Intel386™ EX
Embedded Microprocessor
User’s Manual
Intel386™ EXTB
Embedded
Microprocessor
Intel386™ EXTC
Embedded
Microprocessor
Intel386™ EX
Embedded
Microprocessor
User’s Manual
1996
Information in this document is provided in connection with Intel products. Intel assumes no liability whatsoever, including in-
fringement of any patent or copyright, for sale and use of Intel products except as provided in Intel’s Terms and Conditions of
Sale for such products.
Intel retains the right to make changes to these specifications at any time, without notice. Microcontroller products may have
minor variations to this specification known as errata.
*Other brands and names are the property of their respective owners.
Contact your local Intel sales office or your distributor to obtain the latest specifications before placing your product order.
Copies of documents which have an ordering number and are referenced in this document, or other Intel literature, may be
obtained from:
Intel Corporation
Literature Sales
P.O. Box 7641
Mt. Prospect, IL 60056-7641
or call 1-800-548-4725
COPYRIGHT © INTEL CORPORATION, 1996
CONTENTS
CHAPTER 1
GUIDE TO THIS MANUAL
1.1
1.2
1.3
1.4
1.5
MANUAL CONTENTS................................................................................................... 1-1
NOTATIONAL CONVENTIONS..................................................................................... 1-3
SPECIAL TERMINOLOGY............................................................................................ 1-4
RELATED DOCUMENTS .............................................................................................. 1-5
ELECTRONIC SUPPORT SYSTEMS........................................................................... 1-6
FaxBack Service .......................................................................................................1-6
Bulletin Board System (BBS) ....................................................................................1-7
CompuServe Forums ................................................................................................1-7
World Wide Web .......................................................................................................1-7
TECHNICAL SUPPORT ................................................................................................ 1-7
PRODUCT LITERATURE.............................................................................................. 1-8
1.5.1
1.5.2
1.5.3
1.5.4
1.6
1.7
CHAPTER 2
ARCHITECTURAL OVERVIEW
2.1
2.2
Intel386 EX EMBEDDED PROCESSOR CORE............................................................ 2-1
INTEGRATED PERIPHERALS...................................................................................... 2-3
CHAPTER 3
CORE OVERVIEW
3.1
3.1.1
3.1.2
3.2
3.2.1
3.2.2
3.2.3
3.2.4
3.2.5
3.2.6
Intel386 CX PROCESSOR ENHANCEMENTS............................................................. 3-1
System Management Mode ......................................................................................3-1
Additional Address Lines ..........................................................................................3-1
Intel386 CX PROCESSOR INTERNAL ARCHITECTURE ............................................ 3-2
Core Bus Unit ............................................................................................................3-4
Instruction Prefetch Unit ............................................................................................3-4
Instruction Decode Unit .............................................................................................3-4
Execution Unit ...........................................................................................................3-5
Segmentation Unit ....................................................................................................3-5
Paging Unit ...............................................................................................................3-5
CORE Intel386 EX PROCESSOR INTERFACE............................................................ 3-6
3.3
CHAPTER 4
SYSTEM REGISTER ORGANIZATION
4.1
4.1.1
4.1.2
4.2
OVERVIEW ................................................................................................................... 4-1
Intel386 Processor Core Architecture Registers .......................................................4-2
Intel386 EX Processor Peripheral Registers .............................................................4-2
I/O ADDRESS SPACE FOR PC/AT SYSTEMS ............................................................ 4-2
EXPANDED I/O ADDRESS SPACE.............................................................................. 4-3
ORGANIZATION OF PERIPHERAL REGISTERS........................................................ 4-5
I/O ADDRESS DECODING TECHNIQUES................................................................... 4-6
Address Configuration Register ................................................................................4-6
4.3
4.4
4.5
4.5.1
iii
Intel386™ EX MICROPROCESSOR USER’S MANUAL
4.5.2
4.5.2.1
4.6
4.6.1
4.6.2
4.6.3
4.6.4
Enabling and Disabling the Expanded I/O Space .....................................................4-8
Programming REMAPCFG Example ...................................................................4-8
ADDRESSING MODES................................................................................................. 4-9
DOS-compatible Mode ..............................................................................................4-9
Nonintrusive DOS Mode .........................................................................................4-11
Enhanced DOS Mode .............................................................................................4-11
Non-DOS Mode ......................................................................................................4-11
PERIPHERAL REGISTER ADDRESSES.................................................................... 4-15
4.7
CHAPTER 5
DEVICE CONFIGURATION
5.1
5.2
5.2.1
5.2.1.1
5.2.1.2
5.2.1.3
5.2.1.4
5.2.2
5.2.3
5.2.4
5.2.5
5.2.6
5.2.7
INTRODUCTION ........................................................................................................... 5-1
PERIPHERAL CONFIGURATION................................................................................. 5-3
DMA Controller, Bus Arbiter, and Refresh Unit Configuration ..................................5-3
Using The DMA Unit with External Devices .........................................................5-3
DMA Service to an SIO or SSIO Peripheral .........................................................5-3
Using The Timer To Initiate DMA Transfers .........................................................5-4
Limitations Due To Pin Signal Multiplexing ..........................................................5-4
Interrupt Control Unit Configuration ..........................................................................5-7
Timer/counter Unit Configuration ............................................................................5-11
Asynchronous Serial I/O Configuration ...................................................................5-14
Synchronous Serial I/O Configuration ....................................................................5-18
Chip-select Unit and Clock and Power Management Unit Configuration ................5-19
Core Configuration ..................................................................................................5-21
5.3
PIN CONFIGURATION................................................................................................ 5-23
DEVICE CONFIGURATION PROCEDURE ................................................................ 5-28
CONFIGURATION EXAMPLE..................................................................................... 5-28
Example Design Requirements ...............................................................................5-28
Example Design Solution ........................................................................................5-29
5.4
5.5
5.5.1
5.5.2
CHAPTER 6
BUS INTERFACE UNIT
6.1
6.1.1
6.2
6.2.1
6.2.2
6.2.3
6.2.4
OVERVIEW ................................................................................................................... 6-1
Bus Signal Descriptions ............................................................................................6-3
BUS OPERATION ......................................................................................................... 6-5
Bus States .................................................................................................................6-7
Pipelining ..................................................................................................................6-8
Data Bus Transfers and Operand Alignment ............................................................6-9
Ready Logic ............................................................................................................6-10
BUS CYCLES.............................................................................................................. 6-13
Read Cycle .............................................................................................................6-13
Write Cycle ..............................................................................................................6-16
Pipelined Cycle .......................................................................................................6-19
6.3
6.3.1
6.3.2
6.3.3
iv
CONTENTS
6.3.4
6.3.5
6.3.6
6.3.7
6.3.7.1
6.3.7.2
6.4
6.4.1
Interrupt Acknowledge Cycle ..................................................................................6-23
Halt/Shutdown Cycle ...............................................................................................6-26
Refresh Cycle .........................................................................................................6-28
BS8 Cycle ...............................................................................................................6-31
Write Cycles .......................................................................................................6-31
Read Cycles .......................................................................................................6-31
BUS LOCK................................................................................................................... 6-34
Locked Cycle Activators ..........................................................................................6-34
Locked Cycle Timing ...............................................................................................6-34
LOCK# Signal Duration ...........................................................................................6-35
EXTERNAL BUS MASTER SUPPORT (USING HOLD, HLDA).................................. 6-35
HOLD/HLDA Timing ................................................................................................6-36
HOLD Signal Latency .............................................................................................6-37
DESIGN CONSIDERATIONS...................................................................................... 6-38
Interface To Intel387™ SX Math Coprocessor .......................................................6-38
6.4.2
6.4.3
6.5
6.5.1
6.5.2
6.6
6.6.1
6.6.1.1
6.6.1.2
6.6.2
6.6.3
6.6.4
6.6.5
System Configuration .........................................................................................6-39
Software Considerations ....................................................................................6-40
SRAM/FLASH Interface ..........................................................................................6-41
PSRAM Interface ....................................................................................................6-42
Paged DRAM Interface ...........................................................................................6-43
Non-Paged DRAM Interface ...................................................................................6-44
CHAPTER 7
SYSTEM MANAGEMENT MODE
7.1
7.2
7.2.1
7.2.2
7.2.3
SYSTEM MANAGEMENT MODE OVERVIEW ............................................................. 7-1
SMM HARDWARE INTERFACE................................................................................... 7-1
System Management Interrupt Input (SMI#) .............................................................7-1
SMM Active Output (SMIACT#) ................................................................................7-2
System Management RAM (SMRAM) ......................................................................7-2
SYSTEM MANAGEMENT MODE PROGRAMMING AND CONFIGURATION............. 7-3
Register Status During SMM .....................................................................................7-3
System Management Interrupt ..................................................................................7-4
7.3
7.3.1
7.3.2
7.3.2.1
7.3.2.2
7.3.2.3
7.3.2.4
7.3.2.5
SMI# Priority .........................................................................................................7-7
System Management Interrupt During HALT Cycle .............................................7-8
HALT Restart .......................................................................................................7-9
System Management Interrupt During I/O Instruction ..........................................7-9
I/O Restart ..........................................................................................................7-10
SMM Handler Interruption .......................................................................................7-10
Interrupt During SMM Handler ...........................................................................7-10
HALT During SMM Handler ................................................................................7-11
Idle Mode and Powerdown Mode During SMM ..................................................7-12
SMI# During SMM Operation .............................................................................7-12
SMRAM Programming ............................................................................................7-12
Chip-select Unit Support for SMRAM .................................................................7-12
7.3.3
7.3.3.1
7.3.3.2
7.3.3.3
7.3.3.4
7.3.4
7.3.4.1
v
Intel386™ EX MICROPROCESSOR USER’S MANUAL
7.3.4.2
7.3.5
7.4
7.5
7.5.1
SMRAM State Dump Area .................................................................................7-14
Resume Instruction (RSM) ......................................................................................7-15
THE Intel386 EX PROCESSOR IDENTIFIER REGISTERS ....................................... 7-15
PROGRAMMING CONSIDERATIONS........................................................................ 7-16
System Management Mode Code Example ............................................................7-16
CHAPTER 8
CLOCK AND POWER MANAGEMENT UNIT
8.1
8.1.1
8.1.2
8.1.2.1
8.1.2.2
8.1.2.3
8.1.3
8.2
8.3
8.3.1
OVERVIEW ................................................................................................................... 8-1
Clock Generation Logic .............................................................................................8-1
Power Management Logic ........................................................................................8-3
SMM Interaction with Power Management Modes ...............................................8-4
Bus Interface Unit Operation During Idle Mode ....................................................8-5
Watchdog Timer Unit Operation During Idle Mode ..............................................8-5
Clock and Power Management Registers and Signals .............................................8-6
CONTROLLING THE PSCLK FREQUENCY ................................................................ 8-7
CONTROLLING POWER MANAGEMENT MODES ..................................................... 8-8
Idle Mode ..................................................................................................................8-9
Powerdown Mode ...................................................................................................8-10
Ready Generation During HALT .............................................................................8-10
DESIGN CONSIDERATIONS...................................................................................... 8-11
Reset Considerations ..............................................................................................8-11
Power-up Considerations ........................................................................................8-12
8.3.2
8.3.3
8.4
8.4.1
8.4.2
8.4.2.1
8.4.2.2
8.4.3
8.5
8.5.1
Built-in Self Test .................................................................................................8-12
JTAG Reset ........................................................................................................8-12
Powerdown Mode and Idle Mode Considerations ...................................................8-13
PROGRAMMING CONSIDERATIONS........................................................................ 8-13
Clock and Power Management Unit Code Example ...............................................8-13
CHAPTER 9
INTERRUPT CONTROL UNIT
9.1
9.2
9.2.1
9.2.2
9.2.2.1
9.2.2.2
9.2.3
9.2.4
9.2.5
OVERVIEW ................................................................................................................... 9-1
ICU OPERATION........................................................................................................... 9-4
Interrupt Sources ......................................................................................................9-4
Interrupt Priority ........................................................................................................9-6
Assigning an Interrupt Level .................................................................................9-6
Determining Priority ..............................................................................................9-7
Interrupt Vectors .......................................................................................................9-8
Interrupt Process .......................................................................................................9-9
Poll Mode ................................................................................................................9-14
9.3
9.3.1
9.3.2
REGISTER DEFINITIONS........................................................................................... 9-15
Port 3 Configuration Register (P3CFG) ..................................................................9-18
Interrupt Configuration Register (INTCFG) .............................................................9-19
vi
CONTENTS
9.3.3
9.3.4
9.3.5
9.3.6
9.3.7
9.3.8
9.3.9
Initialization Command Word 1 (ICW1) ...................................................................9-20
Initialization Command Word 2 (ICW2) ...................................................................9-21
Initialization Command Word 3 (ICW3) ...................................................................9-22
Initialization Command Word 4 (ICW4) ...................................................................9-24
Operation Command Word 1 (OCW1) ....................................................................9-25
Operation Command Word 2 (OCW2) ....................................................................9-26
Operation Command Word 3 (OCW3) ....................................................................9-27
9.3.10 Interrupt Request Register (IRR) ............................................................................9-28
9.3.11 In-Service Register (ISR) ........................................................................................9-28
9.3.12 Poll Status Byte (POLL) ..........................................................................................9-28
9.4
DESIGN CONSIDERATIONS...................................................................................... 9-29
Interrupt Acknowledge Cycle ..................................................................................9-29
Interrupt Detection ..................................................................................................9-29
Spurious Interrupts ..................................................................................................9-30
Cascading Interrupt Controllers ..............................................................................9-30
PROGRAMMING CONSIDERATIONS........................................................................ 9-32
Interrupt Control Unit Code Examples ....................................................................9-32
9.4.1
9.4.2
9.4.3
9.4.4
9.5
9.5.1
CHAPTER 10
TIMER/COUNTER UNIT
10.1 OVERVIEW ................................................................................................................. 10-1
10.1.1 TCU Signals and Registers .....................................................................................10-3
10.2 TCU OPERATION ....................................................................................................... 10-5
10.2.1 Mode 0 – Interrupt on Terminal Count ....................................................................10-6
10.2.2 Mode 1 – Hardware Retriggerable One-shot ..........................................................10-8
10.2.3 Mode 2 – Rate Generator .....................................................................................10-10
10.2.4 Mode 3 – Square Wave ........................................................................................10-12
10.2.5 Mode 4 – Software-triggered Strobe .....................................................................10-16
10.2.6 Mode 5 – Hardware-triggered Strobe ....................................................................10-18
10.3 REGISTER DEFINITIONS......................................................................................... 10-20
10.3.1 Configuring the Input and Output Signals .............................................................10-20
10.3.1.1 Hardware Control of GATEn ............................................................................10-20
10.3.1.2 Software Control of GATEn ..............................................................................10-20
10.3.2 Initializing the Counters .........................................................................................10-24
10.3.3 Writing the Counters .............................................................................................10-26
10.3.4 Reading the Counter .............................................................................................10-27
10.3.4.1 Simple Read .....................................................................................................10-27
10.3.4.2 Counter-latch Command ..................................................................................10-27
10.3.4.3 Read-back Command ......................................................................................10-30
10.4 PROGRAMMING CONSIDERATIONS...................................................................... 10-33
10.4.1 Timer/Counter Unit Code Examples .....................................................................10-34
vii
Intel386™ EX MICROPROCESSOR USER’S MANUAL
CHAPTER 11
ASYNCHRONOUS SERIAL I/O UNIT
11.1 OVERVIEW ................................................................................................................. 11-1
11.1.1 SIO Signals .............................................................................................................11-3
11.2 SIO OPERATION ........................................................................................................ 11-4
11.2.1 Baud-rate Generator ...............................................................................................11-4
11.2.2 SIOn Transmitter .....................................................................................................11-6
11.2.3 SIOn Receiver .........................................................................................................11-9
11.2.4 Modem Control .....................................................................................................11-12
11.2.5 Diagnostic Mode ...................................................................................................11-12
11.2.6 SIO Interrupt and DMA Sources ...........................................................................11-13
11.2.6.1 SIO Interrupt Sources ......................................................................................11-13
11.2.6.2 SIO DMA sources ............................................................................................11-13
11.2.7 External UART Support ........................................................................................11-14
11.3 REGISTER DEFINITIONS......................................................................................... 11-15
11.3.1 Pin and Port Configuration Registers (PINCFG and PnCFG [n = 1–3]) ................11-17
11.3.2 SIO and SSIO Configuration Register (SIOCFG) .................................................11-21
11.3.3 Divisor Latch Registers (DLLn and DLHn) ............................................................11-22
11.3.4 Transmit Buffer Register (TBRn) ...........................................................................11-23
11.3.5 Receive Buffer Register (RBRn) ...........................................................................11-24
11.3.6 Serial Line Control Register (LCRn) ......................................................................11-25
11.3.7 Serial Line Status Register (LSRn) .......................................................................11-26
11.3.8 Interrupt Enable Register (IERn) ...........................................................................11-27
11.3.9 Interrupt ID Register (IIRn) ....................................................................................11-28
11.3.10 Modem Control Register (MCRn) ..........................................................................11-29
11.3.11 Modem Status Register (MSRn) ...........................................................................11-31
11.3.12 Scratch Pad Register (SCRn) ...............................................................................11-32
11.4 PROGRAMMING CONSIDERATIONS...................................................................... 11-32
11.4.1 Asynchronous Serial I/O Unit Code Examples ......................................................11-33
CHAPTER 12
DMA CONTROLLER
12.1 OVERVIEW ................................................................................................................. 12-1
12.1.1 DMA Terminology ...................................................................................................12-3
12.1.2 DMA Signals ...........................................................................................................12-4
12.2 DMA OPERATION....................................................................................................... 12-5
12.2.1 DMA Transfers ........................................................................................................12-5
12.2.2 Bus Cycle Options for Data Transfers .....................................................................12-5
12.2.2.1 Fly-By Mode .......................................................................................................12-5
12.2.2.2 Two-Cycle Mode ................................................................................................12-6
12.2.2.3 Programmable DMA Transfer Direction .............................................................12-6
12.2.2.4 Ready Generation For DMA Cycles ...................................................................12-7
12.2.2.5 DMA Usage of the 4-Byte Temporary Register ..................................................12-7
12.2.3 Starting DMA Transfers ..........................................................................................12-9
viii
CONTENTS
12.2.4 Bus Control Arbitration ............................................................................................12-9
12.2.5 Ending DMA Transfers ..........................................................................................12-10
12.2.6 Buffer-transfer Modes ...........................................................................................12-12
12.2.6.1 Single Buffer-Transfer Mode ............................................................................12-12
12.2.6.2 Autoinitialize Buffer-Transfer Mode ..................................................................12-12
12.2.6.3 Chaining Buffer-Transfer Mode ........................................................................12-12
12.2.7 Data-transfer Modes .............................................................................................12-13
12.2.7.1 Single Data-transfer Mode ...............................................................................12-14
12.2.7.2 Block Data-transfer Mode ................................................................................12-18
12.2.7.3 Demand Data-transfer Mode ............................................................................12-21
12.2.8 Cascade Mode ......................................................................................................12-25
12.2.9 DMA Interrupts ......................................................................................................12-26
12.2.10 8237A Compatibility ..............................................................................................12-27
12.3 REGISTER DEFINITIONS......................................................................................... 12-28
12.3.1 Pin Configuration Register (PINCFG) ...................................................................12-31
12.3.2 DMA Configuration Register (DMACFG) ..............................................................12-32
12.3.3 Channel Registers ................................................................................................12-33
12.3.4 Overflow Enable Register (DMAOVFE) ................................................................12-34
12.3.5 Command 1 Register (DMACMD1) .......................................................................12-35
12.3.6 Status Register (DMASTS) ...................................................................................12-36
12.3.7 Command 2 Register (DMACMD2) .......................................................................12-37
12.3.8 Mode 1 Register (DMAMOD1) ..............................................................................12-38
12.3.9 Mode 2 Register (DMAMOD2) ..............................................................................12-40
12.3.10 Software Request Register (DMASRR) ................................................................12-42
12.3.11 Channel Mask and Group Mask Registers (DMAMSK and DMAGRPMSK) .........12-44
12.3.12 Bus Size Register (DMABSR) ...............................................................................12-46
12.3.13 Chaining Register (DMACHR) ..............................................................................12-47
12.3.14 Interrupt Enable Register (DMAIEN) .....................................................................12-48
12.3.15 Interrupt Status Register (DMAIS) ........................................................................12-49
12.3.16 Software Commands ............................................................................................12-50
12.4 DESIGN CONSIDERATIONS.................................................................................... 12-50
12.5 PROGRAMMING CONSIDERATIONS...................................................................... 12-50
12.5.1 DMA Controller Code Examples ...........................................................................12-51
CHAPTER 13
SYNCHRONOUS SERIAL I/O UNIT
13.1 OVERVIEW ................................................................................................................. 13-1
13.1.1 SSIO Signals ...........................................................................................................13-4
13.2 SSIO OPERATION...................................................................................................... 13-5
13.2.1 Baud-rate Generator ...............................................................................................13-5
13.2.2 Transmitter ..............................................................................................................13-6
13.2.2.1 Transmit Mode using Enable Bit ........................................................................13-7
13.2.2.2 Autotransmit Mode ...........................................................................................13-12
13.2.2.3 Slave Mode ......................................................................................................13-12
ix
Intel386™ EX MICROPROCESSOR USER’S MANUAL
13.2.3 Receiver ................................................................................................................13-12
13.3 REGISTER DEFINITIONS......................................................................................... 13-16
13.3.1 Pin Configuration Register (PINCFG) ...................................................................13-17
13.3.2 SIO and SSIO Configuration Register (SIOCFG) .................................................13-18
13.3.3 Prescale Clock Register (CLKPRS) ......................................................................13-19
13.3.4 SSIO Baud-rate Control Register (SSIOBAUD) ....................................................13-20
13.3.5 SSIO Baud-rate Count Down Register (SSIOCTR) ..............................................13-21
13.3.6 SSIO Control 1 Register (SSIOCON1) ..................................................................13-21
13.3.7 SSIO Control 2 Register (SSIOCON2) ..................................................................13-23
13.3.8 SSIO Transmit Holding Buffer (SSIOTBUF) .........................................................13-24
13.3.9 SSIO Receive Holding Buffer (SSIORBUF) ..........................................................13-25
13.4 DESIGN CONSIDERATIONS.................................................................................... 13-25
13.5 PROGRAMMING CONSIDERATIONS...................................................................... 13-26
13.5.1 SSIO Example Code .............................................................................................13-26
CHAPTER 14
CHIP-SELECT UNIT
14.1 OVERVIEW ................................................................................................................. 14-1
14.2 CSU UPON RESET..................................................................................................... 14-2
14.3 CSU OPERATION....................................................................................................... 14-2
14.3.1 Defining a Channel’s Address Block .......................................................................14-2
14.3.2 System Management Mode Support ....................................................................14-10
14.3.3 Bus Cycle Length Control .....................................................................................14-11
14.3.4 Bus Size Control ...................................................................................................14-11
14.3.5 Overlapping Regions ............................................................................................14-11
14.4 REGISTER DEFINITIONS......................................................................................... 14-13
14.4.1 Pin Configuration Register (PINCFG) ...................................................................14-15
14.4.2 Port 2 Configuration Register (P2CFG) ................................................................14-16
14.4.3 Chip-select Address Registers ..............................................................................14-17
14.4.4 Chip-select Mask Registers ..................................................................................14-19
14.5 DESIGN CONSIDERATIONS.................................................................................... 14-21
14.6 PROGRAMMING CONSIDERATIONS...................................................................... 14-22
14.6.1 Chip-Select Unit Code Example ............................................................................14-22
CHAPTER 15
REFRESH CONTROL UNIT
15.1 DYNAMIC MEMORY CONTROL................................................................................. 15-1
15.1.1 Refresh Methods .....................................................................................................15-1
15.2 REFRESH CONTROL UNIT OVERVIEW ................................................................... 15-2
15.2.1 RCU Signals ...........................................................................................................15-4
15.2.2 Refresh Intervals .....................................................................................................15-4
x
CONTENTS
15.2.3 Refresh Addresses .................................................................................................15-4
15.2.4 Bus Arbitration ........................................................................................................15-5
15.3 RCU OPERATION....................................................................................................... 15-5
15.4 REGISTER DEFINITIONS........................................................................................... 15-6
15.4.1 Refresh Clock Interval Register (RFSCIR) ..............................................................15-7
15.4.2 Refresh Control Register (RFSCON) ......................................................................15-8
15.4.3 Refresh Base Address Register (RFSBAD) ............................................................15-9
15.4.4 Refresh Address Register (RFSADD) ...................................................................15-10
15.5 DESIGN CONSIDERATIONS.................................................................................... 15-11
15.6 PROGRAMMING CONSIDERATIONS...................................................................... 15-14
15.6.1 Refresh Control Unit Example Code .....................................................................15-14
CHAPTER 16
INPUT/OUTPUT PORTS
16.1 OVERVIEW ................................................................................................................. 16-1
16.1.1 Port Functionality ....................................................................................................16-2
16.2 REGISTER DEFINITIONS........................................................................................... 16-6
16.2.1 Pin Configuration ....................................................................................................16-7
16.2.2 Initialization Sequence ..........................................................................................16-10
16.3 DESIGN CONSIDERATIONS.................................................................................... 16-10
16.3.1 Pin Status During and After Reset ........................................................................16-10
16.4 PROGRAMMING CONSIDERATIONS...................................................................... 16-11
16.4.1 I/O Ports Code Example .......................................................................................16-11
CHAPTER 17
WATCHDOG TIMER UNIT
17.1 OVERVIEW ................................................................................................................. 17-1
17.1.1 WDT Signals ...........................................................................................................17-3
17.2 WATCHDOG TIMER UNIT OPERATION.................................................................... 17-3
17.2.1 Idle and Powerdown modes ....................................................................................17-4
17.2.2 General-purpose Timer Mode .................................................................................17-4
17.2.3 Software Watchdog Mode .......................................................................................17-5
17.2.4 Bus Monitor Mode ...................................................................................................17-5
17.3 DISABLING THE WDT ................................................................................................ 17-6
17.4 REGISTER DEFINITIONS........................................................................................... 17-7
17.5 DESIGN CONSIDERATIONS.................................................................................... 17-12
17.6 PROGRAMMING CONSIDERATIONS...................................................................... 17-12
17.6.1 Writing to the WDT Reload Registers (WDTRLDH and WDTRLDL) ....................17-12
17.6.2 Minimum Counter Reload Value ...........................................................................17-12
17.6.3 Watchdog Timer Unit Code Examples ..................................................................17-12
xi
Intel386™ EX MICROPROCESSOR USER’S MANUAL
CHAPTER 18
JTAG TEST-LOGIC UNIT
18.1 OVERVIEW ................................................................................................................. 18-1
18.2 TEST-LOGIC UNIT OPERATION................................................................................ 18-3
18.2.1 Test Access Port (TAP) ..........................................................................................18-3
18.2.2 Test Access Port (TAP) Controller ..........................................................................18-4
18.2.3 Instruction Register (IR) ..........................................................................................18-7
18.2.4 Data Registers ........................................................................................................18-8
18.3 TESTING ................................................................................................................... 18-10
18.3.1 Identifying the Device ............................................................................................18-10
18.3.2 Bypassing Devices on a Board .............................................................................18-10
18.3.3 Sampling Device Operation and Preloading Data .................................................18-10
18.3.4 Testing the Interconnections (EXTEST) ................................................................18-10
18.3.5 Disabling the Output Drivers .................................................................................18-11
18.4 TIMING INFORMATION ............................................................................................ 18-12
18.5 DESIGN CONSIDERATIONS.................................................................................... 18-14
APPENDIX A
SIGNAL DESCRIPTIONS
APPENDIX B
COMPATIBILITY WITH THE PC/AT* ARCHITECTURE
B.1
B.1.1
HARDWARE DEPARTURES FROM PC/AT SYSTEM ARCHITECTURE ................... B-1
DMA Unit ................................................................................................................ B-1
Industry Standard Bus (ISA) Signals ...................................................................... B-2
Interrupt Control Unit .............................................................................................. B-4
SIO Units ................................................................................................................ B-4
CPU-only Reset ...................................................................................................... B-4
HOLD, HLDA Pins .................................................................................................. B-4
Port B ...................................................................................................................... B-5
SOFTWARE CONSIDERATIONS FOR A PC/AT SYSTEM ARCHITECTURE............ B-5
Embedded Basic Input Output System (BIOS) ....................................................... B-5
Embedded Disk Operating System (DOS) .............................................................. B-5
Microsoft* Windows* ............................................................................................... B-5
B.1.2
B.1.3
B.1.4
B.1.5
B.1.6
B.1.7
B.2
B.2.1
B.2.2
B.2.3
APPENDIX C
EXAMPLE CODE HEADER FILES
C.1
C.2
REGISTER DEFINITIONS FOR CODE EXAMPLES ................................................... C-1
EXAMPLE CODE DEFINES......................................................................................... C-6
xii
CONTENTS
APPENDIX D
SYSTEM REGISTER QUICK REFERENCE
D.1
D.2
D.3
D.4
D.5
D.6
D.7
D.8
D.9
PERIPHERAL REGISTER ADDRESSES..................................................................... D-1
CLKPRS ....................................................................................................................... D-7
CSnADH (UCSADH)..................................................................................................... D-8
CSnADL (UCSADL)...................................................................................................... D-9
CSnMSKH (UCSMSKH) ............................................................................................. D-10
CSnMSKL (UCSMSKL) .............................................................................................. D-11
DLLn AND DLHn ........................................................................................................ D-12
DMABSR .................................................................................................................... D-13
DMACFG .................................................................................................................... D-14
D.10 DMACHR.................................................................................................................... D-15
D.11 DMACMD1.................................................................................................................. D-16
D.12 DMACMD2.................................................................................................................. D-17
D.13 DMAGRPMSK ............................................................................................................ D-18
D.14 DMAIEN...................................................................................................................... D-19
D.15 DMAIS ........................................................................................................................ D-20
D.16 DMAMOD1 ................................................................................................................. D-21
D.17 DMAMOD2 ................................................................................................................. D-22
D.18 DMAMSK.................................................................................................................... D-23
D.19 DMAnBYCn, DMAnREQn AND DMAnTARn.............................................................. D-24
D.20 DMAOVFE.................................................................................................................. D-25
D.21 DMASRR .................................................................................................................... D-26
D.22 DMASTS..................................................................................................................... D-27
D.23 ICW1 (MASTER AND SLAVE) ................................................................................... D-28
D.24 ICW2 (MASTER AND SLAVE) ................................................................................... D-29
D.25 ICW3 (MASTER)......................................................................................................... D-29
D.26 ICW3 (SLAVE)............................................................................................................ D-30
D.27 ICW4 (MASTER AND SLAVE) ................................................................................... D-30
D.28 IDCODE...................................................................................................................... D-31
D.29 IERn............................................................................................................................ D-32
D.30 IIRn ............................................................................................................................. D-33
D.31 INTCFG ...................................................................................................................... D-34
D.32 IR ................................................................................................................................ D-35
D.33 LCRn........................................................................................................................... D-36
D.34 LSRn........................................................................................................................... D-37
D.35 MCRn.......................................................................................................................... D-38
D.36 MSRn.......................................................................................................................... D-39
xiii
Intel386™ EX MICROPROCESSOR USER’S MANUAL
D.37 OCW1 (MASTER AND SLAVE).................................................................................. D-40
D.38 OCW2 (MASTER AND SLAVE).................................................................................. D-41
D.39 OCW3 (MASTER AND SLAVE).................................................................................. D-42
D.40 P1CFG........................................................................................................................ D-43
D.41 P2CFG........................................................................................................................ D-44
D.42 P3CFG........................................................................................................................ D-45
D.43 PINCFG ...................................................................................................................... D-46
D.44 PnDIR ......................................................................................................................... D-47
D.45 PnLTC......................................................................................................................... D-48
D.46 PnPIN ......................................................................................................................... D-48
D.47 POLL (MASTER AND SLAVE) ................................................................................... D-49
D.48 PORT92...................................................................................................................... D-50
D.49 PWRCON ................................................................................................................... D-51
D.50 RBRn .......................................................................................................................... D-52
D.51 REMAPCFG ............................................................................................................... D-53
D.52 RFSADD..................................................................................................................... D-54
D.53 RFSBAD ..................................................................................................................... D-54
D.54 RFSCIR ...................................................................................................................... D-55
D.55 RFSCON..................................................................................................................... D-55
D.56 SCRn .......................................................................................................................... D-56
D.57 SIOCFG...................................................................................................................... D-57
D.58 SSIOBAUD ................................................................................................................. D-58
D.59 SSIOCON1 ................................................................................................................. D-59
D.60 SSIOCON2 ................................................................................................................. D-60
D.61 SSIOCTR.................................................................................................................... D-61
D.62 SSIORBUF ................................................................................................................. D-61
D.63 SSIOTBUF.................................................................................................................. D-62
D.64 TBRn........................................................................................................................... D-62
D.65 TMRCFG .................................................................................................................... D-63
D.66 TMRCON.................................................................................................................... D-64
D.67 TMRn .......................................................................................................................... D-65
D.68 UCSADH..................................................................................................................... D-67
D.69 UCSADL ..................................................................................................................... D-67
D.70 UCSMSKH.................................................................................................................. D-67
D.71 UCSMSKL .................................................................................................................. D-67
D.72 WDTCNTH AND WDTCNTL....................................................................................... D-68
D.73 WDTRLDH AND WDTRLDL....................................................................................... D-69
D.74 WDTSTATUS.............................................................................................................. D-70
xiv
CONTENTS
APPENDIX E
INSTRUCTION SET SUMMARY
E.1
E.2
E.2.1
E.2.2
E.2.2.1
E.2.2.2
INSTRUCTION ENCODING AND CLOCK COUNT SUMMARY.................................. E-1
INSTRUCTION ENCODING....................................................................................... E-22
32-bit Extensions of the Instruction Set ................................................................ E-23
Encoding of Instruction Fields ............................................................................... E-24
Encoding of Operand Length (w) Field ......................................................... E-24
Encoding of the General Register (reg) Field ............................................... E-24
Encoding of the Segment Register (sreg) Field ............................................ E-25
Encoding of Address Mode .......................................................................... E-26
Encoding of Operation Direction (d) Field .................................................... E-30
Encoding of Sign-Extend (s) Field ................................................................ E-30
Encoding of Conditional Test (tttn) Field ...................................................... E-30
Encoding of Control or Debug or Test Register (eee) Field ......................... E-31
E.2.2.3
E.2.2.4
E.2.2.5
E.2.2.6
E.2.2.7
E.2.2.8
GLOSSARY
INDEX
xv
Intel386™ EX MICROPROCESSOR USER’S MANUAL
FIGURES
Figure
Page
Intel386™ EX Embedded Processor Block Diagram ...................................................2-2
2-1
3-1
3-2
4-1
4-2
4-3
4-4
4-5
4-6
4-7
4-8
5-1
5-2
5-3
5-4
5-5
5-6
Instruction Pipelining ....................................................................................................3-2
The Intel386™ CX Processor Internal Block Diagram .................................................3-3
PC/AT I/O Address Space (10-bit Decode)..................................................................4-3
Expanded I/O Address Space (16-bit Decode) ............................................................4-4
Address Configuration Register (REMAPCFG)............................................................4-7
Setting the ESE Bit Code Example ..............................................................................4-8
DOS-Compatible Mode ..............................................................................................4-10
Example of Nonintrusive DOS-Compatible Mode ......................................................4-12
Enhanced DOS Mode ................................................................................................4-13
NonDOS Mode...........................................................................................................4-14
Peripheral and Pin Connections...................................................................................5-2
Configuration of DMA, Bus Arbiter, and Refresh Unit ..................................................5-5
DMA Configuration Register (DMACFG)......................................................................5-6
Interrupt Control Unit Configuration..............................................................................5-9
Interrupt Configuration Register (INTCFG).................................................................5-10
Timer/Counter Unit Configuration...............................................................................5-12
Timer Configuration Register (TMRCFG)...................................................................5-13
Serial I/O Unit 0 Configuration....................................................................................5-15
Serial I/O Unit 1 Configuration....................................................................................5-16
SIO and SSIO Configuration Register (SIOCFG).......................................................5-17
SSIO Unit Configuration.............................................................................................5-18
Configuration of Chip-select Unit and Clock and Power Management Unit ...............5-20
Core Configuration .....................................................................................................5-21
Port 92 Configuration Register (PORT92)..................................................................5-22
Pin Configuration Register (PINCFG).........................................................................5-24
Port 1 Configuration Register (P1CFG)......................................................................5-25
Port 2 Configuration Register (P2CFG)......................................................................5-26
Port 3 Configuration Register (P3CFG)......................................................................5-27
Basic External Bus Cycles............................................................................................6-6
Simplified Bus State Diagram (Does Not Include Address Pipelining or Hold states)..6-8
Ready Logic ...............................................................................................................6-11
Basic Internal and External Bus Cycles......................................................................6-12
Nonpipelined Address Read Cycles...........................................................................6-15
Nonpipelined Address Write Cycles...........................................................................6-18
Complete Bus States (Including Pipelined Address)..................................................6-20
Pipelined Address Cycles...........................................................................................6-21
Interrupt Acknowledge Cycles....................................................................................6-25
Halt Cycle ...................................................................................................................6-27
Basic Refresh Cycle...................................................................................................6-29
Refresh Cycle During HOLD/HLDA............................................................................6-30
16-bit Cycles to 8-bit Devices (Using BS8#)...............................................................6-33
LOCK# Signal During Address Pipelining ..................................................................6-35
Intel386 EX Processor to Intel387 SX Math Coprocessor Interface...........................6-39
5-7
5-8
5-9
5-10
5-11
5-12
5-13
5-14
5-15
5-16
5-17
5-18
6-1
6-2
6-3
6-4
6-5
6-6
6-7
6-8
6-9
6-10
6-11
6-12
6-13
6-14
6-15
xvi
CONTENTS
Page
FIGURES
Figure
6-16
6-17
6-18
6-19
7-1
7-2
7-3
7-4
7-5
7-6
7-7
8-1
8-2
8-3
8-4
8-5
8-6
8-7
8-8
9-1
9-2
9-3
9-4
9-5
9-6
Intel386 EX Processor to SRAM/FLASH Interface.....................................................6-41
Intel386 EX Processor to PSRAM Interface...............................................................6-42
Intel386 EX Processor to Paged DRAM Interface......................................................6-43
Intel386 EX Processor and Non-Paged DRAM Interface...........................................6-44
Standard SMI# .............................................................................................................7-5
SMIACT# Latency .......................................................................................................7-6
SMI# During HALT ......................................................................................................7-8
SMI# During I/O Instruction..........................................................................................7-9
SMI# Timing ...............................................................................................................7-10
Interrupted SMI# Service............................................................................................7-11
HALT During SMM Handler........................................................................................7-12
Clock and Power Management Unit Connections........................................................8-2
Clock Synchronization..................................................................................................8-3
SMM Interaction with Idle and Powerdown Modes.......................................................8-5
Clock Prescale Register (CLKPRS) .............................................................................8-7
Power Control Register (PWRCON).............................................................................8-8
Timing Diagram, Entering and Leaving Idle Mode .......................................................8-9
Timing Diagram, Entering and Leaving Powerdown Mode ........................................8-11
Reset Synchronization Circuit ....................................................................................8-12
Interrupt Control Unit Configuration..............................................................................9-3
Methods for Changing the Default Interrupt Structure..................................................9-7
Interrupt Process – Master Request from Non-slave Source.....................................9-11
Interrupt Process – Slave Request.............................................................................9-12
Interrupt Process – Master Request from Slave Source ............................................9-13
Port 3 Configuration Register (P3CFG)......................................................................9-18
Interrupt Configuration Register (INTCFG).................................................................9-19
Initialization Command Word 1 Register (ICW1)........................................................9-20
Initialization Command Word 2 Register (ICW2)........................................................9-21
Initialization Command Word 3 Register (ICW3 – Master).........................................9-22
Initialization Command Word 3 Register (ICW3 – Slave)...........................................9-23
Initialization Command Word 4 Register (ICW4)........................................................9-24
Operation Command Word 1 (OCW1) .......................................................................9-25
Operation Command Word 2 (OCW2) .......................................................................9-26
Operation Command Word 3 (OCW3) .......................................................................9-27
Poll Status Byte (POLL) .............................................................................................9-28
Interrupt Acknowledge Cycle......................................................................................9-29
Spurious Interrupts.....................................................................................................9-30
Cascading External 82C59A Interrupt Controllers......................................................9-31
Timer/Counter Unit Signal Connections .....................................................................10-2
Mode 0 – Basic Operation..........................................................................................10-7
Mode 0 – Disabling the Count....................................................................................10-7
Mode 0 – Writing a New Count...................................................................................10-8
Mode 1 – Basic Operation..........................................................................................10-9
Mode 1 – Retriggering the One-shot..........................................................................10-9
9-7
9-8
9-9
9-10
9-11
9-12
9-13
9-14
9-15
9-16
9-17
9-18
9-19
10-1
10-2
10-3
10-4
10-5
10-6
xvii
Intel386™ EX MICROPROCESSOR USER’S MANUAL
FIGURES
Figure
Page
Mode 1 – Writing a New Count.................................................................................10-10
Mode 2 – Basic Operation........................................................................................10-11
Mode 2 – Disabling the Count..................................................................................10-11
10-7
10-8
10-9
10-10 Mode 2 – Writing a New Count.................................................................................10-12
10-11 Mode 3 – Basic Operation (Even Count)..................................................................10-13
10-12 Mode 3 – Basic Operation (Odd Count)...................................................................10-14
10-13 Mode 3 – Disabling the Count..................................................................................10-14
10-14 Mode 3 – Writing a New Count (With a Trigger).......................................................10-15
10-15 Mode 3 – Writing a New Count (Without a Trigger)..................................................10-15
10-16 Mode 4 – Basic Operation........................................................................................10-16
10-17 Mode 4 – Disabling the Count..................................................................................10-17
10-18 Mode 4 – Writing a New Count.................................................................................10-17
10-19 Mode 5 – Basic Operation........................................................................................10-18
10-20 Mode 5 – Retriggering the Strobe ............................................................................10-19
10-21 Mode 5 – Writing a New Count Value ......................................................................10-19
10-22 Timer Configuration Register (TMRCFG).................................................................10-21
10-23 Port 3 Configuration Register (P3CFG)....................................................................10-22
10-24 Pin Configuration Register (PINCFG).......................................................................10-23
10-25 Timer Control Register (TMRCON – Control Word Format).....................................10-25
10-26 Timer n Register (TMRn – Write Format).................................................................10-26
10-27 Timer Control Register (TMRCON – Counter-latch Format) ....................................10-28
10-28 Timer n Register (TMRn – Read Format).................................................................10-29
10-29 Timer Control Register (TMRCON – Read-back Format) ........................................10-30
10-30 Timer n Register (TMRn – Status Format) ...............................................................10-32
11-1
11-2
11-3
11-4
11-5
11-6
11-7
11-8
11-9
Serial I/O Unit 1 Configuration....................................................................................11-2
SIOn Baud-rate Generator Clock Sources.................................................................11-4
SIOn Transmitter ........................................................................................................11-7
SIOn Data Transmission Process Flow......................................................................11-8
SIOn Receiver............................................................................................................11-9
SIOn Data Reception Process Flow.........................................................................11-11
Pin Configuration Register (PINCFG).......................................................................11-17
Port 1 Configuration Register (P1CFG)....................................................................11-18
Port 2 Configuration Register (P2CFG)....................................................................11-19
11-10 Port 3 Configuration Register (P3CFG)....................................................................11-20
11-11 SIO and SSIO Configuration Register (SIOCFG).....................................................11-21
11-12 Divisor Latch Registers (DLLn and DLHn) ...............................................................11-22
11-13 Transmit Buffer Register (TBRn)..............................................................................11-23
11-14 Receive Buffer Register (RBRn)...............................................................................11-24
11-15 Serial Line Control Register (LCRn).........................................................................11-25
11-16 Serial Line Status Register (LSRn)...........................................................................11-26
11-17 Interrupt Enable Register (IERn) ..............................................................................11-27
11-18 Interrupt ID Register (IIRn) .......................................................................................11-28
11-19 Modem Control Signals – Diagnostic Mode Connections ........................................11-29
11-20 Modem Control Signals – Internal Connections.......................................................11-29
xviii
CONTENTS
Page
FIGURES
Figure
11-21 Modem Control Register (MCRn).............................................................................11-30
11-22 Modem Status Register (MSRn)...............................................................................11-31
11-23 Scratch Pad Register (SCRn)...................................................................................11-32
12-1
12-2
12-3
12-4
12-5
12-6
12-7
12-8
12-9
DMA Unit Block Diagram............................................................................................12-2
DMA Temporary Buffer Operation for a Read Transfer..............................................12-8
DMA Temporary Buffer Operation for A Write Transfer .............................................12-8
Start of a Two-cycle DMA Transfer Initiated by DRQn...............................................12-9
Changing the Priority of the DMA Channel and External Bus Requests..................12-10
Buffer Transfer Ended by an Expired Byte Count ....................................................12-11
Buffer Transfer Ended by the EOP# Input................................................................12-11
Single Data-transfer Mode with Single Buffer-transfer Mode ...................................12-15
Single Data-transfer Mode with Autoinitialize Buffer-transfer Mode.........................12-16
12-10 Single Data-transfer Mode with Chaining Buffer-transfer Mode...............................12-17
12-11 Block Data-transfer Mode with Single Buffer-transfer Mode ....................................12-19
12-12 Block Data-transfer Mode with Autoinitialize Buffer-transfer Mode..........................12-20
12-13 Buffer Transfer Suspended by the Deactivation of DRQn........................................12-21
12-14 Demand Data-transfer Mode with Single Buffer-transfer Mode................................12-22
12-15 Demand Data-transfer Mode with Autoinitialize Buffer-transfer Mode .....................12-23
12-16 Demand Data-transfer Mode with Chaining Buffer-transfer Mode ...........................12-24
12-17 Cascade Mode .........................................................................................................12-26
12-18 Pin Configuration Register (PINCFG).......................................................................12-31
12-19 DMA Configuration Register (DMACFG)..................................................................12-32
12-20 DMA Channel Address and Byte Count Registers
(DMAnREQn, DMAnTARn, DMAnBYCn).................................................................12-33
12-21 DMA Overflow Enable Register (DMAOVFE)...........................................................12-34
12-22 DMA Command 1 Register (DMACMD1).................................................................12-35
12-23 DMA Status Register (DMASTS)..............................................................................12-36
12-24 DMA Command 2 Register (DMACMD2).................................................................12-37
12-25 DMA Mode 1 Register (DMAMOD1) ........................................................................12-39
12-26 DMA Mode 2 Register (DMAMOD2) ........................................................................12-41
12-27 DMA Software Request Register (DMASRR – write format)....................................12-42
12-28 DMA Software Request Register (DMASRR – read format) ....................................12-43
12-29 DMA Channel Mask Register (DMAMSK)................................................................12-44
12-30 DMA Group Channel Mask Register (DMAGRPMSK) .............................................12-45
12-31 DMA Bus Size Register (DMABSR) .........................................................................12-46
12-32 DMA Chaining Register (DMACHR).........................................................................12-47
12-33 DMA Interrupt Enable Register (DMAIEN) ...............................................................12-48
12-34 DMA Interrupt Status Register (DMAIS)...................................................................12-49
13-1
13-2
13-3
13-4
13-5
13-6
Transmitter and Receiver in Master Mode .................................................................13-2
Transmitter in Master Mode, Receiver in Slave Mode................................................13-2
Transmitter in Slave Mode, Receiver in Master Mode................................................13-3
Transmitter and Receiver in Slave Mode ...................................................................13-3
Clock Sources for the Baud-rate Generator ...............................................................13-5
SSIO Transmitter with Autotransmit Mode Enabled...................................................13-7
xix
Intel386™ EX MICROPROCESSOR USER’S MANUAL
FIGURES
Figure
Page
SSIO Transmitter with Autotransmit Mode Disabled..................................................13-8
Transmit Data by Polling ............................................................................................13-9
Interrupt Service Routine for Transmitting Data Using Interrupts.............................13-10
13-7
13-8
13-9
13-10 Transmitter Master Mode, Single Word Transfer (Enabled when Clock is High) .....13-11
13-11 Transmitter Master Mode, Single Word Transfer (Enabled when Clock is Low)......13-11
13-12 Receive Data by Polling ...........................................................................................13-13
13-13 Interrupt Service Routine for Receiving Data Using Interrupts.................................13-14
13-14 Receiver Master Mode, Single Word Transfer .........................................................13-15
13-15 Pin Configuration Register (PINCFG).......................................................................13-17
13-16 SIO and SSIO Configuration Register (SIOCFG).....................................................13-18
13-17 Clock Prescale Register (CLKPRS) .........................................................................13-19
13-18 SSIO Baud-rate Control Register (SSIOBAUD) .......................................................13-20
13-19 SSIO Baud-rate Count Down Register (SSIOCTR)..................................................13-21
13-20 SSIO Control 1 Register (SSIOCON1).....................................................................13-22
13-21 SSIO Control 2 Register (SSIOCON2).....................................................................13-23
13-22 SSIO Transmit Holding Buffer (SSIOTBUF).............................................................13-24
13-23 SSIO Receive Holding Buffer (SSIORBUF) .............................................................13-25
14-1
14-2
14-3
14-4
14-5
14-6
14-7
14-8
14-9
15-1
15-2
15-3
15-4
15-5
15-6
15-7
15-8
16-1
16-2
16-3
16-4
16-5
16-6
17-1
17-2
17-3
Channel Address Comparison Logic..........................................................................14-3
Determining a Channel’s Address Block Size............................................................14-4
Bus Cycle Length Adjustments for Overlapping Regions.........................................14-12
Pin Configuration Register (PINCFG).......................................................................14-15
Port 2 Configuration Register (P2CFG)....................................................................14-16
Chip-select High Address Register (CSnADH, UCSADH) .......................................14-17
Chip-select Low Address Register (CSnADL, UCSADL) .........................................14-18
Chip-select High Mask Registers (CSnMSKH, UCSMSKH).....................................14-19
Chip-select Low Mask Registers (CSnMSKL, UCSMSKL).......................................14-20
Refresh Control Unit Connections..............................................................................15-3
Refresh Clock Interval Register (RFSCIR).................................................................15-7
Refresh Control Register (RFSCON) .........................................................................15-8
Refresh Base Address Register (RFSBAD) ...............................................................15-9
Refresh Address Register (RFSADD) ......................................................................15-10
Connections to Ensure Refresh of All Rows in an 8-Bit Wide PSRAM Device ........15-11
RAS# Only Refresh Logic: Paged Mode..................................................................15-13
RAS# Only Refresh Logic: Non-Paged Mode ..........................................................15-14
I/O Port Block Diagram...............................................................................................16-2
Logic Diagram of a Bi-directional Port........................................................................16-3
Port n Configuration Register (PnCFG)......................................................................16-7
Port Direction Register (PnDIR) .................................................................................16-8
Port Data Latch Register (PnLTC)..............................................................................16-8
Port Pin State Register (PnPIN) .................................................................................16-9
Watchdog Timer Unit Connections.............................................................................17-2
WDT Counter Value Registers (WDTCNTH and WDTCNTL)....................................17-8
WDT Status Register (WDTSTATUS)........................................................................17-9
xx
CONTENTS
Page
FIGURES
Figure
17-4
17-5
18-1
18-2
18-3
18-4
18-5
18-6
B-1
WDT Reload Value Registers (WDTRLDH and WDTRLDL)....................................17-10
Power Control Register (PWRCON).........................................................................17-11
Test Logic Unit Connections ......................................................................................18-2
TAP Controller (Finite-State Machine)........................................................................18-6
Instruction Register (IR)..............................................................................................18-7
Identification Code Register (IDCODE)......................................................................18-8
Internal and External Timing for Loading the Instruction Register............................18-12
Internal and External Timing for Loading a Data Register........................................18-13
Derivation of AEN Signal in a Typical PC/AT System ................................................. B-3
Derivation of AEN Signal for Intel386™ EX processor-based Systems...................... B-3
General Instruction Format........................................................................................ E-22
B-2
E-1
xxi
Intel386™ EX MICROPROCESSOR USER’S MANUAL
TABLES
Table
Page
PC-compatible Peripherals...........................................................................................2-3
2-1
2-2
4-1
4-2
5-1
5-2
5-3
5-4
5-5
5-6
5-7
5-8
5-9
5-10
5-11
5-12
5-13
6-1
6-2
6-3
7-1
7-2
7-3
8-1
8-2
9-1
Embedded Application-specific Peripherals .................................................................2-4
Peripheral Register I/O Address Map in Slot 15...........................................................4-5
Peripheral Register Addresses...................................................................................4-15
Master’s IR3 Connections ............................................................................................5-8
Master’s IR4 Connections ............................................................................................5-8
Signal Pairs on Pins without a Multiplexer..................................................................5-23
Example Pin Configuration Registers.........................................................................5-30
Example DMACFG Configuration Register................................................................5-31
Example TMRCFG Configuration Register ................................................................5-32
Example INTCFG Configuration Register ..................................................................5-33
Example SIOCFG Configuration Register..................................................................5-33
Pin Configuration Register Design Woksheet ............................................................5-34
DMACFG Register Design Worksheet .......................................................................5-35
TMRCFG Register Design Worksheet .......................................................................5-36
INTCFG Register Design Worksheet .........................................................................5-37
SIOCFG Register Design Worksheet.........................................................................5-37
Bus Interface Unit Signals............................................................................................6-3
Bus Status Definitions ..................................................................................................6-5
Sequence of Nonaligned Bus Transfers.....................................................................6-10
CR0 Bits Cleared Upon Entering SMM ........................................................................7-3
SMM Processor State Initialization Values...................................................................7-4
Relative Priority of Exceptions and Interrupts...............................................................7-7
Clock and Power Management Registers ....................................................................8-6
Clock and Power Management Signals........................................................................8-6
82C59A Master and Slave Interrupt Sources...............................................................9-5
ICU Registers.............................................................................................................9-16
TCU Signals ...............................................................................................................10-3
TCU Associated Registers .........................................................................................10-4
Operations Caused by GATEn...................................................................................10-6
GATEn Connection Options.....................................................................................10-20
Minimum and Maximum Initial Counts......................................................................10-26
Results of Multiple Read-back Commands Without Reads......................................10-33
SIO Signals ................................................................................................................11-3
Maximum and Minimum Output Bit Rates..................................................................11-5
Divisor Values for Common Bit Rates........................................................................11-5
Status Signal Priorities and Sources ........................................................................11-13
SIO Registers...........................................................................................................11-15
Access to Multiplexed Registers...............................................................................11-16
DMA Signals...............................................................................................................12-4
Operations Performed During Transfer ......................................................................12-6
DMA Registers .........................................................................................................12-28
DMA Software Commands.......................................................................................12-50
SSIO Signals ..............................................................................................................13-4
9-2
10-1
10-2
10-3
10-4
10-5
10-6
11-1
11-2
11-3
11-4
11-5
11-6
12-1
12-2
12-3
12-4
13-1
xxii
CONTENTS
Page
TABLES
Table
13-2
13-3
14-1
14-2
15-1
15-2
16-1
16-2
16-3
17-1
17-2
18-1
18-2
18-3
18-4
18-5
A-1
A-2
A-3
A-4
D-1
E-1
E-2
E-3
E-4
Maximum and Minimum Baud-rate Output Frequencies............................................13-6
SSIO Registers.........................................................................................................13-16
CSU Signals.............................................................................................................14-13
CSU Registers..........................................................................................................14-14
RCU Signals...............................................................................................................15-4
RCU Registers ...........................................................................................................15-6
Pin Multiplexing ..........................................................................................................16-5
I/O Port Registers.......................................................................................................16-6
Control Register Values for I/O Port Pin Configurations.............................................16-7
WDT Signals ..............................................................................................................17-3
WDT Registers ...........................................................................................................17-7
Test Access Port Dedicated Pins...............................................................................18-3
TAP Controller State Descriptions..............................................................................18-4
Example TAP Controller State Selections..................................................................18-5
Test-logic Unit Instructions.........................................................................................18-7
Boundary-scan Register Bit Assignments ..................................................................18-9
Signal Description Abbreviations................................................................................. A-1
Description of Signals Available at the Device Pins.................................................... A-2
Pin State Abbreviations............................................................................................... A-8
Pin States After Reset and During Idle, Powerdown, and Hold................................... A-9
Peripheral Register Addresses....................................................................................D-1
Instruction Set Summary............................................................................................. E-2
Fields Within Instructions........................................................................................... E-23
Encoding of Operand Length (w) Field...................................................................... E-24
Encoding of reg Field When w Field is not Present in Instruction ............................. E-24
Encoding of reg Field When w Field is Present in Instruction ................................... E-25
Encoding of the Segment Register (sreg) Field......................................................... E-25
Encoding of 16-bit Address Mode with “mod r/m” Byte ............................................. E-27
Encoding of 32-bit Address Mode with “mod r/m” Byte (No s-i-b Byte Present)........ E-28
Encoding of 32-bit Address Mode (“mod r/m” Byte and s-i-b Byte Present).............. E-29
Encoding of Operation Direction (d) Field ................................................................. E-30
Encoding of Sign-Extend (s) Field............................................................................. E-30
Encoding of Conditional Test (tttn) Field ................................................................... E-30
When Interpreted as Control Register Field .............................................................. E-31
When Interpreted as Debug Register Field............................................................... E-31
When Interpreted as Test Register Field................................................................... E-31
E-5
E-6
E-7
E-8
E-9
E-10
E-11
E-12
E-13
E-14
E-15
xxiii
1
GUIDE TO THIS
MANUAL
CHAPTER 1
GUIDE TO THIS MANUAL
designers familiar with the principles of microprocessors and with the Intel386 processor archi-
tecture.
This chapter is organized as follows:
• Related Documents (page 1-5)
• Electronic Support Systems (page 1-6)
• Technical Support (page 1-7)
• Product Literature (page 1-8)
1.1 MANUAL CONTENTS
This manual contains 18 chapters and 5 appendixes, a glossary, and an index. This section sum-
marizes the contents of the remaining chapters and appendixes. The remainder of this chapter de-
scribes notational conventions and special terminology used throughout the manual and provides
references to related documentation.
Chapter 2 — Architectural Overview — describes the device features and some potential ap-
plications.
™
Chapter 3 — Core Overview — describes the differences between this device and the Intel386
SX processor core.
Chapter 4 — System Register Organization — describes the organization of the system regis-
ters, the I/O address space, address decoding, and addressing modes.
Chapter 5 — Device Configuration — explains how to configure the device for various appli-
cations.
Chapter 6 — Bus Interface Unit — describes the bus interface logic, bus states, bus cycles, and
instruction pipelining.
Chapter 7 — System Management Mode — describes Intel’s System Management Mode
(SMM).
Chapter 8 — Clock and Power Management Unit — describes the clock generation circuitry,
power management modes, and system reset logic.
1-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Chapter 9 — Interrupt Control Unit — describes the interrupt sources and priority options and
explains how to program the interrupt control unit.
Chapter 10 — Timer/Counter Unit — describes the timer/counters and their available count
formats and operating modes.
Chapter 11 — Asynchronous Serial I/O (SIO) Unit — explains how to use the universal asyn-
chronous receiver/transmitters (UARTs) to transmit and receive serial data.
Chapter 12 — DMA Controller — describes how the enhanced direct memory access controller
allows internal and external devices to transfer data directly to and from the system and explains
how bus control is arbitrated.
Chapter 13 — Synchronous Serial I/O (SSIO) Unit — explains how to transmit and receive
data synchronously.
Chapter 14 — Chip-select Unit — explains how to use the chip-select channels to access vari-
ous external memory and I/O devices.
Chapter 15 — Refresh Control Unit — describes how the refresh control unit generates peri-
odic refresh requests and refresh addresses to simplify the interface to dynamic memory devices.
Chapter 16 — Input/Output Ports — describes the general-purpose I/O ports and explains how
to configure each pin to serve either as an I/O pin or as a pin controlled by an internal peripheral.
Chapter 17 — Watchdog Timer Unit — explains how to use the watchdog timer unit as a soft-
ware watchdog, bus monitor, or general-purpose timer.
Chapter 18 — JTAG Test-logic Unit — describes the independent test-logic unit and explains
how to test the device logic and board-level connections.
Appendix A — Signal Descriptions — describes the device pins and signals and lists pin states
after a system reset and during powerdown, idle, and hold.
Appendix B — Compatibility with PC/AT* Architecture — describes the ways in which the
device is compatible with the standard PC/AT architecture and the ways in which it departs from
the standard.
Appendix C — Example Code Header Files — contains the header files called by the code ex-
amples that are included in several chapters of this manual.
Appendix D — System Register Quick Reference — contains an alphabetical list of registers.
Appendix E — Instruction Set Summary — lists all instructions and their clock counts.
Glossary — defines terms with special meaning used throughout this manual.
Index — lists key topics with page number references.
1-2
GUIDE TO THIS MANUAL
1.2 NOTATIONAL CONVENTIONS
The following notations are used throughout this manual.
#
The pound symbol (#) appended to a signal name indicates that the signal
is active low.
Variables
New Terms
Instructions
Variables are shown in italics. Variables must be replaced with correct
values.
New terms are shown in italics. See the Glossary for a brief definition of
commonly used terms.
Instruction mnemonics are shown in upper case. When you are
programming, instructions are not case sensitive. You may use either
upper or lower case.
Numbers
Hexadecimal numbers are represented by a string of hexadecimal digits
followed by the character H. A zero prefix is added to numbers that begin
with A through F. (For example, FF is shown as 0FFH.) Decimal and
binary numbers are represented by their customary notations. (That is,
255 is a decimal number and 1111 1111 is a binary number. In some cases,
the letter B is added for clarity.)
Units of Measure
The following abbreviations are used to represent units of measure:
A
amps, amperes
gigabytes
Gbyte
Kbyte
KΩ
mA
Mbyte
MHz
ms
kilobytes
kilo-ohms
milliamps, milliamperes
megabytes
megahertz
milliseconds
milliwatts
mW
ns
nanoseconds
picofarads
pF
W
watts
V
volts
µA
µF
µs
µW
microamps, microamperes
microfarads
microseconds
microwatts
1-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Register Bits
When the text refers to more that one bit, the range may appear as two
numbers separated by a colon (example: 7:0 or 15:0). The first bit shown
(7 or 15 in the example) is the most-significant bit and the second bit
shown (0) is the least-significant bit.
Register Names
Signal Names
Register names are shown in upper case. If a register name contains a
lowercase, italic character, it represents more than one register. For
example, PnCFG represents three registers: P1CFG, P2CFG, and P3CFG.
Signal names are shown in upper case. When several signals share a
common name, an individual signal is represented by the signal name
followed by a number, while the group is represented by the signal name
followed by a variable (n). For example, the lower chip-select signals are
named CS0#, CS1#, CS2#, and so on; they are collectively called CSn#.
A pound symbol (#) appended to a signal name identifies an active-low
signal. Port pins are represented by the port abbreviation, a period, and
the pin number (e.g., P1.0, P1.1).
1.3 SPECIAL TERMINOLOGY
The following terms have special meanings in this manual.
Assert and Deassert
The terms assert and deassert refer to the act of making a signal
active and inactive, respectively. The active polarity (high/low) is
defined by the signal name. Active-low signals are designated by a
pound symbol (#) suffix; active-high signals have no suffix. To assert
RD# is to drive it low; to assert HOLD is to drive it high; to deassert
RD# is to drive it high; to deassert HOLD is to drive it low.
DOS I/O Address
Integrated peripherals that are compatible with PC/AT system
architecture can be mapped into DOS (or PC/AT) addresses 0H–
03FFH. In this manual, the terms DOS address and PC/AT address
are synonymous.
Expanded I/O Address All peripheral registers reside at I/O addresses 0F000H–0FFFFH.
PC/AT-compatible integrated peripherals can also be mapped into
DOS (or PC/AT) address space (0H–03FFH).
PC/AT Address
Integrated peripherals that are compatible with PC/AT system
architecture can be mapped into PC/AT (or DOS) addresses 0H–
03FFH. In this manual, the terms DOS address and PC/AT address
are synonymous.
Processor and CPU
Processor refers to the Intel386 EX processor including the
integrated peripherals. CPU refers to the processor core, which is
based on the static Intel386 SX processor.
1-4
GUIDE TO THIS MANUAL
Reserved Bits
Reserved bits are not used in this device, but they may be used in
future implementations. Follow these guidelines to ensure
compatibility with future devices:
• Avoid any software dependence on the state of undefined
register bits.
• Use a read-modify-write sequence to load registers.
• Mask undefined bits when testing the values of defined bits.
• Do not depend on the state of undefined bits when storing
undefined bits to memory or to another register.
• Do not depend on the ability to retain information written to
undefined bits.
Set and Clear
The terms set and clear refer to the value of a bit or the act of giving
it a value. If a bit is set, its value is “1”; setting a bit gives it a “1”
value. If a bit is clear, its value is “0”; clearing a bit gives it a “0”
value.
1.4 RELATED DOCUMENTS
The following documents contain additional information that is useful in designing systems that
incorporate the Intel386 EX processor. To order documents, please call Intel Literature Fulfill-
ment (1-800-548-4725 in the U.S. and Canada; +44(0) 1793-431155 in Europe).
Document Name
Order Number
Intel386™ EX Embedded Microprocessor datasheet
Intel386™ SX Microprocessor datasheet
272420
240187
240331
240332
272326
272520
272587
240800
Intel386™ SX Microprocessor Programmer’s Reference Manual
Intel386™ SX Microprocessor Hardware Reference Manual
Development Tools
Buyer’s Guide for the Intel386™ Embedded Processor Family
Intel386™ EX Microprocessor Pin Multiplexing Map
Packaging
You may also want to refer to Standard 1149.1—1990, IEEE Standard Test Access Port and
Boundary-Scan Architecture and its supplement, Standard 1149.1a—1993.
1-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
1.5 ELECTRONIC SUPPORT SYSTEMS
Intel’s FaxBack* service and application BBS provide up-to-date technical information. Intel also
maintains several forums on CompuServe and offers a variety of information on the World Wide
Web. These systems are available 24 hours a day, 7 days a week, providing technical information
whenever you need it.
1.5.1 FaxBack Service
FaxBack is an on-demand publishing system that sends documents to your fax machine. You can
get product announcements, change notifications, product literature, device characteristics, de-
sign recommendations, and quality and reliability information from FaxBack 24 hours a day, 7
days a week.
1-800-525-3019 (US or Canada)
+44-1793-432509 (Europe)
+65-256-5350 (Singapore)
+852-2-844-4448 (Hong Kong)
+886-2-514-0815 (Taiwan)
+822-767-2594 (Korea)
+61-2-975-3922 (Australia)
1-503-264-6835 (Worldwide)
Think of the FaxBack service as a library of technical documents that you can access with your
phone. Just dial the telephone number and respond to the system prompts. After you select a doc-
ument, the system sends a copy to your fax machine.
Each document has an order number and is listed in a subject catalog. The first time you use Fax-
Back, you should order the appropriate subject catalogs to get a complete list of document order
numbers. Catalogs are updated twice monthly. In addition, daily update catalogs list the title, sta-
tus, and order number of each document that has been added, revised, or deleted during the past
eight weeks. To receive the update for a subject catalog, enter the subject catalog number fol-
lowed by a zero. For example, for the complete microcontroller and flash catalog, request docu-
ment number 2; for the daily update to the microcontroller and flash catalog, request document
number 20.
The following catalogs and information are available at the time of publication:
1. Solutions OEM subscription form
2. Microcontroller and flash catalog
3. Development tools catalog
4. Systems catalog
5. Multimedia catalog
®
6. Multibus and iRMX software catalog and BBS file listings
1-6
GUIDE TO THIS MANUAL
7. Microprocessor, PCI, and peripheral catalog
8. Quality and reliability and change notification catalog
9. iAL (Intel Architecture Labs) technology catalog
1.5.2 Bulletin Board System (BBS)
The bulletin board system (BBS) lets you download files to your computer. The application BBS
has the latest ApBUILDER software, hypertext manuals and datasheets, software drivers, firm-
ware upgrades, code examples, application notes and utilities, and quality and reliability data.
The system supports 1200- through 19200-baud modems. Typical modem settings are 14400
baud, no parity, 8 data bits, and 1 stop bit (14400, N, 8, 1).
To access the BBS, use a terminal program to dial the telephone number given below for your
area; once you are connected, respond to the system prompts. During your first session, enter your
name and location. The system operator will set up your access account within 24 hours. At that
time, you can access the files on the BBS.
503-264-7999
U.S., Canada, Japan, Asia Pacific (up to 19.2 Kbaud)
Europe
44(0)1793-432955
NOTE
If you have problems accessing the BBS, use these settings for your modem:
2400, N, 8, 1. Refer to your terminal software documentation for instructions
on changing these settings.
1.5.3 CompuServe Forums
The CompuServe forums provide a means for you to gather information, share discoveries, and
debate issues. Type “go intel” for access. For information about CompuServe access and service
fees, call CompuServe at 1-800-848-8199 (U.S.) or 614-529-1340 (outside the U.S.).
1.5.4 World Wide Web
We offer a variety of information through the World Wide Web (http://www.intel.com/). Select
“Embedded Design Products” from the Intel home page.
1.6 TECHNICAL SUPPORT
In the U.S. and Canada, technical support representatives are available to answer your questions
between 5 a.m. and 5 p.m. PST. You can also fax your questions to us. (Please include your voice
telephone number and indicate whether you prefer a response by phone or by fax). Outside the
U.S. and Canada, please contact your local distributor.
1-800-628-8686
916-356-7599
U.S. and Canada
U.S. and Canada
U.S. and Canada
916-356-6100 (fax)
1-7
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
1.7 PRODUCT LITERATURE
You can order product literature from the following Intel literature centers.
1-800-548-4725
708-296-9333
U.S. and Canada
U.S. (from overseas)
Europe (U.K.)
Germany
44(0)1793-431155
44(0)1793-421333
44(0)1793-421777
81(0)120-47-88-32
France
Japan (fax only)
1-8
2
ARCHITECTURAL
OVERVIEW
CHAPTER 2
ARCHITECTURAL OVERVIEW
™
The Intel386 EX embedded processor (Figure 2-1) is based on the static Intel386 SX processor.
This highly integrated device retains those personal computer functions that are useful in embed-
ded applications and integrates peripherals that are typically needed in embedded systems. The
Intel386 EX processor provides a PC-compatible development platform in a device that is opti-
mized for embedded applications. Its integrated peripherals and power management options
make the Intel386 EX processor ideal for portable systems.
The integrated peripherals of the Intel386 EX processor are compatible with the standard desktop
PC. This allows existing PC software, including most of the industry’s leading desktop and em-
bedded operating systems, to be easily implemented on an Intel386 EX processor-based platform.
Using PC-compatible peripherals also allows for the development and debugging of application
software on a standard PC platform.
Typical applications using the Intel386 EX processor include automated manufacturing equip-
high-precision industrial flow controllers, interactive television, medical equipment, modems,
and smart copiers.
This chapter is organized as follows:
• Intel386 EX Embedded Processor Core (see below)
• Integrated Peripherals (page 2-3)
2.1 Intel386 EX EMBEDDED PROCESSOR CORE
The Intel386 EX processor contains a modular, fully static Intel386 CX central processing unit
(CPU). The Intel386 CX processor is an enhanced Intel386 SX processor with the addition of
System Management Mode (SMM) and two additional address lines. The Intel386 EX processor
has a 16-bit data bus and a 26-bit address bus, supporting up to 64 Mbytes of memory address
space and 64 Kbytes of I/O address space. The performance of the Intel386 EX processor closely
reflects the Intel386 SX CPU performance at the same speeds.
Chapter 3, “CORE OVERVIEW” describes differences between the Intel386 EX processor core
and the Intel386 SX processor. Please refer to the Intel386™ SX Microprocessor Programmer’s
Reference Manual (order number 240331) for applications and system programming informa-
tion; descriptions of protected, real, and virtual-8086 modes; and details on the instruction set.
2-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Bus Interface
Unit
Chip-select
Unit
JTAG Unit
Address
Data
Clock and Power
Management Unit
™
Intel386 CX Core
Core Enhancements
- A20 Gate
- CPU Reset
- SMM
DRAM Refresh
Control Unit
Watchdog Timer Unit
Bus Monitor
Asynchronous Serial I/O
2 channels
(16450 compatible)
Synchronous Serial I/O
1 channel, full duplex
Timer/counter Unit
3 channels
(82C54 compatible)
I/O Ports
INTR
Interrupt Control Unit
DMA Controller
2 channels
(8237A compatible)
and Bus Arbiter Unit
A2849-02
Figure 2-1. Intel386™ EX Embedded Processor Block Diagram
2-2
ARCHITECTURAL OVERVIEW
2.2 INTEGRATED PERIPHERALS
The Intel386 EX processor integrates both PC-compatible peripherals (Table 2-1) and peripherals
that are specific to embedded applications (Table 2-2).
Table 2-1. PC-compatible Peripherals
Name
Interrupt
Control Unit
(ICU)
Description
Consists of two 82C59A programmable interrupt controllers (PICs) configured as master
and slave. You may cascade up to six external 82C59A PICs to expand the external
interrupt lines to 52. Refer to Chapter 9, “INTERRUPT CONTROL UNIT.”
Timer/counter Provides three independent 16-bit down counters. The programmable TCU is
Unit (TCU)
functionally equivalent to three 82C54 counter/timers with enhancements to allow
remapping of peripheral addresses and interrupt assignments. Refer to Chapter 10,
“TIMER/COUNTER UNIT.”
Asynchronous Features two independent universal asynchronous receiver and transmitter (UART)
Serial I/O
(SIO) Unit
units which are functionally equivalent to National Semiconductor’s NS16450. Each
channel contains a baud-rate generator, transmitter, receiver, and modem control unit.
Receive and transmit interrupt signals can be connected to the ICU controller and DMA
controller. Refer to Chapter 11, “ASYNCHRONOUS SERIAL I/O UNIT.”
DirectMemory Transfers internal or external data between any combination of memory and I/O devices
Access
(DMA)
Controller
for the entire 26-bit address bus. The two independent channels operate in 16- or 8-bit
bus mode. Buffer chaining allows data to be transferred into noncontiguous memory
buffers. The DMA channels can be tied to any of the serial devices to support high data
rates, minimizing processor interruptions. Provides a special two-cycle mode that uses
only one channel for memory-to-memory transfers. Bus arbitration logic resolves priority
conflicts between the DMA channels, the refresh control unit, and an external bus
master. SIO and SSIO interrupts can be connected to DMA for high-speed transfers.
Backward compatible with 8237A. Refer to Chapter 12, “DMA CONTROLLER.”
2-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table 2-2. Embedded Application-specific Peripherals
Name
System
Management
Mode (SMM)
Description
The Intel386 EX processor provides a mechanism for system management with a
combination of hardware and CPU microcode enhancements. An externally generated
system management interrupt (SMI#) allows the execution of system-wide routines that are
independent and transparent to the operating system. The system management mode
(SMM) architectural extensions to the Intel386 CPU are described in Chapter 7, “SYSTEM
MANAGEMENT MODE.”
Clock and
Power
Management
Unit
An external clock source provides the input frequency. The clock and power management
unit generates separate internal clock signals for core and peripherals (half the input
frequency), divides the internal clock by two for baud clock inputs to the SIO and SSIO, and
divides the internal clock by a programmable divisor to provide a prescaled clock signal
(various frequencies) for the TCU and SSIO.
Power management provides idle and powerdown modes (idle stops the CPU clock but
leaves the peripheral clocks running; powerdown stops both CPU and peripheral clocks).
An external clockout signal is also provided. Refer to Chapter 8, “CLOCK AND POWER
MANAGEMENT UNIT.”
Synchronous
Serial I/O
(SSIO) unit
Provides simultaneous, bidirectional high speed serial I/O. Consists of a transmit channel, a
receive channel, and a baud rate generator. Built-in protocols are not included, because
these can be emulated using the CPU. SSIO interrupts can be connected to the DMA unit
for high-speed transfers. Refer to Chapter 13, “SYNCHRONOUS SERIAL I/O UNIT.”
Chip-select
Unit (CSU)
Programmable, eight-channel CSU allows direct access to up to eight devices. Each
channel can operate in 16- or 8-bit bus mode and can generate up to 31 wait states. The
CSU can interface with the fastest memory or the slowest peripheral device. The minimum
address block for memory address-configured channels is 2 Kbytes. The size of these
address blocks can be increased by powers of 2 Kbytes for memory addresses and by
multiples of 2 bytes for I/O addresses. Supports SMM memory addressing and provides
ready generation and programmable wait states. Refer to Chapter 14, “CHIP-SELECT
UNIT.”
Refresh
Control Unit
(RCU)
Provides a means to generate periodic refresh requests and refresh addresses. Consists of
a programmable interval timer unit, a control unit, and an address generation unit. Bus
arbitration logic ensures that refresh requests have the highest priority. The refresh control
unit (RCU) is provided for applications that use DRAMs with a simple EPLD-based DRAM
controller or PSRAMs that do not need a separate controller. Refer to Chapter 15,
“REFRESH CONTROL UNIT.”
Parallel I/O
Ports
Three I/O ports facilitate data transfer between the processor and surrounding system
circuitry. The Intel386 EX processor is unique in that several functions are multiplexed with
each other or with I/O ports. This ensures maximum use of available pins and maintains a
small package. Each multiplexed pin is individually programmable for peripheral or I/O
function. Refer to Chapter 16, “INPUT/OUTPUT PORTS.”
Watchdog
Timer (WDT)
Unit
When enabled, the WDT functions as a general purpose 32-bit timer, a software timer, or a
bus monitor. Refer to Chapter 17, “WATCHDOG TIMER UNIT.”
JTAG Test-
logic Unit
The test-logic unit simplifies board-level testing. Consists of a test access port and a
boundary-scan register. Fully compliant with Standard 1149.1–1990, IEEE Standard Test
Access Port and Boundary-Scan Architecture and its supplement, Standard 1149.1a–1993.
Refer to Chapter 18, “JTAG TEST-LOGIC UNIT.”
2-4
3
CORE OVERVIEW
CHAPTER 3
CORE OVERVIEW
The Intel386™ EX processor core is based upon the Intel386 CX processor, which is an enhanced
ments over the Intel386 SX processor, internal architecture of the Intel386 CX processor, and the
core interface on the Intel386 EX processor.
This chapter is organized as follows:
• Intel386 CX Processor Enhancements (see below)
• Intel386 CX Processor Internal Architecture (page 3-2)
• Core Intel386 EX Processor Interface (page 3-6)
3.1 Intel386 CX PROCESSOR ENHANCEMENTS
The Intel386 CX processor, based on the Intel386 SX processor, adds system management mode
and two additional address lines for a total of 26 address lines.
3.1.1 System Management Mode
The Intel386 CX processor core provides a mechanism for system management with a combina-
tion of hardware and CPU microcode enhancements. An externally generated System Manage-
ment Interrupt (SMI#) allows the execution of system wide routines which are independent and
transparent to the operating system. The System Management Mode (SMM) architecture exten-
sions to the Intel386 SX processor consist of the following elements:
• Interrupt input pin (SMI#) to invoke SMM
• One output pin to identify execution state (SMIACT#)
• One new instruction (RSM, executable only from SMM) to exit SMM
• SMM also added one to four execution clocks to the following instructions: IN, INS, REP
INS, OUT, REP OUT, POPA, HALT, MOV CR0, and SRC. INTR and NMI also need an
additional two clocks for interrupt latency. These cycles were added due to the microcode
modification for the SMM implementation. Refer to Appendix E for the exact execution
times. Otherwise, 100% of the Intel386 SX processor instructions execute on the Intel386
CX processor core.
Please refer to Chapter 7 for more details on System Management Mode.
3.1.2 Additional Address Lines
Two additional address lines were added to the Intel386 CX processor core for a total of 26. This
expands the physical address space from 16 Mbytes to 64 Mbytes.
3-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
3.2 Intel386 CX PROCESSOR INTERNAL ARCHITECTURE
The internal architecture of the Intel386 CX processor consists of functional units that operate in
parallel. Fetching, decoding, execution, memory management and bus accesses for several in-
structions are performed simultaneously. This parallel operation is called pipelined instruction
processing. With pipelining, each instruction is performed in stages, and the processing of several
instructions at different stages may overlap, as shown in Figure 3-1. The pipelined processing of
the Intel386 CX processor results in higher performance and enhanced throughput rate over non-
pipelined processors.
Elapsed Time
Typical
Processor
Fetch 1
Decode 1
Execute 1
Fetch 2
Decode 2
Execute 2
Intel386™ SX CPU/Intel376™ CPU
Store
Result 1
Bus Unit
Fetch 1
Fetch 2
Fetch 3
Fetch 5
Fetch 6
Fetch 4
Decode
Unit
Decode 1
Decode 2
Execute 1
Decode 3
Decode 4
Decode 5
Execution
Unit
Execute 2
Execute 3
Execute 4
Addr &
MMU
Addr &
MMU
MMU
A2850-01
Figure 3-1. Instruction Pipelining
3-2
CORE OVERVIEW
Figure 3-2 shows the internal architecture of the Intel386 CX processor.
Core Plus
Unit
Segmentation Unit
Paging Unit
Adder
HOLD, INTR, NMI,
ERROR#,BUSY#,
Request
Prioritizer
3-Input
Adder
RESET, HLDA,
SMI#, SMIACT#,
PEREQ
Effective Address Bus
Effective Address Bus
32
32
32
Descriptor
Register
Page
Cache
Limit and
Attribute
PLA
Control and
Attribute
PLA
32
32
BE0#, BE1#,
A25:1
Address
Driver
Protection
Test
M/IO#, D/C#,
W/R#, LOCK#,
ADS#, NA#,
READY#
Unit
Internal Control Bus
Pipeline/
Bus Size
Control
D15:0
MUX/
Transceivers
Barrel
Shifter,
Adder
Instruction
Decoder
Prefetcher
Limit
Checker
32
32
Decode
and
Sequencing
Status
Flags
Multiply/
Divide
Code
Stream
3 Decoded
Instruction
Queue
16 Byte
Code
Queue
Control
ROM
32
Register
File
ALU
Control
ALU
Control
Instruction
Predecode
Instruction
Prefetch
Dedicated ALU Bus
32
A2851-02
Figure 3-2. The Intel386™ CX Processor Internal Block Diagram
3-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
The six functional units of the Intel386 CX processor are:
• Core Bus Unit
• Instruction Prefetch Unit
• Instruction Decode Unit
• Execution Unit
• Segmentation Unit
• Paging Unit
3.2.1 Core Bus Unit
The Core Bus Unit provides the interface between the processor and its environment. It accepts
internal requests for instruction fetches (from the Instruction Prefetch Unit) and data transfers
(from the Execution Unit), and prioritizes the requests. At the same time, it generates or processes
the signals to perform the current bus cycle. These signals include the address, data, and control
outputs for accessing external memory and I/O. The Core Bus Unit also controls the interface to
external bus masters and coprocessors.
3.2.2 Instruction Prefetch Unit
The Instruction Prefetch Unit performs the program look ahead function of the CPU. When the
Core Bus Unit is not performing bus cycles to execute an instruction, the Instruction Prefetch Unit
uses the Core Bus Unit to fetch sequentially along the instruction byte stream. These prefetched
instructions are stored in the Instruction Queue to await processing by the Instruction Decode
Unit.
Instruction prefetches are given a lower priority than data transfers; assuming zero wait state
memory access, prefetch activity never delays execution. On the other hand, when there is no data
transfer requested, prefetching uses bus cycles that would otherwise be idle.
3.2.3 Instruction Decode Unit
The Instruction Decode Unit takes instruction stream bytes from the Prefetch Queue and trans-
lates them into microcode. The decoded instructions are then stored in a three-deep Instruction
Queue (FIFO) to await processing by the Execution Unit. Immediate data and opcode offsets are
also taken from the Prefetch Queue. The decode unit works in parallel with the other units and
begins decoding when there is a free slot in the FIFO and there are bytes in the prefetch queue.
Opcodes can be decoded at a rate of one byte per clock. Immediate data and offsets can be decod-
ed in one clock regardless of their length.
3-4
CORE OVERVIEW
3.2.4 Execution Unit
The Execution Unit executes the instructions from the Instruction Queue and therefore commu-
nicates with all other units required to complete the instruction. The functions of its three subunits
are given below.
• The Control Unit contains microcode and special parallel hardware that speeds multiply,
divide, and effective address calculation.
• The Data Unit contains the (Arithmetic Logic Unit) ALU, a file of eight 32-bit general-
purpose registers, and a 64-bit barrel shifter (which performs multiple bit shifts in one
clock). The Data Unit performs data operations requested by the Control Unit.
• The Protection Test Unit checks for segmentation violations under the control of the
microcode.
To speed the execution of memory reference instructions, the Execution Unit partially overlaps
the execution of any memory reference instruction with the previous instruction.
3.2.5 Segmentation Unit
The Segmentation Unit translates logical addresses into linear addresses at the request of the Ex-
ecution Unit. The on-chip Segment Descriptor Cache stores the currently used segment descrip-
tors to speed this translation. At the same time it performs the translation, the Segmentation Unit
checks for bus-cycle segmentation violations. (These checks are separate from the static segmen-
tation violation checks performed by the Protection Test Unit.) The translated linear address is
truncated to a 24-bit physical address.
3.2.6 Paging Unit
When the Intel386 CX processor paging mechanism is enabled, the Paging Unit translates linear
addresses generated by the Segmentation Unit or the Instruction Prefetch Unit into physical ad-
dresses. (When paging is not enabled, the physical address is the same as the linear address, and
no translation is necessary.) The Page Descriptor Cache stores recently used Page Directory and
Page Table entries in its Translation Lookaside Buffer (TLB) to speed this translation. The Paging
Unit forwards physical addresses to the Core Bus Unit to perform memory and I/O accesses.
3-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
3.3 CORE Intel386 EX PROCESSOR INTERFACE
The Intel386 EX processor peripherals are connected to the Intel386 CX processor core through
an internal Bus Interface Unit (BIU). The BIU controls internal peripheral accesses and external
memory and I/O accesses. Because it has the BIU between the Intel386 CX processor core and
the external bus, the Intel386 EX processor bus timings are not identical to those of the Intel386
CX processor or Intel386 SX processor.
The Intel386 CX processor numeric coprocessor interface is maintained and brought out to the
Intel386 EX processor pins. The same I/O addresses used on the Intel386 SX processor are used
on the Intel386 EX processor, even though there are more address lines. The A23 line is high for
coprocessor cycles. Refer to “Interface To Intel387™ SX Math Coprocessor” on page 6-38 for
more details.
3-6
4
SYSTEM
REGISTER
ORGANIZATION
CHAPTER 4
SYSTEM REGISTER ORGANIZATION
cessor, focusing on register organization from an address architecture viewpoint. The chapters
This chapter is organized as follows:
• Organization of Peripheral Registers (page 4-5)
• I/O Address Decoding Techniques (page 4-6)
• Addressing Modes (page 4-9)
• Peripheral Register Addresses (page 4-15)
4.1 OVERVIEW
The Intel386 EX processor has register resources in the following categories:
• Intel386 processor core architecture registers:
— General purpose registers
— Segment registers
— Instruction pointer and flags
— Control registers
— System address registers (protected mode)
— Debug registers
— Test registers
• Intel386 EX processor peripheral registers:
— Configuration space control registers
— Interrupt control unit registers
— Timer/counter unit registers
— DMA unit registers (8237A-compatible and enhanced function registers)
— Asynchronous serial I/O (SIO) registers
— Clock generation selector registers
4-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
— Power management control registers
— Chip-select unit control registers
— Refresh control unit registers
— Watchdog timer control registers
— Synchronous serial I/O control registers
— Parallel I/O port control registers
4.1.1 Intel386 Processor Core Architecture Registers
These registers are a superset of the 8086 and 80286 processor registers. All 16-bit 8086 and
80286 registers are contained within the 32-bit Intel386 processor core registers. A detailed de-
scription of the Intel386 processor architecture base registers can be found in the Intel386™ SX
Microprocessor Programmer’s Reference Manual (order number 240331).
4.1.2 Intel386 EX Processor Peripheral Registers
The Intel386 EX processor contains some peripherals that are common and compatible with the
*
PC/AT system architecture and others that are useful for embedded applications. The peripheral
registers control access to these peripherals and enable you to configure on-chip system resources
such as timer/counters, power management, chip selects, and watchdog timer.
All peripheral registers reside physically in the expanded I/O address space (addresses 0F000H–
0FFFFH). Peripherals that are compatible with PC/AT system architecture can also be mapped
into DOS I/O address space (addresses 0H–03FFH, 10-bit decode). The following rules apply for
accessing peripheral registers after a system reset:
• Registers within the DOS I/O address space are accessible.
• Registers within the expanded I/O address space are accessible only after the expanded I/O
address space is enabled.
4.2 I/O ADDRESS SPACE FOR PC/AT SYSTEMS
The Intel386 EX processor’s I/O address space is 64 Kbytes. On PC/AT platforms, the DOS op-
is used. The first 256 bytes (addresses 00000H–00FFH) are reserved for platform (motherboard)
I/O resources such as the interrupt and DMA controllers, and the remaining 768 bytes (addresses
0100H–03FFH) are available for “general” I/O peripheral card resources. Since only 1 Kbyte of
the address space is supported, add-on I/O peripheral cards typically decode only the lower 10
address lines. Because the upper address lines are not decoded, the 256 platform address locations
and the 768 bus address locations are repeated 64 times (on 1-Kbyte boundaries), covering the
entire 64-Kbyte address space. (See Figure 4-1.)
Generally, add-on I/O peripheral cards do not use the I/O addresses reserved for the platform re-
sources. Software running on the platform can use any of the 64 repetitions of the 256 address
locations reserved for accessing platform resources.
4-2
SYSTEM REGISTER ORGANIZATION
FFFFH (64K)
General Slot I/O
FD00H
Platform I/O (Reserved)
FC00H (63K)
0C00H (3K)
General Slot I/O
0900H
Platform I/O (Reserved)
General Slot I/O
0800H (2K)
0500H
Platform I/O (Reserved)
General Slot I/O
0400H (1K)
0100H (256)
Platform I/O (Reserved)
0000H (0)
A2498-01
Figure 4-1. PC/AT I/O Address Space (10-bit Decode)
4.3 EXPANDED I/O ADDRESS SPACE
The Intel386 EX processor’s I/O address scheme is similar to that of the Extended Industry Stan-
dard Architecture (EISA) bus and the Enhanced - Industry Standard Architecture (E-ISA) bus.
Both standards maintain backward software compatibility with the ISA architecture. The ISA
Platform I/O (0-100H) is accessed with a 16-bit address decode and is located in the first 256 I/O
locations. The General Slot I/O that is typically used by add-in boards is repeated throughout the
64 Kbyte I/O address range due to their 10-bit only decode. This allows 63 of the 64 repetitions
of the first 256 address locations of every 1 Kbyte block to be allocated to specific slots. Each slot
is 4 Kbyte in size, allowing for a total of 16 slots. The partitioning is such that four groups of 256
address locations are assigned to each slot, for a total of 1024 specific address locations per slot.
4-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
(See Figure 4-2.) Thus, each slot has 1 Kbyte addresses (in four 256-byte segments) that can po-
tentially contain extended peripheral registers.
FFFFH (64K)
General Slot I/O
Slot 15
FC00H (63K)
General Slot I/O
Slot 15
F800H (62K)
General Slot I/O
Slot 15
F400H (61K)
General Slot I/O
Slot 15
F000H (60K)
1FFFH (8K)
General Slot I/O
Slot 1
1C00H (7K)
General Slot I/O
Slot 1
1800H (6K)
General Slot I/O
Slot 1
1400H (5K)
General Slot I/O
Slot 1
1000H (4K)
General Slot I/O
Slot 0
0C00H (3K)
General Slot I/O
Slot 0
0800H (2K)
General Slot I/O
Slot 0
0400H (1K)
General Slot I/O
Slot 0
ISA Platform I/O
0000H (0K)
A2499-02
Figure 4-2. Expanded I/O Address Space (16-bit Decode)
4-4
SYSTEM REGISTER ORGANIZATION
The Intel386 EX processor uses slot 15 for the registers needed for integrated peripherals. Using
not use slot 15.
4.4 ORGANIZATION OF PERIPHERAL REGISTERS
The registers associated with the integrated peripherals are physically located in slot 15 of the I/O
space. There are sixteen 4 Kbyte address slots in I/O space. Slot 0 refers to 0H–0FFFH; slot 15
refers to 0F000H–0FFFFH. Table 4-1 shows the address map for the peripheral registers in slot
15. Note that the I/O addresses fall in address ranges 0F000H–0F0FFH, 0F400H–0F4FFH, and
0F800H–0F8FFH; utilizing the unique sets of 256 I/O addresses in Slot 15.
Table 4-1. Peripheral Register I/O Address Map in Slot 15
Register Description
DMA Controller 1
I/O Address Range
0F000H
0F020H
0F040H
0F080H
0F0A0H
0F0F0H
0F400H
0F480H
0F4A0H
0F4C0H
0F4F8H
0F800H
0F810H
0F820H
0F860H
0F8F8H
–
–
–
–
–
–
–
–
–
–
–
–
–
0F01FH
0F03FH
0F05FH
0F09FH
0F0BFH
0F0FFH
0F47FH
0F49FH
0F4BFH
0F4CFH
0F4FFH
0F80FH
0F81FH
0F83FH
0F87FH
0F8FFH
Master Interrupt Controller
Programmable Interval Timer
DMA Page Registers
Slave Interrupt Controller
Math Coprocessor
Chip Select Unit
Synchronous Serial I/O Unit
DRAM Refresh Control Unit
Watchdog Timer Unit
Asynchronous Serial I/O Channel 0 (COM1)
Clock Generation and Power Management Unit
External/Internal Bus Interface Unit
Chip Configuration Registers
Parallel I/O Ports
–
–
–
Asynchronous Serial I/O Channel 1 (COM2)
4-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
4.5 I/O ADDRESS DECODING TECHNIQUES
One of the key features of the Intel386 EX processor is that it is configurable for compatibility
with the standard PC/AT architecture. In a PC/AT system, the platform I/O resources are located
in the slot 0 I/O address space. For the Intel386 EX processor, this means that PC/AT-compatible
internal peripherals should be reflected in slot 0 of the I/O space for DOS operating system and
application software to access and manipulate them properly.
This discussion leads to the concepts of DOS I/O space and expanded I/O space.
DOS I/O Space
DOS I/O space refers to the lower 1 Kbyte of I/O addresses, where
only PC/AT-compatible peripherals can be mapped.
Expanded I/O Space
Expanded I/O space refers to the top 4 Kbytes of I/O addresses,
where all peripheral registers are physically located. The remainder
of this section explains how special I/O address decoding schemes
manipulate register addresses within these two I/O spaces.
4.5.1 Address Configuration Register
I/O address locations 22H and 23H in DOS I/O space offer a special case. These address locations
sor and other integrated PC solutions use them to enable extra address space required for config-
uration registers specific to these products. On the Intel386 EX processor, these address locations
are used to hide the peripheral registers in the expanded I/O space. The expanded I/O space can
be enabled (registers visible) or disabled (registers hidden).
The 16-bit register at I/O location 22H can also be used to control mapping of various internal
peripherals in I/O address space. This register, REMAPCFG, is defined in Figure 4-3.
The remap bits of this register control whether the internal PC compatible peripherals are mapped
into the DOS I/O space. Setting the peripheral bit makes the peripheral accessible only in expand-
ed I/O space. Clearing the peripheral bit makes the peripheral accessible in both DOS I/O space
and expanded I/O space. To access the REMAPCFG register, you must first enable the expanded
I/O address space as described in the next section. At reset, this register is cleared, mapping in-
ternal PC/AT-compatible peripherals into DOS I/O space.
4-6
SYSTEM REGISTER ORGANIZATION
Address Configuration Register
REMAPCFG
Expanded Addr: 0022H
PC/AT Address: 0022H
Reset State:
0000H
15
8
0
ESE
—
—
—
—
—
—
—
—
—
7
S1R
S0R
ISR
IMR
DR
TR
Bit
Number
Bit
Mnemonic
Function
15
ESE
0 = Disables expanded I/O space
1 = Enables expanded I/O space
14–7
6
—
Reserved.
