdd2 tbsd tcasd2 tdqmd twdh2 tbsd tcasd2 column address (1-4) t dakd1 t dakd1 dackn figure 24.28 synchronous dram write bus cycle (burst write (single write 4), rcd = 0, tpc = 1, trwl = 0) 734 ckio a12 or a10 rd/ wr csn ras cas bs dqmxx cke (high) a25 to a16 a15 to a0 tr trw tc1 tc2 tc3 td4 (trwl) (tpc) d31 to d0 row address row address write a command write command row address column address (1-4) tad tad tad tad tad tad tad tad tad tcsd3 tcsd3 trwd trwd trwd trasd2 tdqmd tbsd tbsd twdd2 twdd2 twdh2 tdqmd trasd2 tcasd2 tcasd2 9)d9 9)d9 dackn figure 24.29 synchronous dram write bus cycle (burst mode (single write 4), rcd = 1, tpc = 0, trwl = 0) 735 ckio a12 or a10 rd/ wr csn ras cas bs dqmxx cke a25 to a16 a15 to a0 tnop tc1 tc2/td1 tc3/td2 tc4/td3 td4 d31 to d0 t ad t ad t csd3 t csd3 t rwd t rwd t rasd2 t dqmd t dqmd t bsd t bsd (high) t ad t ad t ad t rds2 t rdh2 t rds2 t rdh2 t casd2 t casd2 t ad row address read command column address t dakd1 t dakd1 dackn figure 24.30 synchronous dram burst read bus cycle (ras down, same row address, cas latency = 1) 736 a25 to a16 (high) t ad t ad t ad t casd2 t csd3 t rwd t dqmd t bsd t rdh2 t rds2 t rdh2 t rds2 t bsd t rasd2 t casd2 t dqmd t rwd t csd3 t ad t ad t ad tnop tc1 tc2 tc3/td1 tc4/td2 td3 td4 ckio a12 or a10 a15 to a0 csn rd/ wr ras cas dqmxx d31 to d0 bs cke row address dackn t dakd1 t dakd1 column address read command figure 24.31 synchronous dram burst read bus cycle (ras down, same row address, cas latency = 2) 737 ckio a12 or a10 rd/ wr csn ras cas bs dqmxx cke a25 to a16 a15 to a0 tp tr tc1 tc2/td1 tc3/td2 tc4/td3 td4 d31 to d0 t ad t ad t csd3 t csd3 t rwd t rwd t rwd t rasd2 t rasd2 t dqmd t dqmd t dqmd t bsd t bsd t dakd1 t dakd1 (high) t ad t ad t ad t ad t ad t ad t rds2 t rdh2 t rds2 t rdh2 t ad row address row address t casd2 t casd2 dackn row address read command column address figure 24.32 synchronous dram burst read bus cycle (ras down, different row address, tpc = 0, rcd = 0, cas latency = 1) 738 %d ) ) #9 wr csn ras cas bs dqmxx cke a25 to a16 a15 to a0 tp tpw tr tc1 tc2/td1 tc3/td2 tc4/td3 d31 to d0 t ad t ad t csd3 t csd3 t rwd t rwd t rwd t rasd2 t rasd2 t rasd2 t rasd2 t dqmd t dqmd t dqmd t bsd t bsd (high) t ad t ad t ad t ad t ad t ad t rds2 t rdh2 t rds2 t rdh2 t ad td4 row address t casd2 t casd2 row address row address t dakd1 t dakd1 dackn column address read command figure 24.33 synchronous dram burst read bus cycle (ras down, different row address, tpc = 1, rcd = 0, cas latency = 1) 739 ckio a12 or a10 rd/ wr csn ras cas bs dqmxx cke a25 to a16 a15 to a0 tc1 tc2 tc3 tc4 d31 to d0 t ad t ad t csd3 t csd3 t rwd t rwd t rasd2 t rasd2 t dqmd t dqmd t wdd2 t wdd2 t bsd t bsd (high) t ad t ad t ad t casd2 t casd2 t ad row address write command column address t dakd1 t dakd1 dackn figure 24.34 synchronous dram burst write bus cycle (ras down, same row address) 740 t wdd2 t wdd2 ckio a12 or a10 rd/ wr csn ras cas bs dqmxx cke a25 to a16 a15 to a0 tp tr tc1 tc2 tc3 tc4 9 9 )9 )9 %.9 %.9 #f9 #f9 #f9 #f9 #).9 #).9 919 919 919 @.9 @.9 <8< )9 )9 )9 )9 )9 )9 )9 # // f ,,!/ # address row address column address t casd2 t casd2 t dakd1 t dakd1 dackn figure 24.35 synchronous dram burst write bus cycle (ras down, different row address, tpc = 0, rcd = 0) 741 %d ) ) #9 wr csn ras cas bs dqmxx cke a25 to a16 a15 to a0 tp tpw tr trw tc1 tc2 tc3 d31 to d0 t ad t ad t csd3 t csd3 t rwd t rwd t rwd t rwd t rasd2 t rasd2 t rasd2 t rasd2 t dqmd t dqmd t dqmd t wdd2 t wdd2 t bsd t bsd (high) t ad t ad t ad t ad t ad t ad td4 write command column address t casd2 t casd2 row address row address t ad t ad # // 9)d9 9)d9 dackn figure 24.36 synchronous dram burst write bus cycle (ras down, different row address, tpc = 1, rcd = 1) 742 ckio csn rd/wr casxx cke ras3x tp tpc trr trrw trrw (tpc) (tpc) t csd3 t csd3 t csd3 t csd3 t rasd2 t rasd2 t rasd2 t rasd2 t casd2 t casd2 t rwd t rwd (high) figure 24.37 synchronous dram auto-refresh timing (tras = 1, tpc = 1) 743 trs1 tcked tcked ckio rd/wr csn ras cas cke tcsd3 trasd2 tcasd2 tcsd3 trasd2 tcasd2 trwd (trs2) (trs2) trs3 (tpc) (tpc) tpc tp tcsd3 tcsd3 trasd2 trasd2 trwd trwd figure 24.38 synchronous dram self-refresh cycle (tras = 1, tpc = 1) 744 ckio a12 or a10 rd/ wr csn ras casxx d31 to d0 a13 or a11 a11 to a2 or a9 to a2 trp1 trp2 trp3 trp4 tmw1 tmw2 tmw3 tmw4 (high) cke )9 )9 )9 )9 )9 )9 )9 )9 )9 )9 )9 %.9 %.9 #f9 #f9 #f9 #).9 #).9 #).9 #).9 %).9 %).9 9)d9 9)d9 dackn figure 24.39 synchronous dram mode register write cycle 745 24.3.7 pcmcia timing t pcm1 t pcm2 ckio a25 to a0 cexx rd/ wr rd d15 to d0 we1 d15 to d0 bs dackn t ad t ad t csd1 t csd1 t rwd t rsd t rsd t rwd t dakd1 t dakd1 t wed t wdd1 t wed t rds1 t rdh1 t bsd t bsd t wdh4 t wdh1 (read) (read) (write) (write) figure 24.40 pcmcia memory bus cycle (ted = 0, teh = 0, no wait) 746 ckio t pcm0 t pcm0w t pcm1 t pcm1w t pcm1w t pcm2 t pcm2w a25 to a0 cexx rd/ wr rd (read) d15 to d0 (read) we1 (write) d15 to d0 (write) bs dackn wait t ad t csd1 t rwd t ad t csd1 t rwd t wdh4 t rsd t rsd t dakd1 t dakd1 t wed t wdd1 t wed t wdh1 t rdh1 t bsd t wts t wth t wts t wth t rds1 t bsd figure 24.41 pcmcia memory bus cycle (ted = 2, teh = 1, one wait, external wait, waitsel = 1) 747 ckio t pcm1 t pcm2 t pcm1 t pcm2 t pcm1 t pcm2 t pcm1 t pcm2 a25 to a4 a3 to a0 cexx rd/ wr rd d15 to d0 bs dackn t ad t ad t csd1 t rwd t csd1 t rwd t ad t ad t ad t ad t dakd1 t rsd t rsd t rdh1 t rdh1 t rsd t rsd t bsd t bsd t bsd t bsd t rds1 t rds1 note: even though burst mode is set, the write cycle operation is the same as in normal mode. (read) (read) t dakd1 figure 24.42 pcmcia memory bus cycle (burst read, ted = 0, teh = 0, no wait) 748 ckio t pcm0 t pcm1 t pcm1w t pcm1w t pcm1w t pcm2 t pcm1 t pcm1w t pcm2 t pcm2w a25 to a4 a3 to a0 cexx rd/ wr rd (read) d15 to d0 (read) bs dackn wait t ad t ad t csd1 t rwd t csd1 t dakd1 t rwd t ad t ad t ad t rsd t rsd t rsd t rsd t dakd1 t bsd t bsd t bsd t rds1 t rdh1 t rdh1 t rds1 t wts t wth t wts t wts t wth t wth note: even though burst mode is set, the write cycle operation is the same as in normal mode. t bsd figure 24.43 pcmcia memory bus cycle (burst read, ted = 1, teh = 1, two waits, burst pitch = 3, waitsel = 1) 749 t pci1 t pci2 ckio a25 to a0 cexx rd/ wr iciord (read) d15 to d0 (read) iciowr (write) d15 to d0 (write) bs dackn t ad t ad t csd1 t csd1 t rwd t icrsd t icrsd t rwd t dakd1 t dakd1 t icwsd t wdd1 t icwsd t rdh1 t rds1 t bsd t bsd t wdh1 t wdh4 figure 24.44 pcmcia i/o bus cycle (ted = 0, teh = 0, no wait) 750 ckio t pci0 t pci0w t pci1 t pci1w t pci1w t pci2 t pci2w a25 to a0 cexx rd/ wr iciord (read) d15 to d0 (read) iciowr (write) d15 to d0 (write) bs dackn wait iois16 t ad t csd1 t rwd t ad t csd1 t rwd t icrsd t icrsd t dakd1 t dakd1 t icwsd t wdd1 t icwsd t wdh1 t wdh4 t rdh1 t bsd t bsd t wts t wth t wts t wth t io16s t io16h t rds1 figure 24.45 pcmcia i/o bus cycle (ted = 2, teh = 1, one wait, external wait, waitsel = 1) 751 ckio t pci0 t pci1 t pci1w t pci2 t pci1 t pci1w t pci2 t pci2w a25 to a4 a0 cexx rd/ wr iciord d15 to d0 iciowr d15 to d0 bs wait iois16 dackn t ad t ad t csd1 t csd1 t rwd t rwd t wdd1 t wdh4 t bsd t ad t ad t icrsd t icrsd t icrsd t icrsd t icwsd t wts t wth t wth t io16s t io16h t ad t rds1 t dakd1 t dakd1 t rds1 t icwsd t icwsd t rdh1 t rdh1 t wdh1 t bsd t wts (read) (read) (write) (write) t wdd1 t wdh4 t bsd t bsd t icwsd t csd1 figure 24.46 pcmcia i/o bus cycle (ted = 1, teh = 1, one wait, bus sizing, waitsel = 1) 752 24.3.8 peripheral module signal timing table 24.8 peripheral module signal timing (vccq = 3.3 0.3 v, vcc = 1.55 v to 2.15 v, avcc = 3.3 0.3 v, ta = C20 to 75c) C66 module item symbol min max unit figure tmu, timer input setup time t tclks 15 ns 24.47 rtc timer clock input setup time t tcks 15 24.48 timer clock edge specification t tckwh 1.5 pcyc* 1 pulse width both-edge specification t tckwl 2.5 oscillation settling time t rosc 3 s 24.49 sci input clock asynchronous t scyc 4 pcyc* 1 24.50, cycle synchronous 6 24.51 input clock rise time t sckr 1.5 24.50 input clock fall time t sckf 1.5 input clock pulse width t sckw 0.4 0.6 tscyc transmit data delay time t txd 100 ns 24.51 receive data setup time (synchronous) t rxs 100 receive data hold time (synchronous) t rxh 100 rts delay time t rtsd 100 cts setup time (synchronous) t ctss 100 cts hold time (synchronous) t ctsh 100 port output data delay time t portd 17 ns 24.52 input data setup time t ports1 15 input data hold time t porth1 8 input data setup time t ports2 tcyc + 15 input data hold time t porth2 8 input data setup time t ports3 3 tcyc + 15 input data hold time t porth3 8 dmac dreq setup time t dres 6 ns 24.53 dreq hold time t dreqh 4 drak delay time t drakd 10 24.54 753 t tclks ckio tclk (input) figure 24.47 tclk input timing t tcks t tcks t tckwh t tckwl ckio tclk (input) figure 24.48 tclk clock input timing rtc crystal oscillator stable oscillation v cc v ccmin t rosc figure 24.49 rtc crystal oscillator oscillation settling time at power-on t sckw t sckr t sckf t scyc sck figure 24.50 sck input clock timing 754 t scyc t txd sck txd (data trans- mission) rxd (data reception) t rxh t rxs t rtsd rts cts t ctsh t ctss figure 24.51 sci i/o timing in synchronous mode ckio port 7 to 0 (write) port 7 to 0 (read) (b:p clock ratio = 1:1) port 7 to 0 (read) (b:p clock ratio = 1:2) port 7 to 0 (read) (b:p clock ratio = 1:4) t ports1 t porth1 t ports2 t porth2 t portd t ports3 t porth3 figure 24.52 i/o port timing 755 t pte[1]s t pte[1]h resetp pte[1] dreqn ckio t drqs t drqh figure 24.53 dreq input timing drak0/1 ckio t drakd t drakd figure 24.54 drak output timing 756 24.3.9 h-udi-related pin timing table 24.9 h-udi-related pin timing (vccq = 3.3 0.3 v, vcc = 1.55 v to 2.15 v, avcc = 3.3 0.3 v, ta = C20 to 75c) item symbol min max unit figure tck cycle time t tckcyc 50 ns 24.55 tck high pulse width t tckh 12 ns tck low pulse width t tckl 12 ns tck rise/fall time t tckf 4 ns trst setup time t trsts 12 ns 24.56 trst hold time t trsth 50 t cyc tdi setup time t tdis 10 ns 24.57 tdi hold time t tdih 10 ns tms setup time t tmss 10 ns tms hold time t tmsh 10 ns tdo delay time t tdod 16ns asemd0 setup time t asemdh 12 ns 24.58 asemd0 hold time t asemds 12 ns t tckl t tckf v il v il v ih v ih v il 1/2 vcoq 1/2 vcoq tck (input) t tckf t tckh t tckcyc figure 24.55 tck input timing 757 resetp t trsts t trsth trst figure 24.56 trst input timing (reset hold) tck tdi tms t tdis t tmss t tdih t tckcyc t tmsh t tdod tdo figure 24.57 h-udi data transfer timing t asemd0s t asemd0h resetp asemd0 figure 24.58 asemd0 input timing 758 24.3.10 ac characteristic test conditions ? i/o signal reference level: vccq/2 (vccq = 3.3 0.3 v, vcc = 1.55 v to 2.15 v) ? input pulse level: vss to 3.0 v (where resetp , resetm , asemd0, irl3 to irl0, irls3 to irls0, adtrg, pint15 to pint0, trst, rxd1, ca, nmi, irq5 C irq0 , ckio, and md5Cmd0 are within vss to vcc) ? input rise and fall times: 1 ns i ol i oh c l v ref sh7729r output pin dut output notes: 1. c l is the total value that includes the capacitance of measurement instruments, etc., and is set as follows for each pin. 30pf: ckio, ras , cas , cs0 , cs2 C cs6 , ce2a , ce2b , back 50pf: all other pins 2. i ol and i oh are the values shown in table 24.3. figure 24.59 output load circuit 759 24.3.11 delay time variation due to load capacitance a graph (reference data) of the variation in delay time when a load capacitance greater than that stipulated (30 pf) is connected to the sh7729rs pins is shown below. the graph shown in figure 24.60 should be taken into consideration in the design process if the stipulated capacitance is exceeded in connecting an external device. if the connected load capacitance exceeds the range shown in figure 24.60, the graph will not be a straight line. +3 +2 +1 +0 +0 +10 +20 +30 +40 +50 load capacitance [pf] delay time [ns] figure 24.60 load capacitance vs. delay time 760 24.4 a/d converter characteristics table 24.10 lists the a/d converter characteristics. table 24.10 a/d converter characteristics (vccq = 3.3 0.3 v, vcc = 1.55 v to 2.15 v, avcc = 3.3 0.3 v, ta = C20 to 75c) item min typ max unit resolution 10 10 10 bits conversion time 16 s analog input capacitance 20 pf permissible signal-source (single- source) impedance 5 k ? nonlinearity error 3 lsb offset error 2 lsb full-scale error 2 lsb quantization error 0.5 lsb absolute accuracy 4 lsb 24.5 d/a converter characteristics table 24.11 lists the d/a converter characteristics. table 24.11 d/a converter characteristics (vccq = 3.3 0.3 v, vcc = 1.55 v to 2.15 v, avcc = 3.3 0.3 v, ta = C20 to 75c) item min typ max unit test conditions resolution 888bits conversion time 10.0 s 20-pf capacitive load absolute accuracy 2.5 4.0 lsb 2-m ? resistance load 761 appendix a pin functions a.1 pin states table a.1 shows pin states during resets, power-down states, and the bus-released state. table a.1 pin states during resets, power-down states, and bus-released state reset power-down category pin power-on reset manual reset standby sleep bus released clock extal i iiii xtal o* 1 o* 1 o* 1 o* 1 o* 1 ckio io* 1 io* 1 io* 1 io* 1 io* 1 extal2 i iiii xtal2 o oooo cap1, cap2 system control resetp i iiii resetm i iiii breq i iiii back o oool md[5:0] i iiii ca i iiii status[1:0]/ptj[7:6] o op* 3 op* 3 op* 3 op* 3 interrupt irq[3:0]/ irl[3:0] / pth[3:0] v* 8 iiii irq4/ pth[4] v* 8 iiii nmi i iiii irls[3:0] /ptf[3:0]/ pint[11:8] viizii mcs [7:0]/ptc[7:0]/ pint[7:0] v op* 3 zh* 11 k* 3 op* 3 zp* 3 tck/ptf4/pint12 iv i iz i i tdi/ptfs/pint13 iv i iz i i tms/ptf6/pint14 iv i iz i i trst /ptf7/pint15 iv i iz i i irqout o oooo 762 table a.1 pin states during resets, power-down states, and bus-released state (cont) reset power-down category pin power-on reset manual reset standby sleep bus released address bus a[25:0] z o zl* 10 oz data bus d[15:0] z i z io z d[23:16]/pta[7:0] z ip* 3 zk* 3 iop* 3 zp* 3 d[31:24]/ptb[7;0] z ip* 3 zk* 3 iop* 3 zp* 3 bus control cs0 / mcs0 h o zh* 11 oz cs[2:4] /ptk[0:2] h op* 3 zh* 11 k* 3 op* 3 zp* 3 cs5 / ce1a /ptk[3] h op* 3 zh* 11 k* 3 op* 3 zp* 3 cs6 / ce1b h o zh* 11 oz bs /ptk[4] h op* 3 zh* 11 k* 3 op* 3 zp* 3 ras3l /ptj[0] h op* 3 zok* 4 op* 3 zop* 4 ras3u /pte[2] v op* 3 zok* 4 op* 3 zop* 4 casl /ptj[2] h op* 3 zok* 4 op* 3 zop* 4 casu /ptj[3] h op* 3 zok* 4 op* 3 zop* 4 we0 /dqmll h o zh* 11 oz we1 /dqmlu/ we h o zh* 11 oz we2 /dqmul/ iciord / ptk[6] h op* 3 zh* 11 k* 3 op* 3 zp* 3 we3 /dqmuu/ iciowr / ptk[7] h op* 3 zh* 11 k* 3 op* 3 zp* 3 rd/ wr h o zh* 11 oz rd h o zh* 11 oz cke/ptk[5] h op* 3 ok* 3 op* 3 op* 3 wait ziziz dmac dreq0 /ptd[4] v zi* 7 zi i dack0 /ptd[5] v op* 3 zk* 3 op* 3 op* 3 drak0/ptd[1] v op* 3 zh* 11 k* 3 op* 3 op* 3 dreq1 /ptd[6] v zi* 7 zi i dack1 /ptd[7] v op* 3 zk* 3 op* 3 op* 3 drak1/ptd[0] v op* 3 zh* 11 k* 3 op* 3 op* 3 timer tclk/pth[7] v zp iop* 5 iop* 5 iop* 5 763 table a.1 pin states during resets, power-down states, and bus-released state (cont) reset power-down category pin power-on reset manual reset standby sleep bus released sci/ smart card without fifo rxd0/scpt[0] txd0/scpt[0] sck0/scpt[1] z z v zi* 7 zo* 7 zp* 3 z zk* 3 zk* 3 iz* 6 oz* 6 iop* 5 iz* 6 oz* 6 iop* 5 scif/irda rxd1/scpt[2] z zi* 7 z iz* 6 iz* 6 with fifo txd1/scpt[2] z zo* 7 zk* 3 oz* 6 oz* 6 sck1/scpt[3] v zp* 3 zk* 3 iop* 5 iop* 5 scif with fifo rxd2/scpt[4] z zi* 7 z iz* 6 iz* 6 txd2/scpt[4] z zo* 7 zk* 3 oz* 6 oz* 6 sck2/scpt[5] v zp* 3 zk* 3 iop* 5 iop* 5 rts2 /scpt[6] v op* 3 zk* 3 op* 3 op* 3 cts2 /irq5/scpt[7] v* 8 zi* 7 iii port audsync /pte[7] ov op* 3 ok* 3 op* 3 op* 3 ce2b /pte[5] v op* 3 zh* 11 k* 3 op* 3 zp* 3 ce2a /pte[4] v op* 3 zh* 11 k* 3 op* 3 zp* 3 tdo/pte[0] ov op* 3 ok* 3 op* 3 op* 3 iois16 /ptg[7] v i z i i ptg[6:0] v i z i i audck/pht[6] v i z i i adtrg /pth[5] v* 8 iizii wakeup /ptd[3] v op* 3 ok* 3 op* 3 zp* 3 resetout /ptd[2] o op* 3 zk* 3 op* 3 op* 3 audata[3:0]/ptg[3:0] iv port i iz port i i ckio2/ptg[4] ov port oi port oz port oi port oi port asebrkak /ptg[5] ov port oi port oz port oi port oi port asemd0 /ptg[6] i (asemd) i z i i ptj[1] h op* 3 zok* 4 op* 3 zop* 4 pte[1] v* 13 op* 3 zok* 4 op* 3 zop* 4 pte[6] v op* 3 zok* 4 op* 