S1R
0 = Makes serial channel 1 (COM2) accessible in both DOS I/O space
and expanded I/O space
1 = Remaps serial channel 1 (COM2) address into expanded I/O space
5
4
S0R
ISR
0 = Makes serial channel 0 (COM1) accessible in both DOS I/O space
and expanded I/O space
1 = Remaps serial channel 0 (COM1) address into expanded I/O space
0 = Makes the slave 82C59A interrupt controller accessible in both DOS
I/O space and expanded I/O space
1 = Remaps slave 82C59A interrupt controller address into expanded
I/O space
3
2
IMR
DR
0 = Makes the master 82C59A interrupt controller accessible in both
DOS I/O space and expanded I/O space
1 = Remaps master 82C59A interrupt controller address into expanded
I/O space
0 = Makes the DMA address accessible in both DOS I/O space and
expanded I/O space
1 = Remaps DMA address into expanded I/O space
1
0
—
Reserved.
TR
0 = Makes the timer control unit accessible in both DOS I/O space and
expanded I/O space
1 = Remaps timer control unit address into expanded I/O space
Figure 4-3. Address Configuration Register (REMAPCFG)
4-7
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
4.5.2 Enabling and Disabling the Expanded I/O Space
The Intel386 EX processor’s expanded I/O space is enabled by a specific write sequence to I/O
addresses 22H and 23H (Figure 4-4). Once the expanded I/O space is enabled, internal peripher-
als (timers, DMA, interrupt controllers and serial communication channels) can be mapped out
of DOS I/O space (using the REMAPCFG register) and registers associated with other internal
peripherals (such as the chip-select unit, power management unit, watchdog timer) can be access-
ed.
4.5.2.1
Programming REMAPCFG Example
The expanded I/O space enable (ESE) bit in the REMAPCFG register can be set only by three
sequential write operations to I/O addresses 22H and 23H as described in Figure 4-4. Once ESE
is set, REMAPCFG and all the on-chip registers in the expanded I/O address range 0F000H–
0FFFFH can be accessed. The remap bits in REMAPCFG are still in effect even after the ESE bit
is cleared.
;;disable interrupts
CLI
; Enable expanded I/O space of Intel386(tm) EX processor
; for peripheral initialization.
MOV AX, 08000H
OUT 23H, AL
XCHG AL, AH
OUT 22H, AL
OUT 22H, AX
; Enable expanded I/O space
; and unlock the re-map bits
;; at this point PC/AT peripherals can be mapped out
;; For example,
;; Map out the on-chip DMA channels from the DOS I/O space (slot 0)
MOV AL, 04H
OUT 22H, AL
; Disables expanded I/O space
MOV AL, 00H
OUT 23H, AL
;; Re-enable Interrupts
STI
Figure 4-4. Setting the ESE Bit Code Example
The REMAPCFG register is write-protected until the expanded I/O space is enabled. When the
enabling write sequence is executed, it sets the ESE bit. A program can check this bit to see
whether it has access to the expanded I/O space registers. Clearing the ESE bit disables the ex-
panded I/O space. This can be done by a byte write with a value of 0 to I/O address 23H. This
again locks the REMAPCFG register and makes it read-only.
4-8
SYSTEM REGISTER ORGANIZATION
4.6 ADDRESSING MODES
Combinations of the value of ESE bit and the individual remap bits in the REMAPCFG register
yield four different peripheral addressing modes for I/O address decoding.
4.6.1 DOS-compatible Mode
DOS-compatible mode is achieved by clearing ESE and all the peripheral remap bits. In this
mode, all PC/AT-compatible peripherals are mapped into the DOS I/O space. Only address lines
A9:0 are decoded for internal peripherals. Accesses to PC/AT-compatible peripherals are valid,
while all other internal peripherals are inaccessible (see Figure 4-5).
This mode is useful for accessing the internal timer, interrupt controller, serial I/O ports, or DMA
controller in a DOS-compatible environment.
4-9
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
3FFH
On-chip UART-0
On-chip UART-1
FFFFH
On-chip 8259A-2
On-chip Timer
0
0
0
23H
22H
REMAPCFG
Register
0
0
0
0
0
0
0
F000H
Expanded I/O Space
On-chip 8259A-1
Note:
Shaded area indicates that
expanded I/O space peripherals
are not accessible
On-chip DMA
0H
DOS I/O Space
A2495-02
Figure 4-5. DOS-Compatible Mode
4-10
SYSTEM REGISTER ORGANIZATION
4.6.2 Nonintrusive DOS Mode
This mode is achieved by first setting the ESE bit (using the three sequential writes), setting the
are set are mapped out of DOS I/O space. Like DOS-compatible mode, only address lines A9:0
are decoded internally. This mode is useful for connecting an external peripheral instead of using
the integrated peripheral. For example, a system might use an external 8237A DMA rather than
using the internal DMA unit. For this configuration, set the ESE bit, set the remap bit associated
with the DMA unit and then clear the ESE bit. In this case, the external 8237A is accessible in
the DOS I/O space, while the internal DMA can be accessed only after the expanded I/O space is
enabled. (See Figure 4-6.)
4.6.3 Enhanced DOS Mode
This mode is achieved by setting the ESE bit and clearing all PC/AT-compatible peripherals’
remap bits. Address lines A15:0 are decoded internally. The expanded I/O space is enabled and
the PC/AT-compatible internal peripherals are accessible in either DOS I/O space or expanded
I/O space. (See Figure 4-7.) If an application frequently requires the additional peripherals, but
at the same time wants to maintain DOS compatibility for ease of development, this is the most
useful mode.
4.6.4 Non-DOS Mode
This mode is achieved by setting the ESE bit and setting all peripherals’ remap bits. Address lines
A15:0 are decoded internally. The expanded I/O space is enabled and all peripherals can be ac-
cessed only in expanded I/O space. This mode is useful for systems that don’t require DOS com-
patibility and have other custom peripherals in slot 0 of the I/O space. (See Figure 4-8.)
For all DOS peripherals, the lower 10 bits in the DOS I/O space and in the expanded I/O space
are identical (except the UARTs, whose lower 8 bits are identical). This makes correlation of their
respective offsets in DOS and expanded I/O spaces easier. Also, the UARTs have fixed I/O ad-
dresses. This differs from standard PC/AT configurations, in which these address ranges are pro-
grammable.
4-11
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
3FFH
On-chip UART-0
On-chip UART-1
FFFFH
On-chip 8259A-2
On-chip Timer
0
23H
22H
0
0
REMAPCFG
Register
0
0
0
0
1
0
0
F000H
Expanded I/O Space
On-chip 8259A-1
Note:
Internal DMA
Shaded area indicates that the on-chip
DMA and expanded I/O space
peripherals are not accessible
0H
DOS I/O Space
A2496-02
Figure 4-6. Example of Nonintrusive DOS-Compatible Mode
4-12
SYSTEM REGISTER ORGANIZATION
3FFH
On-chip UART-2
On-chip UART-1
FFFFH
Other Peripherals
UART-0
UART-1
Timer
On-chip 8259A-2
On-chip Timer
0
8259A-2
8259A-1
1
0
23H
22H
REMAPCFG
Register
On-chip DMA
F000H
0
0
0
0
0
0
0
Expanded I/O Space
On-chip 8259A-1
On-chip DMA
0H
DOS I/O Space
A2501-02
Figure 4-7. Enhanced DOS Mode
4-13
SYSTEM REGISTER ORGANIZATION
4.7 PERIPHERAL REGISTER ADDRESSES
Table 4-2 lists the addresses and names of all user-accessible peripheral registers. I/O Registers
can be accessed as bytes or words. Word accesses to byte registers result in two sequential 8-bit
I/O transfers. The default (reset) value of each register is shown in the Reset Value column. An X
in this column signifies that the register bits are undefined. Some address values do not access
registers, but are decoded to provide a logic control signal. These addresses are listed as Not a
register in the Reset column.
Table 4-2. Peripheral Register Addresses (Sheet 1 of 6)
Expanded
Address
PC/AT
Address
Access Type
(Byte/Word)
Register Name
Reset Value
DMA Controller and Bus Arbiter
F000H
F001H
F002H
F003H
F004H
F005H
F006H
F007H
F008H
F009H
F00AH
F00BH
F00CH
F00DH
F00EH
F00FH
F010H
F011H
F012H
F013H
F014H
F015H
F016H
F017H
NOTES:
0000H
Byte
Byte
Byte
Byte
DMA0TAR0/1 (Note 1)
XX
0001H
0002H
0003H
0004H
0005H
0006H
0007H
0008H
0009H
000AH
000BH
000CH
000DH
000EH
000FH
DMA0BYC0/1 (Note 1)
DMA1TAR0/1 (Note 1)
DMA1BYC0/1 (Note 1)
Reserved
XX
XX
XX
Reserved
Reserved
Reserved
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
DMACMD1/DMASTS
DMASRR
00H
00H
DMAMSK
04H
DMAMOD1
00H
DMACLRBP
DMACLR
Not a register
Not a register
DMACLRMSK
DMAGRPMSK
DMA0REQ0/1
DMA0REQ2/3
DMA1REQ0/1
DMA1REQ2/3
Reserved
Not a register
03H
XX
XX
XX
XX
Reserved
Reserved
Reserved
1. Byte pointer in flip-flop in DMA determines which register is accessed.
2. Shaded rows indicate reserved areas.
4-15
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table 4-2. Peripheral Register Addresses (Sheet 2 of 6)
Expanded
Address
PC/AT
Address
Access Type
(Byte/Word)
Register Name
Reset Value
F018H
F019H
F01AH
F01BH
F01CH
F01DH
F01EH
Byte
Byte
Byte
Byte
Byte
Byte
Byte
DMABSR
X1X10000B
00H
DMACHR/DMAIS
DMACMD2
DMAMOD2
DMAIEN
08H
00H
00H
DMAOVFE
DMACLRTC
0AH
Not a register
Master Interrupt Controller
F020H
F021H
0020H
Byte
ICW1m/IRRm/ISRm/
OCW2m/OCW3m
XX
XX
0021H
0022H
Byte
ICW2m/ICW3m/ICW4m/
OCW1m/POLLm
Address Configuration Register
Word REMAPCFG
Timer/counter Unit
0022H
0000H
F040H
F041H
F042H
F043H
0040H
0041H
0042H
0043H
Byte
Byte
Byte
Byte
TMR0
XX
XX
XX
XX
TMR1
TMR2
TMRCON
DMA Page Registers
Reserved
F080H
F081H
F082H
F083H
F084H
F085H
F086H
F087H
F088H
F089H
F08AH
F08BH
F08CH
NOTES:
0081H
0082H
0083H
Reserved
Reserved
DMA1TAR2
Reserved
DMA1TAR3
DMA0TAR3
DMA0TAR2
Reserved
Reserved
Reserved
Reserved
Reserved
Byte
XX
Byte
Byte
Byte
XX
XX
XX
0087H
0089H
008AH
008BH
1. Byte pointer in flip-flop in DMA determines which register is accessed.
2. Shaded rows indicate reserved areas.
4-16
SYSTEM REGISTER ORGANIZATION
Table 4-2. Peripheral Register Addresses (Sheet 3 of 6)
Expanded
Address
PC/AT
Address
Access Type
(Byte/Word)
Register Name
Reset Value
F08DH
F08EH
F08FH
F098H
F099H
F09AH
F09BH
Reserved
Reserved
Reserved
DMA0BYC2
DMA1BYC2
Reserved
Reserved
Byte
Byte
XX
XX
A20GATE and Fast CPU Reset
Byte PORT92
Slave Interrupt Controller
F092H
0092H
XXXXXX10B
F0A0H
F0A1H
00A0H
00A1H
Byte
ICW1s/IRRs/ISRs/
OCW2s/OCW3s
XX
XX
Byte
ICW2s/ICW3s/ICW4s/
OCW1s/POLLs
Chip-select Unit
CS0ADL
F400H
F402H
F404H
F406H
F408H
F40AH
F40CH
F40EH
F410H
F412H
F414H
F416H
F418H
F41AH
F41CH
F41EH
F420H
F422H
NOTES:
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
CS0ADH
CS0MSKL
CS0MSKH
CS1ADL
CS1ADH
CS1MSKL
CS1MSKH
CS2ADL
CS2ADH
CS2MSKL
CS2MSKH
CS3ADL
CS3ADH
CS3MSKL
CS3MSKH
CS4ADL
CS4ADH
1. Byte pointer in flip-flop in DMA determines which register is accessed.
2. Shaded rows indicate reserved areas.
4-17
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table 4-2. Peripheral Register Addresses (Sheet 4 of 6)
Expanded
Address
PC/AT
Address
Access Type
(Byte/Word)
Register Name
Reset Value
F424H
F426H
F428H
F42AH
F42CH
F42EH
F430H
F432H
F434H
F436H
F438H
F43AH
F43CH
F43EH
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
CS4MSKL
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
FF6FH
FFFFH
FFFFH
FFFFH
CS4MSKH
CS5ADL
CS5ADH
CS5MSKL
CS5MSKH
CS6ADL
CS6ADH
CS6MSKL
CS6MSKH
UCSADL
UCSADH
UCSMSKL
UCSMSKH
Synchronous Serial I/O Unit
F480H
F482H
F484H
F486H
F488H
F48AH
Word
Word
Byte
Byte
Byte
Byte
SSIOTBUF
0000H
0000H
00H
SSIORBUF
SSIOBAUD
SSIOCON1
SSIOCON2
SSIOCTR
C0H
00H
00H
Refresh Control Unit
F4A0H
F4A2H
F4A4H
F4A6H
Word
Word
Word
Word
RFSBAD
0000H
0000H
0000H
00FFH
RFSCIR
RFSCON
RFSADD
Watchdog Timer Unit
F4C0H
F4C2H
F4C4H
F4C6H
F4C8H
NOTES:
Word
Word
Word
Word
Word
WDTRLDH
003FH
WDTRLDL
WDTCNTH
WDTCNTL
WDTCLR
FFFFH
003FH
FFFFH
Not a register
1. Byte pointer in flip-flop in DMA determines which register is accessed.
2. Shaded rows indicate reserved areas.
4-18
SYSTEM REGISTER ORGANIZATION
Table 4-2. Peripheral Register Addresses (Sheet 5 of 6)
Expanded
Address
PC/AT
Address
Access Type
(Byte/Word)
Register Name
Reset Value
F4CAH
Byte
WDTSTATUS
00H
Asynchronous Serial I/O Channel 0 (COM1)
F4F8H
F4F9H
F4FAH
F4FBH
F4FCH
F4FDH
F4FEH
F4FFH
03F8H
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
RBR0/TBR0/DLL0
IER0/DLH0
IIR0
XX/XX/02H
00H/00H
01H
03F9H
03FAH
03FBH
03FCH
03FDH
03FEH
03FFH
LCR0
00H
MCR0
00H
LSR0
60H
MSR0
X0H
SCR0
XX
Clock Generation and Power Management
F800H
F804H
Byte
PWRCON
CLKPRS
00H
Word
0000H
Device Configuration Registers
F820H
F822H
F824H
F826H
F830H
F832H
F834H
F836H
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
P1CFG
00H
00H
00H
00H
00H
00H
00H
00H
P2CFG
P3CFG
PINCFG
DMACFG
INTCFG
TMRCFG
SIOCFG
Parallel I/O Ports
F860H
F862H
F864H
F868H
F86AH
F86CH
F870H
F872H
F874H
NOTES:
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
P1PIN
XX
P1LTC
P1DIR
P2PIN
P2LTC
P2DIR
P3PIN
P3LTC
P3DIR
FFH
FFH
XX
FFH
FFH
XX
FFH
FFH
1. Byte pointer in flip-flop in DMA determines which register is accessed.
2. Shaded rows indicate reserved areas.
4-19
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table 4-2. Peripheral Register Addresses (Sheet 6 of 6)
Expanded
Address
PC/AT
Address
Access Type
(Byte/Word)
Register Name
Reset Value
Asynchronous Serial I/O Channel 1 (COM2)
F8F8H
F8F9H
F8FAH
F8FBH
F8FCH
F8FDH
F8FEH
F8FFH
NOTES:
02F8H
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
RBR1/TBR1/DLL1
IER1/DLH1
IIR1
XX/XX/02H
00H/00H
01H
02F9H
02FAH
02FBH
02FCH
02FDH
02FEH
02FFH
LCR1
00H
MCR1
00H
LSR1
60H
MSR1
X0H
SCR1
XX
1. Byte pointer in flip-flop in DMA determines which register is accessed.
2. Shaded rows indicate reserved areas.
4-20
5
DEVICE
CONFIGURATION
CHAPTER 5
DEVICE CONFIGURATION
eral to peripheral connections. This chapter describes the available configurations and how to
configure them.
This chapter is organized as follows:
• Peripheral Configuration (page 5-3)
• Pin Configuration (page 5-23)
• Device Configuration Procedure (page 5-28)
• Configuration Example (page 5-28)
5.1 INTRODUCTION
†
Device configuration is the process of setting up the microprocessor’s on-chip peripherals for a
particular system design. Specifically, device configuration consists of programming registers to
connect peripheral signals to the package pins and interconnect the peripherals. The peripherals
include the following:
• DMA Controller (DMA)
• Interrupt Control Unit (ICU)
• Timer/counter Unit (TCU)
• Asynchronous Serial I/O Units (SIO0, SIO1)
• Synchronous Serial I/O Unit (SSIO)
• Refresh Control Unit (RCU)
• Chip-select Unit (CSU)
• Watchdog Timer Unit (WDT)
In addition, the pin configuration registers control connections from the coprocessor to the core
and pin connections to the bus arbiter.
†
In this chapter, the terms “peripheral” and “on-chip peripheral” are used interchangeably. An “off-chip peripheral” is
external to the Intel386 EX processor.
5-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Figure 5-1 shows Peripheral A and its connections to other peripherals and the package pins. The
“Internal Connection Logic” provides three kinds of connections:
• Connections between peripherals
• Connections to package pins via multiplexers
• Direct connections to package pins without multiplexers
The internal connection logic is controlled by the Peripheral A configuration register.
Each pin multiplexer (“Pin Mux”) connects one of two internal signals to a pin. One is a periph-
eral signal. The second signal can be an I/O port signal or a signal from/to another peripheral. The
pin multiplexers are controlled by the pin configuration registers. Some input-only pins without
multiplexers (“Shared Pins w/o Muxes”) are routed to two different peripherals. Your design
should use only one of the inputs and disable or ignore the input going to the second peripheral.
Together, the peripheral configuration registers and the pin configuration registers allow you to
select the peripherals to be used, to interconnect them as your design requires, and to bring se-
lected signals to the package pins.
Peripherals B, C, D, ...
Pins
with
Muxes
Microprocessor
Peripheral A
Pin
Muxes
Internal
Connection
Logic
Shared Pins
w/o Muxes
Control
Peripheral A
Configuration
Register
Control
Pin Configuration Registers
A2535-01
Figure 5-1. Peripheral and Pin Connections
5-2
DEVICE CONFIGURATION
5.2 PERIPHERAL CONFIGURATION
This section describes the configuration of each on-chip peripheral. For more detailed informa-
tion on the peripheral itself, see the chapter describing that peripheral.
The symbology used for signals that share a device pin is shown in Figure 5-2. Of the two signal
in parentheses is the alternate signal, which connects to a different peripheral or the core. When
a pin has a multiplexer, it is shown as a switch, and the register bit that controls it is noted above
the switch.
5.2.1 DMA Controller, Bus Arbiter, and Refresh Unit Configuration
Figure 5-2 shows the DMA controller, bus arbiter, and refresh unit configuration. Requests for a
DMA data transfer are shown as inputs to the multiplexer:
• A synchronous serial I/O transmitter (SSTBE) or receiver (SSRBF)
• A timer (OUT1, OUT2)
• An external source (DRQ0, DRQ1)
The inputs are selected by the DMA configuration register (see Figure 5-3).
5.2.1.1
Using The DMA Unit with External Devices
For each DMA channel, three bits in the DMA configuration register (Figure 5-3) select the ex-
ternal request input or one of seven request inputs from the peripherals. Another bit enables or
disables that channel’s DMA acknowledge signal (DACKn#) at the device pin. Enable the
DACKn# signal only when you are using the external request signal (DRQn) and need DACKn#.
The acknowledge signals are not routed to the on-chip peripherals, and therefore, these peripher-
als cannot initiate single-cycle (fly-by) DMA transfers.
An external bus master cannot talk directly to internal peripheral modules because the external
address lines are outputs only. However, an external device could use a DMA channel to transfer
data to or from an internal peripheral because the DMA generates the addresses. This transaction
would be a two-cycle DMA bus transaction.
5.2.1.2
DMA Service to an SIO or SSIO Peripheral
A DMA unit is useful for servicing an SIO or SSIO peripheral operating at a high baud rate. At
high baud rates, the interrupt response time of the core may be too long to allow the serial
channels to use an interrupt to service the receive-buffer-full condition. By the time the interrupt
service routine (ISR) is ready to transfer the receive-buffer data to memory, new data would have
been loaded into the buffer. The issue is the interrupt latency which is the amount of time the
processor takes from recognizing the interrupt to executing the first line of code in the ISR. This
interrupt latency needs to be calculated to determine if an ISR can handle the high baud rate. If
the Interrupt Latency is too high, data transfers to and from the serial channels can occur within
a few bus cycles of the time that a serial unit is ready to move data by using an appropriately
5-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
configured DMA channel. SIO and SSIO inputs to the DMA are selected by the DMA configu-
ration register (Figure 5-3).
5.2.1.3
Using The Timer To Initiate DMA Transfers
A timer output (OUT1, OUT2) can initiate periodic data transfers by the DMA. A DMA channel
is programmed for the transfer, then a timer output pulse triggers the transfer. The most useful
DMA and timer combinations for this type of transfer are the periodic timer modes (mode 2 and
UNIT,” and Chapter 12, “DMA CONTROLLER,” for more information on how to program the
peripherals.
5.2.1.4
Limitations Due To Pin Signal Multiplexing
Pin signal multiplexing can preclude the simultaneous use of a DMA channel and another periph-
eral or specific peripheral signal (see Figure 5-2). For example, using DMA channel 1 with an
external requester device precludes using SIO channel 1 due to the multiplexed signal pairs
DRQ1/RXD1 and DACK1#/TXD1. Please refer to the Intel386™ EX Microprocessor Pin Mul-
tiplexing Map (Order Number 272587) for a complete diagram of multiplexed signals.
5-4
DEVICE CONFIGURATION
DMACFG.2:0
DMA
DRQ0
(DCD1#)†
3
0
RBFDMA0 (SIO0)
TXEDMA1 (SIO1)
SSTBE (SSIO)
OUT1 (TCU)
RBFDMA1 (SIO1)
TXEDMA0 (SIO0)
SSRBF (SSIO)
1
2
3
4
5
6
7
To SIO1
DREQ0
DMACFG.3
PINCFG.4
1
0
DACK0#
(CS5#)
DMAACK0#
DREQ1
From CSU
DMACFG.6:4
3
0
DRQ1
(RXD1)
RBFDMA1 (SIO1 )
TXEDMA0 (SIO0)
SSRBF (SSIO)
OUT2 (TCU)
RBFDMA0 (SIO0)
TXEDMA1 (SIO1)
SSTBE (SSIO)
1
2
3
4
5
6
7
To SIO1
DMACFG.7
PINCFG.2
0
DACK1#
(TXD1)
DMAACK1#
From SIO1
1
To ICU
DMAINT
0
1
PINCFG.3
End of Process
EOP#
(CTS1#)
From SIO1
1
P1CFG.6
HOLD
HOLD
(P1.6)
To/From I/O Port 1
0
Bus Arbiter
HLDA
To
Core
P1CFG.7
1
1
HOLD
HLDA
(P1.7)
To/From I/O Port 1
From CSU
0
Refresh Unit
REFRESH#
PINCFG.6
From
Core
HLDA
REFRESH#
(CS6#)
0
† Alternate pin signals are in parentheses.
A2516-02
Figure 5-2. Configuration of DMA, Bus Arbiter, and Refresh Unit
5-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
DMA Configuration
DMACFG
Expanded Addr: F830H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
D1MSK
D1REQ2
D1REQ1
D1REQ0
D0MSK
D0REQ2
D0REQ1
D0REQ0
Bit
Number
Bit
Mnemonic
Function
7
D1MSK
DMA Acknowledge 1 Mask:
0 = DMA channel 1’s acknowledge (DMAACK1#) signal is not masked.
1 = Masks DMA channel 1’s acknowledge (DMAACK1#) signal. Useful
when channel 1’s request (DREQ1) input is connected to an internal
peripheral.
6–4
D1REQ2:0
DMA Channel 1 Request Connection:
Connects one of the eight possible hardware sources to channel 1’s
request input (DREQ1).
000 = DRQ1 pin (external peripheral)
001 = SIO channel 1’s receive buffer full signal (RBFDMA1)
010 = SIO channel 0’s transmit buffer empty signal (TXEDMA0)
011 = SSIO receive holding buffer full signal (SSRBF)
100 = TCU counter 2’s output signal (OUT2)
101 = SIO channel 0’s receive buffer full signal (RBFDMA0)
110 = SIO channel 1’s transmit buffer empty signal (TXEDMA1)
111 = SSIO transmit holding buffer empty signal (SSTBE)
3
D0MSK
DMA Acknowledge 0 Mask:
0 = DMA channel 0’s acknowledge (DMAACK0#) signal is not masked.
1 = Masks DMA channel 0’s acknowledge (DMAACK0#) signal. Useful
when channel 0’s request (DREQ0) input is connected to an internal
peripheral.
2–0
D0REQ2:0
DMA Channel 0 Request Connection:
Connects one of the eight possible hardware sources to channel 0’s
request input (DREQ0).
000 = DRQ0 pin (external peripheral)
001 = SIO channel 0’s receive buffer full signal (RBFDMA0)
010 = SIO channel 1’s transmit buffer empty signal (TXEDMA1)
011 = SSIO transmit holding buffer empty signal (SSTBE)
100 = TCU counter 1’s output signal (OUT1)
101 = SIO channel 1’s receive buffer full signal (RBFDMA1)
110 = SIO channel 0’s transmit buffer empty signal (TXEDMA0)
111 = SSIO receive holding buffer full signal (SSRBF)
Figure 5-3. DMA Configuration Register (DMACFG)
5-6
DEVICE CONFIGURATION
5.2.2 Interrupt Control Unit Configuration
The interrupt control unit (ICU) comprises two 82C59A interrupt controllers connected in cas-
cade, as shown in Figure 5-4. (See Chapter 9 for more information.) Figure 5-5 describes the in-
terrupt configuration register (INTCFG).
The ICU receives requests from eight internal sources:
• Three outputs from the timer/counter unit (OUT2:0)
• An output from each of the serial I/O units (SIOINT1:0)
• An output from the synchronous serial I/O unit (SSIOINT)
• An output from the DMA unit (DMAINT)
• An output from the WDT unit (WDTOUT#)
In addition, the ICU controls the interrupt sources on ten external pins:
• INT3:0 (multiplexed with I/O port signals P3.5:2) are enabled or disabled by the P3CFG
register (see Figure 5-18).
• INT7:4 share their package pins with four TCU inputs: TMRGATE1, TMRCLK1,
TMRGATE0, and TMRCLK0. These signal pairs are not multiplexed; however, the pin
• INT9:8 share their pins with TMROUT1, TMROUT0, P3.1, P3.0
The three cascade outputs (CAS2:0) should be enabled when an external 82C59A module is con-
nected to one of the INT9:8 or INT3:0 signals. The cascade outputs are ORed with address lines
A18:16. See “Interrupt Acknowledge Cycle” on page 6-23 for details.
Use Tables 5-1 and 5-2 to configure the functionality of the master 82C59A’s IR3, IR4 inputs, and
the associated external pins.
5-7
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table 5-1. Master’s IR3 Connections
Function
INTCFG.6
MCR1.3
P3CFG.1
IR3 connected to SIOINT1
P3.1 selected at pin (P3.1)
0
X
0
IR3 connected to SIOINT1
0
X
1
OUT1 connected to pin (TMROUT1)
IR3 internally driven low
1
1
1
0
0
1
0
1
0
P3.1 selected at pin (P3.1)
IR3 connected to pin (INT8)
IR3 connected to SIOINT1
P3.1 selected at pin (P3.1)
IR3 connected to SIOINT1
1
1
1
pin (INT8) must not be left floating
NOTE: X is a don’t care
Table 5-2. Master’s IR4 Connections
Function
INTCFG.5
MCR0.3
P3CFG.0
IR4 connected to SIOINT0
P3.0 selected at pin (P3.0)
0
X
X
0
IR4 connected to SIOINT0
0
1
OUT0 connected to pin (TMROUT0)
IR4 internally driven low
1
1
1
0
0
1
0
1
0
P3.0 selected at pin (P3.0)
IR4 connected to pin (INT9)
IR4 connected to SIOINT0
P3.0 selected at pin (P3.0)
IR4 connected to SIOINT0
1
1
1
pin (INT9) must not be left floating
NOTE: X is a don’t care
5-8
DEVICE CONFIGURATION
IR0
OUT0 (TCU)
VSS
8259A
Master
P3CFG.2
INTCFG.6
P3CFG.2
INT0
1
0
1
IR1
IR2
(P3.2)†
To/From I/O Port 3
0
INT
MCR1.3
SIOINT1
1
INTR
(to
core)
0
1
SIOINT1
SIOINT0
1
0
IR3
INTCFG.6
1
INT8
TMROUT1
(P3.1)
P3CFG.1
0
OUT1(TCU)
MCR0.3
P3.1
0
INTCFG.5
0
1
SIOINT0
IR4
IR5
1
0
INTCFG.5
1
P3GFG.0
0
1
INT9
TMROUT0
(P3.0)
P3CFG.3
P3CFG.4
0
OUT0(TCU)
To/From I/O Port 3
To/From I/O Port 3
P3.0
VSS
VSS
VSS
0
1
P3CFG.3
1
1
INT1
(P3.3)
0
0
1
P3CFG.4
IR6
IR7
INT2
(P3.4)
CAS2:0
0
P3CFG.5
INTCFG.0
INTCFG.1
0
1
P3CFG.5
1
INT3
(P3.5)
To/From I/O Port 3
To TCU
0
VSS
0
1
IR0
IR1
INT4
INT
(TMRCLK0)
8259A
Slave
SSIOINT
0
1
INT5
To TCU
(TMRGATE0)
OUT1(TCU)
OUT2(TCU)
IR2
IR3
INTCFG.4
DMAINT
To TCU
0
1
IR4
IR5
INT6
(TMRCLK1)
0
1
VSS
0
1
CAS2:0
INTCFG.3
INTCFG.2
VSS
0
1
IR6
IR7
INT7
To TCU
(TMRGATE1)
WDTOUT#
VSS
INTCFG.7
0
1
3
CAS2:0
(A18:16)
A18:16
† Alternate pin signals are in parentheses
Heavier lines indicate multiple signals.
A2522-03
Figure 5-4. Interrupt Control Unit Configuration
5-9
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Interrupt Configuration
INTCFG
Expanded Addr: F832H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
CE
IR3
IR4
SWAP
IR6
IR5/IR4
IR1
IR0
Bit
Number
Bit
Mnemonic
Function
7
CE
Cascade Enable:
0 = Disables the cascade signals CAS2:0 from appearing on the A18:16
address lines during interrupt acknowledge cycles.
1 = Enables the cascade signals CAS2:0, providing access to external
slave 82C59A devices. The cascade signals are used to address
specific slaves. If enabled, slave IDs appear on the A18:16 address
lines during interrupt acknowledge cycles, but are high during idle
cycles.
6
5
4
IR3
Internal Master IR3 Connection:
See Table 5-1 on page 5-8 for all the IR3 configuration options.
IR4
Internal Master IR4 Connection:
See Table 5-2 on page 5-8 for all the IR4 configuration options.
SWAP
INT6/DMAINT Connection:
0 = Connects DMAINT to the slave IR4. Connects INT6 to the slave IR5.
1 = Connects the INT6 pin to the slave IR4. Connects DMAINT to the slave
IR5.
3
2
IR6
Internal Slave IR6 Connection:
0 = Connects VSS to the slave IR6 signal.
1 = Connects the INT7 pin to the slave IR6 signal.
IR5/IR4
Internal Slave IR4 or IR5 Connection:
These depend on whether INTCFG.4 is set or clear.
0 = Connects VSS to the slave IR5 signal.
1 = Connects either the INT6 pin or DMAINT to the slave IR5 signal.
1
0
IR1
IR0
Internal Slave IR1 Connection:
0 = Connects the SSIO interrupt signal (SSIOINT) to the slave IR1 signal.
1 = Connects the INT5 pin to the slave IR1 signal.
Internal Slave IR0 Connection:
0 = Connects VSS to the slave IR0 signal.
1 = Connects the INT4 pin to the slave IR0 signal.
Figure 5-5. Interrupt Configuration Register (INTCFG)
5-10
DEVICE CONFIGURATION
The three-channel Timer/counter Unit (TCU) and its configuration register (TMRCFG) are
shown in Figure 5-6 and Figure 5-7. The clock inputs can be external signals (TMRCLK2:0) or
the on-chip programmable clock (PSCLK). All clock inputs can be held low by programming bits
in the TMRCFG register. The gate inputs can be controlled through software using TMRCFG.6
and the appropriate GTnCON bits in the TMRCFG register. Several of the timer signals go to the
interrupt control unit (see Figure 5-4).
The Timer/counter0 and Timer/counter1 signals are selected individually. In contrast, the Tim-
er/counter2 signals (TMRCLK2, TMRGATE2, TMROUT2) are selected as a group. Note that us-
ing the Timer/counter2 signals precludes use of the coprocessor signals (PEREQ, BUSY#, and
ERROR#).
The CLKINn and GATEn inputs of Timer/counter0 and Timer/counter1 are routed directly to
shared input pins, TMRCLK0/INT4, TMRCLK1/INT6, TMRGATE0/INT5 and
TMRGATE1/INT7. The OUTn inputs of these two counters can be connected to pins
TMROUT0/INT9/P3.0 and TMROUT1/INT8/P3.1 respectively, using bits in registers P3CFG
and INTCFG.
5-11
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
TMRCFG.7
Timer/Counter
TMRCFG.0
PSCLK
Unit
0
1
CLKIN0
TMRCLK0
(INT4)†
To ICU
VCC
TMRCFG.6
TMRCFG.1
0
1
0
1
GATE0
TMRGATE0
(INT5)
TMRCFG.1
To ICU
To ICU
P3CFG.0
0
1
OUT0
TMROUT0
(INT9)
(P3.0)
To/From I/O Port 3
TMRCFG.2
PSCLK
0
1
CLKIN1
TMRCLK1
(INT6)
To ICU
VCC
TMRCFG.6
TMRCFG.3
0
1
0
1
GATE1
OUT1
TMRGATE1
(INT7)
TMRCFG.3
To ICU
P3CFG.1
To ICU, DMA
1
1
TMROUT1
(INT8)
(P3.1)
To/From I/O Port 3
TMRCFG.4
PSCLK
0
CLKIN2
PINCFG.5
0
1
TMRCLK2
(PEREQ)
To Core
0
TMRCFG.6
TMRCFG.5
VCC
0
1
0
1
1
GATE2
OUT2
TMRGATE2
(BUSY#)
TMRCFG.5
To Core
0
To ICU, DMA
1
TMROUT2
(ERROR#)
To Core
0
† Alternate pin signals are in parentheses.
A2517-03
Figure 5-6. Timer/Counter Unit Configuration
5-12
DEVICE CONFIGURATION
.
Timer Configuration
TMRCFG
Expanded Addr: F834H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
TMRDIS
SWGTEN GT2CON CK2CON
GT1CON CK1CON GT0CON CK0CON
Bit
Number
Bit
Mnemonic
Function
7
TMRDIS
Timer Disable:
0 = Enables the CLKINn signals.
1 = Disables the CLKINn signals.
6
5
SWGTEN
Software GATEn Enable
0 = Connects GATEn to either the VCC pin or the TMRGATEn pin.
1 = Enables GT2CON, GT1CON, and GT0CON to control the
connections to GATE2, GATE1 and GATE0 respectively.
GT2CON
Gate 2 Connection:
SWGTEN
GT2CON
0
0
1
1
0
1
0
1
Connects GATE2 to VCC.
Connects GATE2 to the TMRGATE2 pin.
Turns GATE2 off.
Turns GATE2 on.
4
3
CK2CON
GT1CON
Clock 2 Connection:
0 = Connects CLKIN2 to the internal PSCLK signal.
1 = Connects CLKIN2 to the TMRCLK2 pin.
Gate 1 Connection:
SWGTEN
GT1CON
0
0
1
1
0
1
0
1
Connects GATE1 to VCC.
Connects GATE1 to the TMRGATE1 pin.
Turns GATE1 off.
Turns GATE1 on.
2
1
CK1CON
GT0CON
Clock 1 Connection:
0 = Connects CLKIN1 to the internal PSCLK signal.
1 = Connects CLKIN1 to the TMRCLK1 pin.
Gate 0 Connection:
SWGTEN
GT0CON
0
0
1
1
0
1
0
1
Connects GATE0 to VCC.
Connects GATE0 to the TMRGATE1 pin.
Turns GATE0 off.
Turns GATE0 on.
0
CK0CON
Clock 0 Connection:
0 = Connects CLKIN0 to the internal PSCLK signal.
1 = Connects CLKIN0 to the TMRCLK0 pin.
Figure 5-7. Timer Configuration Register (TMRCFG)
5-13
5.2.4 Asynchronous Serial I/O Configuration
Figures 5-8 and 5-9 show the asynchronous serial I/O unit configuration, consisting of channels
SIO0 and SIO1. Each channel has one output (SIOINT0, SIOINT1) to the interrupt control unit
(see Figure 5-4) and two outputs to the DMA unit. (These signals do not go to package pins.)
SIOINTn is active when any one of the SIO status signals (receiver line status, receiver buffer
full, transmit buffer empty, modem status) is set and enabled. All SIO0 pins are multiplexed with
I/O port signals.
Using SIO1 precludes using DMA channel 1 for external DMA requests due to the multiplexing
of the transmit and receive signals with DMA signals (RXD1/DRQ1, TXD1/DACK1#).
NOTE
Using SIO1 modem signals RTS1#, DSR1#, DTR1#, and RI1# precludes use
of the SSIO unit.
5-14
DEVICE CONFIGURATION
SIOCFG.0
SIO0
P3CFG.7
1
COMCLK
(P3.7)†
0
1
BCLKIN
To/From I/O Port 3
To/From I/O Port 2
SERCLK
0
P2CFG.5
1
Receive Data
RXD0
(P2.5)
SIOINT0
RBFDMA0
To ICU
0
To DMA
To DMA
TXEDMA0
P2CFG.6
1
1
Transmit Data
TXD0
(P2.6)
SIOCFG.6
To/From I/O Port 2
To/From I/O Port 2
0
P2CFG.7
CTS0#
(P2.7)
0
1
Clear to Send
0
P1CFG.1
1
1
Request to Send
Data Set Ready
RTS0#
(P1.1)
To/From I/O Port 1
To/From I/O Port 1
0
P1CFG.3
DSR0#
(P1.3)
0
1
0
P1CFG.0
0
1
DCD0#
(P1.0)
0
1
Data Carrier
Detect
To/From I/O Port 1
P1CFG.2
1
1
Data Terminal
Ready
DTR0#
(P1.2)
To/From I/O Port 1
To/From I/O Port 1
0
P1CFG.4
RI0#
(P1.4)
0
1
Ring Indicator
0
VCC
† Alternate pin signals are in parentheses.
A2521-02
Figure 5-8. Serial I/O Unit 0 Configuration
5-15
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
SIOCFG.1
SIO1
P3CFG.7
0
1
COMCLK
(P3.7)†
0
1
BCLKIN
To/From I/O Port 3
To DMA
SERCLK
Receive Data
RXD1
(DRQ1)
SIOINT1
RBFDMA1
TXEDMA1
To ICU
To DMA
To DMA
PINCFG.2
1
1
Transmit Data
TXD1
(DACK1#)
SIOCFG.7
From DMA
0
PINCFG.3
CTS1#
(EOP#)
0
1
Clear to Send
To/From DMA
0
PINCFG.0
0
1
Request to Send
Data Set Ready
RTS1#
(SSIOTX)
From SSIO
DSR1#
(STXCLK)
0
1
To/From SSIO
DCD1#
(DRQ0)
0
1
Data Carrier
Detect
To DMA
PINCFG.1
0
1
Data Terminal
Ready
DTR1#
(SRXCLK)
To/From SSIO
To SSIO
RI1#
(SSIORX)
0
1
Ring Indicator
VCC
† Alternate pin signals are in parentheses.
A2519-02
Figure 5-9. Serial I/O Unit 1 Configuration
5-16
DEVICE CONFIGURATION
SIO and SSIO Configuration
SIOCFG
Expanded Addr: F836H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
S1M
S0M
—
—
—
SSBSRC
S1BSRC
S0BSRC
Bit
Number
Bit
Mnemonic
Function
SIO1 Modem Signal Connections:
7
S1M
0 = Connects the SIO1 modem input signals to the package pins.
1 = Connects the SIO1 modem input signals internally.
6
S0M
SIO0 Modem Signal Connections:
0 = Connects the SIO0 modem input signals to the package pins.
1 = Connects the SIO0 modem input signals internally.
5–3
2
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
SSBSRC
SSIO Baud-rate Generator Clock Source:
0 = Connects the internal PSCLK signal to the SSIO baud-rate
generator.
1 = Connects the internal SERCLK signal to the SSIO baud-rate
generator.
1
0
S1BSRC
S0BSRC
SIO1 Baud-rate Generator Clock Source:
0 = Connects the COMCLK pin to the SIO1 baud-rate generator.
1 = Connects the internal SERCLK signal to the SIO1 baud-rate
generator.
SIO0 Baud-rate Generator Clock Source:
0 = Connects the COMCLK pin to the SIO0 baud-rate generator.
1 = Connects the internal SERCLK signal to the SIO0 baud-rate
generator.
Figure 5-10. SIO and SSIO Configuration Register (SIOCFG)
5-17
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
5.2.5 Synchronous Serial I/O Configuration
The synchronous serial I/O unit (SSIO) is shown in Figure 5-11. Its single configuration register
bit is in the SIOCFG register (Figure 5-10). The transmit buffer empty and receive buffer full sig-
nals (SSTBE and SSRBF) go to the DMA unit (Figure 5-2), and an interrupt signal (SSIOINT)
goes to the ICU (Figure 5-4). Depending on the settings in the SSIOCON1 register (see Chapter
13), SSIOINT is asserted for one of two conditions: the receive buffer is full or the transmit buffer
is empty. Note that using the SSIO signals precludes the use of four of the SIO1 modem signals.
SIOCFG.2
SSIO
PSCLK
0
1
BCLKIN
SERCLK
SSTBE
SSRBF
To DMA
To DMA
To ICU
SSIOINT
Receive Data
SSIORX
(RI1#)*
To SIO1
PINCFG.0
1
0
Transmit Data
Transmit Clock
SSIOTX
(RTS1#)
From SIO1
STXCLK
(DSR1#)
To SIO1
PINCFG.1
1
0
Receive Clock
SRXCLK
(DTR1#)
From SSIO1
*Alternate pin signals are in parentheses.
A2518-02
Figure 5-11. SSIO Unit Configuration
5-18
DEVICE CONFIGURATION
Management Unit.
Refresh Control Unit and the DACK0# signal from the DMA Unit, respectively. Bits 6 and 4 in
the PINCFG register (see Figure 5-15) control these multiplexers. CS3#, CS2#, CS1# and CS0#
are multiplexed with I/O Port 2 signals, P2.3, P2.2, P2.1 and P2.0, respectively. Bits 4:0 in the
P2CFG register (see Figure 5-17) control these multiplexers.
The PWRDOWN output signal of the Clock and Power Management Unit is multiplexed with
I/O Port 3 signal, P3.6. Bit 6 in the P3CFG register (see Figure 5-18) controls this multiplexer.
5-19
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
CSU
P2CFG.0
1
1
1
CSO#
(P2.0)
CS0#
CS1#
CS2#
CS3#
CS4#
CS5#
CS6#
To/From I/O Port 2
To/From I/O Port 2
To/From I/O Port 2
0
P2CFG.1
CS1#
(P2.1)
0
P2CFG.2
CS2#
(P2.2)
0
P2CFG.3
1
1
1
CS3#
(P2.3)
To/From I/O Port 2
To/From I/O Port 2
DACK0# (DMA)
0
P2CFG.4
CS4#
(P2.4)
0
PINCFG.4
CS5#
(DACK0#)
0
PINCFG.6
1
CS6#
(REFRESH#)
REFRESH# (RCU)
0
Clock and
Power
Management
Unit
P3CFG.6
0
1
PWRDOWN
(P3.6)
PWRDOWN
To/From I/O Port 3
A3380-01
Figure 5-12. Configuration of Chip-select Unit and Clock and Power Management Unit
5-20
DEVICE CONFIGURATION
5.2.7 Core Configuration
Three coprocessor signals (ERROR#, PEREQ, and BUSY# in Figure 5-13) can be routed to the
core, as determined by bit 5 of the PINCFG register (see Figure 5-15). Due to signal multiplexing
at the pins, the coprocessor and Timer/counter2 cannot be used simultaneously.
PINCFG.5
Core
PINCFG.5
0
ERROR#
0
1
ERROR#
PEREQ
(TMROUT2)†
From TCU
VCC
1
0
PEREQ
(TMRCLK2)
0
1
To TCU
VSS
1
0
BUSY#
0
1
BUSY#
RESET
(TMRGATE2)
To TCU
VCC
1
RESET Timing
Generation
PORT92.0
From Chip RESET Pin
PORT92.1
To Chip-select Unit
and A20 Pin
A20
P1CFG.5
1
LOCK#
LOCK#
(P1.5)
To/From I/O Port 1
0
† Alternate pin signals are in parentheses.
A2520-02
Figure 5-13. Core Configuration
5-21
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Setting bit 0 in the PORT92 register (see Figure 5-14) resets the core without resetting the periph-
erals. Unlike the RESET pin, which is asynchronous and can be used to synchronize internal
clocks to CLK2, this core-only reset is synchronized with the on-chip clocks and does not affect
the on-chip clock synchronization. After the CPU-RESET this bit is still set to 1. It must be
cleared and then set to cause another core-only reset.
Clearing bit 1 in the PORT92 register forces address line A20 to 0. This bit affects only addresses
generated by the core; addresses generated by the DMA and the refresh control unit are not af-
fected.
Port 92 Configuration
PORT92
Expanded Addr: F092H
ISA Addr:
0092H
(read/write)
Reset State:
XXXXXX10B
7
0
—
—
—
—
—
—
A20G
CPURST
Bit
Number
Bit
Mnemonic
Function
7–2
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
1
0
A20G
A20 Grounded:
0 = Clearing this bit forces address line A20 to 0. This bit affects
addresses generated only by the core. Addresses generated by the
DMA and the Refresh Unit are not affected by this bit.
1 = Setting this bit leaves core-generated addresses unmodified.
CPURST
CPU Reset:
0 = Clearing this bit performs no operation.
1 = Setting this bit resets the core without resetting the peripherals.
This bit must be cleared before issuing another reset.
Figure 5-14. Port 92 Configuration Register (PORT92)
5-22
DEVICE CONFIGURATION
5.3 PIN CONFIGURATION
are routed to two peripheral inputs without the use of a multiplexer. These input-signal pairs are
listed in Table 5-3. The pin is connected to both peripheral inputs.
The remaining pins supporting two signals have multiplexers. For each such pin, a bit in a pin
configuration register enables one of the signals. Table 5-9 lists the bits in each of the four pin
configuration registers. These abbreviated register tables are discussed in “Configuration Exam-
ple” on page 5-28.
When configuring ports to use INT8 or INT9, first set the appropriate INTCFG bit, then the
P3CFG bit. Setting the bits in this order avoids any potential contention on INT8 or INT9.
Table 5-3. Signal Pairs on Pins without a Multiplexer
Names
DRQ0/
Signal Descriptions
DMA External Request 0 indicates that an off-chip peripheral requires DMA service.
DCD1#
Data Carrier Detect SIO1 indicates that the modem or data set has detected the
asynchronous serial channel’s data carrier.
DRQ1/
RXD1
DMA External Request1 indicates that an off-chip peripheral requires DMA service.
Receive Data SIO1 accepts serial data from the modem or data set to the
asynchronous serial channel SIO1.
DSR1#/
Data Set Ready SIO1 indicates that the modem or data set is ready to establish a
STXCLK
communication link with asynchronous serial channel SIO1.
SSIO Transmit Clock synchronizes data being sent by the synchronous serial port.
RI1#/
Ring Indicator SIO1 indicates that the modem or data set has received a telephone
SSIORX
ringing signal.
SSIO Receive Serial Data accepts serial data (most-significant bit first) being sent to
the synchronous serial port.
TMRCLK0/
INT4
Timer/Counter0 Clock Input can serve as an external clock input for timer/counter0.
(The timer/counters can also be clocked internally.)
Interrupt 4 is an undedicated external interrupt.
TMRGATE0/
INT5
Timer/Counter0 Gate Input can control timer/counter0’s counting (enable, disable, or
trigger, depending on the programmed mode).
Interrupt 5 is an undedicated external interrupt.
TMRCLK1/
INT6
Timer/Counter1 Clock Input can serve as an external clock input for timer/counter1.
(The timer/counters can also be clocked internally.)
Interrupt 6 is an undedicated external interrupt.
TMRGATE1/
INT7
Timer/Counter1 Gate Input can control timer/counter1’s counting (enable, disable, or
trigger, depending on the programmed mode).
Interrupt 7 is an undedicated external interrupt.
5-23
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Pin Configuration
PINCFG
Expanded Addr: F826H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
—
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7
—
Reserved. This bit is undefined; for compatibility with future devices, do
not modify this bit.
6
5
PM6
PM5
Pin Mode:
0 = Selects CS6# at the package pin.
1 = Selects REFRESH# at the package pin.
Pin Mode:
0 = Selects the coprocessor signals, PEREQ, BUSY#, and ERROR#, at
the package pins.
1 = Selects the timer control unit signals, TMROUT2, TMRCLK2, and
TMRGATE2, at the package pins.
4
3
2
1
0
PM4
PM3
PM2
PM1
PM0
Pin Mode:
0 = Selects DACK0# at the package pin.
1 = Selects CS5# at the package pin.
Pin Mode:
0 = Selects EOP# at the package pin.
1 = Selects CTS1# at the package pin.
Pin Mode:
0 = Selects DACK1# at the package pin.
1 = Selects TXD1 at the package pin.
Pin Mode:
0 = Selects SRXCLK at the package pin.
1 = Selects DTR1# at the package pin.
Pin Mode:
0 = Selects SSIOTX at the package pin.
1 = Selects RTS1# at the package pin.
Figure 5-15. Pin Configuration Register (PINCFG)
5-24
DEVICE CONFIGURATION
Port 1 Configuration
P1CFG
Expanded Addr: F820H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
PM7
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7
PM7
Pin Mode:
0 = Selects P1.7 at the package pin.
1 = Selects HLDA at the package pin.
6
5
4
3
2
1
0
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Pin Mode:
0 = Selects P1.6 at the package pin.
1 = Selects HOLD at the package pin.
Pin Mode:
0 = Selects P1.5 at the package pin.
1 = Selects LOCK# at the package pin.
Pin Mode:
0 = Selects P1.4 at the package pin.
1 = Selects RI0# at the package pin.
Pin Mode:
0 = Selects P1.3 at the package pin.
1 = Selects DSR0# at the package pin.
Pin Mode:
0 = Selects P1.2 at the package pin.
1 = Selects DTR0# at the package pin.
Pin Mode:
0 = Selects P1.1 at the package pin.
1 = Selects RTS0# at the package pin.
Pin Mode:
0 = Selects P1.0 at the package pin.
1 = Selects DCD0# at the package pin.
Figure 5-16. Port 1 Configuration Register (P1CFG)
5-25
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Port 2 Configuration
P2CFG
Expanded Addr: F822H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
PM7
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7
PM7
Pin Mode:
0 = Selects P2.7 at the package pin.
1 = Selects CTS0# at the package pin.
6
5
4
3
2
1
0
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Pin Mode:
0 = Selects P2.6 at the package pin.
1 = Selects TXD0 at the package pin.
Pin Mode:
0 = Selects P2.5 at the package pin.
1 = Selects RXD0 at the package pin.
Pin Mode:
0 = Selects P2.4 at the package pin.
1 = Selects CS4# at the package pin.
Pin Mode:
0 = Selects P2.3 at the package pin.
1 = Selects CS3# at the package pin.
Pin Mode:
0 = Selects P2.2 at the package pin.
1 = Selects CS2# at the package pin.
Pin Mode:
0 = Selects P2.1 at the package pin.
1 = Selects CS1# at the package pin.
Pin Mode:
0 = Selects P2.0 at the package pin.
1 = Selects CS0# at the package pin.
Figure 5-17. Port 2 Configuration Register (P2CFG)
5-26
DEVICE CONFIGURATION
Port 3 Configuration
P3CFG
Expanded Addr: F824H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
PM7
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7
PM7
Pin Mode:
0 = Selects P3.7 at the package pin.
1 = Selects COMCLK at the package pin.
6
5
4
3
2
PM6
PM5
PM4
PM3
PM2
Pin Mode:
0 = Selects P3.6 at the package pin.
1 = Selects PWRDOWN at the package pin.
Pin Mode:
0 = Selects P3.5 at the package pin.
1 = Connects master IR7 to the package pin (INT3).
Pin Mode:
0 = Selects P3.4 at the package pin.
1 = Connects master IR6 to the package pin (INT2).
Pin Mode:
0 = Selects P3.3 at the package pin.
1 = Connects master IR5 to the package pin (INT1).
Pin Mode:
0 = Selects P3.2 at the package pin.
1 = Connects master IR1 to the package pin (INT0).
1
0
PM1
PM0
Pin Mode:
See Table 5-1 on page 5-8 for all the PM1 configuration options.
Pin Mode:
See Table 5-1 on page 5-8 for all the PM0 configuration options.
Figure 5-18. Port 3 Configuration Register (P3CFG)
5-27
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
5.4 DEVICE CONFIGURATION PROCEDURE
Before configuring the microprocessor, make the following selections:
• The set of peripherals to be used
• The signals to be available at the package pins
• The desired peripheral-peripheral and peripheral-core connections
Although final decisions regarding these selections may be influenced by the possible configura-
tions, we recommend that you initially make the selections without regard to limitations on the
configurations.
We suggest the following procedure for configuring the device for your design. An aide for re-
cording the steps in the procedure and an example configuration are given in “Configuration Ex-
ample” on page 5-28.
1. Pin Configuration. For each desired pin signal, consult the peripheral configuration
diagram to find the bit value in the pin configuration register that connects the signal to a
device pin. When the signal shares a pin that has no multiplexer, make a note of its
companion signal.
2. Peripheral Configuration. For each peripheral in your design, consult the peripheral
configuration diagram and the peripheral configuration register to find the bit values for
your desired internal connections.
3. Configuration Review. Review the results of steps 1 and 2 to see if the configuration
registers have conflicting bit values. If conflicts exist, follow steps 3a and 3b.
a. Attempt to resolve the pin configuration conflicts first. In some cases you may find that
using a different peripheral channel resolves the conflict.
b. Attempt to resolve peripheral configuration conflicts.
If conflicts remain, consider peripheral substitutions.
5.5 CONFIGURATION EXAMPLE
This section presents an example of a PC/AT*-compatible configuration. The last set of tables are
blank; you can use them as worksheets as you follow the steps in the configuration process.
5.5.1 Example Design Requirements
The example is a PC/AT-compatible design with the following requirements:
• Interrupt Control Unit:
— External interrupt inputs available at package pins: INT1:0, INT7:4
• Timer Control Unit:
— Counters 0, 1: Clock input is on-chip programmable clock (PSCLK); no signals
connected externally.
5-28
DEVICE CONFIGURATION
— Counter 2: Clock input is on-chip programmable clock (PSCLK); no signals connected
to package pins
• DMA Unit:
— Not Used
• Asynchronous Serial I/O channel 0 (SIO0):
— Clock input is the internal clock SERCLK
— RXD0, TXD0 connected to package pins
— Modem Signals connected internally.
• Asynchronous Serial I/O channel 1 (SIO1):
— Clock input is the internal clock SERCLK
— Modem signals externally connected
• Synchronous Serial I/O (SSIO):
— Not Used
• Chip Select:
— Chip select signals CS6#, CS5:1#, UCS# connected to package pins
• Core and Bus Arbiter:
— Coprocessor signals connected to package pins
— HOLD and HLDA not connected to package pins
5.5.2 Example Design Solution
The configuration register bit values for the example design are recorded in the following abbre-
viated register tables. Blank worksheets are provided for you to use when designing your system.
Table 5-4 summarizes the bit selections you would need to make in the pin configuration registers
to implement the example design. Tables 5-5 through 5-8 summarize the bit selections you would
make in the peripheral configuration registers.
5-29
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Bit #
P1CFG
Value
0
Bit #
P2CFG
Value
0
Bit #
P3CFG
0 = P3.7
Value
0
7
0 = P1.7
7
0 = P2.7
1 = CTS0#
0 = P2.6
1 = TXD0
0 = P2.5
1 = RXD0
0 = P2.4
1 = CS4#
0 = P2.3
1 = CS3#
0 = P2.2
1 = CS2#
0 = P2.1
1 = CS1#
0 = P2.0
1 = CS0#
7
1 = HLDA
0 = P1.6
1 = COMCLK
0 = P3.6
6
5
4
3
2
1
0
0
0
0
0
0
0
0
6
5
4
3
2
1
0
1
1
1
1
1
1
0
6
5
4
3
2
1
0
0
0
0
1
1
0
0
1 = HOLD
0 = P1.5
1 = PWRDOWN
0 = P3.5
1 = LOCK#
0 = P1.4
1 = INT3
0 = P3.4
1 = RIO#
0 = P1.3
1 = INT2
0 = P3.3
1 = DSR0#
0 = P1.2
1 = INT1
0 = P3.2
1 = DTR0#
0 = P1.1
1 = INT0
0 = P3.1
1 = RTS0#
0 = P1.0
1 = mux
0 = P3.0
1 = DCD0#
1 = mux
Bit #
PINCFG
Value
Pins w/o Muxes
X
Pins w/o Muxes
X
7
6
Reserved
R
0
DRQ0
X
TMRCLK0
INT4
0 = CS6#
DCD1#
DRQ1
X
X
X
X
1 = REFRESH#
TMRGATE0
INT5
5
4
3
2
1
0
0 = Coprocessor Sigs.1
1 = TMR2 Signals2
0 = DACK0#
1 = CS5#
0
1
1
1
1
1
RXD1
X
X
X
DSR1#
STXCLK
RI1#
TMRCLK1
INT6
TMRGATE1
INT7
0 = EOP#
SSIORX
1 = CTS1#
0 = DACK1#
1 = TXD1
NOTES:
0 = SRXCLK
1 = DTR1#
1 PEREQ, BUSY#, ERROR#
2 TMROUT2, TMRCLK2, TMRGATE2
0 = SSIOTX
1 = RTS1#
Table 5-4. Example Pin Configuration Registers
5-30
DEVICE CONFIGURATION
Bit #
DMACFG
Value
1
7
0 = Enables DACK1# at chip pin
1 = Disables DACK1# at chip pin
6–4
000 = DRQ1 pin (external peripheral) connected to DREQ1
000
001 = SIO channel 1’s receive buffer full signal (RBFDMA1) connected to DREQ1
010 = SIO channel 0’s transmit buffer empty signal (TXEDMA0) to DREQ1
011 =SSIO receive holding buffer full signal (SSRBF) to DREQ1
100 = TCU counter 2’s output signal (OUT2) to DREQ1
101 = SIO channel 0’s receive buffer full signal (RBFDMA0) to DREQ1
110 = SIO channel 1’s transmit buffer empty signal (TXEDMA1) to DREQ1
111 = SSIO transmit holding buffer empty signal (SSTBE) to DREQ1
0 = Enables DACK0# at chip pin
3
1
1 = Disables DACK0# at chip pin
2–0
000 = DRQ0 pin (external peripheral) connected to DREQ0
000
001 = SIO channel 0’s receive buffer full signal (RBFDMA0) connected to DREQ0
010 = SIO channel 1’s transmit buffer empty signal (TXEDMA1) connected to DREQ0
011 = SSIO transmit holding buffer empty signal (THBE) connected to DREQ0
100 = TCU counter 1’s output signal (OUT1) connected to DREQ0
101 = SIO channel 1’s receive buffer full signal (RBFDMA1) connected to DREQ0
110 = SIO channel 0’s transmit buffer empty signal (TXEDMA0) connected to DREQ0
111 = SSIO receive holding buffer full signal (RHBF) connected to DREQ0
Table 5-5. Example DMACFG Configuration Register
5-31
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Bit #
TMRCFG
Value
0
7
0 = All clock inputs enabled
1 = CLK2, CLK1, CLK0 forced to 0
6
5
4
3
2
1
0
0 = Connects GATEn to either the VCC pin or the TMRGATEn pin
1 = Turns GATEn on or off, depending on whether bits 1, 3, and 5 are set or clear
0 = With bit 6 clear: VCC to GATE2; with bit 6 set: GATE2 off.
1 = With bit 6 clear: TMRGATE2 pin conn. to GATE2; with bit 6 set: GATE2 on.
0 = PSCLK connected to CLK2
0
0
0
0
0
0
0
1 = TMRCLK2 connected to CLK2
0 = With bit 6 clear: VCC to GATE1; with bit 6 set: GATE1 turned off.
1 = With bit 6 clear: TMRGATE1 pin conn. to GATE1; with bit 6 set: GATE1 on.
0 = PSCLK connected to CLK1
1 = TMRCLK1 connected to CLK1
0 = With bit 6 clear: VCC to GATE0; with bit 6 set: GATE0 turned off.
1 = With bit 6 clear: TMRGATE0 pin conn. to GATE0; with bit 6 set: GATE0 on.
0 = PSCLK connected to CLK0
1 = TMRCLK0 connected to CLK0
Table 5-6. Example TMRCFG Configuration Register
5-32
DEVICE CONFIGURATION
Bit #
INTCFG
Value
0
7
0 = CAS2:0 disabled to pins
1 = CAS2:0 enabled from pins
6
5
4
3
2
1
0
0 = SIOINT1 connected to master IR3
1 = P3.1 connected to IR3
0
0
0 = SIOINT0 connected to master IR4
1 = P3.0 connected to IR4
0 = DMAINT connected to slave IR4. INT6 connected to slave IR5.
1 = INT6 connected to slave IR4. DMAINT connected to slave IR5.
0 = VSS connected to slave IR6
1
1
1
1
1
1 = INT7 connected to slave IR6
0 = VSS connected to slave IR5
1 = INT6 connected to slave IR5
0 = SSIO Interrupt to slave IR1
1 = INT5 connected to slave IR1
0 = VSS connected to slave IR0
1 = INT4 connected to slave IR0
Table 5-7. Example INTCFG Configuration Register
SIOCFG
7
6
0 = SIO1 modem sigs. conn. to pin muxes
1 = SIO1 modem signals internal
1
0
0 = SIO0 modem sigs. conn. to pin muxes
1 = SIO0 modem signals internal
5–3
2
Reserved
R
1
0 = PSCLK connected to SSIO BLKIN
1 = SERCLK connected to SSIO BCLKIN
0 = COMCLK connected to SIO1 BCLKIN
1 = SERCLK connected to SIO1 BCLKIN
0 = COMCLK connected to SIO0 BCLKIN
1 = SERCLK connected to SIO0 BCLKIN
1
0
0
0
Table 5-8. Example SIOCFG Configuration Register
5-33
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Bit #
P1CFG
Value
Bit #
P2CFG
Value
Bit #
P3CFG
0 = P3.7
Value
7
0 = P1.7
7
0 = P2.7
1 = CTS0#
0 = P2.6
1 = TXD0
0 = P2.5
1 = RXD0
0 = P2.4
1 = CS4#
0 = P2.3
1 = CS3#
0 = P2.2
1 = CS2#
0 = P2.1
1 = CS1#
0 = P2.0
1 = CS0#
7
1 = HLDA
0 = P1.6
1 = COMCLK
0 = P3.6
6
5
4
3
2
1
0
6
5
4
3
2
1
0
6
5
4
3
2
1
0
1 = HOLD
0 = P1.5
1 = PWRDOWN
0 = P3.5
1 = LOCK#
0 = P1.4
1 = INT3
0 = P3.4
1 = RIO#
0 = P1.3
1 = INT2
0 = P3.3
1 = DSR0#
0 = P1.2
1 = INT1
0 = P3.2
1 = DTR0#
0 = P1.1
1 = INT0
0 = P3.1
1 = RTS0#
0 = P1.0
1 = mux
0 = P3.0
1 = DCD0#
1 = mux
Bit #
PINCFG
Value
Pins w/o Muxes
X
Pins w/o Muxes
X
7
6
Reserved
DRQ0
TMRCLK0
INT4
0 = CS6#
DCD1#
DRQ1
1 = REFRESH#
TMRGATE0
INT5
5
4
3
2
1
0
0 = Coprocessor Sigs.1
1 = TMR2 Signals2
0 = DACK0#
1 = CS5#
RXD1
DSR1#
STXCLK
RI1#
TMRCLK1
INT6
TMRGATE1
INT7
0 = EOP#
SSIORX
1 = CTS1#
0 = DACK1#
1 = TXD1
NOTES:
0 = SRXCLK
1 = DTR1#
1 PEREQ, BUSY#, ERROR#
2 TMROUT2, TMRCLK2, TMRGATE2
0 = SSIOTX
1 = RTS1#
Table 5-9. Pin Configuration Register Design Woksheet
5-34
DEVICE CONFIGURATION
Bit #
DMACFG
Value
7
0 = Enables DACK1# at chip pin
1 = Disables DACK1# at chip pin
6–4
000 = DRQ1 pin (external peripheral) connected to DREQ1
001 = SIO channel 1’s receive buffer full signal (RBFDMA1) connected to DREQ1
010 = SIO channel 0’s transmit buffer empty signal (TXEDMA0) to DREQ1
011 =SSIO receive holding buffer full signal (SSRBF) to DREQ1
100 = TCU counter 2’s output signal (OUT2) to DREQ1
101 = SIO channel 0’s receive buffer full signal (RBFDMA0) to DREQ1
110 = SIO channel 1’s transmit buffer empty signal (TXEDMA1) to DREQ1
111 = SSIO transmit holding buffer empty signal (SSTBE) to DREQ1
0 = Enables DACK0# at chip pin
3
1 = Disables DACK0# at chip pin
2–0
000 = DRQ0 pin (external peripheral) connected to DREQ0
001 = SIO channel 0’s receive buffer full signal (RBFDMA0) connected to DREQ0
010 = SIO channel 1’s transmit buffer empty signal (TXEDMA1) connected to DREQ0
011 = SSIO transmit holding buffer empty signal (THBE) connected to DREQ0
100 = TCU counter 1’s output signal (OUT1) connected to DREQ0
101 = SIO channel 1’s receive buffer full signal (RBFDMA1) connected to DREQ0
110 = SIO channel 0’s transmit buffer empty signal (TXEDMA0) connected to DREQ0
111 = SSIO receive holding buffer full signal (RHBF) connected to DREQ0
Table 5-10. DMACFG Register Design Worksheet
5-35
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Bit #
TMRCFG
Value
7
0 = All clock inputs enabled
1 = CLK2, CLK1, CLK0 forced to 0
6
5
4
3
2
1
0
0 = Connects GATEn to either the VCC pin or the TMRGATEn pin.
1 = Turns GATEn on or off, depending on whether bits 1, 3, and 5 are set or clear.
0 = With bit 6 clear: VCC to GATE2; with bit 6 set: GATE2 off.
1 = With bit 6 clear: TMRGATE2 pin conn. to GATE2; with bit 6 set: GATE2 on.
0 = PSCLK connected to CLK2
1 = TMRCLK2 connected to CLK2
0 = With bit 6 clear: VCC to GATE1; with bit 6 set: GATE1 turned off.
1 = With bit 6 clear: TMRGATE1 pin conn. to GATE1; with bit 6 set: GATE1 on.
0 = PSCLK connected to CLK1
1 = TMRCLK1 connected to CLK1
0 = With bit 6 clear: VCC to GATE0; with bit 6 set: GATE0 turned off.
1 = With bit 6 clear: TMRGATE0 pin conn. to GATE0; with bit 6 set: GATE0 on.
0 = PSCLK connected to CLK0
1 = TMRCLK0 connected to CLK0
Table 5-11. TMRCFG Register Design Worksheet
5-36
DEVICE CONFIGURATION
Bit #
INTCFG
Value
7
0 = CAS2:0 disabled to pins
1 = CAS2:0 enabled from pins
6
5
4
3
2
1
0
0 = SIOINT1 connected to master IR3
1 = P3.1 connected to IR3
0 = SIOINT0 connected to master IR4
1 = P3.0 connected to IR4
0 = DMAINT connected to slave IR4. INT6 connected to slave IR5.
1 = INT6 connected to slave IR4. DMAINT connected to slave IR5.
0 = VSS connected to slave IR6
1 = INT7 connected to slave IR6
0 = VSS connected to slave IR5
1 = INT6 connected to slave IR5
0 = SSIO Interrupt to slave IR1
1 = INT5 connected to slave IR1
0 = VSS connected to slave IR0
1 = INT4 connected to slave IR0
Table 5-12. INTCFG Register Design Worksheet
SIOCFG
7
6
0 = SIO1 modem sigs. conn. to pin muxes
1 = SIO1 modem signals internal
0 = SIO0 modem sigs. conn. to pin muxes
1 = SIO0 modem signals internal
5–3
2
Reserved
0 = PSCLK connected to SSIO BLKIN
1 = SERCLK connected to SSIO BCLKIN
0 = COMCLK connected to SIO1 BCLKIN
1 = SERCLK connected to SIO1 BCLKIN
0 = COMCLK connected to SIO0 BCLKIN
1 = SERCLK connected to SIO0 BCLKIN
1
0
Table 5-13. SIOCFG Register Design Worksheet
5-37
6
BUS INTERFACE
UNIT
CHAPTER 6
BUS INTERFACE UNIT
The processor communicates with memory, I/O, and other devices through bus operations. Ad-
dress, data, status, and control information define a bus cycle. The Bus Interface Unit supports
execute memory read/write cycles, I/O read/write cycles, interrupt acknowledge cycles, refresh
cycles and processor halt/shutdown cycles.
This chapter is organized as follows:
• Bus Cycles (page 6-13)
• Bus Lock (page 6-34)
• External Bus Master Support (Using HOLD, HLDA) (page 6-35)
• Design Considerations (page 6-38)
6.1 OVERVIEW
The Intel386™ EX processor’s external bus is controlled by the bus interface unit (BIU). To com-
municate with memory and I/O, the external bus consists of a data bus, a separate address bus,
seven bus status pins, two data status pins, and three control pins.
• Bidirectional data bus (D15:0) can transfer 8 or 16 bits of data per cycle.
• Address bus includes the address pins (A25:1), a high-byte-enable pin (BHE#), and a low-
byte-enable pin (BLE#). Address pins select a word in memory, and byte-enable pins select
the byte within the word to access.
• Bus status pins include:
— ADS# indicates the start of a bus cycle and valid address bus outputs.
— W/R# identifies the bus cycle as a write or a read.
— M/IO# identifies the bus cycle as a memory or I/O access.
— D/C# identifies the bus cycle as a data or control cycle.
— LOCK# identifies a locked bus cycle.
— LBA# indicates that the processor generates an internal READY# for the current bus
cycle.
— REFRESH# identifies a refresh bus cycle.
6-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
• Data status pins indicate that data is available on the data bus for a write (WR#) or that the
processor is ready to accept data for a read (RD#). These pins are available so that certain
system configurations can easily connect the processor directly to memory or I/O without
external logic.
• Bus control pins allow external logic to control the bus cycle on a cycle-by-cycle basis:
— READY# indicates that internal logic has completed the current bus cycle or that
external hardware has terminated it.
— NA# requests the next address to be put on the bus during a pipelined bus cycle.
— BS8# indicates that the current bus transaction is for an 8-bit data bus.
The remaining external bus pins interface to external bus masters and external logic for transfer-
ring control of the bus.
• An external bus master activates the HOLD pin to request the external bus.
— The internal bus arbiter arbitrates between the HOLD input and other potential requests
(DMA Units 0 and 1, Refresh Control Unit) based on their priorities.
— When another unit has control of the bus, the bus is released to the external bus master
based on the arbiter’s arbitration scheme (refer to “Bus Control Arbitration” on page
12-9 for information on internal bus masters also controlled by the internal bus arbiter
and the arbitration protocol used by the arbiter).
— When the core has control of the bus, the arbiter passes the request on to the core by
asserting the core HOLD signal.
— The core finishes the current nonlocked bus transfer and releases the bus signals.
— The core asserts the core HLDA signal to indicate that the bus has been released.
— The arbiter then asserts the HLDA pin to indicate to the external bus master that the bus
has been released.
6-2
BUS INTERFACE UNIT
6.1.1 Bus Signal Descriptions
Table 6-1 describes the signals associated with the BIU.
Table 6-1. Bus Interface Unit Signals (Sheet 1 of 2)
Device Pin or
Internal Signal
only
Signal
Description
A25:1
Device pins
Address Bus:
Outputs physical memory or I/O addresses. These signals are valid
when ADS# is active and remain valid until the next T1, T2P, or Ti.
ADS#
Device pin
Address Strobe:
Indicates that the processor is driving a valid bus-cycle definition and
address. (The processor is driving W/R#, D/C#, M/IO#, WR#, RD#,
UCS#, CS6:0#, LOCK#, REFRESH#, A25:1, BHE#, and BLE# on its
pins.)
BHE#
BLE#
Device pins
Byte Enable Outputs:
Indicates which byte of the 16-bit data bus of the processor is being
transferred.
BHE# BLE# Output
0
0
1
1
0
1
0
1
word transfer
upper byte (D15:8) transfer
lower byte (D7:0) transfer
refresh cycle
BS8#
D15:0
Device pin
Bus Size:
Indicates that the currently addressed device is an 8-bit device.
Data Bus:
Device pins
Inputs data during memory read, I/O read, and interrupt acknowledge
cycles; outputs data during memory write and I/O write cycles. During
reads, data is latched at the falling edge of phase 2 (coincides with
rising edge of PH1) of T2, T2P, or T2i when READY# is sampled
active. During writes, this bus is driven during phase 2 of T1 and T1P
and remains active until phase 2 of the next T1, T1P, or Ti.
LBA#
Device pin
Local Bus Access:
Indicates that the processor provides the READY# signal internally to
terminate a bus transaction. This signal is active when the processor
accesses an internal peripheral or when the chip-select unit generates
the READY# signal for accesses to an external peripheral. LBA# is
also active when internal READY# generation is enabled for
Halt/Shutdown cycles and the Watchdog Timer Unit’s Bus Monitor
Mode timeouts.
The LBA# signal goes active in the first T2 state and stays active until
the first T2, T2i or T2P state of the next cycle that does not have
internal READY# generation.
LOCK#
Device pin
Bus Lock:
Prevents other bus masters from gaining control of the system bus.
6-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table 6-1. Bus Interface Unit Signals (Sheet 2 of 2)
Device Pin or
Signal
Internal Signal
Description
only
M/IO#
D/C#
Device pins
Bus Cycle Definition Signals (Memory/IO, Data/Control, Write/Read,
and Refresh):
W/R#
REFRESH#
These four status outputs define the current bus cycle type, as shown
in Table 6-2.
NA#
Device pin
Device pin
Next Address:
Requests address pipelining.
RD#
Read Enable:
Indicates that the current bus cycle is a read cycle and the data bus is
able to accept data.
READY#
Device pin
Ready:
This bidirectional pin is used to terminate the current bus cycle. The
processor drives READY# when LBA# is active. The processor
samples the READY# pin at the falling edge of Phase 2 of T2, T2P or
T2i.
The READY# signal is also used to deassert the WR# signal (Refer to
“Write Cycle” on page 6-16).
WR#
Device pin
Write Enable:
Indicates that the current bus cycle is a write cycle and valid data is
present on the data bus.
6-4
BUS INTERFACE UNIT
6.2 BUS OPERATION
The processor generates eight different types of bus operations:
• Memory data read (data fetch)
• Memory data write
• Memory code read (instruction fetch)
• I/O data read (data fetch)
• I/O data write
• Halt or shutdown
• Refresh
• Interrupt acknowledge
These operations are defined by the states of four bus status pins: M/IO#, D/C#, W/R# and RE-
FRESH#. Table 6-2 lists the various combinations and their definitions.
Table 6-2. Bus Status Definitions
M/IO#
D/C#
W/R#
REFRESH#
Bus Operation
0
0
0
0
1
1
1
1
1
0
0
1
1
0
0
1
1
1
0
1
0
1
0
1
0
0
1
1
1
1
1
1
1
0
1
1
interrupt acknowledge cycle
never occurs
I/O data read
I/O data write
memory code read
halt or shutdown cycle*
refresh cycle
memory data read
memory data write
*The byte address is 2 for a halt and 0 for a shutdown. For both conditions, BHE# is high and BLE# is low.
6-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Cycle 1
Cycle 2
Idle
Cycle 3
Cycle 4
Nonpipelined Nonpipelined Cycle Nonpipelined Nonpipelined
External
(Write)
External
(Read)
External
(Write)
External
(Read)
[Late Ready]
[Late Ready]
State
CLK2
T1
T2
T1
T2
Ti
T1
T2
T1
T2
CLKOUT
A25:1, BHE#
BLE#, D/C#
M/IO#
Valid 1
Valid 2
Valid 3
Valid 4
REFRESH#
W/R#
WR#
RD#
ADS#
NA#
READY#
LBA#
BS8#
LOCK#
Valid 1
Valid 2
Valid 3
Valid 4
Out 1
In 2
Out 3
In 4
D15:0
A2305-02
Figure 6-1. Basic External Bus Cycles
6-6
BUS INTERFACE UNIT
6.2.1 Bus States
The processor uses a double-frequency clock input (CLK2). This clock is internally divided by
two and synchronized to the falling edge of RESET (see Figure 8-2 in Chapter 8) to generate the
internal processor clock signal. Each processor clock cycle is two CLK2 cycles wide.
Each bus cycle is composed of at least two bus states: T1 and T2. Each bus state in turn consists
of two CLK2 cycles, which can be thought of as Phase 1 (PH1) and Phase 2 (PH2) of the bus state.
External circuitry can use the CLKOUT signal (generated by the processor) to synchronize itself
with the processor. This signal is a replica of the PH1P clock, which is the PH1 clock that is used
by the internal peripherals. (For more information, refer to Chapter 8, “CLOCK AND POWER
MANAGEMENT UNIT.”) The CLKOUT signal is used as a phase status indicator for external
circuitry. All device inputs are sampled and outputs are activated at CLK2 rising edges. This
makes synchronous circuit design easy through the use of rising-edge-triggered, registered logic
(such as PALs, PLDs and EPLDs).
Many signals are sampled by the processor on every other CLK2 rising edge: some are sampled
on the CLK2 edge when CLKOUT is going high, while others are sampled on the CLK2 edge
when PH1 is going low.
The maximum data transfer rate for a bus operation is 16 bits for every two processor clock cycles
(two CLKOUT cycles).
During the first bus state (T1), address and bus status pins go active. During the second bus state
(T2), external logic and devices respond.
• When the READY# input is sampled low at the falling edge of PH2 in T2, the bus cycle
terminates.
• When READY# is high when sampled, the bus cycle continues for an additional T2 state,
called a wait-state, and READY# is sampled again. This process continues until READY#
Wait-states are added until READY# is sampled low. READY# is sampled externally when the
LBA# signal is inactive. When the LBA# signal is active, the processor is generating the
READY# signal internally. READY# can be generated internally by either an internal peripheral
or the chip-select unit’s wait-state generator. When no bus cycles are needed (no bus requests are
pending), the processor remains in the idle bus state, Ti. The relationship between T1, T2, and Ti
is shown in Figure 6-2.
From an idle bus, the processor begins a bus cycle by first driving a valid address and bus cycle
status onto the address and status buses. Hardware can distinguish the difference between an idle
cycle and an active bus cycle by the address strobe (ADS#) signal being driven active. The ADS#
signal remains active for only the first T-state of the bus cycle, while the address signals and status
signals remain active until the bus cycle is terminated by an active READY# signal or the bus
cycle is pipelined. Pipelined bus cycles are discussed in “Pipelining” on page 6-8. Basic bus cy-
cles are illustrated in Figure 6-1. The bus status signals indicate the type of bus cycle the proces-
sor is executing. Notice that the signal combinations marked as invalid states may occur when the
bus is idle and ADS# is inactive.
6-7
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Memory read and memory write cycles can be locked to prevent another bus master from using
the local bus. This allows for indivisible read-modify-write operations.
Reset Asserted
READY# Asserted • No Request
Always
No Request
Ti
T1
T2
Request Pending
READY# Asserted •
Request Pending
Bus States:
T1 - First clock of a non-pipelined bus cycle (CPU drives
new address and asserts ADS#).
READY# Negated
T2 - Subsequent clocks of a bus cycle when NA# has not
been sampled asserted in the current bus cycle.
Ti - Idle State
The fastest bus cycle consists of two states: T1 and T2.
A2484-02
Figure 6-2. Simplified Bus State Diagram
(Does Not Include Address Pipelining or Hold states)
6.2.2 Pipelining
The processor can control the address and status outputs so that the outputs for the next bus cycle
become valid before the end of the present bus cycle. This technique, allowing bus cycles to over-
lap, is called pipelining.
Pipelining increases bus throughput without decreasing allowable memory or I/O access time,
thus allowing high bandwidth with relatively slow, inexpensive components. In addition, using
pipelining to address slower devices can yield the same throughput as addressing faster devices
with no pipelining. With pipelining, a device operating at 33 MHz (CLK2 = 66 MHz) can transfer
data at 33 Mbytes per second while requiring a device with access time of approximately 3 T-
states (90 ns at 33 MHz, neglecting signal delays). Without address pipelining, the access time
has to be approximately 2 T-states (60 ns at 25 MHz). Therefore, when pipelining is used, slower
devices can be used in the system to achieve performance similar to a faster device in a non-pipe-
lined system.
Pipelining is not supported during I/O bus cycles and BS8 cycles (16-bit accesses to 8-bit devic-
es).
NOTE
During I/O cycles, NA# is ignored. NA# must be kept deasserted (blocked
externally) during the T2 states of BS8 memory cycles.
6-8
BUS INTERFACE UNIT
NOTE
Pipelining is also supported during memory cycles initiated by the two
integrated DMA units.
Refer to “Pipelined Cycle” on page 6-19 for a description of pipelined cycles.
6.2.3 Data Bus Transfers and Operand Alignment
The processor can address up to 64 Mbytes (226 bytes, addresses 0000000H–3FFFFFFH) of
physical memory and up to 64 Kbytes (216 bytes, addresses 0000H–FFFFH) of I/O. The device
maintains separate physical memory and I/O spaces.
A programmer views the address space (memory or I/O) as a sequence of bytes:
• Words consist of 2 consecutive bytes
• Doublewords consist of 4 consecutive bytes
However, in the system hardware, address space is implemented in 2-byte portions. When the
processor reads a word, it accesses a byte from each portion of the 16-bit data bus. The processor
automatically translates the programmer’s view of consecutive bytes into this hardware imple-
mentation.
Memory and I/O spaces are organized physically as sequences of 16-bit words (225 16-bit mem-
ory locations and 215 16-bit I/O ports maximum). Each word starts at a physical address that is a
multiple of 2 and has 2 individually addressable bytes at consecutive addresses.
Pins A25:1 correspond to the most-significant bits of the physical address; these pins address
words of memory. The least-significant bit of the physical address is used internally to activate
the appropriate byte enable outputs (BHE# or BLE# or both).
Data can be transferred in quantities of either 8 or 16 bits for each bus cycle of a data transfer.
When a data transfer can be completed in a single cycle, the transfer is said to be aligned. For
example, a word transfer involving D15:0 and activating BHE# and BLE# is aligned.
Word transfers that cross a word boundary or doubleword transfers that cross two word bound-
aries are called nonaligned transfers. Nonaligned word transfers require two bus cycles, while
nonaligned doubleword transfers require three. The processor automatically generates these cy-
cles. For example:
• A word (16-bit) transfer at (byte) address 03H requires two byte transfers:
— The first activates word address 04H and uses D7:0 (to write or read the upper byte of
the 16-bit word)
— The second activates word address 02H and uses D15:8 (to write or read the lower byte
of the 16-bit word)
6-9
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
• A doubleword (32-bit) transfer at (byte) address 03H requires three transfers, one word
transfer and two byte transfers:
— The first word transfer activates word address 04H and uses D15:0 (to write or read the
— The next transfer activates word address 06H and uses D7:0 (to write or read the upper
byte of the 32-bit word)
— The last transfer activates word address 02H and uses D15:8 (to write or read the lower
byte of the 32-bit word)
Table 6-3 shows the sequence of bus cycles for all possible alignments and operand length trans-
fers. Even though nonaligned transfers are transparent to a program, they are slower than aligned
transfers (due to the extra cycles needed) and should be avoided.
Table 6-3. Sequence of Nonaligned Bus Transfers
First Cycle
Second Cycle
Third Cycle
Transfer
Type
Physical
Address
Address
Byte
Address Byte
Address
Byte
Bus
Enable
Bus
Enable
Bus
Enable
word
4N+1
4N+3
4N+1
4N+2
4N+3
4N
BHE#
BLE#
BLE#
both
4N+2
4N+2
4N
BLE#
BHE#
BHE#
both
word
4N+4
4N+4
4N+4
4N+4
doubleword
doubleword
doubleword
4N+2
4N+3
both
4N+2
4N+6
both
BLE#
BHE#
6.2.4 Ready Logic
A bus cycle is terminated externally by asserting the READY# pin or internally by either an in-
ternal peripheral or the Chip-select Unit’s wait-state logic. When an access is to an internal pe-
decodes a match to the address and drives the READY# pin, contention occurs on the signal. The
LBA# pin should be used to alleviate the possibility of contention on the READY# pin. The
READY# pin is an output of the processor whenever LBA# is asserted and an input to the pro-
cessor whenever LBA# is deasserted.
The LBA# pin becomes active when the processor is generating the READY# internally. Figure
6-3 shows the implementation of the READY# signal using the LBA# signal. If you wish to sim-
plify decoding of address space and overlap internal I/O registers, you need to provide external
logic to monitor LBA# and end the bus cycle externally when the processor generates the
READY# internally.
NOTE
Since LBA# may be used as an output-enable by both the internal and external
READY# buffers, care must be taken in selecting the external READY# buffer
to minimize contention on the READY# signal caused by differences in buffer
characteristics.
6-10
BUS INTERFACE UNIT
LBA#
Bus
Unit
READY#
To Internal Units
Chip Boundary
A2485-01
Figure 6-3. Ready Logic
When an internal cycle occurs, the LBA# signal becomes active in Phase 1 of the first T2 state.
It then stays active until the rising edge of PH1 of the first T2, T2i or T2P state of the next bus
cycle that requires external READY# to terminate the bus cycle. For example, the processor may
start an internal bus cycle, go through a few idle states, perform another internal cycle, then a cy-
cle in which the Chip-select Unit generates READY#, run through a few more idle states and then
finally do a cycle in which READY# needs to be generated by external logic. LBA# goes active
in the first T2 state of the first internal cycle, and stay active through the next two cycles (even
during all the idle states in between) and go inactive at the rising edge of PH1 in the first T2, T2i
or T2P state of the final cycle (the one that requires an external READY# to terminate).
NOTE
LBA# is deasserted during HOLD cycles.
6-11
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Figure 6-4 shows internal and external bus cycles.
Idle
Cycle 1
Cycle 2
Cycle 3
Idle
Cycle 4
Idle
Cycle
Nonpipelined Nonpipelined Nonpipelined Cycle Nonpipelined Cycle
External
(Write)
Internal
(Read)
Internal
(Write)
External
(Read)
[Late Ready]
[Early Ready]
T1
T2
T1
T2
T1
T2
Ti
T1
T2
Ti
Ti
State
CLK2
CLKOUT
A25:1, BHE#
BLE#, D/C#
M/IO#
Valid 1
Valid 2
Valid 3
Valid 4
REFRESH#
W/R#
WR#
RD#
ADS#
NA#
READY#
End Cycle 2
End Cycle 3
End Cycle 1
End Cycle 4
LBA#
BS8#
Valid 1
Valid 2
Valid 3
Out 3
Valid 4
LOCK#
In
In
Out 1
D15:0
2
4
A2486-03
Figure 6-4. Basic Internal and External Bus Cycles
6-12
BUS INTERFACE UNIT
6.3 BUS CYCLES
The processor executes five types of bus cycles:
• Read
• Write
• Interrupt
• Halt/shutdown
• Refresh
6.3.1 Read Cycle
Read cycles are of two types:
• In a pipelined cycle, the address and status signals are output in the previous bus cycle, to
allow longer memory access times. Pipelined cycles are described in “Pipelined Cycle” on
page 6-19.
• In a nonpipelined cycle, the address and status signals become valid during the first T-state
of the cycle (T1). Figure 6-5 shows the timing for two nonpipelined read cycles (one with
and one without a wait-state).
The sequence of signals for the nonpipelined read cycle is as follows:
1. The processor initiates the cycle by driving the address bus and the status signals active
and asserting ADS#. The type of bus cycle occurring is determined by the states of the
address bus (A25:1), byte enable pins (BLE# and BHE#), and bus status outputs (W/R#,
M/IO#, D/C#, REFRESH#, and LOCK#). Because of output delays, these signals should
be sampled at the rising edge of the CLK2 signal that coincides with the falling edge of
PH2, when ADS# is definitely active. For a read cycle, the bus status outputs have the
following states:
•
•
•
•
•
W/R# is low
M/IO# is high for a memory read and low for an I/O read
D/C# is high for a memory or I/O data read and low for a memory code read
REFRESH# is deasserted
LOCK# is asserted for a locked cycle and deasserted for a nonlocked cycle. In a read-
modify-write sequence, both the memory data read and memory data write cycles are
locked. No other bus master should be permitted to control the bus between two
locked bus cycles.
The address bus, byte enable pins, and bus status pins (with the exception of ADS#)
remain active through the end of the read cycle.
2. At the start of phase 2 of T1, RD# becomes active as the processor prepares the data bus
for input. This indicates that the processor is ready to accept data.
6-13
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
3. When a chip-select region is enabled for the current read cycle but internal READY#
generation is disabled for that region, and the Chip-select Unit is programmed to insert
wait-states, the READY# signal is ignored (not sampled) by the processor until the
programmed number of wait-states are inserted into the cycle.
4. At the falling edge of PH2 in every T2 state (after the wait-states, if any are programmed
in the Chip-select Unit, have expired), READY# is sampled. If READY# is active, the
processor reads the input data on the data bus and deactivates RD#.
5. If READY# is high, wait states are added (additional T2 states for nonpipelined cycles)
until READY# is sampled low. READY# is sampled at the end of each T2 state (at the
falling edge of PH2).
6. Once READY# is sampled low, the processor reads the input data, deactivates RD#, and
terminates the read cycle. If a new bus cycle is pending, it begins on the next T-state.
6-14
BUS INTERFACE UNIT
Idle
Ti
Cycle 1
Non-pipelined
External
Cycle 2
Non-pipelined
External
Idle
Ti
(Read)
(Read)
T1
T2
T1
T2
T2
CLK2
CLKOUT
BHE#, BLE#, A25:1
M/IO#, D/C#
Valid1
Valid2
REFRESH#
W/R#
WR#
RD#
ADS#
NA#
READY#
End Cycle
End Cycle
LBA#
BS8#
LOCK#
D15:0
Valid1
Valid2
In1
In2
A2487-03
Figure 6-5. Nonpipelined Address Read Cycles
6-15
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
6.3.2 Write Cycle
Write cycles are of two types:
• Pipelined. Pipelined write cycles are described in “Pipelined Cycle” on page 6-19.
• Nonpipelined. Figure 6-6 shows two nonpipelined write cycles (one with and one without a
wait state).
The sequence of signals for a nonpipelined write cycle is as follows:
1. The processor initiates the cycle by driving the address bus and the status signals active
and asserting ADS#. The type of bus cycle occurring is determined by the states of the
address bus (A25:1), byte enable pins (BLE# and BHE#), and bus status outputs (W/R#,
M/IO#, D/C#, REFRESH#, and LOCK#). Because of output delays, these signals should
be sampled at the rising edge of the CLK2 signal that coincides with the falling edge of
PH2, when ADS# is definitely active. For a write cycle, the bus status outputs have the
following states:
•
•
•
W/R# is high
M/IO# is high for a memory write and low for an I/O write
D/C# is high for a memory write or I/O write cycle. During halt and shutdown cycles,
D/C# is low. Unless D/C# is decoded by external chip-select logic, the shutdown or
halt cycle looks like a memory write cycle to byte address zero or two, respectively.
Therefore, the signal D/C# needs to be decoded for memory device chip-selects in
this address range (normally SRAM or DRAM devices) in order to recognize halt and
shutdown cycles, thus preventing incorrect write cycles to memory
•
•
REFRESH# is deasserted
LOCK# is asserted for a locked cycle and deasserted for an unlocked cycle. In a read-
modify-write sequence, both the memory data read and memory data write cycles are
locked. No other bus master should be permitted to control the bus between two
locked bus cycles.
The address bus, byte enable pins, and bus status pins (with the exception of ADS# and
WR#) remain active through the end of the write cycle.
2. At the start of Phase 2 in T1, the WR# signal is asserted and the CPU begins to drive
output data on its data pins. The data remains valid until the start of phase 2 in the T-State
after the present bus cycle has terminated.
3. If a chip-select region is enabled for the current read cycle but internal READY#
generation is disabled for that region, and the Chip-select Unit is programmed to insert
wait-states, then the READY# signal is ignored (not sampled) by the processor until the
programmed number of wait-states are inserted into the cycle.
6-16
BUS INTERFACE UNIT
4. The WR# signal can be deasserted in two ways.
•
Early Ready: WR# is deasserted at the rising edge of CLK2 in the middle of the T2
state, after any wait states programmed in the Chip-select Unit have expired.
At the rising edge of PH2, READY# is sampled. If it is found active, WR# is
synchronously deasserted in the middle of T2, driven inactive by the rising edge of
the PH2 clock. The write cycle is then terminated at the end of the T2 state.
NOTE
When READY# is generated by the processor (e.g., when the Chip-select Unit
generates it), then the write cycle is always an Early Ready cycle.
•
Late Ready: When READY# goes low after the rising edge of PH2 of the T2 state
(after the wait-states, if any are programmed in the Chip-select Unit, have expired),
WR# is asynchronously deasserted as soon as READY# is asserted (after a small
delay caused by the logic). The write cycle is then terminated at the end of the T2
state.
The WR# signal operates in this manner to ensure sufficient address and chip-select hold
time during write cycles (required by many memory and I/O devices). In the first case, the
address and chip-select hold time is approximately one CLK2 cycle.
5. When READY# is high, wait-states are added (additional T2 states for nonpipelined
cycles) until READY# is sampled low. READY# is sampled in each T2 state (starting at
the rising edge of PH2) to deassert the WR# signal appropriately, and at the end of each T2
state (at the falling edge of PH2) to terminate the cycle.
6. Once READY# is sampled low, the write cycle terminates. If a new bus cycle is pending,
it begins on the next T-state.
6-17
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Idle
Ti
Cycle 1
Nonpipelined
External
(Write)
[Late Ready]
Cycle 2
Nonpipelined
External
(Write)
[Early Ready]
Idle
Ti
T1
T2
T1
T2
T2
CLK2
CLKOUT
BHE#, BLE#, A25:1
M/IO#, D/C#
Valid1
Valid2
REFRESH#
W/R#
WR#
RD#
ADS#
NA#
READY#
End Cycle 1
End Cycle 2
LBA#
BS8#
Valid 2
LOCK#
D15:0
Valid 1
Out 1
Out 2
A2488-02
Figure 6-6. Nonpipelined Address Write Cycles
6-18
BUS INTERFACE UNIT
6.3.3 Pipelined Cycle
The pipelining feature of the processor is normally used to achieve zero-wait-state memory sub-
systems using devices that are slower than those in a zero-wait-state non-pipelined system. Pipe-
lining allows bus cycles to be overlapped, increasing the amount of time available for the memory
or I/O device to respond. The next address (NA#) input controls pipelining. NA# is generated by
logic in the system to indicate that the address and status buses are no longer needed by the sys-
tem. When pipelining is not desired in a system, the NA# input should be tied inactive.