3 zop* 4 pte[3] v op* 3 zok* 4 op* 3 zop* 4 ptj[4] h op* 3 zok* 4 op* 3 zop* 4 ptj[5] h op* 3 zok* 4 op* 3 zop* 4 764 table a.1 pin states during resets, power-down states, and bus-released state (cont) reset power-down category pin power-on reset manual reset standby sleep bus released analog an[5:0]/ptl[5:0] z zi* 7 zi i an[6:7]/da[1:0]/ ptl[6:7] z zi* 7 i io* 9 io* 9 i: input o: output h: high-level output l: low-level output z: high impedance p: input or output depending on register setting k: input pin is high impedance, output pin holds its state v: i/o buffer off, pull-up mos on notes: *1 depending on the clock mode (md2Cmd0 setting). *2 z when the port function is used. *3 k or p when the port function is used. *4 k or p when the port function is used. z or o when the port function is not used depending on register setting. *5 k or p when the port function is used. i or o when the port function is not used depending on register setting. *6 depending on register setting. *7 i or o when the port function is used. *8 input schmitt buffers and pull-up mos of irq[5:0] and adtrg are on; other inputs are off. *9 i when the port function is used. i or o when the port function is not used, depending on register setting. *10 in standby mode, z or l depending on register setting. *11 in standby mode, z or h depending on register setting. *12 o when da output is enabled; z otherwise. *13 in a power-on reset, leave open or input a high level. 765 a.2 pin specifications table a.2 shows the pin specifications. table a.2 pin specifications pin pin no. i/o function md5 197 i operating mode pin (endian mode) md4, md3 196, 195 i operating mode pin (area 0 bus width) md2 to md0 2, 1, 144 i operating mode pin (clock mode) ras3l /ptj[0] 106 i/o ras (sdram) / i/o port ptj[1] 107 i/o i/o port ce2a /pte[4] 103 i/o pcmcia ce2a / i/o port ce2b /pte[5] 104 i/o pcmcia ce2b / i/o port rxd0/scpt[0] 171 i serial port 0 data input / input port rxd1/scpt[2] 172 i serial port 1 data input / input port rxd2/scpt[4] 174 i serial port 2 data input / input port txd0/scpt[0] 164 o serial port 0 data output / output port txd1/scpt[2] 166 o serial port 1 data output / output port txd2/scpt[4] 168 o serial port 2 data output / output port sck0/scpt[1] 165 i/o serial port 0 clock input/output / i/o port sck1/scpt[3] 167 i/o serial port 1 clock input/output / i/o port sck2/scpt[5] 169 i/o serial port 2 clock input/output / i/o port rts2 /scpt[6] 170 i/o serial port 2 transfer request / i/o port status1/ptj[7] 158 i/o processor state / i/o port status0/ptj[6] 157 i/o processor state / i/o port a25 to a0 86, 84, 82, 80, 78 to 72, 70, 68 to 60, 58, 56 to 53 o address bus d31 to d24/ ptb[7] to ptb[0] 13 to 18, 20, 22 i/o data bus / i/o port d23 to d16/ pta[7] to pta[0] 23 to 26, 28, 30 to 32 i/o data bus / i/o port d15 to d0 34, 36 to 44, 46, 48 to 52 i/o data bus mcs[7:0] /ptc[7:0]/ pint[7:0] 177 to 180,185 to 188 i/o mask rom chip select 7-0 / i/o port / port interrupt request 766 table a.2 pin specifications (cont) pin pin no. i/o function wakeup /ptd[3] 182 i/o wakeup / i/o port resetout / ptd[2] 184 i/o reset output / i/o port drak0/ptd[1] 189 i/o dma control pin / i/o port drak1/ptd[0] 190 i/o dma control pin / i/o port dreq0 /ptd[4] 191 i dma transfer request 0 / input port dreq1 /ptd[6] 192 i dma transfer request 1 / input port an[5:0]/ptl[5:0] 204 to 199 i analog input pin / input port an[7:6]/da[1:0]/ ptl[7:6] 207, 206 i/o analog i/o pin / input port cs6 / ce1b 102 o chip select 6 / pcmcia ce1b cs5 / ce1a / ptk[3] 101 i/o chip select 5 / pcmcia ce2b / i/o port cs4 /ptk[2] 100 i/o chip select 4 / i/o port cs3 /ptk[1] 99 i/o chip select 3 / i/o port cs2 /ptk[0] 98 i/o chip select 2 / i/o port cs0 / mcs0 96 o chip select 0 / mask rom chip select 0 bs /ptk[4] 87 i/o bus cycle start / i/o port ptj[5] 113 i/o i/o port ptj[4] 112 i/o i/o port casu /ptj[3] 110 i/o i/o port casl /ptj[2] 108 i/o cas(sdram) / i/o port dack0 /ptd[5] 114 i/o dma transfer strobe 0 / i/o port dack1 /ptd[7] 115 i/o dma transfer strobe 1 / i/o port rd 88 o read strobe pin we0 / dqmll 89 o d7Cd0 select signal/ dqm(sdram) we1 /dqmlu/ we 90 o d15Cd8 select signal / dqm(sdram)/ pcmcia we signal we2 /dqmul/ iciord /ptk[6] 91 i/o d23Cd16 select signal / dqm(sdram) / pcmcia iord signal / i/o port we3 /dqmuu/ iciowr /ptk[7] 92 i/o d31Cd24 select signal /dqm(sdram) / pcmcia iowr signal / i/o port rd/ wr 93 o read/write select signal 767 table a.2 pin specifications (cont) pin pin no. i/o function audsync/ pte[7] 94 i/o aud synchronous i/o port pte[6] 116 i/o i/o port pte[3] 117 i/o i/o port ras3u /pte[2] 118 i/o ras(sdram) / i/o port pte[1] 119 i/o i/o port *2 tdo/pte[0] 120 i/o test data output i/o port resetm 124 i manual reset input adtrg /pth[5] 125 i adc trigger request / input port iois16 /ptg[7] 126 i i/o for pc card / input port asmd0 /ptg[6] 127 i ase mode / input port asebrkak / ptg[5] 128 i ase break accept / input port ckio2/ptg[4] 129 o/i system clock output / input port audata[3]/ ptg[3] 130 i aud data / input port audata[2]/ ptg[2] 131 i aud data / input port audata[1]/ ptg[1] 133 i aud data / input port audata[0]/ ptg[0] 135 i aud data / input port trst /ptf[7]/ pint[15] 136 i test reset / input port / port interrupt request trst/ptf[6]/ pint[14] 137 i test mode switch / input port / port interrupt request trst/ptf[5]/ pint[13] 138 i test data input / input port / port interrupt request trst/ptf[4]/ pint[12] 139 i test clock / input port / port interrupt request irls [3:0]/ ptf[3:0]/ pint[11:8] 140 to 143 i external interrupt request / input port / port interrupt request audck/pth[6] 151 i aud clock / input port wait 123 i hardware wait request 768 table a.2 pin specifications (cont) pin pin no. i/o function breq 122 i bus request back 121 o bus acknowledge irqout 160 o interrupt / refresh request output resetp 193 i power-on reset input nmi 7 i nonmaskable interrupt request irq[3:0]/ irl[3:0] / pth[3:0] 11 to 8 i external interrupt request / external interrupt source / input port irq4/ pth[4] 12 i external interrupt request / input port cts2 /irq5/ scpt[7] 176 i serial port 2 transfer enable / external interrupt request / input port tclk/pth[7] 159 i/o clock i/o (for tmu/rtc) / i/o port extal 156 i external clock / crystal oscillator pin xtal 155 o crystal oscillator pin cap1 146 external capacitance pin (for pll1) cap2 149 external capacitance pin (for pll2) ckio 162 i/o system clock i/o xtal2 4 o crystal oscillator pin (for on-chip rtc) extal2 5 i crystal oscillator pin (for on-chip rtc) cke/ptk[5] 105 i/o ck enable for sdram / i/o port ca 194 i setting hardware standby v cc q 21, 35, 47, 59, 71, 85, 97, 111, 163, 183 power supply power supply (3.3 v), i/o buffer v cc Crtc 3 power supply rtc oscillator power supply (1.9/1.8 v) v cc Cpll1 v cc Cpll2 145 150 power supply pll power supply (1.9/1.8 v) av cc 205 power supply analog power supply (3.3 v) v ss q 19, 33, 45, 57, 69, 83, 95, 109, 161, 181 power supply power supply (0 v), i/o buffer v ss 27, 79, 132, 152, 153, 173 power supply internal power supply (0 v) v cc 29, 81, 134, 154, 175 power supply internal power supply (1.9/1.8 v) 769 table a.2 pin specifications (cont) pin pin no. i/o function v ss Crtc 6 power supply rtc-oscillator power supply (0 v) v ss Cpll1 v ss Cpll2 147 148 power supply pll power supply (0 v) av ss 198, 208 power supply analog power supply (0 v) notes: *1 except in hardware standby mode, power must be supplied constantly to all power supply pins. in hardware standby mode, power must be supplied to vcc-rtc and vss- rtc at least. *2 in a power-on reset, leave open or input a high level. a.3 treatment of unused pins ? when rtc is not used ? extal2: pull up to v cc (1.9/1.8 v) ? xtal2: leave unconnected ? v cc Crtc: power supply (1.9/1.8 v) ? v ss Crtc: power supply (0 v) ? when pll1 is not used ? cap1: leave unconnected ? v cc Cpll1: power supply (1.9/1.8 v) ? v ss Cpll1: power supply (0 v) ? when pll2 is not used ? cap2: leave unconnected ? v cc Cpll2: power supply (1.9/1.8 v) ? v ss Cpll2: power supply (0 v) ? when