During any particular bus cycle, NA# is sampled only after the address and status have been valid
for one T-state (the T1P state of pipelined cycles or the first T2 state of nonpipelined cycles) and
is continuously sampled in each subsequent T-state until it is found active or the bus cycle is ter-
minated. In particular, NA# is sampled at the rising CLK2 edge in the middle of the T-state (rising
edge of Phase 2).
When the system is designed to assert NA#, pipelining may be dynamically requested on a cycle-
by-cycle basis by asserting NA#. Typically, only some devices in a system are pipelined.
NOTE
Asserting the NA# pin is a request for pipelining. Asserting NA# during a bus
cycle does not guarantee that the next cycle is pipelined. NA# is ignored
during I/O cycles and must be kept deasserted during the T2 states of BS8
memory cycles.
During the T2 state of a nonpipelined cycle, if NA# is sampled active, one of four states occur:
• If a bus cycle is internally pending in the processor and READY# is returned inactive to the
processor and the HOLD input is inactive, then the address, byte enables, and bus status
signals for the next bus cycle are driven and the processor bus unit enters a T2P state. T2P
states are repeated until the bus cycle is terminated.
• If a bus cycle is internally pending in the processor and READY# is returned active to the
processor and the HOLD input is inactive, then the address, byte enables, and bus status
signals for the next bus cycle are driven and the processor bus unit enters a T1
(nonpipelined) state. In effect, the NA# input is ignored in this case.
• If READY# is returned inactive and either a bus cycle is not internally pending or the
HOLD input is active, then the address and byte enables enter an unknown state, the bus
terminated, then the next state is either a T2P state or a T2i state depending on whether a
bus cycle is pending.
• If HOLD is asserted to the processor and READY# is returned active, then the Th state is
entered from a T2 state regardless of whether an internal bus cycle is pending.
Figure 6-8 illustrates the effect of NA# (Figure 6-7 shows the full bus state diagram including the
states related to pipelining). During the second T-state (T2) of a nonpipelined read cycle (cycle
2), NA# is sampled low. A bus cycle was pending internally (cycle 3) and the address, byte en-
ables, and bus status signals for this pending bus cycle (cycle 3) are driven during the next T2P
state (the first wait state of the current bus cycle). The RD# and WR# signals do not change until
READY# is sampled low.
6-19
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
HOLD Asserted
READY# Asserted •
HOLD Asserted
Th
READY# Asserted •
HOLD Asserted
HOLD Negated •
Request Pending
Reset
Asserted
L
D
a
HOLD Negated •
No Request
A
NA# Negated
Always
Ti
T1
T2
T1P
Request Pending •
HOLD Negated
READY# Asserted •
HOLD Negated •
Request Pending
READY# Asserted •
HOLD Negated •
Request Pending
READY# Negated •
NA# Negated
T2i
READY# Asserted •
HOLD Negated •
No Request
Bus States:
T1—first clock of a non-pipelined bus cycle.
READY# Negated •
(No Request +
HOLD Asserted)
T2—subsequent clock of a bus cycle when NA# has
not been sampled active in the current bus cycle.
T2i—subsequent clocks of a bus cycle when NA# has
been sampled active in the current bus cycle and there
is not yet an internal bus request pending.
T2P
T2P—subsequent clocks of a bus cycle when NA# has
been sampled active in the current bus cycle and there
is an internal bus request pending.
T1P—first clock of a pipelined bus cycle.
Ti—idle state.
Th—hold acknowledge state.
READY# Negated
A2376-02
Figure 6-7. Complete Bus States (Including Pipelined Address)
6-20
BUS INTERFACE UNIT
Cycle 1
Pipelined
(Write)
Cycle 2
Non-pipelined
(Read)
Cycle 3
Cycle 4
Pipelined
(Read)
Pipelined
(Write)
[Late Ready]
[Late Ready]
T1P
T2P
T2P
T1P
T2
T2P
T1P
T2i
T2P
T1P
T2
CLK2
CLKOUT
BHE#, BLE#, A25:1,
M/IO#, D/C#
Valid2
Valid3
Valid4
ADS# is asserted as
soon as the CPU has
another bus cycle to
perform, which is not
always immediately
after NA# is asserted.
W/R#
WR#
RD#
ADS#
Note ADS# is
asserted in
every T2P state.
As long as the CPU enters the T2P
state during Cycle 3, address
pipelining is maintained in Cycle 4.
NA#
Asserting NA# more
than once during
any cycle has no
additional effects
NA# could have been asserted in T1P
if desired. Assertion now is the latest
time possible to allow the CPU to enter
T2P state to maintain pipelining in cycle 3.
READY#
LBA#
BS8#
Valid 1
LOCK#
D15:0
Valid 2
Valid 3
Valid 4
In
Out 1
Out 3
Out
2
A2477-03
Figure 6-8. Pipelined Address Cycles
6-21
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
In cycle 3, NA# is sampled in the first T-state (T1P); the address and status have been valid for
one previous T-state and this is a new bus cycle. NA# is sampled active and — because a bus cy-
cle (cycle 4) is pending internally — the address, byte enables, and bus status signals for this
pending bus cycle (cycle 4) are driven during the next T2P state.
In cycle 4, NA# is sampled in the first T-state (T1P); the address and status have been valid for
one previous T-state, and this is a new bus cycle. NA# is sampled active and — because a bus
cycle is not internally pending — the address and byte enables go to an unknown state and the
bus status signals go inactive in the next T2i state. When this cycle is terminated by an active
READY# signal, there is no bus cycle pending internally and the bus enters the idle state (Ti).
From an idle bus, an additional overhead of one clock cycle is required to start a pipelined bus
cycle (this is true with all pipelined bus architectures). This additional clock is used to pipeline
the address and status signals for the first bus cycle in a train of pipelined bus cycles. As long as
back-to-back bus cycles are executed, the pipelined bus can maintain the same throughput as the
nonpipelined bus. Only when the bus pipeline gets broken (by entering an idle or hold state) is
the additional one-clock overhead required to start the pipe again for the next train of pipelined
bus cycles.
The first bus cycle after an idle bus state is always nonpipelined. Systems that use pipelining typ-
ically assert NA# during this cycle to enter pipelining. To initiate pipelining, this nonpipelined
cycle must be extended by at least one T-state so that the address and status can be pipelined be-
fore the end of the cycle. Subsequent cycles can be pipelined as long as no idle bus cycles occur.
Specifically, NA# is sampled at the start of phase 2 of any T-state in which the address and status
signals have been active for one T-state and a new cycle has begun:
• The first T2 state of a nonpipelined cycle (the second T-state)
• The T1P state of a pipelined cycle (the first T-state)
• Any wait state of a nonpipelined or pipelined cycle unless NA# has already been sampled
active
Once NA# is sampled active, it remains active internally throughout the current bus cycle. When
NA# and READY# are active in the same T2 state, the state of NA# is irrelevant because
READY# causes the start of a new bus cycle. Therefore, the new address and status signals are
always driven, regardless of the state of NA#. NA# has no effect on a refresh cycle because the
refresh cycle is entered from an idle bus state and exits to an idle bus state.
With this processor, address pipelining is optional so that bus cycle timing can be closely tailored
to the access time of the memory device.
• Pipelining can be activated once the address is latched externally.
• Pipelining can be not activated if the address is not latched.
For systems that use address pipelining, the great majority of accesses are pipelined. Very few
idle states occur in an Intel386 EX processor system. This means that once the processor has en-
tered pipelining, another bus cycle request is almost always internally pending, resulting in a con-
tinuous train of pipelined cycles. In measured systems, about 85% of bus cycles are pipelined.
6-22
BUS INTERFACE UNIT
™
A complete discussion of the considerations for using pipelining can be found in the Intel386
SX Processor datasheet (order number 240187) or the Intel386™ SX Microprocessor Hardware
Reference Manual (order number 240332).
6.3.4 Interrupt Acknowledge Cycle
An interrupt causes the processor to suspend execution of the current program and execute in-
structions from another program called an interrupt service routine. Interrupts are described in
Chapter 9.
The interrupt control unit coordinates the interrupts of several devices, internal and external. It
contains two 82C59A programmable interrupt controllers (PICs) connected in cascade. The slave
82C59A module controls up to five internal interrupt sources and up to four external interrupt
sources depending upon the configuration programmed. The master 82C59A module controls the
slave 82C59A, three internal interrupt sources and up to six external interrupt sources depending
upon the configuration programmed. When a device signals an interrupt request, the interrupt
control unit activates the processor’s INTR input.
Interrupt acknowledge cycles are special bus cycles that enable the interrupt control unit to output
a service-routine vector onto the data bus. The processor performs two back-to-back interrupt ac-
knowledge cycles in response to an active INTR input (as long as the interrupt flag is enabled).
Interrupt acknowledge cycles are similar to regular bus cycles in that the processor initiates each
bus cycle and an active READY# terminates each bus cycle. The cycles are shown in Figure 6-9.
The sequence of signals for an interrupt acknowledge cycle is as follows:
1. The address and status signals are driven active and ADS# is driven low to start each bus
cycle.
•
Status signals M/IO#, D/C#, and W/R# are low to indicate an interrupt acknowledge
bus cycle. These signals must be decoded to generate the INTA input signal for an
external 82C59A, if an external cascaded 82C59A is used. The REFRESH# signal is
high.
•
LOCK# is active from the beginning of the first cycle to the end of the second. HOLD
requests from other bus masters are not recognized until after the second interrupt
acknowledge cycle is completed.
•
•
NA# is ignored.
The byte address driven during the first cycle is 4; during the second cycle the byte
address is 0. BHE# is high, BLE# is low, and A25:3 and A1 are low for both cycles;
A2 is high for the first cycle and low for the second. If the CAS enable bit in the
interrupt control unit’s configuration register is set (INTCFG.7=1), address bits
A18:16 reflect the status of the CAS lines. The CAS lines go valid at the rising edge
of PH2 of the T1 state of the first interrupt acknowledge cycle. They then go invalid
at the rising edge of PH2 of the next Ti state. At the rising edge of PH2 of the T1 state
of the second interrupt acknowledge cycle, the CAS lines go valid again. They then
go invalid at the rising edge of PH2 of the next Ti state.
6-23
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
NOTE
Since the CAS lines are invalid in the Ti states between the two interrupt
acknowledge cycles, cascading of external 82C59A devices requires latching
the CAS lines. This ensures that the CAS lines remain valid during these Ti
states to fulfill the requirements of the external 82C59A devices.
2. The processor floats D15:0 for both cycles; however, at the end of the second cycle, if the
interrupt is from an external cascaded 82C59A, the service-routine vector number driven
on the lower data bus by the 82C59A is read by the processor on data pins D7:0.
Otherwise, the active internal 82C59A sends the vector to the processor.
3. The first cycle is always an internal cycle and the second may be internal or external.
Therefore, READY# is generated internally for the first cycle and for the second cycle, if
the interrupt request is from one of the internal 82C59A modules. If the interrupt is from a
cascaded external 82C59A, external logic must assert READY# to terminate the second
cycle. The internal Chip-select Unit can not generate READY# for the second interrupt
acknowledge cycle.
6-24
BUS INTERFACE UNIT
Previous Interrupt
Idle
Interrupt
Acknowledge
Cycle 2
Idle
Ti
Acknowledge
Cycle 1
Cycle
(Four bus states)
(Internal)
(Internal)
T2
T1
T2
Ti
Ti
Ti
Ti
T1
T2
Ti
CLK2
CLKOUT
BHE#
BLE#, A25:A3, A1
M/IO#, D/C#, W/R#
A2
WR#
RD#
ADS#
READY#
LBA#
LOCK#
A2490-03
Figure 6-9. Interrupt Acknowledge Cycles
6-25
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
6.3.5 Halt/Shutdown Cycle
The halt condition occurs in response to a HALT instruction. The shutdown condition occurs
when the processor is processing a double fault and encounters a protection fault; the processor
cannot recover and therefore, shuts down. Externally, a shutdown cycle differs from a halt cycle
only in the resulting address bus outputs. The sequence of signals for a halt cycle is as follows:
1. As with other bus cycles, a halt or shutdown cycle is initiated by driving the address and
status signals active and asserting ADS#. Figure 6-10 shows a halt bus cycle. The address
and status signals are driven to the following active states:
•
M/IO# and W/R# are driven high and D/C# is driven low to indicate a halt cycle or a
shutdown cycle.
•
The address bus outputs a byte address of 2 for a halt condition and a byte address of
0 for a shutdown condition. These signals are used by external devices to respond to
the halt or shutdown cycle.
NOTE
The halt or shutdown bus cycle appears as a memory write operation to byte
address 0 or 2 (depending on whether a shutdown or halt cycle is being
performed) if the D/C# signal is not decoded. External address decoders need
to decode the D/C# signal to avoid erroneous writes to devices in this address
region; otherwise, a halt or shutdown cycle corrupts the data at those
addresses. RD#, WR# and the chip-select signals, UCS# and CS6:0#, are
inactive during halt cycles.
2. READY# can be generated externally or internally to terminate a Halt/Shutdown cycle.
The HSREADY bit in the Power Control Register (PWRCON, see Figure 8-5 in Chapter
8), can be set to generate an internal READY# for halt/shutdown cycles. If internal
READY# generation is enabled, then the LBA# signal goes active and behaves as
described in “Ready Logic” on page 6-10. Also, the cycle is always a zero-wait-state
cycle. When external READY# is required to terminate the halt/shutdown cycle, then
READY# may be delayed to add wait-states. The processor remains in the halt or
shutdown condition until one of the following occurs:
•
•
•
NMI goes active; the processor then services the interrupt.
RESET goes active; the processor is reinitialized.
In the halt condition (but not in the shutdown condition), if maskable interrupts are
enabled, an active INTR input causes the processor to end the halt cycle and service
the interrupt. The processor can service processor extension (PEREQ) requests and
hold (HOLD) requests while in the halt or shutdown condition.
•
The processor is in the halt condition and SMI# goes active; the processor then
services the SMI#. When the processor is in the shutdown condition, SMI# has no
effect.
6-26
BUS INTERFACE UNIT
Cycle 1
Nonpipelined Nonpipelined
(Write)
[Late Ready]
Cycle 2
Idle
Ti
(Halt)
T1
T2
T1
T2
Ti
Ti
Ti
CLK2
CLKOUT
CPU remains halted until INTR, SMI#,
NMI, or RESET is asserted.
BHE#, A1, M/IO#, W/R#
Valid 1
Valid 1
CPU responds to HOLD input
while in the HALT state.
A25:2, BLE#, D/C#
WR#
RD#
ADS#
NA#
READY#
LBA#
†
Valid 1
Valid 2
LOCK#
D15:0
Float
Undefined
Out
† HALT cycle must be acknowledged by READY# asserted. This READY# could be
generated internally or externally.
A2492-02
Figure 6-10. Halt Cycle
6-27
6.3.6 Refresh Cycle
The refresh control unit simplifies dynamic memory controller design by issuing dummy read cy-
cles at specified intervals. (For more information, refer to Chapter 15, “REFRESH CONTROL
UNIT.”) Figure 6-11 shows a basic refresh cycle. The sequence of signals for a refresh cycle is as
follows:
1. Like a read cycle, the refresh cycle is initiated by asserting ADS# and completed by
asserting READY#. The address and status pins are driven to the following values:
•
M/IO# and D/C# are driven high and W/R# and REFRESH# are driven low to
indicate a memory refresh.
•
Address lines are driven to the current refresh address (the value in the Refresh
Address Counter in the Refresh Control Unit), while the BHE# and BLE# are driven
high.
2. To complete the refresh cycle, either READY# must be asserted externally or the chip
select unit must be programmed to generate READY# for the address region specified in
relinquishes control to the current internal bus master until the next refresh cycle is
needed.
During hold acknowledge cycles with the HLDA pin active, a refresh request causes the internal
bus arbiter to deassert the HLDA pin. The processor then waits for the HOLD pin to be deasserted
for at least one processor clock cycle. Once HOLD is deasserted, the processor begins the refresh
cycle. Figure 6-12 shows a refresh cycle during a HOLD/HLDA condition.
NOTE
BS8# is ignored during refresh cycles. It has no effect on a refresh cycle.
CAUTION
External bus arbitration logic should monitor the HLDA signal when the
refresh control unit is being used. If a refresh request is not serviced (by
performing a refresh cycle) because an external master does not give up the
bus, the DRAM devices may lose data.
6-28
BUS INTERFACE UNIT
Idle
Ti
Cycle 1
Nonpipelined
External
Idle
Ti
Cycle 2
Refresh
Idle
Cycle 3
Nonpipelined
External
(Write)
[Late Ready]
(Read)
T1
T2
T1
T2
T2
Ti
Ti
T1
T2
CLK2
CLKOUT
BHE#, BLE#
M/IO#, D/C#
Valid 1
Valid 3
Valid 1
Valid 2
Valid 3
A25:1
REFRESH#
W/R#
WR#
RD#
ADS#
NA#
READY#
LBA#
Valid 1
Valid 2
Out
LOCK#
D15:0
Float
In
HOLD
HLDA
A2491-02
Figure 6-11. Basic Refresh Cycle
6-29
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Idle
Ti
HOLD
Acknowledge
Idle
Ti
Idle
HOLD
Acknowledge
Cycle 1
Refresh
Th
Th
Th
T1
T2
Ti
Ti
Th
Th
CLK2
CLKOUT
Floating
Floating
BHE#, BLE#
M/IO#, D/C#
REFRESH#
A25:1
Floating
Floating
Floating
Floating
Valid 1
W/R#
WR#
RD#
ADS#
NA#
READY#
LBA#
Floating
Floating
Floating
LOCK#
D15:0
HOLD
HLDA
Due to refresh pending.
A2493-02
Figure 6-12. Refresh Cycle During HOLD/HLDA
6-30
BUS INTERFACE UNIT
6.3.7 BS8 Cycle
The BS8 cycle allows external logic to dynamically switch between an 8-bit data bus size and a
16-bit data bus size, by using the BS8# signal. Figure 6-13 shows a word access to an 8-bit pe-
ripheral.
To use the dynamic 8-bit bus sizing, an external memory or I/O should connect to the lower eight
bits of the data bus (D7:0), use the BLE# as address bit 0, and assert BS8# (at the BS8# pin) in
T2 of a memory or I/O access. A BS8 cycle can also be generated by the internal chip-select unit
(Refer to Chapter 14, “CHIP-SELECT UNIT”). In this case, the Chip Select Unit generates the
BS8# signal internally.
Depending upon the current bus access width and address and the state of the BS8# signal, the
processor performs the actions described in the next two sections.
6.3.7.1
Write Cycles
• If the current bus cycle is a byte write with BHE# active and BLE# inactive, the processor
copies the upper eight bits of the data bus (D15:8) to the lower eight bits of the data bus
(D7:0), i.e. the byte appears on both the upper and lower data buses.
• If the current bus cycle is a byte write with BHE# inactive and BLE# active, the processor
ignores the state of the BS8# signal.
• If the current bus cycle is a word write with both BHE# and BLE# active and the processor
samples the BS8# signal active at the end of the last T2 (when READY# is sampled active),
the processor waits for the current bus to complete and then executes another write cycle
with the upper eight bits of the data bus (D15:8) copied to the lower eight bits of the data
bus (D7:0). The processor deactivates BLE# on the second cycle (BLE# is used as address
A0 to an 8-bit device; this translates to A0=0 for the first cycle and A0=1 for the second).
6.3.7.2
Read Cycles
• If the current bus cycle is a byte read with BHE# active and BLE# inactive, and the
processor samples the BS8# signal active at the end of the last T2 (when READY# is
sampled active), the processor latches the data on the lower eight bits of the data bus (D7:0)
and internally routes this data to the upper data bus of the core.
• If the current bus cycle is a byte read with BHE# inactive and BLE# active, the processor
ignores the state of the BS8# signal.
• If the current bus cycle is a word read with both BHE# and BLE# active and the processor
samples the BS8# signal active at the end of the last T2 (when READY# is sampled active),
the processor waits for the current bus cycle to complete and latches the data on the lower
eight bits of the data bus (D7:0). It then executes another read cycle, with BLE# inactive
(BLE# is used as address A0 to an 8-bit device; this translates to A0=0 for the first cycle
and A0=1 for the second), latching the data on the lower eight bits of the data bus (D7:0)
again and using it.
6-31
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
The BS8 cycle generates additional bus cycles for read and write cycles only. For interrupt and
halt/shutdown cycles, the accesses are byte wide and the BS8# signal is ignored. For a refresh
cycle, the byte enables are both disabled and the BS8# signal is ignored.
NOTE
If a BS8 cycle requires an additional bus cycle, the processor retains the
current address for the second cycle. Address pipelining cannot be used with
BS8 cycles because address pipelining requires that the next address be
generated on the bus before the end of the current bus cycle. NA# must be kept
deasserted during the T2 states of BS8 memory cycles. NA# is ignored in all
I/O cycles.
NOTE
BS8# must be inactive at the falling edge of PH2 of the T1 state of a non-BS8
cycle; for example, if the current cycle is a BS8 cycle (BS8# asserted) and the
next cycle is not a BS8 cycle, BS8# must be deasserted before the end of the
T1 state of the next cycle, i.e. the non-BS8 cycle.
6-32
BUS INTERFACE UNIT
Low Byte
Write
[Late Ready]
High Byte
Write
[Late Ready]
Low Byte
Read
High Byte
Read
Idle
Cycles
State
CLK2
T1
T2
T1
T2
T1
T2
T1
T2
Ti
Ti
CLKOUT
A25:1
M/IO#
D/C#
Valid 1
Valid 2
BLE#
BHE#
W/R#
WR#
RD#
ADS#
NA#
Must be high
READY#
BS8#
Valid 1
Valid 2
LOCK#
D15:8
D7:0
Data Out High
Data
In
Data
Data Out
Low
Data Out
High
In
Low
High
A3375-01
Figure 6-13. 16-bit Cycles to 8-bit Devices (Using BS8#)
6-33
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
6.4 BUS LOCK
In a system in which more than one device (a bus master) may control the local bus, locked cycles
are used to make sequential bus cycles indivisible. Otherwise, the cycles may be separated by a
cycle from another bus master.
Any bus cycles that must be performed back-to-back, without any intervening bus cycles by other
bus masters, must be locked. The use of a semaphore is one example of this concept. The value
of a semaphore indicates a condition such as the availability of a device. If the CPU reads a sema-
phore to determine that a device is available, then writes a new value to the semaphore to indicate
that it intends to take control of the device, the read cycle and write cycle should be locked to
prevent another bus master from reading from or writing to the semaphore in between the two
cycles.
The LOCK# output indicates, to the other bus masters, that they may not gain control of the bus.
In addition, when LOCK# is asserted, the processor does not recognize a HOLD request from an-
other bus master.
6.4.1 Locked Cycle Activators
The LOCK# signal is activated explicitly by the LOCK prefix on certain instructions. (The in-
structions are listed in the Intel386™ SX Microprocessor Programmer’s Reference Manual, order
number 240331). LOCK# is also asserted automatically for XCHG instructions, descriptor up-
dates, and interrupt acknowledge cycles.
6.4.2 Locked Cycle Timing
LOCK# is activated on the CLK2 edge that begins the first locked bus cycle and deactivated when
READY# is sampled active at the end of the last bus cycle to be locked. LOCK# is activated and
deactivated on these CLK2 edges regardless of address pipelining. If address pipelining is used,
LOCK# remains active until the current bus cycle is completed (READY# sampled active for the
current bus cycle). Consequently, the LOCK# signal can extend into the next memory access cy-
cle that does not need to be locked. (See Figure 6-14). The result is that the use of the bus by an-
other bus master is delayed by one bus cycle.
6-34
BUS INTERFACE UNIT
Unlocked
Bus Cycle
Locked
Bus Cycle
Locked
Bus Cycle
Unlocked
Bus Cycle
CLKOUT
Address Asserted
BLE#, BHE#, A25:1
LOCK Deasserted
LOCK#
READY#
A2489-02
Figure 6-14. LOCK# Signal During Address Pipelining
6.4.3 LOCK# Signal Duration
The maximum duration of the LOCK# signal affects the maximum HOLD request latency be-
cause HOLD is recognized only after LOCK# goes inactive. The duration of LOCK# depends on
the instruction being executed and the number of wait states per cycle. The longest duration of
LOCK# is 9 bus cycles plus approximately 15 clocks. This occurs when an interrupt (hardware
or software) occurs and the processor performs a Locked read of the gate in the interrupt descrip-
tor table (8 bytes), a read of the target descriptor (8 bytes), and a write of the accessed bit in the
target descriptor.
6.5 EXTERNAL BUS MASTER SUPPORT (USING HOLD, HLDA)
The processor provides internal arbitration logic that supports a protocol for transferring control
of the processor bus to an external bus master. This protocol is implemented through the HOLD
input and the HLDA output. The internal arbitration logic of the processor consists of a bus arbi-
ter. This arbiter supports the core and four other bus masters, i.e. external bus master using
HOLD, two internal DMA Units and the Refresh Control Unit. For a description of the protocol
of the internal bus arbiter, refer to “Bus Control Arbitration” on page 12-9.
When the internal bus arbiter receives a request through one of its four possible request signals,
it asserts the HOLD signal to the core. The core then completes its current nonlocked bus cycle
and asserts its HLDA signal, thus informing the arbiter that control of the bus can now be turned
over to the requester.The arbiter then asserts its appropriate acknowledge signal to the requester.
For example, if an external bus master requests the bus using the HOLD input pin, then the arbiter
asserts the HLDA output.
6-35
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
6.5.1 HOLD/HLDA Timing
To gain control of the local bus, the requesting bus master drives the HOLD input active. This
signal can be asynchronous to the processor’s CLK2 input. The processor responds by:
• completing its current bus cycle
• deasserting WR#, RD#, LBA#, SMIACT#, UCS#, CS6:0# and REFRESH# and three-
stating all other bus outputs except HLDA (effectively removing itself from the bus)
• driving HLDA active to signal the requesting bus master that it may take control of the bus
The requesting bus master must maintain HOLD active until it no longer needs the bus. When
HOLD goes low, the processor drives HLDA low and starts a bus cycle (if one is pending).
For valid system operation, the requesting bus master must not take control of the bus until it re-
ceives the HLDA signal and must remove itself from the bus before deasserting the HOLD signal.
Setup and hold times relative to CLK2 for both rising and falling transitions of the HOLD signal
must be met.
If the internal refresh control unit is used, the HLDA signal may drop while an external master
has control of the bus, in which case the external bus master may or may not drop HOLD to allow
the processor to perform the refresh cycle. If the latter occurs, the memory device(s) may lose
data because the refresh cycle could not execute.
When the processor receives an active HOLD input, it completes the current bus cycle before re-
linquishing control of the bus. Figure 6-7 shows the state diagram for the bus including the HOLD
state.
During HOLD, the processor can continue executing instructions that are already in its prefetch
queue. Program execution is delayed if a read cycle is needed while the processor is in the HOLD
state. The processor can queue one write cycle internally, pending the return of bus access; if more
than one write cycle is needed, program execution is delayed until HOLD is released and the pro-
cessor regains control of the bus.
HOLD has priority over most core bus cycles, but is not recognized under certain conditions:
• During locked cycles
• Between two interrupt acknowledge cycles (LOCK# asserted)
• During misaligned word transfers (LOCK# not asserted)
• During doubleword (32-bit) transfers (LOCK# not asserted)
• During misaligned doubleword transfers (LOCK# not asserted)
• During an active RESET signal (HOLD is recognized during the time between the falling
edge of RESET and the first instruction fetch)
All inputs are ignored while the processor is in the HOLD state, except for the following:
• HOLD pin - It is monitored to determine when the processor may regain control of the bus.
• RESET pin - It is of a higher priority than HOLD. An active RESET input reinitializes the
device.
6-36
BUS INTERFACE UNIT
• NMI pin - The request is recognized and latched. It is serviced after HOLD is released.
• SMI# pin - The request is recognized and latched. It is serviced after HOLD is released.
6.5.2 HOLD Signal Latency
Because other bus masters may be used in time-critical applications, the amount of time the bus
master must wait for bus access (HOLD latency) can be a critical design consideration. Because
a bus cycle must be terminated before HLDA can go active, the maximum possible latency occurs
when a bus-cycle instruction is being executed or a DMA block mode transfer is in progress. Wait
states increase latency, and HOLD is not recognized between locked bus cycles and interrupt ac-
knowledge cycles. The internal DMA may also contribute to the latency.
The HOLD latency is dependent on a number of parameters:
• The instruction being executed at the time the HOLD request occurs.
• The number of wait states during various access cycles, including the following:
— Memory wait states
— Code fetch wait states
— Interrupt acknowledge wait states
— Refresh wait states
• The priority of the requester.
• The mode of the DMA:
— Block mode
— Single cycle mode
— Demand transfer mode
6-37
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
6.6 DESIGN CONSIDERATIONS
• Upon reset, UCS# is configured as a 16-bit chip-select signal. If the Boot device is only an
8-bit device, then BS8# must be asserted whenever UCS# is active (until the UCS region
can be reprogrammed to reflect an 8-bit region). One way of doing this is by connecting the
UCS# pin directly to the BS8# pin, if there are no other devices that need to use the BS8#
pin. If UCS# is tied directly to BS8#, then the UCS region need not be programmed to
reflect an 8-bit region.
• Since LBA# may be used as an output-enable by both the internal and external READY#
buffers, care must be taken in selecting the external READY# buffer to minimize
contention on the READY# signal caused by differences in buffer characteristics.
6.6.1 Interface To Intel387™ SX Math Coprocessor
The Intel387 SX Math Coprocessor is an extension to the Intel386 EX embedded processor ar-
chitecture. The combination of the Intel387 SX Math Coprocessor with the Intel386 EX embed-
ded processor dramatically increases the processing speed of computer application software that
uses high performance floating-point operations.
An internal Power Management Unit enables the Intel387 SX Math Coprocessor to perform float-
ing-point operations while maintaining very low power consumption. The internal Power Man-
agement Unit effectively reduces power consumption by 95% when the coprocessor is idle.
This section describes special considerations for interfacing the Intel387 SX Math Coprocessor
with the Intel386 EX embedded processor. For complete information, refer to the Intel387™ SX
Math Coprocessor datasheet (Order number 240225).
6-38
BUS INTERFACE UNIT
6.6.1.1
System Configuration
The Intel387 SX Math Coprocessor can be interfaced to the Intel386 EX embedded processor as
shown in Figure 6-15.
16
W/R#
ADS#
M/IO#
A23
W/R#
ADS#
NPS1#
NPS2
CMD0#
A2
Clock
Generator
CLK2
CPUCLK2
RESETIN
D15:0
RESET
V
CC
Synchronous
Reset
BUSY#
PEREQ
ERROR#
READY#
BUSY#
CKM
STEN
PEREQ
ERROR#
READY#
D15:0
LBA#
NUMCLK2
READYO#
Intel386™ EX
Embedded Processor
Intel387™ SX
Math Coprocessor
A2852-02
Figure 6-15. Intel386 EX Processor to Intel387 SX Math Coprocessor Interface
A dedicated communication protocol makes possible high-speed transfer of opcodes and oper-
ands between the Intel386 EX processor and the Intel387 SX math coprocessor. Most control pins
of the Intel387 SX Math Coprocessor are connected directly to Intel386 EX processor pins.
6-39
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
The interface has these characteristics:
• The Intel387 SX Math Coprocessor shares the local bus of the Intel386 EX processor.
• The Intel386 EX processor and Intel387 SX Math Coprocessor share the same reset signals.
They also share the same clock input.
• The corresponding BUSY#, ERROR#, and PEREQ pins are connected together.
• The Status Enable (STEN) selects the math coprocessor. It causes the chip to recognize
other chip select inputs. STEN is tied high.
• CKM is tied high to select the synchronous mode of operation for the coprocessor.
• The math coprocessor NPS1# and NPS2 inputs are connected to the Intel386 EX processor
M/IO# and A23 inputs respectively. For math coprocessor cycles, M/IO# is always LOW
and A23 always HIGH.
• The math coprocessor input CMD0 is connected to the A2 output. The Intel386 EX
embedded processor generates address 8000F8H when writing a command and address
8000FCH or 8000FEH (treated as 8000FCH by the Intel387 SX Math Coprocessor) when
writing or reading data. It does not generate any other addresses during Intel387 SX Math
Coprocessor bus cycles.
CAUTION
A chip-select signal could go active during coprocessor cycles if a match for
the lower 16 bits of address is found in one of the chip-select regions of the
Chip-select Unit. This can happen because only the lower 16 bits are decoded
by the Chip-select Unit during I/O cycles.
• The READYO# pin of the coprocessor must be sent through a buffer to prevent the Intel386
EX processor and coprocessor from simultaneously driving the READY# pin. The buffer is
enabled using the LBA# pin. During internal bus cycles, the LBA# pin is asserted and the
Intel386 EX processor provides the READY# signal. In a coprocessor access, the LBA# is
deasserted, the external buffer is enabled, and the coprocessor provides the READY# signal
to the Intel386 EX processor.
6.6.1.2
Software Considerations
To enable math-coprocessor support in the Intel386 EX processor, you must set the MP (Math
Present) bit and clear the EM (Coprocessor Emulation) bit in the Machine Status Word (lower
half of the CR0 register in the core). This can be done using the following code:
smsw
or
ax
ax, 2
;; Store Machine Status Word into AX
;; Set MP bit
and
lmsw
ax, 0fffbh
ax
;; Clear EM bit
;; Load AX into Machine Status Word
6-40
BUS INTERFACE UNIT
Also, bit 5 in the PINCFG register (Figure 5-15 on page 5-24) must be cleared, to connect the
coprocessor-related signals of the core to the package pins.
Below is an example of a simple routine that can be executed using the math-coprocessor:
fninit
fldpi
fld1
fadd
fist
;; Initialize Math Coprocessor
;; Load (Push on to the 387 stack) “Pi”
;; Load (Push on to the 387 stack) “1”
;; Add the two values, i.e. Pi + 1
;; Convert to integer and Store at
6.6.2 SRAM/FLASH Interface
SRAM and FLASH devices can be connected directly to the Intel386 EX processor as shown in
Figure 6-16. Separate CSn#, RD# and WR# strobes enable a “glueless” interface. The WR# sig-
nal, when used with an “EARLY READY#” (described in “Write Cycle” on page 6-16), guaran-
tees the ‘WE#-Inactive-to-Address-Invalid’ time of most SRAM and FLASH devices.
SRAM
or FLASH
Intel386™ EX
Embedded
Processor
Address
Data
RD#
OE#
CSn#
WR#
CE#
WE#
A2853-02
Figure 6-16. Intel386 EX Processor to SRAM/FLASH Interface
6-41
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
6.6.3 PSRAM Interface
Pseudo SRAM (PSRAM) devices can be easily interfaced (Figure 6-17) to the Intel386 EX pro-
cessor. PSRAM devices have an interface that is similar to SRAM devices (They are also pin-
compatible in many cases). The two major differences between PSRAM and SRAM devices are:
• PSRAM devices require a CE# precharge (inactive) time between access cycles. Since the
Intel386 EX processor does not guarantee a minimum inactive time on it’s CSn# signals,
control logic is required to satisfy the PSRAM device’s CE# precharge time.
• PSRAM devices have a RFSH# input pin. This signal activates an internal refresh cycle.
The REFRESH# output of the Intel386 EX processor can be connected directly to the
PSRAM device’s RFSH# pin.
PSRAM
Intel386™ EX
Embedded
Processor
Address
Data
Control
Logic
CSn#
CE#
RD#
WR#
OE#
WE#
REFRESH#
RFSH#
Note:
Control logic is necessary to satisfy the precharge time for the CE# signal of the
PSRAM. The precharge time is specified by the PSRAM manufacturer.
A2854-02
Figure 6-17. Intel386 EX Processor to PSRAM Interface
6-42
BUS INTERFACE UNIT
6.6.4 Paged DRAM Interface
External logic is required to interface the Intel386 EX processor to DRAM devices, as shown in
Figure 6-18. The PLD generates the RAS# and CAS# signals.
If RAS#-Only Refresh is being performed (using the Refresh Control Unit of the processor), then
during a Refresh Cycle, the PLD enables the Column Address Buffer and asserts the RAS# signal
(shaded sections in the figure). Refer to Chapter 6, “BUS INTERFACE UNIT,” for more infor-
mation.
A single multiplexer can be used instead of the separate row and column address buffers.
Row
Address
Row
Address
Buffer
Upper Address
OE_ROW#
Address
REFRESH#
BHE#
CS
BLE#
RAS#
CAS#
Paged
DRAM
Intel386™ EX
Embedded Processor
PLD
n#
OE_COL#
Column
Address
Buffer
Lower Address
Column
Address
Note:
A single mux can be used in place of the row and column address buffers.
A3264-02
Figure 6-18. Intel386 EX Processor to Paged DRAM Interface
6-43
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
6.6.5 Non-Paged DRAM Interface
This interface is similar to the Paged DRAM Interface, except that in this case, the lower address
bits are routed to the Row Address Buffer and the higher address bits to the Column Address
Buffer. This is done to simplify the RAS#-Only Refresh logic. The PLD in this case enables the
Row Address Buffer and asserts the RAS# signal (shaded sections in the figure) during a Refresh
Cycle. Refer to Chapter 15, “REFRESH CONTROL UNIT,” for more information.
A single multiplexer can be used instead of the separate row and column address buffers.
Row
Address
Row
Address
Buffer
Lower Address
OE_ROW#
Address
REFRESH#
BHE#
CS
BLE#
RAS#
CAS#
Non-paged
DRAM
Intel386™ EX
Embedded Processor
PLD
n#
OE_COL#
Column
Address
Buffer
Upper Address
Column
Address
Note:
A single mux can be used in place of the row and column address buffers.
A3265-02
Figure 6-19. Intel386 EX Processor and Non-Paged DRAM Interface
6-44
7
SYSTEM
MANAGEMENT
MODE
CHAPTER 7
SYSTEM MANAGEMENT MODE
The Intel386™ EX processor provides a mechanism for system management with a combination
agement is not critical, SMM may be used for other functions such as alternate operating systems,
debuggers, hard disk drive backup, or virtual I/O.
This chapter is organized as follows:
• SMM Hardware Interface (page 7-1)
• System Management Mode Programming and Configuration (page 7-3)
• The Intel386 EX Processor Identifier Registers (page 7-15)
• Programming Considerations (page 7-16)
7.1 SYSTEM MANAGEMENT MODE OVERVIEW
An externally generated system management interrupt (SMI#) allows the execution of system-
wide routines that are independent and transparent to the operating system. The system manage-
ment mode (SMM) architectural extensions to the Intel386 CPU consist of the following ele-
ments:
• An interrupt input pin (SMI#) to invoke SMM
• An output pin (SMIACT#) to identify execution state
• A new instruction (RSM, executable only from SMM) to exit SMM
7.2 SMM HARDWARE INTERFACE
The Intel386 EX processor provides two pins for use in SMM systems: SMI# and SMIACT#.
7.2.1 System Management Interrupt Input (SMI#)
The SMI# input signal is used to invoke system management mode. SMI# is a falling edge
triggered interrupt input signal and is the highest priority of all external interrupt sources. SMI#
forces the core into SMM at the completion of the current instruction. SMI# has these
characteristics:
• SMI# is not maskable.
• SMI# is recognized on an instruction boundary and at each iteration for repeat string
instructions.
• SMI# does not break locked bus cycles.
7-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
• SMI# cannot interrupt currently executing SMM code. The processor latches the falling
edge of a pending SMI# signal while the Intel386 EX processor is executing an existing
SMI# (this allows one level of buffering). The nested SMI# is not recognized until after the
execution of a resume instruction (RSM).
• SMI# brings the processor out of idle or powerdown mode.
7.2.2 SMM Active Output (SMIACT#)
This output indicates that the processor is operating in system management mode. It is asserted
when the CPU initiates the SMM sequence and remains active (low) until the processor executes
the RSM instruction (described in “Resume Instruction (RSM)” on page 7-15) to leave SMM. Be-
fore SMIACT# is asserted, the CPU waits until the end of the instruction boundary. SMIACT# is
used to establish a new memory map for SMM operation. The processor supports this function
by an extension to the internal chip-select unit. In addition, external logic can use this pin to qual-
ify RESET and SMI#. SMIACT# never transitions during a pipelined bus cycle.
7.2.3 System Management RAM (SMRAM)
The SMM architecture requires that a partition of memory be set aside for the SMM driver. This
is called the SMRAM. Several requirements must be met by the system:
• The address range of this partition must be, as a minimum, from 038000H to 03FFFFH
(32 Kbytes).
• The address range from 03FE00H to 03FFFFH (512 bytes) is reserved for the CPU and
must be RAM.
• The SMM handler must start execution at location 038000H. It is not relocatable.
• During normal operation the SMRAM is only accessible when the system is in SMM.
• During system initialization it must be possible to access the SMRAM in order to initialize
it and possibly to install the SMM driver. Obviously, this must be done outside of SMM.
• When the SMRAM overlays other memory in the system, then address decoding and chip
selects must allow the SMM driver to access the shadowed memory locations while in
SMM.
• The SMRAM should not be accessible to alternate bus masters such as DMA.
These requirements are made to ensure that the SMM remains transparent to non-SMM code and
to maintain uniformity across the various Intel processors that support this mode.
It is possible for the designer of an embedded system to place the SMM driver
code in read-only storage, as long as the address space between 03FE00H and
03FFFFH is writable.
The Intel386 EX processor does not support SMRAM relocation. Bit 17 of the SMM Revision
Identifier (see “SMRAM State Dump Area” on page 7-14) indicates whether the processor sup-
7-2
SYSTEM MANAGEMENT MODE
ports the relocation of SMRAM. When this bit is set (1), the processor supports SMRAM reloca-
tion. When this bit is cleared (0), then the processor does not support SMRAM relocation. Since
this device doesn’t support SMRAM relocation, bit 17 of the SMM Revision Identifier is cleared.
The SMRAM address space is fixed from 38000H to 3FFFFH.
7.3 SYSTEM MANAGEMENT MODE PROGRAMMING AND CONFIGURATION
7.3.1 Register Status During SMM
When the CPU recognizes SMI# on an instruction boundary, it waits for all write cycles to com-
plete and asserts the SMIACT# pin. The processor then saves its register state to SMRAM space
and begins to execute the SMM handler. The RSM instruction restores the registers, deasserts the
SMIACT# pin, and returns to the user program.
Upon entering SMM, the processor’s PE, MP, EM, TS and PG bits in CR0 are cleared, as shown
in Table 7-1.
Table 7-1. CR0 Bits Cleared Upon Entering SMM
CR0 Bit
Mnemonic
Description
Function
0 = protection disabled
0
PE
Protection Enable
1 = protection enabled
1
2
MP
EM
TS
Math Coprocessor Present 0 = coprocessor not present
1 = coprocessor present
Emulate Coprocessor
0 = coprocessor opcodes execute
1 = coprocessor opcodes generate a fault
3
Task Switched
0 = coprocessor ESC opcode does not cause fault
1 = coprocessor ESC opcode causes fault
31
PG
Paging Enable
0 = paging disabled
1 = paging enabled
Debug register DR7 is also cleared, except for bits 11–15.
Internally, a descriptor register (invisible to the programmer) is associated with each program-
mer-visible segment register. Each descriptor register holds a 32-bit segment base address, a 32-
bit segment limit, and other necessary segment attributes. When a selector value is loaded into a
segment register, the associated descriptor register is automatically updated with the correct in-
formation. In Real mode, only the base address is updated directly (by shifting the selector value
four bits to the left), since the segment maximum limit and attributes are fixed in Real mode. In
Protected mode, the base address, the limit, and the attributes are all updated per the contents of
the segment descriptor indexed by the selector. After saving the CPU state, the SMM State Save
sequence sets the appropriate bits in the segment descriptor, placing the core in an environment
similar to Real mode, without the 64 Kbyte limit checking.
In SMM, the CPU executes in a Real-like mode. In this mode, the CPU can access (read and
write) any location within the 4 Gbyte logical address space. The physical address space is
64 Mbytes. The CPU can also perform a jump and a call anywhere within a 1 Mbyte boundary
address space. In SMM, the processor generates addresses as it does in real mode; however, there
7-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
is no 64 Kbyte limit. The value loaded into the selector register is shifted to the left four bits and
moved into its corresponding descriptor base, then added to the effective address. The effective
address can be generated indirectly, using a 32-bit register. However, only 16 bits of the Extended
Instruction Pointer (EIP) register are pushed onto the stack during calls, exceptions and INTR ser-
vices. Therefore, when returning from calls, exceptions or INTRs, the upper 16 bits of the 32-bit
EIP are zero. In an SMI# handler, the EIP should not be over the 64 Kbyte boundary. The 16-bit
CS allows addressing within a 1 Mbyte boundary.
Instructions that explicitly access the stack, such as MOV instructions, can access the entire
4 Gbytes of logical address space by using a 32-bit address size prefix. However, instructions that
implicitly access the stack, such as POP, PUSH, CALL, and RET, still have the 64 Kbytes limit.
After SMI# is recognized and the processor state is saved, the processor state is initialized to the
default values shown in Table 7-2.
Table 7-2. SMM Processor State Initialization Values
Register
Content
Unpredictable
General Purpose Register
EFLAGS
00000002H
00008000H
3000H
EIP
CS Selector
DS,ES,FS,GS,SS Selectors
CS Descriptor Base
DS,ES,FS,GS,SS Descriptor Base
0000H
00030000H
00000000H
CS,DS,ES,FS,GS,SS Descriptor Limit 0FFFFFH
DS,ES,FS,GS,SS Attributes
16-bit
CR0
DR6
DR7
Bits 0, 1, 2, 3, 31 cleared
Unpredictable
Bits 0–10,16–31 cleared
When a valid SMI# is recognized on an instruction execution boundary, the CPU immediately
begins execution of the SMM State Save sequence, asserting SMIACT# low (unless the CPU is
in a shutdown condition). The CPU then starts SMI# handler execution. An SMI# cannot inter-
rupt a CPU shutdown. The SMI# handler always starts at 38000H. When there are multiple causes
of SMI#s, only one SMI# is generated, thereby ensuring that SMI#s are not nested.
The Intel386 EX processor extends the standard Intel386 microprocessor architecture by adding
a new feature called the system management interrupt (SMI#). This section describes in detail
how the system designer uses SMI#.
The execution unit recognizes an SMI# (falling edge) on an instruction boundary (see instruction
#3 in Figure 7-1). After all CPU bus cycles have completed, including pipelined cycles, the state
7-4
SYSTEM MANAGEMENT MODE
of the CPU is saved to the SMM State Dump Area. After executing a RSM instruction, the CPU
proceeds to the next application code instruction (see instruction #4 in Figure 7-1). SMM latency
is measured from the falling edge of SMI# to the first ADS# where SMIACT# is active (see Fig-
ure 7-2).
SMI#
Instr Instr Instr
Instr Instr
#1
#2
#3
#4
#5
State
Save
SMM
Handler
State
Resume
Interrupts
Blocked
Interrupts
Blocked
SMI
Latency
SMI#
SMIACT#
2nd SMI# is blocked
A2510-02
Figure 7-1. Standard SMI#
handler is entered. (Note that the CPU does not recognize NMI while executing the SMM State
Save sequence or SMM State Resume sequence.) NMI is always enabled following the comple-
tion of the first interrupt service routine (ISR) or exception handler.
Even when the processor is in SMM, address pipelined bus cycles can be performed correctly by
asserting NA#. Pipelined bus cycles can also be performed immediately before and after SMI-
ACT# assertion. The numbers in Figure 7-2 also reflect a pipelined bus cycle.
7-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
CLK2
T1 T2
CLKOUT
SMI#
ADS#
B
D
READY#
SMIACT#
C
A
State Save, SMM Handler,
State Restore
Normal State
Normal State
A = 1 CLK min, B = 20 CLK min, C = 16 CLK min, D = 4 CLK min
A2512-02
Figure 7-2. SMIACT# Latency
NOTE
Even if bus cycles are pipelined, the minimum clock numbers are guaranteed.
7-6
SYSTEM MANAGEMENT MODE
7.3.2.1
SMI# Priority
When more than one exception or interrupt is pending at an instruction boundary, the processor
services them in a predictable order. The priority among classes of exception and interrupt sourc-
es is shown in Table 7-3. The processor first services a pending exception or interrupt from the
class that has the highest priority, transferring execution to the first instruction of the handler.
Lower priority exceptions are discarded; lower priority interrupts are held pending. Discarded ex-
ceptions are reissued when the interrupt handler returns execution to the point of interruption.
SMI# has the following relative priority, where 1 is highest and 11 is lowest:
Table 7-3. Relative Priority of Exceptions and Interrupts
1
Double Fault
(Highest priority)
2
3
Segmentation Violation
Page Fault
Divide-by-zero
SMI#
4
5
6
Single-step
Debug
7
8
ICE Break
NMI
9
10
INTR
11
I/O Lock
(Lowest Priority)
7-7
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
7.3.2.2
System Management Interrupt During HALT Cycle
Since SMI# is an asynchronous signal, it may be generated at any time. A condition of interest
arises when an SMI# occurs while the CPU is in a HALT state. To give the system designer max-
imum flexibility, the processor allows an SMI# to optionally exit the HALT state. Figure 7-3
shows that the CPU normally re-executes the HALT instruction after RSM; however, by modify-
ing the HALT restart slot in the SMM State Dump area, the SMM handler can redirect the instruc-
tion pointer past the HALT instruction.
SMI#
Instr HALT Halted State
Instr Instr
#1
#2
#3
#4
Option
State
Save
SMM
Handler
State
Resume
A2508-01
Figure 7-3. SMI# During HALT
7-8
SYSTEM MANAGEMENT MODE
7.3.2.3
HALT Restart
It is possible for SMI# to break into the HALT state. In some cases the application might want to
return to the HALT state after RSM. The SMM architecture provides the option of restarting the
HALT instruction after RSM.
The word at address 03FF02H is the HALT restart slot. The processor sets bit 0 of this location
when the processor is in the HALT state while the SMI# occurred. If the SMM driver leaves this
bit set, then the processor re-enters the HALT state when it exits from SMM. When the driver
clears this bit, the processor continues execution with the instruction just after the interrupted
7.3.2.4
System Management Interrupt During I/O Instruction
Like the HALT restart feature, the processor allows restarting I/O cycles which have been inter-
rupted by an SMI#. This gives the system designer the option of performing a hardware I/O cycle
restart without having to modify either application, operating system, or BIOS software. (See Fig-
ure 7-4.)
When a SMI# occurs during an I/O cycle, it then becomes the responsibility of the SMM handler
to determine the source of the SMI#. If, for example, the source is the powered down I/O device,
the SMM handler would power up the I/O device and reinitialize it. The SMM handler would then
write 0FFH to the I/O restart slot in the SMM State Dump area and the RSM instruction would
then restart the I/O instruction.
SMI#
Instr Instr I/O Instr
Instr Instr
#1
#2
#3
#4
#5
Option
State
Save
SMM
Handler
State
Resume
A2509-01
Figure 7-4. SMI# During I/O Instruction
The SMI# input signal can be asynchronous; as a result, SMI# must be valid at least three clock
periods before READY# is asserted for it to be recognized right after the current bus cycle. SMI#
must be sampled valid for at least two clocks, with the other clock used to internally arbitrate for
control. See Figure 7-5 for details. (Note that this diagram is only for I/O cycles and memory data
read cycles.)
7-9
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Priority Arbitration
CLK2
SMI#
Sampled
SMI#
T
T
su
hold
RDY#
T
= SMI# setup time, T
= SMI# hold time
hold
su
A2511-02
Figure 7-5. SMI# Timing
7.3.2.5
I/O Restart
Bit 16 of the SMM Revision Identifier is set (1) indicating that this device does support the I/O
trap restart extension to the SMM base architecture.
The I/O trap restart slot provides the SMM handler the option of automatically re-executing an
interrupted I/O instruction using the RSM instruction. When the RSM instruction is executed
with the I/O trap restart slot set to a value of 0FFH, the CPU automatically re-executes the I/O
instruction that the SMI# has trapped. If the slot contains 00H when the RSM instruction is exe-
cuted, the CPU does not re-execute the I/O instruction. This slot is initialized to 00H during an
SMI#. It is the SMM handler’s responsibility to load the I/O trap restart slot with 0FFH when re-
start is desired.
NOTE
The SMM handler must not set the I/O trap restart slot to 0FFH when the
SMI# is not asserted on an I/O instruction boundary, because this causes
unpredictable results.
7.3.3 SMM Handler Interruption
7.3.3.1
Interrupt During SMM Handler
When the CPU enters SMM, both INTR and NMI are disabled (Figure 7-6). The SMM handler
may enable INTR by executing the STI instruction. NMI is enabled after the completion of the
first interrupt service routine (software or hardware initiated ISR) or exception handler within the
SMM handler. Software interrupt and exception instructions are not blocked during the SMM
handler.
The SMM feature can be used without any other interrupts. INTR and NMI are blocked by the
system during SMI#, unless enabled by software. If INTR or NMI are not enabled during SMM,
7-10
SYSTEM MANAGEMENT MODE
then any pending INTR and NMI is serviced after completion of RSM instruction execution.
Only one INTR and one NMI can be pending.
The SMM handler may choose to enable interrupts to take advantage of device drivers. Since in-
terrupts were enabled while under control of the SMM handler, the signal SMIACT# continues
to be asserted. If the system designer wants to take advantage of existing device drivers that le-
verage interrupts, the memory controller must take this into account.
SMM
Intr
SMM
Application
Application
Instr Instr
Handler
Service
Handler
Instr Instr Instr
Instr Instr Instr Instr Instr Instr
SMI
Latency
SMI#
State
Restore
State
Save
SMM Handler
RSM
SMIACT#
INTR
NMI
RESET
NMI is Blocked
A2505-02
Figure 7-6. Interrupted SMI# Service
HALT During SMM Handler
7.3.3.2
The system designer may wish to place the system into a HALT condition while in SMM. The
ternally blocks INTR and NMI from being recognized until after the RSM instruction is executed.
When a HALT needs to be breakable in SMM, the SMM handler must enable INTR and NMI
before a HALT instruction execution. NMI is enabled after the completion of the first interrupt
service routine within the SMM handler.
After the SMM handler has enabled INTR and NMI, the CPU exits the HALT state and returns
to the SMM handler when INTR or NMI occurs. See Figure 7-7 for details.
7-11
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
SMI#
Instr Instr
Instr Instr
INTR or NMI
#1
#2
#3
#4
State
Save
SMM
Enable
HALT Halted
State
SMM
State
Handler INTR & NMI
Handler Resume
Interrupt
Handler
A2507-01
Figure 7-7. HALT During SMM Handler
Idle Mode and Powerdown Mode During SMM
7.3.3.3
Both Idle Mode and Powerdown Mode may be used while in SMM. Entering and exiting either
of these power management modes from SMM is identical to entering or exiting from normal
mode. The interaction between SMM and power management modes is described in Chapter 8.
7.3.3.4
SMI# During SMM Operation
If the SMI# request is asserted during SMM operation, the second SMI# cannot nest the currently
executing SMM. The second SMI# request is latched, and held pending by the CPU. Only one
SMI# request can be pending. After RSM execution is completed, the pending SMI# is serviced.
At this time, SMIACT# is deasserted once at completion of RSM, then asserted again for the sec-
ond SMI#.
When the SMM handler polls the various SMI# sources for one of the SMI# triggers, and two
SMI# sources are found in the SMI# generation circuit, the SMM handler services both SMI#
sources and executes a RSM instruction. In this SMM handler, if the SMI# generation circuit as-
serts the second SMI# during the first SMI# service routine, the second SMI# is pending. Next,
the SMM handler finds and services two SMI# sources. After the CPU completes the RSM exe-
cution, the pending SMI# (second SMI#) is generated, but there is nothing to service because the
second SMI# was serviced during the first SMM handler. This unnecessary SMI# transaction re-
quires a few hundred clocks. There may be some performance degradation if this example occurs
frequently. For good performance, it is the responsibility of the SMI# generation circuitry to man-
age multiple SMI# assertions.
7.3.4 SMRAM Programming
7.3.4.1
Chip-select Unit Support for SMRAM
The internal chip-select unit (CSU) has been extended to support the SMRAM by using bit 10 in
each Low Address (CASMM) and Low Mask register (CMSMM). The CSU acts on these bits
7-12
SYSTEM MANAGEMENT MODE
exactly as if they represented another address line. The following options are supported by the
chip select unit:
CASMM
CMSMM
Chip select active:
0
1
X
0
0
1
During normal mode only
During SMM only
During normal mode or SMM
To see how this extension of the CSU supports the SMRAM requirements, consider an embedded
system which has 1 Mbyte of 16-bit wide EPROM in the region 03F00000H to 03FFFFFFH and
1 Mbyte of 16-bit wide RAM in the region 00000000H to 000FFFFFH. A single 32 Kbyte RAM
in the region 00038000H to 0003FFFFH is added to support SMM. The chip selects for this sys-
tem during normal operation would be programmed as follows:
REGION
CA25:11
CM25:11
CASMM
CMSMM
BS16
EPROM
RAM
11 1111 0000 0000 0
00 0000 0000 0000 0
00 0000 0011 1000 0
00 0000 1111 1111 1
00 0000 1111 1111 1
00 0000 0000 0111 1
0
0
1
0
0
0
1
1
0
SMRAM
Each row in the above table represents a region of memory and its associated chip select logic.
During initialization, these same chip selects could be programmed as follows:
REGION
CA25:11
CM25:11
CASMM
CMSMM
BS16
EPROM
RAM
11 1111 0000 0000 0
00 0000 0000 0000 0
00 0001 0011 1000 0
00 0000 1111 1111 1
00 0000 1111 1111 1
00 0000 0000 0111 1
0
0
0
0
0
0
1
1
0
SMRAM
Only the SMRAM row has been changed; the SMRAM chip select has been redirected to the re-
gion 013F800H to 013FFFFH and the CASMM bit has been cleared. This allows the initialization
software to set up the SMRAM without entering the SMM. Note that the external design of the
system must guarantee that an SMI# cannot occur while the SMRAM is being initialized.
If the SMM driver needs to access the memory shadowed under the SMRAM, the chip selects
can be reconfigured as follows:
REGION
CA25:11
CM25:11
CASMM
CMSMM
BS16
EPROM
RAM
11 1111 0000 0000 0
00 0001 0000 0000 0
00 0000 0011 1000 0
00 0000 1111 1111 1
00 0000 1111 1111 1
00 0000 0000 0111 1
0
0
1
0
1
0
1
1
0
SMRAM
This leaves the SMRAM in place but moves the normal RAM into the partition 0100000H to
01FFFFFH. The CASMM bit is masked so that the RAM is selected independent of SMM.
7-13
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
7.3.4.2
SMRAM State Dump Area
The SMM State Save sequence asserts SMIACT#. This mechanism indicates to internal modules
that the CPU has entered and is currently executing SMM. The resume (RSM) instruction is only
valid when in SMM. SMRAM space is an area located in the memory address range 38000H–
3FFFFH. The SMRAM area cannot be relocated internally. SMRAM space is intended for access
by the CPU only, and should be accessible only when SMM is enabled. This area is used by the
SMM State Save sequence to save the CPU state in a stack-like fashion from the top of the
SMRAM area downward.
The CPU state dump area always starts at 3FFFFH and ends at 3FE00H. The following is a map
of the CPU state dump in the SMRAM.
Hex Address
Name
CR0
Description
03FFFC
03FFF8
Control flags that affect the processor state
Page directory base register
General condition and control flags
Instruction pointer
CR3
EFLGS
EIP
EDI
ESI
EBP
ESP
EBX
EDX
ECX
EAX
DR6
DR7
TR
03FFF4
03FFF0
03FFEC
03FFE8
03FFE4
03FFE0
03FFDC
03FFC8
03FFD4
03FFD0
03FFCC
03FFC8
03FFC4
03FFC0
03FFBC
03FFB8
03FFB4
03FFB0
03FFAC
03FFA8
Destination index
Source index
Base pointer
Stack pointer
General register
General register
General register
General register
Debug register; contains status at exception
Debug register; controls breakpoints
Task register; used to access current task descriptor
Local descriptor table pointer
General-purpose segment register
General-purpose segment register
Data segment register
LDTR
GS
FS
DS
SS
Stack segment register
Code segment register
General-purpose segment register
Reserved
CS
ES
03FFA7–03FF04
03FF02
—
—
Halt restart slot
03FF00
—
I/O trap restart slot
03FEFC
03FEFB–03FE00
—
SMM revision identifier (10000H)
Reserved
—
7-14
SYSTEM MANAGEMENT MODE
The programmer should not modify the contents of this area in SMRAM space directly. SMRAM
space is reserved for CPU access only and is intended to be used only when the processor is in
SMM.
7.3.5 Resume Instruction (RSM)
After an SMI# request is serviced, the RSM instruction must be executed to allow the CPU to
return to an application transparently after servicing the SMI#. When the RSM instruction is ex-
ecuted, it restores the CPU state from SMRAM and passes control back to the operating system.
The RSM instruction uses the special opcode of 0FAAH. The RSM instruction is reserved for the
SMI# handler and should only be executed by the SMI# handler. Any attempt to execute the RSM
outside of SMM mode results in an invalid opcode exception. At the end of the RSM instruction,
the processor drives SMIACT# high, indicating the end of an SMM routine.
7.4 THE Intel386 EX PROCESSOR IDENTIFIER REGISTERS
The processor has two identifier registers: the Component and Revision ID register and the SMM
Revision ID register. The component ID is 23H; the component revision ID is 09H. This register
can be read as 2309H. The SMM revision identifier is 10000H.
7-15
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
7.5 PROGRAMMING CONSIDERATIONS
7.5.1 System Management Mode Code Example
The following code example contains these software routines.
SerialWriteStr2
Located in SMRAM upon program execution, this routine loops
endlessly while writing a character “X” out the serial port on the
EV386EX board.
SerialWriteStr
InitSIO
Located in the main program in FLASH, this routine loops endlessly
while writing a string out the serial port before entering SMM.
Initializes the serial port including the mode, baud rate, and clock
rate.
MAIN
Executes the program once it is located in FLASH. It also configures
chip selects, copies SMM handler to SMRAM, and loops endlessly
until an SMI# is issued.
See Appendix C for the included header files.
#include “80386EX.h”
#include “EV386EX.h”
#include <string.h>
#include <conio.h>
#include <dos.h>
#if _DEBUG_ == 0
#define SIO_PORT SIO_1
#else
// _DEBUG_ must be defined on the command line
// The debugger uses SIO_0 for host communications
// Under the debugger we must avoid using SIO_0
#define SIO_PORT SIO_0
#endif
#define BAUD_CLKIN 1843200L // Clock rate of COMCLK, i.e., External clocking,
extern char far SMMString[];
extern void InitEXSystem(void);
int
DataSeg;
// For assembly data segment register init.
BYTE Buf[20];
/*********************** Function SerialWriteStr2 **************************
Parameters:
None
Returns:
None
Assumptions:
Not called from main. This function is used as a jump point and is
relocated by the main to 38000H (SRAM) for SMM.
Real/Protected Mode:
No changes required
7-16
SYSTEM MANAGEMENT MODE
---------------------------------------------------------------------------*/
void SerialWriteStr2()
/* Loops while writing a char out to the serial port */
{
_asm
{
mov
mov
ax,0x3900
ss,ax
mov sp,0x100
Forever:
mov
TstStatus:
in
dx,0xf4fd
al,dx
testal,0x20
je
TstStatus
// Code below is same as _SetEXRegByte(TransmitPortAddr,’X’)
mov
mov
out
jmp
ax,’X’
dx,0xf4f8
dx, al
Forever
}
}
/*********************** Function SerialWriteStr **************************
Parameters:
Unit
Unit number of the serial port. 0 for SIO port 0, 1 for SIO
port 1.
*str
Returns:
None
Character string to be written out the serial port.
Assumptions:
None
Real/Protected Mode
------------------------------------------------------------------------*/
void SerialWriteStr(int Unit, const char far *str)
{
WORD TransmitPortAddr;
WORD StatusPortAddr;
// Set Port base, based on serial port used
TransmitPortAddr = (Unit ? TBR1 : TBR0);
StatusPortAddr = (Unit ? LSR1 : LSR0);
for( ; *str != ‘\0’; str++)
{
// Wait until buffer is empty
while(!(_GetEXRegByte(StatusPortAddr) & SIO_TX_BUF_EMPTY)) ;
// Write Character
_SetEXRegByte(TransmitPortAddr,*str);
}
7-17
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
}
/*************************** Function InitSIO *******************************
Parameters:
Unit
Unit number of the serial port. 0 for SIO port 0, 1 for SIO
port 1.
Mode
Defines parity, number of data bits, number of stop bits...
Reference Serial Line Control register for various options
ModemCntrl Defines the operation of the modem control lines
BaudRate
Specifies baud rate. The baud divisor value is calculated
based on clocking source and clock frequency. The clocking
frequency is set by calling the InitializeLibrary function.
Specifies the serial port clocking rate, for internal clocking
= CLK2 for external = COMCLK
ClockRate
Returns:
Error Codes
E_INVAILD_DEVICE
E_OK
-- Unit number specifies a non-existing device
-- Initialized OK, No error.
Assumptions:
SIOCFG Has already been configured for Clocking source and Modem control
source
REMAPCFG register has Expanded I/O space access enabled (ESE bit set).
The processor Port pin are initialized separately.
Real/Protected Mode
No changes required.
--------------------------------------------------------------------------*/
int InitSIO(int Unit,
BYTE Mode,
BYTE ModemCntrl,
DWORD BaudRate,
DWORD BaudClkIn)
{
WORD SIOPortBase;
WORD BaudDivisor;
// Check for valid unit
if(Unit > 1)
return E_INVALID_DEVICE;
// Set Port base based on serial port used
SIOPortBase = (Unit ? SIO1_BASE : SIO0_BASE);
// Initialized Serial Port registers
// Calculate the baud divisor value, based on baud clocking
BaudDivisor = (WORD)(BaudClkIn / (16*BaudRate));
// Turn on access to baud divisor register
_SetEXRegByte(SIOPortBase + LCR, 0x80);
// Set the baud rate divisor register, High byte first
7-18
SYSTEM MANAGEMENT MODE
_SetEXRegByte(SIOPortBase + DLH, HIBYTE(BaudDivisor) );
_SetEXRegByte(SIOPortBase + DLL, LOBYTE(BaudDivisor) );
// Set Serial Line control register
_SetEXRegByte(SIOPortBase + LCR, Mode); // Sets Mode and resets the
// Divisor latch
// Set modem control bits
_SetEXRegByte(SIOPortBase + MCR, ModemCntrl);
return E_OK;
}
/******************************* MAIN ***********************************/
Parameters:
None
Returns:
None
Assumptions:
None
Real/Protected Mode
No changes required.
--------------------------------------------------------------------------*/
#ifndef SetEXRegWordInline
#define SetEXRegWordInline(address, word) \
_asm mov dx, address;
_asm mov ax, word;
_asm out dx, ax;
\
\
#endif
void main(void)
{
InitSIO(SIO_PORT,
SIO_8N1,
// Which Serial Port
// Mode, 8-data, no parity, 1-stop
// Modem line controls
// Baud Rate
SIO_MCR_RTS+SIO_MCR_DTR,
9600,
BAUD_CLKIN);
// Baud Clocking Rate
_asm
// Store registers to preserve values
{
push DI
push SI
push DS
push ES
}
SetEXRegWordInline(CS4ADL, 0x702);
SetEXRegWordInline(CS4ADH, 0x0);
SetEXRegWordInline(CS4MSKL, 0xFC01);
SetEXRegWordInline(CS4MSKH, 0x0);
// Configure chip select 4
7-19
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
SetEXRegWordInline(CS2ADL,0x08700);
SetEXRegWordInline(CS2ADH,0x3);
SetEXRegWordInline(CS2MSKL,0x07C01);
SetEXRegWordInline(CS2MSKH,0x00);
// Enables SRAM as memory
_asm
// Copy SMM_EXAM.BIN code into SRAM
{
mov ax,0x3800
mov es,ax
// Starting address for SMM_EXAM file
// to be placed
mov ax,seg SerialWriteStr2
mov ds,ax
// Address where SMM_EXAM is located
// Length of SMM_EXAM file in bytes
mov cx,0x100
mov si,offset SerialWriteStr2
mov di,0
rep movsb
}
SetEXRegWordInline(CS2MSKL,0x7801); // Resets SRAM to enabled in SMM only
_asm
{
// Restore register values
pop DI
pop SI
pop DS
pop ES
}
// Loop endlessly and display another serial message
// Serial Write Loop
while(1)
{
SerialWriteStr(SIO_PORT,SMMString);
}
}
/**************************** END MAIN **********************************/
7-20
8
CLOCK AND
POWER
MANAGEMENT
UNIT
CHAPTER 8
CLOCK AND POWER MANAGEMENT UNIT
tegrated peripherals. The power management features control the clock signals to provide power
conservation options.
This chapter is organized as follows:
• Controlling the PSCLK Frequency (page 8-7)
• Controlling Power Management Modes (page 8-8)
• Design Considerations (page 8-11)
• Programming Considerations (page 8-13)
8.1 OVERVIEW
ment, and system reset circuitry. It also provides a clock output signal (CLKOUT) for synchro-
nizing external logic to the processor’s system clock. CLKOUT is the PH1P clock.
8.1.1 Clock Generation Logic
An external oscillator must provide an input signal to CLK2, which provides the fundamental
timing for the processor. As Figure 8-1 shows, the clock generation circuitry includes two divide-
by-two counters and a programmable clock divider. The first divide-by-two counter divides the
CLK2 frequency to generate two clocks (PH1 and PH2). For power management, independent
clock signals are routed to the core (PH1C and PH2C) and to the internal peripherals (PH1P and
PH2P).
The second divide-by-two counter divides the processor clock to generate a clock input (SER-
CLK) for the baud-rate generators of the asynchronous and synchronous serial I/O units. The
SERCLK frequency is half the internal clock frequency, or CLK2/4.
The programmable divider generates a prescaled clock (PSCLK) input for the timer/counter and
synchronous serial I/O units. The maximum PSCLK frequency is the internal clock frequency di-
vided by 2 (CLK2/4) and the minimum is the internal clock frequency divided by 513
(CLK2/1026).
8-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Three of the internal peripherals have selectable clock sources.
• The asynchronous serial I/O (SIO) unit can use either the SERCLK signal or an external
clock (connected to the COMCLK pin) as its clock source.
• The synchronous serial I/O (SSIO) unit can use either the SERCLK signal or the PSCLK
signal.
• The timer/counters can use either the PSCLK signal or an external clock connected to the
TMRCLKn input pin.
The individual peripheral chapters explain how to select the clock inputs.
INT
Power
(From ICU)
Management
IDLE
To WDT
NMI
SMI#
PWRCON
PWRDN
PWRDOWN
(pin mux)
RESET
Async Reset
Processor Clock
PH1/PH2
PH1C
PH2C
To Core
To Core
Core
÷2
CLK2
Buffer
PH1P
PH2P
To Peripherals
CLKOUT
Peripheral
Buffer
To Peripherals
SERCLK
÷2
To SIO0
To SIO1
To SSIO
Programmable
Divider
PSCLK
To Timer
To SSIO
CLKPRS
A2470-02
Figure 8-1. Clock and Power Management Unit Connections
8-2
CLOCK AND POWER MANAGEMENT UNIT
The signal from the RESET pin is also routed to the clock generation unit, which synchronizes
the processor clock with the falling edge of the RESET signal and provides a synchronous inter-
nal RESET signal to the rest of the device. The RESET falling edge can occur in either PH1 or
PH2. If RESET falls during PH1, the clock generation circuitry inserts a PH2, so that the next
phase is PH1 (Figure 8-2). If it falls during PH2, the next phase is automatically PH1.
NOTE
The RESET signal must be high for 16 CLK2 cycles to properly reset the
processor.
?
?
?
PH2
PH1
PH2
CLK2
PH1
PH2
RESET
A2467-01
Figure 8-2. Clock Synchronization
In addition to internal synchronization, a CLKOUT (PH1P) clock output is provided to enable
external circuitry to maintain synchronization with the Intel386 EX processor. Since it is one of
the peripheral clock signals, it remains active during idle mode, but is driven low during power-
down mode.
8.1.2 Power Management Logic
The power management circuitry provides two power management modes:
Idle Mode
Idle mode freezes the core clocks, but leaves the peripheral clocks
running. Idle mode can reduce power consumption by about half,
depending on peripheral usage.
Powerdown mode
Powerdown mode freezes both the core and peripheral clocks,
reducing current to leakage current (microamps). Peripherals that are
clocked externally (SIO, Timers, SSIO) continue to run. If inputs are
toggling, power consumption is higher.
To prepare for a power management mode, program the power control register as described in
“Controlling Power Management Modes” on page 8-8, then execute a HALT instruction. The de-
8-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
vice enters the programmed mode when the HALT cycle is terminated by a valid READY#. This
READY# may be generated either internally or externally.
A device reset, an NMI or SMI#, or any unmasked interrupt request from the interrupt control
unit causes the device to exit the power management mode. After a reset, the CPU starts execut-
ing instructions at 3FFFFF0H and the device remains in normal operation. After an interrupt, the
CPU executes the interrupt service routine, then returns to the instruction following the HALT
that prompted the power management mode. Unless software modifies the power control register,
the next HALT instruction returns the device to the programmed power management mode.
8.1.2.1
When the processor receives an SMI# interrupt while it is in idle or powerdown mode, it exits the
power management mode and enters System Management Mode (SMM). Upon exiting SMM,
software can check whether the processor was in a halt state before entering SMM. If it was, soft-
ware can set a flag that returns the processor to the halt state when it exits SMM. Assuming the
power control register bits were not altered in SMM, the processor re-enters idle or powerdown
when it exits SMM. Figure 8-3 illustrates the relationships among these modes.
8-4
CLOCK AND POWER MANAGEMENT UNIT
Halt Instruction
with Powerdown
Flag Set
Halt Instruction
with Idle Flag Set
Normal
Operation
U
n
r
R
a
e
s
s
I
n
I
e
I
o
o
R
e
r
r
k
N
t
r
r
r
o
Powerdown
Mode
Idle
Mode
RSM with
Powerdown Flag
and Halt Restart
Slot Set
RSM Instruction
with Idle Flag and
Halt Restart Slot Set
System
Management
Mode
Reset or
RSM Instruction
with Halt Restart
Slot Clear
A2229-03
Figure 8-3. SMM Interaction with Idle and Powerdown Modes
Bus Interface Unit Operation During Idle Mode
8.1.2.2
The bus interface unit (BIU) can process DMA, DRAM refresh, and external hold requests during
idle mode. When the first request occurs, the core wakes up long enough to relinquish bus control
to the bus arbiter, then returns to idle mode. For the remaining time in idle mode, the bus arbiter
controls the bus. DMA, DRAM refresh, and external hold requests are processed in the same way
as during normal operation.
8.1.2.3
Watchdog Timer Unit Operation During Idle Mode
When the watchdog timer unit is in system watchdog mode, idle mode stops the down-counter.
Since no software can run while the CPU is idle, a software watchdog is not needed. When it is
in bus monitor or general-purpose timer mode, the watchdog timer unit continues to run while the
device is in idle mode. (Chapter 17 describes the watchdog timer unit.)
8-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
8.1.3 Clock and Power Management Registers and Signals
Table 8-1 lists the registers and Table 8-2 list the signals associated with the clock and power man-
agement unit.
Table 8-1. Clock and Power Management Registers
Expanded
Address
Register
CLKPRS
PWRCON
Description
Clock Prescale:
0F804H
This register contains the programmed divisor value used to generate PSCLK
from the internal clock.
Power Control:
0F800H
This register selects the power management mode and internal ready options.
Table 8-2. Clock and Power Management Signals
Device Pin or
Signal
Description
Internal Signal
Input Clock:
CLK2
Device pin
Connect an external clock to this pin to provide the fundamental timing for
the microprocessor.
Output Clock:
CLKOUT
IDLE
Device pin
Internal signal
Internal signal
Device pin
CLKOUT is a Phase 1 output clock (PH1P)
Idle Output (to the Watchdog Timer Unit):
IDLE indicates that the device is in idle mode.
Interrupt Input (from the Interrupt Control Unit):
INTR
INT causes the device to exit powerdown or idle mode.
Nonmaskable Interrupt Input:
NMI
NMI causes the device to exit powerdown or idle mode.
Prescaled Clock Output:
PSCLK is one of two possible clock inputs for the SSIO baud-rate
generator and the Timer/counter Unit. The PSCLK frequency is controlled
by the CLKPRS register.
PSCLK
Internal signal
Powerdown Output (multiplexed with P3.6):
PWRDOWN
RESET
Device pin
Device pin
A high state on the PWRDOWN pin indicates that the device is in
powerdown mode.
System Reset Input:
This signal resets the processor and causes the device to exit powerdown
or idle mode.
Serial Clock Output:
SERCLK
SMI#
Internal signal
Device pin
SERCLK is one of two possible clock inputs for the SIO or SSIO baud-
rate generator. The SERCLK frequency is one-fourth the CLK2 frequency.
System Management Interrupt Input:
SMI# causes the device to exit powerdown or idle mode and causes the
processor to enter System Management Mode.
8-6
CLOCK AND POWER MANAGEMENT UNIT
8.2 CONTROLLING THE PSCLK FREQUENCY
The PSCLK signal can provide a 50% duty cycle prescaled clock to the timer/counter and SSIO
units. This feature is useful for providing various frequencies, including a 1.19318 MHz output
for a PC-compatible system timer, or speaker tone generator. Determine the required prescale val-
ue using the following formula, then write this value to the CLKPRS register (Figure 8-4).
internal clock frequency (CLK2/2)
Prescale value = ---------------------------------------------------------------------------------------- – 2
desired PSCLK frequency
Clock Prescale Register
CLKPRS
Expanded Addr: F804H
ISA Addr:
—
(read/write)
Reset State:
0000H
15
8
0
—
—
—
—
—
—
—
PS8
PS0
7
PS7
PS6
PS5
PS4
PS3
PS2
PS1
Bit
Number
Bit
Mnemonic
Function
15–9
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
8–0
PS8:0
Prescale Value:
These bits determine the divisor that is used to generate PSCLK. Legal
values are from 0000H (divide by 2) to 01FFH (divide by 513).
divisor = PS8:0 + 2
Figure 8-4. Clock Prescale Register (CLKPRS)
To change the frequency of PSCLK, write a new value to the CLKPRS register. The new frequen-
cy takes effect at the first high-to-low transition of PSCLK after CLKPRS has been changed.
8-7
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
8.3 CONTROLLING POWER MANAGEMENT MODES
Two power management modes are available: idle and powerdown. These modes are clock dis-
tribution functions controlled by the power control register (PWRCON), shown in Figure 8-5.
Power Control Register
PWRCON
Expanded Addr:
ISA Addr:
F800H
—
(read/write)
Reset State:
00H
7
0
—
—
—
—
WDTRDY HSREADY
PC1
PC0
Bit
Number
Bit
Mnemonic
Function
7–4
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
3
WDTRDY
HSREADY
PC1:0
Watch Dog Timer Ready:
0 = An external READY must be generated to terminate the cycle when
the WDT times out in Bus Monitor Mode.
1 = Internal logic generates READY# to terminate the cycle when the
WDT times out in Bus Monitor Mode.
2
Halt/Shutdown Ready:
0 = An external ready must be generated to terminate a HALT/Shutdown
cycle.
1 = Internal logic generates READY# to terminate a HALT/Shutdown
cycle.
1–0
Power Control:
Program these bits, then execute a HALT instruction. The device enters
the programmed mode when READY# (internal or external) terminates
the halt bus cycle. When these bits have equal values, the HALT
instruction causes a normal halt and the device remains in active mode.
PC1 PC0
0
1
0
1
0
0
1
1
active mode
idle mode
powerdown mode
active mode
Figure 8-5. Power Control Register (PWRCON)
8-8
CLOCK AND POWER MANAGEMENT UNIT
8.3.1 Idle Mode
Idle mode freezes the core clocks (PH1C low and PH2C) high, and leaves the peripheral clocks
(PH1P and PH2P) toggling. To enter idle mode:
1. Program the PWRCON register (Figure 8-5).
2. Execute a HALT instruction.
3. The CPU enters idle mode when READY# terminates the halt bus cycle.
NOTE
CLKOUT continues to run while the CPU is in idle mode.
PH1
PH2
?
?
CLK2
PH1C
PH2C
CLKOUT/PH1P
PH2P
PH2
PH1
PH2
PH1
CLK2
PH1C
PH2C
CLKOUT/PH1P
PH2P
A2468-02
Figure 8-6. Timing Diagram, Entering and Leaving Idle Mode
8-9
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
8.3.2 Powerdown Mode
Powerdown mode freezes both the core clocks and the peripheral clocks (PH1C and PH1P low,
PH2C and PH2P high). The BIU cannot acknowledge DMA, refresh, and external hold requests
in powerdown mode, since all the clocks are frozen.
To enter powerdown mode, follow these steps:
1. Program the PWRCON register (Figure 8-5).
2. Execute a HALT instruction.
3. The CPU enters powerdown mode when READY# (internal or external) terminates the
halt bus cycle.
When P3.6/PWRDOWN is configured as a peripheral pin, the pin goes high when the clocks stop,
to indicate that the device is in powerdown mode. (Chapter 16 explains how to configure the pin
as either a peripheral pin or a general-purpose I/O port pin.)
8.3.3 Ready Generation During HALT
A halt cycle, like all other CPU bus cycles, requires a valid READY# to complete. This ready can
be generated by either external logic, or from the internal bus interface unit (BIU). Setting bit 2
of the PWRCON causes the READY# to be generated by the internal BIU, and clearing bit 2
requires it to be generated by external logic. When READY# is generated internally the LBA#
signal is driven low.
External logic can use the PWRDOWN output to control other system components and prevent
DMA and hold requests.
NOTE
When the processor exits Powerdown Mode, use the CLKOUT pin for
external synchronization with the processor clock.
8-10
CLOCK AND POWER MANAGEMENT UNIT
PH1
PH2
?
?
CLK2
CLKOUT/PH1P/PH1C
PH2P/PH2C
PWRDOWN
PH2
PH1
PH2
PH1
CLK2
CLKOUT/PH1P/PH1C
PH2P/PH2C
PWRDOWN
A2469-02
Figure 8-7. Timing Diagram, Entering and Leaving Powerdown Mode
8.4 DESIGN CONSIDERATIONS
This section outlines design considerations for the clock and power management unit.
8.4.1 Reset Considerations
External circuitry must provide an input to the RESET pin. The RESET input must remain high
for at least 16 CLK2 cycles to reset the chip properly.
The RESET pin signal is routed directly to the device’s bidirectional pins. Even in idle or power-
down, a device reset floats the bidirectional pins and turns on the weak pull-up or pull-down tran-
sistors.
The clock generation logic generates a synchronous internal RESET signal for the internal pe-
ripherals. If you need a synchronous RESET signal for other system components, you can use a
simple circuit such as the one shown in Figure 8-8 to generate it. Otherwise, the CPU does not
need a synchronous reset.
8-11
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Synchronous
Reset Signal
to chip and
other system
logic.
Asynchronous RESET
CLK2
D
Q
D
Q
A2465-02
Figure 8-8. Reset Synchronization Circuit
8.4.2 Power-up Considerations
8.4.2.1 Built-in Self Test
The Intel386 EX processor supports the Intel386 SX processor built-in self-test (BIST) mode for
testing core functions. To initiate the self test, follow these steps:
1. Hold the RESET pin high for a minimum of 80 CLK2 cycles.
2. Transition the RESET pin from high to low while keeping the BUSY# pin asserted. The
BUSY# input should be asserted at least eight CLK2 cycles before the falling edge of
RESET and must be kept asserted for at least eight CLK2 cycles after the falling edge of
RESET.
20
Once BIST has been initiated, it takes approximately 2 processor clock cycles to complete. At
the completion of the BIST, the processor performs an internal reset and begins normal operation.
8.4.2.2
JTAG Reset
The processor supports an IEEE 1149.1 compliant JTAG boundary scan. The JTAG unit has its
own clock and RESET signals, independent from the rest of the processor. The processor requires
that the JTAG unit be reset before normal operation can begin. To reset the JTAG unit, invert the
processor RESET signal and connect this inverted RESET signal to the TRST# pin.
8-12
CLOCK AND POWER MANAGEMENT UNIT
8.4.3 Powerdown Mode and Idle Mode Considerations
• The “wake-up” signals (INT, NMI, and SMI#) are level-sensitive inputs to the wake-up
circuitry. The active state of any of these inputs prevents the device from entering
powerdown or idle mode.
• The refresh control unit cannot perform DRAM refreshes during powerdown.
• Powerdown mode freezes PSCLK and SERCLK.
• When the device exits powerdown mode, the PWRDOWN signal is synchronized with
CLK2 (at the falling edge of PWRDOWN) so that other devices in the system exit
powerdown at the same internal clock phase as the processor.
• The INTR output of the ICU cannot be masked off to the power management unit using the
CLI instruction. If it is necessary to mask off INTR to the power management unit, all the
interrupt inputs to the 82C59As must be masked. This applies to both powerdown and idle
modes.
8.5 PROGRAMMING CONSIDERATIONS
8.5.1 Clock and Power Management Unit Code Example
This section contains these software routines:
Set_Prescale_Value
Enter_Idle_Mode
Sets the clock prescale value.
Programs the Intel386 EX processor for idle mode.
Enter_Powerdown_Mode
Programs the Intel386 EX processor for powerdown
mode.
Mode_Setting_to_Active
Returns the Intel386 EX processor to active mode.
See Appendix C for the included header files.
#include <conio.h>
#include “80386ex.h”
#include “EV386EX.h”
/*****************************************************************************
Set_Prescale_Value:
Description:
This function sets the clock prescale value.
Parameters:
Prescale
Prescale value
Returns: Error Codes
E_BAD_VECTOR
E_OK
-- Specified Prescale is invalid
-- Initialized OK, No error.
Assumptions:
8-13
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
None
Syntax:
int error;
WORD psclk = 0x02;
error = Set_Prescale_Value(psclk);
Real/Protected Mode:
No changes required.
******************************************************************************/
int Set_Prescale_Value(WORD Prescale)
{
WORD clkprs = 0x0000;
clkprs = _GetEXRegWord(CLKPRS);
/* clear lowest nine bits of clkprs */
clkprs = clkprs & 0xfe00;
/* check that prescale value is only 9 bits in length */
if (Prescale != (Prescale & 0x01ff))
return(E_BADVECTOR);
_SetEXRegWord(CLKPRS, (clkprs | Prescale));
return(E_OK);
}/*Set_Prescale_Value*/
/******************************************************************************
Enter_Idle_Mode:
Description:
This function programs the 386EX for Idle mode. This freezes the
core clocks while leaving the peripheral clocks toggling.
Parameters:
None
Returns:
None
Assumptions:
None
Syntax:
Enter_Idle_Mode();
Real/Protected Mode:
8-14
CLOCK AND POWER MANAGEMENT UNIT
No changes required.
******************************************************************************/
void Enter_Idle_Mode(void)
{
BYTE pwrcon = 0x00;
pwrcon = _GetEXRegByte(PWRCON);
/* clear lowest two bits of pwrcon */
pwrcon = pwrcon & 0xfc;
/* Set mode to idle */
_SetEXRegByte(PWRCON, (pwrcon | IDLE));
/* call HALT instruction to execute IDLE mode */
_asm {
HLT
}
}/* Enter_Idle_Mode */
/*****************************************************************************
Enter_Powerdown_Mode:
Description:
This function programs the 386EX for Powerdown mode. This freezes
both the core and peripheral clocks.
Parameters:
None
Returns:
None
Assumptions:
None
Syntax:
Enter_Powerdown_Mode();
Real/Protected Mode:
No changes required.
*****************************************************************************/
void Enter_Powerdown_Mode(void)
{
BYTE pwrcon = 0x00;
pwrcon = _GetEXRegByte(PWRCON);
8-15
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
/* clear lowest two bits of pwrcon */
pwrcon = pwrcon & 0xfc;
/* Set mode to powerdown */
_SetEXRegByte(PWRCON, pwrcon | PWDWN);
/* call HALT instruction to execute POWERDOWN mode */
_asm {
HLT
}
}/* Enter_Powerdown_Mode */
/*****************************************************************************
Mode_Setting_To_Active:
Description:
This function returns the 386EX to Active mode. Thus, the next
HALT instruction will not invoke the Idle or Powerdown Mode.
Parameters:
None
Returns:
None
Assumptions:
None
Syntax:
Mode_Setting_To_Active();
Real/Protected Mode:
No changes required.
******************************************************************************/
void Mode_Setting_To_Active(void)
{
BYTE pwrcon = 0x00;
pwrcon = _GetEXRegByte(PWRCON);
/* clear lowest two bits of pwrcon */
pwrcon = pwrcon & 0xfc;
/* Set mode to active */
_SetEXRegByte(PWRCON, pwrcon | ACTIVE);
}/*Mode_Setting_To_Active*/
8-16
9
INTERRUPT
CONTROL UNIT
CHAPTER 9
INTERRUPT CONTROL UNIT
The Interrupt Control Unit (ICU) consists of two cascaded interrupt controllers, a master and a
slave, that allow internal peripherals and external devices (through interrupt pins) to interrupt the
core through its interrupt input.
The interrupt control unit is functionally identical to two industry-standard 82C59As connected
in cascade. The system supports a maximum of 15 simultaneous interrupt sources, which can be
individually or globally disabled. The ICU passes the interrupts on to the core based on a pro-
grammable priority structure.
Though the ICU can only handle a maximum of 15 simultaneous sources, a total of 18 interrupt
you can cascade additional 82C59As to six of the external interrupt pins (the pins that connect to
the master 82C59A only).
• ICU operation (page 9-4)
• Register Definitions (page 9-15)
• Design Considerations (page 9-29)
• Programming Considerations (page 9-32)
9.1 OVERVIEW
The ICU consists of two 82C59As configured as master and slave. Each 82C59A has eight inter-
rupt request (IR) signals. The master has seven interrupt sources and a slave 82C59A connected
to its IR signals. The slave has nine interrupt sources connected to its IR signals (two sources are
multiplexed into IR1). The interrupts can be globally or individually enabled or disabled.
The master can receive multiple interrupt requests at once. It can also receive a request while the
core is already processing another interrupt. The master uses a programmable priority structure
that determines:
• The order in which to process multiple interrupt requests
• Which requests can interrupt the processing of other requests
When the master receives an interrupt request, it checks to see that the interrupt is enabled and
determines its priority. If the interrupt is enabled and has sufficient priority, the master sends the
request to the core. This causes the core to initiate an internal interrupt acknowledge cycle.
9-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
The slave 82C59A is cascaded from (or connected to) the master’s IR2 signal. Like the master,
the slave uses a programmable priority structure. When the slave receives an interrupt request, it
sends the request to the master (assuming the request is enabled and has sufficient priority). The
master sees the slave request as a request on its IR2 line. The master then sends the request to the
core (assuming the request is enabled and has sufficient priority) and the core initiates an internal
interrupt acknowledge cycle.
The internal interrupt acknowledge cycle consists of two pulses that are sent to the 82C59A IN-
TA# inputs. This cycle causes the 82C59A that received the original interrupt request to put the
request’s vector number on the bus. The master’s cascade signals (CAS2:0) determine which
82C59A is being acknowledged (i.e., which 82C59A needs to put the vector number on the bus).
The core uses its processing mode (real or protected) and the vector number to find the address
of the interrupt service routine.
The master 82C59A has six device pins (INT9:8, INT3:0) connected to it. You can cascade addi-
tional external 82C59A slaves to these pins to increase the number of possible interrupt sources.
The external interrupt signals, INT9:8, are multiplexed with the internal asynchronous serial I/O
interrupt signals, SIOINT0 and SIOINT1. On the slave 82C59A, the external interrupt signal,
INT6, and the DMA Unit’s DMAINT signal, can be swapped before connecting to the slave’s IR4
and IR5 inputs (see Figure 9-1). The core initiates interrupt acknowledge cycles for the internal
82C59As. External logic must decode the bus signals (M/IO#, D/C#, W/R# and REFRESH#, see
Table 6-2 on page 6-5) to generate external interrupt acknowledge signals. Since the cascade bus
determines which 82C59A is being acknowledged, each external slave must monitor the master’s
cascade signals to determine whether it is the acknowledged slave. For external slaves, the mas-
ter’s cascade signals (CAS2:0) can be driven (using bit 7 of the INTCFG register) onto the
A18:16 address pins.
NOTE
Since external 82C59As require the CAS2:0 signals to stay valid through the
idle states that occur between the two interrupt acknowledge cycles, and since
the processor drives these lines high during these idle states, the CAS2:0 lines
must be latched externally to ensure validity during the idle states.
9-2
INTERRUPT CONTROL UNIT
IR0
OUT0 (TCU)
VSS
8259A
Master
P3CFG.2
INTCFG.6
P3CFG.2
1
INT0
(P3.2)†
0
1
IR1
IR2
To/From I/O Port 3
0
INT
MCR1.3
SIOINT1
1
INTR
(to
core)
0
1
SIOINT1
SIOINT0
1
0
IR3
INTCFG.6
1
INT8
TMROUT1
(P3.1)
P3CFG.1
0
OUT1(TCU)
MCR0.3
P3.1
0
INTCFG.5
0
1
SIOINT0
IR4
IR5
1
0
INTCFG.5
1
P3GFG.0
0
1
INT9
TMROUT0
(P3.0)
P3CFG.3
P3CFG.4
0
OUT0(TCU)
To/From I/O Port 3
To/From I/O Port 3
P3.0
VSS
VSS
VSS
0
1
P3CFG.3
1
1
INT1
(P3.3)
0
0
1
P3CFG.4
IR6
IR7
INT2
(P3.4)
CAS2:0
0
P3CFG.5
INTCFG.0
INTCFG.1
0
1
P3CFG.5
1
INT3
(P3.5)
To/From I/O Port 3
To TCU
0
VSS
0
1
IR0
IR1
INT4
INT
(TMRCLK0)
8259A
Slave
SSIOINT
0
1
INT5
To TCU
(TMRGATE0)
OUT1(TCU)
OUT2(TCU)
IR2
IR3
INTCFG.4
DMAINT
To TCU
0
1
IR4
IR5
INT6
(TMRCLK1)
0
1
VSS
0
1
CAS2:0
INTCFG.3
INTCFG.2
VSS
0
1
IR6
IR7
INT7
To TCU
(TMRGATE1)
WDTOUT#
VSS
INTCFG.7
0
1
3
CAS2:0
(A18:16)
A18:16
† Alternate pin signals are in parentheses
Heavier lines indicate multiple signals.
A2522-03
Figure 9-1. Interrupt Control Unit Configuration
9-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
9.2 ICU OPERATION
The following sections describe the ICU operation. The ICU’s interrupt sources, interrupt priority
structure, interrupt vectors, interrupt processing, and polling mode are discussed.
9.2.1 Interrupt Sources
The ICU support a total of 18 interrupt sources (see Table 9-1) but only a maximum of 15 simul-
taneous sources. Eight of these sources are internal peripherals and ten are external device pins
(INT9:0). However, IR3 and IR4 of the master can be connected to either SIOINT1 and SIOINT0
(internal Asynchronous Serial I/O interrupts), or to external device pins INT8 and INT9, respec-
tively. Similarly, IR1 of the slave can be connected to either SSIOINT (internal Synchronous Se-
rial I/O interrupt), or to external device pin INT5. On the slave, the external interrupt signal,
INT6, and the DMA Unit’s DMAINT signal can be swapped before connecting to the slave’s IR4
and IR5 inputs
The device pins (INT3:0) are multiplexed with port pins. When the port pin function (rather than
the interrupt function) is enabled at the pin, VSS is internally connected to the ICU’s respective
interrupt request input. The device pins, INT7, INT6, and INT4, must be enabled (using register
bits) in order to be used. The port 3 configuration register (P3CFG) controls INT3:0 interrupt
source connections, and the interrupt configuration register (INTCFG) controls the INT9:4 inter-
rupt source connections. The modem control registers (MCR1 and MCR0) are also used to con-
trol the INT9:8 interrupt source connections.