on-chip crystal oscillator is not used ? xtal: leave unconnected ? when extal pin is not used ? extal: pull up to v cc q (3.3 v) ? when a/d converter is not used ? an[7:0]: leave unconnected ? av cc : power supply (3.3 v) ? av ss : power supply (0 v) 770 a.4 pin states in access to each address space table a.3 pin states (ordinary memory/little endian) 8-bit bus width 16-bit bus width pin byte/word/long- word access byte access (address 2n) byte access (address 2n + 1) word/longword access cs6 to cs2 , cs0 enabled enabled enabled enabled rd r low low low low w high high high high rd/ wr r high high high high w low low low low bs enabled enabled enabled enabled ras3u /pte[2] high high high high ras3l /ptj[0] high high high high casl /ptj[2] high high high high casu /ptj[3] high high high high we0 /dqmll r high high high high w low low high low we1 /dqmlu/ we r high high high high w high high low low we2 /dqmul/ iciord / r high high high high ptk[6] w high high high high we3 /dqmuu/ iciowr / r high high high high ptk[7] w high high high high ce2a /pte[4] high high high high ce2b /pte[5] high high high high cke/ptk[5] disabled disabled disabled disabled wait enabled *1 enabled *1 enabled *1 enabled *1 iois16 /ptg[7] disabled disabled disabled disabled a25 to a0 address address address address d7 to d0 valid data valid data invalid data valid data d15 to d8 high-z *2 invalid data valid data valid data d31 to d16 high-z *2 high-z *2 high-z *2 high-z *2 771 table a.3 pin states (ordinary memory/little endian) (cont) 32-bit bus width pin byte access (address 4n) byte access (address 4n + 1) byte access (address 4n + 2) byte access (address 4n + 3) word access (address 4n) word access (address 4n + 2) longword access cs6 to cs2 , cs0 enabled enabled enabled enabled enabled enabled enabled rd r low low low low low low low w high high high high high high high rd/ wr r high high high high high high high w low low low low low low low bs enabled enabled enabled enabled enabled enabled enabled ras3u /pte[2] high high high high high high high ras3l /ptj[0] high high high high high high high casl /ptj[2] high high high high high high high casu /ptj[3] high high high high high high high we0 /dqmll r high high high high high high high w low high high high low high low we1 /dqmlu/ we r high high high high high high high w high low high high low high low we2 /dqmul/ iciord / r high high high high high high high ptk[6] w high high low high high low low we3 /dqmuu/ iciowr / r high high high high high high high ptk[7] w high high high low high low low ce2a /pte[4] high high high high high high high ce2b /pte[5] high high high high high high high cke/ptk[5] disabled disabled disabled disabled disabled disabled disabled wait enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 iois16 /ptg[7] disabled disabled disabled disabled disabled disabled disabled a25 to a0 address address address address address address address d7 to d0 valid data invalid data invalid data invalid data valid data invalid data valid data d15 to d8 invalid data valid data invalid data invalid data valid data invalid data valid data d23 to d16 invalid data invalid data valid data invalid data invalid data valid data valid data d31 to d24 invalid data invalid data invalid data valid data invalid data valid data valid data notes: *1 disabled when wcr2 register wait setting is 0. *2 unused data pins should be switched to the port function, or pulled up or down. 772 table a.4 pin states (ordinary memory/big endian) 8-bit bus width 16-bit bus width pin byte/word/long- word access byte access (address 2n) byte access (address 2n + 1) word/longword access cs6 to cs2 , cs0 enabled enabled enabled enabled rd r low low low low w high high high high rd/ wr r high high high high w low low low low bs enabled enabled enabled enabled ras3u /pte[2] high high high high ras3l /ptj[0] high high high high casl /ptj[2] high high high high casu /ptj[3] high high high high we0 /dqmll r high high high high w low high low low we1 /dqmlu/ we r high high high high w high low high low we2 /dqmul/ iciord / r high high high high ptk[6] w high high high high we3 /dqmuu/ iciowr / r high high high high ptk[7] w high high high high ce2a /pte[4] high high high high ce2b /pte[5] high high high high cke/ptk[5] disabled disabled disabled disabled wait enabled *1 enabled *1 enabled *1 enabled *1 iois16 /ptg[7] disabled disabled disabled disabled a25 to a0 address address address address d7 to d0 valid data invalid data valid data valid data d15 to d8 high-z *2 valid data invalid data valid data d31 to d16 high-z *2 high-z *2 high-z *2 high-z *2 773 table a.4 pin states (ordinary memory/big endian) (cont) 32-bit bus width pin byte access (address 4n) byte access (address 4n + 1) byte access (address 4n + 2) byte access (address 4n + 3) word access (address 4n) word access (address 4n + 2) longword access cs6 to cs2 , cs0 enabled enabled enabled enabled enabled enabled enabled rd r low low low low low low low w high high high high high high high rd/ wr r high high high high high high high w low low low low low low low bs enabled enabled enabled enabled enabled enabled enabled ras3u /pte[2] high high high high high high high ras3l /ptj[0] high high high high high high high casl /ptj[2] high high high high high high high casu /ptj[3] high high high high high high high we0 /dqmll r high high high high high high high w high high high low high low low we1 /dqmlu/ we r high high high high high high high w high high low high high low low we2 /dqmul/ iciord / r high high high high high high high ptk[6] w high low high high low high low we3 /dqmuu/ iciowr / r high high high high high high high ptk[7] w low high high high low high low ce2a /pte[4] high high high high high high high ce2b /pte[5] high high high high high high high cke/ptk[5] disabled disabled disabled disabled disabled disabled disabled wait enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 iois16 /ptg[7] disabled disabled disabled disabled disabled disabled disabled a25 to a0 address address address address address address address d7 to d0 invalid data invalid data invalid data valid data invalid data valid data valid data d15 to d8 invalid data invalid data valid data invalid data invalid data valid data valid data d23 to d16 invalid data valid data invalid data invalid data valid data invalid data valid data d31 to d24 valid data invalid data invalid data invalid data valid data invalid data valid data notes: *1 disabled when wcr2 register wait setting is 0. *2 unused data pins should be switched to the port function, or pulled up or down. 774 table a.5 pin states (burst rom/little endian) 8-bit bus width 16-bit bus width pin byte/word/long- word access byte access (address 2n) byte access (address 2n + 1) word/longword access cs6 to cs2 , cs0 enabled enabled enabled enabled rd r low low low low w rd/ wr r high high high high w bs enabled enabled enabled enabled ras3u /pte[2] high high high high ras3l /ptj[0] high high high high casl /ptj[2] high high high high casu /ptj[3] high high high high we0 /dqmll r high high high high w we1 /dqmlu/ we r high high high high w we2 /dqmul/ iciord / r high high high high ptk[6] w we3 /dqmuu/ iciowr / r high high high high ptk[7] w ce2a /pte[4] high high high high ce2b /pte[5] high high high high cke/ptk[5] disabled disabled disabled disabled wait enabled *1 enabled *1 enabled *1 enabled *1 iois16 /ptg[7] disabled disabled disabled disabled a25 to a0 address address address address d7 to d0 valid data valid data invalid data valid data d15 to d8 high-z *2 invalid data valid data valid data d31 to d16 high-z *2 high-z *2 high-z *2 high-z *2 775 table a.5 pin states (burst rom/little endian) (cont) 32-bit bus width pin byte access (address 4n) byte access (address 4n + 1) byte access (address 4n + 2) byte access (address 4n + 3) word access (address 4n) word access (address 4n + 2) longword access cs6 to cs2 , cs0 enabled enabled enabled enabled enabled enabled enabled rd r low low low low low low low w rd/ wr r high high high high high high high w bs enabled enabled enabled enabled enabled enabled enabled ras3u /pte[2] high high high high high high high ras3l /ptj[0] high high high high high high high casl /ptj[2] high high high high high high high casu /ptj[3] high high high high high high high we0 /dqmll r high high high high high high high w we1 /dqmlu/ we r high high high high high high high w we2 /dqmul/ iciord / r high high high high high high high ptk[6] w we3 /dqmuu/ iciowr / r high high high high high high high ptk[7] w ce2a /pte[4] high high high high high high high ce2b /pte[5] high high high high high high high cke/ptk[5] disabled disabled disabled disabled disabled disabled disabled wait enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 iois16 /ptg[7] disabled disabled disabled disabled disabled disabled disabled a25 to a0 address address address address address address address d7 to d0 valid data invalid data invalid data invalid data valid data invalid data valid data d15 to d8 invalid data valid data invalid data invalid data valid data invalid data valid data d23 to d16 invalid data invalid data valid data invalid data invalid data valid data valid data d31 to d24 invalid data invalid data invalid data valid data invalid data valid data valid data notes: *1 disabled when wcr2 register wait setting is 0. *2 unused data pins should be switched to the port function, or pulled up or down. 776 table a.6 pin states (burst rom/big endian) 8-bit bus width 16-bit bus width pin byte/word/long- word access byte access (address 2n) byte access (address 2n + 1) word/longword access cs6 to cs2 , cs0 enabled enabled enabled enabled rd r low low low low w rd/ wr r high high high high w bs enabled enabled enabled enabled ras3u /pte[2] high high high high ras3l /ptj[0] high high high high casl /ptj[2] high high high high casu /ptj[3] high high high high we0 /dqmll r high high high high w we1 /dqmlu/ we r high high high high w we2 /dqmul/ iciord / r high high high high ptk[6] w we3 /d qm u u / iciowr / r high high high high ptk[7] w ce2a /pte[4] high high high high ce2b /pte[5] high high high high cke/ptk[5] disabled disabled disabled disabled wait enabled *1 enabled *1 enabled *1 enabled *1 iois16 /ptg[7] disabled disabled disabled disabled a25 to a0 address address address address d7 to d0 valid data invalid data valid data valid data d15 to d8 high-z *2 valid data invalid data valid data d31 to d16 high-z *2 high-z *2 high-z *2 high-z *2 777 table a.6 pin states (burst rom/big endian) (cont) 32-bit bus width pin byte access (address 4n) byte access (address 4n + 1) byte access (address 4n + 2) byte access (address 4n + 3) word access (address 4n) word access (address 4n + 2) longword access cs6 to cs2 , cs0 enabled enabled enabled enabled enabled enabled enabled rd r low low low low low low low w rd/ wr r high high high high high high high w bs enabled enabled enabled enabled enabled enabled enabled ras3u /pte[2] high high high high high high high ras3l /ptj[0] high high high high high high high casl /ptj[2] high high high high high high high casu /ptj[3] high high high high high high high we0 /dqmll r high high high high high high high w we1 /dqmlu/ we r high high high high high high high w we2 /dqmul/ iciord / r high high high high high high high ptk[6] w we3 /d qm u u / iciowr / r high high high high high high high ptk[7] w ce2a /pte[4] high high high high high high high ce2b /pte[5] high high high high high high high cke/ptk[5] disabled disabled disabled disabled disabled disabled disabled wait enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 iois16 /ptg[7] disabled disabled disabled disabled disabled disabled disabled a25 to a0 address address address address address address address d7 to d0 invalid data invalid data invalid data valid data invalid data valid data valid data d15 to d8 invalid data invalid data valid data invalid data invalid data valid data valid data d23 to d16 invalid data valid data invalid data invalid data valid data invalid data valid data d31 to d24 valid data invalid data invalid data invalid data valid data invalid data valid data notes: *1 disabled when wcr2 register wait setting is 0. *2 unused data pins should be switched to the port function, or pulled up or down. 778 table a.7 pin states (synchronous dram/little endian) 32-bit bus width pin byte access (address 4n) byte access (address 4n + 1) byte access (address 4n + 2) byte access (address 4n + 3) word access (address 4n) word access (address 4n + 2) longword access cs6 to cs2 , cs0 enabled enabled enabled enabled enabled enabled enabled rd r high high high high high high high w high high high high high high high rd/ wr r high high high high high high high w low low low low low low low bs enabled enabled enabled enabled enabled enabled enabled ras3u /pte[2] high/low *1 high/low *1 high/low *1 high/low *1 high/low *1 high/low *1 high/low *1 ras3l /ptj[0] low/high *1 low/high *1 low/high *1 low/high *1 low/high *1 low/high *1 low/high *1 casl /ptj[2] low low low low low low low casu /ptj[3] high high high high high high high we0 /dqmll r low high high high low high low w low high high high low high low we1 /dqmlu/ we r high low high high low high low w high low high high low high low we2 /dqmul/ iciord / r high high low high high low low ptk[6] w high high low high high low low we3 /dqmuu/ iciowr / r high high high low high low low ptk[7] w high high high low high low low ce2a /pte[4] high high high high high high high ce2b /pte[5] high high high high high high high cke/ptk[5] high *2 high *2 high *2 high *2 high *2 high *2 high *2 wait disabled disabled disabled disabled disabled disabled disabled iois16 /ptg[7] disabled disabled disabled disabled disabled disabled disabled a25 to a0 address command address command address command address command address command address command address command d7 to d0 valid data invalid data invalid data invalid data valid data invalid data valid data d15 to d8 invalid data valid data invalid data invalid data valid data invalid data valid data d23 to d16 invalid data invalid data valid data invalid data invalid data valid data valid data d31 to d24 invalid data invalid data invalid data valid data invalid data valid data valid data notes: *1 lower 32-mb access/ upper 32-mb access *2 normally high. low in self-refreshing. 779 table a.8 pin states (synchronous dram/big endian) 32-bit bus width pin byte access (address 4n) byte access (address 4n + 1) byte access (address 4n + 2) byte access (address 4n + 3) word access (address 4n) word access (address 4n + 2) longword access cs6 to cs2 , cs0 enabled enabled enabled enabled enabled enabled enabled rd r high high high high high high high w high high high high high high high rd/ wr r high high high high high high high w low low low low low low low bs enabled enabled enabled enabled enabled enabled enabled ras3u /pte[2] high/low *1 high/low *1 high/low *1 high/low *1 high/low *1 high/low *1 high/low *1 ras3l /ptj[0] low/high *1 low/high *1 low/high *1 low/high *1 low/high *1 low/high *1 low/high *1 casl /ptj[2] low low low low low low low casu /ptj[3] high high high high high high high we0 /dqmll r high high high low high low low w high high high low high low low we1 /dqmlu/ we r high high low high high low low w high high low high high low low we2 /dqmul/ iciord / r high low high high low high low ptk[6] w high low high high low high low we3 /dqmuu/ iciowr / r low high high high low high low ptk[7] w low high high high low high low ce2a /pte[4] high high high high high high high ce2b /pte[5] high high high high high high high cke/ptk[5] high *2 high *2 high *2 high *2 high *2 high *2 high *2 wait disabled disabled disabled disabled disabled disabled disabled iois16 /ptg[7] disabled disabled disabled disabled disabled disabled disabled a25 to a0 address command address command address command address command address command address command address command d7 to d0 valid data invalid data invalid data invalid data valid data invalid data valid data d15 to d8 invalid data valid data invalid data invalid data valid data invalid data valid data d23 to d16 invalid data invalid data valid data invalid data invalid data valid data valid data d31 to d24 invalid data invalid data invalid data valid data invalid data valid data valid data notes: *1 lower 32-mb access/ upper 32-mb access *2 normally high. low in self-refreshing. 