9-4
INTERRUPT CONTROL UNIT
Table 9-1. 82C59A Master and Slave Interrupt Sources
Master IR
Line
Connected
by
Slave
IR Line
Connected
by
Source
Source
IR0
TMROUT0
Hardwired
IR0
VSS
INTCFG.0=0
INTCFG.0=1
(timer control unit)
VSS
INT4
(device pin)
IR1
P3CFG.2=0
P3CFG.2=1
Hardwired
IR1
SSIOINT
(SSIO unit)
INTCFG.1=0
INTCFG.1=1
Hardwired
INT0
(device pin)
INT5
(Device pin)
IR2
IR3
Slave 82C59A
Cascade
IR2
IR3
TMROUT1
(timer control unit)
SIOINT1
(SIO unit)
INTCFG.6=0
P3CFG.1=0
TMROUT2
(timer control unit)
Hardwired
INT8
INTCFG.6=1
P3CFG.1=1
MCR0.3=1
(device pin)
IR4
IR5
SIOINT0
(SIO unit)
INTCFG.5=0
P3CFG.0=0
IR4
IR5
DMAINT
(DMA unit)
INTCFG.4=0
INTCFG.4=1
INT9
INTCFG.5=1
P3CFG.0=1
MCR1.3=1
INT6
(device pin)
(device pin)
VSS
P3CFG.3=0
INT6
INTCFG.4=0
INTCFG.2=1
(device pin)
INT1
(device pin)
P3CFG.3=1
DMAINT
(DMA unit)
INTCFG.4=1
INTCFG.2=1
IR6
IR7
VSS
P3CFG.4=0
P3CFG.4=1
IR6
IR7
VSS
INTCFG.3=0
INTCFG.3=1
INT2
(device pin)
INT7
(device pin)
VSS
P3CFG.5=0
P3CFG.5=1
WDTOUT#
(watchdog timer)
Hardwired
INT3
(device pin)
9-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Interrupt processing begins with the assertion of an IR signal. During the ICU initialization pro-
cess (described in “Register Definitions” on page 9-15), you can program the ICU to be either
edge-triggered or level-triggered. See “Interrupt Detection” on page 9-29 for a description of the
difference between level and edge triggered signals.
9.2.2 Interrupt Priority
Each 82C59A contains eight interrupt request signals. An 82C59A can receive several concurrent
interrupt requests or can receive a request while the core is servicing another interrupt. When ei-
ther condition occurs, the 82C59A uses a programmable priority structure to determine the order
in which to process the interrupts. There are two parts to the priority structure:
• Assigning an interrupt level to each IR signal
• Determining their relative priorities
9.2.2.1
Assigning an Interrupt Level
By default, the interrupt structure for each 82C59A is configured so that IR0 has the highest level
and IR7 has the lowest level. Two methods (shown in Figure 9-2) are available for changing this
interrupt structure:
Specific Rotation
This method assigns a specific IR signal as the lowest level. The
other IR signals are automatically rearranged in a circular manner.
For example, if you specify IR5 as the lowest level, IR6 becomes the
highest level, IR7 becomes the second-highest, and so on, with IR4
the second-lowest.
Automatic Rotation
This method assigns an IR signal to the lowest level after the core
services its interrupt. As with specific rotation, the other signals are
automatically rearranged in a circular manner. For example, the IR4
signal is assigned the lowest level after the core services its interrupt,
IR5 becomes the highest level, IR6 becomes the second-highest, and
so on, with IR3 the second-lowest.
9-6
INTERRUPT CONTROL UNIT
Default
Specific
Rotation
Automatic
Rotation
(Before)
Automatic
Rotation
(After)
Becomes
Highest
Level
Highest
Level
Before
Being
Highest
Level
Becomes
Highest
Level
IR6
IR7
IR0
IR1
IR2
IR3
IR4
IR5
IR4
IR5
IR6
IR7
IR0
IR1
IR2
IR3
IR5
IR6
IR7
IR0
IR1
IR2
IR3
IR0
IR1
IR2
IR3
IR4
IR5
IR6
IR7
Serviced
Assigned
Lowest
Level
After Being
Serviced
Specified
Lowest
Level
Lowest
Level
IR4
A2303-02
Figure 9-2. Methods for Changing the Default Interrupt Structure
Determining Priority
9.2.2.2
There are three modes that determine relative priorities, i.e., whether a level higher, lower, or
equal to another level has higher or lower interrupt priority.
Fully nested
In the fully nested mode, higher level IR signals have higher interrupt
priority. In this mode, when an 82C59A receives multiple interrupt
requests, it passes the highest level request to the core (or to the
master if the 82C59A is a slave). The core stops processing the lower
level request, processes the higher level request, then returns to finish
the lower level request.
Special fully nested
The special fully nested mode allows higher or equal level IR signals
to have higher interrupt priority. In this mode, if the core is
processing an interrupt, a higher or equal level interrupt request is
passed through to the core. Also, since all interrupts from the slave
are directed into a single IR line (IR2) on the master (the master does
not know the priorities of the slave interrupts it receives), this mode
enables a higher-level interrupt on the slave to interrupt the
9-7
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
processing of a lower-level slave interrupt. The special fully nested
mode is generally used by the master in a cascaded system.
Special mask
In some applications, you may want to allow lower-level requests
interrupt the processing of higher-level interrupts. The special mask
mode supports these applications. Unlike the special-fully nested and
fully nested modes, which are selected during ICU initialization, the
special mask mode can be enabled and disabled during program
operation. When special mask mode is enabled, only interrupts from
the source currently in service are inhibited. All other interrupt
requests (of both higher or lower levels) are passed on.
When the internal slave receives an interrupt request, it passes that request to the master. The mas-
ter receives all internal slave interrupt requests on its IR2 signal. This means that in fully nested
mode, higher-level slave requests cannot interrupt lower-level slave interrupts. For example, sup-
pose the slave gets an interrupt request on its IR7 signal. The slave sends the interrupt request to
the master’s IR2 signal (assuming the slave’s IR7 interrupt is enabled and has sufficient priority).
The master sends the interrupt request to the core (assume the master’s IR2 interrupt is enabled
and has sufficient priority). The core initiates an interrupt acknowledge cycle and begins process-
ing the interrupt. Next, the slave gets an interrupt request on its IR0 signal (assume IR0 is as-
signed a higher level than IR7). It then sends another IR2 to the master.
When the master is in fully nested mode, it does not relay the request to the core because the core
is in the process of servicing the previous IR2 interrupt and only a higher-level request can inter-
rupt its process (IR2 is not higher than IR2).
When the master is in special fully nested mode, the request is passed through to the core (IR2 is
equal to IR2).
9.2.3 Interrupt Vectors
Each interrupt request has a corresponding interrupt vector number. The interrupt vector number
is a pointer to a location in memory where the address of the interrupt’s service routine is stored.
The relationship between the interrupt vector number and the location in memory of the inter-
rupt’s service routine address depends on the system’s programmed operating mode (real, pro-
tected, or virtual86). Chapter 9 of the Intel386™ SX Microprocessor Programmer’s Reference
Manual explains this relationship.
During an interrupt acknowledge cycle, the ICU puts the interrupt’s vector number on the bus.
From the interrupt vector number and the system’s operating mode, the core determines where to
find the address of the interrupt’s service routine.
You must initialize each 82C59A with an interrupt vector base number. The 82C59As determine
the vector number for each interrupt request from this base number. The base vector number cor-
responds to the IR0 signal’s vector number and must be on an 8-byte boundary.
Other vector numbers are determined by adding the line number of the IR signal to the base. For
example, if the base vector number is 32, the IR5 vector number is 37. Valid vector numbers for
maskable interrupts range from 32 to 255. Because the base vector number must reside on an
8-byte boundary, the valid base vector numbers are 32 + n 8 where 0 ≤ n ≤ 27.
×
9-8
INTERRUPT CONTROL UNIT
9.2.4 Interrupt Process
Each IR signal has a mask, a pending, and an in-service bit associated with it.
• The mask bit disables the IR signal. The respective mask bits provide a way to individually
disable the IR signals. You can globally disable all interrupts to the core using the CLI
instruction. The mask bits reside in the OCW1.
• The pending bit indicates that the IR signal is requesting interrupt service. The pending bit
resides in the IRR (Interrupt Request Register, which is accessed through OCW3).
• The in-service bit indicates that the processor is in the process of servicing the interrupt.
The in-service bit resides in the ISR (Interrupt Service Register, which is accessed through
OCW3).
When the master 82C59A receives an interrupt request, it sets the corresponding pending bit and
sends the request to the core (assuming the request is enabled and has sufficient priority). The
core then initiates an acknowledge cycle: the master clears its pending bit, sets its in-service bit,
and puts the interrupt vector number on the bus.
When the slave 82C59A receives an interrupt request, it sets the corresponding pending bit and
sends the request to the master (assuming the request is enabled and has sufficient priority). When
the master receives the slave request, it sets its IR2 pending bit and sends the IR2 request to the
core (assuming the request is enabled and has sufficient priority). The core initiates an interrupt
acknowledge cycle: the master clears its IR2 pending bit and sets its IR2 in-service bit. The mas-
ter’s cascade bus activates the slave, which responds to the interrupt acknowledge cycle, clears
its pending bit, sets its in-service bit, and puts the interrupt vector number on the bus.
An 82C59A uses its in-service bits and programmed priority structure to determine whether an
interrupt has sufficient priority. The in-service bits indicate which interrupt requests are being ser-
viced. The priority structure determines whether a new interrupt request’s level has sufficient pri-
ority to interrupt the current process.
You can use one of three methods to clear an in-service bit: enable the automatic end-of-interrupt
(AEOI) mode, issue a specific end-of-interrupt (EOI) command, or issue a nonspecific EOI com-
mand. The AEOI mode is available only on the master 82C59A.
AEOI mode
This mode is enabled during system initialization. In
this mode, the 82C59A clears the in-service bit at the
beginning of an interrupt’s processing. This means
that interrupts of any level can interrupt the
processing of other interrupts.
Specific EOI command
This command instructs the 82C59A to clear a
specific IR in-service bit.
Nonspecific EOI command
This command instructs the 82C59A to clear the in-
service bit that corresponds to the highest level IR
signal active at that time.
9-9
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
NOTE
Unlike the AEOI mode (this is a mode, and not a command like specific EOI
interrupt’s service routine.
Figure 9-3 illustrates the process that takes place when the master receives a non-slave interrupt
request (which is a request on any IR signal to the master, that does not have a slave cascaded
from it). Figure 9-4 illustrates the process that occurs when a slave receives an interrupt request.
Figure 9-5 continues by showing what happens when the master receives a slave interrupt request
(for example, an IR2 request).
9-10
INTERRUPT CONTROL UNIT
Master receives an interrupt request. (From a non-slave source.)
Master sets the request's pending bit.
(operating in
fully nested
mode)
Is
Is master
operating in
special-fully
nested
Is
Yes
special
mask mode
enabled?
No
No
request
enabled?
mode?
No
Yes
Yes
Is
the
in-service
bit for this
request
set?
Is
Is
request
request
Yes
No
No
equal or higher
than any set
in-service
bits?
higher level
than any set
in-service
bits?
End
No
Yes
Yes
Master sends request to CPU. CPU initiates interrupt acknowledge cycle.
Master clears request's pending bit, sets its in-service bit, and puts its
interrupt vector number on the bus.
Master clears its in-service bit. The
CPU uses its operating mode and the
interrupt vector number to find the
interrupt service routine's address.
CPU begins processing interrupt.
Is
Yes
master in
AEOI
mode?
No
The CPU uses its operating mode and the interrupt vector number to find
the interrupt service routine's address. CPU begins processing interrupt.
The interrupt service routine sends an EOI command, causing the master
to clear its in-service bit.
An interrupt return instruction is issued, ending the interrupt process.
A2427-01
Figure 9-3. Interrupt Process – Master Request from Non-slave Source
9-11
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Slave receives an interrupt request.
Slave sets the request's pending bit.
(operating in
fully nested mode)
Is
Is
special
mask mode
enabled?
Yes
No
request
enabled?
No
Yes
Is
Is
the
request
higher
than any set
in-service
bits?
No
Yes
in-service
bit for this
request
set?
End
Yes
No
Slave sends request to master.
Note:
See the "Interrupt Process - Master Request from Slave Source" figure for the continuation of this flow chart.
A2428-01
Figure 9-4. Interrupt Process – Slave Request
9-12
INTERRUPT CONTROL UNIT
Master receives IR2 interrupt request.
Master sets its IR2 pending bit.
(operating in
fully nested
mode)
Is
Is master
operating in
special-fully
nested
Is
Yes
No
No
special
mask mode
enabled?
request
enabled?
mode?
No
Yes
Yes
Is
Is
Is
request
request
the
IR2 in-service
bit
Yes
No
No
higher level
than any set
in-service
bits?
equal or higher
than any set
in-service
bits?
End
set?
No
Yes
Yes
Master sends request to CPU. CPU initiates interrupt acknowledge cycle.
Master clears IR2 pending bit and sets IR2 in-service bit.
Slave clears its pending bit, sets its in-service bit, and puts its interrupt
vector number on the bus.
The CPU uses its operating mode and the interrupt vector number to find
the interrupt service routine's address. The CPU processes the interrupt.
Interrupt routine sends an EOI command to the slave, clearing its IR2
in-service bit
Does
slave have
other
in-service bits
set?
No
Interrupt routine sends an
EOI command to the master,
clearing its IR2 in-service bit.
Yes
An interrupt return instruction is issued, ending the interrupt process.
A2429-02
Figure 9-5. Interrupt Process – Master Request from Slave Source
9-13
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
The interrupt’s priority structure determines which EOI command should be used. Use the spe-
cific EOI command for the special mask mode. In this mode, a lower-level interrupt can interrupt
the processing of a higher-level interrupt. The specific EOI command is necessary because it al-
lows you to specifically clear the lower level in-service bit.
The fully nested mode allows only interrupts of higher levels to interrupt the processing of a low-
er-level interrupt. In this mode, the nonspecific EOI command automatically clears the in-service
bit for the current process (because it has the highest level).
Special-fully nested mode allows equal or higher level requests to interrupt the processing of oth-
er interrupts. For this mode, the nonspecific EOI command automatically clears the appropriate
in-service bit. However, when processing master IR2 interrupts, you must make sure all the slave
in-service bits are cleared before issuing the nonspecific EOI command to the master.
9.2.5 Poll Mode
The 82C59A modules can operate in a polling mode. Conventional polling requires the core to
check each peripheral device in the system periodically to see whether it requires servicing. With
the 82C59A’s polling mode, the core, by initiating the polling process, can determine whether any
of the devices attached to the 82C59A require servicing. This improves conventional polling ef-
ficiency by allowing the core to poll only the 82C59A, not each of the devices connected to it.
The polling mode is enabled by setting the polling bit in the Operation Command Word 3 register
(OCW3).
NOTE
After the polling procedure has been executed once, polling is disabled, i.e., it
is a one-shot operation. To repeat the polling procedure, the polling bit must be
set again.
The polling process takes the place of the standard interrupt process. In the standard interrupt pro-
cess, the master sends interrupt requests to the core. In the polling mode, an interrupt request can
be detected by reading the 82C59A’s poll status byte. The poll status byte indicates whether the
82C59A requires servicing. If the 82C59A requires servicing, the poll status byte indicates the
highest-priority pending interrupt request.
Polling is always a two-step process:
• A poll command is issued.
• The poll status byte is read.
When an 82C59A receives an interrupt request before it receives a poll command, it sets the re-
quest’s in-service bit and configures the poll status byte to reflect the interrupt request. The poll
status byte is used to determine which device connected to the 82C59A requires servicing. At the
end of a request’s servicing, you must issue a command to clear the request’s in-service bit.
The polling mode allows expansion of the system’s external interrupt capability. Without polling,
the system can have a maximum of 52 external interrupt sources. This is accomplished by cas-
cading six 82C59As to the master’s six external interrupt pins and using the four external inter-
rupt pins connected to the slave. The polling mode increases the system’s interrupt capability by
9-14
INTERRUPT CONTROL UNIT
configuring more than six external 82C59As. Since the polling mode doesn’t require that the ad-
ditional 82C59As be cascaded from the master, the number of interrupt request sources for a
polled system is limited only by the number of 82C59As that the system can address.
Polling and standard interrupt processing can be used within the same program. Systems that use
polling as the only method of device servicing must still fully initialize the 82C59A modules. Al-
so, the interrupt requests to the core must be disabled using the mask bits or the CLI instruction.
9.3 REGISTER DEFINITIONS
The registers associated with the ICU consist of pin and signal configuration registers, initializa-
tion command words (ICWs), operation command words (OCWs), and status registers.
• The configuration registers enable the external interrupt sources.
• The ICWs initialize the 82C59As during system initialization.
• The OCWs modify an 82C59A’s operation during program execution.
• The status registers reflect pending and in-service interrupts.
NOTE
ICW2, ICW3 and ICW4 of an 82C59A are all at the same address. Therefore a
programming sequence must be followed to program these registers. The first
access goes to ICW2, the second to ICW3 and the third to ICW4. When
completed every time.
When initializing the ICU, write first to ICW1, then to ICW2, ICW3 and
ICW4 in order.
Table 9-2 describes these registers and the following sections contain bit descriptions for each
register.
9-15
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table 9-2. ICU Registers (Sheet 1 of 2)
Expanded
Address
PC/AT*
Address
Register
P3CFG
Function
0F824H
—
Port 3 Configuration:
(read/write)
The INT3:0 signals are multiplexed with P3.5:2. This register
determines which signals are connected to the package pins.
When a P3.n signal rather than an INTn signal is connected to
a package pin, VSS is connected to the master’s IRn signal.
INTCFG
0F832H
—
Interrupt Configuration:
(read/write)
Determines the master’s and the slave’s IR signal
connections: SIOINT1 or INT8; SIOINT0 or INT9; VSS or INT7;
VSS or INT6; SSIOINT or INT5; VSS or INT4. Swaps DMAINT
and INT6. Also enables the master’s cascade bus (CAS2:0).
When enabled, the cascade signals appear on the A18:16
address lines during an interrupt acknowledge cycle.
ICW1 (master)
ICW1 (slave)
0F020H
0F0A0H
0020H
00A0H
Initialization Command Word 1:
Determines whether interrupt request signals are level
sensitive or edge triggered.
(write only)
ICW2 (master)
ICW2 (slave)
0F021H
0F0A1H
0021H
00A1H
Initialization Command Word 2:
Contains the base interrupt vector number for the 82C59A.
The base interrupt vector is the IR0 vector number, the base
plus one is the IR1 vector number, and so on.
(write only)
ICW3 (master)
(write only)
0F021H
0F0A1H
0021H
00A1H
Initialization Command Word 3:
Identifies the master’s IR signals that are connected to slave
82C59A devices. The internal slave is connected to the
master’s IR2 signal. You can connect external slaves to the
master’s IR1, IR3, IR4, IR5, IR6, and IR7 signals.
ICW3 (slave)
(write only)
Initialization Command Word 3:
Indicates that the internal slave is cascaded from the master’s
IR2 signal.
ICW4 (master)
ICW4 (slave)
0F021H
0F0A1H
0021H
00A1H
Initialization Command Word 4:
Selects either special-fully nested or fully nested mode and
enables the automatic end-of-interrupt mode.
(write only)
OCW1 (master) 0F021H
0021H
00A1H
Operation Command Word 1:
OCW1 (slave)
0F0A1H
Masks (disables) individual interrupt request signals.
(read/write)
OCW2 (master) 0F020H
0020H
00A0H
Operation Command Word 2:
OCW2 (slave)
0F0A0H
Changes interrupt levels and sends end-of-interrupt
commands.
(write only)
OCW3 (master) 0F020H
0020H
00A0H
Operation Command Word 3:
OCW3 (slave)
0F0A0H
Enables special mask mode, issues the poll command, and
allows access to the interrupt request and in-service registers.
(write only)
NOTE: All master 82C59A registers are accessed through two expanded or PC/AT addresses; all the slave
registers are accessed through two other expanded or PC/AT addresses. The order in which you write
or read these addresses along with certain register bit settings determines which register is accessed.
9-16
INTERRUPT CONTROL UNIT
Table 9-2. ICU Registers (Sheet 2 of 2)
Expanded
Address
PC/AT*
Address
Register
Function
IRR (master)
IRR (slave)
0F020H
0F0A0H
0020H
00A0H
Interrupt Request:
Indicates pending interrupt requests.
(read only)
ISR (master)
ISR (slave)
0F020H
0F0A0H
0020H
00A0H
In-service:
Indicates the interrupt requests that are currently being
serviced.
(read only)
POLL (master)
0F020H
0F021H
0020H
0021H
Poll Status Byte:
Indicates whether any of the devices connected to the 82C59A
require servicing. If the 82C59A requires servicing, this byte
indicates the highest-priority pending interrupt.
POLL (slave)
(read only)
0F0A0H
0F0A1H
00A0H
00A1H
NOTE: Once the polling bit is set in OCW3, the Poll Status
Byte of a particular 82C59A can be read by doing an access to
any of the four addresses of that 82C59A.
NOTE: All master 82C59A registers are accessed through two expanded or PC/AT addresses; all the slave
registers are accessed through two other expanded or PC/AT addresses. The order in which you write
or read these addresses along with certain register bit settings determines which register is accessed.
To initialize the 82C59As:
1. Globally disable all maskable interrupts to the core using the CLI instruction.
2. Write to the initialization command words.
NOTE
You must initialize both the master and the slave (either can be initialized
first).
The 8259A module has a state machine that controls access to the individual registers. Improper
initialization occurs when the following sequences are not followed:
• To initialize the master, write to its initialization command words in order (ICW1, ICW2,
ICW3, then ICW4).
• To initialize the slave, write to its initialization command words in order (ICW1, ICW2,
ICW3, then ICW4).
9-17
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
9.3.1 Port 3 Configuration Register (P3CFG)
Use the P3CFG register to connect the interrupt request signals (INT3:0) to the package pins.
These signals are multiplexed with port 3 signals, P3.5–2. Connecting a port 3 signal to the pack-
age pin also connects VSS to the corresponding master’s IR signal, disabling the signal.
Port 3 Configuration
P3CFG
Expanded Addr: F824H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
PM7
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7
PM7
Pin Mode:
0 = Selects P3.7 at the package pin.
1 = Selects COMCLK at the package pin.
6
5
4
3
2
PM6
PM5
PM4
PM3
PM2
Pin Mode:
0 = Selects P3.6 at the package pin.
1 = Selects PWRDOWN at the package pin.
Pin Mode:
0 = Selects P3.5 at the package pin.
1 = Connects master IR7 to the package pin (INT3).
Pin Mode:
0 = Selects P3.4 at the package pin.
1 = Connects master IR6 to the package pin (INT2).
Pin Mode:
0 = Selects P3.3 at the package pin.
1 = Connects master IR5 to the package pin (INT1).
Pin Mode:
0 = Selects P3.2 at the package pin.
1 = Connects master IR1 to the package pin (INT0).
1
0
PM1
PM0
Pin Mode:
See Table 5-1 on page 5-8 for all the PM1 configuration options.
Pin Mode:
See Table 5-1 on page 5-8 for all the PM0 configuration options.
Figure 9-6. Port 3 Configuration Register (P3CFG)
9-18
INTERRUPT CONTROL UNIT
9.3.2 Interrupt Configuration Register (INTCFG)
Use the INTCFG register to connect the INT9:4 interrupt request pins to the master’s and the
slave’s IR signals and to enable the master’s external cascade signals. When enabled, the cascade
signals appear on address lines A18:16 during interrupt acknowledge cycles. Every external slave
monitors these lines to determine whether it is the slave being addressed.
Interrupt Configuration
INTCFG
Expanded Addr: F832H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
CE
IR3
IR4
SWAP
IR6
IR5/IR4
IR1
IR0
Bit
Number
Bit
Mnemonic
Function
7
CE
Cascade Enable:
0 = Disables the cascade signals CAS2:0 from appearing on the A18:16
address lines during interrupt acknowledge cycles.
1 = Enables the cascade signals CAS2:0, providing access to external
slave 82C59A devices. The cascade signals are used to address
specific slaves. If enabled, slave IDs appear on the A18:16 address
lines during interrupt acknowledge cycles, but are high during idle
cycles.
6
5
4
IR3
Internal Master IR3 Connection:
See Table 5-1 on page 5-8 for all the IR3 configuration options.
IR4
Internal Master IR4 Connection:
See Table 5-2 on page 5-8 for all the IR4 configuration options.
SWAP
INT6/DMAINT Connection:
0 = Connects DMAINT to the slave IR4. Connects INT6 to the slave IR5.
1 = Connects the INT6 pin to the slave IR4. Connects DMAINT to the slave
IR5.
3
2
IR6
Internal Slave IR6 Connection:
0 = Connects VSS to the slave IR6 signal.
1 = Connects the INT7 pin to the slave IR6 signal.
IR5/IR4
Internal Slave IR4 or IR5 Connection:
These depend on whether INTCFG.4 is set or clear.
0 = Connects VSS to the slave IR5 signal.
1 = Connects either the INT6 pin or DMAINT to the slave IR5 signal.
1
0
IR1
IR0
Internal Slave IR1 Connection:
0 = Connects the SSIO interrupt signal (SSIOINT) to the slave IR1 signal.
1 = Connects the INT5 pin to the slave IR1 signal.
Internal Slave IR0 Connection:
0 = Connects VSS to the slave IR0 signal.
1 = Connects the INT4 pin to the slave IR0 signal.
Figure 9-7. Interrupt Configuration Register (INTCFG)
9-19
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
9.3.3 Initialization Command Word 1 (ICW1)
Initialization begins with writing ICW1. Use ICW1 to select the interrupt request triggering type
(level or edge). The following actions occur within an 82C59A module when its ICW1 is written:
• The interrupt mask register is cleared, enabling all interrupt request signals.
• The IR7 signal is assigned the lowest interrupt level (default).
• Special mask mode is disabled.
Initialization Command Word 1
ICW1 (master and slave)
(write only)
master slave
Expanded Addr: F020H
F0A0H
00A0H
XXH
ISA Addr:
Reset State:
0020H
XXH
7
0
0
0
0
RSEL1
LS
0
0
1
Bit
Number
Bit
Mnemonic
Function
7–5
—
Clear these bits to guarantee device operation.
Register Select 1 (Also see OCW2 and OCW3):
4
RSEL1
ICW1, OCW2, and OCW3 are accessed through the same addresses.
0 = OCW2 or OCW3 is accessed (Figure 9-13 and Figure 9-15).
1 = ICW1 register is accessed.
3
LS
Level/Edge Sensitive:
0 = Selects edge-triggered IR input signals.
1 = Selects level-sensitive IR input signals.
All internal peripherals interface with the 82C59As in edge-triggered
mode only. This is compatible with the PC/AT bus specification. Each
source signal initiates an interrupt request by making a low-to-high
transition. External peripherals interface with the 8259As in edge-
triggered or level-sensitive mode. The modes are selected for the
device, not for individual interrupts.
NOTE: If an internal peripheral interrupt is used, the 8259A that the
interrupt is connected to must be programmed for edge-triggered
interrupts.
2–1
0
—
—
Clear these bits to guarantee device operation.
Set this bit to guarantee device operation.
NOTE: The 82C59A must be initialized before it can be used. After reset, the 82C59A register states are
undefined. The 82C59A modules must be initialized before the IF flag in the core FLAG register is
set. All peripherals that use interrupts connected to the ICU must be initialized before initializing
the ICU.
Figure 9-8. Initialization Command Word 1 Register (ICW1)
9-20
INTERRUPT CONTROL UNIT
9.3.4 Initialization Command Word 2 (ICW2)
Use the ICW2 register to define the base interrupt vector for the 82C59A. Valid vector numbers
for maskable interrupts range from 32 to 255. Because the base vector number must reside on an
8-byte boundary, the valid base vector numbers are 32 + n 8 where 0 ≤ n ≤ 27. Write the base
×
interrupt vector’s five most-significant bits to ICW2’s five most-significant bits. The 82C59A de-
termines specific IR signal vector numbers by adding the number of the IR signal to the base in-
terrupt vector.
Initialization Command Word 2
ICW2 (master and slave)
(write only)
master slave
Expanded Addr: F021H
F0A1H
00A1H
XXH
ISA Addr:
Reset State:
0021H
XXH
7
0
T7
T6
T5
T4
T3
0
0
0
Bit
Number
Bit
Mnemonic
Function
7–3
T7:3
Base Interrupt Type:
Write the base interrupt vector’s five most-significant bits to these bits.
Clear these bits to guarantee device operation.
2–0
T2:0
Figure 9-9. Initialization Command Word 2 Register (ICW2)
9-21
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
9.3.5 Initialization Command Word 3 (ICW3)
The ICW3 register contains information about the master/slave connections. For this reason, the
functions of the master’s ICW3 and the slave’s ICW3 differ.
ICW3 (at 0F021H or 0021H) is the master’s cascade configuration register (Figure 9-11). The
master has an internal slave cascaded from its IR2 signal. You can cascade additional slaves from
the master’s IR7, IR6, IR5, IR4, IR3 and IR1 signals. Setting a bit indicates that a slave 82C59A
is cascaded from the corresponding master’s IR signal.
NOTE
Since the internal slave is cascaded from the master’s IR2 signal, you must set
the S2 bit.
Initialization Command Word 3
ICW3 (master)
Expanded Addr: F021H
ISA Addr:
0021H
(write only)
Reset State:
XXH
7
0
S7
S6
S5
S4
S3
S2
S1
0
Bit
Number
Bit
Mnemonic
Function
7–3
S7:3
Slave IRs
0 = No slave 8259A is attached to the corresponding IR signal of the
master.
1 = A slave 82C59A is attached to the corresponding IR signal of the
master.
2
1
S2
S1
0 = Internal slave not used
1 = Internal slave is cascaded from the master’s IR2 signal.
Slave IRs
0 = No slave 8259A is attached to the master through the IR1 signal of
the master.
1 = A slave 82C59A is attached to the IR1 signal of the master.
0
—
Clear this bit to guarantee device operation.
Figure 9-10. Initialization Command Word 3 Register (ICW3 – Master)
9-22
INTERRUPT CONTROL UNIT
ICW3 (at 0F0A1H or 00A1H) is the internal slave ID register (Figure 9-11). Use this register to
indicate that the slave is cascaded from the master’s IR2 signal. This gives the internal slave an
ID of 2. Each slave device uses the IDs to determine whether it is the slave being addressed. Dur-
ing a slave access, the slave’s ID is driven on the master’s CAS2:0 signals. If these signals are
enabled (bit 7 of INTCFG is 1), they appear on the A18:16 address lines.
Initialization Command Word 3
ICW3 (slave)
(write only)
Expanded Addr: F0A1H
ISA Addr:
00A1H
XXH
Reset State:
7
0
0
0
0
0
0
0
1
0
Bit
Number
Bit
Mnemonic
Function
7–2
—
Clear these bits to guarantee device operation.
Set this bit to guarantee device operation.
Clear this bit to guarantee device operation.
1
0
—
—
Figure 9-11. Initialization Command Word 3 Register (ICW3 – Slave)
9-23
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
9.3.6 Initialization Command Word 4 (ICW4)
Use ICW4 to select the special-fully nested mode or the fully nested mode and to enable the au-
tomatic EOI mode.
Initialization Command Word 4
ICW4 (master and slave)
(write only)
master slave
Expanded Addr: F021H
F0A1H
00A1H
XXH
ISA Addr:
Reset State:
0021H
XXH
7
0
0
0
0
SFNM
0
0
AEOI
1
Bit
Number
Bit
Mnemonic
Function
7–5
—
Write zero to these bits to guarantee device operation.
4
SFNM
Special-fully Nested Mode:
0 = Selects fully nested mode.
1 = Selects special-fully nested mode. Only the master 82C59A can
operate in special-fully nested mode.
3–2
1
—
Write zero to these bits to guarantee device operation.
AEOI
Automatic EOI Mode:
0 = Disables automatic EOI mode.
1 = Enables automatic EOI mode. Only the master 82C59A can operate
in automatic EOI mode.
0
—
Write one to this bit to guarantee device operation.
Figure 9-12. Initialization Command Word 4 Register (ICW4)
9-24
INTERRUPT CONTROL UNIT
9.3.7 Operation Command Word 1 (OCW1)
OCW1 is the interrupt mask register. Setting a bit in the interrupt mask register disables (masks)
interrupts from the corresponding IR signal. For example, setting the master’s OCW1 M3 bit dis-
ables interrupts from the master IR3 signal. Clearing a bit in the interrupt mask register enables
interrupts from the corresponding IR signal.
Operation Command Word 1
OCW1 (master and slave)
(read/write)
master slave
Expanded Addr: F021H
F0A1H
00A1H
XXH
ISA Addr:
Reset State:
0021H
XXH
7
0
M7
M6
M5
M4
M3
M2
M1
M0
Bit
Number
Bit
Mnemonic
Function
7–0
M7:0
Mask IR:
0 = Enables interrupts on the corresponding IR signal.
1 = Disables interrupts on the corresponding IR signal.
NOTE: Setting the mask bit does not clear the respective interrupt
pending bit.
NOTE: The 8259A must be initialized before it can be used. After reset, the 8259A register states are
undefined. The 8259A modules must be initialized before the IF flag in the core FLAG register is
set.
Figure 9-13. Operation Command Word 1 (OCW1)
9-25
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
9.3.8 Operation Command Word 2 (OCW2)
Use OCW2 to change the priority structure and issue EOI commands.
Operation Command Word 2
OCW2 (master and slave)
master slave
Expanded Addr: F020H F0A0H
(write only)
ISA Addr:
Reset State:
0020H
XXH
00A0H
XXH
7
0
R
SL
EOI
RSEL1
RSEL0
L2
L1
L0
Bit
Number
Bit
Mnemonic
Function
7
R
The Rotate (R), Specific Level (SL), and End-of-Interrupt (EOI) Bits:
These bits change the priority structure and/or send an EOI command.
6
5
SL
R
SL EOI Command
EOI
0
0
0
0
1
1
1
1
*
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
Cancel automatic rotation*
Send a nonspecific EOI command
No operation
Send a specific EOI command**
Enable automatic rotation*
Enable automatic rotation and send a nonspecific EOI
Initiate specific rotation**
Initiate specific rotation and send a specific EOI**
These cases allow you to change the priority structure while the
82C59A is operating in the automatic EOI mode.
** The L2:0 bits (see below) specify the specific level for these cases.
Register Select Bits:
4–3
RSEL1:0
ICW1, OCW2 and OCW3 are accessed through the same addresses.
The states of RSEL1:0 determine which register is accessed. Write 00
to these bits to access OCW2.
RSEL1 RSEL0
0
0
1
0
1
X
OCW2
OCW3
ICW1
2–0
L2:0
IR Level:
When you program bits 7–5 to initiate specific rotation, these bits specify
the IR signal that is assigned the lowest level.
When you program bits 7–5 to send a specific EOI command, these bits
specify the IR signal that receives the EOI command.
If SL=0, then these bits have no effect.
Figure 9-14. Operation Command Word 2 (OCW2)
9-26
INTERRUPT CONTROL UNIT
9.3.9 Operation Command Word 3 (OCW3)
Use OCW3 to enable the special mask mode, issue a poll command, and provide access to the
interrupt in-service and request registers (ISR, IRR).
Operation Command Word 3
OCW3 (master and slave)
(write only)
master slave
Expanded Addr: F020H
F0A0H
00A0H
XXH
ISA Addr:
Reset State:
0020H
XXH
7
0
0
ESMM
SMM
RSEL1
RSEL0
POLL
ENRR
RDSEL
Bit
Number
Bit
Mnemonic
Function
Clear this bit to guarantee device operation.
7
—
6
5
ESMM
SMM
Enable Special Mask Mode (ESMM) and Special Mask Mode (SMM):
Use these bits to enable or disable special mask mode.
ESMM SMM
0
0
1
1
0
1
0
1
No action
No action
Disable special mask mode
Enable special mask mode
4–3
RSEL1:0
Register Select:
ICW1, OCW2 and OCW3 are accessed through the same addresses.
The states of RSEL1:0 determine which register is accessed. Write 01
to these bits to access OCW3.
RSEL1 RSEL0
0
0
1
0
1
X
OCW2
OCW3
ICW1
2
POLL
Poll Command:
Set this bit to issue a poll command, thus initiating the polling process.
1
0
ENRR
Enable Register Read Select (ENRR) and Read Register Select
(RDSEL):
RDSEL
These bits select which register is read during the next F020H and
F0A0H (or PC/AT address 0020H, 00A0H) read access.
ENRR
RDSEL Register Read on Next Read Pulse
0
0
1
1
0
1
0
1
No action
No action
Interrupt Request Register
In-service Register
Figure 9-15. Operation Command Word 3 (OCW3)
9-27
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
9.3.10 Interrupt Request Register (IRR)
This 8-bit, read-only register contains the levels requesting an interrupt to be acknowledged. It is
interrupt is acknowledged. Bits 7:0 of this register are the pending bits, respectively, of interrupt
requests IR7:0.
9.3.11 In-Service Register (ISR)
This 8-bit, read-only register contains the priority levels that are being serviced. It is accessed us-
ing OCW3 (see Figure 9-15). The ISR is updated when an End-of-Interrupt command is issued.
Bits 7:0 of this register are the in-service bits, respectively, of interrupt requests IR7:0.
9.3.12 Poll Status Byte (POLL)
Read the poll status byte after issuing a poll command to determine whether any of the devices
connected to the 82C59A require servicing. Once the polling bit is set in OCW3, the Poll Status
Byte of a particular 82C59A can be read by doing an access to any of the four addresses of that
82C59A.
Poll Status Byte
POLL (master and slave)
(read only)
master slave
Expanded Addr: F020H
F0A0H
00A0H
XXH
ISA Addr:
Reset State:
0020H
XXH
7
0
INT
—
—
—
—
L2
L1
L0
Bit
Number
Bit
Mnemonic
Function
7
INT
Interrupt Pending:
0 = No request pending.
1 = Indicates that a device attached to the 82C59A requires servicing.
6–3
2–0
—
Reserved. These bits are undefined.
L2:0
Interrupt Request Level:
When bit 7 is set, these bits indicate the highest-priority IR signal that
requires servicing. When bit 7 is clear, i.e., no request is pending, these
bits are indeterminate.
Figure 9-16. Poll Status Byte (POLL)
9-28
INTERRUPT CONTROL UNIT
9.4 DESIGN CONSIDERATIONS
The following sections discuss some design considerations.
9.4.1 Interrupt Acknowledge Cycle
When the core receives an interrupt request from the master, it completes the instruction in
progress and any succeeding locked instructions, then initiates an interrupt acknowledge cycle.
The interrupt acknowledge cycle generates an internal interrupt acknowledge (INTA#) signal that
consists of two locked pulses (Figure 9-17). This INTA# signal is connected to the internal
82C59A interrupt acknowledge inputs. On the falling edge of the second INTA#, the 82C59A sets
its interrupt in-service bit. It then clears its interrupt pending bit on the rising edge of the second
INTA#. On the second INTA# falling edge, the addressed 82C59A (determined by the master’s
cascade signals) also drives the interrupt vector number on the data bus.
INTA#
Data Bus
valid
Vector Number
A2430-01
Figure 9-17. Interrupt Acknowledge Cycle
9.4.2 Interrupt Detection
The processing of an interrupt begins with the assertion of an interrupt request at one of the IR
signals. During system initialization, you can program the IR signals, as a group, to be either edge
or level triggered (using ICW1 described in Figure 9-8).
Edge triggered
The 82C59A recognizes a rising edge transition on an IR signal as an
interrupt request. A device requesting service must maintain a high state
on an IR signal until after the falling edge of the first INTA# pulse. You
can reset the edge-detection circuit during initialization of the 82C59A
or by deasserting the IR signal. To reset the edge-detection circuit
properly, the interrupt source must hold the IR line low for a minimum
time of 10ns.
9-29
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Level triggered
The 82C59A recognizes a high level on an IR line as an interrupt
request. A device must maintain the high level until after the falling edge
of the first INTA# pulse. Unlike an edge-triggered IR signal, a level-
triggered IR signal continues to generate interrupts as long as it is
asserted. To avoid continuous interrupts from the same source, a device
must deassert a level-sensitive IR signal before the interrupt handler
issues an end-of-interrupt (EOI) command.
compatible with the PC/AT bus specification. Each source signal initiates an interrupt by making
a low-to-high transition.
9.4.3 Spurious Interrupts
For both edge and level-triggered interrupts, a high level must be maintained on the IR line until
after the falling edge of the first INTA# pulse (see Figure 9-18). A spurious interrupt request is
generated if this stipulation is not met. A spurious interrupt on any IR line generates the same vec-
tor number as an IR7 request. The spurious interrupt, however, does not set the in-service bit for
IR7. Therefore, an IR7 interrupt service routine must check the in-service register to determine
whether the interrupt source was a valid IR7 (the in-service bit is set) or a spurious interrupt (the
in-service bit is cleared).
INTA#
IR (Spurious)
IR (Valid)
IR sampled on this edge.
A2431-01
Figure 9-18. Spurious Interrupts
9.4.4 Cascading Interrupt Controllers
Figure 9-19 is a block diagram showing the connections for two cascaded 82C59As. The PLD
generates READY# (for the second Interrupt Acknowledge Cycle) and INTA# to the external
82C59A devices. The PLD also generates appropriate timings for the INTA# signals to satisfy
82C59A specifications.
The RD# and WR# strobes are used to read and write to the 82C59A registers. These strobes are
inactive during Interrupt Acknowledge Cycles.
9-30
INTERRUPT CONTROL UNIT
Intel386™ EX
Processor
PLD
READY#
INTA#
READY#
M/IO#
W/R#
INTA#
and
READY#
State
Machine
D/C#
ADS#
LBA#
CLKOUT
CLK2
CAS0
CAS1
CAS2
External
CAS
Decode
External
82C59As
CAS0
Latch
CAS1
CAS2
A0
INTA#
CS #
x
INT
INT
x
INT
RD#
WR#
y
D7:0
CAS0
CAS1
CAS2
A0
BLE#
CS #
x
INT
INTA#
CS #
y
CS #
y
RD#
RD#
WR#
WR#
D7:0
D7:0
A2857-01
Figure 9-19. Cascading External 82C59A Interrupt Controllers
9-31
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
9.5 PROGRAMMING CONSIDERATIONS
Consider the following when programming the ICU.
• When an 82C59A receives an interrupt request, it sets the request’s pending bit (regardless
of whether the IR signal is masked or not). The pending bit remains set until the interrupt is
serviced.
• When the LS bit in ICW1 is set to edge-triggered during initialization, all the interrupt
pending bits will be cleared.
• In special-fully nested mode, care must be taken when processing interrupt requests from
the master’s internal cascade signal (IR2). At the end of the slave’s interrupt service routine,
first issue a nonspecific EOI to the slave. Before issuing a nonspecific EOI command to the
master, make sure that the slave has no other in-service bits set.
• Systems that use polling as the only method of device servicing must still fully initialize the
82C59A modules. Also, the interrupt requests to the core must be disabled using the mask
bits or the CLI instruction.
• The 8259A must be initialized before it can be used. After reset, the 82C59A register states
are undefined. The 82C59A modules must be initialized before the IF flag in the core FLAG
register is set. All peripherals that use interrupts connected to the ICU must be initialized
before initializing the ICU.
9.5.1 Interrupt Control Unit Code Examples
The example code contains these software routines:
InitICU
Initializes the Master and Slave 82C59A Interrupt Controllers
Initializes the Slave 82C59A Interrupt Controllers
Disables interrupts on the Master and internal Slave
Enables interrupts on the Master and internal Slave
InitICUSlave
Disable8259Interrupt
Enable8259Interrupt
SetIRQVector
Loads the interrupt vector table with the address of the Interrupt
Service Routine
SetInterruptVector
Poll_Command
Called by SetIRQVector to load vector table
Issues a poll command to read the poll status byte of the ICU
See Appendix C for included header files.
#include <conio.h>
#include “80386ex.h”
#include “EV386EX.h”
/* Globals For information about the ICU */
BYTE
BYTE
_IRQ_SlaveBase_= 0x30;
_IRQ_MstrBase_ = 0x20;
9-32
INTERRUPT CONTROL UNIT
BYTE
_CascadeBits_ = 0x4;
/*****************************************************************************
InitICU
Description:
Initialization for both the master and slave Interrupt Control
Units (ICU). tine only initializes the internal interrupt
controllers, external ICUs must be initialized separately. These
should be initialized before interrupts are enabled(i.e., enable()).
Parameters:
MstrMode
MstrBase
Mode of operation for Master ICU
Specifies the base interrupt vector number for the
Master interrupts.
For example, if IR1 of the master goes active and the
MstrBase = 0x20, the processor uses interrupt
vector table entry 0x21.
MstrCascade
SlaveMode
SlaveBase
Which Master IRQs are used for Slave ICUs.
Mode of operation for Slave ICU
Specifies the base interrupt vector number for the
Slave interrupts.
For example, if IR1 of the slave goes active and the
SlaveBase = 0x40, the processor uses interrupt
vector table entry 0x41.
MstrPins
Defines what EX pins are available externally to the
chip for the Master.
SlavePins
Defines what EX pins are available externally to the
chip for the Slave.
Returns:Error Code
E_OK
-- Initialized OK, No error.
Assumptions:
REMAPCFG register has Expanded I/O space access enabled (ESE bit set).
Syntax:
#define ICU_TRIGGER_EDGE
0x0
0x4
0x2
0x1
#define MPIN_INT0
#define MCAS_IR1
#define SPIN_INT4
int error_code;
error_code = InitICU(ICU_TRIGGER_EDGE,
0x20,
MCAS_IR1,
ICU_TRIGGER_EDGE,
0x30,
MPIN_INT0,
SPIN_INT4);
9-33
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Real/Protected Mode
No changes required.
*****************************************************************************/
int InitICU(BYTE MstrMode, BYTE MstrBase, BYTE MstrCascade, BYTE SlaveMode,
BYTE SlaveBase, BYTE MstrPins, BYTE SlavePins)
{
BYTE icw, cfg_pins;
/* Program Slave ICU */
_IRQ_SlaveBase_ = SlaveBase & 0xf8;
_SetEXRegByte(ICW1S, 0x11 | SlaveMode);// Set slave triggering
_SetEXRegByte(ICW2S, _IRQ_SlaveBase_); // Set slave base interrupt type,
// least 3-bit must be 0
_SetEXRegByte(ICW3S, 0x2);
_SetEXRegByte(ICW4S, 0x1);
// Set slave ID
// Set bit 0 to guarantee operation
/* Program Master ICU */
_IRQ_MstrBase_ = MstrBase & 0xf8;
_CascadeBits_ = MstrCascade | 0x4;
icw = (MstrMode & ICU_TRIGGER_LEVEL) ? 0x19 : 0x11;
_SetEXRegByte(ICW1M, icw);
// Set master triggering
_SetEXRegByte(ICW2M, _IRQ_MstrBase_); // Set master base interrupt
// type, least 3-bit must be 0
_SetEXRegByte(ICW3M, _CascadeBits_); // Set master cascade pins,
// Make sure IR2 set for Cascade
icw = (MstrMode & ~ICU_TRIGGER_LEVEL) | 1;// Set bit 0 and remove
// Trigger_level bit (in ICW1)
_SetEXRegByte(ICW4M, icw);
// Set slave IDs in master
/* Program chip configuration registers */
cfg_pins = _GetEXRegByte(INTCFG);
if( (MstrCascade & 0xfb) != 0 )
// bit 2 (IR2) is internal,
// external signals not required
// for just IR2
cfg_pins |= 0x80;
// Using external slaves,
// therefore enable Cascade signals
cfg_pins |= SlavePins;
_SetEXRegByte(INTCFG, SlavePins);
cfg_pins = _GetEXRegByte(P3CFG);
// Set Slave external interrupt pins
// Preserve other set bits
_SetEXRegByte(P3CFG, cfg_pins | MstrPins);// Set Master external
// interrupt pins
return E_OK;
}/* InitICU */
9-34
INTERRUPT CONTROL UNIT
/*****************************************************************************
InitICUSlave
Description:
Initialization only the internal slave Interrupt Control Units (ICU).
This routine only initializes the internal interrupt controller,
external ICUs must be initialized separately.
Parameters:
SlaveMode
SlaveBase
Mode of operation for Slave ICU
Specifies the base interrupt vector number for the
Slave interrupts.
For example, if IR1 of the slave goes active and the
SlaveBase = 0x40 the processor uses interrupt
vector table entry 0x41.
SlavePins
Defines what EX pins are available externally to the
chip for the Slave.
Returns:Error Code
E_OK
-- Initialized OK, No error.
Assumptions:
REMAPCFG register has Expanded I/O space access enabled (ESE bit set).
Syntax:
/* ICU Modes */
#define ICU_SFNM
#define ICU_AUTOEOI
0x10
0x2
#define ICU_TRIGGER_LEVEL 0x8
#define ICU_TRIGGER_EDGE
/* ICU Slave Pins */
#define SPIN_INT4
#define SPIN_INT5
#define SPIN_INT6
#define SPIN_INT7
0x0
0x1
0x2
0x4
0x8
int error_code;
error_code = InitICUSlave(ICU_TRIGGER_EDGE, 0x30, SPIN_INT4);
Real/Protected Mode
No changes required.
*****************************************************************************/
int InitICUSlave(BYTE SlaveMode, BYTE SlaveBase, BYTE SlavePins)
{
BYTE cfg_pins;
/* Program Slave ICU */
9-35
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
_IRQ_SlaveBase_ = SlaveBase & 0xf8;
_SetEXRegByte(ICW1S, 0x11 | SlaveMode); // Set slave triggering
_SetEXRegByte(ICW2S, _IRQ_SlaveBase_); // Set slave base interrupt
// type, least 3-bit must be 0
_SetEXRegByte(ICW3S, 0x2);
_SetEXRegByte(ICW4S, 0x1);
// Set slave ID
// Set bit 0 to guarantee
// operation
cfg_pins = _GetEXRegByte(INTCFG);
cfg_pins |= SlavePins;
_SetEXRegByte(INTCFG, SlavePins);
// Set Slave external interrupt
// pins
return E_OK;
}/* InitICUSlave */
/*****************************************************************************
Disable8259Interrupt:
Description:
Disables 8259a interrupts for the master and the slave.
Parameters:
MstrMask
SlaveMask
Mask value for master ICU
Mask value for slave ICU
Each bit location that is set disables the corresponding
interrupt (by setting the bit in the interrupt control register).
For example, to disable master IR3 and IR5 set MstrMask = 0x28
(bits 3 and 5 are set).
Returns:
None
Assumptions:
REMAPCFG register has Expanded I/O space access enabled (ESE bit set).
Syntax:
/* ICU IRQ Mask Values*/
#define IR0
#define IR1
#define IR2
#define IR3
#define IR4
#define IR5
#define IR6
0x1
0x2
0x4
0x8
0x10
0x20
0x40
9-36
INTERRUPT CONTROL UNIT
#define IR7
0x80
Disable8259Interrupt(IR0 | IR1 | IR3 | IR4 | IR5 | IR6 | IR7,
IR1 | IR2 | IR3 | IR4 |IR5 | IR6);
Real/Protected Mode
No changes required.
*****************************************************************************/
void Disable8259Interrupt(BYTE MstrMask, BYTE SlaveMask)
{
BYTE Mask;
if(MstrMask != 0)
{
Mask = _GetEXRegByte(OCW1M);
_SetEXRegByte(OCW1M, Mask | MstrMask);
}
if(SlaveMask != 0)
{
Mask = _GetEXRegByte(OCW1S);
_SetEXRegByte(OCW1S, Mask | SlaveMask);
}
}/* Disable8259Interrupt */
/*****************************************************************************
Enable8259Interrupt:
Description:
Enables 8259a interrupts for the master and the slave.
Parameters:
MstrMask
SlaveMask
Enable mask value for master ICU
Enable mask value for slave ICU
Each bit location that is set enables the corresponding
interrupt (by clearing the bit in the interrupt control register).
For example, to enable master IR3 and IR5 set MstrMask = 0x28
(bits 3 and 5 are set).
Returns:
None
Assumptions:
REMAPCFG register has Expanded I/O space access enabled (ESE bit set).
Syntax:
9-37
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
/* ICU IRQ Mask Values*/
#define IR0
#define IR1
#define IR2
#define IR3
#define IR4
#define IR5
#define IR6
#define IR7
0x1
0x2
0x4
0x8
0x10
0x20
0x40
0x80
Enable8259Interrupts(IR2, IR0 | IR7); //Enable MasterIR2 for cascading
//Enable INT4 and WDTOUT on Slave
Real/Protected Mode
No changes required.
*****************************************************************************/
void Enable8259Interrupt(BYTE MstrMask, BYTE SlaveMask)
{
BYTE Mask;
if(MstrMask != 0)
{
Mask = _GetEXRegByte(OCW1M);
_SetEXRegByte(OCW1M, Mask & (~MstrMask));
}
if(SlaveMask != 0)
{
Mask = _GetEXRegByte(OCW1S);
_SetEXRegByte(OCW1S, Mask & (~SlaveMask));
}
}/* Enable8259Interrupt */
/****************************************************************************
SetIRQVector:
Description:
Loads the interrupt vector table with the address of the interrupt
routine. The vector table entry number is determined by the vector
number.
Parameters:
InterProc
Address of interrupt function, will be loaded into
the interrupt table.
IRQ
ISR_Type
Hardware Interrupt request number (0-15).
Specifies if the interrupt function should be treated
as a TRAP_ISR or an INTERRUPT_ISR. Real Mode only
9-38
INTERRUPT CONTROL UNIT
supports INTERRUPT_ISR (parameter is ignored).
Protected mode supports both.
Returns:Error Code
E_INVALID_VECTOR
E_BADVECTOR
-- An IRQ of greater than 15 was passed
-- IRQ is used for cascading to a slave interrupt
controller
E_OK
-- Initialized OK, No error.
Assumptions:
Compiler supports far and interrupt keywords
ICU must be configured before this function is call for it to
operate properly
_IRQ_SlaveBase_,_IRQ_MstrBase_,_CascadeBits_ are set before function
is called. These are initialized in the InitICU functions supplied
in this source.
Syntax:
int error_code;
error_code = SetIRQVector(wdtISR,
15,
// Slave IR#’s are offset by 8 in
// Vector Table
INTERRUPT_ISR);
Real/Protected Mode
No changes required. Uses SetInterruptVector which is mode
dependant (separate source)
*****************************************************************************/
int SetIRQVector( void (far interrupt *IntrProc)(void), int IRQ, int IntrType)
{
int Vector;
if(IRQ > 15) return E_INVALID_VECTOR;
if(IRQ > 7)
// Get Vector from Slave
Vector = _IRQ_SlaveBase_ + IRQ - 8;
else
{
// From Master
if((1 << IRQ) & _CascadeBits_) return E_BADVECTOR;
Vector = _IRQ_MstrBase_ + IRQ;
}
SetInterruptVector(IntrProc, Vector, IntrType);
9-39
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
return E_OK;
}/* SetIRQVector */
/*****************************************************************************
SetInterruptVector:
Description:
Loads the interrupt vector table with the address of the interrupt
routine. The vector table entry number is determined by the vector
number.
Parameters:
InterProc
Address of interrupt function, will be loaded into
the interrupt table.
ISR_Type
Specifies if the interrupt function. Real Mode only
supports INTERRUPT_ISR (the parameter is ignored). The
parameter is kept to maintain compatibility with the
protected mode version of this function.
Returns:
None
Assumptions:
Compiler supports far and interrupt keywords
Compiler may issue a warning about IntrType not used.
IntrType is kept for protected mode compatibility.
Syntax:
SetInterruptVector(wdtISR, INTERRUPT_ISR);
Real/Protected Mode
Real Mode only
*****************************************************************************/
void SetInterruptVector( void (far interrupt *IntrProc)(void),
int Vector, int IntrType)
{
(void) IntrType;
// Reference to avoid compiler warning
((unsigned long far *)(0))[Vector] = (unsigned long)IntrProc;
}/* SetInterruptVector */
/*****************************************************************************
9-40
INTERRUPT CONTROL UNIT
Poll_Command:
Description:
This routine issues a poll command which reads the poll status byte
of the ICU.
Parameters:
Master_or_Slave Specifies which interrupt controller is polled
Returns:
Current value of poll status byte
Assumptions:
None
Syntax:
in poll_status;
poll_status = Poll_Command();
Real/Protected Mode:
No changes required.
******************************************************************************/
int Poll_Command( int Master_or_Slave)
{
int poll_status;
if (Master_or_Slave == Master) {
_SetEXRegByte(OCW3M, 0x0c);//Initiate polling sequence
poll_status = _GetEXRegByte(ICW2M);
}
else {
_SetEXRegByte(OCW3S, 0x0c); //Initiate polling sequence
poll_status = _GetEXRegByte(ICW2S);
}
return(poll_status);
} /* Poll_Command */
9-41
10
TIMER/COUNTER
UNIT
CHAPTER 10
TIMER/COUNTER UNIT
The Timer/counter Unit (TCU) has the same basic functionality as the industry-standard 82C54
counter/timer. It contains three independent 16-bit down counters, which can be driven by a pres-
caled value of the processor clock or an external clock. The counters contain two count formats
put signals can appear at device pins, generate interrupt requests, and initiate DMA
transactions.
This chapter is organized as follows:
• Overview (see below)
• TCU Operation (page 10-5)
• Register Definitions (page 10-20)
• Programming Considerations (page 10-33)
10.1 OVERVIEW
The TCU contains control logic and three independent 16-bit down counters (Figure 10-1). Each
counter has two input signals and one output signal:
CLKINn
You can independently connect each counter’s clock input (CLKINn) signal to
either the internal prescaled clock (PSCLK) signal or the external timer clock
(TMRCLKn) pin. This allows you to use either a prescaled value of the
processor’s internal clock or an external clock to drive each counter.
NOTE
The maximum CLKINn frequency, whether connected internally or externally,
is 8 MHz.
GATEn
Each counter has a gate (GATEn) input signal. This signal provides counter
operation control. In some of the counter operating modes, a high level on a
counter’s GATEn signal enables or resumes counting and a low level disables or
suspends counting. In other modes, a rising edge on GATEn loads a new count
value. You can independently connect each counter’s GATEn signal to either VCC
or the external timer gate (TMRGATEn) pin, or you can drive each counter’s
GATEn signal high or low through register bits.
OUTn
Each counter contains an output signal called OUTn. You can independently
connect these signals to the external timer clock output (TMROUTn) pins.
OUT0, OUT1, and OUT2 are routed to the interrupt control unit. OUT1 is also
routed to DMA channel 0, and OUT2 is also routed to DMA channel 1.
10-1
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
Therefore, the OUTn signals can drive external devices, generate interrupt
requests, initiate DMA transactions or combinations of the three.
Each counter operates independently. Six different counting modes are available and two count
formats: binary (16 bits) or BCD (4 decades). Each operating mode allows you to program the
counter with an initial count and to change this value “on the fly.” You can determine the count
and status of each counter without disturbing its current operation.
S
Control Logic
y
s
t
TMRCFG.0
TMRCFG.7
e
m
B
u
s
0
1
PSCLK
CLKIN0
TMRCLK0
P3CFG.0
P3CFG.1
PINCFG.5
TMRCFG.6
TMRCFG.1
OUT0
OUT1
OUT2
TMROUT0
V
0
1
0
CC
To ICU
(Master IR0)
GATE0
CLKIN1
1
TMRGATE0
TMRCFG.2
PSCLK
0
1
TMRCLK1
TMRCFG.3
TMRCFG.6
TMROUT1
V
0
1
CC
0
1
To ICU
(Slave IR2)
GATE1
CLKIN2
TMRGATE1
To DMA
Ch0 MUX
PINCFG.5
TMRCFG.4
PSCLK
0
1
TMRCLK2
TMRCFG.5
TMRCFG.6
PINCFG.5
TMROUT2
V
0
1
0
1
CC
To ICU
(Slave IR3)
GATE2
TMRGATE2
To DMA
Ch1 MUX
A2317-02
Figure 10-1. Timer/Counter Unit Signal Connections
10-2
TIMER/COUNTER UNIT
10.1.1 TCU Signals and Registers
Table 10-1 and Table 10-2 lists the signals and registers associated with the TCU.
Table 10-1. TCU Signals
Device Pin or
Signal
PSCLK
Description
Internal Signal
Internal signal
Prescaled Clock:
This is one of the two possible connections for the counter’s CLKINn
signal. PSCLK is an internal signal that is a prescaled value of the
processor internal clock. The clock and power management unit
contains a programmable divider that determines the PSCLK frequency.
See “Controlling the PSCLK Frequency” on page 8-7, for information on
how to program PSCLK’s frequency.
TMRCLK0
TMRCLK1
TMRCLK2
Device pin
Timer Clock Input:
This is one of the two possible connections for the counter’s CLKINn
signal. You can drive a counter with an external clock source by
connecting the clock source to the counter’s TMRCLKn pin.
TMRGATE0 Device pin
TMRGATE1
TMRGATE2
Timer Gate Input:
This input can be connected to the counter’s GATEn input to control the
counter’s operation. In some of the counter’s operating modes, a high
level on GATEn enables or resumes counting, while a low level disables
or suspends counting. In other modes, a rising edge on GATEn loads a
new count value.
TMROUT0
TMROUT1
TMROUT2
Device pin
Timer Output:
The counter’s OUTn signal can be connected to this pin. The operation,
and consequently the waveform, of the output depends on the counter’s
operating mode.
10-3
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
Table 10-2. TCU Associated Registers
Expanded
Address
PC/AT*
Address
Register
P3CFG
Function
0F824H
0F826H
—
Peripheral Pin Selections:
PINCFG
These registers determine whether a counter’s input and output
signals are connected to package pins.
(read/write)
TMRCFG
0F834H
0F043H
—
Timer Configuration:
(read/write)
Enables the counter’s CLKINn input signal, selects the CLKINn
connection (PSCLK or TMRCLKn) for each counter, and either
connects TMRGATEn or VCC to each counter’s GATEn input signal,
or sets GATEn high or low through register bits.
TMRCON
0043H
TMRCON has three formats: control word, counter-latch, and read-
back. When writing to TMRCON, certain bit settings determine which
format is accessed.
Control Word Format:
Programs a specific counter. Selects a counter’s operating mode and
count format. After programming a counter, you can write a count
value to the counter’s TMRn register at any time.
Counter-latch Format:
Issues a counter-latch command to a specific counter. The counter-
latch command allows you to latch the count of a specified counter.
After issuing a counter-latch command, you can check the counter’s
count by reading the counter’s TMRn register.
Read-back Format:
Issues a read-back command to one or more counters. The read-
back command allows you to latch the count and status of one or
more counters. After issuing the read-back command, you can check
the counter’s status by reading the counter’s TMRn register. After
checking a counter’s status, you can read the counter’s TMRn
register again to check its count.
TMR0
TMR1
TMR2
0F040H
0F041H
0F042H
0040H
0041H
0042H
Status Format:
Read this register after issuing a read-back command to check
counter n’s status. Reading TMRn again accesses its read format.
Read Format:
Read this register to check counter n’s count value.
Write Format:
Write this register at any time after initializing counter n to change the
counter’s count value.
10-4
TIMER/COUNTER UNIT
10.2 TCU OPERATION
Each counter can operate in any one of six operating modes. These modes are described in sec-
tions 10.2.1 through 10.2.6. In all modes, the counters decrement on the falling edge of CLKINn.
In modes 0, 1, 4, and 5, the counters roll over to the highest count, either 0FFFFH for binary
counting or 9999 for BCD counting, and continue counting down. However, the state of the
OUTn is only affected by the first run through the counter and does not change on subsequent
runs. Modes 2 and 3 are periodic modes; in these modes, when the counter reaches terminal count
it is reloaded with the currently programmed count value.
To specify a counter’s operating mode, write to the TMRCON register’s control word format.
Writing to this register initiates counting. To specify a count, write to the counter’s TMRn regis-
ter’s write format. In modes 0 and 4, the count is loaded on the falling edge of CLKINn. Modes
1 and 5 require a rising edge on a counter’s GATEn signal (or gate-trigger) to load the count. In
modes 2 and 3, the count is loaded when the counter reaches terminal count or when the counter
receives a gate-trigger, whichever is first.
The GATEn signal affects the counting operation for each mode differently (Table 10-3). For
modes 0, 2, 3, and 4, GATEn is level sensitive, and the logic level is sampled on the rising edge
of CLKINn. The action then occurs on the falling edge of the next CLKINn. In these modes,
GATEn must be high for counting to begin. During a counting sequence, a low level at GATEn
suspends counting, while a high level at GATEn resumes counting.
For modes 1, 2, 3, and 5, GATEn is rising-edge sensitive. In these modes, a rising edge at GATEn
sets an edge-sensitive flip-flop in the counter. This flip-flop is then sampled on the next rising
edge of CLKINn; the flip-flop is reset immediately after it is sampled. In this way, a trigger is
detected no matter when it occurs - a high level does not have to be maintained until the next ris-
ing edge of CLKINn. Therefore, a rising edge on GATEn that occurs between two rising CLKINn
edges is recognized as a gate-trigger. The operation caused by a gate-trigger occurs on the falling
CLKINn edge following the trigger. Note that in modes 2 and 3, the GATEn input is both edge-
and level-sensitive. In modes 1, 2, 3, and 5, a gate-trigger causes the counter to load new count
values.
10-5
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
Table 10-3. Operations Caused by GATEn
Operating
Low or Falling
Rising
High
Modes
0
1
Disables counting
—
Enables counting
—
—
1) Initiates counting
2) Resets OUTn after next CLKINn
2
3
1) Disables counting
Initiates counting
Initiates counting
Enables counting
Enables counting
2) Sets OUTn immediately high
1) Disables counting
2) Sets OUTn immediately high
4
5
Disables counting
—
—
Enables counting
—
Initiates counting
10.2.1 Mode 0 – Interrupt on Terminal Count
This mode allows you to generate a rising edge on a counter’s OUTn signal. Initializing a counter
es terminal count, OUTn is driven high. At this point, the counter rolls over and continues count-
ing with OUTn high. OUTn stays high and the counter keeps counting down and rolling over until
a new count is written or you reprogram the counter. You can write a new count to the counter at
any time to drive OUTn low and start a new counting sequence. Writing a new control word re-
programs the counter.
Mode 0’s basic operation is outlined below and shown in Figure 10-2.
1. After a control word write, OUTn is driven low.
2. On the CLKINn pulse following a count write, the count is loaded.
3. On each succeeding CLKINn pulse, the count is decremented.
4. When the count reaches terminal count, OUTn is driven high.
NOTE
Writing a count of N causes a rising edge on OUTn in N + 1 CLKINn pulses
(provided GATEn remains high and count was written before the rising edge of
CLKINn).
10-6
TIMER/COUNTER UNIT
Control
Word = 10H
Count = 4
Writes to
Counter n
CLKINn
GATEn
OUTn
?
?
?
?
0004 0003 0002 0001 0000 FFFF FFFE
Count
A2311-01
Figure 10-2. Mode 0 – Basic Operation
Figure 10-3 shows suspending the counting sequence. A low level on GATEn causes the counter
to suspend counting (both the state of OUTn and the count remain unchanged). A high level on
GATEn resumes counting.
Control
Word = 10H
Count = 3
Writes to
Counter n
CLKINn
GATEn
OUTn
?
?
?
?
0003 0002 0002 0002 0001 0000 FFFF
Count
A2394-02
Figure 10-3. Mode 0 – Disabling the Count
10-7
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
Figure 10-4 shows writing a new count before the current count reaches zero. The counter loads
the new count on the CLKINn pulse after you write it, then decrements this new count on each
succeeding CLKINn pulse. OUTn remains low until the new count reaches zero.
Control
Word = 10H
Count = 3
Count = 2
Writes to
Counter n
CLKINn
GATEn
OUTn
?
?
?
?
0003 0002 0001 0002 0001 0000 FFFF
Count
A2395-02
Figure 10-4. Mode 0 – Writing a New Count
10.2.2 Mode 1 – Hardware Retriggerable One-shot
This mode is similar to mode 0; it allows you to generate a rising edge on a counter’s OUTn sig-
driving its OUTn signal low. When the counter reaches zero, OUTn is driven high. At this point,
the counter rolls over and continues counting with OUTn high. OUTn stays high and keeps count-
ing down and rolling over until the counter receives another gate-trigger or you reprogram it. You
can retrigger the one-shot at any time with a gate-trigger, causing the counter to reload the count
and drive OUTn low. Writing a new control word to the counter reprograms it.
Mode 1’s basic operation is outlined below and shown in Figure 10-5.
1. After a control word write, OUTn is driven high.
2. On the CLKINn pulse following a gate-trigger, the count is loaded and OUTn is driven
low.
3. On each succeeding CLKINn pulse, the count is decremented.
4. When the count reaches zero, OUTn is driven high.
NOTE
Writing a count of N causes a rising edge on OUTn in N CLKINn pulses after
the count is loaded (using a gate-trigger).
10-8
TIMER/COUNTER UNIT
Control
Word = 12H
Count = 3
Writes to
Counter n
CLKINn
GATEn
OUTn
Count
?
?
?
?
?
0003 0002 0001 0000 FFFF 0003 0002
A2312-02
Figure 10-5. Mode 1 – Basic Operation
Figure 10-6 shows retriggering the one-shot. On the CLKINn pulse following the retrigger, the
counter reloads the count. The control logic then decrements the count on each succeeding
CLKINn pulse; OUTn remains low until the count reaches zero.
Control
Count = 3
Word = 12H
Writes to
Counter n
CLKINn
GATEn
OUTn
?
?
?
?
?
0003 0002 0001 0003 0002 0001 0000
Count
A2396-01
Figure 10-6. Mode 1 – Retriggering the One-shot
10-9
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
Figure 10-7 shows writing a new count. The counter waits for a gate-trigger to load the new count.
The counter loads the new count on the CLKINn pulse following the trigger, then decrements the
count on each succeeding CLKINn pulse. OUTn remains low until the count reaches zero.
Control
Word = 12H
Count =2
Count = 4
Writes to
Counter n
CLKINn
GATEn
OUTn
Count
?
?
?
?
?
0002 0001 0000 FFFF FFFE 0004 0003
A2397-02
Figure 10-7. Mode 1 – Writing a New Count
10.2.3 Mode 2 – Rate Generator
In this periodic mode, a counter’s OUTn signal remains high until the count reaches one, then
goes low for one clock pulse (CLKINn). After this single clock pulse, OUTn goes high and the
count is reloaded. The cycle then repeats. You can use a gate-trigger to reload the count at any
time. This provides a way to synchronize the counting cycle. A high level on a counter’s GATEn
signal enables counting; a low level on a counter’s GATEn signal disables counting.
Mode 2’s basic operation is outlined below and shown in Figure 10-8.
1. After a control word write, OUTn is driven high.
2. The count is loaded on the CLKINn pulse following one of these events:
•
•
•
A write to a control word followed by a write to count
A gate trigger
The counter reaches one
3. On each succeeding CLKINn pulse, the count is decremented.
4. When the count reaches one, OUTn is driven low.
5. On the following CLKINn pulse, OUTn is driven high and the count is reloaded.
6. The process is repeated from step 3.
10-10
TIMER/COUNTER UNIT
Control
Word= 14H
Count = 3
Writes to
Counter n
CLKINn
GATEn
OUTn
?
?
?
?
0003 0002 0001 0003 0002 0001 0003
Count
A2313-01
Figure 10-8. Mode 2 – Basic Operation
Figure 10-9 shows suspending the counting sequence. A low level on GATEn causes the counter
to suspend counting. The count remains unchanged and OUTn is immediately driven (or stays)
high (If the GATEn goes low when OUTn is low, then OUTn is immediately driven high). A ris-
ing edge on the GATEn causes the counter to be reloaded with the count. A high level on GATEn
resumes counting.
Control
Word = 14H
Count = 3
Writes to
Counter n
CLKINn
GATEn
OUTn
?
?
?
?
0003 0002 0002
0003 0002 0001 0003
Count
A2398-01
Figure 10-9. Mode 2 – Disabling the Count
10-11
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
Figure 10-10 shows writing a new count. The counter loads the new count after the counter reach-
es one. When the counter receives a gate-trigger after a new count was written to it, the counter
loads the new count on the next CLKINn pulse. This allows GATEn to synchronize the counters.
Control
Word = 14H
Count = 4
Count = 5
Writes to
Counter n
CLKINn
GATEn
OUTn
?
?
?
?
0004 0003 0002 0001 0005 0004 0003
Ccount
A2399-01
Figure 10-10. Mode 2 – Writing a New Count
10.2.4 Mode 3 – Square Wave
In this periodic mode, a counter’s OUTn signal remains high for half a specified count, then goes
low for the remainder of the count. A count of N results in a square wave with a period of N
CLKINn pulses. A high level on a counter’s GATEn signal enables counting; a low level on a
counter’s GATEn signal disables counting. The output produced by a counter’s OUTn signal de-
pends on whether a count is odd or even. Mode 3’s basic operation for even and odd counts is
outlined below and shown in Figure 10-11 and Figure 10-12.
Even count basic operation:
1. After a control word write, OUTn is driven high.
2. The count is loaded on the CLKINn pulse following one of these events:
•
•
•
A write to a control word followed by a write to count
A gate trigger
The counter reaches terminal count
3. On each succeeding CLKINn pulse, the count is decremented by two.
4. After the count reaches terminal count, OUTn is driven low and the count is reloaded.
5. On each succeeding CLKINn pulse, the count is decremented by two.
6. After the count reaches terminal count, OUTn is driven high and the count is reloaded.
7. The process is repeated from step 3.
10-12
TIMER/COUNTER UNIT
Control
Word = 16H
Count = 4
Writes to
Counter
n
CLKIN
n
GATE
OUT
n
n
Count
?
?
?
?
0004 0002 0004 0002 0004 0002 0004 0002 0004 0002
A2314-01
Figure 10-11. Mode 3 – Basic Operation (Even Count)
Odd count basic operation:
1. After a control word write, OUTn is driven high.
2. On the CLKINn pulse following a gate-trigger or when the count rolls over, count minus
one is loaded.
3. On each succeeding CLKINn pulse, the count is decremented by two.
4. When the count rolls over, OUTn is driven low and the count minus one is loaded. (This
causes OUTn to stay high for one more CLKINn pulse than it stays low.)
5. On each succeeding CLKINn pulse, the count is decremented by two.
6. After the count reaches terminal count, OUTn is driven high and the count minus one is
loaded.
7. The process is repeated from step 3.
10-13
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
Control
Word = 16H
Count = 5
Writes to
Counter n
CLKINn
GATEn
OUTn
?
?
?
?
0004 0002 0000 0004 0002 0004 0002 0000 0004 0002
Count
A2400-01
Figure 10-12. Mode 3 – Basic Operation (Odd Count)
NOTE
For an even count of N, OUTn remains high for N/2 counts and low for N/2
counts (provided GATEn remains high). For an odd count of N, OUTn remains
high for (N + 1)/2 counts and low for (N – 1)/2 counts (provided GATEn
remains high).
Figure 10-13 shows suspending the counting sequence. A low level on GATEn causes the counter
to drive OUTn active (When OUTn is low, a falling edge on GATEn causes OUTn to be driven
high immediately) and suspend counting. A rising edge on the GATEn causes the counter to be
reloaded with the count. A high level on GATEn resumes counting.
Control
Count = 4
Word = 16H
Writes to
Counter
n
CLKIN
n
GATE
OUT
n
n
?
?
?
?
0004 0002 0004 0002 0002 0002 0004 0002 0004 0002
Count
A2401-02
Figure 10-13. Mode 3 – Disabling the Count
10-14
Figure 10-14 and Figure 10-15 shows writing a new count. If the counter receives a gate-trigger
after writing a new count but before the end of the current half-cycle, the count is loaded on the
next CLKINn pulse and counting continues from the new count (Figure 10-14). Otherwise, the
new count is loaded at the end of the current half-cycle (Figure 10-15).
Control
Word = 16H
Writes to
Count = 8
Count = 10
Counter n
CLKINn
GATEn
OUTn
Count
?
?
?
?
8
6
4
2
8
6
4
10
8
6
A2407-02
Figure 10-14. Mode 3 – Writing a New Count (With a Trigger)
Control
Word = 16H
Count = 4
Count = 8
Writes to
Counter n
CLKINn
GATEn
OUTn
Count
?
?
?
?
4
2
4
2
4
2
8
6
4
2
A2406-01
Figure 10-15. Mode 3 – Writing a New Count (Without a Trigger)
10-15
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
10.2.5 Mode 4 – Software-triggered Strobe
count is loaded on the CLKINn pulse following a count write. When the counter reaches zero,
OUTn strobes low for one clock pulse. The counter rolls over and continues counting, but does
not strobe low when it reaches zero again. The counter strobes low only the first time it reaches
zero after a count write. A high level on a counter’s GATEn signal enables counting; a low level
on a counter’s GATEn signal disables counting.
Mode 4’s basic operation is outlined below and shown in Figure 10-16.
1. After a control word write, OUTn is driven high.
2. On the CLKINn pulse following the count write, the count is loaded.
3. On each succeeding CLKINn pulse, the count is decremented.
4. When the count reaches zero, OUTn is driven low.
5. On the following CLKINn pulse, OUTn is driven high.
NOTE
Writing a count of N causes OUTn to strobe low in N + 1 CLKINn pulses,
provided GATEn remains high. OUTn remains low for one CLKINn pulse,
then goes high.
Control
Count = 3
Word = 18H
Writes to
Counter n
CLKINn
GATEn
OUTn
Count
?
?
?
?
0003 0002 0001 0000 FFFF FFFE FFFD
A2315-01
Figure 10-16. Mode 4 – Basic Operation
10-16
TIMER/COUNTER UNIT
Figure 10-17 shows suspending the counting sequence. A low level on GATEn causes the counter
to suspend counting (both the state of OUTn and the count remain unchanged). A high level on
GATEn resumes counting.
Control
Count = 3
Word = 18H
Writes to
Counter n
CLKINn
GATEn
OUTn
Count
?
?
?
?
0003 0003 0003 0002 0001 0000 FFFF
A2402-01
Figure 10-17. Mode 4 – Disabling the Count
Figure 10-18 shows writing a new count. On the CLKINn pulse following the new count write,
the counter loads the new count and counting continues from the new count.
Control
Word = 18H
Count = 2
Count = 3
Writes to
Counter n
CLKINn
GATEn
OUTn
?
?
?
?
0003 0002 0001
0002 0001 0000 FFFF
Count
A2403-01
Figure 10-18. Mode 4 – Writing a New Count
10-17
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
10.2.6 Mode 5 – Hardware-triggered Strobe
Initializing a counter for mode 5 sets the counter’s OUTn signal, starting the counting sequence.
A gate-trigger loads the programmed count. When the counter reaches zero, OUTn strobes low
for one clock pulse. The counter then rolls over and continues counting, but OUTn does not strobe
low when the count reaches zero. The OUTn strobes low only the first time it reaches zero after
a count is loaded.
Mode 5’s basic operation is outlined below and shown in Figure 10-19.
1. After a control word write, OUTn is driven high.
2. On the CLKINn pulse following a gate-trigger, the count is loaded.
3. On each succeeding CLKINn pulse, the count is decremented.
4. When the count reaches zero, OUTn is driven low.
5. On the following CLKINn pulse, OUTn is driven high.
NOTE
Writing a count of N causes OUTn to strobe low N + 1 CLKINn pulses after
the counter receives a gate-trigger. OUTn remains low for one CLKINn pulse,
then goes high.
Control
Count = 3
Word = 1AH
Writes to
Counter
n
CLKIN
n
GATE
OUT
n
n
Count
?
?
?
?
?
0003 0002 0001 0000 FFFF 0003 0002 0001 0000
A2316-01
Figure 10-19. Mode 5 – Basic Operation
10-18
TIMER/COUNTER UNIT
Figure 10-20 shows retriggering the strobe with a gate-trigger. On the CLKINn pulse following
the retrigger, the counter reloads the count. The control logic then decrements the count on each
succeeding CLKINn pulse. OUTn remains high until the count reaches zero, then strobes low for
one CLKINn pulse.
Control
Word = 1AH
Writes to
Count = 3
Counter
n
CLKIN
n
GATE
OUT
n
n
Count
?
?
?
?
?
?
0003 0002 0003 0002 0001 0000 FFFF FFFE
A2404-01
Figure 10-20. Mode 5 – Retriggering the Strobe
Figure 10-21 shows the writing of a new count value. The counter waits for a gate-trigger to load
the new count; it does not affect the current sequence until the counter receives a trigger. On the
CLKINn pulse following the trigger, the control logic loads the new count. The control logic then
decrements the count on each succeeding CLKINn pulse. OUTn remains high until the count
reaches zero, then strobes low for one CLKINn pulse.
Control
Word = 1AH
Count = 3
Count = 5
Writes to
Counter
n
n
CLKIN
VCC
GATE
n
OUT
n
?
?
?
?
?
0003 0002 0001 0000 FFFF FFFE 0005 0004 0003
Count
A2405-01
Figure 10-21. Mode 5 – Writing a New Count Value
10-19
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
10.3 REGISTER DEFINITIONS
The following sections describe how to configure a counter’s input and output signals, initialize
a counter for a specific operating mode and count format, write count values to a counter, and
read a counter’s status and count.
10.3.1 Configuring the Input and Output Signals
Each counter is driven by a clock pulse on its CLKINn input. You can connect each counter’s
CLKINn input to either its timer clock (TMRCLKn) pin or the prescaled clock (PSCLK) signal.
The counters can handle up to 1/2 the processor clock (CLK2/4) input frequency but only up to
a maximum of 8 MHz (CLKINn frequency can never be more than 8 MHz). PSCLK is an internal
signal that is a prescale value of the processor’s internal clock. The frequency of PSCLK is pro-
grammable. See “Controlling the PSCLK Frequency” on page 8-7.
The GATEn signals of the counters can be controlled through hardware or software, as described
in the next two sections.
10.3.1.1
Hardware Control of GATEn
You can connect each counter’s GATEn signal to:
• Its timer gate (TMRGATEn) pin
• V
CC
Hardware (through a pin or V ) control of the GATEn requires that the SWGTEN bit in the
CC
TMRCFG register be reset.
10.3.1.2
Software Control of GATEn
You can also use the TMRCFG register to drive GATEn high or low through register bits. The
SWGTEN and GTnCON bits are used to control the GATEn signal. If SWGTEN is set, then the
value of the GTnCON bit causes the GATEn input of the counter to be driven to the corresponding
voltage level.
Table 10-4. GATEn Connection Options
GATEn
SWGTEN
GTnCON
connection
0
0
1
0
1
0
VCC
TMRGATEn
0
(Gaten is OFF)
1
1
1
(Gaten is ON)
The timer configuration register (TMRCFG) enables the counter’s CLKINn signals and deter-
mines each counter’s CLKINn and GATEn signal connections or logical value (Figure 10-22).
10-20
TIMER/COUNTER UNIT
.
Timer Configuration
TMRCFG
Expanded Addr: F834H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
TMRDIS
SWGTEN GT2CON CK2CON
GT1CON CK1CON GT0CON CK0CON
Bit
Number
Bit
Mnemonic
Function
7
TMRDIS
Timer Disable:
0 = Enables the CLKINn signals.
1 = Disables the CLKINn signals.
6
5
SWGTEN
Software GATEn Enable
0 = Connects GATEn to either the VCC pin or the TMRGATEn pin.
1 = Enables GT2CON, GT1CON, and GT0CON to control the
connections to GATE2, GATE1 and GATE0 respectively.
GT2CON
Gate 2 Connection:
SWGTEN
GT2CON
0
0
1
1
0
1
0
1
Connects GATE2 to VCC.
Connects GATE2 to the TMRGATE2 pin.
Turns GATE2 off.
Turns GATE2 on.
4
3
CK2CON
GT1CON
Clock 2 Connection:
0 = Connects CLKIN2 to the internal PSCLK signal.
1 = Connects CLKIN2 to the TMRCLK2 pin.
Gate 1 Connection:
SWGTEN
GT1CON
0
0
1
1
0
1
0
1
Connects GATE1 to VCC.
Connects GATE1 to the TMRGATE1 pin.
Turns GATE1 off.
Turns GATE1 on.
2
1
CK1CON
GT0CON
Clock 1 Connection:
0 = Connects CLKIN1 to the internal PSCLK signal.
1 = Connects CLKIN1 to the TMRCLK1 pin.
Gate 0 Connection:
SWGTEN
GT0CON
0
0
1
1
0
1
0
1
Connects GATE0 to VCC.
Connects GATE0 to the TMRGATE1 pin.
Turns GATE0 off.
Turns GATE0 on.
0
CK0CON
Clock 0 Connection:
0 = Connects CLKIN0 to the internal PSCLK signal.
1 = Connects CLKIN0 to the TMRCLK0 pin.
Figure 10-22. Timer Configuration Register (TMRCFG)
10-21
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
The peripheral pin selection registers (P3CFG and PINCFG) determine whether each counter’s
OUTn signal is connected to its TMROUTn pin. See Figure 10-1 for the TCU signal connections.
For details on the P3CFG and PINCFG registers see Figure 10-23 and Figure 10-24. The counter
output signals are automatically connected to the interrupt control unit. Counter 1’s output signal
(OUT1) is automatically connected to DMA channel 0, and counter 2’s output signal (OUT2) is
automatically connected to DMA channel 1.
Use P3CFG bits 0 and 1 to connect TMROUT0 and TMROUT1 to package pins.
Port 3 Configuration
P3CFG
Expanded Addr: F824H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
PM7
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7
PM7
Pin Mode:
0 = Selects P3.7 at the package pin.
1 = Selects COMCLK at the package pin.
6
5
4
3
2
PM6
PM5
PM4
PM3
PM2
Pin Mode:
0 = Selects P3.6 at the package pin.
1 = Selects PWRDOWN at the package pin.
Pin Mode:
0 = Selects P3.5 at the package pin.
1 = Connects master IR7 to the package pin (INT3).
Pin Mode:
0 = Selects P3.4 at the package pin.
1 = Connects master IR6 to the package pin (INT2).
Pin Mode:
0 = Selects P3.3 at the package pin.
1 = Connects master IR5 to the package pin (INT1).
Pin Mode:
0 = Selects P3.2 at the package pin.
1 = Connects master IR1 to the package pin (INT0).
1
0
PM1
PM0
Pin Mode:
See Table 5-1 on page 5-8 for all the PM1 configuration options.
Pin Mode:
See Table 5-1 on page 5-8 for all the PM0 configuration options.
Figure 10-23. Port 3 Configuration Register (P3CFG)
10-22
TIMER/COUNTER UNIT
Use PINCFG bit 5 to connect TMROUT2, TMRCLK2, and TMRGATE2 to package pins.
Pin Configuration
PINCFG
Expanded Addr: F826H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
—
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7
—
Reserved. This bit is undefined; for compatibility with future devices, do
not modify this bit.
6
5
PM6
PM5
Pin Mode:
0 = Selects CS6# at the package pin.
1 = Selects REFRESH# at the package pin.
Pin Mode:
0 = Selects the coprocessor signals, PEREQ, BUSY#, and ERROR#, at
the package pins.
1 = Selects the timer control unit signals, TMROUT2, TMRCLK2, and
TMRGATE2, at the package pins.
4
3
2
1
0
PM4
PM3
PM2
PM1
PM0
Pin Mode:
0 = Selects DACK0# at the package pin.
1 = Selects CS5# at the package pin.
Pin Mode:
0 = Selects EOP# at the package pin.
1 = Selects CTS1# at the package pin.
Pin Mode:
0 = Selects DACK1# at the package pin.
1 = Selects TXD1 at the package pin.
Pin Mode:
0 = Selects SRXCLK at the package pin.
1 = Selects DTR1# at the package pin.
Pin Mode:
0 = Selects SSIOTX at the package pin.
1 = Selects RTS1# at the package pin.
Figure 10-24. Pin Configuration Register (PINCFG)
10-23
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
10.3.2 Initializing the Counters
The timer control register (TMRCON) has three formats: control word, counter-latch, and read-
back. When writing to TMRCON, certain bit settings determine which format is accessed.
Use the TMRCON’s control word format (Figure 10-25) to specify a counter’s count format and
operating mode. Writing the control word forces OUTn to go to an initial state that depends on
the selected operating mode.
10-24
TIMER/COUNTER UNIT
Timer Control (Control Word Format)
TMRCON
Expanded Addr: F043H
ISA Addr:
0043H
XXH
Reset State:
7
0
SC1
SC0
RW1
RW0
M2
M1
M0
CNTFMT
Bit
Number
Bit
Mnemonic
Function
7–6
SC1:0
Select Counter:
Use these bits to specify a particular counter. The selections you make for
bits 5–0 define this counter’s operation.
00 = counter 0
01 = counter 1
10 = counter 2
11 is not an option for TMRCON’s control word format. Selecting 11
accesses TMRCON’s read-back format, which is shown in Figure 10-29.
5–4
RW1:0
Read/Write Select:
These bits select a read/write option for the counter specified by bits 7–6.
01 = read/write least-significant byte only
10 = read/write most-significant byte only
11 = read/write least-significant byte first, then most-significant byte
00 is not an option for TMRCON’s control word format. Selecting 00
accesses TMRCON’s counter-latch format, which is shown in Figure
10-27.
3–1
M2:0
Mode Select:
These bits select an operating mode for the counter specified by bits 7–6.
000 = mode 0
001 = mode 1
X10 = mode 2
X11 = mode 3
100 = mode 4
101 = mode 5
X is a don’t care.
0
CNTFMT
Count Format:
This bit selects the count format for the counter specified by bits 7–6.
0 = binary (16 bits)
1 = binary coded decimal (4 decades)
Figure 10-25. Timer Control Register (TMRCON – Control Word Format)
10-25
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
10.3.3 Writing the Counters
Use the write format of a counter’s Timer n register (TMRn) to specify a counter’s count. The
count must conform to the write selection specified in the control word (least-significant byte
only, most-significant byte only, or least-significant byte followed by the most-significant byte).
You can write a new count to a counter without affecting the counter’s programmed operating
mode. New counts must also conform to the specified write selection.
Timer n (Write Format)
TMRn (n = 0–2)
Expanded Addr: F040H, F041H
F042H
ISA Addr:
0040H, 0041H
0042H
Reset State:
XXH
7
0
CV7
CV6
CV5
CV4
CV3
CV2
CV1
CV0
Bit
Bit
Function
Number Mnemonic
7–0 CV7:0
Count Value:
Write a count value for the counter to these bits. When writing the
counter’s count value, follow the write selection specified in the counter’s
control word.
Figure 10-26. Timer n Register (TMRn – Write Format)
Table 10-5 lists the minimum and maximum initial counts for each mode.
Table 10-5. Minimum and Maximum Initial Counts
Mode
Minimum Count
Maximum Count
0–1
2–3
4–5
1
2
1
0
0
0
NOTE: 0 is equivalent to 216 for binary counting and 104 for BCD counting.
10-26
TIMER/COUNTER UNIT
10.3.4 Reading the Counter
To read the counter you can perform a simple read operation or send a latch command to the
counter. TMRCON contains two formats that allow you to send latch commands to individual
counters: the counter-latch and read-back format. The counter-latch command latches the count
of a specific counter. The read-back command latches the count and/or status of one or more spec-
ified counters.
10.3.4.1
Simple Read
To perform a simple read operation in modes 0, 2, 3 and 4, suspend the counter’s operation (using
you must disable the counter so that the count is not in the process of changing when it is read.
However, in modes 1 and 5, where the counter’s operation can not be suspended, the counter can
still be read. But since the counter is running, there is a minor inaccuracy in the read value.
10.3.4.2
Counter-latch Command
Use the counter-latch format of TMRCON (Figure 10-27) to latch the count of a specific counter.
To issue a counter-latch command to a counter, write to the TMRCON register with bits 5-4 reset
and SC1 and SC0 (bits 7-6) programmed appropriately. A counter continues to run even after the
count is latched. The counter-latch command allows reading the count without disturbing the
count in progress.
10-27
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
Timer Control (Counter-latch Format)
TMRCON
Expanded Addr: F043H
ISA Addr:
0043H
Reset State:
XXH
7
0
SC1
SC0
0
0
0
0
0
0
Bit
Number
Bit
Mnemonic
Function
7–6
SC1:0
Select Counter:
These bits specify the counter that receives the counter-latch command.
00 = counter 0
01 = counter 1
10 = counter 2
11 is not an option for TMRCON’s counter-latch format. Selecting 11
accesses TMRCON’s read-back format, which is shown in Figure 10-29.
5–4
3–0
—
—
Write zeros to these bits to issue a counter-latch command to the
counter specified by bits 7–6.
01, 10, and 11 are not valid options for TMRCON’s counter-latch format.
Reserved; for compatibility with future devices, write zeros to these bits.
NOTE: Bits 5–0 serve another function when you select the read-back command (SC1:0 = 11). See
Figure 10-29 for the read-back bit functions.
Figure 10-27. Timer Control Register (TMRCON – Counter-latch Format)
When a counter receives a counter-latch command, it latches the count. This count remains
latched until you either read the count or reconfigure the counter. When you send multiple
counter-latch commands without reading the counter, only the first counter-latch command latch-
es the count value.
After issuing a counter-latch command, you can read the counter’s TMRn register. When reading
the counter’s TMRn register you must follow the counter’s programmed read selection (least-sig-
nificant byte only, most-significant byte only, or least-significant byte followed by the most-sig-
nificant byte). If the counter is programmed for two-byte counts, you must read two bytes. You
need not read the two bytes consecutively; you may insert read, write, or programming operations
between the byte reads.
10-28
TIMER/COUNTER UNIT
You can interleave reads and writes of the same counter; for example, if the counter is pro-
grammed for the two-byte read/write selection, the following sequence is valid.
1. Read least-significant byte.
2. Write new least-significant byte.
3. Read most-significant byte.
4. Write new most-significant byte.
Timer n (Read Format)
TMRn (n = 0–2)
Expanded Addr: F040H, F041H
F042H
ISA Addr:
0040H, 0041H
0042H
Reset State:
XXH
7
0
CV7
CV6
CV5
CV4
CV3
CV2
CV1
CV0
Bit
Number
Bit
Mnemonic
Function
7–0
CV7:0
Count Value:
These bits contain the counter’s count value. When reading the
counter’s count value, follow the read selection specified in the counter’s
control word.
Figure 10-28. Timer n Register (TMRn – Read Format)
10-29
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
10.3.4.3
Read-back Command
Use the read-back format of TMRCON (Figure 10-29) to latch the count and/or status of one or
more counters. Latch a counter’s status to check its programmed operating mode, count format,
and read/write selection and to determine whether the latest count written to it has been loaded.
Timer Control (Read-back Format)
TMRCON
Expanded Addr: F043H
ISA Addr:
0043H
XXH
Reset State:
7
0
1
1
COUNT#
STAT#
CNT2
CNT1
CNT0
0
Bit
Bit
Function
Write ones to these bits to select the read-back command.
Number Mnemonic
7–6
5
—
00, 01, and 10 are not valid options for TMRCON’s read-back format.
Count Latch:
COUNT#
0 = Clearing this bit latches the count of each selected counter. Use bits
3–1 to select one or more of the counters.
1 = No effect
4
STAT#
Status Latch:
0 = Clearing this bit latches the status of each selected counter. Use bits
3–1 to select one or more of the counters.
1 = No effect
3
2
1
0
CNT2
CNT1
CNT0
—
Counter 2 Select:
0 = The actions specified by bits 5 and 4 do not affect counter 2.
1 = The actions specified by bits 5 and 4 affect counter 2.
Counter 1 Select:
0 = The actions specified by bits 5 and 4 do not affect counter 1.
1 = The actions specified by bits 5 and 4 affect counter 1.
Counter 0 Select:
0 = The actions specified by bits 5 and 4 do not affect counter 0.
1 = The actions specified by bits 5 and 4 affect counter 0.
Reserved; for compatibility with future devices, write zero to this bit.