780 table a.9 pin states (pcmcia/little endian) pcmcia memory interface (area 5) 8-bit bus width 16-bit bus width pin byte/word/long- word access byte access (address 2n) byte access (address 2n + 1) word/longword access cs6 to cs2 , cs0 enabled enabled high enabled rd r low low low low w high high high high rd/ wr r high high high high w low low low low bs enabled enabled enabled enabled ras3u /pte[2] high high high high ras3l /ptj[0] high high high high casl /ptj[2] high high high high casu /ptj[3] high high high high we0 /dqmll r high high high high w high high high high we1 /dqmlu/ we r high high high high w low low low low we2 /dqmul/ iciord / r high high high high ptk[6] w high high high high we3 /dqmuu/ iciowr / r high high high high ptk[7] w high high high high ce2a /pte[4] high high low low ce2b /pte[5] high high high high cke/ptk[5] disabled disabled disabled disabled wait enabled *1 enabled *1 enabled *1 enabled *1 iois16 /ptg[7] disabled disabled disabled disabled a25 to a0 address address address address d7 to d0 valid data valid data invalid data valid data d15 to d8 high-z *2 invalid data valid data valid data d31 to d16 high-z *2 high-z *2 high-z *2 high-z *2 781 table a.9 pin states (pcmcia/little endian) (cont) pcmcia memory interface (area 6) pcmcia/io interface (area 6) 8-bit bus width 16-bit bus width 8-bit bus width 16-bit bus width pin byte/ word/ long- word access byte access (ad- dress 2n) byte access (ad- dress 2n + 1) word/ long- word access byte/ word/ long- word access byte access (ad- dress 2n) byte access (ad- dress 2n+1) word/ long- word access cs6 to cs2 , cs0 enabled enabled high enabled enabled enabled high enabled rd r low low low low high high high high whigh high high high high high high high rd/ wr r high high high high high high high high wlow low low low low low low low bs enabled enabled enabled enabled enabled enabled enabled enabled ras3u /pte[2] high high high high high high high high ras3l /ptj[0] high high high high high high high high casl /ptj[2] high high high high high high high high casu /ptj[3] high high high high high high high high we0 /dqmll r high high high high high high high high whigh high high high high high high high we1 /dqmlu/ we r high high high high high high high high wlow low low low high high high high we2 /dqmul/ iciord / r high high high high low low low low ptk[6] whigh high high high high high high high we3 /dqmuu/ iciowr / r high high high high high high high high ptk[7] whigh high high high low low low low ce2a /pte[4] high high high high high high high high ce2b /pte[5] high high low low high high low low cke/ptk[5] disabled disabled disabled disabled disabled disabled disabled disabled wait enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 iois16 /ptg[7] disabled disabled disabled disabled disabled disabled enabled enabled a25 to a0 address address address address address address address address d7 to d0 valid data valid data invalid data valid data valid data valid data invalid data valid data d15 to d8 high-z* 2 invalid data valid data valid data high-z* 2 invalid data valid data valid data d31 to d16 high-z* 2 high-z* 2 high-z* 2 high-z* 2 high-z* 2 high-z* 2 high-z* 2 high-z* 2 notes: *1 disabled when wcr2 register wait setting is 0. *2 unused data pins should be switched to the port function, or pulled up or down. 782 table a.10 pin states (pcmcia/big endian) pcmcia memory interface (area 5) 8-bit bus width 16-bit bus width pin byte/word/long- word access byte access (address 2n) byte access (address 2n + 1) word/longword access cs6 to cs2 , cs0 enabled enabled high enabled rd r low low low low w high high high high rd/ wr r high high high high w low low low low bs enabled enabled enabled enabled ras3u /pte[2] high high high high ras3l /ptj[0] high high high high casl /ptj[2] high high high high casu /ptj[3] high high high high we0 /dqmll r high high high high w high high high high we1 /dqmlu/ we r high high high high w low low low low we2 /dqmul/ iciord / r high high high high ptk[6] w high high high high we3 /dqmuu/ iciowr / r high high high high ptk[7] w high high low low ce2a /pte[4] high high low low ce2b /pte[5] high high high high cke/ptk[5] disabled disabled disabled disabled wait enabled *1 enabled *1 enabled *1 enabled *1 iois16 /ptg[7] disabled disabled disabled disabled a25 to a0 address address address address d7 to d0 valid data invalid data valid data valid data d15 to d8 high-z *2 valid data invalid data valid data d31 to d16 high-z *2 high-z *2 high-z *2 high-z *2 783 table a.10 pin states (pcmcia/big endian) (cont) pcmcia memory interface (area 6) pcmcia/io interface (area 6) 8-bit bus width 16-bit bus width 8-bit bus width 16-bit bus width pin byte/ word/ long- word access byte access (ad- dress 2n) byte access (ad- dress 2n + 1) word/ long- word access byte/ word/ long- word access byte access (ad- dress 2n) byte access (ad- dress 2n+1) word/ long- word access cs6 to cs2 , cs0 enabled enabled high enabled enabled enabled high enabled rd r low low low low high high high high w high high high high high high high high rd/ wr r high high high high high high high high w low low low low low low low low bs enabled enabled enabled enabled enabled enabled enabled enabled ras3u /pte[2] high high high high high high high high ras3l /ptj[0] high high high high high high high high casl /ptj[2] high high high high high high high high casu /ptj[3] high high high high high high high high we0 /dqmll r high high high high high high high high w high high high high high high high high we1 /dqmlu/ we r high high high high high high high high w low low low low high high high high we2 /dqmul/ iciord / r high high high high low low low low ptk[6] w high high high high high high high high we3 /dqmuu/ iciowr / r high high high high high high high high ptk[7] w high high high high low low low low ce2a /pte[4] high high high high high high high high ce2b /pte[5] high high low low high high low low cke/ptk[5] disabled disabled disabled disabled disabled disabled disabled disabled wait enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 enabled *1 iois16 /ptg[7] disabled disabled disabled disabled disabled disabled enabled enabled a25 to a0 address address address address address address address address d7 to d0 valid data invalid data valid data valid data valid data invalid data invalid data valid data d15 to d8 high-z* 2 valid data invalid data valid data high-z* 2 valid data invalid data valid data d31 to d16 high-z* 2 high-z* 2 high-z* 2 high-z* 2 high-z* 2 high-z* 2 high-z* 2 high-z* 2 notes: *1 disabled when wcr2 register wait setting is 0. *2 unused data pins should be switched to the port function, or pulled up or down. 784 appendix b memory-mapped control registers b.1 register address map table b.1 memory-mapped control registers control register module* 1 bus* 2 address* 4 size (bits) access size (bits)* 3 pteh ccn l h'fffffff0 32 32 ptel ccn l h'fffffff4 32 32 ttb ccn l h'fffffff8 32 32 tea ccn l h'fffffffc 32 32 mmucr ccn l h'ffffffe0 32 32 basra ccn l h'ffffffe4 32 32 basrb ccn l h'ffffffe8 32 32 ccr ccn l h'ffffffec 32 32 ccr2 ccn i h'040000b0 32 32 tra ccn l h'ffffffd0 32 32 expevt ccn l h'ffffffd4 32 32 intevt ccn l h'ffffffd8 32 32 bara ubc l h'ffffffb0 32 32 bamra ubc l h'ffffffb4 8 8 bbra ubc l h'ffffffb8 16 16 barb ubc l h'ffffffa0 32 32 bamrb ubc l h'ffffffa4 8 8 bbrb ubc l h'ffffffa8 16 16 bdrb ubc l h'ffffff90 32 32 bdmrb ubc l h'ffffff94 32 32 brcr ubc l h'ffffff98 16 16 frqcr cpg i h'ffffff80 16 16 stbcr cpg i h'ffffff82 8 8 stbcr2 cpg i h'ffffff88 8 8 wtcnt cpg i h'ffffff84 8 16 wtcsr cpg i h'ffffff86 8 16 bcr1 bsc i h'ffffff60 16 16 bcr2 bsc i h'ffffff62 16 16 785 table b.1 