Figure 10-29. Timer Control Register (TMRCON – Read-back Format)
10-30
TIMER/COUNTER UNIT
mand is functionally equivalent to several counter-latch commands, one for each counter latched.
Each counter's latched count and status is held until it is read or until you reconfigure the counter.
A counter’s latched count or status is automatically unlatched when read, but other counters’
latched values remain latched until they are read.
After latching a counter’s status and count with a read-back command, reading TMRn accesses
its status format (Figure 10-30). Reading TMRn again accesses its read format. When both the
count and status of a counter are latched, the first read of TMRn indicates the counter’s status and
the next one or two reads (depending on the counter’s read selection) indicate the counter’s count.
Subsequent reads return unlatched count values. When only the count of a counter is latched, then
the first one or two reads of TMRn return the counter’s count. When the counter is programmed
for the two-byte read selection, you must read two bytes.
10-31
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
Timer n (Status Format)
TMRn (n = 0–2)
Expanded Addr: F040H, F041H
F042H
ISA Addr:
0040H, 0041H
0042H
Reset State:
XXH
7
0
OUTPUT
NULCNT
RW1
RW0
M2
M1
M0
CNTFMT
Bit
Number
Bit
Mnemonic
Function
7
OUTPUT
Output Status:
This bit indicates the current state of the counter’s output signal.
0 = OUTn is low
1 = OUTn is high
6
NULCNT
Count Status:
This bit indicates whether the latest count written to the counter has
been loaded. Some modes require a gate-trigger before the counter
loads new count values.
0 = the latest count written to the counter has been loaded
1 = a count has been written to the counter but has not yet been loaded
5–4
3–1
RW1:0
M2:0
Read/Write Select Status:
These bits indicate the counter’s programmed read/write selection.
00 = Never occurs
01 = read/write least-significant byte only
10 = read/write most-significant byte only
11 = read/write least-significant byte first, then most-significant byte
Mode Status:
These bits indicate the counter’s programmed operating mode.
000 = mode 0
001 = mode 1
X10 = mode 2
X11 = mode 3
100 = mode 4
101 = mode 5
X is a don’t care.
0
CNTFMT
Counter Format Status:
This bit indicates the counter’s programmed count format.
0 = binary (16 bits)
1 = binary coded decimal (4 decades)
Figure 10-30. Timer n Register (TMRn – Status Format)
10-32
TIMER/COUNTER UNIT
When a counter receives multiple read-back commands, it ignores all but the first command; the
count/status that the core reads is the count/status latched from the first read-back command (see
Table 10-6).
Table 10-6. Results of Multiple Read-back Commands Without Reads
Command
Read-back Command
Command Result
Sequence
1
2
3
Latch counter 0’s count and status. Counter 0’s count and status latched.
Latch counter 1’s status.
Counter 1’s status latched.
Latch counter 2 and 1’s status.
Counter 2’s status latched; counter 1’s status
command ignored because command 2
already latched its status.
4
5
Latch counter 2’s count.
Counter 2’s count latched.
Latch counter 1’s count and status. Counter 1’s count latched; counter 1’s status
command ignored because command 2
already latched its status.
6
Latch counter 0’s count.
Counter 0’s count command ignored because
command 1 already latched its count.
10.4 PROGRAMMING CONSIDERATIONS
Consider the following when programming the TCU.
• The 16-bit counters are read and written a byte at a time. The control word format of
TMRCON selects whether you read or write the least-significant byte only, most-significant
byte only, or least-significant byte then most-significant byte (this is called the counter’s
read/write selection). You must read and write the counters according to their programmed
read/write selections.
• When you program a counter for the two-byte read or write selection, you must read or
write both bytes. If you’re using more than one subroutine to read or write a counter, make
sure that each subroutine reads or writes both bytes before transferring control.
• You can program the counters for either an internal or external clock source (to CLKINn).
The internal source is a prescaled value of the processor clock and therefore, is turned off in
the processor’s powerdown mode (processor clock is off). If an external clock source is
used, it is not affected by the processor’s powerdown mode, because the clock signal is
provided by an off-chip source. “Controlling Power Management Modes” on page 8-8
describes the processor’s powerdown and idle modes.
10-33
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
• With the readback command:
— If both the status and counter values are latched, the user can read the value of the
Read/Write selection bits from the status register to know what bytes of the counter
value are being latched in the TMRn register.
— If only the counter value is latched, you must know the Read/Write selection before the
counter value can be read correctly.
• When a read-back command is issued to latch both the counter and the status of a timer, the
TMRn register holds both of these values. The first read of TMRn accesses the status byte
and the next one or two reads (depending on the R/W format) of TMRn access the timer’s
counter value.
10.4.1 Timer/Counter Unit Code Examples
The example code contains these software routines:
InitTimer
Initializes the specified timer’s mode, counter value, inputs, and outputs
Configures the specified timer(s) for a read-back command
Latches the counter value of the specified timer
SetUp_ReadBack
CounterLatch
ReadCounter
Performs a simple read command on the specified timer’s current counter
value
TimerISR
Interrupt Service Routine for timer-generated interrupts
Code is also included that demonstrates how to change the timer’s counter value and issue a read-
back command. See Appendix C for included header files.
#include <conio.h>
#include “80386ex.h”
#include “ev386ex.h”
/*****************************************************************************
InitTimer:
Description:
This function initializes a timer’s inputs, outputs, operating mode,
and initial counter value.
Parameters:
Unit
Mode
Unit number of the timer. The processor supports 0, 1 or 2
Defines Counter Mode
Inputs
Output
InitCount
Enable
Specifies Input sources
Specifies Which Output to drive
Value to be loaded into count register
Enable (1) or disable (0) Timer
10-34
TIMER/COUNTER UNIT
Returns:Error Codes
E_INVALID_DEVICE
E_OK
-- Unit number specifies a non-existing device
-- Initialized OK, No error.
Assumptions:
REMAPCFG register has Expanded I/O space access enabled (ESE bit set);
This function also initializes the Timer-Counter Unit to be in the
Read/Write Format of least-significant byte first, then most-significant
byte
Syntax:
int error;
error = InitTimer
(TMR_2,
TMR_SQWAVE | TMR_CLK_BIN,
TMR_CLK_INTRN,
TMR_OUT_ENABLE,
0xFFFF,
TMR_ENABLE );
Real/Protected Mode:
No changes required.
*****************************************************************************/
int InitTimer(int Unit, WORD Mode, BYTE Inputs, BYTE Output, WORD InitCount,
int Enable)
{
BYTE
WORD
TmpByte;
TmrCntPort;
if(Unit > 2)
return E_INVALID_DEVICE;
TmrCntPort = 0xf040 + Unit;// Set depending on which timer
/* Set Pin configuration */
if(Unit < 2)
{
TmpByte = _GetEXRegByte(P3CFG) | (Output << Unit);
_SetEXRegByte(P3CFG,TmpByte);
// Bit 0 or 1
}
else
{
TmpByte = _GetEXRegByte(PINCFG) | (Output << 5); // Bit 5
_SetEXRegByte(PINCFG,TmpByte);
}
/* Set Timer Config */
TmpByte = _GetEXRegByte(TMRCFG);
// All Timers share this register,
// Keep previous settings
10-35
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
if(!Enable)
TmpByte |= 0x80;
// Set Timer Disable Bit
TmpByte |= (Inputs << (Unit*2));
_SetEXRegByte(TMRCFG,TmpByte);
// Set CKnCON and GTnCON bits
/* Set Timer Control Register */
TmpByte = Unit << 6;
// Set counter select
TmpByte |= (0x30 | Mode);
// Set R/W low then high byte and Mode bits
_SetEXRegByte(TMRCON,TmpByte);
/* Set Initial Counter Value */
TmpByte = HIBYTE(InitCount);
_SetEXRegByte(TmrCntPort, LOBYTE(InitCount));
_SetEXRegByte(TmrCntPort, TmpByte);
return E_OK;
}/* InitTimer */
/*****************************************************************************
SetUp_ReadBack:
Description:
This routine configures the Control Word for a Read Back Command.
After calling this function, the latched status and counter values
can be read from the TMRn registers. Example code of how to do
this for Timer2 is included after this function.
Parameters:
Timer0
Timer1
Timer2
GetStatus
GetCount
Cleared if Timer0’s values are not to be latched
Cleared if Timer1’s values are not to be latched
Cleared if Timer2’s values are not to be latched
Cleared if Status Byte is not to be latched
Cleared if Count Byte(s) is not to be latched
Returns:
None
Assumptions:
No assumptions have been made for this set-up function. However, if
a user were to latch only the counter value, the configured R/W
Format would have to be known. The setting of the R/W format can be
read from the Status Byte if this value is latched. An example of
this is included after the SetUp_ReadBack function.
Syntax:
#define ENABLE
1
10-36
TIMER/COUNTER UNIT
#define DISABLE 0
SetUp_ReadBack(DISABLE, DISABLE, ENABLE, ENABLE, ENABLE);
Real/Protected Mode:
No changes required
*****************************************************************************/
void SetUp_ReadBack( BYTE Timer0, BYTE Timer1, BYTE Timer2, BYTE GetStatus,
BYTE GetCount )
{
BYTE rb_control = 0;
rb_control |= 0xc0; // Set TMRCON to read-back command
if (GetStatus != 0)
rb_control &= 0xef;
if (GetCount != 0)
rb_control &= 0xdf;
if (Timer0 != 0)
rb_control |= 0x02;
if (Timer1 != 0)
rb_control |= 0x04;
if (Timer2 != 0)
rb_control |= 0x08;
_SetEXRegByte(TMRCON, rb_control);
} /* SetUp_ReadBack */
/*****************************************************************************
CounterLatch:
Description:
This function invokes a counter-latch command for the specified
timer and returns the latched counter value.
Parameters:
Timer Unit number of timer whose counter value is to be latched
10-37
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
Returns:
Counter Value of specified timer
Assumptions:
This function assumes that the R/W format is configured to be LSB
first, then MSB
Syntax:
WORD Counter_Value;
Counter_Value = CounterLatch(TMR_1);
Real/Protected Mode:
No changes required
*****************************************************************************/
WORD CounterLatch( BYTE Timer )
{
BYTE control_word = 0;
BYTE CounterL, CounterH;
WORD Counter;
control_word = Timer << 6;
control_word &= 0xc0;
_SetEXRegByte(TMRCON, control_word); //Select which counter
switch (Timer) {
case TMR_0:
CounterL = _GetEXRegByte(TMR0);
CounterH = _GetEXRegByte(TMR0);
break;
case TMR_1:
CounterL = _GetEXRegByte(TMR1);
CounterH = _GetEXRegByte(TMR1);
break;
case TMR_2:
CounterL = _GetEXRegByte(TMR2);
CounterH = _GetEXRegByte(TMR2);
break;
}
Counter = (((WORD)CounterH << 8) + CounterL);
return(Counter);
}/* CounterLatch */
10-38
TIMER/COUNTER UNIT
/*****************************************************************************
ReadCounter:
Description:
This function performs a simple read operation on the specified
timer. However, because the counter value is not latched, the timer
must be disabled, read, and then re-enabled.
Parameters:
Timer
Unit number of Timer whose count is being read
Returns:
Counter value that was read
Assumptions:
This function assumes that the R/W format is configured to be LSB
first, then MSB
Syntax:
WORD Counter_Value;
Counter_Value = ReadCounter(TMR0);
Real/Protected Mode:
No changes required
*****************************************************************************/
WORD ReadCounter(BYTE Timer)
{
BYTE CountL, CountH;
WORD Count = 0;
DisableTimer();
switch (Timer) {
case TMR_0:
CountL = _GetEXRegByte(TMR0);
CountH = _GetEXRegByte(TMR0);
break;
case TMR_1:
CountL = _GetEXRegByte(TMR1);
CountH = _GetEXRegByte(TMR1);
break;
10-39
Intel386™ EX EMBEDDED PROCESSOR USER’S MANUAL
case TMR_2:
CountL = _GetEXRegByte(TMR2);
CountH = _GetEXRegByte(TMR2);
break;
}
Count = (((WORD)CountH << 8) + CountL);
EnableTimer();
return(Count);
}/* ReadCounter */
/*****************************************************************************
TimerISR:
Description:
Interrupt Service Routine for Timer-generated interrupts.
Parameters:
None
Returns:
None
Assumptions:
None
Syntax:
Not called by user.
Real/Protected Mode:
No changes required
******************************************************************************/
void interrupt far TimerISR(void)
{
/* Write message out to serial port as an example */
SerialWriteStr(SIO_0, “In TimerISR\n”);
NonSpecificEOI();
}/* TimerISR */
// If this ISR services Timer1 or Timer2,
// an EOI is also needed for the Slave 8259
10-40
TIMER/COUNTER UNIT
/*****************************************************************************
Example of how to write a new initial counter value to a timer
This value can be rewritten at any time without affecting the
Counter’s programmed mode.
Before writing an initial count value, the Control Word must be
configured for the proper R/W and Count formats.
-->This example assumes that Timer1 is in the R/W format of LSB first,
then MSB, and that the Count format is binary.
_SetEXRegByte(TMR1, InitialCountL);
_SetEXRegByte(TMR1, InitialCountH);
*****************************************************************************/
/****************************************************************************/
***Example of how to issue a Read Back command for Timer2, latching
both the status and the counter.
BYTE Status, CountL, CountH, RWmode;
WORD Count;
SetUp_ReadBack(0, 0, 1, 1, 1);
Status = GetEXRegByte(TMR2);
//Configure Read Back command for timer2,
latching both status and count
RWmode = Status & 0x30; //Mask off bits that correspond to the Read/Write Mode
switch (RWmode) { //Read Counter Value according to configured R/W format
case 0x10:
//Read/Write least significant byte only
Count = _GetEXRegByte(TMR2);
break;
case 0x20: //Read/Write most significant byte only
CountH = _GetEXRegByte(TMR2);
Count = (WORD)CountH << 8;
break;
case 0x30: //Read/Write LSB first, then MSB
CountL = _GetEXRegByte(TMR2);
CountH = _GetEXRegByte(TMR2);
Count = (((WORD)CountH << 8) + CountL);
break;
}
*****************************************************************************/
10-41
11
ASYNCHRONOUS
SERIAL I/O UNIT
CHAPTER 11
ASYNCHRONOUS SERIAL I/O UNIT
The asynchronous serial I/O (SIO) unit provides a means for the system to communicate with ex-
data characters received from the CPU. The SIO unit consists of two independent SIO channels,
This chapter is organized as follows:
• Overview (see below)
• SIO Operation (page 11-4)
• Programming Considerations (page 11-32)
11.1 OVERVIEW
Each SIO channel contains a baud-rate generator, transmitter, receiver, and modem control unit.
These are shown in Figure 11-1 for SIO Unit 1 (see Figure 5-8 on page 5-15 for the SIO Unit 0
configuration). The baud-rate generator can be clocked by either the internal serial clock (SER-
CLK) signal or the COMCLK pin. The transmitter and receiver contain shift registers and buffers.
Data to be transmitted is written to the transmit buffer. The buffer’s contents are transferred to the
transmit shift register and shifted out via the transmit data pin (TXDn). Data received is shifted
in via the receive data pin (RXDn). When a data byte is received, the contents of the receive shift
register are transferred to the receive buffer. The modem control logic provides interfacing for the
handshaking signals between an SIO channel and a modem or data set.
In addition to the transmit and receive channels, each SIO can generate an interrupt or a request
to the DMA unit, or both. An interrupt can be generated when an error has occurred in the receive
channel, when the transmit channel is ready to transmit another character, when the receive chan-
nel is full, or when a change in any of the modem control signals has occurred. A DMA request
may be issued any time a channel’s receive buffer is full or its transmit buffer is empty. This al-
lows the SIO to run at higher speeds for more efficient processing of serial data.
11-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
SIOCFG.1
SIO1
P3CFG.7
0
1
COMCLK
(P3.7)†
0
1
BCLKIN
To/From I/O Port 3
To DMA
SERCLK
Receive Data
RXD1
(DRQ1)
SIOINT1
RBFDMA1
TXEDMA1
To ICU
To DMA
To DMA
PINCFG.2
1
1
Transmit Data
TXD1
(DACK1#)
SIOCFG.7
From DMA
0
PINCFG.3
CTS1#
(EOP#)
0
1
Clear to Send
To/From DMA
0
PINCFG.0
0
1
Request to Send
Data Set Ready
RTS1#
(SSIOTX)
From SSIO
DSR1#
(STXCLK)
0
1
To/From SSIO
DCD1#
(DRQ0)
0
1
Data Carrier
Detect
To DMA
PINCFG.1
0
1
Data Terminal
Ready
DTR1#
(SRXCLK)
To/From SSIO
To SSIO
RI1#
(SSIORX)
0
1
Ring Indicator
VCC
† Alternate pin signals are in parentheses.
A2519-02
Figure 11-1. Serial I/O Unit 1 Configuration
11-2
ASYNCHRONOUS SERIAL I/O UNIT
11.1.1 SIO Signals
Table 11-1 lists the SIOn signals.
Table 11-1. SIO Signals
Description
Device Pin or
Signal
Internal Signal
Baud-rate
Internal signal
SERCLK:
Generator
Clock Source
Device pin
(input)
This internal signal is the processor’s input clock, CLK2, divided by
four.
COMCLK:
An external source connected to this pin can clock the SIOn baud-rate
generator.
TXDn
Device pin
(output)
Transmit Data:
The transmitter uses this pin to shift serial data out. Data is
transmitted least-significant bit first.
RXDn
CTSn#
DSRn#
Device pin
(input)
Receive Data:
The receiver uses this pin to shift serial data in. Data is received least-
significant bit first.
Device pin
(input)
Clear to Send:
Indicates that the modem or data set is ready to exchange data with
the SIOn channel.
Device pin
(input)
Data Set Ready:
Indicates that the modem or data set is ready to establish the
communications link with the SIOn channel.
DCDn#
RIn#
Device pin
(input)
Data Carrier Detect:
Indicates that the modem or data set has detected the data carrier.
Device pin
(input)
Ring Indicator:
Indicates that the modem or data set has detected a telephone ringing
signal.
RTSn#
DTRn#
SIOINTn
Device pin
(output)
Request to Send:
Indicates to the modem or data set that the SIOn channel is ready to
exchange data.
Device pin
(output)
Data Terminal Ready:
Indicates to the modem or data set that the SIOn channel is ready to
establish a communications link.
Internal Signal
SIOINT:
This signal is connected to the interrupt control unit and is asserted
(HIGH) when any one of the following interrupt types has an active
condition and is enabled via the IER register: Receiver Error flag,
Received Data Available, Transmitter Holding Register Empty, or
Modem Status. The SIOINT signal is deasserted (LOW) upon the
appropriate interrupt service or reset operation.
11-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table 11-1. SIO Signals
Device Pin or
Signal
Description
Internal Signal
TXEDMAn
Internal Signal
Transmitter Empty:
When this signal is high, the Transmitter Holding Register is empty
(transmit data has been loaded into the Transmit Shift Register).
RBFDMAn
Internal Signal
Receiver Full:
When high, this signal indicates that the Receive Buffer has been
loaded with data from the Receive Shift Register.
11.2 SIO OPERATION
The following sections describe the operation of the baud-rate generator, transmitter, and receiver
and discusses the modem control logic, SIO diagnostic mode, and SIO interrupt sources.
11.2.1 Baud-rate Generator
Each SIO channel’s baud-rate generator provides the clocking source for the channel’s transmitter
and receiver. The baud-rate generator can divide its input (BCLKIN) by any divisor from 1 to
(216–1). The output frequency is 16 times the desired bit time. The transmitter shifts data out on
the rising edge of BCLKIN. The receiver samples input data in the middle of a bit time.
The internal serial clock (SERCLK) signal or the COMCLK pin can be connected to the baud-
rate generator’s BCLKIN signal (Figure 11-2). The SIO configuration register (SIOCFG) selects
one of these sources.
SIOCFG.n
SIOn
Baud-rate
Generator
SERCLK
Baud Rate
Generator
Output
1
0
CLK2
÷ 2
÷ 2
BCLKIN
Frequency
COMCLK
(pin mux)
A2524-02
Figure 11-2. SIOn Baud-rate Generator Clock Sources
SERCLK provides a baud-rate input frequency (BCLKIN) of CLK2/4. The COMCLK pin allows
an external source with a maximum frequency of CLK2/4 to provide the baud-rate generator in-
put frequency.
11-4
ASYNCHRONOUS SERIAL I/O UNIT
The baud-rate generator’s output frequency is determined by BCLKIN and a divisor as follows.
BCLKIN frequency
baud-rate generator output frequency = --------------------------------------------------- ,
divisor
baud rate generator output frequency
bit rate = --------------------------------------------------------------------------------------------------
16
The minimum divisor value is 1, giving a maximum baud rate of BCLKIN. The maximum divisor
value is 0FFFFH (65535), giving a minimum of BCLKIN/65535. For example, the maximum and
minimum bit-rate frequencies using SERCLK with a 25 MHz device (CLK2 = 50 MHz) or
COMCLK with a 12.5 MHz input are shown in Table 11-2. Table 11-3 shows the divisor values
required for common baud rates.
Table 11-2. Maximum and Minimum Output Bit Rates
Input Clock (BCLKIN)
12.5 MHz
12.5 MHz
Divisor
Output Bit Rate
0001H
781.25 KHz (max)
11.921 Hz (min)
0FFFFH
Table 11-3. Divisor Values for Common Bit Rates
Divisor
Input Clock (BCLKIN)
Output Bit Rate
% Error
1AEH
6BH
48H
16.5 MHz (processor clock = 33 MHz)
16.5 MHz (processor clock = 33 MHz)
16.5 MHz (processor clock = 33 MHz)
12.5 MHz (processor clock = 25 MHz)
12.5 MHz (processor clock = 25 MHz)
12.5 MHz (processor clock = 25 MHz)
10 MHz (processor clock = 20 MHz)
10 MHz (processor clock = 20 MHz)
10 MHz (processor clock = 20 MHz)
8 MHz (processor clock = 16 MHz)
8 MHz (processor clock = 16 MHz)
8 MHz (processor clock = 16 MHz)
2400 b/s
9600 b/s
14.4 Kb/s
2400 b/s
9600 b/s
14.4 Kb/s
2400 b/s
9600 b/s
14.4 Kb/s
2400 b/s
9600 b/s
14.4 Kb/s
– 0.07
+ 0.39
– 0.54
+ 0.15
+ 0.47
+ 0.46
+ 0.15
+ 0.16
+ 0.94
+ 0.16
+ 0.16
– 0.79
145H
51H
36H
104H
41H
2BH
0D0H
34H
23H
11-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
11.2.2 SIOn Transmitter
The data frame for transmissions is programmable. It consists of a start bit, 5 to 8 data characters,
an optional parity bit, and 1 to 2 stop bits. The transmitter can produce even, odd, forced, or no
parity. The transmitter can also produce break conditions. A break condition forces the serial out-
put (TXDn) to the spacing (logic 0) state for longer than a transmission time (the time of the start
bit + data bits + parity bit + stop bits). On the receiving end, a break condition sets an error flag.
Forced parity (“sticky bit”) allows the SIO to be used in multiprocessor communications. When
using forced parity the serial port uses the parity bit to distinguish between address and data bytes.
Forced parity is enabled in the SIO by setting the PEN and SP bits in the serial line control register
(Figure 11-15). When enabled for forced parity, the bit that is transmitted in the parity bit location
is the complement of the EPS bit (also in the serial line control register). In the receiver, if PEN
and SP are 1, the receiver compares the bit that is received in the parity bit location with the com-
plement of the EPS bit. If the values being compared are not equal, the receiver sets the Parity
Error bit in LSR and causes an error interrupt if line status interrupts are enabled.
For example, if forced parity is enabled and EPS is 0, the receiver expects the bit received at the
parity bit location to be 1. If it is not, the parity error bit is set. By forcing the bit value at the parity
bit location, rather than calculating a parity value, a system with a master transmitter and multiple
receivers can identify some transmitted characters as receiver addresses and the rest of the char-
acters as data. If PEN = 0, the SP bit is ignored.
Each SIO channel transmitter contains a transmit shift register, a transmit buffer, and a transmit
data pin (TXDn). Data to be transmitted is written to the transmit buffer. The transmitter then
transfers the data to the transmit shift register. The transmitter shifts the data along with asynchro-
nous communication bits (start, stop, and parity) out via the TXDn pin. The TXD0 and TXD1
pins are multiplexed with other functions. The pin configuration registers (PINCFG and P2CFG)
determine whether a TXDn signal or an alternate function is connected to the package pin.
11-6
ASYNCHRONOUS SERIAL I/O UNIT
Baud-rate
Clock
SIOn Transmit Shift
Register
S
y
s
TXDn
(pin mux)
t
e
m
Transmit Buffer Empty
(To ICU and DMA)
SIOn Transmit Buffer
B
u
s
A2326-01
Figure 11-3. SIOn Transmitter
At reset, TBE and TE are set, indicating that the transmit buffer and shift register are empty. Writ-
ing data to the transmit buffer clears TBE and TE. When the transmitter transfers data from the
buffer to the shift register, TBE is set. Unless new data is written to the transmit buffer, TE is set
when the transmitter finishes shifting out the shift register’s contents.
The transmitter’s transmit buffer empty signal can be connected to the interrupt control and DMA
units. Figure 11-4 shows the process for transmitting data.
11-7
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Select the BCLKIN source and
the transmitter input baud rate.
Select the data frame. (Word length, number
of stop bits, and type of parity.) Enable interrupts
and/or DMA.
Is
No
transmit buffer
empty?
Yes
Write data to transmit buffer register.
(ISR or DMA cycle)
Transmitter transfers data to shift register
and sets transmit buffer empty flag, causing
an interrupt or DMA request.
More
Data to
Transmit
?
Yes
Transmitter shifts data frame onto
the TXDn pin. Data is transmitted
least-significant bit first.
No
Transmitter shifts out last stop bit
then sets the transmitter empty flag.
End
A2527-02
Figure 11-4. SIOn Data Transmission Process Flow
11-8
ASYNCHRONOUS SERIAL I/O UNIT
11.2.3 SIOn Receiver
The data frame for receptions is programmable, and is identical to the data frame for transmis-
sions. It consists of a start bit, 5 to 8 data characters, an optional parity bit, and 1 to 2 stop bits.
The receiver can be programmed for even, odd, forced, or no parity. When the receiver detects a
parity condition other than what it was programmed for, it sets a parity error flag. In addition to
detecting parity errors, the receiver can detect break conditions, framing errors, and overrun er-
rors.
• A break condition indicates that the received data input is held in the spacing (logic 0) state
for longer than a data transmission time (the time of the start bit + data bits + parity + stop
bits).
• A framing error indicates that the received character did not have a valid stop bit.
• An overrun error indicates that new data overwrote old data before the old data was read.
Each SIO channel receiver contains a receive shift register, a receive buffer, and a receive data
pin (RXDn). Data received is shifted into the receive shift register via the RXDn pin. Once a data
byte has been received, the receiver strips off the asynchronous communication bits (start, stop,
and parity) and transfers the contents of its shift register to the receive buffer.
The RXD0 pin is multiplexed with another function. The pin configuration register (P2CFG) de-
termines whether the RXD0 signal or the alternate function is connected to the package pin.
Baud-rate
Clock
S
y
s
SIOn Receive Shift
Register
RXDn
(pin mux)
Receiver Errors
(To ICU)
t
e
m
SIOn Receive Buffer
Receive Buffer Full
(To ICU and DMA)
B
u
s
A2327-02
Figure 11-5. SIOn Receiver
11-9
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
The receiver contains a receive buffer full (RBF) flag and flags for each of the error conditions
described above. At reset, RBF and each of the error flags (PE, FE, OE, and BI) are clear, indi-
cating that the receive buffer is empty, and no error has occurred. When a character is received
the receiver checks for parity, framing or break errors, and sets the appropriate bits, if necessary.
It then shifts the data into the receive buffer, sets the OE bit if an overrun occurs, and then sets
the RBF flag. Reading the receive buffer clears the RBF flag and any error flags in the LSR that
may have been set, except for the OE bit. The OE bit is cleared by reading the LSR.
High speed serial transfers may require using the DMA to eliminate interrupt latency time in ser-
vicing the SIO. Since the SIO unit clears the error bits in the line status register each time the re-
ceive buffer register is read, it would be impossible to detect an error using DMA. Because of
this, two RBF signals are used:
• One RBF signal (RBFDMA) goes directly to the DMA unit. This signal is blocked when an
error (parity, overrun, break, or framing) occurs. This prevents a DMA request from being
generated by the RBF.
• The other RBF signal (RBFINT) goes directly to the interrupt priority logic and out on
SIOINT if enabled in the Interrupt Enable Register.
When the Interrupt Enable Register is programmed to generate an SIOINT on receiver errors, the
error can be serviced as part of the interrupt handler.
11-10
ASYNCHRONOUS SERIAL I/O UNIT
Select the BCLKIN source and
the receiver input baud rate.
Select the data frame. (Word length,
number of stop bits, and type of
parity.) Enable interrupts and/or DMA.
Receiver shifts data into shift register
from the RXDn pin.
Was
a framing error
detected?
Was
a parity error
detected?
No
No
Was
a break condition
detected?
No
Yes
Yes
Yes
Receiver sets the
parity error flag.
Receiver sets
the framing
error flag.
Receiver sets
the break
interrupt flag.
Any
error flags
set?
Service error
interrupt
(if enabled)
Yes
Is
Yes
No
receive
buffer full flag
set?
Receiver transfers
data to receive
buffer and sets
Service error
interrupt
(if enabled)
overrun error flag.
No
Receiver transfers data to receive
buffer and sets receive buffer full
flag.
End
A2525-02
Figure 11-6. SIOn Data Reception Process Flow
11-11
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
11.2.4 Modem Control
The modem control logic provides interfacing for four input signals and two output signals used
for handshaking and status indication between the SIOn and a modem or data set. An external
modem or data set uses the input signals to inform the SIOn when:
• It is ready to establish a communications link (DSRn#)
• It has detected a data carrier signal (DCDn#)
• It has detected a telephone ringing signal (RIn#)
• It is ready to exchange data (CTSn#)
The SIOn uses its output signals to inform the modem or data set when it is ready to establish a
communication link (DTRn#), and when it is ready to exchange data (RTSn#).
The modem output signals can be internally connected to the modem input signals using the SIO
configuration register. In this case, the modem input signals are disconnected from the pins,
RTSn# is connected to CTSn#, DTRn# is connected to both DSRn# and DCDn#, and VCC is con-
nected RIn#.
The SIO contains status flags that indicate the current state of the modem control input signals
and status flags that indicate whether any of the modem control input signals have changed state.
11.2.5 Diagnostic Mode
The SIO channels provide a diagnostic mode to aid in isolating faults in the communications link.
In this mode, data that is transmitted is immediately received. This feature allows the processor
to verify the internal transmit and receive data paths of an SIOn channel.
The diagnostic mode connections are as follows:
• The transmitter serial output (TXDn) is set to a logic 1 state.
• The receiver serial input (RXDn) is disconnected from the pin.
• The transmit shift register output is “looped back” into the receive shift register.
• The four modem control inputs (CTSn#, DSRn#, DCDn#, and RIn#) are disconnected from
the pins and controlled by modem control register bits.
• The modem control output pins (RTSn#, DTRn#) are forced to their inactive states.
11-12
ASYNCHRONOUS SERIAL I/O UNIT
11.2.6 SIO Interrupt and DMA Sources
11.2.6.1 SIO Interrupt Sources
Each SIO channel has four status signals: receiver line status, receiver buffer full, transmit buffer
empty, and modem status. An overrun error, parity error, framing error, or break condition can
activate the receiver line status signal. When the receiver transfers data from its shift register to
its buffer, it activates the receive buffer full signal. When the transmitter transfers data from its
transmit buffer to its transmit shift register, it activates the transmit buffer empty signal. A change
on any of the modem control input signals activates the modem status signal. When the modem
signals are connected internally either through the configuration register or the diagnostic mode,
changes of state still activate the modem status signal. For these cases, however, the signal values
are controlled by register bits rather than by external input signals.
Each of the four status signals can be used as an interrupt request source for the SIOINTn signal.
The Interrupt Enable register (IER) is used to enable any or all of the status signals as interrupt
sources. When an SIOINTn occurs the IP# bit (bit 0) of the Interrupt ID register (IIRn) is cleared
and the interrupt handler must determine which of the status signals caused the interrupt by read-
ing bits 1 and 2 of the IIRn register (Table 11-4). When more than one status signal is enabled as
an interrupt source and two or more are active at the same time then the source of the interrupt is
based on a fixed priority scheme (Table 11-4).
Table 11-4. Status Signal Priorities and Sources
Interrupt ID Register
Priority
Status Signal
Activated By
Bit 2 Bit 1 Bit 0
1 (Highest)
Receiver Line Status overrun error, parity error, framing error, or
break condition
1
1
0
1
0
1
0
2
Receive Buffer Full
the receiver transferring data from its shift
register to its buffer
0
3
Transmit Buffer Empty the transmitter transmitting data from its
transmit buffer to its transmit shift register
0
0
4 (Lowest)
Modem Status
a change on any of the modem control
input signals (CTSn#, DCDn#, DSRn#, and
RIn#
0
0
11.2.6.2
SIO DMA sources
The transmit and receive channel on each SIO is supported by both DMA channels. The receiver
buffer full and transmit buffer empty signals of the line status register are brought out as
RBFDMAn, and TXEDMAn. The TXEDMAn signal is connected directly to the multiplexers
controlling the source of DREQn for each of the two DMA channels. The RBFDMAn signal is
also connected to the DREQn muxes, but it is qualified by the LSR error conditions so that the
DMA request is blocked if an error has occurred in the reception of a character. This prevents the
DMA from transferring a character from the SIO with an error.
11-13
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
11.2.7 External UART Support
Many PC compatible applications may need to support COM3 and COM4 serial ports. Since the
integrated serial ports are mapped to I/O addresses that support only COM1 and COM2, an inter-
face to support an external serial I/O unit has been included. The master ICU interrupt inputs IR3
and IR4 may be brought out to package pins as INT8 (muxed with P3.1/TMROUT1) and INT9
(muxed with P3.0/TMROUT0), respectively.
In order to select between the internal SIO units and the external SIO units, use the OUT2 bit in
the modem control register (MCR). In normal user mode (with no loopback) clear the OUT2 bit
to enable external SIO support, and set OUT2 to enable internal SIO support. Doing this, com-
bined with the correct settings of the P3CFG.1:0 and INTCFG.6:5 bits connect the INT8 and
INT9 pins to IR3 and IR4 of the master ICU, respectively. Note that the reset state of P3CFG and
INTCFG enables SIOINT and disconnects OUT2 gating. See Chapter 5, “DEVICE CONFIGU-
RATION” (Tables 5-1 and 5-2) for more details on how to select this option.
11-14
ASYNCHRONOUS SERIAL I/O UNIT
11.3 REGISTER DEFINITIONS
Table 11-5 lists the registers associated with the SIO unit and the following sections contain bit
descriptions for each register.
Table 11-5. SIO Registers (Sheet 1 of 2)
Expanded
Address
PC/AT*
Address
Register
PINCFG
Function
0F826H
—
Pin Configuration:
(read/write)
Connects the SIO1 transmit data (TXD1), data terminal ready
(DTR1#), and request to send (RTS1#) signals to package pins.
P1CFG
(read/write)
0F820H
—
—
Port 1 Configuration:
Connects the SIO0 ring indicator (RI0#), data set ready (DSR0#),
data terminal ready (DTR0#), request to send (RTS0#), and data
carrier detected (DCD0#) signals to package pins.
P2CFG
0F822H
Port 2 Configuration:
(read/write)
Connects the SIO0 clear to send (CTS0#), transmit data (TXD0),
and receive data (RXD0) signals to package pins.
P3CFG
(read/write)
0F824H
0F836H
—
—
Port 3 Configuration:
Connects COMCLK to the package pin.
SIOCFG
SIO and SSIO Configuration:
(read/write)
Connects the SIOn modem input signals internally or to package
pins and connects either the internal SERCLK signal or the
COMCLK pin to the SIOn baud-rate generator input.
DLL0
DLL1
(read/write)
0F4F8H
0F8F8H
03F8H
02F8H
Divisor Latch Low:
Stores the lower 8 bits of the SIOn baud-rate generator divisor.
DLH0
DLH1
(read/write)
0F4F9H
0F8F9H
03F9H
02F9H
Divisor Latch High:
Stores the upper 8 bits of the SIOn baud-rate generator divisor.
TBR0
TBR1
(write only)
0F4F8H
0F8F8H
03F8H
02F8H
Transmit Buffer:
Holds the data byte to transmit.
RBR0
RBR1
(read only)
0F4F8H
0F8F8H
03F8H
02F8H
Receiver Buffer:
Holds the data byte received.
LCR0
LCR1
0F4FBH
0F8FBH
03FBH
02FBH
Line Control:
Specifies the data frame (word length, number of stop bits, and
type of parity) for transmissions and receptions. Allows the
transmitter to transmit a break condition.
(read/write)
LSR0
LSR1
(read only)
0F4FDH
0F8FDH
03FDH
02FDH
Line Status:
Contains the transmitter empty, transmit buffer empty, receive
buffer full, and receive error flags.
IER0
IER1
(read/write)
0F4F9H
0F8F9H
03F9H
02F9H
Interrupt Enable:
Independently connects the four signals (modem status, receive
line status, transmit buffer empty, and receive buffer full) to the
interrupt request output (SIOINTn).
11-15
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table 11-5. SIO Registers (Sheet 2 of 2)
Expanded
Address
PC/AT*
Address
Register
Function
IIR0
IIR1
0F4FAH
0F8FAH
03FAH
02FAH
Interrupt ID:
Indicates whether the modem status, transmit buffer empty,
receive buffer full, or receiver line status signal generated an
interrupt request.
(read only)
MCR0
MCR1
(read/write)
0F4FCH
0F8FCH
03FCH
02FCH
Modem Control:
Controls the interface with the modem or data set.
Allows use of external UARTs.
MSR0
MSR1
(read/write)
0F4FEH
0F8FEH
03FEH
02FEH
Modem Status:
Provides the current state of the control lines for the modem or
data set to the CPU.
SCR0
SCR1
(read/write)
0F4FFH
0F8FFH
03FFH
02FFH
Scratch Pad:
An 8-bit read/write register available for use as a scratch pad; has
no effect on SIOn operation.
For PC compatibility, the SIO unit accesses its 11 registers through 8 I/O addresses. The RBRn,
TBRn, and DLLn registers share the same addresses and the IERn and DLHn registers share the
same addresses. Bit 7 (DLAB) of the LCRn determines which register is accessed during a read
or write operation (Table 11-6).
Table 11-6. Access to Multiplexed Registers
Register Accessed
Expanded Address
PC/AT Address
DLAB = 0
DLAB = 1
0F4F8H (read)
03F8H (read)
RBR0
TBR0
IER0
DLL0
DLL0
DLH0
DLL1
DLL1
DLH1
0F4F8H (write)
03F8H (write)
0F4F9H (read/write)
0F8F8H (read)
03F9H (read/write)
02F8H (read)
RBR1
TBR1
IER1
0F8F8H (write)
02F8H (write)
0F8F9H (read/write)
02F9H (read/write)
11-16
ASYNCHRONOUS SERIAL I/O UNIT
11.3.1 Pin and Port Configuration Registers (PINCFG and PnCFG [n = 1–3])
Use PINCFG bits 2:0 to connect the SIO1 signals to package pins.
Pin Configuration
PINCFG
Expanded Addr: F826H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
—
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7
—
Reserved. This bit is undefined; for compatibility with future devices, do
not modify this bit.
6
5
PM6
PM5
Pin Mode:
0 = Selects CS6# at the package pin.
1 = Selects REFRESH# at the package pin.
Pin Mode:
0 = Selects the coprocessor signals, PEREQ, BUSY#, and ERROR#, at
the package pins.
1 = Selects the timer control unit signals, TMROUT2, TMRCLK2, and
TMRGATE2, at the package pins.
4
3
2
1
0
PM4
PM3
PM2
PM1
PM0
Pin Mode:
0 = Selects DACK0# at the package pin.
1 = Selects CS5# at the package pin.
Pin Mode:
0 = Selects EOP# at the package pin.
1 = Selects CTS1# at the package pin.
Pin Mode:
0 = Selects DACK1# at the package pin.
1 = Selects TXD1 at the package pin.
Pin Mode:
0 = Selects SRXCLK at the package pin.
1 = Selects DTR1# at the package pin.
Pin Mode:
0 = Selects SSIOTX at the package pin.
1 = Selects RTS1# at the package pin.
Figure 11-7. Pin Configuration Register (PINCFG)
11-17
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Use P1CFG bits 4:0 to connect SIO0 signals to package pins.
Port 1 Configuration
P1CFG
Expanded Addr: F820H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
PM7
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7
PM7
Pin Mode:
0 = Selects P1.7 at the package pin.
1 = Selects HLDA at the package pin.
6
5
4
3
2
1
0
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Pin Mode:
0 = Selects P1.6 at the package pin.
1 = Selects HOLD at the package pin.
Pin Mode:
0 = Selects P1.5 at the package pin.
1 = Selects LOCK# at the package pin.
Pin Mode:
0 = Selects P1.4 at the package pin.
1 = Selects RI0# at the package pin.
Pin Mode:
0 = Selects P1.3 at the package pin.
1 = Selects DSR0# at the package pin.
Pin Mode:
0 = Selects P1.2 at the package pin.
1 = Selects DTR0# at the package pin.
Pin Mode:
0 = Selects P1.1 at the package pin.
1 = Selects RTS0# at the package pin.
Pin Mode:
0 = Selects P1.0 at the package pin.
1 = Selects DCD0# at the package pin.
Figure 11-8. Port 1 Configuration Register (P1CFG)
11-18
ASYNCHRONOUS SERIAL I/O UNIT
Use P2CFG bits 7–5 to connect SIO0 signals to package pins.
Port 2 Configuration
P2CFG
Expanded Addr: F822H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
PM7
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7
PM7
Pin Mode:
0 = Selects P2.7 at the package pin.
1 = Selects CTS0# at the package pin.
6
5
4
3
2
1
0
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Pin Mode:
0 = Selects P2.6 at the package pin.
1 = Selects TXD0 at the package pin.
Pin Mode:
0 = Selects P2.5 at the package pin.
1 = Selects RXD0 at the package pin.
Pin Mode:
0 = Selects P2.4 at the package pin.
1 = Selects CS4# at the package pin.
Pin Mode:
0 = Selects P2.3 at the package pin.
1 = Selects CS3# at the package pin.
Pin Mode:
0 = Selects P2.2 at the package pin.
1 = Selects CS2# at the package pin.
Pin Mode:
0 = Selects P2.1 at the package pin.
1 = Selects CS1# at the package pin.
Pin Mode:
0 = Selects P2.0 at the package pin.
1 = Selects CS0# at the package pin.
Figure 11-9. Port 2 Configuration Register (P2CFG)
11-19
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Use P3CFG bit 7 to connect the COMCLK pin to the package pin.
Port 3 Configuration
P3CFG
Expanded Addr: F824H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
PM7
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7
PM7
Pin Mode:
0 = Selects P3.7 at the package pin.
1 = Selects COMCLK at the package pin.
6
5
4
3
2
PM6
PM5
PM4
PM3
PM2
Pin Mode:
0 = Selects P3.6 at the package pin.
1 = Selects PWRDOWN at the package pin.
Pin Mode:
0 = Selects P3.5 at the package pin.
1 = Connects master IR7 to the package pin (INT3).
Pin Mode:
0 = Selects P3.4 at the package pin.
1 = Connects master IR6 to the package pin (INT2).
Pin Mode:
0 = Selects P3.3 at the package pin.
1 = Connects master IR5 to the package pin (INT1).
Pin Mode:
0 = Selects P3.2 at the package pin.
1 = Connects master IR1 to the package pin (INT0).
1
0
PM1
PM0
Pin Mode:
See Table 5-1 on page 5-8 for all the PM1 configuration options.
Pin Mode:
See Table 5-1 on page 5-8 for all the PM0 configuration options.
Figure 11-10. Port 3 Configuration Register (P3CFG)
11-20
ASYNCHRONOUS SERIAL I/O UNIT
11.3.2 SIO and SSIO Configuration Register (SIOCFG)
Use SIOCFG to select the baud-rate generator clock source for the SIO channels and to have a
channel’s modem input signals connected internally rather than to package pins. Selecting the in-
ternal modem signal connection option connects RTS# to CTS#, DTR# to DSR# and DCD#, and
V
CC to RI#. The modem signal connections for this internal option are shown in Figure 11-20.
SIO and SSIO Configuration
SIOCFG
Expanded Addr: F836H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
S1M
S0M
—
—
—
SSBSRC
S1BSRC
S0BSRC
Bit
Number
Bit
Mnemonic
Function
SIO1 Modem Signal Connections:
7
S1M
0 = Connects the SIO1 modem input signals to the package pins.
1 = Connects the SIO1 modem input signals internally.
6
S0M
SIO0 Modem Signal Connections:
0 = Connects the SIO0 modem input signals to the package pins.
1 = Connects the SIO0 modem input signals internally.
5–3
2
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
SSBSRC
SSIO Baud-rate Generator Clock Source:
0 = Connects the internal PSCLK signal to the SSIO baud-rate
generator.
1 = Connects the internal SERCLK signal to the SSIO baud-rate
generator.
1
0
S1BSRC
S0BSRC
SIO1 Baud-rate Generator Clock Source:
0 = Connects the COMCLK pin to the SIO1 baud-rate generator.
1 = Connects the internal SERCLK signal to the SIO1 baud-rate
generator.
SIO0 Baud-rate Generator Clock Source:
0 = Connects the COMCLK pin to the SIO0 baud-rate generator.
1 = Connects the internal SERCLK signal to the SIO0 baud-rate
generator.
Figure 11-11. SIO and SSIO Configuration Register (SIOCFG)
11-21
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
11.3.3 Divisor Latch Registers (DLLn and DLHn)
Use these registers to program the baud-rate generator’s output frequency. The baud-rate gener-
ator’s output determines the transmitter and receiver bit times.
Divisor Latch Low
DLL0, DLL1
(read/write)
DLL0
DLL1
F8F8H
02F8H
02H
Expanded Addr: F4F8H
ISA Addr:
Reset State:
03F8H
02H
7
0
0
LD7
LD6
LD5
LD4
LD3
LD2
LD1
LD0
Divisor Latch High
DLH0, DLH1
(read/write)
DLH0
DLH1
F8F9H
02F9H
00H
Expanded Addr: F4F9H
ISA Addr:
Reset State:
03F9H
00H
7
UD15
UD14
UD13
UD12
UD11
UD10
UD9
UD8
Bit
Number
Bit
Mnemonic
Function
Lower 8 Divisor and Upper 8 Divisor Bits:
DLLn
(7–0)
LD7:0
Write the lower 8 divisor bits to DLLn and the upper 8 divisor bits to
DLHn. The baud-rate generator output is a function of the baud-rate
generator input (BCLKIN) and the 16-bit divisor.
DLHn
(7–0)
UD15:8
BCLKIN frequency
baud-rate generator output frequency = ----------------------------------------------------
divisor
bit rate (shifting rate) = baud-rate generator output frequency/16
NOTE: The divisor latch registers share address ports with other SIO registers. Bit 7 (DLAB) of
LCRn must be set in order to access the divisor latch registers.
If DLL = DLH = 00H, baud-rate generator ouput frequency = 0 (stops clock).
Figure 11-12. Divisor Latch Registers (DLLn and DLHn)
11-22
ASYNCHRONOUS SERIAL I/O UNIT
11.3.4 Transmit Buffer Register (TBRn)
Write the data words to be transmitted to TBRn. Use the interrupt control or DMA units or poll
the serial line status register (LSRn) to determine whether the transmit buffer is empty.
Transmit Buffer
TBR0, TBR1
(write only)
TBR0
TBR1
F8F8H
02F8H
XXH
Expanded Addr: F4F8H
ISA Addr:
Reset State:
03F8H
XXH
7
0
TB7
TB6
TB5
TB4
TB3
TB2
TB1
TB0
Bit
Number
Bit
Mnemonic
Function
7–0
TB7:0
Transmit Buffer Bits:
These bits make up the next data word to be transmitted. The transmitter
loads this word into the transmit shift register. The transmit shift register
then shifts the bits out, along with the asynchronous communication bits
(start, stop, and parity). The data bits are shifted out least-significant bit
(TB0) first.
NOTE: The transmit buffer register shares an address port with other SIO registers. You must clear
bit 7 (DLAB) of LCRn before you can write to the transmit buffer register.
Figure 11-13. Transmit Buffer Register (TBRn)
11-23
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
11.3.5 Receive Buffer Register (RBRn)
Read RBRn to obtain the last data word received. Use the interrupt control or DMA units or poll
the serial line status register (LSRn) to determine whether the receive buffer is full.
Receive Buffer
RBR0, RBR1
(read only)
RBR0
RBR1
F8F8H
02F8H
XXH
Expanded Addr: F4F8H
ISA Addr:
Reset State:
03F8H
XXH
7
0
RB7
RB6
RB5
RB4
RB3
RB2
RB1
RB0
Bit
Number
Bit
Mnemonic
Function
7–0
RB7:0
Receive Buffer Bits:
These bits make up the last word received. The receiver shifts bits in,
starting with the least-significant-bit. The receiver then strips off the
asynchronous bits (start, parity, and stop) and transfers the received
data bits from the receive shift register to the receive buffer.
NOTE: The receive buffer register shares an address port with other SIO registers. Bit 7 (DLAB) of
the LCRn must be cleared in order to read the receive buffer register.
Figure 11-14. Receive Buffer Register (RBRn)
11-24
ASYNCHRONOUS SERIAL I/O UNIT
11.3.6 Serial Line Control Register (LCRn)
Use LCRn to provide access to the multiplexed registers, send a break condition, and determine
the data frame for receptions and transmissions.
Serial Line Control
LCR0, LCR1
LCR0
LCR1
Expanded Addr: F4FBH F8FBH
(read/write)
ISA Addr:
Reset State:
03FBH 02FBH
00H 00H
7
0
DLAB
SB
SP
EPS
PEN
STB
WLS1
WLS0
Bit
Number
Bit
Mnemonic
Function
7
DLAB
Divisor Latch Access Bit:
This bit determines which of the multiplexed registers is accessed.
0 = Allows access to the receiver and transmit buffer registers (RBRn and
TBRn) and the interrupt enable register (IERn).
1 = Allows access to the divisor latch registers (DLLn and DLHn).
6
SB
Set Break:
0 = No effect on TXDn.
1 = Forces the TXDn pin to the spacing (logic 0) state for as long as bit is
set.
5
4
3
SP
Sticky Parity, Even Parity Select, and Parity Enable:
These bits determine whether the control logic produces (during
transmission) or checks for (during reception) even, odd, no, or forced
parity.
EPS
PEN
SP
X
0
0
1
EPS PEN Function
X
0
1
0
1
0
1
1
1
1
parity disabled (no parity option)
produce or check for odd parity
produce or check for even parity
produce or check for forced parity (parity bit = 1)
1
produce or check for forced parity (parity bit = 0)
2
STB
Stop Bits:
This bit specifies the number of stop bits transmitted and received in each
serial character.
0 = 1 stop bit
1 = 2 stop bits (1.5 stop bits for 5-bit characters)
1–0
WLS1:0
Word Length Select:
These bits specify the number of data bits in each transmitted or received
serial character.
00 = 5-bit character
01 = 6-bit character
10 = 7-bit character
11 = 8-bit character
Figure 11-15. Serial Line Control Register (LCRn)
11-25
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
11.3.7 Serial Line Status Register (LSRn)
Use LSRn to check the status of the transmitter and receiver.
Serial Line Status
LSR0
LSR1
LSR0, LSR1
(read only)
Expanded Addr: F4FDH F8FDH
ISA Addr:
03FDH 02FDH
60H 60H
Reset State:
7
0
—
TE
TBE
BI
FE
PE
OE
RBF
Bit
Number
Bit
Mnemonic
Function
7
6
—
Reserved. This bit is undefined.
Transmitter Empty:
TE
TBE
BI
The transmitter sets this bit to indicate that the transmit shift register and
transmit buffer register are both empty. Writing to the transmit buffer
register clears this bit.
5
4
Transmit Buffer Empty:
The transmitter sets this bit after it transfers data from the transmit buffer
to the transmit shift register. Writing to the transmit buffer register clears
this bit.
Break Interrupt:
The receiver sets this bit whenever the received data input is held in the
spacing (logic 0) state for longer than a full word transmission time.
Reading the receive buffer register or the serial line status register clears
this bit.
3
FE
Framing Error
The receiver sets this bit to indicate that the received character did not
have a valid stop bit. Reading the receive buffer register or the serial line
status register clears this bit. If data frame is set for two stop bits the
second stop bit is ignored.
2
1
PE
OE
Parity Error:
The receiver sets this bit to indicate that the received data character did
not have the correct parity. Reading the receive buffer register or the
serial line status register clears this bit.
Overrun Error:
The receiver sets this bit to indicate an overrun error. An overrun occurs
when the receiver transfers a received character to the receive buffer
register before the CPU reads the buffer’s old character. Reading the
serial line status register clears this bit.
0
RBF
Receive Buffer Full:
The receiver sets this bit after it transfers a received character from the
receive shift register to the receive buffer register. Reading the receive
buffer register clears this bit.
Figure 11-16. Serial Line Status Register (LSRn)
11-26
ASYNCHRONOUS SERIAL I/O UNIT
11.3.8 Interrupt Enable Register (IERn)
Use IERn to connect the SIOn status signals to the interrupt control unit. All four status signals
can be connected to the interrupt control unit.
Interrupt Enable
IER0, IER1
(read/write)
IER0
IER1
Expanded Addr: F4F9H
F8F9H
02F9H
00H
ISA Addr:
Reset State:
03F9H
00H
7
0
—
—
—
—
MS
RLS
TBE
RBF
Bit
Number
Bit
Mnemonic
Function
7–4
—
Reserved; for compatibility with future devices, write zeros to these bits.
Modem Status Interrupt Enable:
3
MS
0 = Modem input signal changes do not cause interrupts.
1 = Connects the modem status signal to the interrupt control unit’s
SIOINTn output. A change on one or more of the modem input
signals activates the modem status signal.
2
RLS
Receiver Line Status Interrupt Enable:
0 = LSR error conditions do not cause interrupts.
1 = Connects the receiver line status signal to the interrupt control unit’s
SIOINTn output. Sources for this interrupt include overrun error,
parity error, framing error, and break interrupt.
1
0
TBE
RBF
Transmit Buffer Empty Interrupt Enable:
0 = Transmit Buffer Empty signal does not cause interrupts.
1 = Connects the transmit buffer empty signal to the interrupt control
unit’s SIOINTn output.
Receive Buffer Full Interrupt Enable:
0 = Receive buffer full signal does not cause interrupts.
1 = Connects the receive buffer full signal to the interrupt control unit’s
SIOINTn output.
NOTE: The interrupt enable register is multiplexed with the divisor latch high register. You must clear
bit 7 (DLAB) of the serial line control register (LCRn) before you can access the interrupt
control register.
Figure 11-17. Interrupt Enable Register (IERn)
11-27
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
11.3.9 Interrupt ID Register (IIRn)
Use the IIRn to determine whether an interrupt is pending and, if so, which status signal generated
the interrupt request.
Interrupt ID
IIR0, IIR1
IIR0
IIR1
Expanded Addr: F4FAH F8FAH
(read only)
ISA Addr:
Reset State:
03FAH
01H
02FAH
01H
7
0
—
—
—
—
—
IS2
IS1
IP#
Bit
Number
Bit
Mnemonic
Function
7–3
2
—
Reserved. These bits are undefined.
Interrupt Source:
IS2:1
If an interrupt is pending (bit 0 = 0), these bits specify which status signal
caused the pending interrupt.
IS2
IS1
Interrupt Source
0
0
1
1
0
1
0
1
modem status signal*
transmitter buffer empty signal
receive buffer full signal
receiver line status signal**
* When one of the modem input signals (CTSn#, DSRn#, RIn#, and
DCDn#) changes state, the modem status signal is activated.
** A framing error, overrun error, parity error, or break interrupt activates
the receiver line status signal.
Reading the modem status register clears the modem status signal.
Reading the IIRn register or writing to the transmit buffer register clears
the transmit buffer empty signal. Reading the receive buffer register
clears the receive buffer full signal. Reading the receive buffer register or
the serial line status register clears the LSRn error bits, which clears the
receiver line status signal.
0
IP#
Interrupt Pending:
This bit indicates whether an interrupt is pending.
0 = Interrupt is pending
1 = No interrupt is pending
Figure 11-18. Interrupt ID Register (IIRn)
11-28
11.3.10 Modem Control Register (MCRn)
Use MCRn to put the SIOn into a diagnostic test mode. In this mode, the modem input signals
are disconnected from the package pins and controlled by the lower four MCRn bits and the mo-
dem output signals are forced to their inactive states (Figure 11-19). Additionally, the MCRn sig-
nals are also forced into the MSR register bits.
MCRn.1
CTS#
DSR#
DCD#
RI#
RTSn#
(forced high)
RTS#
DTR#
MCRn.0
MCRn.3
MCRn.2
DTRn#
(forced high)
Note: MCRn.1 indicates that modem control register bit 1 controls the CTS input, and so on.
A2529-01
Figure 11-19. Modem Control Signals – Diagnostic Mode Connections
Besides the diagnostic mode, there are two other options for connecting the modem input signals.
You can connect the signals internally using the SIO configuration (SIOCFG) register. The inter-
nal connection mode disconnects the modem input signals from the package pins and connects
the modem output signals to the modem input signals (in this case, the modem output signals re-
main connected to package pins). See Figure 11-20. In this mode, the values you write to MCRn
bits 0 and 1 control the state of the modem’s internal input signals and output pins.
CTS#
RTSn#
DTRn#
RTS#
DTR#
DSR#
DCD#
RI#
Vcc
A2528-01
Figure 11-20. Modem Control Signals – Internal Connections
The other option is standard mode. In standard mode, the modem input and output signals are
connected to the package pins. In this mode, the values you write to MCRn bits 0 and 1 control
the state of the modem’s output pins.
11-29
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Modem Control
MCR0, MCR1
(read/write)
MCR0
F4FCH
03FCH
00H
MCR1
F8FCH
02FCH
00H
Expanded Addr:
ISA Addr:
Reset State:
7
0
—
—
—
LOOP
OUT2
OUT1
RTS
DTR
Bit
Number
Bit
Mnemonic
Function
7–5
4
—
Reserved; for compatibility with future devices, write zeros to these bits.
LOOP
Loop Back Test Mode:
0 = Normal mode
1 = Setting this bit puts the SIOn into diagnostic (or loop back test) mode. This causes
the SIO channel to:
•
•
•
•
set its transmit serial output (TXDn)
disconnect its receive serial input (RXDn) from the package pin
loop back the transmitter shift register’s output to the receive shift register’s input
disconnect the modem control inputs (CTSn#, DSRn#, RIn#, and DCDn#) from the
package pins
•
•
force modem control outputs (RTSn# and DTRn#) to their inactive states
connects MCRn bits to MSRn bits
3–2
OUT2:1
Test Bits:
In diagnostic mode (bit 4=1), these bits control the ring indicator (RIn) and data carrier
detect (DCDn#) modem inputs. Setting OUT1 activates the internal RIn bit; clearing
OUT1 deactivates the internal RIn bit. Setting OUT2 activates the internal DCDn bit;
clear OUT2 deactivates the internal DCDn bit.
In normal user mode (bit 4=0) OUT1 has no effect and OUT2 in conjunction with
INTCFG.5/6 selects internal SIO interrupt or external interrupt. See Table 5-1 on page
5-8 for the configuration options.
1
RTS
Ready to Send:
The function of this bit depends on whether the SIOn is in diagnostic mode
(MCRn.4=1), internal connection mode, or standard mode.
In diagnostic mode, setting this bit activates the internal CTSn bit; clearing this bit
deactivates the internal CTSn bit.
In internal connection mode, setting this bit activates the internal CTSn# signal and the
RTSn# pin; clearing this bit deactivates the internal CTSn# signal and the RTSn# pin.
In standard mode, setting this bit activates the RTSn# pin; clearing this bit deactivates
the RTSn# pin. Note that pin is inverted from bit.
0
DTR
Data Terminal Ready:
The function of this bit depends on whether the SIOn is in diagnostic mode
(MCRn.4=1), internal connection mode, or standard mode.
In diagnostic mode, setting this bit activates the internal DSRn# signal; clearing this bit
deactivates the internal DSRn# signal.
In internal connection mode, setting this bit activates the internal DSRn# and DCDn#
signals and the DTRn# pin; clearing this bit deactivates the internal DSRn# and DCDn#
signals and the DTRn# pin. Note that pin is inverted from bit.
In standard mode, setting this bit activates the DTRn# pin; clearing this bit deactivates
the DTRn# pin. Note that pin is inverted from bit.
Figure 11-21. Modem Control Register (MCRn)
11-30
ASYNCHRONOUS SERIAL I/O UNIT
11.3.11 Modem Status Register (MSRn)
Read MSRn to determine the status of the modem control input signals. The upper four bits reflect
the current state of the modem input signals and the lower four bits indicate whether the inputs
(except for RI#) have changed state since the last time this register was read. These lower four
bits are reset to zero when the CPU reads the Modem Status register.
Modem Status
MSR0, MSR1
(read only)
MSR0
MSR1
Expanded Addr: F4FEH F8FEH
ISA Addr:
Reset State:
03FEH
X0H
02FEH
X0H
7
0
DCD
RI
DSR
CTS
DDCD
TERI
DDSR
DCTS
Bit
Number
Bit
Mnemonic
Function
7
DCD
Data Carrier Detect:
This bit is the complement of the data carrier detect (DCDn#) input. In
diagnostic test mode, this bit is equivalent to MCRn.3 (OUT2).
6
5
4
3
RI
Ring Indicator:
This bit is the complement of the ring indicator (RIn#) input. In diagnostic
test mode, this bit is equivalent to MCRn.2 (OUT1).
DSR
CTS
DDCD
Data Set Ready:
This bit is the complement of the data set ready (DSRn#) input. In
diagnostic test mode, this bit is equivalent to MCRn.0 (DTR).
Clear to Send:
This bit is the complement of the clear to send (CTSn#) input. In
diagnostic test mode, this bit is equivalent to MCRn.1 (RTS).
Delta Data Carrier Detect:
When set, this bit indicates that the DCDn# input has changed state
since the last time this register was read. Reading this register clears
this bit.
2
1
0
TERI
Trailing Edge Ring Indicator:
When set, this bit indicates that the RIn# input has changed from a low
to a high state since the last time this register was read. Reading this
register clears this bit.
DDSR
DCTS
Delta Data Set Ready:
When set, this bit indicates that the DSRn# input has changed state
since the last time this register was read. Reading this register clears
this bit.
Delta Clear to Send:
When set, this bit indicates that the CTSn# input has changed state
since the last time this register was read. Reading this register clears
this bit.
Figure 11-22. Modem Status Register (MSRn)
11-31
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
11.3.12 Scratch Pad Register (SCRn)
SCRn is available for use as a scratch pad. Writing and reading this register has no effect on SIOn
operation.
Scratch Pad
SCR0, SCR1
(read/write)
SCR0
SCR1
F8FFH
02FFH
XXH
Expanded Addr: F4FFH
ISA Addr:
Reset State:
03FFH
XXH
7
0
SP7
SP6
SP5
SP4
SP3
SP2
SP1
SP0
Bit
Number
Bit
Mnemonic
Function
7–0
SP7:0
Writing and reading this register has no effect on SIOn operation.
Figure 11-23. Scratch Pad Register (SCRn)
11.4 PROGRAMMING CONSIDERATIONS
Consider the following when programming the SIO.
• The divisor latch low register (DLLn) is multiplexed with the receive and transmit buffer
registers (RBRn and TBRn) and the divisor latch high register (DLHn) is multiplexed with
the interrupt enable register (IERn). Bit 7 of the serial line control register (LCRn) controls
which register is accessed.
• The SIO contains four status signals: receiver line status, receive buffer full, transmit buffer
empty, and modem status. You can connect (OR) these signals to the interrupt control unit’s
SIOINTn interrupt request signal using the interrupt enable register (IERn). If you receive
an interrupt request on the SIOINTn signal, read the interrupt ID register (IIRn) to
determine which status signal with the highest priority caused the request.
Several sources can activate the receiver line status and the modem status signals. If IIRn
indicates that the receiver line status signal caused an interrupt request, read the serial line
status register (LSRn) to determine the receive error condition that activated the receiver
line status signal. If IIRn indicates that the modem status signal caused an interrupt request,
read the modem status register (MSRn) to determine which modem input signal activated
the modem status signal.
• DMA can be used for servicing the SIO channels for higher baud rates. When doing this,
remember that the isolated RBF and TBE (RBFDMA and TBEDMA) signals are connected
to the DMA DREQ inputs. RBFDMA is blocked if any of the error bits in the LSR are set.
Neither signal is gated by the IERn register.
11-32
ASYNCHRONOUS SERIAL I/O UNIT
11.4.1 Asynchronous Serial I/O Unit Code Examples
The code example contains these software routines:
InitSIO
Initializes the SIO for asynchronous transfers
SerialReadStr
Polled serial read function that reads a specified number of
characters
SerialReadChar
SerialWriteChar
SerialWriteStr
Polled serial read function that reads a single character
Polled serial write function that writes a single character
Polled serial write function that writes out an entire string of
characters
SerialWriteMem
Polled serial write function that writes out a specified number of
characters stored in a buffer
Serial0_ISR
Service_RBF
Template interrupt service routine for SIO_0 interrupts
Service routine for interrupts generated by the Receive Buffer Full
signal
SerialWriteStr_Int
Service_TBE
Interrupt driven serial write function
Service routine for interrupts generated by the Transmit Buffer
Empty signal
The last software routine shows how to use these functions to enable RBF interrupts on the SIO.
See Appendix C for the included header files.
#include <conio.h>
#include <stdio.h>
#include “80386ex.h”
#include “ev386ex.h”
/* Variable Declarations */
int Tbuffer_index = 0;
char trans_buffer[1024];
char rec_buffer;
/*****************************************************************************
InitSIO:
Description:
Initialization routine for Asynchronous Serial I/O Port.
Parameters:
Unit
Unit number of the serial port. 0 for SIO port 0,
1 for SIO port 1.
Mode
Defines Parity, number of data bits, number of stop bits--
Reference Serial Line Control register for various
11-33
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
options
ModemCntrl
BaudRate
Defines the operation of the modem control lines
Specifies baud rate. The baud divisor value is calculated
based on clocking source and clock frequency. The
clocking frequency is set by calling the
InitializeLibrary function.
ClockRate
Specifies the serial port clocking rate, for internal
clocking = CLK2, for external = COMCLK
Returns:Error Codes
E_INVAILD_DEVICE -- Unit number specifies a non-existing device
E_OK
-- Initialized OK, No error.
Assumptions:
SIOCFG Has already been configured for Clocking source and Modem
control source
REMAPCFG register has Expanded I/O space access enabled (ESE bit set).
The processor Port pin are initialized separately.
Syntax:
#define SIO_0
#define SIO_8N1
0x0
(SIO_8DATA | SIO_1STOPBIT | SIO_NOPARITY)
#define SIO_MCR_RTS 0x2
#define SIO_MCR_DTR 0x1
#define SIO_8DATA
0x3
#define SIO_1STOPBIT 0x0
//Clock rate of COMCLK, i.e., External clocking
#define BAUD_CLKIN
int error;
1843200L
error = InitSIO(SIO_0,
SIO_8N1,
// Which Serial Port
// Mode, 8-data, no parity, 1-stop
SIO_MCR_RTS+SIO_MCR_DTR, // Modem line controls
9600,
// Baud Rate
BAUD_CLKIN);
// Baud Clocking Rate
Real/Protected Mode:
No changes required.
*****************************************************************************/
int InitSIO(int Unit, BYTE Mode, BYTE ModemCntrl, DWORD BaudRate,
DWORD BaudClkIn)
{
WORD SIOPortBase;
WORD BaudDivisor;
/* Check for valid unit */
if(Unit > 1)
11-34
ASYNCHRONOUS SERIAL I/O UNIT
return E_INVALID_DEVICE;
/* Set Port base based on serial port used */
SIOPortBase = (Unit ? SIO1_BASE : SIO0_BASE);
/* Initialized Serial Port registers */
/* Calculate the baud divisor value, based on baud clocking */
BaudDivisor = (WORD)(BaudClkIn / (16*BaudRate));
/* Turn on access to baud divisor register */
_SetEXRegByte(SIOPortBase + LCR, 0x80);
/* Set the baud rate divisor register, High byte first */
_SetEXRegByte(SIOPortBase + DLH, HIBYTE(BaudDivisor) );
_SetEXRegByte(SIOPortBase + DLL, LOBYTE(BaudDivisor) );
/*** Set Serial Line control register ***/
_SetEXRegByte(SIOPortBase + LCR, Mode);// Sets Mode and
//reset the Divisor latch
/* Set modem control bits */
_SetEXRegByte(SIOPortBase + MCR, ModemCntrl);
return E_OK;
}/* InitSIO */
/*****************************************************************************
SerialReadStr
Description:
Is a Polled serial port read function that waits forever or until
count characters are read from the serial port.
Parameters:
Unit
Unit number of the serial port. 0 for SIO port 0,
1 for SIO port 1.
str
count
Address of where to place the input data
Number of characters to read before returning.
Returns: Error Code
E_OK or Error code status (value of Line Status Register (LSR)
Assumptions:
REMAPCFG register has Expanded I/O space access enabled (ESE bit set).
The processor Port pin are initialized separately.
Syntax:
11-35
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
#define SIO_0 0
#define LENGTH 32
char String_Read[LENGTH];
int error;
error = SerialReadStr (SIO_0,
String_Read,
LENGTH);
Real/Protected Mode
No changes required.
*****************************************************************************/
int SerialReadStr(int Unit, char far *str, int count)
{
WORD ReceivePortAddr;
WORD StatusPortAddr;
BYTE Status;
int i;
/* Set Port base, based on serial port used */
ReceivePortAddr = (Unit ? RBR1 : RBR0);
StatusPortAddr = (Unit ? LSR1 : LSR0);
for(i=0; i < count-1; i++)
{
// Status register is cleared after read, so we must save
// it’s value when read
while(!((Status=_GetEXRegByte(StatusPortAddr)) & SIO_RX_BUF_FULL))
if( Status & SIO_ERROR_BITS ) /* Error Bit set then return NULL */
{
str[i+1] = ‘\0’;
return Status & SIO_ERROR_BITS;
}
str[i] = _GetEXRegByte(ReceivePortAddr);
}
str[i] = ‘\0’;
return E_OK;
}/* SerialReadStr */
/*****************************************************************************
SerialReadChar:
Description:
Is a Polled serial port read function that waits forever or
11-36
ASYNCHRONOUS SERIAL I/O UNIT
until a character has been received from the serial port.
Parameters:
Unit Unit number of the serial port. 0 for SIO port 0,
1 for SIO port 1.
Returns:
BYTE Read from serial port, if zero an error occurred.
Assumptions:
REMAPCFG register has Expanded I/O space access enabled (ESE bit set).
The processor Port pin are initialized separately.
Syntax:
#define SIO_0 0
BYTE character;
character = SerialReadChar (SIO_0);
Real/Protected Mode
No changes required.
*****************************************************************************/
BYTE SerialReadChar(int Unit)
{
WORD ReceivePortAddr;
WORD StatusPortAddr;
WORD Status;
/* Set Port base, based on serial port used */
ReceivePortAddr = (Unit ? RBR1 : RBR0);
StatusPortAddr = (Unit ? LSR1 : LSR0);
// Status register is cleared after read, so we must save
// it’s value when read
while(!((Status=_GetEXRegByte(StatusPortAddr)) & SIO_RX_BUF_FULL))
if( Status & SIO_ERROR_BITS ) // Error Bit set then return NULL
{
return 0;
}
return _GetEXRegByte(ReceivePortAddr);
}/* SerialReadChar */
/*****************************************************************************
11-37
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
SerialWriteChar:
Description:
Is a Polled serial port write function that waits forever or
until a character has been written to the serial port.
Parameters:
Unit
Unit number of the serial port. 0 for SIO port 0,
1 for SIO port 1.
ch
Character value to be written out
Returns:
None
Assumptions:
REMAPCFG register has Expanded I/O space access enabled (ESE bit set).
The processor Port pin are initialized separately.
Syntax:
#define SIO_0 0
char Char_Out = ‘a’;
SerialWriteChar (SIO_0, Char_Out);
Real/Protected Mode
No changes required.
*****************************************************************************/
void SerialWriteChar(int Unit, BYTE ch)
{
WORD TransmitPortAddr;
WORD StatusPortAddr;
/* Set Port base, based on serial port used */
TransmitPortAddr = (Unit ? TBR1 : TBR0);
StatusPortAddr = (Unit ? LSR1 : LSR0);
/* Wait until buffer is empty */
while(!(_GetEXRegByte(StatusPortAddr) & SIO_TX_BUF_EMPTY)) ;
_SetEXRegByte(TransmitPortAddr,ch);
}/* SerialWriteChar */
/*****************************************************************************
SerialWriteStr:
11-38
ASYNCHRONOUS SERIAL I/O UNIT
Description:
Is a Polled serial port write function that waits forever
or until all characters have been written to the serial port.
The NUL character (‘\0’) is used to indicate end of string.
Parameters:
Unit
Unit number of the serial port. 0 for SIO port 0,
1 for SIO port 1.
str
Address of a zero terminated string to be transmitted
Returns:
None
Assumptions:
REMAPCFG register has Expanded I/O space access enabled (ESE bit set).
The processor Port pin are initialized separately.
Syntax:
#define SIO_0 0
SerialWriteStr (SIO_0,
HelloString);
Real/Protected Mode
No changes required.
*****************************************************************************/
void SerialWriteStr(int Unit, const char far *str)
{
WORD TransmitPortAddr;
WORD StatusPortAddr;
/* Set Port base, based on serial port used */
TransmitPortAddr = (Unit ? TBR1 : TBR0);
StatusPortAddr = (Unit ? LSR1 : LSR0);
for( ; *str != ‘\0’; str++)
{
/* Wait until buffer is empty */
while(!(_GetEXRegByte(StatusPortAddr) & SIO_TX_BUF_EMPTY)) ;
/* Write Character */
_SetEXRegByte(TransmitPortAddr,*str);
}
}/* SerialWriteStr */
11-39
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
/*****************************************************************************
SerialWriteMem:
Description:
Is a Polled serial port write function that waits forever or
until count characters have been written to the serial port.
Parameters:
Unit
Unit number of the serial port. 0 for SIO port 0,
1 for SIO port 1.
mem
Address of a buffer to be transmitted
count Number of characters in buffer to be transmitted
Returns:
None
Assumptions:
REMAPCFG register has Expanded I/O space access enabled (ESE bit set).
The processor Port pin are initialized separately.
Syntax:
#define SIO_0 0
#define COUNT 32
char Buffer[COUNT];
SerialWriteMem
(SIO_0,
Buffer,
COUNT);
Real/Protected Mode
No changes required.
*****************************************************************************/
void SerialWriteMem(int Unit, const char far *mem, int count)
{
WORD TransmitPortAddr;
WORD StatusPortAddr;
int i;
/* Set Port base, based on serial port used */
TransmitPortAddr = (Unit ? TBR1 : TBR0);
StatusPortAddr = (Unit ? LSR1 : LSR0);
for(i=0 ; i < count; i++)
{
/* Wait until buffer is empty */
while(!(_GetEXRegByte(StatusPortAddr) & SIO_TX_BUF_EMPTY)) ;
/* Write Character */
_SetEXRegByte(TransmitPortAddr,mem[i]);
11-40
ASYNCHRONOUS SERIAL I/O UNIT
}
} /* SerialWriteMem */
/*****************************************************************************
Serial0_ISR:
Description:
Template Interrupt Service Routine for Serial Port0 interrupts.
This function identifies the cause of the interrupt and branches
to the corresponding action.
Parameters:
None (Not called by user)
Returns:
None
Assumptions:
None
Syntax:
Not a user function.
Real/Protected Mode:
No changes required.
******************************************************************************/
void interrupt far Serial0_ISR (void)
{
BYTE iir0, lsr0, msr0;
iir0 = _GetEXRegByte(IIR0);
switch ((iir0&0x06) >> 1) {
case 0:
/* modem status signal */
msr0 = _GetEXRegByte(MSR0);
if ((msr0&0x08) && (msr0&0x80)){
/* data carrier detect has been set */
}
11-41
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
if ((msr0 & 0x04) && (msr0 & 0x40)) {
/* ring indicator */
}
if ((msr0 & 0x02) && (msr0 & 0x20)) {
/* data set ready bit has been set */
}
if ((msr0 & 0x01) && (msr0&0x10)) {
/* clear to send signal has been set */
}
break;
case 1:
Service_TBE(); /* Routine for Interrupt driven Serial Writes */
break;
case 2:
/* RBF signal */
Service_RBF(); /* Routine specific to RBF generated interrupts */
break;
case 3:
/* receive line status signal */
lsr0 = _GetEXRegByte(LSR0);
if (lsr0 & 0x10) {
/* break interrupt */
}
if (lsr0 & 0x08) {
/* framing error */
}
if (lsr0 & 0x04) {
/* parity error */
}
if (lsr0 & 0x02) {
/* overrun error */
}
break;
} /* End of switch */
NonSpecificEOI(); // Send End-Of-Interrupt Signal to Master
}/* Serial0_ISR */
11-42
ASYNCHRONOUS SERIAL I/O UNIT
/*****************************************************************************
Service_RBF:
Description:
Service routine for interrupts generated by RBF signal. This
routine is used for Interrupt-Driven Serial Reads. It echoes
the typed character to the screen, stopping when it receives
an ESC character.
Parameters:
None
Returns:
None
Assumptions:
None
Syntax:
Not called by user
Real/Protected Mode:
No changes required.
*****************************************************************************/
void Service_RBF (void)
{
/* Read in contents of RBR0 */
rec_buffer = _GetEXRegByte(RBR0);
SerialWriteChar(SIO_0, rec_buffer); // Echo to screen
if ( rec_buffer == 0x1b ) {
/* ESC character received, disable RBF interrupts*/
_SetEXRegByte(IER0, 0x00);
}
}/* Service_RBF */
/*****************************************************************************
SerialWriteStr_Int:
11-43
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Description:
Is an interrupt driven serial port write function.
The NUL character (‘\0’) is used to indicate end of string.
Parameters:
Unit
Unit number of the serial port. 0 for SIO port 0,
1 for SIO port 1.
str
Address of a zero terminated string to be transmitted
Returns:
None
Assumptions:
REMAPCFG register has Expanded I/O space access enabled (ESE bit set).
The processor Port pin are initialized separately.
Syntax:
#define SIO_0 0
SerialWriteStr_Int (SIO_0, HelloString);
Real/Protected Mode
No changes required.
*****************************************************************************/
void SerialWriteStr_Int(int Unit, const char far *str)
{
BYTE PortIntEnable;
PortIntEnable = (Unit ? IER1 : IER0);
strcpy (trans_buffer, str); // Copy string into buffer
/* Enable TBE interrupts */
_SetEXRegByte(IER0,0x02);
}/* SerialWriteStr_Int */
/*****************************************************************************
Service_TBE:
Description:
Service routine for TBE generated interrupts. This function is used
for Interrupt-Driven Serial Transmits.
Parameters:
11-44
ASYNCHRONOUS SERIAL I/O UNIT
None
Assumptions:
None
Syntax:
Not called by user.
Real/Protected Mode:
No changes required.
******************************************************************************/
void Service_TBE(void)
{
if (trans_buffer[Tbuffer_index] != ‘\0’) {
_SetEXRegByte(TBR0, trans_buffer[Tbuffer_index]);
Tbuffer_index++;
}
else {
/* Disable TBE interrupts */
_SetEXRegByte(IER0, 0x00);
}
}/* Service_TBE */
/****************************************************************************
Example code to show how to set up for a Serial Port interrupt.
This example is for an interrupt on SIO_0 sourced by the
Receive Buffer Full Signal. The source code for the functions “SetIRQVector”
and “Disable8259Interrupt” is included in the Interrupt Control Unit chapter.
SetIRQVector(Serial0_ISR, 4, INTERRUPT_ISR); // Set vector for Interrupt
// on Master line 4
Disable8259Interrupt(IR1+IR5+IR6+IR7, IR0+IR1+IR2+IR3+IR4+IR5+IR6+IR7);
Enable8259Interrupt(IR2+IR4,0);// Enable slave interrupt to master(IR2),
// Enable SIO_0 (IR4)
_enable(); // Enable Interrupts
_SetEXRegByte(IER0, 0x01);
// Enable interrupt on RBF signal
/****************************************************************************/
11-45
12
DMA
CONTROLLER
CHAPTER 12
DMA CONTROLLER
The DMA controller improves system performance by allowing external or internal peripherals
bits). It contains two identical channels. The DMA controller has features that are unavailable on
This chapter is organized as follows:
• DMA Operation (page 12-5)
• Register Definitions (page 12-28)
• Design Considerations (page 12-50)
• Programming Considerations (page 12-50)
12.1 OVERVIEW
Figure 12-1 shows a block diagram of the DMA unit. The DMA channels are independently con-
figurable. Each channel contains a request input (DREQn) and an acknowledge output
(DMAACKn#). An external peripheral (connected to the DRQn pin) or one of the internal pe-
ripherals (asynchronous serial I/O, synchronous serial I/O, or timer control unit) can request
DMA service. The DMA configuration register is used to select one of the possible sources. In
addition to these hardware request sources, each channel contains a software request register that
can be used to initiate software requests. The channels share an end-of-process signal (EOP#).
This signal functions as either an input or an open-drain output. EOP# either terminates a transfer
(as an input) or signals that a transfer is completed (as an output).
12-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
DMACFG.2:0
DMA
DRQ0
(DCD1#)†
3
0
RBFDMA0 (SIO0)
TXEDMA1 (SIO1)
SSTBE (SSIO)
OUT1 (TCU)
RBFDMA1 (SIO1)
TXEDMA0 (SIO0)
1
2
3
4
5
6
7
To SIO1
DREQ0
SSRBF (SSIO)
DMACFG.3
PINCFG.4
0
DACK0#
(CS5#)
DMAACK0#
DREQ1
From CSU
1
DMACFG.6:4
3
0
DRQ1
(RXD1)
RBFDMA1 (SIO1 )
TXEDMA0 (SIO0)
SSRBF (SSIO)
OUT2 (TCU)
1
2
3
4
5
6
7
To SIO1
RBFDMA0 (SIO0)
TXEDMA1 (SIO1)
SSTBE (SSIO)
DMACFG.7
PINCFG.2
0
DACK1#
(TXD1)
DMAACK1#
From SIO1
1
DMAINT
To ICU
PINCFG.3
0
1
End of Process
EOP#
(CTS1#)
From SIO1
1
P1CFG.6
HOLD
Bus Arbiter
HLDA
HOLD
(P1.6)
To/From I/O Port 1
0
P1CFG.7
1
HLDA
(P1.7)
To/From I/O Port 1
0
† Alternate pin signals are in parentheses.
A2531-02
Figure 12-1. DMA Unit Block Diagram
12-2
DMA CONTROLLER
12.1.1 DMA Terminology
This section provides a definition of some of the terms used in this chapter to describe the DMA
controller.
DMA Process
A DMA process is the execution of a programmed DMA task from
beginning to end. Each DMA process requires initial programming
by the Intel386 EX processor.
Buffer
A contiguous block of data.
Buffer Transfer
Data Transfer
The action required by the DMA to transfer an entire buffer.
The DMA action in which a group of bytes or words are moved
between devices by the DMA controller. A data transfer operation
may involve movement of one or many bytes.
Bus Cycle
Requester
Access by the DMA to a single byte or word.
The Requester is the device which requests service by the DMA
controller. All of the control signals which the DMA monitors or
generates for specific channels are logically related to the requester.
Only the requester is considered capable of initiating or terminating a
DMA process. The requester may be either I/O or memory and may
be the Source or the Destination of the transfer or neither.
Target
The Target is the device with which the Requester wishes to
communicate. As far as the DMA process is concerned, the Target is
a slave which is incapable of control over the process. The Target
may be either I/O or memory, and may be either the Source or the
Destination of the transfer.
Source
The Source is the memory or I/O from which data is being read.
The Destination is the memory or I/O to which data is being written.
Destination
12-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
12.1.2 DMA Signals
Table 12-1 describes the DMA signals.
Table 12-1. DMA Signals
Device Pin
Signal
or Internal
Signal
Description
DRQ0
Device pin
(input)
DMA Channel 0 Requests:
The SIO channel 0 receiver, SIO channel 0 transmitter,
SIO channel 1 receiver, SIO channel 1 transmitter, SSIO
transmitter, SSIO receiver, TCU counter 1 output, or an
external device can request DMA channel 0 service.
These sources are referred to as channel 0 hardware
requests. You can also issue channel 0 software
SIO0 RBFDMA0/TXEDMA0
SIO1 TXEDMA1/RBFDMA1
SSIO Transmitter/Receiver
TCU Counter 1
Internal
signals
requests by writing to the DMA software request register.
DRQ1
Device pin
(input)
DMA Channel 1 Requests:
The SIO channel 1 receiver, SIO channel 1 transmitter,
SIO channel 0 transmitter, SIO channel 0 receiver, SSIO
receiver, SSIO transmitter, TCU counter 2 output, or an
external device can request DMA channel 1 service.
These sources are referred to as channel 1 hardware
requests. You can also issue channel 1 software
SIO1 RBFDMA1/TXEDMA1
SIO0 TXEDMA0/RBFDMA0
SSIO Receiver/Transmitter
TCU Counter 2
Internal
signals
requests by writing to the DMA software request register.
DACKn#
EOP#
Device pin
(output)
DMA Channel n Acknowledge:
Indicates that channel n is ready to service the
requesting device. An external device uses the DRQn
pin to request DMA service; the DMA uses the DACKn#
pin to indicate that the request is being serviced.
Device pin
End-of-process:
(input/open-
drain output)
As an input:
Activating this signal terminates a DMA transfer.
As an output:
This signal is activated when a DMA transfer completes.
12-4
DMA CONTROLLER
12.2 DMA OPERATION
The following sections describe the operation of the DMA. See “Register Definitions” on page
12-28 for details on implementing DMA Controller options.
12.2.1 DMA Transfers
The DMA transfers data between a requester and a target. The data can be transferred from the
requester to target or vice versa. The target addresses and requester addresses can be located in
either memory or I/O space, and data transfers can be on a byte or word basis. The requester can
be in external device I/O space, in internal peripheral I/O space, or memory mapped I/O. (Very
simply, the requester is the thing that activated DREQn.) An external device or an internal periph-
eral requests service by activating a channel’s request input (DREQn). A requester in memory re-
quests service through the DMA software request register. The requester either deposits data to
or fetches data from the target.
A channel is programmed by writing to a set of requester address, target address, byte count, and
control registers. The address registers specify base addresses for the target and requester, and the
byte count registers specify the number of bytes that need to be transferred to or from the target.
Typically, a channel is programmed to transfer a block of data. Therefore, it is necessary to dis-
tinguish between the process of transferring one byte or word (data transfer) and the process of
transferring the entire block of data (buffer transfer).
The byte count determines the number of data transfers that make up a buffer transfer. After each
data transfer within a buffer transfer, the byte count is decremented (by 1 for byte transfers and
by 2 for word transfers) and the requester and target addresses are either incremented, decrement-
ed, or left unchanged. When the byte count expires (reaches FFFFFFH), the buffer transfer is
complete. If the channel’s end-of-process (EOP#) signal is activated before the byte count ex-
pires, the buffer transfer is terminated.
NOTE
Since the buffer transfer is complete when the byte count reaches FFFFFFH,
the number of bytes transferred is the byte count + 1.
12.2.2 Bus Cycle Options for Data Transfers
There are two bus cycle options for transferring data, fly-by and two-cycle. Fly-by allows data to
be transferred in one bus cycle. It requires that the requester be in external I/O and the target be
in memory. The two-cycle option allows data to be transferred between any combination of mem-
ory and I/O through the use of a four-byte temporary buffer.
12.2.2.1
Fly-By Mode
The fly-by option performs either a memory write or a memory read bus cycle. A write cycle
transfers data from the requester to the target (memory), and a read cycle transfers data from the
target (memory) to the requester. When a data transfer is initiated, the DMA places the memory
address of the target on the bus and selects the requester by asserting the DACKn# signal. The
requester then either deposits the transfer data on the data bus or fetches the transfer data off the
12-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
data bus, depending on the transfer direction. Since the requester is selected via the DACKn# sig-
nal the requester address is not meaningful in a fly-by mode transfer.
Support logic (either external or built in to the I/O device) must be designed to monitor the
DACKn# signal and accordingly generate the correct control signals to the I/O device, since all
processor signals are used to access memory. This means that if it is an I/O to memory transfer,
this logic generates an I/O read cycle and the processor generates the memory write cycle. If it is
a memory to I/O transfer, the logic generates an I/O write cycle and the processor generates the
memory read cycle. This way the data is driven by the I/O device and latched by the memory de-
vice during an I/O to memory transfer, and driven by the memory device and latched by the I/O
device during a memory to I/O transfer.
12.2.2.2
Two-Cycle Mode
The two-cycle option first fills the four-byte temporary buffer with data from the source, then
writes that data to the destination. This method allows transfers between any combination of
memory and I/O with any combination of data path widths (8- or 16-bit). The amount of data and
the data bus widths determine the number of bus cycles required to transfer data. For example, it
takes six bus cycles to transfer four bytes of data from an 8-bit source to a 16-bit destination: four
read cycles to fill the temporary buffer from the 8-bit source, and two write cycles to transfer the
data to the 16-bit destination.
A buffer transfer can complete, be terminated, or be suspended before the temporary buffer is
filled from the source. If the buffer transfer completes or is terminated before the temporary buff-
er is filled, the DMA writes the partial data to the destination. When a requester suspends a buffer
transfer, the contents of the partially filled temporary buffer are stored until the transfer is restart-
ed. At this point, the DMA performs read cycles until the buffer is full, then performs write cycles
to transfer the data to the destination.
12.2.2.3
Programmable DMA Transfer Direction
The relationship between Requester, Target, Source, and Destination is determined by the pro-
grammable DMA transfer direction. The transfer directions are defined as Write, Read, or Verify.
The following table describes which operations are being performed by the Requester and Target
for each transfer direction. In this table, the device being read is the Source, and the device being
written is the Destination. The Verify cycle is used to perform a data read only. No write cycle is
indicated or assumed in a Verify cycle. The Verify cycle is useful for validating block fill opera-
tions. An external comparator must be provided to do any comparisons on the data read.
Table 12-2. Operations Performed During Transfer
Read
Write
Verify
Requester
Target
Read
Write
Write
Read
Read
Read
A special case not indicated in this table is when the Requester is neither the Source nor Destina-
tion. One example of this case would be when the DMA is being used to transfer data from one
memory or I/O location to another, and one of the timer outputs is being used to initiate that trans-
fer. In this case, the timer output would be selected as the DMA request source (using the
12-6
DMA CONTROLLER
DMACFG register), but the Requester address registers would be programmed with one of the
memory addresses. It doesn’t really matter which memory is the Requester and which is the Tar-
get, as long as the transfer direction is set to provide the correct Source and Destination.
12.2.2.4
Ready Generation For DMA Cycles
DMA cycles are identical to any other type of memory or I/O cycles in terms of how they are
completed. A valid READY# must be sampled at the end of the last T2 state in order to complete
a DMA Read or Write cycle. This READY# may be generated externally, or internally using the
appropriate chip select unit (see Chapter 14, “CHIP-SELECT UNIT” for a description of gener-
ating READY# internally).
12.2.2.5
DMA Usage of the 4-Byte Temporary Register
Each DMA channel has a 4-byte temporary FIFO register used for temporary data storage during
two cycle transfers. The way the DMA channel fills and empties this register depends on the data
transfer mode, the bus sizes of the source and destination, and the data transfer direction. The fol-
lowing describes how the Temporary Register is filled and emptied for the Read and Write Trans-
fer Directions.
Filling the Temporary Register:
Read Cycle In a Read Cycle data is transferred from the Requester to the Target. Each request
(DREQn) in a Read Cycle results in the DMA transferring a byte (if requester is
an 8-bit device) or a word (if the requester is a 16-bit device) from the Source
(Requester) to the temporary register. This continues until either the Temporary
Register is full, or until the byte count or terminal count is reached.
Write Cycle In a Write Cycle data is transferred from the Target to the Requester. The first
request (DREQn) initiates a fill of the temporary register (four byte reads of the
Target if the Target is 8-bit, or two word reads if it is 16-bit). The buffer is
considered full if either four bytes have been stored, or if less than four bytes, the
byte count or terminal count has been reached.
Emptying the Temporary Register:
Read Cycle Once the Temporary Register has been filled the DMA empties it by doing four
byte write cycles (if Target is 8-bit), or two word write cycles (if Target is 16-bit).
This is done in a burst-type fashion since all four requests have already occurred.
The byte counter is decremented after each write has occurred.
Write Cycle Once the Temporary Register has been filled the DMA does a single write cycle
transferring the first byte (if Requester is 8-bit), or the first word (if Requester is
16-bit). This first write cycle happens immediately after the buffer has been
filled. Each subsequent request (DREQn) results in another write cycle
transferring another byte or word from the Temporary Register to the Requester.
This continues until either the Temporary Register is empty, or byte count or
terminal count has been reached.
12-7
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Figures 12-2 and 12-3 are simple diagrams of how the Temporary Register is filled and emptied
for a Read DMA cycle and a Write DMA cycle.
Filling the Temporary Register
Emptying the Temporary Register
DREQn
#1
DREQn
#2
DREQn
#3
DREQn
#4
Four separate requests each with a read
of the requester. Each byte is stored in
the Temporary Register.
Write
#1
Write
#1
Write
#1
Write
#1
Once the Temporary Resister is full, the
DMA does four burst writes to the target
to empty it.
A3381-01
Figure 12-2. DMA Temporary Buffer Operation for a Read Transfer
Filling the Temporary Register
Emptying the Temporary Register
DREQn
#1
A single request with four separate reads of
the target. Each read stores a byte in the
Temporary Register.
DREQn
#2
DREQn
#3
DREQn
#4
Once the Temporary Resister is full, the
DMA does a write cycle to transfer the first
byte from the Temporary Register to the
target. On each subsequent request, the
DMA performs a write cycle transferring a
byte from the Temporary Register to the
target. This continues until empty.
A3382-01
Figure 12-3. DMA Temporary Buffer Operation for A Write Transfer
12-8
DMA CONTROLLER
12.2.3 Starting DMA Transfers
Internal I/O, external I/O, or memory can request DMA service. The internal I/O requesters (the
to the DMA request inputs. You must connect an external I/O source to the DMA DRQn; when
you are using fly-by mode, you must also connect an external I/O source to the DACKn# signals.
In addition, memory mapped I/O peripherals may use DRQn/DACKn#. DACKn# is active during
the entire fly-by mode transfer, but during a two-cycle mode transfer it is only active during the
access to the requester. These sources make up the DMA hardware request sources. The DMA
unit also contains a software request register that allows you to generate software DMA requests.
This allows memory-to-memory transfers. Figure 12-4 shows the timing for the start of a DMA
transfer.
Tx
Tx
Ti
Ti
T1
T2
T1
CLKOUT
DRQn
A25:1
BHE#, BLE#
M/IO#
DACKn#
ADS#
READY#
x Cycle
Transition to DMA
Cycle
DMA Cycle
A2480-02
Figure 12-4. Start of a Two-cycle DMA Transfer Initiated by DRQn
12.2.4 Bus Control Arbitration
The bus arbiter services bus control requests from the two DMA channels, an external device, and
the refresh control unit. The DMA channels interface with the bus arbiter through its DMA chan-
nel request signals (DREQn) and its DMA channel acknowledge signals (DMAACKn#). Other
external bus masters interface with the bus arbiter through similar request and acknowledge sig-
nals, the HOLD and HOLDA signals respectively. The refresh control unit gains bus control
through an internal Refresh request. The REFRESH# status pin indicates that the Refresh Control
Unit has gained bus control and that a valid refresh cycle is being executed. After receiving a bus
12-9
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
control request, the bus arbiter services these requests by issuing an internal hold signal request-
ing control of the bus from the core. The core returns an internal hold acknowledge signal to the
arbiter when bus ownership is granted. The arbiter then issues an acknowledge signal to the re-
questing device.
Refresh requests always have the highest priority, while the priority structure of the other three
requests is configurable. By default, DMA channel 0 requests have the next highest priority, fol-
changing the priority of the DMA and external bus requests, low-priority selection or rotation.
The priority requests are programmed in the DMACMD2 register (see Figure 12-24). The low-
priority selection method allows you to assign a particular request to the lowest priority level.
With the rotation method, a request is automatically assigned to the lowest priority level after it
gains bus control. The rotation method allows requesting devices to share the system bus more
evenly. With both methods, the other request priority levels are adjusted in a circular manner (see
Figure 12-5).
Low-priority
Select
Default
Rotating
Becomes
Highest
Level
Becomes
Highest
Level
Highest
Level
External Bus
Master
DMA
Channel 1
DMA
Channel 0
DMA
Channel 0
DMA
Channel 1
External Bus
Master
DMA
Channel 1
Specified
Lowest
Level
Assigned
Lowest
Level
DMA
Channel 0
Lowest
Level
External Bus
Master
After Gaining
Bus Control
A2532-01
Figure 12-5. Changing the Priority of the DMA Channel and External Bus Requests
12.2.5 Ending DMA Transfers
When a channel’s byte count expires, the buffer transfer is complete and the end-of-process
count expires by activating the EOP# input. The channel can sample the EOP# input synchro-
nously or asynchronously. With synchronous sampling, the channel samples EOP# at the end of
the last state of every data transfer. With asynchronous sampling, the DMA samples the inputs at
the beginning of every state of requester access, then waits until the end of the state to act on the
input. Figure 12-7 illustrates terminating a buffer transfer by activating the EOP# input; the figure
shows both asynchronous and synchronous EOP# sampling. EOP# sampling is programmed in
the DMACMD2 register (Figure 12-24).
12-10
DMA CONTROLLER
Terminating a buffer transfer by deasserting DREQn can also be done either synchronously or
asynchronously. The effect is identical to that of synchronous or asynchronous sampling of
EOP#. When DREQn is used to terminate a DMA transfer in asynchronous mode, DREQn must
be sampled inactive one CLKOUT before READY#. In synchronous mode it must be sampled
inactive at the same time as READY#. When DREQn is sampled active in either of the above
cases another DMA cycle is executed (depending on operating mode).
T2
T1
T2
Ti
Tx
Tx
Tx
CLKOUT
DRQn
ADS#
EOP#
(As an output)
READY#
x Cycle
DMA Cycle
A2483-02
Figure 12-6. Buffer Transfer Ended by an Expired Byte Count
T2
T1
T2
T2
T2
T2
Ti
CLKOUT
ADS#
READY#
EOP# (Async)
EOP# (Sync)
DMA Cycle
A2482-02
Figure 12-7. Buffer Transfer Ended by the EOP# Input
12-11
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
12.2.6 Buffer-transfer Modes
After a buffer transfer is completed or terminated, a channel can either become idle (require re-
programming) or reprogram itself and begin another buffer transfer after it is initiated by a hard-
ware or software request. The DMA’s three buffer-transfer modes (single, autoinitialize, and
chaining) determine whether a channel becomes idle or is reprogrammed after it completes or ter-
minates a buffer transfer.
12.2.6.1
Single Buffer-Transfer Mode
By default (single buffer-transfer mode), the DMA transfers a channel’s buffer only once. When
the entire buffer of data has been transferred, the channel becomes idle and must be repro-
grammed before it can perform another buffer transfer. The single buffer-transfer mode is useful
when you know the exact amount of data to be transferred and you know that there is time to re-
program the channel (requester and target addresses and byte count) before another buffer of data
needs to be transferred.
12.2.6.2
Autoinitialize Buffer-Transfer Mode
When programmed for the autoinitialize buffer-transfer mode, the DMA automatically reloads
the channel with the original transfer information (the requester and target addresses and the byte
count) when the transfer completes. The channel then repeats the original buffer transfer. The au-
toinitialize buffer-transfer mode is useful when you need to transfer a fixed amount of data be-
tween the same locations multiple times.
12.2.6.3
Chaining Buffer-Transfer Mode
This mode is similar to the autoinitialize buffer-transfer mode, in that the DMA automatically re-
programs the channel after the current buffer transfer is complete. The difference is that the au-
toinitialize buffer-transfer mode uses the original transfer information, while the chaining buffer-
transfer mode uses new transfer information. While a channel is performing a chaining buffer
transfer, you write new requester and target addresses and a new byte count to it. This prepares
the channel for the next buffer transfer, without affecting the current buffer transfer. When the
channel completes its current buffer transfer, the channel is automatically programmed with the
new transfer information that you wrote to it. The chaining buffer-transfer mode is useful when
you need to transfer data between multiple requesters and targets.
NOTE
If a channel does not contain new transfer information at the end of its buffer
transfer, the channel becomes idle, ending the chaining process; it must be
reprogrammed before it can perform another buffer transfer.
The Chaining Buffer Transfer Mode is entered from the Single Buffer Transfer Mode. The mode
registers should be programmed first, with all of the transfer modes defined as if the channel were
to operate in the Single Buffer Transfer Mode. The channel’s base and current registers are then
loaded. When the channel has been set up in this way and the chaining interrupt service routine
is in place, the Chaining Buffer Transfer Mode can be entered by programming the Chaining reg-
ister. “Chaining Register (DMACHR)” on page 12-47 describes this process.
12-12
DMA CONTROLLER
The DMAINT signal is active immediately after the Chaining Process has been entered, as the
channel then perceives the Base Registers to be empty and in need of reloading. It is important to
have the interrupt service routine in place at the time the Chaining Process is entered. The inter-
rupt request is removed when the most significant byte of the Base Target Address is loaded.
NOTE
Since the most significant byte of the Base Target Address only exists in
0FXXXH I/O address space, the Chaining Buffer Transfer Mode cannot be
used in a DOS Compatible-only mode.
The interrupt occurs again when the first buffer transfer expires and the Current Registers are
loaded from the Base Registers. The cycle continues until the Chaining Process is disabled, or the
host fails to respond to DMAINT before the Current Buffer expires.
Exiting the Chaining Process can be done by resetting the Chaining Mode Register. If an interrupt
is pending for the channel when the Chaining Register is reset, the interrupt request is removed.
The Chaining Process can be temporarily disabled by setting the channel’s mask bit in the Mask
Register.
The interrupt service routine for DMAINT has the responsibility of reloading the Base Register
as necessary. It should check the status of the channel to determine the cause of the channel ex-
piration, etc. It should also have access to operating system information regarding the channel, if
any exists. The DMAINT service routine should be capable of determining whether the chain
should be continued or terminated and act on that information.
NOTE
The chaining buffer-transfer mode is not useful with block transfer mode since
the CPU must be able to get control of the bus before the end of the “block” in
order to reprogram the new values into the DMA registers. Since block
transfer mode locks out any other bus requests (except refresh) the processor
cannot regain control of the bus until the entire block has been transferred.
12.2.7 Data-transfer Modes
There are three data-transfer modes (single, block, and demand) that determine how the bytes or
words that make up a buffer of data are transferred. The DMAMOD1 register is used to select a
channel’s data transfer mode.
Single Mode
A channel request causes one byte or word (depending on the
selected bus widths) to be transferred. Single mode requires a
channel request for every data transfer within a buffer transfer.
Block Mode
A channel request causes the entire buffer of data to be transferred.
Demand Mode
The amount of buffer data (bytes or words) that the channel transfers
depends on how long the channel request input is held active. In this
mode, the channel continues to transfer data while the channel
request input is held active; when the signal goes inactive, the buffer
12-13
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
transfer is suspended and the channel waits for the request input to be
reactivated before it continues.
12.2.7.1
Single Data-transfer Mode
In single data-transfer mode, a DMA request causes the channel to gain bus control. The channel
transfers data (a byte or a word), decrements the buffer byte count (by 1 for byte transfers and 2
for word transfers), then relinquishes bus control. The channel will then Autoinitialize if it has
been programmed to do so. The channel continues to operate in this manner until the buffer trans-
fer is complete or terminated. In this mode, the channel gives up bus control after every data trans-
channel’s buffer-transfer mode determines whether the channel becomes idle or is reprogrammed
after a buffer transfer completes or is terminated.
The single data-transfer mode is compatible with all of the buffer-transfer modes. The following
flowcharts show the transfer process flow for a channel programmed for single data-transfer
mode with each buffer-transfer mode: single (Figure 12-8), autoinitialize (Figure 12-9), and
chaining (Figure 12-10).
12-14
DMA CONTROLLER
After initialization, the DMA channel is
programmed with the requester and
target addresses and a byte count.
No
DREQn
active?
Yes
DMA gains bus control.
DMA transfers one byte or word of data
and decrements the byte count.
DMA channel relinquishes bus control.
Byte
count = FFFFFFH
or EOP#
active?
No
Yes
Buffer transfer is complete, so channel
becomes idle.
A2331-02
Figure 12-8. Single Data-transfer Mode with Single Buffer-transfer Mode
12-15
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
After initialization, the DMA channel is
programmed with the requester and
target addresses and a byte count.
No
DREQn
active?
Yes
DMA gains bus control.
DMA transfers one byte or word of data
and decrements the byte count.
DMA channel relinquishes bus control.
Byte
count = FFFFFFH
No
or EOP#
active?
Yes
DMA channel is reprogrammed with the
original addresses and byte count.
A2332-02
Figure 12-9. Single Data-transfer Mode with Autoinitialize Buffer-transfer Mode
12-16
DMA CONTROLLER
After initialization, the DMA channel is
programmed with the requester and
target addresses and a byte count.
No
DREQn
active?
Yes
DMA gains bus control, transfers one byte
or word of data, decrements byte count, and
then relinquishes bus control.
Write new requester and
target addresses and a
new byte count.
Is there
a new process
to set up?
Yes
No
Byte
count = FFFFFFH
or EOP#
No
active?
Yes
Was the
channel set up
for a new
DMA is programmed
with the new addresses
and byte count.
Yes
process?
No
No new transfer information, so channel
becomes idle.
A2335-02
Figure 12-10. Single Data-transfer Mode with Chaining Buffer-transfer Mode
12-17
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
12.2.7.2
Block Data-transfer Mode
In block data-transfer mode, a channel request initiates a buffer transfer. The channel gains bus
control, then transfers the entire buffer of data. The DRQn signal only needs to be held active until
DACKn# is active.
NOTE
Block mode, unlike the single mode, only gives up control of the bus for
DRAM refresh cycles.
As with single mode, the channel’s buffer-transfer mode determines whether the channel be-
comes idle or is reprogrammed after the buffer transfer completes or is terminated.
The block data-transfer mode is compatible with the single and autoinitialize buffer-transfer
modes, but not with the chaining buffer-transfer mode. The chaining buffer-transfer mode
requires that the transfer information for the next buffer transfer be written to the channel before
the current buffer transfer completes. This is impossible with block data-transfer mode, because
the channel only relinquishes control of the bus for DRAM refresh cycles during the buffer
transfer. The following flowcharts show the transfer process flow for a channel programmed for
the block data-transfer mode with the single (Figure 12-11) and autoinitialize (Figure 12-12)
buffer-transfer modes.
12-18
DMA CONTROLLER
After initialization, the DMA channel is
programmed with the requester and target
addresses and a byte count.
No
DREQn
active?
Yes
DMA gains bus control.
DMA transfers one byte or word of data and
decrements the byte count.
Byte
count = FFFFFFH
or EOP#
active?
No
Yes
DMA channel relinquishes bus control.
Buffer transfer is complete, so channel
becomes idle.
A2334-02
Figure 12-11. Block Data-transfer Mode with Single Buffer-transfer Mode
12-19
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
After initiallization, the DMA channel is
programmed with the requester and target
addresses and a byte count.
No
DREQn
active?
Yes
DMA gains bus control.
DMA transfers one byte or word of data
and decrements the byte count
Byte
count = FFFFFFH
or EOP#
active?
No
Yes
DMA channel relinquishes bus control.
DMA channel is reprogrammed with the
original addresses and byte count.
A2333-02
Figure 12-12. Block Data-transfer Mode with Autoinitialize Buffer-transfer Mode
12-20
DMA CONTROLLER
12.2.7.3
Demand Data-transfer Mode
In demand data-transfer mode, a channel request initiates a buffer transfer. The channel gains bus
control and begins the buffer transfer. As long as the request signal (DRQn) remains active, the
channel continues to perform data transfers. When the DRQn signal goes inactive, the channel
completes its current bus cycle and relinquishes bus control, suspending the buffer transfer. In this
way, the demand mode allows peripherals to access memory in small, irregular bursts without
wasting bus control time. As in other data-transfer modes, a buffer transfer is completed when the
buffer’s byte count expires or is terminated if the EOP# input is activated. At this point, the chan-
nel’s buffer-transfer mode determines whether the channel becomes idle or is reprogrammed.
Since DRQn going inactive suspends a buffer transfer, the channel continually samples DRQn
during a demand buffer transfer. During a buffer transfer, the channel can sample DRQn synchro-
nously or asynchronously (it always samples DRQn asynchronously at the start of a buffer trans-
fer). With synchronous sampling, the channel samples DRQn at the end of the last state of every
data transfer. With asynchronous sampling, the channel samples DRQn at the beginning of every
state, then waits until the end of the state to act on the input. See Figure 12-13. The DRQn sam-
pling is programmed in the DMACMD2 register (Figure 12-24).
T2
T1
T2
T2
T2
T2
Ti
CLKOUT
ADS#
READY#
DRQn (Async)
DRQn (Sync)
x
DMA Cycle
Cycle
A2481-02
Figure 12-13. Buffer Transfer Suspended by the Deactivation of DRQn
The demand data-transfer mode is compatible with all of the buffer-transfer modes. The follow-
ing flowcharts show the transfer process flow for a channel programmed for the demand data-
transfer mode with each buffer-transfer mode: single (Figure 12-14), autoinitialize (Figure
12-15), and chaining (Figure 12-16).
12-21
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
After initialization, the DMA channel is
programmed with the requester and target
addresses and a byte count.
No
DREQn
active?
Yes
DMA gains bus control.
DMA transfers one byte or word of data
and decrements the byte count.
Yes
Byte
count = FFFFFFH
DMA channel
relinquishes
bus control.
DREQn
active?
or EOP#
active?
No
No
Yes
DMA channel relinquishes bus control.
Buffer transfer is complete, so channel
becomes idle.
A2338-02
Figure 12-14. Demand Data-transfer Mode with Single Buffer-transfer Mode
12-22
DMA CONTROLLER
After initialization, the DMA channel is
programmed with the requester and target
addresses and a byte count.
No
DREQn
active?
Yes
DMA gains bus control.
DMA transfers one byte or word of data
and decrements the byte count.
Yes
Byte
count = FFFFFFH
DMA channel
relinquishes
bus control.
DREQn
active?
or EOP#
active?
No
No
Yes
DMA channel relinquishes bus control.
DMA channel is reprogrammed with the
original addresses and byte count.
A2339-02
Figure 12-15. Demand Data-transfer Mode with Autoinitialize Buffer-transfer Mode
12-23
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
After initialization, the DAM channel is
programmed with the requester and target
addresses and a byte count.
No
DREQn
No
active?
Write new
requester
and target
addresses
and a new
byte count.
Is there
a new process
to set up?
Yes
Yes
DMA gains bus control.
DMA transfers one byte or word of data
and decrements the byte count.
Yes
Byte
count = FFFFFFH
DMA channel
relinquishes
bus control.
DREQn
active?
or EOP#
active?
No
No
Yes
DMA channel relinquishes bus control.
Was the
channel set up
for a new
DMA is
programmed
with the new
addresses and
byte count.
Yes
process?
No
No new transfer information, so channel
becomes idle.
A2336-02
Figure 12-16. Demand Data-transfer Mode with Chaining Buffer-transfer Mode
12-24
DMA CONTROLLER
12.2.8 Cascade Mode
Cascade mode allows an external 8237A or another DMA-type device to gain bus control. A cas-
caded device requests bus control by holding a channel’s request input (DRQn) active. Once
granted bus control, the cascaded device remains bus master until it relinquishes bus control by
deactivating DRQn.
If a refresh request occurs while a cascaded device has bus control, the cascaded device must
deassert its request or the refresh cycle will be missed. The following steps take place in response
to a refresh request.
1. The channel deasserts its acknowledge signal (DACKn#) to the cascaded device.
— At this point, the cascaded device should relinquish bus control by removing DRQn.
2. As soon as DRQn is removed, the refresh cycle is started.
it must reassert DRQn.
3. If the cascaded device has reasserted DRQn when the refresh cycle is complete, the
channel reasserts DACKn#, giving bus control back to the cascaded device without bus
priority arbitration.
The following flowchart (Figure 12-17) shows this process flow.
12-25
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
After initialization, the DMA channel
is programmed.
No
DRQn
active?
Yes
Cascaded device gains bus control.
No
Cascaded
device
deasserts
DRQn,
Refresh
cycle is
performed.
Refresh
request?
DRQn
active?
Yes
Yes
relinquishing
bus control.
No
Cascaded device relinquishes bus
control.
Cascade cycle complete.
A2337-02
Figure 12-17. Cascade Mode
12.2.9 DMA Interrupts
Each channel contains two interrupt causing signals, chaining status and transfer complete. When
a channel is configured for the chaining buffer-transfer mode, the chaining status signal indicates
that the channel has started its buffer transfer and new transfer information can be written without
affecting the current buffer transfer. Once activated, the chaining status signal remains active until
the most significant byte of the base target address is written, or resetting the chaining enable bit.
The transfer complete status signal indicates that the channel has finished a buffer transfer —
either the channel’s byte count has expired or the buffer transfer was terminated by an EOP#
input. DMACLRTC clears the DMAINT signal going to the Interrupt Control Unit. DMACLRTC
is executed by writing to location F01EH; the data written to the location is immaterial — writing
any data to the location causes the DMA to deactivate the transfer complete status signal.
12-26
DMA CONTROLLER
The four interrupt source signals (two per channel) are internally connected (ORed) to the inter-
rupt request output (DMAINT). When an interrupt from DMAINT is detected, you can determine
which signal caused the request by reading the DMA interrupt status register.
12.2.10 8237A Compatibility
Although the DMA is an enhancement over the 8237A, you can configure it to operate in an
8237A-compatible mode. A list of the features common to the DMA and 8237A and a list of
DMA enhancements follow.
Features common to the DMA and 8237A:
• Data-transfer modes (single, block, and demand)
• Buffer-transfer modes (single and autoinitialize)
• Fly-by data transfer bus cycle option
• Programmed via 8-bit registers
• Transfers between memory and I/O (target must be in memory and requester must be
external)
DMA enhancements:
• Chaining buffer-transfer mode
• Two-cycle data transfer bus cycle option (provides byte assembly and allows memory-to-
memory transfers using only one channel)
• Transfers between any combination of memory and I/O
• Address registers for both the target and the requester; addresses can be incremented,
decremented, or left unchanged during a buffer transfer
A channel is configured for 8237A compatibility by enabling only the common features and lim-
iting the byte count and the target address modification capability. The 8237A uses a 16-bit target
address and a 16-bit byte count, while the DMA uses a 26-bit target address and a 24-bit byte
count. Therefore, for compatibility, the DMA contains an overflow register that allows you to
configure the target and byte count so that only the lower 16 bits are modified during buffer trans-
fers. With this configuration, the upper byte count bits are ignored; the byte count expires when
it is decremented from 0000H to FFFFH (16-bit versus 24-bit rollovers).
12-27
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
12.3 REGISTER DEFINITIONS
Table 12-3 lists the registers associated with the DMA unit, and the following sections contain bit
descriptions for each register.
Table 12-3. DMA Registers (Sheet 1 of 3)
Expanded
Address
PC/AT*
Address
Register
PINCFG
Description
F826H
F830H
F008H
—
—
Pin Configuration:
(read/write)
Connects the DMA channel acknowledge
(DMAACK0#, DMAACK1#) and end-of-process
signals to package pins DACK0#, DACK1# and
EOP#, respectively.
DMACFG
(read/write)
DMA Configuration:
Determines which signal is connected to the DMA
channel request inputs (DREQn). Masks the channel
acknowledge signals (DMAACK0#, DMAACK1#),
which is useful when using internal requesters.
DMACMD1
(write only)
0008H
DMA Command 1:
Simultaneously enables or disables both DMA
channels. Enables the rotating method for changing
the bus control priority structure.
DMA0REQ0
DMA0REQ1
DMA0REQ2
DMA0REQ3
F010H
F010H
F011H
F011H
—
—
—
—
Channel 0 and 1 Requester Address:
Contains channel n’s 26-bit requester address.
During a buffer transfer, this address may be
incremented, decremented, or left unchanged.
Reading these registers returns the current address.
DMA1REQ0
DMA1REQ1
DMA1REQ2
DMA1REQ3
(read/write)
F012H
F012H
F013H
F013H
—
—
—
—
DMA0TAR0
DMA0TAR1
DMA0TAR2
DMA0TAR3
F000H
F000H
F087H
F086H
0000H
0000H
0087H
—
Channel 0 and 1 Target Address:
Contains channel n’s 26-bit target address. During a
buffer transfer, this address may be incremented,
decremented, or left unchanged. Reading these
registers returns the current address.
DMA1TAR0
DMA1TAR1
DMA1TAR2
DMA1TAR3
(read/write)
F002H
F002H
F083H
F085H
0002H
0002H
0083H
—
DMA0BYC0
DMA0BYC1
DMA0BYC2
F001H
F001H
F098H
0001H
0001H
—
Channel 0 and 1 Byte Count:
Contains channel n’s 24-bit byte count. During a
buffer transfer, this byte count is decremented.
Reading these registers returns the current byte
count.
DMA1BYC0
DMA1BYC1
DMA1BYC2
(read/write)
F003H
F003H
F099H
0003H
0003H
—
12-28
DMA CONTROLLER
Table 12-3. DMA Registers (Sheet 2 of 3)
Expanded
Address
PC/AT*
Address
Register
DMASTS
Description
F008H
0008H
DMA Status:
(read only)
Indicates whether a hardware request is pending on
channel 0 and 1. Indicates whether channel 0’s or
channel 1’s byte count has expired.
DMACMD2
(write only)
F01AH
F00BH
—
DMA Command 2:
Assigns a bus control requester (DMA channel 0,
DMA channel 1, or external bus master) to the lowest
priority level. Selects the type of sampling for the end-
of-process (EOP#) and the DMA request (DRQn)
inputs. The DMA can sample these signals
asynchronously or synchronously.
DMAMOD1
(write only)
000BH
DMA Mode 1:
Determines the data-transfer mode. Enables the
autoinitialize buffer-transfer mode. Determines the
transfer direction (whether the target is the
destination or source for a transfer). Determines
whether the DMA increments or decrements the
target address during a buffer transfer (only if the
DMA is set up to modify the target address; see
DMAMOD2).
DMAMOD2
(write only)
F01BH
—
DMA Mode 2:
Selects the data transfer bus cycle option. Specifies
whether the requester and target are in memory or
I/O. Determines whether the DMA modifies the target
and requester addresses. Determines whether the
DMA increments or decrements the requester
address during a buffer transfer (only if the DMA is set
up to modify the requester address).
DMASRR
F009H
0009H
DMA Software Request:
(read/write)
Write Format
Generates a channel 0 and/or a channel 1 software
request.
Read Format
Indicates whether a software request is pending on
DMA channel 0 or 1.
DMAMSK
(write only)
F00AH
F00FH
F018H
000AH
000FH
—
DMA Individual Channel Mask:
Individually masks (disables) channel 0’s and 1’s
hardware request input (DREQ0 and DREQ1). This
does not mask software requests.
DMAGRPMSK
(read/write)
DMA Group Channel Mask:
Simultaneously masks (disables) both channels’
hardware request inputs (DREQ0 and DREQ1). This
does not mask software requests.
DMABSR
DMA Bus Size:
(write only)
Determines the requester and target data bus widths
(8 or 16 bits).
12-29
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table 12-3. DMA Registers (Sheet 3 of 3)
Expanded
Address
PC/AT*
Address
Register
Description
DMACHR
(write only)
F019H
F01CH
F019H
—
—
—
DMA Chaining:
Enables chaining buffer-transfer mode for a specified
channel.
DMAIEN
(read/write)
DMA Interrupt Enable:
Connects the channel transfer complete status
signals to the interrupt request output (DMAINT).
DMAIS
DMA Interrupt Status:
(read only)
Indicates which signal generated an interrupt request:
channel 0 transfer complete, channel 1 transfer
complete, channel 0 chaining, or channel 1 chaining
status.
DMAOVFE
(read/write)
F01DH
—
DMA Overflow Enable:
Included for 8237A compatibility. Controls whether all
26 bits or only the lower 16 bits of the requester and
target addresses are incremented or decremented
during buffer transfers. Controls whether the byte
count is 24 bits or 16 bits.
12-30
DMA CONTROLLER
12.3.1 Pin Configuration Register (PINCFG)
Use PINCFG to connect DACK0#, EOP#, and DACK1# to package pins.
Pin Configuration
PINCFG
Expanded Addr: F826H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
—
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7
—
Reserved. This bit is undefined; for compatibility with future devices, do
not modify this bit.
6
5
PM6
PM5
Pin Mode:
0 = Selects CS6# at the package pin.
1 = Selects REFRESH# at the package pin.
Pin Mode:
0 = Selects the coprocessor signals, PEREQ, BUSY#, and ERROR#, at
the package pins.
1 = Selects the timer control unit signals, TMROUT2, TMRCLK2, and
TMRGATE2, at the package pins.
4
3
2
1
0
PM4
PM3
PM2
PM1
PM0
Pin Mode:
0 = Selects DACK0# at the package pin.
1 = Selects CS5# at the package pin.
Pin Mode:
0 = Selects EOP# at the package pin.
1 = Selects CTS1# at the package pin.
Pin Mode:
0 = Selects DACK1# at the package pin.
1 = Selects TXD1 at the package pin.
Pin Mode:
0 = Selects SRXCLK at the package pin.
1 = Selects DTR1# at the package pin.
Pin Mode:
0 = Selects SSIOTX at the package pin.
1 = Selects RTS1# at the package pin.
Figure 12-18. Pin Configuration Register (PINCFG)
12-31
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
12.3.2 DMA Configuration Register (DMACFG)
Use DMACFG to select one of the hardware sources for each channel and to mask the DMA ac-
knowledge (DMAACKn#) signals when using internal requesters.
DMA Configuration
DMACFG
Expanded Addr: F830H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
D1MSK
D1REQ2
D1REQ1
D1REQ0
D0MSK
D0REQ2
D0REQ1
D0REQ0
Bit
Number
Bit
Mnemonic
Function
7
D1MSK
DMA Acknowledge 1 Mask:
0 = DMA channel 1’s acknowledge (DMAACK1#) signal is not masked.
1 = Masks DMA channel 1’s acknowledge (DMAACK1#) signal. Useful
when channel 1’s request (DREQ1) input is connected to an internal
peripheral.
6–4
D1REQ2:0
DMA Channel 1 Request Connection:
Connects one of the eight possible hardware sources to channel 1’s
request input (DREQ1).
000 = DRQ1 pin (external peripheral)
001 = SIO channel 1’s receive buffer full signal (RBFDMA1)
010 = SIO channel 0’s transmit buffer empty signal (TXEDMA0)
011 = SSIO receive holding buffer full signal (SSRBF)
100 = TCU counter 2’s output signal (OUT2)
101 = SIO channel 0’s receive buffer full signal (RBFDMA0)
110 = SIO channel 1’s transmit buffer empty signal (TXEDMA1)
111 = SSIO transmit holding buffer empty signal (SSTBE)
3
D0MSK
DMA Acknowledge 0 Mask:
0 = DMA channel 0’s acknowledge (DMAACK0#) signal is not masked.
1 = Masks DMA channel 0’s acknowledge (DMAACK0#) signal. Useful
when channel 0’s request (DREQ0) input is connected to an internal
peripheral.
2–0
D0REQ2:0
DMA Channel 0 Request Connection:
Connects one of the eight possible hardware sources to channel 0’s
request input (DREQ0).
000 = DRQ0 pin (external peripheral)
001 = SIO channel 0’s receive buffer full signal (RBFDMA0)
010 = SIO channel 1’s transmit buffer empty signal (TXEDMA1)
011 = SSIO transmit holding buffer empty signal (SSTBE)
100 = TCU counter 1’s output signal (OUT1)
101 = SIO channel 1’s receive buffer full signal (RBFDMA1)
110 = SIO channel 0’s transmit buffer empty signal (TXEDMA0)
111 = SSIO receive holding buffer full signal (SSRBF)
Figure 12-19. DMA Configuration Register (DMACFG)
12-32
DMA CONTROLLER
12.3.3 Channel Registers
To program a DMA channel’s requester and target addresses and its byte count, write to the DMA
channel registers. Some of the channel registers require the use of a byte pointer (BP) flip-flop to
control the access to the upper and lower bytes. After you write or read a register that requires a
byte pointer specification, the DMA toggles the byte pointer. For example, writing to
DMA0TAR0 with BP=0 causes the DMA to set BP. The clear byte pointer software command
(DMACLRBP) is available so that you can force BP to a known state (0) before writing to the
channel registers. Issue DMACLRBP by writing to location F00CH or 000CH; the data written
to the location doesn’t matter —writing to the location is all that is necessary to cause the DMA
to clear the byte pointer.
DMA Channel 0
24
DMA0REQ3
16
DMA0REQ2
8
0
Requester Address
Target Address
Byte Count
DMA0REQ1
DMA0REQ0
F011H (BP=1)
F011H (BP=0)
F010H (BP=1)
F010H (BP=0)
24
DMA0TAR3
F086H
16
DMA0TAR2
F087H
8
DMA0TAR1
F000H (BP=1)
0
DMA0TAR0
F000H (BP=0)
16
DMA0BYC2
F098H
8
DMA0BYC1
F001H (BP=1)
0
DMA0BYC0
F001H (BP=0)
DMA Channel 1
24
DMA1REQ3
16
DMA1REQ2
F013H (BP=0)
8
DMA1REQ1
F012H (BP=1)
0
DMA1REQ0
F012H (BP=0)
Requester Address
Target Address
Byte Count
F013H (BP=1)
24
DMA1TAR3
F085H
16
DMA1TAR2
F083H
8
DMA1TAR1
F002H (BP=1)
0
DMA1TAR0
F002H (BP=0)
16
DMA1BYC2
F099H
8
DMA1BYC1
F003H (BP=1)
0
DMA1BYC0
F003H (BP=0)
Figure 12-20. DMA Channel Address and Byte Count Registers
(DMAnREQn, DMAnTARn, DMAnBYCn)
12-33
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
NOTE
The value you write to the byte count register must be one less than the
number of bytes to be transferred. To transfer one byte, write zero to the byte
count register (byte count = number of bytes – 1). To transfer one word, write
one (byte) to the byte count register (byte count = [number of words X 2] – 1).
12.3.4 Overflow Enable Register (DMAOVFE)
Use DMAOVFE to specify whether all 26 bits or only the lower 16 bits of the target and requester
addresses are incremented or decremented during buffer transfers and to determine whether all
24 bits of the byte count or only the lower 16 bits of the byte count are decremented during buffer
transfers. A byte count configured for 16-bit decrementing expires when it is decremented from
0000H to 0FFFFH.
DMA Overflow Enable
DMAOVFE
Expanded Addr: F01DH
ISA Addr:
—
(read/write)
Reset State:
0AH
7
0
—
—
—
—
ROV1
TOV1
ROV0
TOV0
Bit
Number
Bit
Mnemonic
Function
7–4
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
3
2
ROV1
TOV1
Channel 1 Requester Overflow Enable:
0 = lowest 16 bits of requester address increment/decrement
1 = all bits of requester address increment/decrement
Channel 1 Target & Byte Counter Overflow Enable:
0 = lowest 16 bits of target address and byte count
increment/decrement
1 = all bits of target address and byte count increment/decrement
1
0
ROV0
TOV0
Channel 0 Requester Overflow Enable:
0 = lowest 16 bits of requester address increment/decrement
1 = all bits of requester address increment/decrement
Channel 0 Target & Byte Counter Overflow Enable:
0 = lowest 16 bits of target address and byte count
increment/decrement
1 = all bits of target address and byte count increment/decrement
Figure 12-21. DMA Overflow Enable Register (DMAOVFE)
12-34
DMA CONTROLLER
12.3.5 Command 1 Register (DMACMD1)
Use DMACMD1 to enable both channels and to select the rotating method for changing the bus
control priority structure.
DMA Command 1
DMACMD1
Expanded Addr: F008H
ISA Addr:
0008H
(write only)
Reset State:
00H
7
0
—
—
—
PRE
—
CE
—
—
Bit
Number
Bit
Mnemonic
Function
7–5
—
Reserved; for compatibility with future devices, write zeros to these bits.
Priority Rotation Enable:
4
PRE
0 = Priority is fixed based on value in DMACMD2.
1 = Enables the rotation method for changing the bus control priority
structure. That is, after the external bus master or one of the DMA
channels is given bus control, it is assigned to the lowest priority
level.
3
2
—
Reserved; for compatibility with future devices, write zero to this bit.
Channel Enable:
CE
0 = Enables channel 0 and 1.
1 = Disables the channels.
1–0
—
Reserved; for compatibility with future devices, write zeros to these bits.
Figure 12-22. DMA Command 1 Register (DMACMD1)
12-35
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
12.3.6 Status Register (DMASTS)
Use DMASTS to check the status of the channels individually. The DMA sets bits in this register
to indicate that a channel has a hardware request pending or that a channel’s byte count has ex-
pired.
DMA Status
DMASTS
(read only)
Expanded Addr: F008H
ISA Addr:
0008H
00H
Reset State:
7
0
—
—
R1
R0
—
—
TC1
TC0
Bit
Bit
Function
Number Mnemonic
7–6
5
—
Reserved. These bits are undefined.
Request 1:
R1
When set, this bit indicates that channel 1 has a hardware request
pending. When the request is removed, this bit is cleared.
4
R0
Request 0:
When set, this bit indicates that channel 0 has a hardware request
pending. When the request is removed, this bit is cleared.
3–2
1
—
Reserved. These bits are undefined.
Transfer Complete 1:
TC1
When set, this bit indicates that channel 1 has completed a buffer
transfer (either its byte count expired or it received an EOP# input).
Reading this register clears this bit and clears TC1 in DMAIS.
0
TC0
Transfer Complete 0:
When set, this bit indicates that channel 0 has completed a buffer
transfer (either its byte count expired or it received an EOP# input).
Reading this register clears this bit and clears TC0 in DMAIS.
Figure 12-23. DMA Status Register (DMASTS)
12-36
12.3.7 Command 2 Register (DMACMD2)
Use DMACMD2 to select the DREQn and EOP# sampling: asynchronous or synchronous. Bus
timing diagrams that show the differences between asynchronous and synchronous sampling are
shown in Figure 12-5 on page 12-10 and Figure 12-13 on page 12-21. Also, use DMACMD2 to
assign a particular bus request to the lowest priority level for fixed priority mode.
DMA Command 2
DMACMD2
Expanded Addr: F01AH
ISA Addr:
—
(write only)
Reset State:
08H
7
0
—
—
—
—
PL1
PL0
ES
DS
Bit
Number
Bit
Mnemonic
Function
7–4
—
Reserved; for compatibility with future devices, write zeros to these bits.
Low Priority Level Set:
3–2
PL1:0
Use these bits to assign a particular bus request to the lowest priority
level in fixed priority mode.
00 = Assigns channel 0’s request (DREQ0) to the lowest priority level
01 = Assigns channel 1’s request (DREQ1) to the lowest priority level
10 = Assigns the external bus master request (HOLD) to the lowest
priority level
11 = Reserved
1
0
ES
DS
EOP# Sampling:
0 = Causes the DMA to sample the EOP# input asynchronously.
1 = Causes the DMA to sample the end-of-process (EOP#) input
synchronously.
DREQn Sampling:
0 = Causes the DMA to sample the DREQn inputs asynchronously.
1 = Causes the DMA to sample the channel request (DREQn) inputs
synchronously.
Figure 12-24. DMA Command 2 Register (DMACMD2)
12-37
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
12.3.8 Mode 1 Register (DMAMOD1)
Use DMAMOD1 to select a particular channel’s data-transfer mode and transfer direction and to
enable the channel’s auto-initialize buffer-transfer mode. You can configure the DMA to modify
the target address during a buffer transfer by clearing DMAMOD2.2, then use DMAMOD1.5 to
specify how the channel modifies the address.
12-38
DMA CONTROLLER
DMA Mode 1
DMAMOD1
(write only)
Expanded Addr: F00BH
ISA Addr:
000BH
Reset State:
00H
7
0
DTM1
DTM0
TI
AI
TD1
TD0
0
CS
Bit
Number
Bit
Mnemonic
Function
7–6
DTM1:0
Data-transfer Mode:
00 = Demand
01 = Single
10 = Block
11 = Cascade
5
TI
Target Increment/Decrement:
0 = Causes the target address to be incremented after each data
transfer in a buffer transfer.
1 = Causes the target address for the channel specified by bit 0 to be
decremented after each data transfer in a buffer transfer. Note that it
does not decrement words. When decrementing it will do two byte
transfers for a word.
Note: When the target address is programmed to remain constant
(DMAMOD2.2 = 1), this bit is a don’t care.
4
AI
Autoinitialize:
0 = Disables the autoinitialize buffer-transfer mode for the channel
specified by bit 0.
1 = Enables the autoinitialize buffer-transfer mode for the channel
specified by bit 0.
3–2
TD1:0
Transfer Direction:
Determines the transfer direction for the channel specified by bit 0.
00 = Target is read; nothing is written (used for testing)
01 = Data is transferred from the requester to the target
10 = Data is transferred from the target to the requester
11 = Reserved
Note: In cascade mode, these bits become don’t cares.
Must be 0 for correct operation.
Channel Select:
1
0
0
CS
0 = The selections for bits 7–2 affect channel 0.
1 = The selections for bits 7–2 affect channel 1.
Figure 12-25. DMA Mode 1 Register (DMAMOD1)
12-39
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
12.3.9 Mode 2 Register (DMAMOD2)
Use DMAMOD2 to select the data transfer bus cycle option, specify whether the requester and
target are in memory or I/O, and determine whether the DMA modifies the target and requester
addresses. If you set up the DMA to modify the requester address, use DMAMOD2 to determine
whether the DMA increments or decrements the requester address during a buffer transfer.
12-40
DMA CONTROLLER
DMA Mode 2
DMAMOD2
(write only)
Expanded Addr: F01BH
ISA Addr:
—
Reset State:
00H
7
0
BCO
RD
TD
RH
RI
TH
0
CS
Bit
Number
Bit
Mnemonic
Function
7
BCO
Bus Cycle Option:
0 = Selects the fly-by data transfer bus cycle option for the channel specified
by bit 0.
1 = Selects the two-cycle data transfer bus cycle option for the channel
specified by bit 0.
6
RD
Requester Device Type:
0 = Clear this bit when the requester for the channel specified by bit 0 is in
memory space.
1 = Set this bit when the requester for the channel specified by bit 0 is in I/O
space.
This bit is ignored if BCO is cleared.
Target Device Type:
5
4
3
TD
RH
RI
0 = Clear this bit when the target for the channel specified by bit 0 is in
memory space.
1 = Set this bit when the target for the channel specified by bit 0 is in I/O
space.
Requester Address Hold:
0 = Causes the address to be modified (incremented or decremented,
depending on DMAMOD2.3).
1 = Causes the requester’s address for the channel specified by bit 0 to
remain constant during a buffer transfer.
Requester Address Increment/Decrement:
0 = Causes the requester address to be incremented after each data transfer
in a buffer transfer.
1 = Causes the requester address for the channel specified by bit 0 to be
decremented after each data transfer in a buffer transfer. Note that it
does not decrement words. When decrementing it will do two byte
transfers for a word.
Note: When the target address is programmed to remain constant
(DMAMOD2.4 = 1), this bit is a don’t care.
2
TH
Target Address Hold:
0 = Causes the address to be modified (incremented or decremented,
depending on DMAMOD1.5).
1 = Causes the target’s address for the channel specified by bit 0 to remain
constant during a buffer transfer.
1
0
0
Must be 0 for correct operation.
Channel Select:
CS
0 = The selections for bits 7–2 affect channel 0.
1 = The selections for bits 7–2 affect channel 1.
Figure 12-26. DMA Mode 2 Register (DMAMOD2)
12-41
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
12.3.10 Software Request Register (DMASRR)
Write DMASRR to issue software DMA service requests. Software requests are subject to bus
control priority arbitration with all other software and hardware requests. A software request ac-
tivates the internal channel request signal. This signal remains active until the channel completes
its buffer transfer (either by an expired byte count or an EOP# input). In the demand data-transfer
mode, a buffer transfer is suspended by deactivating the channel request signal. Because you can-
not deactivate the internal channel request signal before the end of a buffer transfer, you cannot
use software requests with demand data-transfer mode.
DMA Software Request (write format)
DMASRR
Expanded Addr: F009H
ISA Addr:
0009H
Reset State:
00H
7
0
—
—
—
—
—
SR
0
CS
Bit
Number
Bit
Mnemonic
Function
7–3
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
2
SR
Software Request:
Setting this bit generates a software request for the channel specified by
bit 0. When the channel’s buffer transfer completes, this bit is cleared.
1
0
0
Must be 0 for correct operation.
Channel Select:
CS
0 = The selection for bit 2 affects channel 0.
1 = The selection for bit 2 affects channel 1.
Figure 12-27. DMA Software Request Register (DMASRR – write format)
12-42
DMA CONTROLLER
Read DMASRR to see whether a software request for a particular channel is pending. Each re-
quest bit is cleared upon Terminal Count or external EOP#. When in auto-initialize mode, both
bits are cleared when a Terminal Count or external EOP# occurs.
DMA Software Request (read format)
DMASRR
Expanded Addr: F009H
ISA Addr:
0009H
00H
Reset State:
7
0
—
—
—
—
—
—
SR1
SR0
Bit
Number
Bit
Mnemonic
Function
7–2
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
1
0
SR1
SR0
Software Request 1:
When set, this bit indicates that channel 1 has a software request
pending.
Software Request 0:
When set, this bit indicates that channel 0 has a software request
pending.
Figure 12-28. DMA Software Request Register (DMASRR – read format)
12-43
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
12.3.11 Channel Mask and Group Mask Registers (DMAMSK and DMAGRPMSK)
Use the DMAMSK and DMAGRPMSK registers to disable (mask) or enable channel hardware
requests. DMAMSK allows you to disable or enable hardware requests for only one channel at a
time, while DMAGRPMSK allows you to disable or enable hardware requests for both channels
at once.
NOTE
Each mask bit is set when its associated channel produces an End-of-Process if
the channel is not programmed for Autoinitialize. Software must then clear the
appropriate mask bit to allow further DREQn requests from initiating
transfers.
DMA Individual Channel Mask
DMAMSK
Expanded Addr: F00AH
ISA Addr:
000AH
(write only)
Reset State:
04H
7
0
—
—
—
—
—
HRM
0
CS
Bit
Number
Bit
Mnemonic
Function
7–3
—
Reserved; for compatibility with future devices, write zeros to these bits.
Hardware Request Mask:
2
HRM
0 = Unmasks (enables) hardware requests for the channel specified by
bit 0.
1 = Masks (disables) hardware requests for the channel specified by
bit 0.
NOTE: When this bit is set, the channel can still receive software
requests.
1
0
0
Must be 0 for correct operation.
Channel Select:
CS
0 = The selection for bit 2 affects channel 0.
1 = The selection for bit 2 affects channel 1.
Figure 12-29. DMA Channel Mask Register (DMAMSK)
12-44
DMA CONTROLLER
DMA Group Channel Mask
DMAGRPMSK
Expanded Addr: F00FH
ISA Addr:
000FH
(read/write)
Reset State:
03H
7
0
—
—
—
—
—
—
HRM1
HRM0
Bit
Number
Bit
Mnemonic
Function
7–2
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
1
0
HRM1
HRM0
Hardware Request Mask 1:
0 = Channel 1’s hardware requests are not masked.
1 = Masks (disables) channel 1’s hardware requests. When this bit is
set, channel 1 can still receive software requests.
Hardware Request Mask 0:
0 = Channel 0’s hardware requests are not masked.
1 = Masks (disables) channel 0’s hardware requests. When this bit is
set, channel 0 can still receive software requests.
Figure 12-30. DMA Group Channel Mask Register (DMAGRPMSK)
12-45
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
12.3.12 Bus Size Register (DMABSR)
Use DMABSR to determine the requester and target data bus widths (8 or 16 bits).
DMA Bus Size
DMABSR
Expanded Addr: F018H
ISA Addr:
—
(write only)
Reset State:
X1X10000B
7
0
—
RBS
—
TBS
CS
—
—
0
Bit
Number
Bit
Mnemonic
Function
7
—
Reserved; for compatibility with future devices, write zero to this bit.
Requester Bus Size:
6
RBS
Specifies the requester’s data bus width for the channel specified by bit
0.
0 = 16-bit bus
1 = 8-bit bus
5
4
—
Reserved; for compatibility with future devices, write zero to this bit.
TBS
Target Bus Size:
Specifies the target’s data bus width for the channel specified by bit 0.
0 = 16-bit bus
1 = 8-bit bus
3–1
0
Must be 0 for correct operation.
0
CS
Channel Select:
0 = The selections for bits 7–4 affect channel 0.
1 = The selections for bits 7–4 affect channel 1.
Figure 12-31. DMA Bus Size Register (DMABSR)
12-46
DMA CONTROLLER
12.3.13 Chaining Register (DMACHR)
Use DMACHR to enable or disable the chaining buffer-transfer mode for a selected channel. The
following steps describe how to set up a channel to perform chaining buffer transfers.
1. Set up the chaining interrupt (DMAINT) service routine.
2. Configure the channel for the single buffer-transfer mode.
3. Program the mode registers.
4. Program the target address, requester address, and byte count registers.
5. Enable the channel for the chaining buffer-transfer mode. (This enables the DMAINT
output.)
6. Enable the DMAINT interrupt in the ICU and service it. (The service routine should load
the transfer information for the next buffer transfer.)
7. Enable the channel by unmasking DREQn and setting bit 2 in DMACMD1.
From this point, the chaining interrupt indicates each time the channel requires new transfer in-
formation. The cycle continues as long as the chaining buffer-transfer mode is enabled and new
transfer information is written to the channel. New transfer information must be written to the
channel before the channel’s current buffer transfer completes.
DMA Chaining
DMACHR
Expanded Addr: F019H
ISA Addr:
—
(write only)
Reset State:
00H
7
0
—
—
—
—
—
CE
0
CS
Bit
Number
Bit
Mnemonic
Function
7–3
—
Reserved; for compatibility with future devices, write zeros to these bits.
Chaining Enable:
2
CE
0 = Disables the chaining buffer-transfer mode for the channel specified
by bit 0.
1 = Enables the chaining buffer-transfer mode for the channel specified
by bit 0.
1
0
0
Must be 0 for correct operation.
Channel Select:
CS
0 = The selection for bit 2 affects channel 0.
1 = The selection for bit 2 affects channel 1.
Figure 12-32. DMA Chaining Register (DMACHR)
12-47
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
12.3.14 Interrupt Enable Register (DMAIEN)
Use DMAIEN to individually connect channel 0’s and 1’s transfer complete signal to the
DMAINT interrupt request output.
DMA Interrupt Enable
DMAIEN
Expanded Addr: F01CH
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
—
—
—
—
—
—
TC1
TC0
Bit
Number
Bit
Mnemonic
Function
7–2
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
1
TC1
Transfer Complete 1:
0 = Disables Transfer Complete interrupts.
1 = Connects channel 1’s transfer complete signal to the interrupt
control unit’s DMAINT input.
Note: When channel 1 is in chaining mode (DMACHR.2=1 and
DMACHR.0=1), this bit is a don’t care.
0
TC0
Transfer Complete 0:
0 = Disables Transfer Complete interrupts.
1 = Connects channel 0’s transfer complete signal to the interrupt
control unit’s DMAINT input.
Note: When channel 0 is in chaining mode (DMACHR.2=1 and
DMACHR.0=0), this bit is a don’t care.
Figure 12-33. DMA Interrupt Enable Register (DMAIEN)
12-48
DMA CONTROLLER
12.3.15 Interrupt Status Register (DMAIS)
DMAIS indicates which source activated the DMA interrupt request signal (channel 0 transfer
complete, channel 1 transfer complete, channel 0 chaining, or channel 1 chaining).
DMA Interrupt Status
DMAIS
Expanded Addr: F019H
ISA Addr:
—
(read only)
Reset State:
00H
7
0
TC1
TC0
CI1
CI0
—
—
—
—
Bit
Number
Bit
Mnemonic
Function
7–6
Reserved. These bits are undefined.
Transfer Complete 1:
—
5
TC1
When set, this bit indicates that channel 1 has completed a buffer
transfer (either its byte count expired or it received an EOP# input). This
bit is set only if bit 1 of the interrupt enable register is set. Reading the
DMA status register (DMASTS) clears this bit.
Note: In chaining mode, this bit becomes a don’t care.
Transfer Complete 0:
4
TC0
When set, this bit indicates that channel 0 has completed a buffer
transfer (either its byte count expired or it received an EOP# input). This
bit is set only if bit 0 of the interrupt enable register is set. Reading the
DMA status register (DMASTS) clears this bit.
Note: In chaining mode, this bit becomes a don’t care.
Reserved. These bits are undefined.
3–2
1
—
CI1
Chaining Interrupt 1:
When set, this bit indicates that new requester and target addresses and
a new byte count should be written to channel 1. This bit is cleared when
new transfer information is written to the channel. (Writing to the most-
significant byte of the target address clears this bit.)
Note: Outside chaining mode, this bit becomes a don’t care.
Chaining Interrupt 0:
0
CI0
When set, this bit indicates that new requester and target addresses and
a new byte count should be written to channel 0. This bit is cleared when
new transfer information is written to the channel. (Writing to the most-
significant byte of the target address clears this bit.)
Note: Outside chaining mode, this bit becomes a don’t care.
Figure 12-34. DMA Interrupt Status Register (DMAIS)
12-49
12.3.16 Software Commands
The DMA contains four software commands: clear byte pointer, clear DMA, clear mask register,
and clear transfer complete signal. Each software command has an I/O address associated with it
(see Table 12-4). To issue a software command, write to its I/O address; the data written doesn’t
matter —writing to the location is all that is necessary.
Table 12-4. DMA Software Commands
Name
(Address)
Command
Functions
DMACLRBP
(0F00CH or
000CH)
Clear byte pointer
Resets the byte pointer flip-flop. Perform this
command at the beginning of any access to the
channel registers, to ensure a predictable place in
the register programming sequence.
DMACLR
(0F00DH or
000DH)
Clear DMA
Sets all DMA functions to their default states.
DMACLRMSK
(0F00EH or
000EH)
Clear mask register
Clear transfer complete signal
Simultaneously clears the mask bits of all channels
(enabling all channels).
DMACLRTC
(0F01EH)
Resets the transfer complete signal (DMAINT).
Allows the source of the DMA request (hardware or
software) to acknowledge the completion of a
transfer process.
12.4 DESIGN CONSIDERATIONS
EOP# requires an external pull-up resistor. To determine the maximum value, the rise time must
be less than three bus cycles. To determine the minimum value, use the I specification from the
OL
Intel386™ EX Embedded Microprocessor datasheet (order number 272420).
12.5 PROGRAMMING CONSIDERATIONS
Consider the following when programming the DMA.
• The DMA transfers data between a requester and a target. The transfer direction is
programmable and determines whether the requester or the target is the source or
destination of a transfer.
• The two-cycle data transfer bus cycle option uses a four-byte temporary buffer. During a
buffer transfer, the channel fills the temporary buffer from the source before writing any
data to the destination. Therefore, the number of bus cycles that it takes to transfer data
from the source to the destination depends on the amount of data to transfer and the source
and destination data bus widths.
• Each channel contains a 26-bit requester address, 26-bit target address, and 24-bit byte
count. These values are programmed through the use of 8-bit registers, some of which are
multiplexed at the same addresses. A byte pointer (BP) controls the access to these
multiplexed registers. After you write or read a register that requires a byte pointer
specification, the channel toggles the byte pointer. For example, writing to DMA0TAR0
12-50
DMA CONTROLLER
with BP=0 causes the DMA to set BP. The clear byte pointer software command
(DMACLRBP) allows you to force BP to a known state (0) before writing to the registers.
• The target and requester addresses are incremented, decremented, or left unchanged and the
byte count is decremented after each data transfer within a buffer transfer. Reading a
register returns the current (or modified) value rather than the original programmed values.
• The chaining buffer-transfer mode requires that you write new transfer information to the
channel before the current buffer transfer completes. The channel determines whether new
transfer information was written to it by checking the most-significant byte of the target
address. Writing to this byte sets an internal flag that tells the channel that new transfer
information was written to it. Therefore, it is only necessary to change the target address
between chaining buffer transfers. If you want to change the requester address and byte
count also, you should write these values before writing the most-significant byte of the
target address.
• If a channel is configured to increment the requester address and the requester’s bus size is
selected as 16 bits, the channel increments the requester address by two after each data
transfer. However, if the channel is configured to decrement the requester address, the
channel only decrements the address by one. This is true for the target also. In other words,
the channels cannot decrement by words. When a channel is configured to decrement the
requester or target address and transfer words, the correct number of words is transferred;
however, the transfers are on a byte basis.
• Enabling both the autoinitialize and chaining buffer-transfer modes will have unpredictable
results.
• The DMA controller does not allow programming one channel while another channel is
active. If both channels are being used, the programmer must mask an active channel before
reprogramming the other channel. Failure to do this may result in incorrect DMA transfers.
12.5.1 DMA Controller Code Examples
This section contains these software routines:
EnableDMAHWRequests
DisableDMAHWRequests
SetDMAReqIOAddr
SetDMATargMemAddr
SetDMAXferCount
InitDMA
Enables channel hardware requests for the given
DMA channel
Disables channel hardware requests for the specified
DMA channel
Sets the requester to an I/O port address for the
specified channel
Sets the target memory address for the specified DMA
channel
Sets the target memory device for the specified DMA
channel
Initializes the DMA
12-51
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
InitDMA1ForSSIXmitterToMem
DMAInterrupt
Initializes DMA channel1 for transfers between the
SIO transmitter port and memory
Interrupt Service Routine for DMA generated
interrupts
See Appendix C for included header files.
#include “80386ex.h”
#include “ev386ex.h”
#include “dma.h”
#include <DOS.h>
#pragma warning(disable:4704)
/*Disable optimization warning*/
/*****************************************************************************
EnableDMAHWRequests:
int EnableDMAHWRequests(int nChannel)
Description:
Enables channel hardware requests for the given DMA channel.
Parameters:
nChannel
--channel to enable hardware requests
Returns:
Error Code
Assumptions:
None
Syntax:
int error_code;
error_code = EnableDMAHWRequests (DMA_Channel0);
Real/Protected Mode:
No changes required
*****************************************************************************/
int EnableDMAHWRequests(int nChannel)
{
BYTE regDMAMSK = 0; /*Clear regDMAMSK[HRM]*/
/*Check input*/
if ( (nChannel != DMA_Channel0) && (nChannel != DMA_Channel1) )
return ERR_BADINPUT;
regDMAMSK = nChannel; /*Set regDMAMSK[CS] to channel*/
_SetEXRegByte(DMAMSK, regDMAMSK); /*Clear hardware request mask for*/
12-52
DMA CONTROLLER
/*
given channel*/
}
/*****************************************************************************
DisableDMAHWRequests:
Description:
Disables channel hardware requests for the given DMA channel.
The channel, however, can still receive software requests.
Parameters:
nChannel
--channel to mask hardware requests
Returns:
Error Code
Assumptions:
None
Syntax:
int error_code;
error_code = DisableDMAHWRequests(DMA_Channel0);
Real/Protected Mode:
No changes required
*****************************************************************************/
int DisableDMAHWRequests(int nChannel)
{
WORD regDMAMSK = 0;
//Check input
if ( (nChannel != DMA_Channel0) && (nChannel != DMA_Channel1) )
return ERR_BADINPUT;
regDMAMSK = nChannel;
regDMAMSK &= 0x04;
//Set regDMAMSK[CS] to channel
//Set regDMAMSK[HRM]
_SetEXRegByte(DMAMSK, regDMAMSK);
//Set hw request mask for given
//channel
return ERR_NONE;
}
/*****************************************************************************
SetDMAReqIOAddr:
12-53
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Description:
Sets the requester to an I/O port address, wIO, for the DMA
channel specified by nChannel.
Parameters:
nChannel
wIO
--channel for which to set Requester I/O port address
--I/O address
Returns:
None
Assumptions:
None
Syntax:
SetDMAReqIOAddr(DMA_Channel1, TBR0);
//Sets Req. to Serial Receiver
Real/Protected Mode:
No changes required
*****************************************************************************/
int SetDMAReqIOAddr(int nChannel, WORD wIO)
{
WORD addrDMAReq0_1;
WORD addrDMAReq2_3;
/*Check input*/
if ( (nChannel != DMA_Channel0) && (nChannel != DMA_Channel1) )
return ERR_BADINPUT;
/*Set registers to correct channel*/
addrDMAReq0_1 = ( nChannel == DMA_Channel0 ? DMA0REQ0_1 : DMA1REQ0_1);
addrDMAReq2_3 = ( nChannel == DMA_Channel0 ? DMA0REQ2_3 : DMA1REQ2_3);
_SetEXRegByte(DMACLRBP, 0x0); /* Clear the byte pointer flip-flop */
/* Write requester I/O address, bits 0-7 */
_SetEXRegByte(addrDMAReq0_1, (BYTE) (wIO & 0xFF));
/* Write requester I/O address, bits 8-15 */
_SetEXRegByte(addrDMAReq0_1, (BYTE) ((wIO >> 8) & 0xFF));
_SetEXRegByte(addrDMAReq2_3, 0x00); /* Zero requester address bits 16-23 */
_SetEXRegByte(addrDMAReq2_3, 0x00); /* Zero requester address bits 24-25 */
return ERR_NONE;
}
/*****************************************************************************
SetDMATargMemAddr:
12-54
DMA CONTROLLER
Description:
Sets the target memory address for the DMA channel specified
by nChannel.
Parameters:
nChannel
ptMemory
--channel for which to set target address
--pointer to target memory location
Returns:
None
Assumptions:
Processor is in real mode.
Syntax:
static char lpsz[]=”Hello World”;
SetDMATargMemAddr(DMA_Channel1, lpsz);
Real/Protected Mode:
The address calculation from ptMemory assumes the processor is in real
mode.
*****************************************************************************/
int SetDMATargMemAddr(int nChannel, void *ptMemory)
{
WORD addrDMATar0_1;
WORD addrDMATar2;
WORD addrDMATar3;
WORD wSegment;
WORD wOffset;
DWORD lAddress;
/*Check input*/
if ( (nChannel != DMA_Channel0) && (nChannel != DMA_Channel1) )
return ERR_BADINPUT;
/*Set registers to correct channel*/
addrDMATar0_1 = ( nChannel == DMA_Channel0 ? DMA0TAR0_1 : DMA1TAR0_1);
addrDMATar2 = ( nChannel == DMA_Channel0 ? DMA0TAR2 : DMA1TAR2);
addrDMATar3 = ( nChannel == DMA_Channel0 ? DMA0TAR3 : DMA1TAR3);
/*If in tiny, small, or medium model,*/
#if defined(M_I86TM) || defined(M_I86SM) || defined(M_I86MM)
_asm
{
/*...then grab our segment from DS*/
mov ax, ds
mov wSegment, ds
}
wOffset = (WORD) ptMemory; /*...and our offset from the pointer*/
12-55
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
#else
/*Else in compact, large, or huge memory model*/
wSegment = _FP_SEG(ptMemory); /*...grab the segment from the pointer*/
wOffset = _FP_OFF(ptMemory);
#endif
address*/
lAddress = ((DWORD) wSegment << 4) + wOffset;
/*...and the offset from the pointer*/
/*Assuming real mode, compute our physical
_SetEXRegByte(DMACLRBP, 0x0);
/*Clear the byte pointer flip-flop */
/* Write target address, bits 0-7 */
_SetEXRegByte(addrDMATar0_1, (BYTE) (lAddress & 0xFF));
/* Write target address, bits 8-15 */
_SetEXRegByte(addrDMATar0_1, (BYTE) ((lAddress >> 8) & 0xFF));
/* Write target address, bits 16-23 */
_SetEXRegByte(addrDMATar2, (BYTE) ((lAddress >> 16) & 0xFF));
/* Write target address, bits 24-25 */
_SetEXRegByte(addrDMATar3, (BYTE) ((lAddress >> 24) & 0x03));
return ERR_NONE;
}
/*****************************************************************************
SetDMAXferCount:
Description:
Sets the target memory device for the DMA channel specified
by nChannel
PARAMETERS:
nChannel
ptMemory
--channel for which to set target address
--pointer to target memory location
Returns:
None
Assumptions:
Processor is in real mode.
Syntax:
static char lpsz[]=”Hello World”;
SetDMATargMemAddr(DMA_Channel0, lpsz);
Real/Protected Mode:
The address calculation from ptMemory assumes the processor is in real
mode.
12-56
DMA CONTROLLER
*****************************************************************************/
int SetDMAXferCount(int nChannel, DWORD lCount)
{
WORD addrDMAByc0_1;
WORD addrDMAByc2;
/*Check input*/
if ( (nChannel != DMA_Channel0) && (nChannel != DMA_Channel1) )
return ERR_BADINPUT;
/*Set registers to correct channel*/
addrDMAByc0_1 = (nChannel == DMA_Channel0 ? DMA0BYC0_1 : DMA1BYC0_1);
addrDMAByc2 = (nChannel == DMA_Channel0 ? DMA0BYC2 : DMA1BYC2);
_SetEXRegByte(DMACLRBP, 0x0); /* Clear the byte pointer flip-flop */
/* Write count, bits 0-7 */
_SetEXRegByte(addrDMAByc0_1, (BYTE) ( lCount & 0xFF));
/* Write count, bits 8-15 */
_SetEXRegByte(addrDMAByc0_1, (BYTE) ((lCount >> 8) & 0xFF));
/* Write count, bits 16-23 */
_SetEXRegByte(addrDMAByc2, (BYTE) ((lCount >> 16) & 0xFF));
return ERR_NONE;
}
/*****************************************************************************
InitDMA:
Description:
Enables the DMA and initializes settings independent of the
two channels:
bus arbitration--set to no rotation, external bus master
request(HOLD) assigned to lowest priority level
EOP# sampling--set to asynch. (no effect when DMA is used
with internal peripherals)
DRQn sampling--set to synch. (no effect when DMA is used
with internal peripherals)
Parameters:
None
Returns:
None
Assumptions:
None
12-57
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Syntax:
InitDMA();
//Initialize DMA peripheral
Real/Protected Mode:
No changes required
*****************************************************************************/
void InitDMA(void)
{
_SetEXRegByte(DMACLR, 0x0);
_SetEXRegByte(DMACMD1, 0x0); /*DMACMD1[7:5]=0: reserved*/
/*DMACMD1[4]=0: disable priority rotation*/
/* enable*/
/*Resets DMA peripheral*/
/*DMACMD1[2]=0: enable channel’s 0 and 1*/
/*DMACMD1[1:0]=0: reserved*/
_SetEXRegByte(DMACMD2, 0x8); /*DMACMD2[7:4]=0: reserved*/
/*DMACMD2[3:2]=2: assign HOLD to the lowest*/
/* priority level*/
/*DMACMD2[1]=0: EOP# samples input async.*/
/*DMACMD2[0]=0: DRQn samples input async.*/
}
/*****************************************************************************
InitDMA1ForSSIXmitterToMem:
Description:
This function prepares DMA channel 1 for transfers between the async.
serial port transmitter (channel 0) and memory. After calling this
function, a DMA transfer can be initiated by setting the Target address,
setting the transfer count, and clearing the hardware request mask
for this DMA channel.
Parameters:
None
Returns:
None
Assumptions:
InitDMA() has been called to enable the peripheral.
Syntax:
static char lpsz[]=”Hello World”;
InitDMA();
InitDMA1ForSerialXmitter();
//Initialize DMA peripheral
//Initialize DMA channel 1
.
.
12-58
DMA CONTROLLER
.
SetDMATargMemAddr(DMA_Channel1, lpsz); //Set target memory address
//Set transfer count
SetDMAXferCount(DMA_Channel1, strlen(lpsz) );
EnableDMAHWRequests(DMA_Channel1);
//Begin transfer at SIO request
Real/Protected Mode:
No changes required
*****************************************************************************/
void InitDMA1ForSerialXmitter(void)
{
BYTE regDMACfg;
BYTE regDMAIE;
BYTE regDMAOvfE;
DisableDMAHWRequests(DMA_Channel1); /*Disable channel 1 Hardware requests*/
regDMACfg = (_GetEXRegByte(DMACFG) & 0x0F) | 0xA0;
_SetEXRegByte(DMACFG, regDMACfg); /*DMACFG[7]=1: mask DMA Acknowledge for*/
/*
channel 1*/
/*DMACFG[6:4]=3: set channel request to*/
/*
/*
SIO’s channel 0’s transmit buffer*/
empty signal*/
/*DMAMSK[3:0]=unmodified: channel 0*/
/* settings*/
_SetEXRegByte(DMAMOD1, 0x9); /*DMAMOD1[7:6]=0: set to demand data-xfer*/
/* mode*/
/*DMAMOD1[5]=0: increment target*/
/*DMAMOD1[4]=0: disable autoinitialize*/
/*
/*DMAMOD1[3:2]=2: data is xfer’d from targ.*/
/* to req.*/
buffer-xfer mode*/
/*DMAMOD1[1]=0: reserved*/
/*DMAMOD1[0]=1: selections for bits 7-2*/
/*
affect channel 1*/
_SetEXRegByte(DMAMOD2, 0xD1); /*DMAMOD2[7]=1: Select 2-cycle data xfer*/
/*DMAMOD2[6]=1: Requester is in I/O space*/
/*DMAMOD2[5]=0: Target is in memory space*/
/*DMAMOD2[4]=1: Requester is held constant*/
/*
/*DMAMOD2[3]=x: Req. Inc/Dec...see*/
/* DMAMOD2[4]*/
/*DMAMOD2[2]=0: Target address is*/
/* modified...see DMAMOD1[5]*/
thru xfer*/
/*DMAMOD2[1]=0: reserved*/
/*DMAMOD2[0]=1: selections for bits 7-2*/
/*
affect channel 1*/
12-59
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
_SetEXRegByte(DMABSR, 0x51); /*DMABSR[7]=0: reserved*/
/*DMABSR[6]=1: sets req.’s bus size to 8-bit*/
/*DMABSR[5]=0: reserved*/
/*DMABSR[4]=1: sets tar.’s bus size to 8-bit*/
/*DMABSR[3:2]=0: reserved*/
/*DMABSR[0]=1: selections for bits 7-2*/
/*
affect channel 1*/
_SetEXRegByte(DMACHR, 0x1); /*DMACHR[7:3]=0: reserved*/
/*DMACHR[2]=0: disable chaining buffer-xfer*/
/* mode*/
/*DMACHR[1]=0: reserved*/
/*DMACHR[0]=1: selections for bits 7-2 affect*/
/*
channel 1*/
regDMAIE = _GetEXRegByte(DMAIEN) & 0x1;
_SetEXRegByte(DMAIEN, regDMAIE); /*DMAIE[7:2]=untouched: reserved*/
/*DMAIE[1]=0: masks channel 1’s transfer*/
/*
/*
complete signal from interrupt*/
controller*/
/*DMAIE[0]=untouched: channel 0 setting*/
regDMAOvfE = _GetEXRegByte(DMAOVFE) | 0xC;
_SetEXRegByte(DMAOVFE, regDMAOvfE); /*DMAOVFE[7:4]=untouched: reserved*/
/*DMAOVFE[3]=1: all bits of channel 1
/*
/*
req. address are inc/dec*/
(see DMAMOD[4])*/
/*DMAOVFE[2]=1: all bits of channel 1*/
/* target addr. are inc/dec*/
/*DMAOVFE[1:0]=untouched: channel 0*/
/* settings*/
SetDMAReqIOAddr(DMA_Channel1, TBR0); /*Sets Req. I/O address to Serial*/
/*Receiver*/
}
/*****************************************************************************
DMAInterrupt:
Description:
This function is called by the DMA unit when it either completes a
transfer or (in chaining xfer mode) when a new requester, target, and
byte count should be written to the device.
Parameters:
None
Returns:
None
Assumptions:
12-60
DMA CONTROLLER
None
Syntax:
regDMAIE = _GetEXRegByte(DMAIEN) | 0x2; //Enable tc interrupt for
// channel 0
_SetEXRegByte(DMAIEN, regDMAIE);
//Set interrupt routine
SetIRQVector(DMAInterrupt, 12, INTERRUPT_ISR);
Enable8259Interrupt(0, IR4); //Enable slave IR4, DMA interrupt
NonSpecificEOI(); //Clear all interrupts
Real/Protected Mode:
No changes required
*****************************************************************************/
void interrupt far DMAInterrupt(void)
{
WORD regDMAIS;
regDMAIS = _GetEXRegByte(DMAIS); /*Get interrupt status register*/
if (regDMAIS & 0x10)
{
}
/*Transfer Complete, channel 0*/
_SetEXRegByte(DMACLRTC, 0x00); /*Clear transfer complete signal*/
if (regDMAIS & 0x20)
{
/*Transfer Complete, channel 1*/
_SetEXRegByte(DMACLRTC, 0x00); /*Clear transfer complete signal*/
}
if (regDMAIS & 0x1)
{
/*Chaining Interrupt, channel 0*/
/*Chaining Interrupt, channel 1*/
}
if (regDMAIS & 0x2)
{
}
NonSpecificEOI(); /*Send End-Of-Interrupt Signal to Master/Slave*/
}
12-61
13
SYNCHRONOUS
SERIAL I/O UNIT
CHAPTER 13
SYNCHRONOUS SERIAL I/O UNIT
The synchronous serial I/O (SSIO) unit provides 16-bit bidirectional serial communications. The
full-duplex communications. Either channel can originate the clocking signal or receive an exter-
nally generated clocking signal.
This chapter is organized as follows:
• SSIO Operation (page 13-5)
• Register Definitions (page 13-16)
• Design Considerations (page 13-25)
• Programming Considerations (page 13-26)
13.1 OVERVIEW
The SSIO unit contains a baud-rate generator, transmitter, and receiver. The baud-rate generator
has two possible internal clock sources (PSCLK or SERCLK). The transmitter and receiver are
double buffered. They contain 16-bit holding buffers and 16-bit shift registers. Data to be trans-
mitted is written to the transmit holding buffer. The buffer’s contents are transferred to the trans-
mit shift register and shifted out via the serial data transmit pin (SSIOTX). Data received is
shifted in via the serial data receive pin (SSIORX). Once 16 bits have been received, the contents
of the receive shift register are transferred to the receive buffer.
Both the transmitter and receiver can operate in either master or slave mode. In master mode, the
ter or receiver. If the transmitter or receiver is enabled in master mode, the baud-rate generator’s
signal appears on the transmit or receive clock pin, and is available for clocking an external slave
transmitter or receiver. In slave mode, an external master device controls the serial communica-
tions. An input on the external transmit or receive clock pin clocks the transmitter or receiver. The
transmitter and receiver need not operate in the same mode. This allows the transmitter and re-
ceiver to operate at different frequencies (an internal and an external clock source or two different
external clock sources can be used). Figures 13-1 through 13-4 illustrate the various transmit-
ter/receiver master/slave combinations.
13-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Baud-rate
Generator
Transmitter
Receiver
Clock Source
(PSCLK or SERCLK)
S
y
STXCLK
SSTBE
(to DMA controller)
s
SSIOCON.5 (TIE)
t
SSIOTX
(pin mux)
e
m
SSIOINT
(to Slave interrupt
controller IR1)
B
u
s
SRXCLK
(pin mux)
SSIOCON.1 (RIE)
SSRBF
(to DMA controller)
SSIORX
A2434-02
Figure 13-1. Transmitter and Receiver in Master Mode
Baud-rate
Generator
Clock Source
(PSCLK or SERCLK)
S
STXCLK
y
SSTBE
(to DMA controller)
Transmitter
s
SSIOCON.5 (TIE)
SSIOTX
(pin mux)
t
e
m
SSIOINT
(to Slave interrupt
controller IR1)
B
u
SSIOCON.1 (RIE)
s
Receiver
SSIORX
SSRBF
(to DMA controller)
SRXCLK
(pin mux)
A2435-02
Figure 13-2. Transmitter in Master Mode, Receiver in Slave Mode
13-2
SYNCHRONOUS SERIAL I/O UNIT
Clock Source
(PSCLK or SERCLK)
Baud-rate
Generator
STXCLK
S
y
Transmitter
SSTBE
(to DMA controller)
s
SSIOCON.5 (TIE)
SSIOCON.1 (RIE)
SSIOTX
(pin mux)
t
e
m
SSIOINT
(to Slave interrupt
controller IR1)
B
u
s
SRXCLK
(pin mux)
Receiver
SSIORX
SSRBF
(to DMA controller)
A2436-02
Figure 13-3. Transmitter in Slave Mode, Receiver in Master Mode
STXCLK
S
y
SSTBE
(to DMA controller)
Transmitter
s
SSIOCON.5 (TIE)
SSIOTX
(pin mux)
t
e
m
SSIOINT
(to Slave interrupt
controller IR1)
B
u
s
Receiver
SSIOCON.1 (RIE)
SSIORX
SSRBF
(to DMA controller)
SRXCLK
(pin mux)
A2437-02
Figure 13-4. Transmitter and Receiver in Slave Mode
13-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
13.1.1 SSIO Signals
Table 13-1 lists the SSIO signals.
Table 13-1. SSIO Signals
Device Pin or
Signal
Description
Internal Signal
STXCLK
Device pin
Serial Transmit Clock:
(input or output)
This pin functions as either an output or an input, depending on whether
the transmitter is operating in master or slave mode.
In master mode, STXCLK functions as an output. The baud-rate
generator’s output appears on this pin through the transmitter and can be
used to clock a slave receiver.
In slave mode, STXCLK functions as an input clock for the transmitter.
Serial Receive Clock:
SRXCLK
Device pin
(input or output)
This pin functions as either an output or an input, depending on whether
the receiver is operating in master or slave mode.
In master mode, SRXCLK functions as an output. The baud-rate
generator’s output appears on this pin through the receiver and can be
used to clock a slave transmitter.
In slave mode, SRXCLK functions as an input clock for the receiver.
Transmit Serial Data:
SSIOTX
SSIORX
SSRBF
SSTBE
Device pin
(output)
The transmitter uses this pin to shift serial data out of the device. Data is
transmitted most-significant bit first.
Device pin
(input)
Receive Serial Data:
The receiver uses this pin to shift serial data into the device. Data is
received most-significant bit first.
Internal signal
(output)
Receive Buffer Full:
This internal signal is used to indicate that received serial data has been
transferred from the receive shift register to the receive holding buffer.
Internal signal
(output)
Transmit Buffer Empty:
This internal signal is used to indicate that serial data has been shifted
from the transmit holding register to the transmit shift register.
SSIOINT
BCLKIN
Internal signal
(output)
SSIO Interrupt:
This internal signal goes active when either the transmit holding register
is empty or the receive holding register is full.
Internal signal
(input)
Prescaled Clock (PSCLK):
This internal signal is a prescaled value of the internal clock frequency
(CLK2/2). PSCLK is programmable for a range of divide-by values.
Serial Clock (SERCLK):
This internal signal is half the internal clock frequency (CLK2/4).
13-4
SYNCHRONOUS SERIAL I/O UNIT
13.2 SSIO OPERATION
The following sections describe the operation of the baud-rate generator, transmitter, and receiv-
er.
13.2.1 Baud-rate Generator
Either the prescaled clock or the serial clock (PSCLK or SERCLK) can drive the baud-rate gen-
erator (Figure 13-5). The SIO and SSIO configuration register (SIOCFG) selects one of these
sources.
SIOCFG.2
Baud-rate
Generator
SERCLK
CLK2
÷ 2
÷ 2
1
0
PSCLK
BCLKIN
SSIOBAUD
BV6:0
9-bit Programmable Divider
CLKPRS
A2443-02
Figure 13-5. Clock Sources for the Baud-rate Generator
CLK2
BCLKIN = SERCLK = ---------------
4
OR
CLK2/2
BCLKIN = PSCLK = -----------------------------------------------
prescale value + 2
SERCLK provides a baud-rate input frequency (BCLKIN) of CLK2/4. The PSCLK frequency
depends on the 9-bit programmable divider. The input to the programmable divider is divided by
a 9-bit prescale value + 2.
A prescale value of 0 gives the maximum PSCLK frequency (CLK2/4) and a prescale value of
1FFH (511) gives the minimum PSCLK frequency (CLK2/1026).
13-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
The baud-rate generator contains a seven-bit down counter. A programmable baud-rate value
(BV) is the reload value for the counter. The counter counts down from BV to zero, toggles the
baud-rate generator output, then reloads the BV and counts down again. The baud-rate genera-
tor’s output is a function of BV and BCLKIN as follows.
BCLKIN
baud-rate output frequency = ---------------------
2BV + 2
A BV of 0 gives the maximum output frequency (BCLKIN/2) and a BV of 7FH (127) gives the
minimum output frequency (BCLKIN/256).
BCLKIN
---------------------------------------------------------------------------------
BV =
– 1
2 × baud-rate output frequency
The maximum and minimum baud-rate output frequencies with a 33MHz (CLK2 = 66MHz) de-
vice are shown in Table 13-2.
Table 13-2. Maximum and Minimum Baud-rate Output Frequencies
Baud-rate Value
Input Frequency (BCLKIN)
Output Frequency
(BV)
0
16.5 MHz
8.25 MHz
(using either SERCLK or PSCLK with a
prescale value of 0)
7FH
64.327 KHz
251.277 Hz
(using PSCLK with a prescale value of
1FFH)
13.2.2 Transmitter
The transmitter contains a 16-bit buffer and a 16-bit shift register. When the transmitter is en-
abled, the contents of the buffer are immediately transferred to the shift register. The shift register
shifts data out via SSIOTX. Either the internal baud-rate generator (master mode) or an input sig-
nal on the STXCLK pin (slave mode) drives the transmitter. The maximum transmitter input fre-
quency is 8.25 MHz with a 33MHz processor clock (CLK2 = 66MHz). In master mode, the baud-
rate generator must be programmed and enabled prior to enabling the transmitter. In slave mode,
the transmitter must be enabled prior to the application of an external clock.
13-6
SYNCHRONOUS SERIAL I/O UNIT
13.2.2.1
Transmit Mode using Enable Bit
The transmitter contains a transmit holding buffer empty (THBE) flag and a transmit underrun
error (TUE) flag. At reset, THBE is set, indicating that the buffer is empty. Writing data to the
buffer clears THBE. When the transmitter transfers data from the buffer to the shift register,
THBE is set. If the transmitter is enabled (TEN bit is set, AUTOTXM is clear), it transfers the
new contents of the transmit buffer to the shift register each time the shift register finishes shifting
its current contents.
If the shift register finishes shifting out its current contents before a new value is written to the
transmit buffer, it reloads the old value and shifts it out again. This condition is known as a trans-
mitter underrun error. TUE is set to indicate an underrun error. For high speed transfers this can
be a problem, since the Baud-rate generator clock may be too fast; it may not allow enough time
to control the TEN bit for each word transfer. This could cause the same word to be transmitted
problem.
The transmitter also has a transmit holding buffer empty signal (SSTBE). This signal can be con-
nected to the interrupt control and DMA units. This allows you to use either an interrupt service
routine or a DMA transfer to load new data in the transmit holding buffer.
Figures 13-6 and 13-7 are simple descriptions of the SSIO transmitter state-machine when Au-
totransmit mode is enabled or disabled.
E
B
1
T
O
T
U
A
HOLD
Initialize
SSIO
THBE = 0
Data Written Into
SSIOTBUF
Clears THBE
Data In Buffer
Moved To
Shift Register
THBE Set
THBE = 1
A3400-01
Figure 13-6. SSIO Transmitter with Autotransmit Mode Enabled
13-7
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
0
T
TEN = 1
THBE = 1
HOLD
Initialize
SSIO
TEN = 1
THBE = 0
TUE
Set
Data Written Into
SSIOTBUF
Clears THBE
TUE = 1
Data In Buffer
Moved To
Shift Register
THBE Set
THBE = 1
A3399-01
Figure 13-7. SSIO Transmitter with Autotransmit Mode Disabled
The SSIO Unit can be operated either by using a polling method or through interrupts.
• Figure 13-8 shows a basic flowchart for using the polling method to transmit data through
the SSIO.
• Figure 13-9 shows a basic flowchart for the Interrupt Service Routine necessary when using
interrupts to transmit data through the SSIO. If interrupts are used, follow the below
sequence for initialization:
1. Initialize the SSIO.
2. Initialize the interrupts - ICU initialization, Interrupt Service Routine, etc.
3. Unmask the interrupts on the ICU.
13-8
SYNCHRONOUS SERIAL I/O UNIT
Initialize SSIO
Write Data to Buffer
(SSIOTBUF)
No
Yes
AUTOTXM=1
?
Enable Transmitter
TEN=1
No
THBE=1
?
Yes
Yes
AUTOTXM=1
?
No
Delay To Allow Transmitter
To Shift First Bit Out
Disable Transmitter
TEN=0
Yes
No
TUE=1
?
Error
Routine
A3394-01
Figure 13-8. Transmit Data by Polling
13-9
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
SSIO Transmitter
Causes Interrupt
Disable Interrupts While
Transmitting Data
No
Error
Routine
THBE=1
?
Yes
Write Data to Buffer
(SSIOTBUF)
Enable Transmitter
TEN=1
Delay To Allow Transmitter
To Shift First Bit Out
Disable Transmitter
TEN=0
Enable Interrupts
Exit Interrupt Service Routine
A3398-01
Figure 13-9. Interrupt Service Routine for Transmitting Data Using Interrupts
13-10
SYNCHRONOUS SERIAL I/O UNIT
If the transmitter is disabled while a data value in the shift register is being shifted out, it continues
running until the last bit is shifted out. Then the shift register stops and the data and clock pins
the shift register.
If the transmitter is disabled then re-enabled before the current value has been shifted out, it con-
tinues as if it were never disabled.
If you enable the transmitter while the baud-rate generator clock is high, the data and clock pin
states are as shown in Figure 13-10. If you enable the transmitter while the baud-rate generator
clock is low, the data and clock pin states are as shown in Figure 13-11. These figures show master
mode, single word transfers. At the end of transmission, STXCLK and SSIOTX are three-stated
and require external pull-up resistors. For single word transfers, you must enable the transmitter,
which starts the shifting process, then disable the transmitter before 16 bits are shifted out. For
high baud rates use the Autotransmit mode.
Baud-rate
Generator Clock
Transmitter Enable
Float
Float
Float
Float
STXCLK
SSIOTX
TB15
TB14
TB13
TB1
TB0
A2445-01
Figure 13-10. Transmitter Master Mode, Single Word Transfer
(Enabled when Clock is High)
Baud-rate
Generator Clock
Transmitter Enable
Float
Float
Float
Float
STXCLK
SSIOTX
TB15
TB14
TB1
TB0
A2444-01
Figure 13-11. Transmitter Master Mode, Single Word Transfer
(Enabled when Clock is Low)
13-11
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
13.2.2.2
Autotransmit Mode
Set the AUTOTXM bit (SSIOCON2.2) and the TXMM bit (SSIOCON2.1) to enable Autotrans-
mit mode. When the AUTOTXM bit is set, the word is automatically transferred to the shift reg-
ister and the THBE bit is set. In this mode the TEN bit is ignored. Once the data is transferred to
the shift register, the word is shifted out. If no new data has been written into the buffer, the trans-
mitter stops.
The Transmit Underrun Error (TUE) bit is not used in Autotransmit mode.
The Autotransmit mode eliminates the problem of controlling the TEN bit during high-speed data
transfers using polling or interrupts to move new data to the transmit buffer (SSIOTBUF).
13.2.2.3
Slave Mode
Operation in transmitter slave mode is similar to master mode, except the transmitter is clocked
from the STXCLK pin. When the transmitter is enabled any time during the STXCLK clock cy-
cle, TB15 appears on the SSIOTX pin and remains on the pin until the second falling edge of
STXCLK.
13.2.3 Receiver
The receiver contains a 16-bit holding buffer and a 16-bit shift register. When enabled, the shift
register shifts data in via the SSIORX pin. After the receiver shifts in 16 bits of data, the contents
of the shift register are transferred to the buffer. Either the internal baud-rate generator (master
mode) or an input signal on the SRXCLK pin (slave mode) can clock the receiver.
The receiver contains a receive holding buffer full flag (RHBF) and a receive overflow error flag
(ROE). At reset, RHBF is clear, indicating that the buffer is empty. When the receiver transfers
data from the shift register to the buffer, RHBF is set. Reading the buffer clears RHBF. When the
receiver is enabled, it transfers the contents of the shift register to the receive buffer each time the
shift register finishes shifting its current contents. If the shift register finishes shifting in its cur-
rent contents before the old value is read from the receive buffer, the receiver transfers the new
value into the buffer, overwriting the old value and sets the ROE flag. This condition is known as
a receive overflow error.
The receiver also has an internal receive holding buffer full signal (SSRBF). This signal can be
connected to the DMA unit for DMA initiated transfers. The SSRBF signal is also ORed with the
SSTBE signal to generate the SSIOINT signal which is sent to the interrupt controller. Before the
SSRBF signal is ORed it is masked with the Receive Interrupt Bit (RIE) in the SSIOCON1 reg-
ister. These options allow you to use either an interrupt service routine or a DMA transfer to read
data from the receive holding buffer.
13-12
SYNCHRONOUS SERIAL I/O UNIT
The SSIO Unit can be operated either by using a polling method or through interrupts.
• Figure 13-12 shows a basic flowchart for using the polling method to receive data through
the SSIO.
• Figure 13-13 shows a basic flowchart for the Interrupt Service Routine necessary when
using interrupts to receive data through the SSIO. If interrupts are used, follow the below
sequence for initialization:
1. Initialize the SSIO.
2. Initialize the interrupts - ICU initialization, Interrupt Service Routine, etc.
3. Unmask the interrupts on the ICU.
Initialize SSIO
No
RHBF=1
?
Yes
Read Data
(SSIORBUF)
Yes
No
ROE=0
?
Error
Routine
A3396-01
Figure 13-12. Receive Data by Polling
13-13
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
SSIO Receiver
Causes Interrupt
Disable Interrupts While
Receiving Data
Read Data From Buffer
(SSIORBUF)
No
Error
Routine
ROE=0
?
Yes
Enable Interrupts
Exit Interrupt Service Routine
A3397-01
Figure 13-13. Interrupt Service Routine for Receiving Data Using Interrupts
13-14
SYNCHRONOUS SERIAL I/O UNIT
If the receiver is disabled while a data value is being shifted into the shift register, it continues
running until the last bit is shifted in. Then the shift register is loaded into the buffer register, the
shift register stops and the clock pin (SRXCLK) is three-stated if in the master mode.
If the receiver is disabled then enabled before the current word has been shifted in, it continues
as if it were never disabled.
Figure 13-14 shows the serial receive data (SSIORX) pin values for a master mode, single word
transfer. For single word transfers, it is necessary to enable the receiver thus starting the shifting
process, then disable the receiver before 16 bits are shifted in.
Baud-rate
Generator Clock
Receiver Enable
Float
Float
SRXCLK
SSIORX
Ignored
Ignored
RB15
RB14
RB1
RB0
A2446-01
Figure 13-14. Receiver Master Mode, Single Word Transfer
Operation in receiver slave mode is similar to master mode, except the receiver is clocked from
the SRXCLK pin. When the receiver is enabled any time during the SRXCLK clock cycle, data
on the SSIORX pin is latched into the shift register at the next rising edge of SRXCLK. The SRX-
CLK and SSIORX pins are three-stated.
13-15
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
13.3 REGISTER DEFINITIONS
Table 13-3 list the registers associated with the SSIO and the following sections contain bit de-
scriptions for each register.
Table 13-3. SSIO Registers
Expanded
Address
Register
Function
PINCFG
F826H
Pin Configuration:
(read/write)
Connects the serial receive clock signal (SRXCLK) and the transmit serial
data signal (SSIOTX) to the package pin.
SIOCFG
(read/write)
F836H
F804H
F484H
SIO and SSIO Configuration:
Selects the baud-rate generator’s clock source, SERCLK or PSCLK.
CLKPRS
(read/write)
Clock Prescale:
Controls the frequency of PSCLK.
SSIOBAUD
(read/write)
SSIO Baud-rate Control:
Enables the baud-rate generator and determines its baud rate. In master
mode, the transmitter and receiver are clocked by the baud-rate generator.
SSIOCTR
(read only)
F48AH
F486H
SSIO Baud-rate Count Down:
Indicates whether the baud-rate generator is enabled and reflects the current
value of the baud-rate down-counter.
SSIOCON1
(read/write)
SSIO Control 1:
Enables the transmitter and receiver, indicates when the transmit buffer is
empty and the receive buffer is full. Enables or disables the transmitter or
receiver interrupts. SSIOCON1 also indicates two error conditions: the
transmit underrun and receiver overflow.
SSIOCON2
(read/write)
F488H
SSIO Control 2:
Selects whether the transmitter and receiver are in master or slave mode. In
master mode, the baud-rate generator clocks the transmitter or receiver. In
slave mode, an external master clocks the transmitter or receiver. Also
controls the enabling of the Automatic Transmit mode.
SSIOTBUF
(read/write)
F480H
F482H
SSIO Transmit Buffer:
Holds the 16-bit data word to transmit. Data is transmitted most-significant bit
first.
SSIORBUF
(read only)
SSIO Receive Buffer:
Holds the 16-bit data word received. Data is received most-significant bit first.
13-16
SYNCHRONOUS SERIAL I/O UNIT
13.3.1 Pin Configuration Register (PINCFG)
The serial receive clock (SRXCLK) and transmit serial data (SSIOTX) pins are multiplexed with
other functions. Use PINCFG bits 0 and 1 to select the pin functions.
Pin Configuration
PINCFG
Expanded Addr: F826H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
—
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7
—
Reserved. This bit is undefined; for compatibility with future devices, do
not modify this bit.
6
5
PM6
PM5
Pin Mode:
0 = Selects CS6# at the package pin.
1 = Selects REFRESH# at the package pin.
Pin Mode:
0 = Selects the coprocessor signals, PEREQ, BUSY#, and ERROR#, at
the package pins.
1 = Selects the timer control unit signals, TMROUT2, TMRCLK2, and
TMRGATE2, at the package pins.
4
3
2
1
0
PM4
PM3
PM2
PM1
PM0
Pin Mode:
0 = Selects DACK0# at the package pin.
1 = Selects CS5# at the package pin.
Pin Mode:
0 = Selects EOP# at the package pin.
1 = Selects CTS1# at the package pin.
Pin Mode:
0 = Selects DACK1# at the package pin.
1 = Selects TXD1 at the package pin.
Pin Mode:
0 = Selects SRXCLK at the package pin.
1 = Selects DTR1# at the package pin.
Pin Mode:
0 = Selects SSIOTX at the package pin.
1 = Selects RTS1# at the package pin.
Figure 13-15. Pin Configuration Register (PINCFG)
13-17
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
13.3.2 SIO and SSIO Configuration Register (SIOCFG)
Use SIOCFG bit 2 to connect either PSCLK or SERCLK to the baud-rate generator’s input
(BCLKIN).
SIO and SSIO Configuration
SIOCFG
Expanded Addr: F836H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
S1M
S0M
—
—
—
SSBSRC
S1BSRC
S0BSRC
Bit
Number
Bit
Mnemonic
Function
SIO1 Modem Signal Connections:
7
S1M
0 = Connects the SIO1 modem input signals to the package pins.
1 = Connects the SIO1 modem input signals internally.
6
S0M
SIO0 Modem Signal Connections:
0 = Connects the SIO0 modem input signals to the package pins.
1 = Connects the SIO0 modem input signals internally.
5–3
2
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
SSBSRC
SSIO Baud-rate Generator Clock Source:
0 = Connects the internal PSCLK signal to the SSIO baud-rate
generator.
1 = Connects the internal SERCLK signal to the SSIO baud-rate
generator.
1
0
S1BSRC
S0BSRC
SIO1 Baud-rate Generator Clock Source:
0 = Connects the COMCLK pin to the SIO1 baud-rate generator.
1 = Connects the internal SERCLK signal to the SIO1 baud-rate
generator.
SIO0 Baud-rate Generator Clock Source:
0 = Connects the COMCLK pin to the SIO0 baud-rate generator.
1 = Connects the internal SERCLK signal to the SIO0 baud-rate
generator.
Figure 13-16. SIO and SSIO Configuration Register (SIOCFG)
13-18
SYNCHRONOUS SERIAL I/O UNIT
13.3.3 Prescale Clock Register (CLKPRS)
Use CLKPRS to program the PSCLK frequency.
Clock Prescale Register
CLKPRS
Expanded Addr: F804H
ISA Addr:
—
(read/write)
Reset State:
0000H
15
8
0
—
—
—
—
—
—
—
PS8
PS0
7
PS7
PS6
PS5
PS4
PS3
PS2
PS1
Bit
Number
Bit
Mnemonic
Function
15–9
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
8–0
PS8:0
Prescale Value:
These bits determine the divisor that is used to generate PSCLK. Legal
values are from 0000H (divide by 2) to 01FFH (divide by 513).
divisor = PS8:0 + 2
Figure 13-17. Clock Prescale Register (CLKPRS)
13-19
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
13.3.4 SSIO Baud-rate Control Register (SSIOBAUD)
Use SSIOBAUD to enable the baud-rate generator and determine the baud-rate generator’s sev-
en-bit down counter’s reload value (BV).
SSIO Baud-rate Control
SSIOBAUD
Expanded Addr: F484H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
BEN
BV6
BV5
BV4
BV3
BV2
BV1
BV0
Bit
Number
Bit
Mnemonic
Function
7
BEN
Baud-rate Generator Enable:
Setting this bit enables the baud-rate generator. Clearing this bit disables
the baud-rate generator, clears the baud-rate count value, and forces the
baud rate clock to zero.
6–0
BV6:0
Baud-rate Value:
The baud-rate value (BV) is the reload value for the baud-rate
generator’s seven-bit down counter. The baud-rate generator’s output is
a function of BV and the baud-rate generator’s input (BCLKIN), as
follows.
BCLKIN
baud-rate output frequency (Hz) = --------------------- (Hz)
2BV + 2
If you know the desired output baud-rate frequency, you can determine
BV as follows.
BCLKIN
---------------------------------------------------------------------------------
BV =
– 1
2 × baud-rate output frequency
Figure 13-18. SSIO Baud-rate Control Register (SSIOBAUD)
13-20
SYNCHRONOUS SERIAL I/O UNIT
13.3.5 SSIO Baud-rate Count Down Register (SSIOCTR)
Read SSIOCTR to determine the status of the baud-rate generator. The down counter is reloaded
when CV6:0 reaches zero or when a new value is written to SSIOBAUD.
Baud-rate Count Down
SSIOCTR
Expanded Addr: F48AH
ISA Addr:
—
(read only)
Reset State:
00H
7
0
BSTAT
CV6
CV5
CV4
CV3
CV2
CV1
CV0
Bit
Number
Bit
Mnemonic
Function
7
BSTAT
Baud-rate Generator Status:
0 = The baud-rate generator is disabled.
1 = The baud-rate generator is enabled.
6–0
CV6:0
Current Value:
These bits indicate the current value of the baud-rate down counter.