memory-mapped control registers (cont) control register module* 1 bus* 2 address* 4 size (bits) access size (bits)* 3 wcr1 bsc i h'ffffff64 16 16 wcr2 bsc i h'ffffff66 16 16 mcr bsc i h'ffffff68 16 16 pcr bsc i h'ffffff6c 16 16 rtcsr bsc i h'ffffff6e 16 16 rtcnt bsc i h'ffffff70 16 16 rtcor bsc i h'ffffff72 16 16 rfcr bsc i h'ffffff74 16 16 sdmr bsc i h'ffffd000C h'ffffeffe 8 mcscr0 bsc i h'ffffff50 16 16 mcscr1 bsc i h'ffffff52 16 16 mcscr2 bsc i h'ffffff54 16 16 mcscr3 bsc i h'ffffff56 16 16 mcscr4 bsc i h'ffffff58 16 16 mcscr5bsc i h'ffffff5a 16 16 mcscr6 bsc i h'ffffff5c 16 16 mcscr7 bsc i h'ffffff5e 16 16 r64cnt rtc p h'fffffec0 8 8 rseccnt rtc p h'fffffec2 8 8 rmincnt rtc p h'fffffec4 8 8 rhrcnt rtc p h'fffffec6 8 8 rwkcnt rtc p h'fffffec8 8 8 rdaycnt rtc p h'fffffeca 8 8 rmoncnt rtc p h'fffffecc 8 8 ryrcnt rtc p h'fffffece 8 8 rsecar rtc p h'fffffed0 8 8 rminar rtc p h'fffffed2 8 8 rhrar rtc p h'fffffed4 8 8 rwkar rtc p h'fffffed6 8 8 rdayar rtc p h'fffffed8 8 8 rmonar rtc p h'fffffeda 8 8 786 table b.1 memory-mapped control registers (cont) control register module* 1 bus* 2 address* 4 size (bits) access size (bits)* 3 rcr1 rtc p h'fffffedc 8 8 rcr2 rtc p h'fffffede 8 8 icr0 intc i h'fffffee0 16 16 ipra intc i h'fffffee2 16 16 iprb intc i h'fffffee4 16 16 tocr tmu p h'fffffe90 8 8 tstr tmu p h'fffffe92 8 8 tcor0 tmu p h'fffffe94 32 32 tcnt0 tmu p h'fffffe98 32 32 tcr0 tmu p h'fffffe9c 16 16 tcor1 tmu p h'fffffea0 32 32 tcnt1 tmu p h'fffffea4 32 32 tcr1 tmu p h'fffffea8 16 16 tcor2 tmu p h'fffffeac 32 32 tcnt2 tmu p h'fffffeb0 32 32 tcr2 tmu p h'fffffeb4 16 16 tcpr2 tmu p h'fffffeb8 32 32 scsmr sci p h'fffffe80 8 8 scbrr sci p h'fffffe82 8 8 scscr sci p h'fffffe84 8 8 sctdr sci p h'fffffe86 8 8 scssr sci p h'fffffe88 8 8 scrdr sci p h'fffffe8a 8 8 intevt2 intc i h'04000000 32 32 irr0 intc i h'04000004 16 8 irr1 intc i h'04000006 16 8 irr2 intc i h'04000008 16 8 icr1 intc i h'04000010 16 16 icr2 intc i h'04000012 16 16 inter intc i h'04000014 16 16 iprc intc i h'04000016 16 16 iprd intc i h'04000018 16 16 787 table b.1 memory-mapped control registers (cont) control register module* 1 bus* 2 address* 4 size (bits) access size (bits)* 3 ipre intc i h'0400001a 16 16 sar0 dmac p h'04000020 32 16,32 dar0 dmac p h'04000024 32 16,32 dmatcr0 dmac p h'04000028 32 16,32 chcr0 dmac p h'0400002c 32 8,16,32 sar1 dmac p h'04000030 32 16,32 dar1 dmac p h'04000034 32 16,32 dmatcr1 dmac p h'04000038 32 16,32 chcr1 dmac p h'0400003c 32 8,16,32 sar2 dmac p h'04000040 32 16,32 dar2 dmac p h'04000044 32 16,32 dmatcr2 dmac p h'04000048 32 16,32 chcr2 dmac p h'0400004c 32 8,16,32 sar3 dmac p h'04000050 32 16,32 dar3 dmac p h'04000054 32 16,32 dmatcr3 dmac p h'04000058 32 16,32 chcr3 dmac p h'0400005c 32 8,16,32 dmaor dmac p h'04000060 16 8,16 cmstr cmt p h'04000070 16 8,16,32 cmcsr cmt p h'04000072 16 8,16,32 cmcnt cmt p h'04000074 16 8,16,32 cmcor cmt p h'04000076 16 8,16,32 addrah a/d p h'04000080 8 8,16,32* 5, * 6 addral a/d p h'04000082 8 8,16* 5 addrbh a/d p h'04000084 8 8,16,32* 5, * 6 addrbl a/d p h'04000086 8 8,16* 5 addrch a/d p h'04000088 8 8,16,32* 5, * 6 addrcl a/d p h'0400008a 8 8,16* 5 addrdh a/d p h'0400008c 8 8,16,32* 5, * 6 addrdl a/d p h'0400008e 8 8,16* 5 adcsr a/d p h'04000090 8 8,16,32* 5, * 6 adcr a/d p h'04000092 8 8,16 788 table b.1 memory-mapped control registers (cont) control register module* 1 bus* 2 address* 4 size (bits) access size (bits)* 3 dadr0 d/a p h'040000a0 8 8,16,32* 5, * 6 dadr1 d/a p h'040000a2 8 8,16* 5 dacr d/a p h'040000a4 8 8,16,32 pacr port p h'04000100 16 16 pbcr port p h'04000102 16 16 pccr port p h'04000104 16 16 pdcr port p h'04000106 16 16 pecr port p h'04000108 16 16 pfcr port p h'0400010a 16 16 pgcr port p h'0400010c 16 16 phcr port p h'0400010e 16 16 pjcr port p h'04000110 16 16 pkcr port p h'04000112 16 16 plcr port p h'04000114 16 16 scpcr port p h'04000116 16 16 padr port p h'04000120 8 8 pbdr port p h'04000122 8 8 pcdr port p h'04000124 8 8 pddr port p h'04000126 8 8 pedr port p h'04000128 8 8 pfdr port p h'0400012a 8 8 pgdr port p h'0400012c 8 8 phdr port p h'0400012e 8 8 pjdr port p h'04000130 8 8 pkdr port p h'04000132 8 8 pldr port p h'04000134 8 8 scpdr port p h'04000136 8 8 scsmr1 irda p h'04000140 8 8 scbrr1 irda p h'04000142 8 8 scscr1 irda p h'04000144 8 8 scftdr1 irda p h'04000146 8 8 scssr1 irda p h'04000148 16 16 789 table b.1 memory-mapped control registers (cont) control register module* 1 bus* 2 address* 4 size (bits) access size (bits)* 3 scfrdr1 irda p h'0400014a 8 8 scfcr1 irda p h'0400014c 8 8 scfdr1 irda p h'0400014e 16 16 scsmr2 scif p h'04000150 8 8 scbrr2 scif p h'04000152 8 8 scscr2 scif p h'04000154 8 8 scftdr2 scif p h'04000156 8 8 scssr2 scif p h'04000158 16 16 scfrdr2 scif p h'0400015a 8 8 scfcr2 scif p h'0400015c 8 8 scfdr2 scif p h'0400015e 16 16 sdir udi i h'04000200 16 16 sdsr udi i h'04000204 16 16 sddr/sddrh udi i h'04000208 16/32 16/32 sddrl udi i h'0400020a 16 16 sdar udi i h'0400020c 16 16 sdare udi i h'04000210 16 16 notes: *1 modules: ccn: cache controller ubc: user break controller cpg: clock pulse generator bsc: bus state controller rtc: realtime clock intc: interrupt controller tmu: timer unit sci: serial communication interface *2 internal buses: l: cpu, ccn, cache, tlb, and dsp connected i: bsc, cache, dmac, intc, cpg, and h-udi connected p: bsc and peripheral modules (rtc, tmu, sci, scif, irda, a/d, d/a, dmac, ports, cmt) connected *3 the access size shown is for control register access (read/write). an incorrect result will be obtained if a different size from that shown is used for access. *4 to exclude area 1 control registers from address translation by the mmu, set the first 3 bits of the logical address to 101, to locate the registers in the p2 space. *5with 16-bit access, it is not possible to read data in two registers simultaneously. *6 with 32-bit access, it is possible to read data in the register at [accessed address + 2] simultaneously. 790 b.2 register bits table b.2 register bits register bit 7 bit 6 bit 5 bit 4 bit 3 bit 2 bit 1 bit 0 module sdmr bsc scsmr c/a chr pe o/e stop mp cks1 cks0 sci scbrr sci scscr tie rie te re mpie teie cke1 cke0 sci sctdr sci scssr tdre rdrf orer fer per tend mpb mpbt sci scrdr sci scscmr sdir sinv smif sci tocr tcoe tmu tstr str2 str1 str0 tmu tcor0 tmu tcnt0 tmu tcr0 unftmu unie ckeg1 ckeg0 tpsc2 tpsc1 tpsc0 tcor1 tmu tcnt1 tmu 791 table b.2 register bits (cont) register bit 7 bit 6 bit 5 bit 4 bit 3 bit 2 bit 1 bit 0 module tcr1 unftmu unie ckeg1 ckeg0 tpsc2 tpsc1 tpsc0 tcor2 tmu tcnt2 tmu tcr2 icpf unf tmu icpe1 icpe0 unie ckeg1 ckeg0 tpsc2 tpsc1 tpsc0 tcpr2 tmu r64cnt 1 hz 2 hz 4 hz 8 hz 16 hz 32 hz 64 hz rtc rseccnt 10 sec 1 sec rtc rmincnt 10 min 1 min rtc rhrcnt 10 hours 1 hour rtc rwkcnt day of week rtc rdaycnt 10 days 1 day rtc rmoncnt 10 months 1 month rtc ryrcnt 10 years 1 year rtc rsecar enb 10 sec 1 sec rtc rminar enb 10 min 1 min rtc rhrar enb 10 hours 1 hour rtc rwkar enb day of week rtc rdayar enb 10 days 1 day rtc rmonar enb 10 months 1 month rtc 792 table b.2 register bits (cont) register bit 7 bit 6 bit 5 bit 4 bit 3 bit 2 bit 1 bit 0 module rcr1 cf cie aie af rtc rcr2 pef pes2 pes1 pes0 rtcen adj reset start rtc icr0 nml nmie intc ipra tmu0 tmu1 intc tmu2 rtc iprb wdt ref intc sci bcr1 pula puld hizmem hizcnt endian a0bst1 a0bst0 a5bst1 bsc a5bst0 a6bst1 a6bst0 dramt p2 dramt p1 dramt p0 a5pcm a6pcm bcr2 a6sz1 a6sz0 a5sz1 a5sz0 a4sz1 a4sz0 bsc a3sz1 a3sz0 a2sz1 a2sz0 wcr1 waitsel a6iw1 a6iw0 a5iw1 a5iw0 a4iw1 a4iw0 bsc a3iw1 a3iw0 a2iw1 a2iw0 a0iw1 a0iw0 wcr2 a6w2 a6w1 a6w0 a5w2 a5w1 a5w0 a4w2 a4w1 bsc a4w0 a3w1 a3w0 a2w1 a2w0 a0w2 a0w1 a0w0 mcr tpc1 tpc0 rcd1 rcd0 trwl1 trwl0 tras1 tras0 bsc rasd amx3 amx2 amx1 amx0 rfsh rmode edomo de pcr a6w3 a5w3 a5ted2 a6ted2 a5teh2 a6teh2 bsc a5ted1 a5ted0 a6ted1 a6ted0 a5teh1 a5teh0 a6teh1 a6teh0 rtcsr bsc cmf cmie cks2 cks1 cks0 ovf ovie lmts rtcnt bsc rtcor bsc 793 table b.2 register bits (cont) register bit 7 bit 6 bit 5 bit 4 bit 3 bit 2 bit 1 bit 0 module rfcr bsc frqcr stc2 ifc2 pfc2 slpfr q ckoen cpg pllen pstby stc1 stc0 ifc1 ifc0 pfc1 