Figure 13-19. SSIO Baud-rate Count Down Register (SSIOCTR)
13.3.6 SSIO Control 1 Register (SSIOCON1)
SSIOCON1 contains both transmit and receive control and status bits. Use the control bits to en-
able the receiver and transmitter and to connect the transmit buffer empty and receive buffer full
signals to the interrupt control unit. The status bits indicate that the transmit buffer is empty, a
transmit underrun error occurred, the receive buffer is full, or a receive overflow error occurred.
Both the transmit buffer empty and the receive buffer full signals can be connected (ORed) to the
interrupt request source (SSIOINT). When an interrupt request from this source is detected, you
can determine which signal caused the request by reading the SSIOCON1 receive buffer full and
transmit buffer empty status bits.
13-21
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
SSIO Control 1
SSIOCON1
Expanded Addr: F486H
ISA Addr:
—
(read/write)
Reset State:
C0H
7
0
TUE
THBE
TIE
TEN
ROE
RHBF
RIE
REN
Bit
Number
Bit
Mnemonic
Function
7
TUE
Transmit Underrun Error:
The transmitter sets this bit to indicate a transmit underrun error in the
TEN transfer mode. Clear this bit to clear the error flag. If a one is written
to TUE, it is ignored and TUE retains its previous value.
6
5
THBE
(read only bit)
Transmit Holding Buffer Empty:
The transmitter sets this bit when the transmit buffer contents have been
transferred to the transmit shift register, indicating that the buffer is now
ready to accept new data. Writing data to the transmit buffer clears
THBE. When this bit is clear, the buffer is not ready to accept any new
data.
TIE
Transmitter Interrupt Enable:
0 = Clearing this bit prevents the Interrupt Control Unit from sensing
when the transmit buffer is empty.
1 = Setting this bit connects the transmit buffer empty internal signal to
the Interrupt Control Unit.
4
3
TEN
ROE
Transmitter Enable:
0 = Disables the transmitter.
1 = Enables the transmitter.
Receive Overflow Error:
The receiver sets this bit to indicate a receiver overflow error. Write zero
to this bit to clear the flag.
If a one is written to ROE, the one is ignored and ROE retains its
previous value.
2
1
RHBF
(read only bit)
Receive Holding Buffer Full:
The receiver sets this bit when the receive shift register contents have
been transferred to the receive buffer.
Reading the buffer clears this bit.
Receive Interrupt Enable:
RIE
0 = Clearing this bit prevents the Interrupt Control Unit from sensing
when the receive buffer is full.
1 = Setting this bit connects the receiver buffer full internal signal to the
Interrupt Control Unit.
0
REN
Receiver Enable:
0 = Clearing this bit disables the receiver.
1 = Setting this bit enables the receiver.
Figure 13-20. SSIO Control 1 Register (SSIOCON1)
13-22
SYNCHRONOUS SERIAL I/O UNIT
13.3.7 SSIO Control 2 Register (SSIOCON2)
Use the control bits TXMM and RXMM in SSIOCON2 to put the transmitter or receiver in mas-
ter or slave mode. The AUTOTXM bit is used to determine if the TEN bit controls the transmit-
ting of the data.
SSIO Control 2
SSIOCON2
Expanded Addr: F488H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
—
—
—
—
—
AUTOTXM
TXMM
RXMM
Bit
Number
Bit
Mnemonic
Function
7–3
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
2
1
AUTOTXM
Automatic Transmit off mode for master mode
0 = Clearing this bit puts the TEN bit into normal operation
1 = Setting this bit and the TXMM bit causes TEN to be ignored. Every
time a word is loaded into the transmit shift register from the transmit
holding buffer it is transmitted out and then stops.
TXMM
Transmit Master Mode:
0 = Clearing this bit puts the transmitter in slave mode. In slave mode, an
external device controls the transmit serial communications. An input
on the STXCLK pin clocks the transmitter.
1 = Setting this bit puts the transmitter in master mode. In master mode,
the internal baud-rate generator controls the transmit serial
communications. The baud-rate generator’s output clocks the
internal transmitter and appears on the STXCLK pin.
0
RXMM
Receive Master Mode:
0 = Clearing this bit puts the receiver in slave mode. In slave mode, an
external device controls the receive serial communications. An input
on the SRXCLK pin clocks the receiver.
1 = Setting this bit puts the receiver in master mode. In master mode, the
internal baud-rate generator controls the receive serial
communications. The baud-rate generator’s output clocks the
internal receiver and appears on the SRXCLK pin.
Figure 13-21. SSIO Control 2 Register (SSIOCON2)
13-23
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
13.3.8 SSIO Transmit Holding Buffer (SSIOTBUF)
Write the data words to be transmitted to SSIOTBUF. Use the interrupt controller, DMA unit or
polling (read SSIOCON1) to determine when to write to the transmit buffer.
Transmit Holding Buffer
SSIOTBUF
Expanded Addr: F480H
ISA Addr:
—
(read/write)
Reset State:
0000H
15
8
0
TB15
TB7
TB14
TB6
TB13
TB5
TB12
TB4
TB11
TB10
TB2
TB9
TB1
TB8
TB0
7
TB3
Bit
Number
Bit
Mnemonic
Function
15–0
TB15:0
Transmit Buffer Bits:
These bits make up the next data word to be transmitted. The control
logic loads this word into the transmit shift register. The transmit shift
register shifts the bits out on the falling edge of the tranmitter clock pin.
The word is transmitted out starting with the most-significant bit (TB15).
Figure 13-22. SSIO Transmit Holding Buffer (SSIOTBUF)
13-24
SYNCHRONOUS SERIAL I/O UNIT
13.3.9 SSIO Receive Holding Buffer (SSIORBUF)
Read SSIORBUF to obtain the last data word received. Use the interrupt controller, DMA unit or
polling (read SSIOCON1) to determine when to read the receive buffer.
Receive Holding Buffer
SSIORBUF
Expanded Addr: F482H
ISA Addr:
—
(read only)
Reset State:
0000H
15
8
0
RB15
RB14
RB6
RB13
RB5
RB12
RB4
RB11
RB10
RB2
RB9
RB1
RB8
RB0
7
RB7
RB3
Bit
Number
Bit
Mnemonic
Function
15–0
RB15:0
Receive Buffer Bits:
This register contains the last word received. The receive shift register
shifts bits in with the rising edge of the receiver clock pin. Data is shifted
in starting with the most-significant bit. The control logic then transfers
the received word from the receive shift register to SSIORBUF.
Figure 13-23. SSIO Receive Holding Buffer (SSIORBUF)
13.4 DESIGN CONSIDERATIONS
The transmit buffer empty signal can be connected to the interrupt control and DMA units. How-
ever, at high baud-rates interrupt latency is too long to prevent a transmit underrun error. For these
cases, use the DMA to load the data to be transmitted into the transmit buffer.
To illustrate this point, assume the maximum input transmit baud-rate of 8.25 MHz. To prevent a
transmit underrun error, a new 16-bit data word must be written to the transmit buffer before the
transmit shift register shifts out 16 bits.
1
-------------------------
16 bits ×
= 16 × 121 ns = 1939 ns
8.25 MHz
At 33 MHz, one clock is 30 ns. The transmit buffer must be reloaded within 64 clocks (1939/30),
but interrupt latency is longer than 64 clocks. Therefore, the DMA unit is required to load the
transmit buffer.
13-25
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
13.5 PROGRAMMING CONSIDERATIONS
• When operating the transmitter in Master mode, and not in Autotransmit mode, you must
ensure that the last character to be transmitted is in the process of being shifted out before
disabling the transmitter. If the transmitter is disabled before the character has begun
shifting, the character remains in the shift register and is shifted out when the transmitter is
re-enabled. At high baud rates this can be a problem and using the Autotransmit mode is
recommended.
• The SSIO interrupt line is multiplexed with INT5. When configuring your system for SSIO-
generated interrupts, you must clear INTCFG.1 to connect the SSIO interrupt signal to the
ICU.
• The serial receive clock (SRXCLK) and transmit serial data (SSIOTX) pins are multiplexed
with other functions. Use PINCFG bits 0 and 1 to select the pin functions.
• No register programming is required for the shared signal pairs RI1#/SSIORX and
DSR1#/STXCLK. Both do not have multiplexers since one of the shared signals is a
dedicated input.
13.5.1 SSIO Example Code
This section includes these software routines:
InitSSIO
Initializes the SSIO for synchronous transfers
SSerialReadWord
SSerialWriteWord
SSIO_ISR
Polled serial read function that receives a single character
Polled serial write function that transmits a single character
Interrupt Service Routine for interrupts generated by the SSIO
Service routine for interrupts generated by the RHBF signal
Service routine for interrupts generated by the THBE signal
Service_RHBF
Service_THBE
The final code example shows an SSIO transfer in which the transmitter is interrupt-driven and
the receiver is polled. See Appendix C for included header files.
#include <conio.h>
#include “80386EX.h”
#include “EV386EX.h”
WORD value = ‘1’;
BYTE Control;
BYTE poll;
/*****************************************************************************
InitSSIO:
Description:
13-26
SYNCHRONOUS SERIAL I/O UNIT
Initialization routine for Synchronous Serial I/O Port.
Parameters:
Mode
Enables receiver and transmitter; Enables TBE and RHBF
interrupts
MasterTxRx
BaudValue
PreScale
Defines whether Tx and/or Rx are in Master Mode
Enables Baud-rate generator and sets Baud-rate Value
9-bit Clock prescale value
Returns:
None
Assumptions:
PINCFG & SIOCFG should be configured before this is called.
Prescale is only used if SIOCFG.2 is clear.
Syntax:
#define SSIO_TX_MASTR
0x2
0x1
0
// Transmit Master Mode
// Receive Master Mode
// Transmit Slave Mode
// Receive Slave Mode
// Transmit Interrupt Enable
// Transmitter Enable
// Receive Interrupt Enable
// Receiver Enable
// Enable Baud Rate Generator
// Baud Rate Clocking Source:
// SERCLK = CLK2/4
#define SSIO_RX_MASTR
#define SSIO_TX_SLAVE
#define SSIO_RX_SLAVE
#define SSIO_TX_IE
#define SSIO_TX_ENAB
#define SSIO_RX_IE
0
0x20
0x10
0x2
0x1
0x80
0x1
#define SSIO_RX_ENAB
#define SSIO_BAUD_ENAB
#define SSIO_CLK_SERCLK
#define SSIO_CLK_PSCLK
0x0
// Baud Rate Clocking Source:
// PSCLK = (CLK2/2) /
(CLKPRS+2)
InitSSIO (SSIO_TX_IE| SSIO_TX_ENAB | SSIO_RX_ENAB,
SSIO_RX_MASTR | SSIO_TX_SLAVE,
SSIO_BAUD_ENAB,
SSIO_CLK_PSCLK);
Real/Protected Mode:
No changes required.
*****************************************************************************/
void InitSSIO(BYTE Mode, BYTE MasterTxRx, BYTE BaudValue, BYTE PreScale)
{
/*** Set clocking iff either TX or RX is a master ***/
if(MasterTxRx != 0)
{
/* If 0 using PSCLK, therefore set PreScale */
if((_GetEXRegByte(SIOCFG) & BIT2MSK) == 0)
_SetEXRegByte(CLKPRS, PreScale);
13-27
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
/* Init Baud Rate Generator */
_SetEXRegByte(SSIOBAUD,BaudValue);
}
_SetEXRegByte(SSIOCON1,Mode);
_SetEXRegByte(SSIOCON2,MasterTxRx);
}/* InitSSIO */
/*****************************************************************************
SSerialReadWord:
Description:
Is a Polled serial port read function that will wait forever
or until a character has been received from the serial port.
Parameters:
MasterSlave
Defines if receiver is in Master or Slave mode
Returns:
Word read from serial port
Assumptions:
In Slave Mode, receiver must be enabled prior to this function call.
Syntax:
#define SSIO_RX_MASTR
#define SSIO_RX_SLAVE
0x1
0x0
WORD character;
character = SSerialReadWord( SSIO_RX_MASTR );
Real/Protected Mode:
No changes required.
******************************************************************************/
WORD SSerialReadWord(BYTE MasterSlave)
{
register BYTE SSControl;
if(MasterSlave == SSIO_RX_MASTR)
{
/* Save Control Register */
SSControl = _GetEXRegByte(SSIOCON1);
/* Get Control Register Ready to disable */
SSControl &= (~SSIO_RX_ENAB);// Clear the bit
/* Enable Receiver */
_SetEXRegByte(SSIOCON1, SSControl | SSIO_RX_ENAB);
/* Wait until Receive Holding Buffer is Full */
while( !(_GetEXRegByte(SSIOCON1) & SSIO_RHBF) );
13-28
SYNCHRONOUS SERIAL I/O UNIT
/* Disable Receiver */
_SetEXRegByte(SSIOCON1, SSControl);
}
else { // Slave Receiver, Receiver MUST already be Enabled
/* Wait until Receive Holding Buffer is Full */
while(!(_GetEXRegByte(SSIOCON1) & SSIO_RHBF) );
}
return (WORD)_GetEXRegWord(SSIORBUF);
}/* SSerialReadWord */
/*****************************************************************************
SSerialWriteWord:
Description:
Is a Polled serial port write function that will wait forever
or until a character has been written to the serial port.
Parameters:
Ch
Word to be written out to serial port
MasterSlave
Defines whether transmitter is Master or Slave
Returns:
None
Assumptions:
If transmitter is in Slave mode, it must already be enabled.
Syntax:
#define SSIO_TX_MASTR
#define SSIO_TX_SLAVE
0x2
0x0
char Ch = ‘a’;
SSerialWriteWord( (WORD)Ch, SSIO_TX_MASTR );
Real/Protected Mode:
No changes required.
******************************************************************************/
void SSerialWriteWord(WORD Ch,BYTE MasterSlave)
{
register BYTE SSControl;
unsigned int i;
if(MasterSlave == SSIO_TX_MASTR)
{
/* Save Control Register */
SSControl = _GetEXRegByte(SSIOCON1);
13-29
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
/* Get Control Register Ready to disable */
SSControl &= (~SSIO_TX_ENAB);
// Clear the bit
/* Set Buffer to Character */
_SetEXRegWord(SSIOTBUF,Ch);
/* Enable Transmitter */
_SetEXRegByte(SSIOCON1, SSControl | SSIO_TX_ENAB);
/* Wait until Transmit Holding Buffer is empty */
while( !(_GetEXRegByte(SSIOCON1) & SSIO_THBE) );
for(i=0;i < 4000; i++) { // Delay so transmit begins before disable
_asm {
nop
}
}
/* Disable Transmitter */
_SetEXRegByte(SSIOCON1, SSControl);
}
else
{
// Slave, Transmitter MUST already be Enabled
/* Wait until Transmit Holding Buffer is empty */
while( !(_GetEXRegByte(SSIOCON1) & SSIO_THBE) );
_SetEXRegWord(SSIOTBUF,Ch);// Write to Buffer
}
}/* SSerialWriteWord */
/******************************************************************************
SSIO_ISR:
Description:
Interrupt Service Routine for SSIO generated interrupts. This ISR
identifies the cause of the interrupt and calls the appropriate
routine.
Parameters:
None
Returns:
None
Assumptions:
It is assumed that the Slave 8259 is operating in Fully Nested Mode.
If the Slave were in SMM, a Specific EOI would have to be sent to the
13-30
SYNCHRONOUS SERIAL I/O UNIT
Slave to clear the in-service bit.
It is also assumed that the Master is not operating in AEOI, SFNM, or
SMM. If the Master were in SMM or SFNM, a Specific EOI would have to
be used. On the other hand, if the Master were operating in AEOI mode,
no EOI signal would have to be sent.
Syntax:
Not called by user
Real/Protected Mode:
No changes required.
******************************************************************************/
void interrupt far SSIO_ISR (void)
{
Control = _GetEXRegByte(SSIOCON1);
/* If THBE is set and Transmitter Interrupts are enabled */
if ((Control & SSIO_THBE) && (Control & SSIO_TX_IE)) {
Service_THBE(); // Service routine specific to THBE interrupts
}
/* Else if RHBF is set and Receiver Interrupts are enabled */
else if ((Control & SSIO_RHBF) && (Control & SSIO_RX_IE)) {
Service_RHBF(); // Service routine specific to RHBF interrupts
}
NonSpecificEOI(); // For Slave
NonSpecificEOI(); // For Master
}/* SSIO_ISR */
/*****************************************************************************
Service_RHBF:
Description:
Service Routine for SSIO interrupts generated by the RHBF signal.
Parameters:
None
Returns:
None
Assumptions:
13-31
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
None
Syntax:
Not called by user
Real/Protected Mode:
No changes required.
******************************************************************************/
void Service_RHBF(void)
{
WORD buffer;
buffer = _GetEXRegWord(SSIORBUF);
/* Display received character on the screen */
SerialWriteChar(SIO_0, (BYTE)buffer);
}/* Service_RHBF */
/*****************************************************************************
Service_THBE:
Description:
Service routine for SSIO interrupts generated by THBE signal.
Parameters:
None
Returns:
None
Assumptions:
None
Syntax:
Not called by user
Real/Protected Mode:
No changes required.
******************************************************************************/
void Service_THBE(void)
{
int i;
if ( value <= ‘9’) {
13-32
SYNCHRONOUS SERIAL I/O UNIT
_SetEXRegWord(SSIOTBUF, value);
value++;
}
else {
/* Disable Transmitter and Transmitter interrupts */
for(i=0;i < 4000; i++) { // Delay so transmit begins before disable
_asm {
nop
}
}
_SetEXRegByte(SSIOCON1,_GetEXRegByte(SSIOCON1) & 0xcf); // Clear TEN, TIE
}
} /* Service_THBE */
/*****************************************************************************
Example Code showing SSIO transfer in which the transmitter is
interrupt-driven and the receiver is polled:
InitSSIO(SSIO_RX_ENAB | SSIO_TX_ENAB | SSIO_TX_IE,
SSIO_TX_MASTR | SSIO_RX_SLAVE, 0xF0, 0);
// Setup SSIO interrupts
_SetEXRegByte(INTCFG, _GetEXRegByte(INTCFG) & 0xfd); // Slave IR1 is
// multiplexed
SetIRQVector(SSIO_ISR, 9, INTERRUPT_ISR); // SSIO IR will be generated
// on Slave IR1
Disable8259Interrupt(IR1+IR3+IR4+IR5+IR6+IR7, IR0+IR2+IR3+IR4+IR5+IR6+IR7);
Enable8259Interrupt(IR2,IR1); // Enable slave interrupt to master(IR2),
// Enable slave IR1
_enable(); // Enable Interrupts
// Initialize SSIO Ports
_SetEXRegByte(PINCFG, _GetEXRegByte(PINCFG) & 0xfc);
_SetEXRegByte(SIOCFG, _GetEXRegByte(SIOCFG) & 0xfb);
// Fill up transmit buffer with first character
_SetEXRegWord(SSIOTBUF, ‘a’);
// Use Polled SSIO receiver function to receive character
while ( input < ‘z’ ) {
input = SSerialReadWord(SSIO_RX_SLAVE);
SerialWriteChar(SIO_0, (BYTE)input); // Print to screen
}
******************************************************************************/
13-33
14
CHIP-SELECT
UNIT
CHAPTER 14
CHIP-SELECT UNIT
The Chip-select Unit (CSU) of the processor can be used to eliminate external address and bus-
cycle decoders in your system. The chip-selects generated by this unit can simplify external “glue
logic” by providing signals that can be connected directly to the chip-enable inputs of external
memory and I/O devices. If a particular device or address region does not require a chip-enable
signal, a chip-select region can be programmed only to enable termination of accesses to that
region. A chip-select region can also be programmed to generate a chip-enable signal and
terminate accesses to that region.
riety of devices. Each channel can operate in either 16-bit or 8-bit bus mode, generate up to 31
This chapter is organized as follows:
• CSU Operation (page 14-2)
• Register Definitions (page 14-13)
• Design Considerations (page 14-21)
• Programming Considerations (page 14-22)
14.1 OVERVIEW
Each chip-select channel consists of address and mask registers and an output signal. The address
and mask registers allow you to define memory or I/O address blocks for each channel. You also
specify whether or not the chip-select is activated when the processor is operating in system man-
agement mode. When the processor accesses a channel’s address block, the CSU activates the
channel’s output signal. Connecting a channel’s output to a memory or I/O device simplifies
memory and I/O interfacing by removing the need for and delay of decoding addresses externally.
NOTE
Chip-select channels are not activated during interrupt acknowledge cycles
and halt and shutdown cycles.
14-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
14.2 CSU UPON RESET
Upon reset of the processor, only the UCS channel is enabled and all other chip-selects are dis-
abled. UCS is enabled for the entire memory space of the processor.
The UCS region is initialized upon reset with the following settings:
• Mask set to 7FFFH (CM25:11 in UCSMSKH and UCSMSKL registers)
• CMSMM set
• 16-bit bus size
• Memory access
• External READY# ignored
• 15 wait states
With all the UCS mask bits set to 1, the UCS# pin is active for the entire 64 MBytes of the pro-
cessor’s memory address space. The UCS region can be programmed for a smaller size during
initialization. Normally, UCS# is used to select non-volatile memory devices, such as ROM and
FLASH, at the top of the memory address space so that the processor can fetch the first instruction
from address 3FFFFF0H after RESET. If the Port92 CPU-only RESET is used (described in
Chapter 5), the UCS channel must remain enabled for the top of the memory address space (a
CPU-only RESET does not affect the chip-select registers) and therefore, the UCS channel does
not re-initialize to its reset state.
14.3 CSU OPERATION
Each chip-select channel functions independently. The following sections describe chip-select
channel address blocks, system management mode support, and bus cycle length and bus size
control.
A 15-bit channel address and mask are used to specify a channel’s active address block. When
the processor accesses an address in memory or I/O, the upper 15 bits of the address are compared
to the chip-select channel address and OR’d with the channel mask. This means that the CSU
compares the channel address and ORs the channel mask to A25:11 for memory addresses and
A15:1 for I/O addresses. Ones in the channel’s mask exclude the corresponding bits from address
comparisons. Figure 14-1 shows the logic for determining address equality.
14-2
CHIP-SELECT UNIT
15-bit Channel Address
bit x
Address
bit x
Chip-select
Channel Output
bit x
15-bit Channel Mask
A2533-01
Figure 14-1. Channel Address Comparison Logic
The lower address bits are excluded from address comparisons (only 15 bits are compared). For
memory addresses which have 26-bit addresses, the minimum channel address block size is 2
Kbytes; for I/O addresses with 16-bit addresses, the minimum channel address block size is 2
bytes.
NOTE
The starting address of any channel address block must be a multiple of the
block size. For example, a 256 Kbyte block can only start at an address that is
a multiple of 256 Kbytes (0H, 4000H, 8000H, etc.).
Because you can set ones in the channel mask to exclude certain address bits from comparisons,
you can increase the size of a channel’s address block (by powers of 2 in Kbytes for memory ad-
dresses and by powers of 2 in bytes for I/O addresses). Figure 14-2 illustrates how memory ad-
dress block sizes are determined from the channel’s mask; the concept is the same for I/O address
block sizes (replace Kbyte with byte). As shown in Figure 14-2, the bit location of the right-most
zero in the channel mask determines the channel’s active address block size.
14-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
15-bit Channel Mask
Block Size
1
15
1
X
X
X
X X X X X X X X X X X X X 0
2 = 2 Kbyte
2
X X X X X X X X X X X X 0 1
X X X X X X X X X X X 0 1 1
2 = 4 Kbyte
3
2 = 8 Kbyte
15
0
1
1
1
1 1
1 1
1 1 1 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1 1
2
= 32768 Kbyte
16
2
= 65536 Kbyte
A2534-01
Figure 14-2. Determining a Channel’s Address Block Size
Any ones that are to the left of the right-most zero determine the number of blocks and the loca-
tions where the blocks are repeated. This is best illustrated by the following four examples. The
examples assume the channel is configured for memory addresses; however, the concepts dis-
cussed also apply to I/O-configured channels.
14-4
CHIP-SELECT UNIT
Example 1
This example establishes a single 32-Kbyte address block starting at 1340000H (a 32-Kbyte
boundary). In this example, the 15-bit channel address is the starting address of the channel’s ac-
tive address block (because there are no 1’s in the channel mask where there are 1’s in the channel
address)
.
15
0 1 0 0 1 1 0 1 0 0 0 0 0 0 0
1
15-bit Channel Address
15-bit Channel Mask
0 0 0 0 0 0 0 0 0 0 0 1 1 1 1
25
0 1 0 0 1 1 0 1 0 0 0 X X X X
0
Channel Active Address
X X X X X X X X X X X
Because the least-significant 0 in the channel’s mask is in bit position 5, this channel’s active ad-
dress block size is 25 = 32 Kbytes. Because there are no 1’s to the left of the right-most 0 in the
channel’s mask, the block is not repeated.
1347FFFH
Active
1340000H
14-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Example 2
This example establishes four 4-Kbyte address blocks starting at 0000000H, 0002000H,
0004000H, and 0006000H (4-Kbyte boundaries).
15
1
15-bit Channel Address
15-bit Channel Mask
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 1 1 0 1
25
0
Channel Active Address
0 0 0 0 0 0 0 0 0 0 0 X X 0 X
X X X X X X X X X X X
Because the least-significant 0 in the channel’s mask is in bit position 2, this channel’s active ad-
dress block size is 22 = 4 Kbytes. Because there are two 1’s to the left of the right-most 0 in the
channel’s mask, the block is repeated 22 = 4 times. Also, because there are no 1’s in the channel
mask where there are 1’s in the channel address, the channel address is the starting address of the
lowest active address block. In this example, each active 4-Kbyte address block in memory is fol-
lowed by an inactive 4-Kbyte address block and each active address block starts on a 4-Kbyte
address boundary.
Maximum
Memory Address
0007000H
Active
Active
Active
Active
0006FFFH
0006000H
0005FFFH
0005000H
0004FFFH
0004000H
0003FFFH
0003000H
0002FFFH
0002000H
0001FFFH
0001000H
0000FFFH
0000000H
14-6
CHIP-SELECT UNIT
Example 3
This example establishes four 2-Kbyte address blocks starting at 2413000H, 2433000H,
2613000H, and 2633000H.
15
1 0 0 1 0 0 0 0 0 1 0 0 1 1 0
1
15-bit Channel Address
15-bit Channel Mask
0 0 0 0 1 0 0 0 1 0 0 0 0 0 0
25
1 0 0 1 X 0 0 0 X 1 0 0 1 1 0
0
Channel Active Address
X X X X X X X X X X X
Because the least-significant 0 in the channel’s mask is in bit position 1, this channel’s active ad-
dress block size is 21 = 2 Kbytes. Because there are two 1’s to the left of the right-most 0 in the
channel’s mask, the address block is repeated 22 = 4 times. Also, because there are no 1’s in the
channel mask where there are 1’s in the channel address, the channel address is the starting ad-
dress of the lowest active address block. In this example, each active 2-Kbyte address block in
memory is followed by an inactive 2-Kbyte address block and each active address block starts at
a 2-Kbyte boundary.
14-7
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Maximum Memory
Address
2633800H
Active
Active
Active
Active
26337FFH
2633000H
2613FFFH
2613800H
26137FFH
2613000H
2433FFFH
2433800H
24337FFH
2433000H
2432FFFH
2413800H
24137FFH
2413000H
14-8
CHIP-SELECT UNIT
Example 4
This example establishes two 16-Kbyte address blocks starting at 0E08000H and 0E28000H (16-
Kbyte boundaries).
15
0 0 1 1 1 0 0 0 1 0 1 0 0 0 0
1
15-bit Channel Address
15-bit Channel Mask
0 0 0 0 0 0 0 0 1 0 0 0 1 1 1
25
0 0 1 1 1 0 0 0 X 0 1 0 X X X
0
Channel Active Address
X X X X X X X X X X X
Because the least-significant 0 in the channel mask is in bit position 4, this channel’s active ad-
dress block size is 24 = 16 Kbytes. Because there is one 1 to the left of the right-most 0 in the
channel mask, the address block is repeated 21 = 2 times. Unlike the other examples, there is a 1
in the channel mask where there is a 1 in the channel address. For this reason, the channel address
is not the starting address of the lowest active address block. In this example, each active 16-
Kbyte address block is followed by an inactive 16-Kbyte address block and each block starts at
a 16-Kbyte address boundary.
Maximum
Memory Address
0E2C000H
0E2BFFFH
Active
0E28000H
0E0C000H
0E0BFFFH
Active
0E08000H
14-9
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
14.3.2 System Management Mode Support
The processor supports four operating modes: system management mode (SMM), protected, real
and virtual-86 mode. In order for a system to operate correctly in SMM, it must meet several re-
quirements. The CSU provides support for some of these requirements. To use SMM, you must
set aside a partition of memory, called SMRAM, for the SMM driver. SMRAM must meet the
following conditions:
• Located at 38000H–3FFFFH (32 Kbytes)
• Accessible only when the processor is in SMM during normal operation
• Accessible during system initialization when the processor is not in SMM
The CSU allows you to specify an address block and control whether or not the chip-select is ac-
tivated while the processor is in SMM. While in SMM (with CASMM=1), the chip-select is ac-
tive only when the processor is the bus master, such as when the processor is not in a hold state.
Refer to Chapter 7 (“Programming Considerations” on page 7-16) for a code example of pro-
gramming chip-selects to support SMM.
14-10
CHIP-SELECT UNIT
14.3.3 Bus Cycle Length Control
Each chip-select channel controls how bus cycles to its address block terminate. Each channel can
generate up to 31 wait states and then unconditionally terminate or wait for an external bus ready
signal to terminate. If the channel is programmed for wait states and to sample external READY#,
the external READY# is ignored until the programmed number of wait states has been inserted
into the cycle. If greater than 31 wait states are required, ready must be generated externally, and
the external READY# option must be selected.
NOTE
When a chip-select region overlaps on-chip peripheral addresses, the on-chip
peripheral always generates READY# and overrides the channel’s
configuration.
14.3.4 Bus Size Control
The processor assumes that the currently addressed device requires a 16-bit data bus unless the
bus size control pin (BS8#) is asserted. When asserted, BS8# tells the processor that the addressed
device requires an 8-bit data bus. You can program a chip-select channel specifically for 8-bit de-
14.3.5 Overlapping Regions
You can configure CSU channels to have overlapping address blocks. When channels with over-
lapping address blocks have different bus cycle length and bus size configurations, the CSU must
adjust these parameters. Figure 14-3 shows how the CSU adjusts the bus cycle length. In the case
of different bus sizes, the CSU defaults to an 8-bit bus size.
If one overlapping chip-select region has the RDY bit set and the other overlapping region does
not, the CSU defaults to the ‘RDY Bit Set’ operation; in this case an external READY# is neces-
sary to terminate accesses to the address locations in which the two chip-selects overlap.
NOTE
If a bus cycle address activates multiple overlapping CSU channels, all the
enabled chip-select signals of those channels go active. To avoid contention on
the data bus, care must be taken when using these chip-select signals
externally.
14-11
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Is any
channel
dependent on
external
Wait for smallest number
of all overlapping regions'
wait state values.
Yes
ready?
No
Wait
State
READY#
asserted?
Wait for largest number of
all overlapping regions'
wait state values.
No
Yes
Complete bus cycle.
A2392-02
Figure 14-3. Bus Cycle Length Adjustments for Overlapping Regions
14-12
CHIP-SELECT UNIT
14.4 REGISTER DEFINITIONS
Table 14-1 and Table 14-2 list the signals and registers associated with the chip-select unit. There
are seven general-purpose chip-select channels (CSn) and one upper chip-select channel (UCS).
Upon reset, the UCS is enabled with the entire 64 Mbyte memory address space as its address
block. The UCS can be used to select a memory device at the top of the memory address space
so that the processor can fetch the first instruction from address 3FFFFF0H after reset.
Table 14-1. CSU Signals
Device Pin or Internal
Signal
Description
Signal
CS6:0#
UCS#
Device pins
(output)
Chip-select Signals:
Indicates that the memory or I/O address that the
processor is accessing is in channel n’s active address
region.
14-13
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table 14-2. CSU Registers
Expanded
Register
PINCFG
Description
Address
0F826H
Pin Configuration:
(read/write)
Connects the CS6:5# signals to package pins.
P2CFG
0F822H
Port 2 Configuration:
(read/write)
Connects the CS4:0# signals to package pins.
CS0ADH
CS1ADH
CS2ADH
CS3ADH
CS4ADH
CS5ADH
CS6ADH
UCSADH
(read/write)
0F402H
0F40AH
0F412H
0F41AH
0F422H
0F42AH
0F432H
0F43AH
Chip-select High Address:
Defines the upper 10 bits of the chip-select channel address. The
processor uses a chip-select’s channel address to determine the starting
location of the channel’s active address block.
CS0ADL
CS1ADL
CS2ADL
CS3ADL
CS4ADL
CS5ADL
CS6ADL
UCSADL
(read/write)
0F400H
0F408H
0F410H
0F418H
0F420H
0F428H
0F430H
0F438H
Chip-select Low Address:
Defines the lower 5 bits of the chip-select channel address. Configures
the channel for memory or I/O addresses, determines whether or not the
channel is activated when the processor is operating in system
management mode, configures the channel’s bus size, defines the
minimum number of wait states inserted into the bus cycle, and defines
whether an external READY# is required to terminate the bus cycle.
CS0MSKH
CS1MSKH
CS2MSKH
CS3MSKH
CS4MSKH
CS5MSKH
CS6MSKH
UCSMSKH
(read/write)
0F406H
0F40EH
0F416H
0F41EH
0F426H
0F42EH
0F436H
0F43EH
Chip-select High Mask:
Defines the upper 10 bits of the chip-select channel mask. The processor
uses a chip-select’s channel mask to determine the size of the channel’s
active address block and if the address block is repeated.
CS0MSKL
CS1MSKL
CS2MSKL
CS3MSKL
CS4MSKL
CS5MSKL
CS6MSKL
UCSMSKL
(read/write)
0F404H
0F40CH
0F414H
0F41CH
0F424H
0F42CH
0F434H
0F43CH
Chip-select Low Mask:
Defines the lower 5 bits of the chip-select channel mask and enables the
channel’s output pin.
14-14
CHIP-SELECT UNIT
14.4.1 Pin Configuration Register (PINCFG)
Use PINCFG bits 6 and 4 to connect the CS6# and CS5# signals to package pins.
Pin Configuration
PINCFG
Expanded Addr: F826H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
—
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7
—
Reserved. This bit is undefined; for compatibility with future devices, do
not modify this bit.
6
5
PM6
PM5
Pin Mode:
0 = Selects CS6# at the package pin.
1 = Selects REFRESH# at the package pin.
Pin Mode:
0 = Selects the coprocessor signals, PEREQ, BUSY#, and ERROR#, at
the package pins.
1 = Selects the timer control unit signals, TMROUT2, TMRCLK2, and
TMRGATE2, at the package pins.
4
3
2
1
0
PM4
PM3
PM2
PM1
PM0
Pin Mode:
0 = Selects DACK0# at the package pin.
1 = Selects CS5# at the package pin.
Pin Mode:
0 = Selects EOP# at the package pin.
1 = Selects CTS1# at the package pin.
Pin Mode:
0 = Selects DACK1# at the package pin.
1 = Selects TXD1 at the package pin.
Pin Mode:
0 = Selects SRXCLK at the package pin.
1 = Selects DTR1# at the package pin.
Pin Mode:
0 = Selects SSIOTX at the package pin.
1 = Selects RTS1# at the package pin.
Figure 14-4. Pin Configuration Register (PINCFG)
14-15
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
14.4.2 Port 2 Configuration Register (P2CFG)
Use P2CFG bits 4–0 to connect the CS4:0# signals to package pins.
Port 2 Configuration
P2CFG
Expanded Addr: F822H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
PM7
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7
PM7
Pin Mode:
0 = Selects P2.7 at the package pin.
1 = Selects CTS0# at the package pin.
6
5
4
3
2
1
0
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Pin Mode:
0 = Selects P2.6 at the package pin.
1 = Selects TXD0 at the package pin.
Pin Mode:
0 = Selects P2.5 at the package pin.
1 = Selects RXD0 at the package pin.
Pin Mode:
0 = Selects P2.4 at the package pin.
1 = Selects CS4# at the package pin.
Pin Mode:
0 = Selects P2.3 at the package pin.
1 = Selects CS3# at the package pin.
Pin Mode:
0 = Selects P2.2 at the package pin.
1 = Selects CS2# at the package pin.
Pin Mode:
0 = Selects P2.1 at the package pin.
1 = Selects CS1# at the package pin.
Pin Mode:
0 = Selects P2.0 at the package pin.
1 = Selects CS0# at the package pin.
Figure 14-5. Port 2 Configuration Register (P2CFG)
14-16
CHIP-SELECT UNIT
14.4.3 Chip-select Address Registers
The Address Register of each chip-select channel defines the address block that the channel re-
sponds to during an access. The value in this register is compared to A25:11 of the processor bus
during a memory access and to A15:1 during an I/O access. A bus cycle whose address matches
the non-masked (see “Chip-select Mask Registers” on page 14-19) bits of the Address Register
causes the respective chip-select channel to have an address match. Even if there is an address
match, whether or not the CSU activates the channel depends on the values of the channel’s SMM
address and mask bits (CASMM and CMSMM) and the chip-select channel enable bit (CSEN).
The CASMM and CMSMM bits determine whether or not the channel is activated when the pro-
cessor is operating in SMM.
Write a channel’s 15-bit address to the chip-select address registers. These bits are masked by the
channel’s 15-bit mask.
NOTE
When a chip-select channel is activated, it either asserts a chip-select signal,
controls wait states and READY# generation, or both.
Chip-select High Address
CSnADH (n = 0–6), UCSADH
(read/write)
Expanded Addr: F402H, F40AH
F412H, F41AH
F422H, F42AH
F432H, F43AH
ISA Addr:
—
Reset State:
0000H (CSnADH)
FFFFH (UCSADH)
15
8
0
—
—
—
—
—
—
CA15
CA7
CA14
7
CA13
CA12
CA11
CA10
CA9
CA8
CA6
Bit
Number
Bit
Mnemonic
Function
15–10
9–0
—
Reserved; for compatibility with future devices, write zeros to these bits.
Chip-select Channel Address Upper Bits:
CA15:6
Defines the upper 10 bits of the channel’s 15-bit address. The address
bits CA15:6 and the mask bits CM15:6 form a masked address that is
compared to memory address bits A25:16 or I/O address bits A15:6.
Figure 14-6. Chip-select High Address Register (CSnADH, UCSADH)
14-17
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Chip-select Low Address
CSnADL (n = 0–6), UCSADL
(read/write)
Expanded Addr: F400H, F408H
F410H, F418H
F420H, F428H
F430H, F438H
—
ISA Addr:
Reset State:
0000H (CSnADL)
FF6FH (UCSADL)
15
8
0
CA5
CA4
—
CA3
—
CA2
CA1
CASMM
WS2
BS16
WS1
MEM
7
RDY
WS4
WS3
WS0
Bit
Number
Bit
Mnemonic
Function
15–11
CA5:1
Chip-select Address Value Lower Bits:
Defines the lower 5 bits of the channel’s 15-bit address. The address bits
CA5:1 and the mask bits CM5:1 form a masked address that is compared to
memory address bits A15:11 or I/O address bits A5:1.
10
CASMM
SMM Address Bit:
If this bit is set (and unmasked), the CSU activates the chip-select channel
only while the processor is in SMM (and not in a hold state). Otherwise, the
CSU activates the channel only when processor is operating in a mode
other than SMM.
Setting the SMM mask bit in the channel’s mask low register masks this bit.
When this bit is masked, an address match activates the chip-select,
regardless of whether the processor is in SMM or not.
9
BS16
Bus Size 16-bit:
0 = All bus cycles to addresses in the channel’s address block are byte-
wide.
1 = Bus cycles are 16 bits unless the bus size control pin (BS8#) is
asserted.
8
7
MEM
RDY
Bus Cycle Type:
0 = Configures the channel for an I/O addresses
1 = Configures the channel for memory addresses
Bus Ready Enable:
0 = External READY# is ignored. READY# generated by CSU to terminate
the bus cycle.
1 = Requires that external READY# be active to complete a bus cycle. This
bit must be set to extend wait states beyond the number determined by
WS4:0 (see “Bus Cycle Length Control” on page 14-11).
6–5
4–0
—
Reserved; for compatibility with future devices, write zeros to these bits.
Wait State Value:
WS4:0
WS4:0 defines the minimum number of wait states inserted into the bus
cycle. A zero value means no wait states.
Figure 14-7. Chip-select Low Address Register (CSnADL, UCSADL)
14-18
CHIP-SELECT UNIT
14.4.4 Chip-select Mask Registers
The Mask Register of each chip-select region is used to prevent bits from being compared with
the starting address, thus masking them from the comparison. This masking allows you to specify
the size of the region being defined. The mask should be set such that it masks the lower address
bits being compared, up to the size that you would like the block to be.
Write a channel’s 15-bit mask to the chip-select mask registers. Also, use the chip-select low
mask register to enable the channel and to mask the channel’s SMM address bit. When the chan-
nel’s SMM address bit is masked, the CSU activates the channel even if the channel is operating
in SMM.
Chip-select High Mask
CSnMSKH (n = 0–6), UCSMSKH
(read/write)
Expanded Addr: F406H, F40EH
F416H, F41EH
F426H, F42EH
F436H, F43EH
ISA Addr:
—
Reset State:
0000H (CSnMSKH)
FFFFH (UCSMSKH)
15
8
0
—
—
—
—
—
—
CM15
CM7
CM14
CM6
7
CM13
CM12
CM11
CM10
CM9
CM8
Bit
Number
Bit
Mnemonic
Function
15–10
9–0
—
Reserved; for compatibility with future devices, write zeros to these bits.
Mask Value Upper Bits:
CM15:6
Defines the upper 10 bits of the channel’s 15-bit mask. The mask bits
CM15:6 and the address bits CA15:6 form a masked address that is
compared to memory address bits A25:16 or I/O address bits A15:6.
Figure 14-8. Chip-select High Mask Registers (CSnMSKH, UCSMSKH)
14-19
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Chip-select Low Mask
CSnMSKL (n = 0–6), UCSMSKL
(read/write)
Expanded Addr: F404H, F40CH
F414H, F41CH
F424H, F42CH
F434H, F43CH
—
ISA Addr:
Reset State:
0000H (CSnMSKL)
FFFFH (UCSMSKL)
15
8
0
CM5
CM4
—
CM3
—
CM2
—
CM1
CMSMM
—
—
—
7
—
—
—
CSEN
Bit
Number
Bit
Mnemonic
Function
Chip-select Mask Value Lower Bits:
15–11
CM5:1
Defines the lower 5 bits of the channel’s 15-bit mask. The mask bits
CM5:1 and the address bits CA5:1 form a masked address that is
compared to memory address bits A15:11 or I/O address bits A5:1.
10
CMSMM
SMM Mask Bit:
0 = The SMM address bit is not masked.
1 = Masks the SMM address bit in the channel’s Chip-Select Low
Address register. When the SMM address bit is masked, an address
match activates the chip-select, regardless of whether the processor
is in SMM.
9–1
0
—
Reserved; for compatibility with future devices, write zeros to these bits.
Chip-select Enable:
CSEN
0 = Disables the chip-select channel.
1 = Enables the chip-select channel.
Figure 14-9. Chip-select Low Mask Registers (CSnMSKL, UCSMSKL)
14-20
CHIP-SELECT UNIT
14.5 DESIGN CONSIDERATIONS
When designing with the CSU, consider the following:
• Upon reset, UCS# is configured as a 16-bit chip-select signal. If the Boot device is only an
8-bit device, then BS8# must be asserted whenever UCS# is active (until the UCS channel
can be reprogrammed to reflect an 8-bit region). One way of doing this is by connecting the
UCS# pin directly to the BS8# pin, if there are no other devices that need to use the BS8#
pin. If UCS# is tied directly to BS8#, then the UCS channel need not be programmed to
reflect an 8-bit region.
• If the Port92 CPU-only RESET is used (described in Chapter 5), the UCS channel must
remain enabled for the top of the memory address space (a CPU-only RESET does not
affect the chip-select registers) and therefore, the UCS channel does not re-initialize to its
reset state.
• If arbitrary chip-select regions are required to access external memory and I/O devices and
a single channel can not be programmed to accommodate the address space of these
regions, multiple chip-select signals can be “ORed” to create a single chip-enable to a
device. For example a 512 Kbyte region chip-select signal starting on a 256 Kbyte boundary
can be created by “ORing” two 256 Kbyte chip-select signals.
• Refer to Chapter 6 (“Design Considerations” on page 6-38) for examples of using chip-
select signals to access external devices.
14-21
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
14.6 PROGRAMMING CONSIDERATIONS
When programming the CSU, consider the following:
This ensures that all other bits are properly programmed before the region is enabled. When
reprogramming the channel, always disable the channel before changing anything else.
• A chip-select channel is enabled by setting bit 0 of its Chip-Select Low Mask register and
its output signal is connected to the package pin by setting or clearing the appropriate
PINCFG or P2CFG register bit. The PINCFG and P2CFG registers are shown in Figures
14-4 and 14-5.
• The minimum address block for memory address-configured channels is 2 Kbytes and for
I/O address-configured channels is 2 bytes. The size of these address blocks can be
increased by powers of 2 Kbytes for memory addresses and by powers of 2 bytes for I/O
addresses.
• A channel’s address block of size n always starts on an n address boundary.
14.6.1 Chip-Select Unit Code Example
This following code example initializes the UCS and CS4 channels of the CSU. See Appendix C
for the included header files.
#include <conio.h>
#include “80386ex.h”
#include “EV386EX.h”
/*
Description:
Initialize Chip Select Unit for:
UCS:
Start address is 00H.
Region size is 512 Kbytes.
0 wait states.
Upper chip select is Enabled.
16 bit data bus size in memory space.
External bus ready is Disabled.
SMM region is accessible during SMI access and memory access.
CS4:
Start address is 080000H.
Region size is 512 Kbytes.
0 wait states.
Chip select 4 is Enabled.
16 bit data bus size in memory space.
External bus ready is Disabled.
SMM region is accessible during SMI access only.
Parameters:
None
14-22
CHIP-SELECT UNIT
Assumptions:
REMAPCFG register has Expanded I/O space access enabled (ESE bit set).
*/
void Init_CSU(void)
{
_SetEXRegWord(UCSADL, 0x700); /* Configure the upper chip select */
_SetEXRegWord(UCSADH, 0x0);
_SetEXRegWord(UCSMSKL, 0xFC01);
_SetEXRegWord(UCSMSKH, 0x7);
_SetEXRegWord(CS4ADL, 0x300); /* Configure chip select 4 */
_SetEXRegWord(CS4ADH, 0x8);
_SetEXRegWord(CS4MSKL, 0xF801);
_SetEXRegWord(CS4MSKH, 0x7);
}
14-23
15
REFRESH
CONTROL UNIT
CHAPTER 15
REFRESH CONTROL UNIT
The Refresh Control Unit (RCU) simplifies the interface between the processor and a dynamic
random access memory (DRAM) device by providing a way to generate periodic refresh requests
and refresh addresses. These refresh requests and addresses can then be used by an external
DRAM controller to generate the appropriate DRAM signals and addresses needed to perform
refresh operations. The RCU can be used in conjunction with the Chip-select Unit to generate
chip select signals for DRAM regions; these signals can be used by the external DRAM controller
to initiate refresh cycles.
This type of memory has an interface similar to a static random access memory (SRAM), but re-
quires a periodic refresh similar to DRAM.
This chapter is organized as follows:
• RCU Operation (page 15-5)
• Register Definitions (page 15-6)
• Design Considerations (page 15-11)
• Programming Considerations (page 15-14)
15.1 DYNAMIC MEMORY CONTROL
Typical DRAM devices require control logic to enable read, write, and refresh operations. The
RCU simplifies control logic design requirements by providing the necessary cell access require-
ments for refresh operations.
DRAM devices are built as matrices of memory cells. Therefore, each memory cell has a row and
column address associated with it. A typical controller design strobes addresses into a DRAM de-
vice through the use of two control lines: a row address strobe (RAS#) and a column address
strobe (CAS#). The controller presents lower (or row) address bits during RAS# and upper (or
column) address bits during CAS#. Activating RAS# accesses all cells within the specified row.
Accessing a cell refreshes it; therefore, cycling through the row addresses refreshes a DRAM de-
vice.
15.1.1 Refresh Methods
There are two common methods for refreshing a DRAM device: RAS#-only and CAS#-before-
RAS#. The DRAM controller design requirements are simpler for RAS#-only than for CAS#-be-
fore-RAS#.
15-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
The RAS#-only method requires that the DRAM controller activate its RAS# signal when the
RCU activates its REFRESH# signal. This causes the controller to drive the refresh address gen-
erated by the RCU onto the DRAM address inputs, refreshing the specified DRAM row. With this
method, the controller need not assert the CAS# signal whenever the REFRESH# signal is active.
The CAS#-before-RAS# method requires that the DRAM device contain an internal counter to
determine the DRAM row addresses. To perform a refresh cycle using the CAS#-before-RAS#
method, the controller must generate a CAS# signal followed by a RAS# signal when the RCU
activates its REFRESH# signal. With this method, the DRAM device generates its own refresh
addresses and the RCU provides the REFRESH# signal.
If the CS6#/REFRESH# pin is being used for its CS6# function, another way of identifying a re-
fresh cycle is to look at the states of the bus status signals, M/IO#, D/C# and W/R#, (shown in
W/R# is low, and both BHE# and BLE# are inactive during a refresh cycle. These signals can be
used by the DRAM controller to initiate a DRAM refresh cycle.
15.2 REFRESH CONTROL UNIT OVERVIEW
The RCU includes an interval timer unit, a control unit, and an address generation unit (Figure
15-1). The interval timer unit uses a refresh clock interval register and a 10-bit interval counter
to create a periodic signal (timeout). The control unit uses this signal to initiate periodic refresh
requests. The address generation unit uses a refresh base address register and a 13-bit address
counter to generate DRAM refresh addresses. The DRAM device can use these addresses as row
addresses during RAS-only refresh cycles. Each time the interval timer unit times out, a new re-
fresh address is generated.
15-2
REFRESH CONTROL UNIT
Interval Timer Unit
Refresh Clock Interval Register
10-bit Interval Counter
Timeout
Processor Clock
(CLK2/2)
REFRESH#
(pin mux)
Control Unit
Refresh
Request
Refresh Control Register
S
y
s
Refresh
Acknowledge
t
e
m
Address Generation Unit
B
u
s
A25:14
A13:1
Refresh Base Address Register
13-bit Address Counter
Refresh Address Register
A2341-01
Figure 15-1. Refresh Control Unit Connections
15-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
15.2.1 RCU Signals
Table 15-1 describes the signals associated with the RCU.
Table 15-1. RCU Signals
Device Pin or Internal
Signal
CLKOUT
Description
Signal
Device Pin
Processor Clock:
(from Clock and Power
Management Unit)
Provides the clocking signal for the interval counter. The
interval timer unit loads and decrements the counter on
the falling edges of the processor clock.
Timeout
Internal signal
Timeout:
(from the interval counter to
the control unit)
Indicates that the interval counter has reached one. The
control unit initiates a refresh request when it detects
this signal, unless a refresh request is pending, in which
case it ignores this signal.
REFRESH#
Device pin
(output)
External Refresh:
Indicates that a refresh bus cycle is in progress and that
the refresh address is on the bus.
Refresh
Request
Internal signal
Internal signal
Refresh Request:
Indicates that the control unit is requesting bus
ownership.
Refresh
Refresh Acknowledge:
Acknowledge
Indicates that the refresh control unit is being granted
bus ownership.
A25:1
Device pins
(output)
Address Bus:
Contains the refresh address during refresh cycles. This
address can be used by the DRAM device to refresh a
single row.
15.2.2 Refresh Intervals
The interval timer unit controls the rate at which the control unit generates refresh requests. Re-
fresh intervals are programmable through the use of a refresh control interval register (RFSCIR)
and a 10-bit down counter. The counter is loaded from RFSCIR, then decremented on each CLK-
OUT falling edge. When the counter reaches one, the interval timer unit reloads the counter from
the RFSCIR and asserts its timeout signal. The timeout signal causes the control unit to initiate a
refresh request, provided there is not one already pending. (The RCU must complete the present
refresh cycle before the control logic can generate a new refresh request). The control unit ignores
the timeout signal if it already has a refresh request pending.
15.2.3 Refresh Addresses
The physical address generated during a refresh bus cycle has two components: address bits
A25:14 (from the refresh base address register) and address bits A13:1 (from the 13-bit address
counter).
15-4
REFRESH CONTROL UNIT
The 13-bit address counter is a combination of a binary counter and a 7-bit linear-feedback shift
register. The binary counter produces address bits A13:8 and the linear-feedback shift register
produces address bits A7:1. The shift register nonsequentially produces all 128 (27) possible com-
binations. Each time the lower seven bits cycle through all 128 combinations, the binary counter
increments the upper 6 bits. This continues until the 13-bit address counter cycles through 8192
(213) address combinations. The counter then rolls over to its original value and the process re-
peats.
15.2.4 Bus Arbitration
Because the two DMA channels, an external device (via the HOLD pin), and the refresh control
unit can all request bus control, bus control priority must be arbitrated. Refresh requests always
have the highest priority. “Bus Control Arbitration” on page 12-9 discusses the priority structure
of the other bus control requests.)
When a refresh occurs while a DMA channel is performing a transfer, the RCU “steals” a bus cy-
cle to perform a refresh. An external device can gain bus control through either the HOLD signal
or the DMA cascade mode. In this case, a refresh request causes the HLDA or DMACKn# signal
to be deasserted. When this happens, the external device should deassert its request line (HOLD
or DRQn) to allow the RCU to perform a refresh cycle. The refresh cycle is not executed until the
external device deasserts its request. If the external device reasserts its request signal before the
RCU completes the refresh cycle, bus control is given back to the external device after the refresh
cycle completes, without further arbitration.
15.3 RCU OPERATION
The following steps describe the basic refresh cycle, which is initiated every time the interval
counter reaches one.
1. The interval timer unit asserts the timeout signal and reloads the interval counter with the
refresh clock interval register value. The interval counter decrements on each succeeding
processor clock falling edge.
2. The RCU requests bus ownership.
3. Bus ownership is given to the control unit.
4. The control unit asserts the REFRESH# signal and a bus memory read cycle (with neither
Byte-enable signal active) is executed with the address supplied by the RCU.
5. The DRAM controller asserts RAS#, latching the row address inside the DRAM device.
This refreshes the row.
6. The control unit deasserts REFRESH#, and the process repeats from step 1 when the
interval counter reaches one again.
Once enabled, the DRAM refresh process continues until you reprogram the RCU, a reset occurs,
or the processor enters powerdown mode.
15-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
15.4 REGISTER DEFINITIONS
Table 15-2 provides an overview of the registers associated with the RCU. The following sections
provide specific programming information for each register.
Table 15-2. RCU Registers
Expanded
Register
RFSCIR
Description
Address
0F4A2H
Refresh Clock Interval:
(read/write)
Determines the processor clock (CLK2/2) count between refresh requests.
Refresh Control:
RFSCON
(read/write)
0F4A4H
0F4A0H
0F4A6H
Enables the refresh control unit. Reading this register also provides the
current value of the interval counter.
RFSBAD
(read/write)
Refresh Base Address:
Contains the A25:14 address bits of the refresh address. This establishes
a memory region for refreshing.
RFSADD
Refresh Address:
(read/write)
Contains the A13:1 address bits of the refresh address. The 13-bit address
counter generates these values.
15-6
REFRESH CONTROL UNIT
15.4.1 Refresh Clock Interval Register (RFSCIR)
Use RFSCIR to program the interval timer unit’s 10-bit down counter. The refresh counter value
is a function of DRAM specifications and processor frequency as follows:
DRAM refresh period (µs) × processor clock (MHz)
counter value =
---------------------------------------------------------------------------------------------------------------------------------------- ,
X
where X = 128 or the # of DRAM rows, whichever is greater.
The DRAM refresh period is the time required to refresh all rows in the DRAM device.
NOTE
Because the lower seven address bits come from a linear-feedback shift
register, which generates all address bit combinations in a nonsequential order,
X in the equation above must never be less than 128 to ensure proper refresh of
all the rows in a DRAM device that has less than 128 rows.
Refresh Clock Interval
RFSCIR
Expanded Addr: F4A2H
ISA Addr:
—
(read/write)
Reset State:
0000H
15
8
0
—
—
—
—
—
—
RC9
RC1
RC8
RC0
7
RC7
RC6
RC5
RC4
RC3
RC2
Bit
Number
Bit
Mnemonic
Function
15–10
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
9–0
RC9:0
Refresh Counter Value:
Write the counter value to these ten bits. The interval counter counts
down from this value. When the interval counter reaches one, the control
unit initiates a refresh request (provided it does not have a request
pending). The counter value is a function of DRAM specifications and
processor frequency (see the equation above).
Figure 15-2. Refresh Clock Interval Register (RFSCIR)
15-7
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
15.4.2 Refresh Control Register (RFSCON)
Use RFSCON to enable and disable the refresh control unit and to check the current interval
counter value.
Refresh Control
RFSCON
Expanded Addr: F4A4H
ISA Addr:
—
(read/write)
Reset State:
0000H
15
8
0
REN
CV7
—
—
—
—
—
CV9
CV1
CV8
CV0
7
CV6
CV5
CV4
CV3
CV2
Bit
Number
Bit
Mnemonic
Function
15
REN
Refresh Control Unit Enable:
This bit enables or disables the refresh control unit.
0 = Disables refresh control unit
1 = Enables refresh control unit
14–10
9–0
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
CV9:0
Counter Value:
These read-only bits represent the current value of the interval counter.
Write operations to these bits have no effect.
Figure 15-3. Refresh Control Register (RFSCON)
15-8
REFRESH CONTROL UNIT
15.4.3 Refresh Base Address Register (RFSBAD)
Use RFSBAD to set up the memory region that needs refreshing. The value written to this register
forms the upper bits (A25:14) of the refresh address. The RFSBAD register can be used in con-
junction with the Chip Select Unit (CSU) to generate a chip-select for the DRAM region during
refresh cycles. If the address in the RFSBAD matches the region programmed in the CSU for
DRAM, then the DRAM chip-select is generated for both access and refresh cycles.
By programming two separate regions in the CSU, one for DRAM access cycles and the other for
DRAM refresh cycles, separate chip-selects can be generated for the two types of cycles. In this
case, the RFSBAD needs to be programmed with an address that matches the CSU region that is
programmed for the refresh cycle chip-select.
Refresh Base Address
RFSBAD
Expanded Addr: F4A0H
ISA Addr:
—
(read/write)
Reset State:
0000H
15
8
0
—
—
—
—
RA25
RA24
RA16
RA23
RA15
RA22
RA14
7
RA21
RA20
RA19
RA18
RA17
Bit
Number
Bit
Mnemonic
Function
15–12
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
11–0
RA25:14
Refresh Base:
These bits make up the A25:14 address bits of the refresh address. This
establishes a memory region for refreshing.
Figure 15-4. Refresh Base Address Register (RFSBAD)
15-9
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
15.4.4 Refresh Address Register (RFSADD)
RFSADD contains the bits A13:1 of the refresh address. The lowest address bit is not used be-
cause most DRAM devices contain word-wide memory arrays; for all refresh operations, the low-
est address bit remains set.
Refresh Address
RFSADD
Expanded Addr: F4A6H
ISA Addr:
—
(read/write)
Reset State:
00FFH
15
8
0
—
—
RA13
RA5
RA12
RA4
RA11
RA10
RA2
RA9
RA1
RA8
1
7
RA7
RA6
RA3
Bit
Number
Bit
Mnemonic
Function
15–14
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
13–1
0
RA13:1
—
Refresh Address Bits:
These bits comprise A13:1 of the refresh address.
Refresh Bit 0:
A0 of the refresh address. This bit is always 1 and is read-only.
Figure 15-5. Refresh Address Register (RFSADD)
15-10
REFRESH CONTROL UNIT
15.5 DESIGN CONSIDERATIONS
Consider the following when programming the RCU.
• The system address bus does not contain an address A0 signal; instead, it uses the BLE#
and the BHE# pins to generate the lowest address bit. During all refresh operations, BLE#
and BHE# are driven high.
This needs to be noted especially when interfacing to an 8-bit wide Pseudo Static RAM
(PSRAM) device. The lowest address bit generated by the refresh address counter is A1. A
circuit like the one shown in Figure 15-6 can be used to ensure the refresh of all rows. Here
BLE# is connected to an address line of the PSRAM that is not used during refresh. Address
A1 of the processor is connected to A0 of the PSRAM and so forth. For example, when
using a 128Kx8-bit PSRAM device (refresh cycles only use the address present on inputs
A8:0), connect A1 of the processor to A0 of the PSRAM, A2 to A1 and so on, until A9 to
A8. Then connect BLE# of the processor to any one of the A16:9 address lines of the
PSRAM. Since PSRAM is random access memory, this scheme works. During access
cycles, sequential accesses by the processor go to non-contiguous addresses in the PSRAM,
but since the processor does both the read and write cycles, this does not pose a problem.
Intel386™ EX
Embedded
Processor
PSRAM
BLE# can be
connected to
any of these
address lines
An
.
.
.
.
BLE#
Am+1
Am+1
Am
.
.
.
.
.
.
Used
during
refresh
cycles
A2
A1
A1
A0
A3352-02
Figure 15-6. Connections to Ensure Refresh of All Rows in an 8-Bit Wide PSRAM Device
• An external device can gain bus control through either the HOLD signal or the DMA
cascade mode. In this case, a refresh request causes the HLDA or DACKn# signal to be
deasserted. When this happens, the external device must drop its request line (HOLD or
DRQn) to allow the RCU to perform a refresh cycle. The refresh request remains pending
until the RCU gets control of the bus.
15-11
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
• If the counter value stored in the Refresh Clock Interval Register (RFSCIR) is <8 and the
RCU is enabled, the RCU always has bus control and other devices will never gain access
to the bus. This is because refresh requests have the highest priority in the bus arbitration
scheme and you are requesting the bus too often.
• There are two common methods of refreshing DRAM: RAS#-only and CAS#-before-
RAS#.
— RAS#-only refresh takes advantage of the Intel386 EX Embedded Processor’s built in
refresh address counter (RFSADD).
— In a CAS#-before-RAS# refresh, the DRAM provides the row address for the refresh
cycle. The RCU counter still generates the row addresses, but they are disregarded by
the DRAM. The only external logic required is a PLD to recognize a refresh cycle and
provide the CAS# and RAS# signals to the DRAM.
Page Mode
A paged DRAM access uses the upper address lines for the row
addresses and the lower lines for the column addresses. On the
Intel386 EX embedded processor, the lower address lines are
connected to the Refresh Address Counter Register (RFSADD). The
RFSADD increments through a set sequence at each refresh request.
Because the lower address bits (wired to the Column Address Buffer)
change with each refresh request, the PLD must enable this buffer
when RAS# is asserted during a refresh cycle. Figure 15-7 shows the
external logic needed for paged RAS#-only refresh cycles. The PLD
can determine a refresh cycle by monitoring BHE# and BLE# (they
are both inactive during a refresh cycle), or by an active signal on the
REFRESH# pin. The buffer and lines that are active during this type
of refresh have a shaded background in Figure 15-7.
15-12
REFRESH CONTROL UNIT
Row
Address
Row
Address
Buffer
Upper Address
OE_ROW#
Address
REFRESH#
BHE#
CS
BLE#
RAS#
CAS#
Paged
DRAM
Intel386™ EX
Embedded Processor
PLD
n#
OE_COL#
Column
Address
Buffer
Lower Address
Column
Address
Note:
A single mux can be used in place of the row and column address buffers.
A3264-02
Figure 15-7. RAS# Only Refresh Logic: Paged Mode
Non-page Mode
In non-paged mode, the row address buffer can be connected to the
lower address lines and the column address buffer to the upper lines.
Figure 15-8 illustrates the hardware configuration for non-paged
DRAM accesses. The lines and buffer that are enabled in this type of
refresh are highlighted in the figure. The lower address bits are
connected to the Row Address Buffer and the upper address bits are
connected to the Column Address Buffer. As in Page Mode, the PLD
recognizes a refresh request by sampling both BHE# and BLE# (they
are both inactive during a refresh cycle), or by detecting an active
signal on the REFRESH# pin. The buffer and lines that are active
during this type of refresh have a shaded background in Figure 15-8.
15-13
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Row
Address
Row
Address
Buffer
Lower Address
OE_ROW#
Address
REFRESH#
BHE#
CS
BLE#
RAS#
CAS#
Non-paged
DRAM
Intel386™ EX
Embedded Processor
PLD
n#
OE_COL#
Column
Address
Buffer
Upper Address
Column
Address
Note:
A single mux can be used in place of the row and column address buffers.
A3265-02
Figure 15-8. RAS# Only Refresh Logic: Non-Paged Mode
15.6 PROGRAMMING CONSIDERATIONS
REFRESH# and CS6# share a package pin. To select the REFRESH# signal at this pin, set bit 6
in the PINCFG register:
_SetEXRegByte(PINCFG, (_GetEXRegByte(PINCFG) | 0x40));
15.6.1 Refresh Control Unit Example Code
The following code example contains software routines that initialize the refresh control unit and
retrieve the current value of the refresh interval timer. See Appendix C for the included header
files.
#include <conio.h>
#include “80386ex.h”
#include “ev386ex.h”
/*****************************************************************************
InitRCU:
Description:
Initializes the Refresh Control Unit
15-14
REFRESH CONTROL UNIT
Parameters:
Counter_Value Value of the refresh interval
Returns:
Error Codes:
E_BADVECTOR
E_OK
User input an invalid parameter
Executed correctly
Assumptions:
None
Syntax:
#define REFRESH_INTERVAL 0x186
int error_code;
//Counter value for DRAM with
// 1024 rows and a refresh period
// of 16 msec (25 MHz Processor Clock)
error_code = InitRCU(REFRESH_INTERVAL);
Real/Protected Mode:
No changes required
*****************************************************************************/
extern int InitRCU(WORD Counter_Value)
{
/* Check that Counter_Value is 10 bits in length */
if (Counter_Value != (Counter_Value & 0x03ff) )
return(E_BADVECTOR);
/* Clear lower 10 bits of RFSCIR */
_SetEXRegWord(RFSCIR, 0xfc00);
/* Set lower 10 bits of RFSCIR to Counter_Value */
_SetEXRegWord(RFSCIR, _GetEXRegWord(RFSCIR) | Counter_Value);
/* Enable Refresh Unit */
_SetEXRegWord(RFSCON, _GetEXRegWord(RFSCON) | 0x8000);
return(E_OK);
}/* InitRCU */
/*****************************************************************************
Get_RCUCounterValue:
Description:
This function returns the current value of the refresh interval timer.
15-15
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Parameters:
None
Returns:
Refresh Interval Counter Value
Assumptions:
NONE
Syntax:
WORD CounterValue;
CounterValue = Get_RCUCounterValue();
Real/Protected Mode:
No changes required
******************************************************************************/
extern WORD Get_RCUCounterValue(void)
{
WORD Counter_Value;
Counter_Value = _GetEXRegWord(RFSCON) & 0x3ff; // Counter value contained
// in bits RFSCON9:0
return(Counter_Value);
}/* Get_RCUCounterValue */
15-16
16
INPUT/OUTPUT
PORTS
CHAPTER 16
INPUT/OUTPUT PORTS
Input/Output (I/O) ports allow you to transfer information between the processor and the sur-
rounding system circuitry. I/O ports are typically used to read system status, monitor system op-
eration, output device status, configure system options, and generate control signals.
The Intel386™ EX processor’s I/O port pins are multiplexed with peripheral pin functions. With
and use any remaining pins for general-purpose I/O. For example, this device offers eight chip-
not need all eight chip-selects, you can use up to five pins (P2.0–P2.4) for I/O.
This chapter describes the I/O ports and explains how to configure them. The information is ar-
ranged as follows:
• Overview (see below)
• Register Definitions (page 16-6)
• Programming Considerations (page 16-11)
16.1 OVERVIEW
The Intel386 EX processor has three 8-bit bidirectional I/O ports, all of which are functionally
identical (Figure 16-1). Each port has three control registers and a status register.
All three ports share pins with internal peripherals (see Table 16-1). If your design does not re-
quire a pin’s peripheral function, you can configure that pin for use as an I/O port. For example,
if you don’t need serial channel 0, you can use P1.4–P1.0 and P2.7–P2.5 as I/O ports and still
allow the bus interface unit to use P1.7–P1.5 and the chip-select unit to use P2.4–P2.0.
Each pin can operate either in I/O mode or in peripheral mode. In I/O mode, a pin has three pos-
sible configurations:
• open-drain output (requires an external pull-up resistor)
• complementary output
In I/O mode, register bits control the direction (input or output) of each pin and the value of each
output pin. In peripheral mode, the internal peripheral controls the operation (input or output) of
the pin. Table 16-1 lists the port pins with their reset status, multiplexed peripheral functions, di-
rection (input or output), and associated internal peripheral.
16-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
PnCFG.x
I
n
t
e
r
n
a
l
Internal Peripherals
PnDIR.x
1
0
Pn.x
B
u
s
PnLTC.x
PnPIN.x
A2393-01
Figure 16-1. I/O Port Block Diagram
16.1.1 Port Functionality
The function of a bi-directional port pin is controlled by the state of the Port Control Latch (Pn-
LTC). This is shown in Figure 16-2.
16-2
INPUT/OUTPUT PORTS
From Internal
Peripheral
Read Port
Data latch
0
1
Pin
PnLTC
Q
D
S
Q#
CK
Write Port
Data Latch
Read Port
Pin State
PnPIN
SYNC
0
To Internal
Peripheral
V
or
CC
V
SS†
1
PnDIR
Q
S
Internal Data
Bus (F-Bus)
D
0
1
Q#
CK
Write Port
Direction
S
Read Port
Direction
From Internal
Peripheral
Direction
PnCFG
Control
Q
D
Q#
CK
Write Port
Control
Read Port
Control
† Depends on peripheral's inactive state
A3266-01
Figure 16-2. Logic Diagram of a Bi-directional Port
16-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
The output of the Pin Configuration latch (PnCFG) selects whether the I/O port or peripheral is
connected to the pin. When the port is programmed to act as a peripheral pin, both the data for
the pin and the directional control signal for the pin come from the associated integrated periph-
eral. When a bi-directional port pin is programmed as an I/O port, all port parameters are under
software control.
The output of the Port Direction latch (PnDIR) enables or disables the three-state output driver
when the pin is programmed as an I/O port. The three-state output driver is enabled by clearing
the Port Direction latch. The data driven to an output port pin is held in the Port Data latch. Setting
the Port Direction latch disables the three-state output driver making the pin an input.
The signal present on the pin is routed through a synchronizer to a three-state buffer that connects
the I/O port path to the internal data bus. Not all peripheral input functions are synchronous. For
example, the interrupt pins (INT9-INT0) are asynchronous so that they can wake up the chip from
Powerdown mode when the clocks are stopped.
The state of the pin can be read at any time regardless of whether the pin is used as an I/O port or
for a peripheral function.
16-4
INPUT/OUTPUT PORTS
Table 16-1. Pin Multiplexing
Peripheral Function
Port Pin
Reset Status(1)
Internal
Peripheral
Pin
Signal
Direction(2)
P1.0
P1.1
P1.2
P1.3
P1.4
P1.5
P1.6
P1.7
P2.0
P2.1
P2.2
P2.3
P2.4
P2.5
P2.6
P2.7
P3.0
P3.1
P3.2
P3.3
P3.4
P3.5
P3.6
P3.7
wk 1
wk 1
wk 1
wk 1
wk 1
wk 1
wk 0
wk 0
wk 1
wk 1
wk 1
wk 1
wk 1
wk 0
wk 0
wk 1
wk 0
wk 0
wk 0
wk 0
wk 0
wk 0
wk 0
wk 0
DCD0#
RTS0#
DTR0#
DSR0#
RI0#
I
O
O
I
SIO0
SIO0
SIO0
SIO0
SIO0
BIU
I
LOCK#
HOLD
HLDA
O
I
BIU
O
O
O
O
O
O
I
BIU
CS0#
CSU
CS1#
CSU
CS2#
CSU
CS3#
CSU
CS4#
CSU
RXD0
SIO0
SIO0
SIO0
Timer 0
Timer 1
ICU
TXD0
O
I
CTS0#
TMROUT0
TMROUT1
INT0
O
O
I
INT1
I
ICU
INT2
I
ICU
INT3
I
ICU
PWRDOWN
COMCLK
O
I
CLK & PM
SIO0, SIO1
NOTES:
1. wk 0 = weakly pulled down; wk 1 = weakly pulled up.
2. I = input; O = output.
16-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
16.2 REGISTER DEFINITIONS
Each port has three control registers and a status register associated with it (Table 16-2). The con-
trol registers (PnCFG, PnDIR, and PnLTC) can be both read and written. The status register (Pn-
PIN) can only be read. All four registers reside in I/O address space.
Table 16-2. I/O Port Registers
Register
P1CFG
P2CFG
P3CFG
Address
Description
0F820H
0F822H
0F824H
Port n Mode Configuration:
Each bit controls the mode of the associated pin.
0 = Selects I/O mode.
(read/write)
1 = Selects peripheral mode.
P1DIR
P2DIR
P3DIR
0F864H
0F86CH
0F874H
Port n Direction:
Each bit controls the direction of a pin that is in I/O mode.
0 = Configures a pin as a complementary output. If a pin is in peripheral
mode, this value is ignored.
(read/write)
1 = Configures a pin as either an input or an open-drain output.
Port n Data Latch:
P1LTC
P2LTC
P3LTC
0F862H
0F86AH
0F872H
Each bit contains data to be driven onto an output pin that is in I/O mode. Write
the desired pin state value to this register. If a pin is in peripheral mode, this
value is ignored.
(read/write)
Reading this register returns the value in the register—not the actual pin state.
Port n Pin State:
P1PIN
P2PIN
P3PIN
0F860H
0F868H
0F870H
Each bit of this read-only register reflects the state of the associated pin.
Reading this register returns the current pin state value, regardless of the pin’s
mode and direction.
(read only)
16-6
INPUT/OUTPUT PORTS
16.2.1 Pin Configuration
You select the operating mode of each pin by writing to the associated bit in the PnCFG registers
(Figure 16-3 gives an abbreviated version of these registers; for the complete register descrip-
tions, see Appendix D). Setting a bit selects peripheral mode; clearing a bit selects I/O mode. In-
ternal peripherals control pins configured for peripheral mode, while the PnDIR (Figure 16-4)
and PnLTC (Figure 16-5) registers control pins configured for I/O mode. Table 16-3 shows the
PnDIR and PnLTC register values that determine the pin direction and state.
NOTE
You must program both registers to correctly configure the pins.
Table 16-3. Control Register Values for I/O Port Pin Configurations
Desired Pin Configuration
Desired Pin State
PnDIR PnLTC
High-impedance input
high impedance
1
1
1
0
0
1
1
0
1
0
high impedance
Open-drain output
0
1
0
Complementary output
Regardless of the pin’s configuration, you can read the PnPIN registers (Figure 16-6) to determine
the current pin state.
Port n Configuration
PnCFG (n=1–3)
(read/write)
Expanded Addr: F820H, F822H, F824H
ISA Addr:
—
Reset State:
00H
7
0
PM7
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7–0
PM7:0
Pin Mode:
0 = Places pin in I/O mode, controlled by PnDIR and PnLTC registers.
1 = Places pin in peripheral mode, controlled by the internal peripheral
Figure 16-3. Port n Configuration Register (PnCFG)
16-7
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Port DIrection
PnDIR (n=1–3)
(read/write)
Expanded Addr: F864H, F86CH, F874H
ISA Addr:
—
Reset State:
FFH
7
0
PD7
PD6
PD5
PD4
PD3
PD2
PD1
PD0
Bit
Number
Bit
Mnemonic
Function
7–0
PD7:0
Pin Direction:
0 = Configures the pin as a complementary output.
1 = Configures the pin as an open-drain output or high-impedance input.
Figure 16-4. Port Direction Register (PnDIR)
Port Data Latch
PnLTC (n=1–3)
(read/write)
Expanded Addr: F862H, F86AH, F872H
ISA Addr:
—
Reset State:
FFH
7
0
PL7
PL6
PL5
PL4
PL3
PL2
PL1
PL0
Bit
Number
Bit
Mnemonic
Function
7–0
PL7:0
Port Data Latch:
Writing a value to a PL bit causes that value to be driven onto the
corresponding pin.
For a complementary output, write the desired pin value to its PL bit.
This value is strongly driven onto the pin.
For an open-drain output, a one results in a high-impedance (input) state
at the pin.
For a high-impedance input, write a one to the corresponding PL bit. A
one results in a high-impedance state at the pin, allowing external
hardware to drive it.
Figure 16-5. Port Data Latch Register (PnLTC)
16-8
INPUT/OUTPUT PORTS
Port Pin State
PnPIN (n=1–3)
(read only)
Expanded Addr: F860H, F868H, F870H
ISA Addr:
—
Reset State:
XXH
7
0
PS7
PS6
PS5
PS4
PS3
PS2
PS1
PS0
Bit
Number
Bit
Mnemonic
Function
7–0
PS7:0
Pin State:
Reading a PS bit returns the logic state present on the associated port
pin.
Figure 16-6. Port Pin State Register (PnPIN)
16-9
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
16.2.2 Initialization Sequence
After a device reset, a weak pull-up or pull-down resistor holds each pin high or low until user
software writes to the PnCFG register. The pins are configured as inputs in I/O port mode. To en-
sure that the pins are initialized correctly and that the weak resistors are turned off, follow this
suggested initialization sequence.
NOTE
Even if you want to use the entire port as I/O (its default configuration after
reset), you must write to PnCFG to turn off the weak pull-up and pull-down
resistors.
1. Write to PnLTC to specify the pin value. Writing to PnLTC before PnDIR ensures that
output pins initialize to known values.
•
•
For an output pin, write the data that is to be driven by the pin to its PnLTC bit.
For an input pin, set its PnLTC bit.
2. Write to PnDIR to specify the pin direction.
•
•
To configure a pin as a complementary output, clear its PnDIR bit.
To configure a pin as an input or open-drain output, set its PnDIR bit.
3. Write to PnCFG to turn off the weak resistors and select either I/O or peripheral mode.
•
•
To configure a pin for I/O mode, clear its PnCFG bit.
To configure a pin for peripheral mode, set its PnCFG bit.
16.3 DESIGN CONSIDERATIONS
This section outlines design considerations for the I/O ports.
• Source and sink current are different between the three ports. Consult the latest Intel386™
EX Embedded Microprocessor datasheet (order number 272420) for exact specifications.
• Use read/modify/write operations to set and clear bits.
16.3.1 Pin Status During and After Reset
A device reset applies an asynchronous reset signal to the port pins. To avoid contention with ex-
ternal drivers, the pins are configured as inputs in I/O port mode. To prevent pins from floating,
a weak pull-up or pull-down resistor holds each pin high or low (Table 16-1). Writing to the
PnCFG register (regardless of the value written) turns off these resistors. For example, writing
any value to P1CFG after a reset turns off the weak pull-down resistors on P1.7–P1.6 and the
weak pull-up resistors on P1.5–P1.0. The resistors remain off until the next reset.
16-10
INPUT/OUTPUT PORTS
16.4 PROGRAMMING CONSIDERATIONS
16.4.1 I/O Ports Code Example
The following code example contains a software routine that initializes the I/O port pins. See Ap-
pendix C for the included header files.
#include <conio.h>
#include “80386ex.h”
#include “ev386ex.h”
/*******************************************************************************
Init_IOPorts:
Description:
This function initializes the direction and mode of the I/O port pins.
Although the pins are default configured to the peripheral state after
RESET, they must still be initialized to turn off the weak resistors.
Parameters:
Port1
Port2
Port3
PortDir1
PortDir2
PortDir3
PortLtc1
PortLtc2
PortLtc3
Port1 Mode Configuration
Port2 Mode Configuration
Port3 Mode Configuration
Port1 Direction
Port2 Direction
Port3 Direction
Port1 Data Latch Value
Port2 Data Latch Value
Port3 Data Latch Value
Returns:
None
Assumptions:
None
Syntax:
// Port 1 configuration defines
#define DCD0
#define RTS0
#define DTR0
#define DSR0
#define RI0
#define LOCK
#define HOLD
#define HOLDACK
0x1
0x2
0x4
0x8
0X10
0x20
0X40
0X80
// Port 2 configuration defines
#define CS0
#define CS1
0x1
0x2
16-11
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
#define CS2
#define CS3
#define CS4
#define RXD0
#define TXD0
#define CTS0
0x4
0x8
0X10
0x20
0X40
0X80
// Port 3 configuration defines
#define TMROUT0
#define TMROUT1
#define INT0
0x1
0x2
0x4
#define INT1
0x8
#define INT2
#define INT3
#define PWRDWN
#define COMCLK
0x10
0x20
0x40
0x80
// Port Direction defines
#define P0_IN
#define P1_IN
#define P2_IN
#define P3_IN
#define P4_IN
#define P5_IN
#define P6_IN
#define P7_IN
#define Px_OUT
0x1
0x2
0x4
0x8
0x10
0x20
0x40
0x80
0
//Initialize SIO0 pins, DRAM and SRAM Chip Selects, Interrupt Signals,
//and TimerOut Signals to be in peripheral mode
Init_IOPorts( DCD0|RTS0|DTR0|DSR0|RI0,
CS2|CS4|RXD0|TXD0|CTS0,
TMROUT0|TMROUT1|INT0|INT1|INT2|INT3|PWRDWN|COMCLK,
Px_OUT,
Px_OUT,
Px_OUT,
0xff, // This example shows all output pins being
// initially 1
0xff,
// Note: Input pins must be given an initial
0xff ); // value of 1 whereas peripheral pins initially
// can be set or cleared
Real/Protected Mode:
No changes required.
********************************************************************************
/
extern void Init_IOPorts(BYTE Port1, BYTE Port2, BYTE Port3, BYTE PortDir1,
BYTE PortDir2, BYTE PortDir3, BYTE PortLtc1,
BYTE PortLtc2, BYTE PortLtc3)
16-12
INPUT/OUTPUT PORTS
{
/* Select pin values */
_SetEXRegByte(P1LTC, PortLtc1);
_SetEXRegByte(P2LTC, PortLtc2);
_SetEXRegByte(P3LTC, PortLtc3);
/* Select pin directions */
_SetEXRegByte(P1DIR, PortDir1);
_SetEXRegByte(P2DIR, PortDir2);
_SetEXRegByte(P3DIR, PortDir3);
/* Turn off weak resistors and select either I/O or peripheral mode */
_SetEXRegByte(P1CFG, Port1);
_SetEXRegByte(P2CFG, Port2);
_SetEXRegByte(P3CFG, Port3);
} /* Init_IOPorts */
16-13
17
WATCHDOG
TIMER UNIT
CHAPTER 17
WATCHDOG TIMER UNIT
The watchdog timer (WDT) unit can function as a general-purpose timer, a software watchdog
timer, or a bus monitor, or it can be disabled.
This chapter is organized as follows:
• Disabling the WDT (page 17-6)
• Register Definitions (page 17-7)
• Design Considerations (page 17-12)
• Programming Considerations (page 17-12)
17.1 OVERVIEW
The watchdog timer unit (Figure 17-1) includes a 32-bit reload register, a 32-bit down-counter,
an 8-state binary counter, a readable counter value register, and a status register.
The watchdog timer can operate in three modes:
• General-purpose 32-bit timer/counter mode (default mode)
• Watchdog mode
• Bus-monitor mode
Only a single mode can be active at one time. If you have no need for any of its functions, you
can disable the unit entirely.
Watchdog mode protects systems from software upsets. In watchdog mode, system software must
reload the down-counter at regular intervals. If it fails to do so, the timer expires and asserts
WDTOUT. For example, the watchdog times out if the software goes into an endless loop.
Some possible uses of this feature include:
• Connecting WDTOUT to the NMI pin to generate a non-maskable interrupt
• Connecting the WDTOUT signal to the RESET pin to reset the processor (and possibly the
entire system)
In watchdog mode only, idle mode stops the down-counter. Since no software can execute while
the CPU is idle, a software watchdog is unnecessary. (Chapter 8, “CLOCK AND POWER MAN-
AGEMENT UNIT,” discusses idle mode.)
Bus monitor mode protects normally not-ready systems from ready-hang conditions. (A normally
not-ready system is one in which a bus cycle continues until the accessed device asserts
17-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
READY#). In bus monitor mode, the ADS# signal from the bus interface unit (BIU) reloads the
down-counter and the READY# signal stops it. The READY# signal can be generated either ex-
ternally or internally, using the WDTRDY bit in the PWRCON register (Figure 17-5). If this bit
is deasserted, then an external READY# is required to terminate the cycle when the WDT times
out (WDTOUT is asserted) in Bus Monitor mode. In this case, if a READY# is never generated
by external logic, the processor hangs (since the bus cycle never terminates). If the WDTRDY bit
is set, the processor generates an internal READY# to terminate the cycle upon time-out
(WDTOUT is asserted) in Bus Monitor mode.
The WDT circuitry correctly matches each READY# with a corresponding ADS# (even in pipe-
lined mode when two ADS# pulses occur before the first READY# pulse).
Reload
Registers
32-Bit
Down Counter
WDTRLDH
WDTRLDL
E
I
WDTCNTH
B
u
WDTCNTL
8-State
Binary
s
WDTOUT
Counter
Connect
to NMI
WDTCLR
or RESET
WDTOUT#
to IR7 of
WDTSTATUS
Slave 8259A
A2330-02
Figure 17-1. Watchdog Timer Unit Connections
17-2
WATCHDOG TIMER UNIT
17.1.1 WDT Signals
Table 17-1 describes the signals associated with the WDT.
Table 17-1. WDT Signals
Device Pin or
Signal
Description
Internal Signal
ADS#
IDLE
Device pin
Address Status (from the bus interface unit):
Indicates that the processor is driving a valid bus-cycle definition
and address onto its pins. Bus monitor mode reloads and starts the
down-counter each time ADS# is asserted.
Internal signal
Idle (from the clock and power management unit):
Indicates that the device is in idle mode (core clocks stopped and
peripheral clocks running). In watchdog mode, the down-counter
stops when the core is idle. In bus monitor or general-purpose
timer mode, the WDT continues to run while the core is idle.
READY#
WDTOUT
Device pin
Device pin
Ready (from the bus interface unit):
Indicates that the current bus cycle has completed. Bus monitor
mode stops the down-counter when READY# is asserted.
Watchdog Timer Output:
Indicates that the down-counter has timed out. If you want a WDT
timeout to reset the device, connect WDTOUT to the RESET input.
If you want a WDT timeout to generate a nonmaskable interrupt,
connect WDTOUT to the NMI input.
An internal signal carries the inverted value of WDTOUT to the
interrupt control unit (the slave’s IR7 line). If you want a WDT
timeout to cause a maskable interrupt, enable the interrupt.
(Chapter 8, “Interrupt Control Unit,” explains how to do this.)
17.2 WATCHDOG TIMER UNIT OPERATION
After a device reset, the WDT begins counting down in general-purpose timer mode. Unless you
change the mode, change the reload value, or disable it, the WDT times out and asserts WDTOUT
22
after 4 million (2 ) processor clock cycles (PH1 or CLKOUT cycles).
The 32-bit down-counter decrements on every processor clock cycle. When the down-counter
reaches zero, the 8-state binary counter drives the WDTOUT pin high for eight processor clock
cycles (16 CLK2 cycles) to signal the timeout. An internal signal carries the inverted value of the
WDTOUT pin to the interrupt control unit (the slave’s IR7 line). A WDT timeout can reset the
system or generate an interrupt request, depending on how WDTOUT is used in your system.
17-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
The reload registers hold a user-defined value that reloads the down-counter when one of the fol-
lowing reload events occurs:
• In watchdog mode, when system software executes a specific instruction sequence (called a
lockout sequence) to the WDTCLR location
• In bus monitor mode, when the bus interface unit asserts ADS#
• In all modes, when the down-counter reaches zero
Software can read the status register to determine the mode of the WDT, and can read the count
registers to determine the current value of the down-counter.
17.2.1 Idle and Powerdown modes
In CPU-idle mode, the WDT is disabled only if it is in watchdog mode. Since no software can
execute while the CPU is in idle mode, the software watchdog is unnecessary. The WDT operates
normally in general-purpose timer and bus-monitor modes if the CPU is in idle mode.
In CPU-powerdown mode, the WDT unit is disabled, like all other peripherals.
17.2.2 General-purpose Timer Mode
The WDT defaults to general-purpose timer mode after reset. If your system has no requirement
for a software watchdog or a bus monitor, you can use the WDT in this mode. At reset, the down-
counter begins decrementing once every clock cycle, beginning at 3FFFFFH (the initial values of
22
the reload and count registers). Unless you intervene, the WDT times out after 4 million (2 ) pro-
cessor clock cycles.
Software can read the count registers (WDTCNTH and WDTCNTL) at any time to determine the
current value of the down-counter. You might, for example, read the count when one event occurs,
When the down-counter reaches zero, the 8-state binary counter drives the WDTOUT pin high
for eight processor clock cycles (16 CLK2 cycles). During the clock cycle immediately after the
down-counter reaches zero, the down-counter is reloaded with the contents of the reload registers.
22
If you want fewer than 4 million (2 ) processor clock cycles between WDT timeouts, write a 32-
bit reload value to the reload registers (Figure 17-4):
1. Write the upper 16 bits of the reload value to WDTRLDH.
2. Write the lower 16 bits of the reload value to WDTRLDL.
In the general-purpose timer mode, you cannot reload the counter except on a WDT timeout.
However, you can force a reload by entering bus monitor mode, allowing an ADS# to reload the
counter, then switching back to general-purpose timer mode.
17-4
WATCHDOG TIMER UNIT
17.2.3 Software Watchdog Mode
In software watchdog mode, system software must periodically reload the down-counter with a
reload value or the timer expires and asserts WDTOUT. The reload value depends on the design
and some experimentation.
After reset, the WDT defaults to general-purpose timer mode. Unless you intervene, the WDT
22
times out after 4 million (2 ) processor clock cycles. If you want to use the WDT as a system
watchdog, use this sequence to enable watchdog mode:
1. Write the upper 16 bits of the reload value to WDTRLDH (Figure 17-4).
2. Write the lower 16 bits of the reload value to WDTRLDL (Figure 17-4).
3. Write two sequential words, 0F01EH followed by 0FE1H, to the WDTCLR location
(0F4C8H). This sequence (called a lockout sequence) sets the WDTEN bit in the
watchdog status register and loads the contents of the reload value register into the down-
counter.
Regardless of the values of the two control bits (BUSMON and CLKDIS) in the WDTSTATUS
register (Figure 17-3), the lockout sequence sets the WDTEN bit and clears the remaining bits.
The lockout sequence prohibits writes to the WDTSTATUS and reload registers; only a system
reset can change them. This reduces the possibility for errant software to duplicate the instruc-
tions and illegally reload the timer.
The same lockout sequence that enables the watchdog reloads the down-counter. Write two se-
quential words, 0F01EH followed immediately by 0FE1H, to the WDTCLR location (0F4C8H).
17.2.4 Bus Monitor Mode
In bus monitor mode, ADS# reloads and starts the down-counter and READY# stops it. The ini-
tial values of the reload register and down-counter are 3FFFFFH.
CAUTION
For correct operation in Bus Monitor mode (see “Overview” on page 17-1),
you must have a minimum reload value = (Maximum number of wait-states in
your system + 12). For example, if the slowest device in your system requires
8 wait-states during an access, the reload value must be greater than or equal to
20.
Use this sequence to enable bus monitor mode:
1. Write the upper word of the reload value to WDTRLDH (Figure 17-4).
2. Write the lower word of the reload value to WDTRLDL (Figure 17-4).
3. Set the bus monitor bit (BUSMON) in WDTSTATUS (Figure 17-3).
Because you never execute the lockout sequence in bus monitor mode, you can change the reload
value and enable or disable the mode at any time.
17-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
• To change the reload value, write the new values to the WDTRLDH and WDTRLDL
registers, as described in steps 1 and 2 above.
• To disable or enable bus monitor mode, write to the bus monitor bit (BUSMON):
— 0 = disabled
— 1 = enabled
17.3 DISABLING THE WDT
If your system has no need for the WDT, when the unit is in bus monitor or general-purpose timer
mode, you can disable the unit by setting the CLKDIS bit in the WDTSTATUS register (Figure
17-3), which stops the clock to the WDT. In this configuration, the WDT consumes minimal pow-
er, but you can re-enable the unit at any time.
If the WDT is in watchdog mode, you cannot write to the WDTSTATUS register to stop the clock
and therefore cannot disable the unit.
17-6
WATCHDOG TIMER UNIT
17.4 REGISTER DEFINITIONS
This section describes the registers associated with the WDT, and explains how these registers
can be used to enable and use each WDT mode.
Table 17-2 describes the registers associated with the WDT.
Table 17-2. WDT Registers
Register
WDTCLR
Address
Description
0F4C8H
Watchdog Timer Clear:
Write the lockout sequence to this location. Circuitry at this address decodes
the lockout sequence to enable watchdog mode, reload the counter, or both.
This location is used only for watchdog mode.
WDTCNTH
WDTCNTL
0F4C4H
0F4C6H
WDT Counter:
can read them to determine the current count value. Any reload event
reloads these registers with the contents of WDTRLDH and WDTRLDL.
(read only)
WDTRLDH
WDTRLDL
0F4C0H
0F4C2H
WDT Reload Value:
Write the reload value to these registers, using two word writes. After a
lockout sequence is issued, these registers cannot be written again until after
a device reset. A reload event (each WDT mode has its own; refer to
Sections 17.2.2 through 17.2.4) reloads WDTCNTH and WDTCNTL with the
contents of these registers.
(read/write)
WDTSTATUS
(read/write)
0F4CAH
WDT Status:
This register contains one read-only bit (WDTEN) that indicates whether
watchdog mode is enabled and two read/write bits that control bus monitor
mode and the WDT clock. A lockout sequence sets the WDTEN bit and
clears the two read/write bits, disabling bus monitor mode and enabling the
WDT clock. After a lockout sequence is issued, a write to this register has no
effect unless the device is reset.
Software can read this register to determine the current status of the WDT
and (unless a lockout sequence has been issued) can set the BUSMON bit to
enable bus monitor mode or set the CLKDIS bit to disable the WDT.
PWRCON
0F800H
Power Control register:
(read/write)
This register holds the WDTRDY bit that is used to enable/disable internal
READY# generation for the WDT Bus Monitor mode.
17-7
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
WDT Counter Value (High)
WDTCNTH
Expanded Addr: F4C4H
ISA Addr:
—
(read only)
Reset State:
003FH
15
8
0
WC31
WC23
WC30
WC22
WC29
WC21
WC28
WC20
WC27
WC26
WC18
WC25
WC24
WC16
7
WC19
WC17
WDT Counter Value (Low)
WDTCNTL
Expanded Addr: F4C6H
ISA Addr:
—
(read only)
Reset State:
FFFFH
15
8
0
WC15
WC7
WC14
WC6
WC13
WC5
WC12
WC4
WC11
WC10
WC2
WC9
WC1
WC8
WC0
7
WC3
Bit
Mnemonic
Bit Number
Function
WDT Counter Value High Word and Low Word:
High 15–0
Low 15–0
WC31:16
WC15:0
Read the high word of the counter value from WDTCNTH and the low
word from WDTCNTL.
Figure 17-2. WDT Counter Value Registers (WDTCNTH and WDTCNTL)
17-8
WATCHDOG TIMER UNIT
WDT Status
WDTSTATUS
(read/write)
Expanded Addr: F4CAH
ISA Addr:
—
Reset State:
00H
7
0
WDTEN
—
—
—
—
—
BUSMON
CLKDIS
Bit
Number
Bit
Mnemonic
Function
7
WDTEN
Watchdog Mode Enabled:
This read-only bit indicates whether watchdog mode is enabled. Only a
lockout sequence can set this bit and only a device reset can clear it.
0 = Watchdog mode disabled
1 = Watchdog mode enabled
6–2
1
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
BUSMON
Bus Monitor Enable:
0 = Disables bus monitor mode
1 = Enables bus monitor mode
Read this bit to determine the current status. A lockout sequence clears
BUSMON and prevents writes to the WDTSTATUS register.