pfc0 stbcr stby stbxtl mstp2 mstp1 mstp0 cpg stbcr2 mstp9 mdchg mstp8 mstp7 mstp6 mstp5mstp4 mstp3 cpg wtcnt cpg wtcsr tme wt/it rsts wovf iovf cks2 cks1 cks0 cpg bdrb ubc bdmrb ubc brcr ubc basma basmb scmfca scmfcb scmfda scmfdb pcte pcba dbeb pcbb seq etbe barb ubc bamrb basm bam bam ubc bbrb xye xys ubc cdb1 cdb0 idb1 idb0 rwb1 rwb0 szb1 szb0 794 table b.2 register bits (cont) register bit 7 bit 6 bit 5 bit 4 bit 3 bit 2 bit 1 bit 0 module bara ubc bamra ubc bbra ubc cda1 cda0 ida1 ida0 rwa1 rwa0 sza1 sza0 betr ubc brsr svf pid2 pid1 pid0 bsa27 bsa26 bsa25bsa24 ubc bsa23 bsa22 bsa21 bsa20 bsa19 bsa18 bsa17 bsa16 bsa15bsa14 bsa13 bsa12 bsa11 bsa10 bsa9 bsa8 bsa7 bsa6 bsa5bsa4 bsa3 bsa2 bsa1 bsa0 brdr dvf bda27 bda26 bda25bda24 ubc bda23 bda22 bda21 bda20 bda19 bda18 bda17 bda16 bda15bda14 bda13 bda12 bda11 bda10 bda9 bda8 bda7 bda6 bda5bda4 bda3 bda2 bda1 bda0 tra ccn expevt ccn intevt ccn 795 table b.2 register bits (cont) register bit 7 bit 6 bit 5 bit 4 bit 3 bit 2 bit 1 bit 0 module mmucr ccn sv rc rc tf ix at basra ubc basrb ubc ccr ccn 00cfcbwtce cdr2 ccn w3load w3lock w2load w2lock pteh ccn ptel ccn v prprsz c d sh ttb ccn tea ccn 796 table b.2 register bits (cont) register bit 7 bit 6 bit 5 bit 4 bit 3 bit 2 bit 1 bit 0 module intevt2 intc irr0 intc pint0r pint1r irq5r irq4r irq3r irq2r irq1r irq0r irr1 intc txi1r bri1r rxi1r eri1r dei3r dei2r dei1r dei0r irr2 intc adir txi2r bri2r rxi2r eri2r icr1 mai irqlvl blmsk irq51s irq50s irq41s irq40s intc irq31s irq30s irq21s irq20s irq11s irq10s irq01s irq00s icr2 int15s int14s int13s int12s int11s int10s int9s int8s intc int7s int6s int5s int4s int3s int2s int1s int0s inter int15e int14e int13e int12e int11e int10e int9e int8e intc int7e int6e int5e int4e int3e int2e int1e int0e iprc irq3 level irq2 level intc irq1 level irq0 level iprd pint0 to 7 level pint8 to 15 level intc irq5 level irq4 level ipre dmac level irda level intc scif level a/d level sar0 dmac dar0 dmac 797 table b.2 register bits (cont) register bit 7 bit 6 bit 5 bit 4 bit 3 bit 2 bit 1 bit 0 module dmatcr0 dmac chcr0 dmac diroamal dm1 dm0 sm1 sm0 rs3 rs2 rs1 rs0 ds tm ts1 ts0 ie te de sar1 dmac dar1 dmac dmatcr1 dmac chcr1 dmac diroamal dm1 dm0 sm1 sm0 rs3 rs2 rs1 rs0 ds tm ts1 ts0 ie te de sar2 dmac dar2 dmac 798 table b.2 register bits (cont) register bit 7 bit 6 bit 5 bit 4 bit 3 bit 2 bit 1 bit 0 module dmatcr2 dmac chcr2 dmac dirorlamal dm1 dm0 sm1 sm0 rs3 rs2 rs1 rs0 ds tm ts1 ts0 ie te de sar3 dmac dar3 dmac dmatcr3 dmac chcr3 dmac dirorlamal dm1 dm0 sm1 sm0 rs3 rs2 rs1 rs0 ds tm ts1 ts0 ie te de dmaor pr1pr0 dmac ae nmif dme cmstr cmt str cmcsr cmt cmf cks1 cks0 cmcnt cmt 799 table b.2 register bits (cont) register bit 7 bit 6 bit 5 bit 4 bit 3 bit 2 bit 1 bit 0 module cmcor cmt addrah ad9 ad8 ad7 ad6 ad5ad4 ad3 ad2 a/d addral ad1 ad0 a/d addrbh ad9 ad8 ad7 ad6 ad5ad4 ad3 ad2 a/d addrbl ad1 ad0 a/d addrch ad9 ad8 ad7 ad6 ad5ad4 ad3 ad2 a/d addrcl ad1 ad0 a/d addrdh ad9 ad8 ad7 ad6 ad5ad4 ad3 ad2 a/d addrdl ad1 ad0 a/d adcsr adf ade adst scan cks ch2 ch1 ch0 a/d adcr trge a/d dadr0 d/a dadr1 d/a dacr daoe1 daoe0 dae d/a pacr pa7m d1 pa7m d0 pa6m d1 pa6m d0 pa5m d1 pa5m d0 pa4m d1 pa4m d0 port pa3m d1 pa3m d0 pa2m d1 pa2m d0 pa1m d1 pa1m d0 pa0m d1 pa0m d0 pbcr pb7m d1 pb7m d0 pb6m d1 pb6m d0 pb5m d1 pb5m d0 pb4m d1 pb4m d0 port pb3m d1 pb3m d0 pb2m d1 pb2m d0 pb1m d1 pb1m d0 pb0m d1 pb0m d0 pcdr pc7m d1 pc7m d0 pc6m d1 pc6m d0 pc5m d1 pc5m d0 pc4m d1 pc4m d0 port pc3m d1 pc3m d0 pc2m d1 pc2m d0 pc1m d1 pc1m d0 pc0m d1 pc0m d0 pdcr pd7m d1 pd7m d0 pd6m d1 pd6m d0 pd5m d1 pd5m d0 pd4m d1 pd4m d0 port pecr pe7m d1 pe7m d0 pe6m d1 pe6m d0 pe5m d1 pe5m d0 pe4m d1 pe4m d0 port pe3m d1 pe3m d0 pe2m d1 pe2m d0 pe1m d1 pe1m d0 pe0m d1 pe0m d0 800 table b.2 register bits (cont) register bit 7 bit 6 bit 5 bit 4 bit 3 bit 2 bit 1 bit 0 module pfcr pf7m d1 pf7m d0 pf6m d1 pf6m d0 pf5m d1 pf5m d0 pf4m d1 pf4m d0 port pf3m d1 pf3m d0 pf2m d1 pf2m d0 pf1m d1 pf1m d0 pf0m d1 pf0m d0 pgcr pg7m d1 pg7m d0 pg6m d1 pg6m d0 pg5m d1 pg5m d0 pg4m d1 pg4m d0 port pg3m d1 pg3m d0 pg2m d1 pg2m d0 pg1m d1 pg1m d0 pg0m d1 pg0m d0 phcr ph7m d1 ph7m d0 ph6m d1 ph6m d0 ph5m d1 ph5m d0 ph4m d1 ph4m d0 port ph3m d1 ph3m d0 ph2m d1 ph2m d0 ph1m d1 ph1m d0 ph0m d1 ph0m d0 pjcr pj7m d1 pj7m d0 pj6m d1 pj6m d0 pj5m d1 pj5m d0 pj4m d1 pj4m d0 port pj3m d1 pj3m d0 pj2m d1 pj2m d0 pj1m d1 pj1m d0 pj0m d1 pj0m d0 pkcr pk7m d1 pk7m d0 pk6m d1 pk6m d0 pk5m d1 pk5m d0 pk4m d1 pk4m d0 port pk3m d1 pk3m d0 pk2m d1 pk2m d0 pk1m d1 pk1m d0 pk0m d1 pk0m d0 plcr pl7m d1 pl7m d0 pl6m d1 pl6m d0 pl5m d1 pl5m d0 pl4m d1 pl4m d0 port pl3m d1 pl3m d0 pl2m d1 pl2m d0 pl1m d1 pl1m d0 pl0m d1 pl0m d0 scpcr scp7m d1 scp7m d0 scp6m d1 scp6m d0 scp5m d1 scp5m d0 scp4m d1 scp4m d0 port scp3m d1 scp3m d0 scp2m d1 scp2m d0 scp1m d1 scp1m d0 scp0m d1 scp0m d0 padr pa7dt pa6dt pa5dt pa4dt pa3dt pa2dt pa1dt pa0dt port pbdr pb7dt pb6dt pb5dt pb4dt pb3dt pb2dt pb1dt pb0dt port pcdr pc7dt pc6dt pc5dt pc4dt pc3dt pc2dt pc1dt pc0dt port pddr pd7dt pd6dt pd5dt pd4dt pd3dt pd2dt pd1dt pd0dt port pedr pe7dt pe6dt pe5dt pe4dt pe3dt pe2dt pe1dt pe0dt port pfdr pf7dt pf6dt pf5dt pf4dt pf3dt pf2dt pf1dt pf0dt port pgdr pg7dt pg6dt pg5dt pg4dt pg3dt pg2dt pg1dt pg0dt port phdr ph7dt ph6dt ph5dt ph4dt ph3dt ph2dt ph1dt ph0dt port 801 table b.2 register bits (cont) register bit7 bit6 bit5 bit4 bit3 bit2 bit1 bit0 module pjdr pj7dt pj6dt pj5dt pj4dt pj3dt pj2dt pj1dt pj0dt port pkdr pk7dt pk6dt pk5dt pk4dt pk3dt pk2dt pk1dt pk0dt port pldr pl7dt pl6dt pl5dt pl4dt pl3dt pl2dt pl1dt pl0dt port scpdr scp7dt scp6dt scp5dt scp4dt scp3dt scp2dt scp1dt scp0dt port sdir ti3 ti2 ti1 ti0 h-udi sdsr pr3 pr2 pr1 pr0 vr3 vr2 vr1 vr0 h-udi asemw brkaf sdtrf sddr h-udi (sddrl) sdar ar9ar8 h-udi ar7 ar6 ar5ar4 ar3 ar2 sdare are9 are8 h-udi are7 are6 are5are4 are3 are2 802 appendix c product lineup table c.1 sh7729r models power supply voltage operating abbr. i/o internal frequency model marking package sh7729r 3.3 0.3 v 2.0 0.15 v 200 mhz HD6417729Rhf200 208-pin plastic hqfp (fp-208e) 1.9 0.15 v 167 mhz HD6417729Rf167 208-pin plastic lqfp (fp-208c) HD6417729Rbp167v 240-pin csp (bp-240a) 1.8 + 0.25 v 1.8 - 0.15v 133 mhz HD6417729Rf133 208-pin plastic lqfp (fp-208c) HD6417729Rbp133v 240-pin csp (bp-240a) 1.7 + 0.25 v 1.7 - 0.15v 100 mhz HD6417729Rf100 208-pin plastic lqfp (fp-208c) HD6417729Rbp100v 240-pin csp (bp-240a) 803 appendix d package dimensions figure d.1, d.2, and d.3 show the sh7729r package dimensions. hitachi code jedec eiaj weight (reference value) fp-208c conforms 2.7 g unit: mm *dimension including the plating thickness base material dimension 30.0 0.2 30.0 0.2 0.5 1.70 max 0 C 8 *0.17 0.05 156 105 104 52 1 157 208 53 *0.22 0.05 0.08 m 0.08 1.40 0.5 0.1 1.0 28 0.10 0.05 1.25 0.20 0.04 0.15 0.04 figure d.1 package dimensions (fp-208c) 804 hitachi code jedec eiaj weight (reference value) fp-208e conforms 5.3 g unit: mm *dimension including the plating thickness base material dimension 30.6 0.2 30.6 0.2 0.5 3.56 max 0 C 8 *0.17 0.05 156 105 104 52 1 157 208 53 *0.22 0.05 0.10 m 0.10 3.20 0.5 0.1 1.3 28 0.15 +0.10 C0.15 1.25 0.20 0.04 0.15 0.04 figure d.2 package dimensions (fp-208e) 805 13.00 0.20 c a 13.00 0.65 0.20 c a 0.15 0.2 c c 4 ? 0.10 c 1.40max 0.33 0.05 b a 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 a b c d e f g h j k l m n p r t u v w 0.65 0.65 0.65 ?.08 c mab 240 ? ?.40 0.05 preliminary figure d.3 package dimensions (bp-240av) 806 sh7729r hardware manual publication date: 1st edition, august 2001 published by: customer service division semiconductor & integrated circuits hitachi, ltd. edited by: technical documentation group hitachi kodaira semiconductor co., ltd. co py ri g ht ? hitachi, ltd., 2001. all ri g hts reserved. printed in ja p an. |
Price & Availability of HD6417729R
 |
|
|
|
|
All Rights Reserved © IC-ON-LINE 2003 - 2022 |
| [Add Bookmark] [Contact Us] [Link exchange] [Privacy policy] |
|
Mirror Sites : [www.datasheet.hk]
[www.maxim4u.com] [www.ic-on-line.cn]
[www.ic-on-line.com] [www.ic-on-line.net]
[www.alldatasheet.com.cn]
[www.gdcy.com]
[www.gdcy.net] |