0
CLKDIS
Clock Disable:
Write to this bit to stop or restart the clock to the WDT; read it to
determine the current clock status. A lockout sequence clears CLKDIS
and prevents writing to this register.
0 = Clock enabled
1 = Processor clock (frequency=CLK2/2) disabled (stopped)
Figure 17-3. WDT Status Register (WDTSTATUS)
17-9
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
WDT Reload Value (High)
WDTRLDH
Expanded Addr: F4C0H
ISA Addr:
—
(read/write)
Reset State:
003FH
15
8
0
WR31
WR30
WR22
WR29
WR21
WR28
WR20
WR27
WR26
WR18
WR25
WR24
WR16
7
WR23
WR19
WR17
WDT Reload Value (Low)
WDTRLDL
Expanded Addr: F4C2H
ISA Addr:
—
(read/write)
Reset State:
FFFFH
15
8
0
WR15
WR14
WR6
WR13
WR5
WR12
WR4
WR11
WR10
WR2
WR9
WR1
WR8
WR0
7
WR7
WR3
Bit
Number
Bit
Mnemonic
Function
WDT Reload Value (High Word and Low Word):
High 15–0 WR31:16
Low 15–0 WR15:0
Write the high word of the reload value to WDTRLDH and the low word
to the WDTRLDL.
Figure 17-4. WDT Reload Value Registers (WDTRLDH and WDTRLDL)
17-10
WATCHDOG TIMER UNIT
Power Control Register
PWRCON
Expanded Addr:
ISA Addr:
F800H
—
(read/write)
Reset State:
00H
7
0
—
—
—
—
WDTRDY HSREADY
PC1
PC0
Bit
Number
Bit
Mnemonic
Function
7–4
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
3
WDTRDY
HSREADY
PC1:0
Watch Dog Timer Ready:
0 = An external READY must be generated to terminate the cycle when
the WDT times out in Bus Monitor Mode.
1 = Internal logic generates READY# to terminate the cycle when the
WDT times out in Bus Monitor Mode.
2
Halt/Shutdown Ready:
0 = An external ready must be generated to terminate a HALT/Shutdown
cycle.
1 = Internal logic generates READY# to terminate a HALT/Shutdown
cycle.
1–0
Power Control:
Program these bits, then execute a HALT instruction. The device enters
the programmed mode when READY# (internal or external) terminates
the halt bus cycle. When these bits have equal values, the HALT
instruction causes a normal halt and the device remains in active mode.
PC1 PC0
0
1
0
1
0
0
1
1
active mode
idle mode
powerdown mode
active mode
Figure 17-5. Power Control Register (PWRCON)
17-11
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
17.5 DESIGN CONSIDERATIONS
This section outlines design considerations for the watchdog timer unit.
Depending on the system configuration, a WDT timeout can cause a maskable interrupt, a non-
maskable interrupt, or a system reset.
Maskable interrupt
The WDT timeout signal is internally inverted and connected to the
interrupt control unit’s slave IR7 line. If you want a WDT timeout to
generate a slave IR7 interrupt (maskable interrupt), you need only
enable (unmask) the interrupt (Refer to Chapter 9, for details).
Ensure that the slave 8259A is configured for edge-triggered
interrupts (refer to Chapter 9, Interrupt Control Unit) if IR7 is
unmasked. Otherwise, the WDT generates continuous interrupts.
Nonmaskable interrupt If you want a WDT timeout to cause a nonmaskable interrupt,
connect the WDTOUT pin to the NMI input pin.
Reset
If you want a WDT timeout to reset the system, connect the
WDTOUT pin to the RESET input pin.
17.6 PROGRAMMING CONSIDERATIONS
This section outlines programming considerations for the watchdog timer unit.
17.6.1 Writing to the WDT Reload Registers (WDTRLDH and WDTRLDL)
WDTRLDH and WDTRLDL are 16 bit registers at addresses 0F4C0H and 0F4C2H respectively.
Therefore, when using a 32-bit write to load the two registers, the lower 16 bits should contain
the data for WDTRLDH and the higher 16 bits should contain the data for WDTRLDL.
For example, 4321H can be written to WDTRLDH and 0CCCCH to WDTRLDL using a 32-bit
write of the number 0CCCC4321H to I/O address 0F4C0H.
17.6.2 Minimum Counter Reload Value
To ensure correct operation of the Watchdog Timer, the WDT’s counter should never be reloaded
with a value less than 8.
17.6.3 Watchdog Timer Unit Code Examples
This section includes these software routines:
ReLoadDownCounter Initiates a lockout sequence
GetWDT_Count
Reads the value of the counter
WDT_BusMonitor
Places the WDT in Bus Monitor Mode
EnableWDTInterrupt Enables WDT interrupts
17-12
WATCHDOG TIMER UNIT
See Appendix C for included header files.
#include <dos.h>
#include <conio.h>
#include “80386ex.h”
#include “ev386ex.h”
/*****************************************************************************
ReLoadDownCounter:
Description:
This function initiates a lockout sequence which results in the
setting of the WDTEN bit in the status register. By setting
WDTEN, the software watchdog mode is enabled.
Parameters:
None
Returns:
None
Assumptions:
None
Syntax:
ReloadDownCounter();
Real/Protected Mode:
No changes required
*****************************************************************************/
void ReLoadDownCounter(void)
{
_disable(); /* Disable interrupts */
_SetEXRegWord(WDTCLR,0xf01e);
_SetEXRegWord(WDTCLR,0xfe1);
_enable(); /* Enable interrupts */
}/* ReLoadDownCounter */
/*****************************************************************************
GetWDT_Count:
Description:
Returns current value of watch dog counter.
17-13
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Parameters:
None
Returns:
16-bit down-counter value
Assumptions:
None
Syntax:
WORD counter_value;
counter_value = GetWDT_Count();
Real/Protected Mode:
No changes required.
******************************************************************************/
DWORD GetWDT_Count(void)
{
WORD LowWord, HiWord;
LowWord = _GetEXRegWord(WDTCNTL);
HiWord = _GetEXRegWord(WDTCNTH);
return (((DWORD)HiWord << 16) + LowWord);
}/* GetWDT_Count */
/*****************************************************************************
WDT_BusMonitor:
Description:
Enables the bus monitor mode of the Watch Dog Timer.
Parameters:
EnableDisable
Nonzero if bus monitor mode is to be enabled,
Zero if it is to be disabled
Returns:
None
Assumptions:
None
Syntax:
#define Enable 0x01
#define Disable 0x00
17-14
WATCHDOG TIMER UNIT
WDT_BusMonitor(Enable);
Real/Protected Mode:
No changes required.
******************************************************************************/
void WDT_BusMonitor(BYTE EnableDisable)
{
BYTE Status;
Status = _GetEXRegByte(WDTSTATUS);
if(EnableDisable) /* If true, Enable */
_SetEXRegByte(WDTSTATUS, Status | BIT1MSK);
/* Set Bit */
else /* else,
Disable */
_SetEXRegByte(WDTSTATUS, Status & ~BIT1MSK); /* Clear Bit */
}/* WDT_BusMonitor */
/*****************************************************************************
EnableWDTInterrupt:
Description:
Enables a maskable interrupt on the assertion of WDTOUT
Parameters:
None
Returns:
None
Assumptions:
None
Syntax:
EnableWDTInterrupt();
Real/Protected Mode:
No changes required
******************************************************************************/
extern void EnableWDTInterrupt(void)
{
InitICUSlave(ICU_TRIGGER_EDGE, 0x30, 0); /* Initialize Slave ICU */
SetIRQVector(wdtISR, 15, INTERRUPT_ISR); /* Puts address of interrupt
service
17-15
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
routine in Interrupt Vector Table */
Enable8259Interrupt(IR2,IR7);
/* Enable slave interrupt to master(IR2),
Enable slave IR2 */
_enable(); /* Enable Interrupts */
} /* EnableWDTInterrupt */
/*****************************************************************************
wdtISR:
Description:
Interrupt Service Routine for Watchdog Timer
Parameters:
None
Returns:
None
Assumptions:
None
Syntax:
Not called by user; Interrupt Control Unit executes this
routine upon acknowledgment of a WDT interrupt
Real/Protected Mode:
No changes required
*****************************************************************************/
void interrupt far wdtISR(void)
{
SerialWriteStr(SIO_PORT,”Executing in WDT_ISR”); /* Prints out to Serial
Port as a demonstration */
NonSpecificEOI();
} /* wdtISR */
17-16
18
JTAG TEST-LOGIC
UNIT
CHAPTER 18
JTAG TEST-LOGIC UNIT
The JTAG test-logic unit enables you to test both the device logic and the interconnections be-
tween the device and the board (system) it is plugged into. The term JTAG refers to the Joint Test
Action Group, the IEEE technical subcommittee that developed the testability standard published
†
as Standard 1149.1-1990, IEEE Standard Test Access Port and Boundary-Scan Architecture and
TM
its supplement, Standard 1149.1a-1993. The Intel386 EX Embedded Processor JTAG test-log-
ic unit is fully compliant with this standard.
ming of flash memory; refer to AP-720, Programming Flash Memory through the Intel386™ EX
This chapter is organized as follows:
• Test-Logic Unit Operation (page 18-3)
• Testing (page 18-10)
• Timing Information (page 18-12)
• Design Considerations (page 18-14)
18.1 OVERVIEW
As the title of the IEEE standard suggests, two major components of the test-logic unit are the test
access port and the boundary-scan register. The term test access port (TAP) refers to the dedicat-
ed input and output pins through which a tester communicates with the test-logic unit. The term
boundary-scan refers to the ability to scan (observe) the signals at the boundary (the pins) of a
device. A boundary-scan cell resides at each pin. These cells are connected serially to form the
boundary-scan register, which allows you to control or observe every device pin except the clock
pin, the power and ground pins, and the test access port pins.
The test-logic unit allows a tester to perform these tasks:
• Identify a component on a board (manufacturer, part number, and version)
• Bypass one or more components on a board while testing others
• Preload a pin state for a test or read the current pin state
• Perform static (slow-speed) testing of this device
• Test off-chip circuitry and board-level interconnections
†
Some of the figures and tables in this chapter were reproduced from Standard 1149.1-1990, IEEE Standard Test
Access Port and Boundary-Scan Architecture, Copyright 1993 by the Institute of Electrical and Electronics Engineers,
Inc., with the permission of the IEEE.
18-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
• Place all device output pins into their inactive drive (high-impedance) state, allowing
external hardware to drive connections that the processor normally drives
The test-logic unit (Figure 18-1) is fully compliant with IEEE Standard 1149.1. It consists of the
test access port (TAP), the test access port controller, the instruction register (IR), and three data
registers (IDCODE, BYPASS, and BOUND). It also includes logic for generating necessary
clock and control signals.
TDI
IR Register
TCK
TMS
TAP
Controller
TRST#
IDCODE Register
BYPASS Register
Output
Stage
TDO
BOUND Register
. . . . . .
A2340-01
Figure 18-1. Test Logic Unit Connections
18-2
JTAG TEST-LOGIC UNIT
18.2 TEST-LOGIC UNIT OPERATION
18.2.1 Test Access Port (TAP)
The test access port consists of five dedicated pins (four inputs and one output). It is through these
pins that all communication with the test-logic unit takes place. This unit has its own clock (TCK)
and reset (TRST#) pins, so it is independent of the rest of the device. The test-logic unit can read
or write its registers even if the rest of the device is in reset or powerdown.
CAUTION
The JTAG Test-Logic Unit must be reset upon power-up using the TRST# pin.
(To do this, invert the RESET signal and send this inverted RESET to the
TRST# pin). If this is not done, the processor may power-up with the JTAG
test-logic unit in control of the device pins, and the system does not initialize
properly.
The test-logic unit allows you to shift test instructions and test data into the device and to read the
results of the test. A tester (that is, an external bus master such as automatic test equipment or a
component that interfaces to a higher-level test bus) controls the TAP controller’s operation by
applying signals to the clock (TCK) and test-mode-select (TMS) inputs. Instructions and data are
shifted serially from the test-data input (TDI) to the test-data output (TDO). Table 18-1 describes
the test access port pins.
Table 18-1. Test Access Port Dedicated Pins
Pin
Description
Name
TCK
Test Clock Input:
Provides the clock input for the test-logic unit. An external signal must provide a maximum
input frequency of one-half the CLK2 input frequency. TCK is driven by the test-logic unit’s
control circuitry.
TDI
Test Data Input:
Serial input for test instructions and data. Sampled on the rising edge of TCK; valid only when
either the instruction register or a data register is being serially loaded (SHIFT-IR, SHIFT-DR).
TDO
Test Data Output:
Serial output for test instructions and data. TDO shifts out the contents of the instruction
register or the selected data register (LSB first) on the falling edge of TCK. If serial shifting is
not taking place, TDO floats.
TMS
Test Mode Select Input:
Controls the sequence of the TAP controller’s states. Sampled on the rising edge of TCK.
TRST#
Test Reset Input:
Resets the TAP controller. Asynchronously clears the data registers and initializes the
instruction register to 0010 (the IDCODE instruction opcode).
NOTE: The JTAG Test-Logic Unit must be reset upon power-up using the TRST# pin. If this is not done
the processor may power-up with the JTAG test-logic unit in control of the device pins, and the
system does not initialize properly.
18-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
18.2.2 Test Access Port (TAP) Controller
The TAP controller is a finite-state machine that is capable of 16 states (Figure 18-2). Three of its
states provide the basic actions required for testing:
• Applying stimulus (update-data-register)
• Executing a test (run-test/idle)
• Capturing the response (capture-data-register)
Its remaining states support loading instructions, shifting information toward TDO, scanning
pins, and pausing to allow time for the tester to perform other operations.
The TAP controller changes state only in response to the assertion of the test-reset input (TRST#)
or the state of the mode-select pin (TMS) on the rising edge of TCK. TRST# causes the TAP con-
troller to enter its test-logic-reset state, and the state of TMS on the rising edge of TCK controls
the subsequent states.Table 18-2 describes the states and Figure 18-2 illustrates how the TAP state
machine moves from one state to another.
Table 18-2. TAP Controller State Descriptions (Sheet 1 of 2)
Next State
(on TCK Rising Edge)
State
Description
TMS = 0
TMS = 1
Resets the test-logic unit and forces the IDCODE
instruction into the instruction register. (In
Test-Logic-Reset
components that have no IDCODE instruction, the
BYPASS instruction is loaded instead.) Test logic is
disabled; the device is in normal operating mode.
Run-Test/Idle Test-Logic-Reset
Run-Test/Idle Select-DR-Scan
Run-Test/Idle
Executes a test or disables the test logic.
Selects the data register to be placed in the serial
path between TDI and TDO.
Select-DR-Scan
Capture-DR
Shift-DR
Select-IR-Scan
Exit1-DR
Parallel loads data into the active data register, if
necessary. Otherwise, the active register retains its
previous state.
Capture-DR
The active register shifts data one stage toward
TDO on each TCK rising edge.
Shift-DR
Exit1-DR
Pause-DR
Exit2-DR
Shift-DR
Pause-DR
Pause-DR
Shift-DR
Exit1-DR
Update-DR
Exit2-DR
The active register retains its previous state.
The active register temporarily stops shifting data
and retains its previous state.
The active register retains its previous state.
Update-DR
Applies stimulus to the device. Data is latched onto
the active register’s parallel output on the falling
edge of TCK. If the register has no parallel output, it
retains its previous state.
Update-DR
Run-Test/Idle Select-DR-Scan
NOTE: By convention, the abbreviation DR stands for data register, and IR stands for instruction register.
The active register is the register that the current instruction has placed in the serial path between
TDI and TDO.
18-4
JTAG TEST-LOGIC UNIT
Table 18-2. TAP Controller State Descriptions (Sheet 2 of 2)
Next State
(on TCK Rising Edge)
State
Description
TMS = 0
TMS = 1
Test-logic is idle and the instruction register retains
its previous state.
Select-IR-Scan
Capture-IR
Capture-IR
Shift-IR
Test-Logic-Reset
Loads the SAMPLE/PRELOAD instruction
instruction (0001) into the instruction register.
Exit1-IR
Exit1-IR
Shifts the SAMPLE/PRELOAD instruction one
stage toward TDO while shifting the new instruction
in from TDI on each rising edge of TCK.
Shift-IR
Shift-IR
Exit1-IR
Pause-IR
Exit2-IR
The instruction register retains its previous state.
Pause-IR
Pause-IR
Shift-IR
Update-IR
Exit2-IR
The instruction register temporarily stops shifting
and retains its previous state.
The instruction register retains its previous state.
Update-IR
Latches the current instruction onto the instruction
register’s parallel output on the falling edge of TCK.
Update-IR
Run-Test/Idle Select-DR-Scan
NOTE: By convention, the abbreviation DR stands for data register, and IR stands for instruction register.
The active register is the register that the current instruction has placed in the serial path between
TDI and TDO.
For example, assume that the TAP controller is in its test-logic-reset state and you want it to start
shifting the contents of the instruction register from TDI toward TDO (Shift-IR state). This state
change requires a zero, two ones, then two zeros on TMS at the next five rising edges of TCK
(see Table 18-3). By supplying the proper values in the correct sequence, you can move the TAP
controller from any state to any other state.
Table 18-3. Example TAP Controller State Selections
Initial State
TMS Value at TCK Rising Edge Resulting State
0
Test-Logic-
Reset
Run-Test/Idle
1
1
Run-Test/Idle
Select-DR-Scan
Select-IR-Scan
Select-DR-
Scan
0
0
Select-IR-
Scan
Capture-IR
Shift-IR
Capture-IR
18-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Test - Logic
1
- Reset
0
1
1
1
Run - Test/
Idle
Select -
DR - Scan
Select -
IR - Scan
0
0
0
1
1
Capture -
DR
Capture -
IR
0
0
0
0
Shift -
DR
Shift -
IR
1
1
1
0
1
0
Exit1 -
DR
Exit1 -
IR
0
0
Pause -
DR
Pause -
IR
1
1
0
0
Exit2 -
DR
Exit2 -
IR
1
1
Update -
DR
Update -
IR
1
0
1
0
A2356-01
Figure 18-2. TAP Controller (Finite-State Machine)
18-6
JTAG TEST-LOGIC UNIT
18.2.3 Instruction Register (IR)
An instruction opcode is clocked serially through the TDI pin into the four-bit instruction register
(Figure 18-3). The instruction determines which data register is affected. Table 18-4 lists the in-
structions with their binary opcodes, descriptions, and associated registers.
Instruction Register
IR
Reset State
(Using TRST#):
02H
3
0
INST3
INST2
INST1
INST0
Bit
Number
Bit
Mnemonic
Function
Instruction opcode. At reset (using TRST#, or after 5 TCK cycles with
3–0
INST3:0
TMS held low), this field is loaded with 0010, the opcode for the IDCODE
instruction. Instructions are shifted into this field serially through the TDI
pin. (Table 18-4 lists the valid instruction opcodes.)
Figure 18-3. Instruction Register (IR)
Table 18-4. Test-logic Unit Instructions
Affected
Register
Mnemonic
BYPASS
EXTEST
Opcode†
Description
1111
Bypass on-chip system logic (mandatory instruction).
Used for those components that are not being tested.
BYPASS
0000
0001
Off-chip circuitry test (mandatory instruction).
BOUND
BOUND
IDCODE
Used for testing device interconnections on a board.
Sample pins/preload data (mandatory instruction).
SAMPRE
IDCODE
Used for controlling (preload) or observing (sample) the signals at
device pins. This test has no effect on system operation.
0010
1000
ID code test (optional instruction).
Used to identify devices on a board.
High-impedance/On-Circuit Emulation (ONCE) mode test
(optional instruction).
HIGHZ
BYPASS
Used to place device pins into their inactive drive states. Allows
external components to drive signals onto connections that the
processor normally drives.
†
The opcode is the sequence of data bits shifted serially into the instruction register (IR) from the TDI input.
The opcodes for EXTEST and BYPASS are mandated by IEEE 1149.1, so they should be the same for all
JTAG-compliant devices. The remaining opcodes are designer-defined, so they may vary among devices.
All unlisted opcodes are reserved. Use of reserved opcodes could cause the device to enter reserved
factory-test modes.
18-7
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
18.2.4 Data Registers
The test-logic unit uses three data registers: bypass, identification code, and boundary-scan. The
instruction determines which data register is used.
The single-bit bypass register (BYPASS) provides a minimal-length serial path between TDI and
TDO. During board-level testing, you can use this path for any devices that are not currently un-
der test. This speeds access to the data registers for the devices that are being tested.
The 32-bit identification code register (IDCODE) identifies a device by manufacturer, part num-
ber, and version number. Figure 18-4 describes the register and shows the values for the Intel386
EX processor.
Identification Code Register
IDCODE
2027 0013H (3V)
2827 0013H (5V)
Reset State:
31
24
0
0
1
0
0 (3V)
0
0
0
1 (5V)
23
15
7
16
8
0
0
0
0
0
0
1
0
0
0
0
1
0
0
0
1
0
0
1
0
1
1
0
1
0
Bit
Number
Bit
Mnemonic
Function
31–28
27–12
11–1
0
V3:0
Device version number.
Device part number.
PN15:0
MFR10:0
IDP
Manufacturer identification (compressed JEDEC106-A code).
Identification Present. Always true for this device.
This is the first data bit shifted out of the device during a data scan
immediately following an exit from the test-logic-reset state. A one
indicates that an IDCODE register is present. (A zero originates from the
BYPASS register and indicates that the device being interrogated has no
IDCODE register.)
Figure 18-4. Identification Code Register (IDCODE)
18-8
JTAG TEST-LOGIC UNIT
The boundary-scan register (BOUND) holds data to be applied to the pins or data observed at the
pins. Each bit corresponds to a specific pin (Table 18-5).
Table 18-5. Boundary-scan Register Bit Assignments
Bit
Pin
Bit
Pin
Bit
Pin
Bit
Pin
0
1
M/IO#
D/C#
W/R#
READY#
BS8#
RD#
25
26
27
28
29
30
31
32
33
34
35
36
37
A15
A16/CAS0
A17/CAS1
A18/CAS2
A19
50
51
52
53
54
55
56
57
58
59
60
61
62
TMROUT2
TMRGATE2
INT4/TMRCLK0
INT5/TMRGATE0
INT6/TMRCLK1
INT7/TMRGATE1
STXCLK
75
76
77
78
79
80
81
82
83
84
85
86
87
P2.2
P2.3
2
P2.4
3
DACK0#
P2.5/RXD0
P2.6/TXD0
P2.7
4
5
A20
6
WR#
A21
7
BLE#
BHE#
ADS#
NA#
A22
FLT#
UCS#
8
A23
P1.0
CS6#/REFRESH#
LBA#
9
A24
P1.1
10
11
12
A25
P1.2
D0
A1
SMI#
P1.3
D1
P3.0/TMROUT0/
INT9
A2
A3
P1.4
P1.5
D2
D3
13
38
P3.1/TMROUT1/
INT8
63
88
14
15
16
17
18
19
20
21
22
23
24
A4
A5
39
40
41
42
43
44
45
46
47
SRXCLK
SSIORX
64
65
66
67
68
69
70
71
72
P1.6/HOLD
RESET
89
90
91
92
93
94
95
96
97
98
99
100
D4
D5
A6
SSIOTX
P1.7/HLDA
DACK1#/TXD1
EOP#
D6
A7
P3.2/INT0
D7
A8
P3.3/INT1
D8
A9
P3.4/INT2
WDTOUT
DRQ0
D9
A10
A11
A12
A13
A14
P3.5/INT3
D10
D11
D12
D13
D14
D15
P3.6/PWRDOWN
P3.7/SERCLK
DRQ1/RXD1
SMIACT#
P2.0
48 PEREQ/TMRCLK2 73
49 NMI 74
P2.1
NOTES:
1. Bit 0 is closest to TDI; bit 100 is closest to TDO.
2. The boundary-scan chain consists of 101 bits; however, each bit has both a control cell and a data cell,
so an EXTEST instruction requires 202 shifts (101 bits × 2 cells).
18-9
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
18.3 TESTING
This section explains how to use the test-logic unit to test the device and the board interconnec-
tions. For any test, you must load an instruction and perform an instruction-scan cycle, then sup-
ply the correct sequence of ones and zeros to move the TAP controller through the required states
to perform the test.
18.3.1 Identifying the Device
The IDCODE instruction allows you to determine the contents of a device’s IDCODE register.
When TRST# is asserted, the test-logic-reset state forces the IDCODE instruction into the in-
struction register’s parallel output latches. You can also load this instruction like any other, by ma-
nipulating the TDI input to supply the binary opcode (0010). The Capture-DR state loads the
identification code into the IDCODE register, and the Shift-DR state shifts the value out.
18.3.2 Bypassing Devices on a Board
The BYPASS instruction allows you to bypass one or more devices on a board while testing oth-
ers. This significantly reduces the time required for a test. For example, assume that a board has
100 devices, each of which has 101 bits in its boundary-scan register. If the boundary-scan cells
are all connected in series, the boundary-scan path is 10,100 stages long. Bypassing devices al-
lows you to shorten the path considerably. If you set 99 of the devices to shift through their bypass
registers and only a single chip to shift through its boundary-scan register (101 bits in this case),
the serial path is only 200 stages long.
You load the BYPASS instruction by manipulating TDI to supply the binary opcode (1111). The
Capture-DR state loads a logic 0 into the bypass register and the Shift-DR state shifts the value
out.
18.3.3 Sampling Device Operation and Preloading Data
The SAMPLE/PRELOAD instruction has two functions: SAMPLE takes a snapshot of data flow-
ing from (or to) the system pins to (or from) on-chip system logic, while PRELOAD places an
initial data pattern at the latched parallel outputs of the boundary-scan register cells in preparation
for another boundary-scan test operation.
You load the SAMPLE/PRELOAD instruction by manipulating TDI to supply the binary opcode
(0001). The Shift-DR state places the boundary-scan register in the serial path between TDI and
TDO, the Capture-DR state loads the pin states into the boundary-scan register, and the Update-
DR state loads the shift-register contents into the boundary-scan register’s parallel outputs.
18.3.4 Testing the Interconnections (EXTEST)
The EXTEST instruction allows testing of off-chip circuitry and board-level interconnections.
Boundary-scan cells at the system outputs are used to apply test stimuli, while cells at system in-
puts capture the results. The Capture-DR state captures input pins into the chain; the Update-DR
state drives the new values of the parallel output onto the output pins.
18-10
JTAG TEST-LOGIC UNIT
Typically, you would use the SAMPLE/PRELOAD instruction to load data onto the boundary-
scan register’s latched parallel outputs before loading the EXTEST instruction. You load the EX-
TEST instruction by manipulating TDI to supply the binary opcode (0000). The Update-DR state
drives the preloaded data onto the pins for the first test. Stimuli for the remaining tests are shifted
in while the results for the completed tests are shifted out.
18.3.5 Disabling the Output Drivers
The HIGHZ instruction places all system logic outputs into an inactive drive (high impedance)
state. This state allows an in-circuit emulator to drive signals onto connections that processor out-
puts normally drive, without risk of damaging the processor. It also allows you to connect a data
source (such as a test chip) to board-level signals (such as an array of memory devices) that the
processor outputs normally drive. During normal operation, the processor outputs would be ac-
tive, while the test chip outputs would be inactive. During testing, you would use the HIGHZ in-
struction to place the processor outputs into an inactive drive state, then enable the test chip to
drive the connections.
You load the HIGHZ instruction by manipulating the TDI input to supply the binary opcode
(1000). The Capture-DR state loads a logic 0 into the bypass register, and the Shift-DR state shifts
the value out.
18-11
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
18.4 TIMING INFORMATION
The test-logic unit’s input/output timing is as specified in IEEE 1149.1. Figure 18-5 shows the
pin timing associated with loading the instruction register and Figure 18-6 shows the timing for
loading a given data register.
TCK
TMS
Controller State
TDI
Data Input to IR
IR Shift-Register
Parallel Output of IR
Data Input to TDR
TDR Shift-Register
Parallel Output of TDR
Register Selected
TDO Enable
IDCode
New Instruction
Old Data
Inactive
Instruction Register
Inactive
Inactive
Active
Active
TDO
= Don't care or undefined.
A2361-01
Figure 18-5. Internal and External Timing for Loading the Instruction Register
18-12
JTAG TEST-LOGIC UNIT
TCK
TMS
Controller State
TDI
Data Input to IR
IR Shift-Register
Parallel Output of IR
Data Input to TDR
TDR Shift-Register
Parallel Output of TDR
Instruction Register
TDO Enable
IDCode
Instruction
New Data
Inactive
Old Data
Test Data Register
Inactive
Active
Active
Inactive
TDO
= Don't care or undefined.
A2362-01
Figure 18-6. Internal and External Timing for Loading a Data Register
18-13
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
18.5 DESIGN CONSIDERATIONS
This section outlines considerations for the test-logic unit.
• The JTAG Test-Logic Unit must be reset upon power-up using the TRST# pin. (To do this,
invert the RESET signal and send this inverted RESET to the TRST# pin). If this is not
done, the processor may power-up with the JTAG test-logic unit in control of the device
pins, and the system does not initialize properly.
• For system-level in-circuit emulation, use the HIGHZ instruction to enter ONCE mode. For
device-level in-circuit emulation, you assert the FLT# pin to enter ONCE mode. This
method can interfere with the test-logic unit’s parallel functions, although it does not affect
the shifting functions or the TDO output.
18-14
A
SIGNAL
DESCRIPTIONS
APPENDIX A
states of certain pins during reset, idle, powerdown, and hold. The information is presented in four
tables:
• Table A-1 defines the abbreviations used in Table A-2 to describe the signals.
• Table A-2 describes each signal.
• Table A-3 defines the abbreviations used in Table A-4 to describe the pin states.
• Table A-4 lists the states of output and bidirectional pins after reset and during idle mode,
powerdown, and hold.
Table A-1. Signal Description Abbreviations
Abbreviation
Definition
signal is active low
#
—
I
not applicable or none
standard TTL input
standard CMOS output
open-drain output
bidirectional (input and output)
Schmitt-trigger input
power pin
O
OD
I/O
ST
P
G
ground pin
A-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table A-2 is an alphabetical list of the signals available at the device pins. The Multiplexed With
column lists other signals that share a pin with the signal listed in the Signal column.
Table A-2. Description of Signals Available at the Device Pins (Sheet 1 of 6)
Multiplexed With
Signal
Type
Name and Description
(Alternate
Function)
A25:19
O
Address Bus:
—
A18:16
A15:1
CAS2:0
—
Outputs physical memory or port I/O addresses. These
signals are valid when ADS# is active and remain valid until
the next T1, T2P, or Ti.
ADS#
O
Address Status:
—
Indicates that the processor is driving a valid bus-cycle
definition and address (W/R#, D/C#, M/IO#, A25:1, BHE#,
BLE#) onto its pins.
BHE#
BLE#
O
O
I
Byte High Enable:
—
Indicates that the processor is transferring a high data byte.
Byte Low Enable:
—
Indicates that the processor is transferring a low data byte.
BS8#
Bus Size:
—
Indicates that an 8-bit device is currently being addressed.
BUSY#
I
Busy:
TMRGATE2
Indicates that the math coprocessor is busy. If BUSY# is
sampled low at the falling edge of RESET, the processor
performs an internal self test.
CAS2:0
CLK2
O
Cascade Address:
A18:16
—
Carries the slave address information from the master
8259A interrupt module during interrupt acknowledge bus
cycles.
ST
Input Clock:
Is connected to an external clock that provides the
fundamental timing for the microprocessor. The internal
processor clock frequency is half the CLK2 frequency.
CLKOUT
COMCLK
O
I
Clock Output:
—
Use this output to synchronize external devices with the
processor.
SIO Baud Clock:
P3.7
An external source connected to this pin can clock the SIOn
baud-rate generator.
CS6#
CS5#
CS4#
CS3#
CS2#
CS1#
CS0#
O
Chip-selects:
REFRESH#
DACK0#
P2.4
P2.3
P2.2
Activated when the address of a memory or I/O bus cycle is
within the address region programmed by the user.
P2.1
P2.0
A-2
SIGNAL DESCRIPTIONS
Table A-2. Description of Signals Available at the Device Pins (Sheet 2 of 6)
Multiplexed With
(Alternate
Signal
Type
Name and Description
Function)
CTS1#
CTS0#
I
Clear to Send:
EOP#
P2.7
Indicates that the modem or data set is ready to exchange
data with the SIO channel.
D15:0
I/O
Data Bus:
—
Inputs data during memory read, I/O read, and interrupt
acknowledge cycles; outputs data during memory write and
I/O write cycles. During reads, data is latched during the
falling edge of phase 2 of T2, T2P, or T2i. During writes, this
bus is driven during phase 2 of T1 and remains active until
phase 2 of the next T1, T1P, or Ti.
DACK1#
DACK0#
O
DMA Channel Acknowledge:
TXD1
CS5#
Indicates that the DMA channel is ready to service the
requesting device. An external device uses the DRQn pin to
request DMA service; the DMA uses the DACKn pin to
indicate that the request is being serviced.
D/C#
O
I
Data/Control:
—
Indicates whether the current bus cycle is a data cycle
(memory or I/O read or write) or a control cycle (interrupt
acknowledge, halt/shutdown, or code fetch).
DCD1#
DCD0#
Data Carrier Detect:
DRQ0
P1.0
Indicates that the modem or data set has detected the SIO
channel’s data carrier.
DRQ1
DRQ0
I
I
DMA External Request:
RXD1
DCD1#
Indicates that an external device requires DMA service.
DSR1#
DSR0#
Data Set Ready:
STXCLK
P1.3
Indicates that the modem or data set is ready to establish
the communications link with the SIO channel.
DTR1#
DTR0#
O
Data Terminal Ready:
SRXCLK
P1.2
Indicates that the SIO channel is ready to establish a
communications link with the modem or data set.
EOP#
I/OD End-of-process:
As an input, this signal terminates a DMA transfer. As an
CTS1#
ouput, it indicates that a DMA transfer has completed.
ERROR#
FLT#
I
I
Error:
TMROUT2
—
Indicates the the math coprocessor has an error condition.
Float:
Forces all bidirectional and output signals except TDO to a
high-impedance state.
HLDA
O
Hold Acknowledge:
P1.7
Indicates that the processor has relinquished local bus
control to another bus master in response to a HOLD
request.
A-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table A-2. Description of Signals Available at the Device Pins (Sheet 3 of 6)
Multiplexed With
Signal
Type
Name and Description
(Alternate
Function)
HOLD
I
Hold Request:
P1.6
An external bus master asserts HOLD to request control of
the local bus. The processor finishes the current nonlocked
bus transfer, releases the bus signals, and asserts HLDA.
INT9
INT8
INT7
INT6
INT5
INT4
INT3
INT2
INT1
INT0
I
Interrupt Requests:
P3.0/TMROUT0
P3.1/TMROUT1
TMRGATE1
TMRCLK1
TMRGATE0
TMRCLK0
P3.5
P3.4
P3.3
P3.2
These maskable inputs cause the processor to suspend
execution of the current program and execute an interrupt
acknowledge cycle.
LBA#
O
Local Bus Access:
—
Indicates that the processor provides the READY# signal
internally to terminate a bus transaction. This signal is
active when the processor accesses an internal peripheral
or when the chip-select unit provides the READY# signal for
an external peripheral.
LOCK#
M/IO#
O
O
Bus Lock:
P1.5
—
Prevents other bus masters from gaining control of the bus.
Memory/IO:
Indicates whether the current bus cycle is a memory cycle
or an I/O cycle.
NA#
NMI
I
Next Address:
—
—
Requests address pipelining.
ST
Nonmaskable Interrupt Request:
Causes the processor to suspend execution of the current
program and execute an interrupt acknowledge cycle.
PEREQ
I
Processor Extension Request:
TMRCLK2
Indicates that the math coprocessor has data to transfer to
the processor.
P1.7
P1.6
P1.5
P1.4
P1.3
P1.2
P1.1
P1.0
I/O
Port 1:
HLDA
HOLD
LOCK#
RI0#
DSR0#
DTR0#
RTS0#
DCD0#
General-purpose, bidirectional I/O port.
A-4
SIGNAL DESCRIPTIONS
Table A-2. Description of Signals Available at the Device Pins (Sheet 4 of 6)
Multiplexed With
(Alternate
Signal
Type
Name and Description
Function)
P2.7
I/O
Port 2:
CTS0#
TXD0
RXD0
CS4#
CS3#
CS2#
CS1#
CS0#
P2.6
P2.5
P2.4
P2.3
P2.2
P2.1
P2.0
General-purpose, bidirectional I/O port.
P3.7
P3.6
P3.5
P3.4
P3.3
P3.2
P3.1
P3.0
I/O
Port 3:
COMCLK
PWRDOWN
INT3
INT2
INT1
INT0
TMROUT1/INT8
TMROUT0/INT9
General-purpose, bidirectional I/O port.
PWRDOWN
O
O
Powerdown Output:
P3.6
Indicates that the device is in powerdown mode.
RD#
Read Enable:
Indicates that the current bus cycle is a read cycle and the
data bus is able to accept data.
READY#
I/O
Ready:
—
Terminates the current bus cycle. The processor drives
READY# when LBA# is active; otherwise, the processor
samples READY# on the falling edge of phase 2 of T2, T2P
or T2i.
REFRESH#
RESET
O
ST
I
Refresh:
CS6#
—
Indicates that a refresh bus cycle is in progress and that the
refresh address is on the bus for the DRAM controller.
System Reset Input:
Suspends any operation in progress and places the
processor into a known reset state.
RI1#
RI0#
Ring Indicator:
SSIORX
P1.4
Indicates that the modem or data set has received a
telephone ringing signal.
RTS1#
RTS0#
O
I
Request to Send:
SSIOTX
P1.1
Indicates that the SIO channel is ready to exchange data
with the modem or data set.
RXD1
RXD0
Receive Data:
DRQ1
P2.5
Accepts data from the modem or data set to the SIO
channel.
A-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table A-2. Description of Signals Available at the Device Pins (Sheet 5 of 6)
Multiplexed With
Signal
Type
Name and Description
(Alternate
Function)
SMI#
ST
System Management Interrupt:
—
—
Causes the device to enter System Management Mode.
SMI# is the highest priority external interrupt.
SMIACT#
SRXCLK
O
System Management Interrupt Active:
Indicates that the processor is in System Management
Mode.
I/O
SSIO Receive Clock:
DTR1#
In master mode, the baud-rate generator’s output appears
on SRXCLK and can be used to clock a slave transmitter. In
slave mode, SRXCLK functions as an input clock for the
receiver.
SSIORX
SSIOTX
STXCLK
I
SSIO Receive Serial Data:
RI1#
Accepts serial data (most-significant bit first) into the SSIO.
O
SSIO Transmit Serial Data:
RTS1#
DSR1
Sends serial data (most-significant bit first) from the SSIO.
I/O
SSIO Transmit Clock:
In master mode, the baud-rate generator’s output appears
on STXCLK and can be used to clock a slave receiver. In
slave mode, STXCLK functions as an input clock for the
transmitter.
TCK
TDI
I
I
Test Clock Input:
—
—
Provides the clock input for the test-logic unit.
Test Data Input:
Serial input for test instructions and data. Sampled on the
rising edge of TCK; valid only when either the instruction
register or a data register is being serially loaded.
TDO
O
Test Data Output:
—
Serial output for test instructions and data. TDO shifts out
the contents of the instruction register or the selected data
register (LSB first) on the falling edge of TCK. If serial
shifting is not taking place, TDO floats.
TMRCLK2
TMRCLK1
TMRCLK0
I
I
Timer/Counter Clock Input:
PEREQ
INT6
INT4
An external clock source connected to the TMRCLKn pin
can drive the corresponding timer/counter.
TMRGATE2
TMRGATE1
TMRGATE0
Timer/Counter Gate Input:
BUSY#
INT7
INT5
Can control the counter’s operation (enable, disable, or
trigger, depending on the programmed mode).
TMROUT2
TMROUT1
TMROUT0
O
Timer/Counter Output:
ERROR#
P3.1/INT8
P3.0/INT9
Can provide the timer/counter’s output. The form of the
output depends on the programmed mode.
A-6
SIGNAL DESCRIPTIONS
Table A-2. Description of Signals Available at the Device Pins (Sheet 6 of 6)
Multiplexed With
(Alternate
Signal
Type
Name and Description
Function)
TMS
I
Test Mode Select:
—
Controls the sequence of the test-logic unit’s TAP controller
states. Sampled on the rising edge of TCK.
TRST#
ST
Test Reset:
—
Resets the test-logic unit’s TAP controller. Asynchronously
clears the data registers and initializes the instruction
register to 0010 (the IDCODE instruction opcode).
TXD1
TXD0
O
O
Transmit Data:
DACK1#
P2.6
Transmits serial data from the corresponding SIO channel.
UCS#
Upper Chip-select:
—
—
—
Activated when the address of a memory or I/O bus cycle is
within the address region programmed by the user.
VCC
P
System Power:
Provides the nominal DC supply input. Connected
externally to a VCC board plane.
VSS
G
System Ground:
Provides the 0 volt connection from which all inputs and
outputs are measured. Connected externally to a ground
board plane.
WDTOUT
W/R#
O
O
Watchdog Timer Output:
—
—
Indicates that the watchdog timer has expired.
Write/Read:
Indicates whether the current bus cycle is a write cycle or a
read cycle.
WR#
O
Write Enable:
—
Indicates that the current bus cycle is a write cycle.
A-7
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table A-3 defines the abbreviations used in Table A-4 to describe the pin states.
Table A-3. Pin State Abbreviations
Abbreviation
Description
Output driven to VCC
1
0
Output driven to VSS
Output floats
Z
Q
Output remains active
Output retains current state
X
WH
Pin floats and has a temporary weak
pull-up
WL
Pin floats and has a temporary weak
pull-down
A-8
SIGNAL DESCRIPTIONS
Table A-4 lists the states of output and bidirectional pins after reset and during idle mode, pow-
erdown, and hold.
Table A-4. Pin States After Reset and During Idle, Powerdown, and Hold (Sheet 1 of 2)
Pin State
Symbol
Type
Reset
Idle
Powerdown
Hold
A25:1
ADS#
BHE#
BLE#
O
O
1
1
1
1
1
Z
Z
Z
Z
Z
Q
1
1
O
0
X
X
1
0
O
0
1
CAS2:0
CLKOUT
CS4:0#
CS6:5#
D15:0
O
1
1
O
Q
Q
Q
Q
Z
0
O
WH
1
X
O
X
1
I/O
O
Z
Z
Z
1
DACK1:0#
D/C#
1
Q
0
X
O
1
0
Z
X
Z
1
DTR1:0
EOP#
O
WH
WH
WL
1
X
Z
X
I/OD
O
Z
HLDA
Q
Q
X
1
X
LBA#
O
X
1
LOCK#
M/IO#
O
WH
0
X
Z
Z
X
X
X
X
X
X
Q
1
O
1
P1.5:0
I/O
I/O
I/O
I/O
I/O
I/O
O
WH
WL
WH
WL
WH
WL
WL
1
X
X
X
X
X
X
X
1
X
P1.7:6
X
P2.4:0
X
P2.6:5
X
P2.7
X
P3.7:0
X
PWRDWN
RD#
1
O
1
READY#
REFRESH#
RTS1#
RTS0#
SMIACT#
SRXCLK
SSIOTX
NOTES:
I/O
O
Z
Z
Z
Z
1
1
Q
X
X
X
Q
Q
X
O
WL
WH
1
X
X
X
X
1
O
O
X
I/O
O
WH
WL
X or Q(1)
X or Q(1)
Q
Q
1. X if clock source is internal; Q if clock source is external.
2. Q when shifting data out through the JTAG port, otherwise Z.
A-9
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table A-4. Pin States After Reset and During Idle, Powerdown, and Hold (Sheet 2 of 2)
STXCLK
TDO
I/O
O
O
O
O
O
O
O
O
O
WH
Q
X or Q(1)
Z or Q(2)
X or Q(1)
X or Q(1)
X or Q(1)
X or Q(1)
X
Q
Z or Q(2)
Z or Q(2)
Z or Q(2)
TMROUT2
TMROUT1:0
TXD1
WH
WL
1
Q
Q
Q
Q
Q
Q
1
Q
Q
Q
Q
1
TXD0
WL
0
UCS#
WDTOUT
W/R#
0
X
Q
Z
0
1
WR#
1
1
1
1
NOTES:
1. X if clock source is internal; Q if clock source is external.
2. Q when shifting data out through the JTAG port, otherwise Z.
The following input pins have permanent weak pull-up resistors: TCK, TDI, TMS, TRST#,
SMI#, PEREQ/TMRCLK2, and FLT#.
A-10
B
COMPATIBILITY
WITH THE PC/AT*
ARCHITECTURE
APPENDIX B
COMPATIBILITY WITH THE PC/AT*
ARCHITECTURE
The Intel386™ EX embedded processor is NOT 100% PC/AT* compatible. Due to compatibility
issues, not all PC software executes on the Intel386 EX processor. In addition, not all ISA/PC-
104 cards operate in an Intel386 EX processor system.
It is the responsibility of the designer to determine if a specific PC/AT software or hardware pack-
functionality than a desktop PC. It is important for the designer to evaluate the requirements of
the PC software or hardware that is expected to operate on the Intel386 EX processor system.
This appendix is organized as follows:
• Hardware Departures from PC/AT System Architecture (see below)
• Software Considerations for a PC/AT System Architecture (page B-5)
B.1 HARDWARE DEPARTURES FROM PC/AT SYSTEM ARCHITECTURE
This appendix describes the areas in which the Intel386 EX processor departs from a standard
PC/AT system architecture and explains how to work around those departures if necessary. Chap-
ter 5, “DEVICE CONFIGURATION”, shows an example configuration for a PC/AT-compatible
system
B.1.1 DMA Unit
The PC/AT architecture uses two 8237A DMA controllers, connected in cascade, for a total of
seven channels. One DMA controller allows byte transfers and the other allows word transfers.
However, the 8237A has two major restrictions:
• It has only 16-bit addressing capability. This requires a page register to allow address
extension for a system based on a processor like the Intel386 EX processor, with 26-bit (64
Mbyte) physical memory addressing capability. A page register implementation is
cumbersome and degrades the system performance.
• The 8237A has no natural two-cycle data transfer mode to allow memory-to-memory
transfers. Instead, two DMA channels have to be used in a very specific manner.
Transferring data between memory and memory-mapped I/O devices, common in
embedded applications, would not be easy.
B-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
To eliminate these problems with an 8237A DMA controller, the Intel386 EX processor inte-
grates a DMA controller unit that differs from the 8237A DMA in these ways:
• It provides two channels, each capable of either byte or word transfers.
• Each channel can transfer data between any combination of memory and I/O. The Bus
Interface Unit supports both external fly-by and two-cycle operation.
• For programming compatibility, the internal DMA unit preserves all of the 8-bit registers of
the 8237A. The 8237A’s command register bits that affect two-channel memory-to-memory
transfers, compressed timing, and DREQ/DACK signal polarity selection are not supported
by the internal DMA.
• The internal DMA uses 26-bit address registers to support the 26-bit address bus and uses
24-bit byte count registers to support larger data blocks than are possible with the 8237A.
However, each channel can be configured to look like an 8237A with page registers (i.e.,
16-bit address and byte count registers).
Chapter 12, “DMA CONTROLLER,” describes the DMA unit’s features in detail.
While the internal DMA offers a comprehensive set of features to meet the needs of most embed-
ded applications, strict DOS compatibility may be critical to some. A PC/AT compatible system’s
Basic Input Output System (BIOS) only uses the DMA for floppy disk access. Since both MS-
DOS* and Microsoft* Windows* make calls to the BIOS for disk access, it is possible to modify
the BIOS. The floppy disk controller allows data transfers to occur using DMA, Polling, or Inter-
rupt based. A few BIOS vendors have implemented the transfers using polling for disk transfers.
Some programs bypass the BIOS and go directly to the hardware; typically, these are disk inten-
sive programs like hard disk backup software or disk management software.
If more DMA channels are required for compatibility, external controllers could be added. The
Intel386 EX processor’s flexible address remapping scheme enables you to map the internal
DMA out of the DOS I/O space and then connect an external 8237A to achieve PC/AT compati-
bility. The internal DMA can still be used for other non-DOS related functions.
B.1.2 Industry Standard Bus (ISA) Signals
The address, data, and control signals, along with the interrupt and DMA control signals, do not
directly conform to the PC/AT ISA bus. (They more closely match the Intel386™ SX processor
local bus signals.) However, you can easily construct a subset PC/AT ISA bus from these signals
or a combination of these signals. For example, the AEN signal is typically generated as shown
in Figure B-1 in a PC/AT-compatible system.
B-2
COMPATIBILITY WITH THE PC/AT* ARCHITECTURE
HLDA
Processor
AEN
MASTER#
(From PC/AT* Bus)
A2504-01
Figure B-1. Derivation of AEN Signal in a Typical PC/AT System
For systems based on Intel386 EX processor, the AEN signal could be derived as shown in Figure
B-2. Notice that since the DMA acknowledge signals are used instead of a generic HLDA, there
is no need to incorporate the REFRESH# signal in the logic.
DACK0#
DACK
DACK1#
Processor
AEN
MASTER# (From PC/AT* Bus)
A2503-01
Figure B-2. Derivation of AEN Signal for Intel386™ EX processor-based Systems
In a PC/AT system using the 8237A DMA controller in fly-by mode, the 8237A generates appro-
priate control signals for memory (MEMR# or MEMW#) and for I/O (IOW# and IOR#). The
Intel386 EX processor’s internal DMA, during fly-by transfers, generates control signals (M/IO#
and W/R#) that apply to the memory device. There needs to be some external logic that can detect
the DMA operation (through the AEN signal) and generate a complementary I/O cycle. For ex-
ample, if the DMA is generating a memory read cycle and AEN is active, then the logic should
drive the IOW# signal on the PC/AT bus. Actually, the internal DMA could be programmed in a
two-cycle mode eliminating the need for external logic. This will not have a significant impact
on the performance — of the two cycles required to complete the transfer, the I/O cycle is the long
one (meeting PC/AT timings) while the memory cycle is relatively very quick.
The drive capability and the operating frequency of the Intel386 EX processor signals are differ-
ent from the standard PC/AT bus, which requires 24 mA drive capacity at 200 pF capacitive load.
Most PC/AT systems presently operate in a “quiet bus” mode so that non-ISA cycles are not re-
flected on the ISA bus. In a typical implementation, the address/data buses may change states, but
the control signals are not strobed if a non-ISA cycle is detected. External three-state buffers and
some decoding logic are needed to implement this scheme. The EV386EX (evaluation board for
B-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
the Intel386 EX embedded processor) demonstrates the design of a Synchronous Expansion Bus
that is very similar to the ISA bus. The Intel386 EX processor is not capable of providing a 100%
compatible ISA bus due to its lack of DMA channels and interrupt inputs.
B.1.3 Interrupt Control Unit
Interrupt signals IRQ10, IRQ11, and IRQ15 found on an ISA bus are not directly available for
external interrupt connections in systems based on an Intel386 EX processor. If an application
intends to use these IRQn signals, then they can be rerouted to other IRQ signals available in an
Intel386 EX processor architecture, and the respective interrupt handler routines assigned accord-
ingly.
B.1.4 SIO Units
In the modem control register (MCR), the OUT1 register bit is used only in loopback tests. The
OUT2 bit in the MCR is used as an SIO interrupt enable control signal. This allows two additional
UARTs to be added externally as COM3 and COM4.
The SIO units (COM1 and COM2) are connected to the equivalent of a PC’s local bus, not the
ISA bus. However, this does not affect the compatibility with DOS application software in any
form.
B.1.5 CPU-only Reset
The RESET pin on the Intel386 EX processor can be considered to function as a system reset
function because all of the on-chip peripheral units, as well as the CPU core, are initialized to a
known start-up state. There is no separate reset pin that goes only to the CPU. Some CPU-only
reset modes, such as a keyboard controller generated CPU-only reset, will not function as expect-
ed.
A CPU-only reset can be implemented by routing the reset signal to either the NMI or SMI# sig-
nal, and the appropriate handler code could then generate a corresponding CPU-Only-Reset func-
tion by setting bit 0 of the PORT92H register.
B.1.6 HOLD, HLDA Pins
These pins do not connect directly to the CPU. Instead they go to the Bus Arbiter which controls
the internal HOLD and HLDA signals connected to the CPU core. However the presence of the
bus arbiter is transparent as far as functionality of the external HOLD and HLDA pins of Intel386
EX processor are concerned.
In a PC/AT system, if an external bus master gains the bus by raising HOLD to the CPU or raising
DREQ in DMA cascade mode, the corresponding HLDA or DACK signal stays active until the
bus master drops HOLD or DREQ. In the Intel386 EX processor, when the refresh control unit
requests the bus, the bus arbiter deactivates the signals on the HLDA or DACK# pins while the
external bus master still has the bus (HOLD or DREQ is high). At this point, the external bus mas-
ter or DMA must deassert its HOLD or DREQ signal for a minimum of one CPU clock cycle and
it can then assert the signal again.
B-4
COMPATIBILITY WITH THE PC/AT* ARCHITECTURE
B.1.7 Port B
The Port B register found on the PC/AT is not supported on the Intel386 EX processor. It can be
implemented externally with a PLD. The EXPLR1 (Explorer Evaluation board) supports this
Port B.
B.2 SOFTWARE CONSIDERATIONS FOR A PC/AT SYSTEM ARCHITECTURE
B.2.1 Embedded Basic Input Output System (BIOS)
The BIOS provides low-level drivers to interface to the hardware. The BIOS is hardware depen-
dent and typically requires changes for the embedded design. There are several third party BIOS
vendors that support the Intel386 EX processor.
Embedded PC features supported include PCMCIA, Flash, Advanced Power Management
(APM), Source, Remote Floppy, OEM Configurable, and Video/Keyboard rerouted through the
serial port. For a complete list of vendors and their features, call the Intel BBS as described in
“Electronic Support Systems” on page 1-6. A good evaluation vehicle is the EV386EX Evalua-
tion Board which comes with five different third party BIOS demonstrations. New and updated
demonstrations are also available on the Intel BBS.
B.2.2 Embedded Disk Operating System (DOS)
The DOS operating system offers functions for I/O communication, floppy/hard disk, video, key-
board, program handling, memory management, and network support. All these are available to
the Intel386 EX embedded processor user.
Embedded PC DOS features include Advanced Power Management (APM) support, ROMable,
Source, Disk Compression, and XIP. A variety of third party DOS vendors support the Intel386
EX processor. For a complete list of vendors and their features, call the Intel BBS. The EV386EX
Evaluation Board also provides a variety of DOS demonstrations. New and updated DOS dem-
onstrations are also available on the Intel BBS.
B.2.3 Microsoft* Windows*
The Intel386 EX processor can run both Microsoft Windows 3.1 and Microsoft ROM Windows.
Both require RAM and disk space to execute. Other hardware such as a keyboard controller, real
time clock and video controller may be required. For more information on implementing ROM
Windows, refer to Mobile Intel486™ SX CPU PC Designs Using FlashFile™ Components (Or-
der Number 292149).
B-5
C
EXAMPLE CODE
HEADER FILES
APPENDIX C
EXAMPLE CODE HEADER FILES
This appendix contains the header files called by the code examples that are included in several
chapters of this manual. Section C.1 contains the register definitions for each code routine. Sec-
tion C.2 contains the variable definitions.
C.1 REGISTER DEFINITIONS FOR CODE EXAMPLES
/* 80386EX REGISTER DEFINITIONS */
#define _SetEXRegWord(reg,val)
#define _SetEXRegByte(reg,val)
#define _ReadEXRegWord(val,reg)
#define _GetEXRegByte(reg)
(outpw(reg,val))
(outp(reg,val))
(val=inpw(reg))
inp(reg)
#define _GetEXRegWord(reg)
inpw(reg)
/* REMAP ADDRESSING Registers */
#define REMAPCFGH
#define REMAPCFGL
#define REMAPCFG
0x0023
0x0022
0x0022
/* INTERRUPT CONTROL REGISTERS -- SLOT 15 ADDRESSES */
#define ICW1M
#define ICW1S
#define ICW2M
#define ICW2S
#define ICW3M
#define ICW3S
#define ICW4M
#define ICW4S
#define OCW1M
#define OCW1S
#define OCW2M
#define OCW2S
#define OCW3M
#define OCW3S
0xF020
0xF0A0
0xF021
0xF0A1
0xF021
0xF0A1
0xF021
0xF0A1
0xF021
0xF0A1
0xF020
0xF0A0
0xF020
0xF0A0
/* INTERRUPT CONTROL REGISTERS -- SLOT 0 ADDRESSES */
#define ICW1MDOS
#define ICW1SDOS
#define ICW2MDOS
#define ICW2SDOS
#define ICW3MDOS
#define ICW3SDOS
#define ICW4MDOS
#define ICW4SDOS
#define OCW1MDOS
#define OCW1SDOS
#define OCW2MDOS
0x0020
0x00A0
0x0021
0x00A1
0x0021
0x00A1
0x0021
0x00A1
0x0021
0x00A1
0x0020
C-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
#define OCW2SDOS
#define OCW3MDOS
#define OCW3SDOS
0x00A0
0x0020
0x00A0
/* CONFIGURATION Registers */
#define DMACFG
#define INTCFG
#define TMRCFG
#define SIOCFG
#define P1CFG
#define P2CFG
#define P3CFG
#define PINCFG
0xF830
0xF832
0xF834
0xF836
0xF820
0xF822
0xF824
0xF826
/* WATCHDOG TIMER Registers */
#define WDTRLDH
#define WDTRLDL
#define WDTCNTH
#define WDTCNTL
#define WDTCLR
#define WDTSTATUS
0xF4C0
0xF4C2
0xF4C4
0xF4C6
0xF4C8
0xF4CA
/* TIMER CONTROL REGISTERS -- SLOT 15 ADDRESSES */
#define TMR0
#define TMR1
#define TMR2
#define TMRCON
0xF040
0xF041
0xF042
0xF043
/* TIMER CONTROL REGISTERS -- SLOT 0 ADDRESSES */
#define TMR0DOS
#define TMR1DOS
#define TMR2DOS
#define TMRCONDOS
0x0040
0x0041
0x0042
0x0043
/* INPUT/OUTPUT PORT UNIT Registers */
#define P1PIN
#define P1LTC
#define P1DIR
#define P2PIN
#define P2LTC
#define P2DIR
#define P3PIN
#define P3LTC
#define P3DIR
0xF860
0xF862
0xF864
0xF868
0xF86A
0xF86C
0xF870
0xF872
0xF874
/* ASYNCHRONOUS SERIAL CHANNEL 0 -- SLOT 15 ADDRESSES */
#define RBR0
#define THR0
#define TBR0
#define DLL0
#define IER0
#define DLH0
0xF4F8
0xF4F8
0xF4F8
0xF4F8
0xF4F9
0xF4F9
C-2
EXAMPLE CODE HEADER FILES
#define IIR0
#define LCR0
#define MCR0
#define LSR0
#define MSR0
#define SCR0
0xF4FA
0xF4FB
0xF4FC
0xF4FD
0xF4FE
0xF4FF
/* ASYNCHRONOUS SERIAL CHANNEL 0 -- SLOT 0 ADDRESSES */
#define RBR0DOS
#define THR0DOS
#define TBR0DOS
#define DLL0DOS
#define IER0DOS
#define DLH0DOS
#define IIR0DOS
#define LCR0DOS
#define MCR0DOS
#define LSR0DOS
#define MSR0DOS
#define SCR0DOS
0x03F8
0x03F8
0x03F8
0x03F8
0x03F9
0x03F9
0x03FA
0x03FB
0x03FC
0x03FD
0x03FE
0x03FF
/* ASYNCHRONOUS SERIAL CHANNEL 1 -- SLOT 15 ADDRESSES */
#define RBR1
#define THR1
#define TBR1
#define DLL1
#define IER1
#define DLH1
#define IIR1
#define LCR1
#define MCR1
#define LSR1
#define MSR1
#define SCR1
0xF8F8
0xF8F8
0XF8F8
0xF8F8
0xF8F9
0xF8F9
0xF8FA
0xF8FB
0xF8FC
0xF8FD
0xF8FE
0xF8FF
/* ASYNCHRONOUS SERIAL CHANNEL 1 -- SLOT 0 ADDRESSES */
#define RBR1DOS
#define THR1DOS
#define TBR1DOS
#define DLL1DOS
#define IER1DOS
#define DLH1DOS
#define IIR1DOS
#define LCR1DOS
#define MCR1DOS
#define LSR1DOS
#define MSR1DOS
#define SCR1DOS
0x02F8
0x02F8
0x02F8
0x02F8
0x02F9
0x02F9
0x02FA
0x02FB
0x02FC
0x02FD
0x02FE
0x02FF
/* SYNCHRONOUS SERIAL CHANNEL REGISTERS */
#define SSIOTBUF
#define SSIORBUF
#define SSIOBAUD
#define SSIOCON1
0xF480
0xF482
0xF484
0xF486
C-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
#define SSIOCON2
#define SSIOCTR
0xF488
0xF48A
/* CHIP SELECT UNIT Registers */
#define CS0ADL
#define CS0ADH
#define CS0MSKL
#define CS0MSKH
#define CS1ADL
#define CS1ADH
#define CS1MSKL
#define CS1MSKH
#define CS2ADL
#define CS2ADH
#define CS2MSKL
#define CS2MSKH
#define CS3ADL
#define CS3ADH
#define CS3MSKL
#define CS3MSKH
#define CS4ADL
#define CS4ADH
#define CS4MSKL
#define CS4MSKH
#define CS5ADL
#define CS5ADH
#define CS5MSKL
#define CS5MSKH
#define CS6ADL
#define CS6ADH
#define CS6MSKL
#define CS6MSKH
#define UCSADL
#define UCSADH
#define UCSMSKL
#define UCSMSKH
0xF400
0xF402
0xF404
0xF406
0xF408
0xF40A
0xF40C
0xF40E
0xF410
0xF412
0xF414
0xF416
0xF418
0xF41A
0xF41C
0xF41E
0xF420
0xF422
0xF424
0xF426
0xF428
0xF42A
0xF42C
0xF42E
0xF430
0xF432
0xF434
0xF436
0xF438
0xF43A
0xF43C
0xF43E
/* REFRESH CONTROL UNIT Registers */
#define RFSBAD
#define RFSCIR
#define RFSCON
#define RFSADD
0xF4A0
0xF4A2
0xF4A4
0xF4A6
/* POWER MANAGEMENT CONTROL Registers */
#define PWRCON
#define CLKPRS
0xF800
0xF804
/* DMA UNIT REGISTERS -- SLOT 15 ADDRESSES */
#define DMA0TAR
#define DMA0BYC
#define DMA1TAR
#define DMA1BYC
0xF000
0xF001
0xF002
0xF003
C-4
EXAMPLE CODE HEADER FILES
#define DMACMD1
#define DMASTS
#define DMASRR
0xF008
0xF008
0xF009
0xF00A
0xF00B
0xF00C
0xF00D
0xF00E
0xF00F
0xF010
0xF011
0xF012
0xF013
0xF018
0xF019
0xF019
0xF01A
0xF01B
0xF01C
0xF01D
0xF01E
0xF083
0xF085
0xF086
0xF087
0xF098
0xF099
#define DMAMSK
#define DMAMOD1
#define DMACLRBP
#define DMACLR
#define DMACLRMSK
#define DMAGRPMSK
#define DMA0REQL
#define DMA0REQH
#define DMA1REQL
#define DMA1REQH
#define DMABSR
#define DMACHR
#define DMAIS
#define DMACMD2
#define DMAMOD2
#define DMAIEN
#define DMAOVFE
#define DMACLRTC
#define DMA1TARPL
#define DMA1TARPH
#define DMA0TARPH
#define DMA0TARPL
#define DMA0BYCH
#define DMA1BYCH
/* DMA UNIT REGISTERS -- SLOT 0 ADDRESSES */
#define DMA0TARDOS
#define DMA0BYCDOS
#define DMA1TARDOS
#define DMA1BYCDOS
#define DMACMD1DOS
#define DMASTSDOS
#define DMASRRDOS
#define DMAMSKDOS
#define DMAMOD1DOS
#define DMACLRBPDOS
#define DMACLRDOS
0x0000
0x0001
0x0002
0x0003
0x0008
0x0008
0x0009
0x000A
0x000B
0x000C
0x000D
#define DMACLRMSKDOS 0x000E
#define DMAGRPMSKDOS 0x000F
#define DMA1TARPLDOS 0x0083
#define DMA0TARPLDOS 0x0087
/* A20GATE AND FAST CPU RESET -- SLOT 15 ADDRESS */
#define PORT92
/* A20GATE AND FAST CPU RESET -- SLOT 0 ADDRESS */
#define PORT92DOS 0x0092
0xF092
C-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
C.2 EXAMPLE CODE DEFINES
/*********************** Global typedef **********************/
typedef unsigned char
typedef unsigned short WORD;
typedef unsigned long
BYTE;
/* 8-bit value */
/* 16-bit value */
DWORD; /* 32-bit value */
/******************** Global Used defines ********************/
/* Error Flags */
#define E_OK
0
1
2
3
#define E_INVALID_DEVICE
#define E_INVALID_VECTOR
#define E_BADVECTOR
#define INTERRUPT_ISR
#define TRAP_ISR
1
2
#define IDT_ALIAS
#define TRAP_TYPE
#define INTR_TYPE
2
/* Only valid for protected mode */
/* Only valid for protected mode */
/* Only valid for protected mode */
0x8f00
0x8e00
#define LOBYTE(w)
#define HIBYTE(w)
((BYTE)(w))
((BYTE)(((WORD)(w) >> 8) & 0xFF))
#define LOWORD(l)
#define HIWORD(l)
((WORD)(DWORD)(l))
((WORD)((((DWORD)(l)) >> 16) & 0xFFFF))
/*** Bit Masks ***/
#define BIT0MSK
#define BIT1MSK
#define BIT2MSK
#define BIT3MSK
#define BIT4MSK
#define BIT5MSK
#define BIT6MSK
#define BIT7MSK
0x1
0x2
0x4
0x8
0x10
0x20
0x40
0x80
/*** Global Function ***/
extern void _EnableExtIOMem(void);
/******* Interrupt Control Unit configuration defines ********/
/* ICU Modes */
#define ICU_SFNM
0x10
#define ICU_AUTOEOI
0x2
0x8
0x0
#define ICU_TRIGGER_LEVEL
#define ICU_TRIGGER_EDGE
/* ICU Master Pins */
C-6
EXAMPLE CODE HEADER FILES
#define MPIN_INT0
#define MPIN_INT1
#define MPIN_INT2
#define MPIN_INT3
0x4
0x8
0x10
0x20
/* ICU Master External Cascade IRs */
#define MCAS_IR1
#define MCAS_IR2
#define MCAS_IR5
#define MCAS_IR6
#define MCAS_IR7
/* ICU Slave Pins */
#define SPIN_INT4
#define SPIN_INT5
#define SPIN_INT6
#define SPIN_INT7
/* ICU IRQ Mask Values*/
#define IR0
0x2
0x4
0x20
0x40
0x80
0x1
0x2
0x4
0x8
0x1
#define IR1
0x2
#define IR2
0x4
#define IR3
0x8
#define IR4
#define IR5
#define IR6
#define IR7
0x10
0x20
0x40
0x80
/* ICU EOI Types */
#define NONSPECIFIC_EOI
#define SPECIFIC_EOI
#define NonSpecificEOI()
0x20
0x60
_SetEXRegByte(OCW2S,NONSPECIFIC_EOI);
_SetEXRegByte(OCW2M,NONSPECIFIC_EOI)
#define MstrSpecificEOI(irq)
)
_SetEXRegByte(OCW2M, 0x60 | ((BYTE)((irq) & 0x7))
#define SlaveSpecificEOI(irq) _SetEXRegByte(OCW2S, 0x60 | ((BYTE)((irq) & 0x7))
)
#define Master
#define Slave
1
0
/* ICU Function Definitions */
extern int InitICU (BYTE MstrMode, BYTE MstrBase, BYTE MstrCascade,
BYTE SlaveMode, BYTE SlaveBase,BYTE MstrPins,
BYTE SlavePins);
extern int InitICUSlave(BYTE SlaveMode, BYTE SlaveBase, BYTE SlavePins);
extern void SetInterruptVector(void (far interrupt *IntrProc)(void),
int Vector, int IntrType);
extern int SetIRQVector(void (far interrupt *IntrProc)(void), int IRQ,
int IntrType);
extern void Enable8259Interrupt(BYTE MstrMask, BYTE SlaveMask);
extern void Disable8259Interrupt(BYTE MstrMask, BYTE SlaveMask);
extern int Poll_Command(int Master_or_Slave);
C-7
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
/************ Asynchronous Serial I/O Port defines ***********/
#define SIO_0
#define SIO_1
0
1
#define SIO0_IRQ
#define SIO1_IRQ
4
3
/* IRQ # Master IRQ4 */
/* IRQ # Master IRQ3 */
#define SIO_5DATA
#define SIO_6DATA
#define SIO_7DATA
#define SIO_8DATA
0x0
0x1
0x2
0x3
#define SIO_1STOPBIT
#define SIO_2STOPBIT
0x0
0x4
#define SIO_NOPARITY
#define SIO_ODDPARITY
#define SIO_EVNPARITY
#define SIO_FRC0PARITY
#define SIO_FRC1PARITY
0x0
0x8
0x18
0x28
0x38
#define SIO_SETBREAK
0x40
#define SIO_INTERNAL_SRC
#define SIO_EXTERNAL_SRC
#define SIO_CLKSRC_CLK2
0x1
0x0
0x1
#define SIO_CLKSRC_COMCLK 0x0
#define SIO_INTR_NONE
#define SIO_INTR_RBF
#define SIO_INTR_TBE
#define SIO_INTR_RLS
#define SIO_INTR_MS
0
0x1
0x2
0x4
0x8
#define SIO_MCR_LOOP_BACK 0x10
#define SIO_MCR_OUT2
#define SIO_MCR_OUT1
#define SIO_MCR_RTS
0x8
0x4
0x2
0x1
#define SIO_MCR_DTR
#define SIO_8N1
#define SIO_7N1
(SIO_8DATA | SIO_1STOPBIT | SIO_NOPARITY)
(SIO_7DATA | SIO_1STOPBIT | SIO_NOPARITY)
/* Status Bits */
#define SIO_ERROR_BITS
#define SIO_RX_BUF_FULL
#define SIO_OVERRUN
0x1e
0x1
0x2
#define SIO_PARITY_ERR
#define SIO_FRAMING_ERR
#define SIO_BREAK_INTR
#define SIO_TX_BUF_EMPTY
0x4
0x8
0x10
0x20
C-8
EXAMPLE CODE HEADER FILES
#define SIO_TX_EMPTY
0x40
/* Offsets from beginning of SIO port addresses */
#define RBR
0
#define TBR
#define DLL
#define IER
#define DLH
#define IIR
#define LCR
#define MCR
#define LSR
#define MSR
#define SCR
#define SIO0_BASE
#define SIO1_BASE
0
0
1
1
2
3
4
5
6
7
0xF4F8
0xF8F8
/* Define Function Macros */
#define GetSIO0Status()
#define GetSIO1Status()
#define GetSIO0InterruptID()
#define GetSIO1InterruptID()
#define GetSIO0ModemStatus()
#define GetSIO1ModemStatus()
#define GetSIO0Char()
_GetEXRegByte(LSR0)
_GetEXRegByte(LSR1)
_GetEXRegByte(IIR0)
_GetEXRegByte(IIR1)
_GetEXRegByte(MSR0)
_GetEXRegByte(MSR1)
_GetEXRegByte(RBR0)
_GetEXRegByte(RBR1)
_SetEXRegByte(IER0,src)
_SetEXRegByte(IER1,src)
_SetEXRegByte(LCR0,Mode)
#define GetSIO1Char()
#define ChangeSIO0IntrSrc(src)
#define ChangeSIO1IntrSrc(src)
#define ChangeSIO0Mode(Mode)
#define ChangeSIO1Mode(Mode)
_SetEXRegByte(LCR1,Mode)
#define DisableSIO0Interrupt(src) _SetEXRegByte(IER0,_GetEXRegByte(IER0) &
!(src))
#define DisableSIO1Interrupt(src) _SetEXRegByte(IER1,_GetEXRegByte(IER1) &
!(src))
/* SIO Function Definitions */
extern int InitSIO
(int Unit, BYTE Mode, BYTE ModemCntrl, DWORD BaudRate,
DWORD BaudClkIn);
extern BYTE SerialReadChar(int Unit);
extern int SerialReadStr(int Unit, char far *str, int count);
extern void SerialWriteChar(int Unit, BYTE ch);
extern void SerialWriteStr(int Unit, const char far *str);
extern void SerialWriteMem(int Unit, const char far *mem, int count);
void interrupt far Serial0_ISR(void);
extern void Service_RBF (void);
extern void SerialWriteStr_Int(int Unit, const char far *str);
extern void Service_TBE(void);
/***************** DMA configuration defines *****************/
C-9
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
typedef enum
{
DMA_Channel0 = 0,
DMA_Channel1 = 1
} DMAChannelEnum;
typedef enum
{
ERR_NONE = 0,
ERR_BADINPUT = -1
} ERREnum;
/* DMA Function Definitions */
int SetDMAReqIOAddr(int nChannel, WORD wIO);
int SetDMATargMemAddr(int nChannel, void *ptMemory);
int SetDMAXferCount(int nChannel, DWORD lCount);
int EnableDMAHWRequests(int nChannel);
int DisableDMAHWRequests(int nChannel);
void InitDMA(void);
void InitDMA1ForSerialXmitter(void);
/*************** Port I/O configuration defines **************/
/* Port 1 configuration defines */
#define DCD0
#define RTS0
#define DTR0
#define DSR0
#define RI0
#define LOCK
#define HOLD
#define HOLDACK
0x1
0x2
0x4
0x8
0X10
0x20
0X40
0X80
/* Port 2 configuration defines */
#define CS0
0x1
#define CS1
#define CS2
#define CS3
#define CS4
#define RXD0
#define TXD0
#define CTS0
0x2
0x4
0x8
0X10
0x20
0X40
0X80
/* Port 3 configuration defines */
#define TMROUT0
#define TMROUT1
#define INT0
0x1
0x2
0x4
#define INT1
0x8
#define INT2
#define INT3
#define PWRDWN
#define COMCLK
0x10
0x20
0x40
0x80
/* Port Direction defines */
#define P0_IN
0x1
C-10
EXAMPLE CODE HEADER FILES
#define P1_IN
#define P2_IN
#define P3_IN
#define P4_IN
#define P5_IN
#define P6_IN
#define P7_IN
#define Px_OUT
0x2
0x4
0x8
0x10
0x20
0x40
0x80
0
/* Pin configuration defines */
#define RTS1
0x1
0
0x2
0
0x4
0
0x8
0
#define SSIOTX
#define DTR1
#define SRXCLK
#define TXD1
#define DACK1
#define CTS1
#define EOP
#define CS5
0x10
0
0x20
0
0x40
0
#define DACK0
#define TIMER2
#define COPROC
#define REFRESH
#define CS6
/* Port I/O Function Definitions */
extern void Init_IOPorts
(BYTE Port1, BYTE Port2, BYTE Port3, BYTE PortDir1,
BYTE PortDir2, BYTE PortDir3, BYTE PortLtc1,
BYTE PortLtc2, BYTE PortLtc3);
/**************** Timer configuration defines ****************/
#define TMR_0
0
#define TMR_1
1
#define TMR_2
2
#define TMR0_IRQ
#define TMR1_IRQ
#define TMR2_IRQ
/* Timer Modes */
#define TMR_TERMCNT
#define TMR_1SHOT
#define TMR_RATEGEN
#define TMR_SQWAVE
#define TMR_SW_TRIGGER
#define TMR_HW_TRIGGER
/* Count Type */
#define TMR_CLK_BCD
#define TMR_CLK_BIN
0
10
11
/* IRQ # Master IRQ0 */
/* IRQ # Slave IRQ2 */
/* IRQ # Slave IRQ3 */
0
(1<<1)
(2<<1)
(3<<1)
(4<<1)
(5<<1)
1
0
/* Timer Pin Configuration */
#define TMR_CLK_INTRN
#define TMR_CLK_EXTRN
#define TMR_GATE_VCC
0
0x1
0
C-11
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
#define TMR_GATE_EXTRN
#define TMR_OUT_ENABLE
#define TMR_OUT_DISABLE
0x2
0x1
0
#define TMR_ENABLE
#define TMR_DISABLE
1
0
/* Timer Macros Definitions */
#define DisableTimer() \
_SetEXRegByte( TMRCFG, (_GetEXRegByte(TMRCFG) | 0x80))
#define EnableTimer() \
_SetEXRegByte( TMRCFG, (_GetEXRegByte(TMRCFG) & 0x7f))
/* Timer Function Definitions */
extern int InitTimer
(int Unit, WORD Mode, BYTE Inputs, BYTE Output,
WORD InitCount, int Enable);
extern void SetUp_ReadBack (BYTE Timer0, BYTE Timer1, BYTE Timer2,
BYTE GetStatus, BYTE GetCount);
extern WORD CounterLatch(BYTE Timer);
extern WORD ReadCounter(BYTE Timer);
void interrupt far TimerISR(void);
/**************** SSIO configuration defines *****************/
#define SSIO_TUE
0x80
0x40
0x20
0x10
0x08
0x04
0x02
0x01
/* Transmit Underflow Error */
/* Transmit Holding Buffer Empty */
/* Transmit Interrupt Enable */
/* Transmitter Enable */
/* Receive Overflow Error */
/* Receive Holding Buffer Full */
/* Receive Interrupt Enable */
/* Receiver Enable */
#define SSIO_THBE
#define SSIO_TX_IE
#define SSIO_TX_ENAB
#define SSIO_ROE
#define SSIO_RHBF
#define SSIO_RX_IE
#define SSIO_RX_ENAB
#define SSIO_TX_MASTR
#define SSIO_RX_MASTR
#define SSIO_TX_SLAVE
#define SSIO_RX_SLAVE
0x02
0x01
0
/* Transmit Master Mode */
/* Receive Master Mode */
0
#define SSIO_CLK_SERCLK
#define SSIO_CLK_PSCLK
#define SSIO_BAUD_ENAB
0x01
0x00
0x80
/* Baud Rate Clocking Source:
SERCLK = CLK2/4 */
/* Baud Rate Clocking Source:
PSCLK = (CLK2/2) / (CLKPRS+2) */
/* Enable Baud Rate Generator */
/* SSIO Function Definitions */
extern void InitSSIO
(BYTE Mode, BYTE MasterTxRx, BYTE BaudValue,
C-12
EXAMPLE CODE HEADER FILES
BYTE PreScale);
extern WORD SSerialReadWord(BYTE MasterSlave);
extern void SSerialWriteWord(WORD Ch,BYTE MasterSlave);
void interrupt far SSIO_ISR(void);
extern void Service_RHBF(void);
extern void Service_THBE(void);
/********************* Watch Dog Timer ***********************/
#define SetWatchDogReload(ReloadHi,ReloadLow) \
_SetEXRegWord(WDTRLDL,ReloadLow);_SetEXRegWord(WDTRLDH,ReloadHi);
#define WatchDogClockDisable()\
_SetEXRegByte(WDTSTATUS, _GetEXRegByte(WDTSTATUS) | BIT0MSK)
#define WatchDogClockEnable()\
_SetEXRegByte(WDTSTATUS, _GetEXRegByte(WDTSTATUS) & ~BIT0MSK)
/* Watch Dog Timer Function Definitions */
extern void ReLoadDownCounter(void);
extern DWORD GetWDT_Count(void);
extern void WDT_BusMonitor(BYTE EnableDisable);
extern void EnableWDTInterrupt(void);
void interrupt far wdtISR(void);
/********************Refresh Control Unit*********************/
#define EnableRCU() \
_SetEXRegWord(RDFSCON, _GetEXRegWord(RDFSCON) | 0x8000)
#define DisableRCU() \
_SetEXRegWord(RDFSCON, _GetEXRegWord(RDFSCON) & 0x7fff)
/* Refresh Control Unit Function Definitions */
extern int InitRCU(WORD counter_value);
extern WORD Get_RCUCounterValue(void);
/****************Clock and Power Management Unit**************/
#define IDLE
#define PWDWN
#define ACTIVE
0x02
0x01
0x00
/* Clock and Power Management Function Definitions */
extern int Set_Prescale_Value(WORD prescale);
extern void Enter_Idle_Mode(void);
extern void Enter_Powerdown_Mode(void);
extern void Mode_Setting_To_Active(void);
C-13
D
SYSTEM
REGISTER QUICK
REFERENCE
APPENDIX D
SYSTEM REGISTER QUICK REFERENCE
D.1 PERIPHERAL REGISTER ADDRESSES
Table D-1. Peripheral Register Addresses (Sheet 1 of 6)
Expanded
Address
PC/AT
Address
Access Type
(Byte/Word)
Register Name
Reset Value
DMA Controller and Bus Arbiter
F000H
F001H
F002H
F003H
F004H
F005H
F006H
F007H
F008H
F009H
F00AH
F00BH
F00CH
F00DH
F00EH
F00FH
F010H
F011H
F012H
F013H
F014H
F015H
F016H
F017H
F018H
NOTES:
0000H
Byte
Byte
Byte
Byte
DMA0TAR0/1 (Note 1)
XX
0001H
0002H
0003H
0004H
0005H
0006H
0007H
0008H
0009H
000AH
000BH
000CH
000DH
000EH
000FH
DMA0BYC0/1 (Note 1)
DMA1TAR0/1 (Note 1)
DMA1BYC0/1 (Note 1)
Reserved
XX
XX
XX
Reserved
Reserved
Reserved
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
DMACMD1/DMASTS
DMASRR
00H
00H
DMAMSK
04H
DMAMOD1
00H
DMACLRBP
DMACLR
Not a register
Not a register
DMACLRMSK
DMAGRPMSK
DMA0REQ0/1
DMA0REQ2/3
DMA1REQ0/1
DMA1REQ2/3
Reserved
Not a register
03H
XX
XX
XX
XX
Reserved
Reserved
Reserved
Byte
DMABSR
X1X10000B
1. Byte pointer in flip-flop in DMA determines which register is accessed.
2. Shaded rows indicate reserved areas.
D-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table D-1. Peripheral Register Addresses (Sheet 2 of 6)
Expanded
Address
PC/AT
Address
Access Type
(Byte/Word)
Register Name
Reset Value
F019H
F01AH
F01BH
F01CH
F01DH
F01EH
Byte
Byte
Byte
Byte
Byte
Byte
DMACHR/DMAIS
00H
DMACMD2
DMAMOD2
DMAIEN
08H
00H
00H
DMAOVFE
DMACLRTC
0AH
Not a register
Master Interrupt Controller
F020H
F021H
0020H
Byte
ICW1m/IRRm/ISRm/
OCW2m/OCW3m
XX
XX
0021H
0022H
Byte
ICW2m/ICW3m/ICW4m/
OCW1m/POLLm
Address Configuration Register
Word REMAPCFG
Timer/counter Unit
0022H
0000H
F040H
F041H
F042H
F043H
0040H
0041H
0042H
0043H
Byte
Byte
Byte
Byte
TMR0
XX
XX
XX
XX
TMR1
TMR2
TMRCON
DMA Page Registers
Reserved
F080H
F081H
F082H
F083H
F084H
F085H
F086H
F087H
F088H
F089H
F08AH
F08BH
F08CH
F08DH
NOTES:
0081H
0082H
0083H
Reserved
Reserved
DMA1TAR2
Reserved
DMA1TAR3
DMA0TAR3
DMA0TAR2
Reserved
Reserved
Reserved
Reserved
Reserved
Reserved
Byte
XX
Byte
Byte
Byte
XX
XX
XX
0087H
0089H
008AH
008BH
1. Byte pointer in flip-flop in DMA determines which register is accessed.
2. Shaded rows indicate reserved areas.
D-2
SYSTEM REGISTER QUICK REFERENCE
Table D-1. Peripheral Register Addresses (Sheet 3 of 6)
Expanded
Address
PC/AT
Address
Access Type
(Byte/Word)
Register Name
Reset Value
F08EH
F08FH
F098H
F099H
F09AH
F09BH
Reserved
Reserved
Byte
Byte
DMA0BYC2
DMA1BYC2
Reserved
XX
XX
Reserved
A20GATE and Fast CPU Reset
Byte PORT92
Slave Interrupt Controller
F092H
0092H
XXXXXX10B
F0A0H
F0A1H
00A0H
00A1H
Byte
ICW1s/IRRs/ISRs/
OCW2s/OCW3s
XX
XX
Byte
ICW2s/ICW3s/ICW4s/
OCW1s/POLLs
Chip-select Unit
CS0ADL
F400H
F402H
F404H
F406H
F408H
F40AH
F40CH
F40EH
F410H
F412H
F414H
F416H
F418H
F41AH
F41CH
F41EH
F420H
F422H
F424H
NOTES:
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
CS0ADH
CS0MSKL
CS0MSKH
CS1ADL
CS1ADH
CS1MSKL
CS1MSKH
CS2ADL
CS2ADH
CS2MSKL
CS2MSKH
CS3ADL
CS3ADH
CS3MSKL
CS3MSKH
CS4ADL
CS4ADH
CS4MSKL
1. Byte pointer in flip-flop in DMA determines which register is accessed.
2. Shaded rows indicate reserved areas.
D-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table D-1. Peripheral Register Addresses (Sheet 4 of 6)
Expanded
Address
PC/AT
Address
Access Type
(Byte/Word)
Register Name
Reset Value
F426H
F428H
F42AH
F42CH
F42EH
F430H
F432H
F434H
F436H
F438H
F43AH
F43CH
F43EH
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
Word
CS4MSKH
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
0000H
FF6FH
FFFFH
FFFFH
FFFFH
CS5ADL
CS5ADH
CS5MSKL
CS5MSKH
CS6ADL
CS6ADH
CS6MSKL
CS6MSKH
UCSADL
UCSADH
UCSMSKL
UCSMSKH
Synchronous Serial I/O Unit
F480H
F482H
F484H
F486H
F488H
F48AH
Word
Word
Byte
Byte
Byte
Byte
SSIOTBUF
0000H
0000H
00H
SSIORBUF
SSIOBAUD
SSIOCON1
SSIOCON2
SSIOCTR
C0H
00H
00H
Refresh Control Unit
F4A0H
F4A2H
F4A4H
F4A6H
Word
Word
Word
Word
RFSBAD
0000H
0000H
0000H
00FFH
RFSCIR
RFSCON
RFSADD
Watchdog Timer Unit
F4C0H
F4C2H
F4C4H
F4C6H
F4C8H
F4CAH
NOTES:
Word
Word
Word
Word
Word
Byte
WDTRLDH
003FH
WDTRLDL
WDTCNTH
WDTCNTL
WDTCLR
FFFFH
003FH
FFFFH
Not a register
00H
WDTSTATUS
1. Byte pointer in flip-flop in DMA determines which register is accessed.
2. Shaded rows indicate reserved areas.
D-4
SYSTEM REGISTER QUICK REFERENCE
Table D-1. Peripheral Register Addresses (Sheet 5 of 6)
Expanded
Address
PC/AT
Address
Access Type
(Byte/Word)
Register Name
Reset Value
Asynchronous Serial I/O Channel 0 (COM1)
F4F8H
F4F9H
F4FAH
F4FBH
F4FCH
F4FDH
F4FEH
F4FFH
03F8H
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
RBR0/TBR0/DLL0
IER0/DLH0
IIR0
XX/XX/02H
00H/00H
01H
03F9H
03FAH
03FBH
03FCH
03FDH
03FEH
03FFH
LCR0
00H
MCR0
00H
LSR0
60H
MSR0
X0H
SCR0
XX
Clock Generation and Power Management
F800H
F804H
Byte
PWRCON
CLKPRS
00H
Word
0000H
Device Configuration Registers
F820H
F822H
F824H
F826H
F830H
F832H
F834H
F836H
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
P1CFG
00H
00H
00H
00H
00H
00H
00H
00H
P2CFG
P3CFG
PINCFG
DMACFG
INTCFG
TMRCFG
SIOCFG
Parallel I/O Ports
F860H
F862H
F864H
F868H
F86AH
F86CH
F870H
F872H
F874H
NOTES:
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
P1PIN
XX
P1LTC
P1DIR
P2PIN
P2LTC
P2DIR
P3PIN
P3LTC
P3DIR
FFH
FFH
XX
FFH
FFH
XX
FFH
FFH
1. Byte pointer in flip-flop in DMA determines which register is accessed.
2. Shaded rows indicate reserved areas.
D-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table D-1. Peripheral Register Addresses (Sheet 6 of 6)
Expanded
Address
PC/AT
Address
Access Type
(Byte/Word)
Register Name
Reset Value
Asynchronous Serial I/O Channel 1 (COM2)
F8F8H
F8F9H
F8FAH
F8FBH
F8FCH
F8FDH
F8FEH
F8FFH
NOTES:
02F8H
Byte
Byte
Byte
Byte
Byte
Byte
Byte
Byte
RBR1/TBR1/DLL1
IER1/DLH1
IIR1
XX/XX/02H
00H/00H
01H
02F9H
02FAH
02FBH
02FCH
02FDH
02FEH
02FFH
LCR1
00H
MCR1
00H
LSR1
60H
MSR1
X0H
SCR1
XX
1. Byte pointer in flip-flop in DMA determines which register is accessed.
2. Shaded rows indicate reserved areas.
D-6
SYSTEM REGISTER QUICK REFERENCE
D.2 CLKPRS
Clock Prescale Register
CLKPRS
Expanded Addr: F804H
ISA Addr:
—
(read/write)
Reset State:
0000H
15
8
0
—
—
—
—
—
—
—
PS8
PS0
7
PS7
PS6
PS5
PS4
PS3
PS2
PS1
Bit
Number
Bit
Mnemonic
Function
15–9
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
8–0
PS8:0
Prescale Value:
These bits determine the divisor that is used to generate PSCLK. Legal
values are from 0000H (divide by 2) to 01FFH (divide by 513).
divisor = PS8:0 + 2
D-7
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.3 CSnADH (UCSADH)
Chip-select High Address
CSnADH (n = 0–6), UCSADH
(read/write)
Expanded Addr: F402H, F40AH
F412H, F41AH
F422H, F42AH
F432H, F43AH
—
ISA Addr:
Reset State:
0000H (CSnADH)
FFFFH (UCSADH)
15
8
0
—
—
—
—
—
—
CA15
CA7
CA14
7
CA13
CA12
CA11
CA10
CA9
CA8
CA6
Bit
Number
Bit
Mnemonic
Function
15–10
9–0
—
Reserved; for compatibility with future devices, write zeros to these bits.
Chip-select Channel Address Upper Bits:
CA15:6
Defines the upper 10 bits of the channel’s 15-bit address. The address
bits CA15:6 and the mask bits CM15:6 form a masked address that is
compared to memory address bits A25:16 or I/O address bits A15:6.
D-8
SYSTEM REGISTER QUICK REFERENCE
D.4 CSnADL (UCSADL)
Chip-select Low Address
CSnADL (n = 0–6), UCSADL
(read/write)
Expanded Addr: F400H, F408H
F410H, F418H
F420H, F428H
F430H, F438H
ISA Addr:
—
Reset State:
0000H (CSnADL)
FF6FH (UCSADL)
15
8
0
CA5
CA4
—
CA3
—
CA2
CA1
CASMM
WS2
BS16
WS1
MEM
7
RDY
WS4
WS3
WS0
Bit
Number
Bit
Mnemonic
Function
15–11
CA5:1
Chip-select Address Value Lower Bits:
Defines the lower 5 bits of the channel’s 15-bit address. The address bits
CA5:1 and the mask bits CM5:1 form a masked address that is compared to
memory address bits A15:11 or I/O address bits A5:1.
10
CASMM
SMM Address Bit:
If this bit is set (and unmasked), the CSU activates the chip-select channel
only while the processor is in SMM (and not in a hold state). Otherwise, the
CSU activates the channel only when processor is operating in a mode
other than SMM.
Setting the SMM mask bit in the channel’s mask low register masks this bit.
When this bit is masked, an address match activates the chip-select,
regardless of whether the processor is in SMM or not.
9
BS16
Bus Size 16-bit:
0 = All bus cycles to addresses in the channel’s address block are byte-
wide.
1 = Bus cycles are 16 bits unless the bus size control pin (BS8#) is
asserted.
8
7
MEM
RDY
Bus Cycle Type:
0 = Configures the channel for an I/O addresses
1 = Configures the channel for memory addresses
Bus Ready Enable:
0 = External READY# is ignored. READY# generated by CSU to terminate
the bus cycle.
1 = Requires that external READY# be active to complete a bus cycle. This
bit must be set to extend wait states beyond the number determined by
WS4:0 (see “Bus Cycle Length Control” on page 14-11).
6–5
4–0
—
Reserved; for compatibility with future devices, write zeros to these bits.
Wait State Value:
WS4:0
WS4:0 defines the minimum number of wait states inserted into the bus
cycle. A zero value means no wait states.
D-9
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.5 CSnMSKH (UCSMSKH)
Chip-select High Mask
CSnMSKH (n = 0–6), UCSMSKH
(read/write)
Expanded Addr: F406H, F40EH
F416H, F41EH
F426H, F42EH
F436H, F43EH
—
ISA Addr:
Reset State:
0000H (CSnMSKH)
FFFFH (UCSMSKH)
15
8
0
—
—
—
—
—
—
CM15
CM7
CM14
CM6
7
CM13
CM12
CM11
CM10
CM9
CM8
Bit
Number
Bit
Mnemonic
Function
15–10
9–0
—
Reserved; for compatibility with future devices, write zeros to these bits.
Mask Value Upper Bits:
CM15:6
Defines the upper 10 bits of the channel’s 15-bit mask. The mask bits
CM15:6 and the address bits CA15:6 form a masked address that is
compared to memory address bits A25:16 or I/O address bits A15:6.
D-10
SYSTEM REGISTER QUICK REFERENCE
D.6 CSnMSKL (UCSMSKL)
Chip-select Low Mask
CSnMSKL (n = 0–6), UCSMSKL
(read/write)
Expanded Addr: F404H, F40CH
F414H, F41CH
F424H, F42CH
F434H, F43CH
ISA Addr:
—
Reset State:
0000H (CSnMSKL)
FFFFH (UCSMSKL)
15
8
0
CM5
CM4
—
CM3
—
CM2
—
CM1
CMSMM
—
—
—
7
—
—
—
CSEN
Bit
Number
Bit
Mnemonic
Function
Chip-select Mask Value Lower Bits:
15–11
CM5:1
Defines the lower 5 bits of the channel’s 15-bit mask. The mask bits
CM5:1 and the address bits CA5:1 form a masked address that is
compared to memory address bits A15:11 or I/O address bits A5:1.
10
CMSMM
SMM Mask Bit:
0 = The SMM address bit is not masked.
1 = Masks the SMM address bit in the channel’s Chip-Select Low
Address register. When the SMM address bit is masked, an address
match activates the chip-select, regardless of whether the processor
is in SMM.
9–1
0
—
Reserved; for compatibility with future devices, write zeros to these bits.
Chip-select Enable:
CSEN
0 = Disables the chip-select channel.
1 = Enables the chip-select channel.
D-11
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.7 DLLn AND DLHn
Divisor Latch Low
DLL0, DLL1
(read/write)
DLL0
F4F8H
03F8H
02H
DLL1
F8F8H
02F8H
02H
Expanded Addr:
ISA Addr:
Reset State:
7
0
0
LD7
LD6
LD5
LD4
LD3
LD2
LD1
LD0
Divisor Latch High
DLH0, DLH1
(read/write)
DLH0
F4F9H
03F9H
00H
DLH1
F8F9H
02F9H
00H
Expanded Addr:
ISA Addr:
Reset State:
7
UD15
UD14
UD13
UD12
UD11
UD10
UD9
UD8
Bit
Number
Bit
Mnemonic
Function
Lower 8 Divisor and Upper 8 Divisor Bits:
DLLn
(7–0)
LD7:0
Write the lower 8 divisor bits to DLLn and the upper 8 divisor bits to
DLHn. The baud-rate generator output is a function of the baud-rate
generator input (BCLKIN) and the 16-bit divisor.
DLHn
(7–0)
UD15:8
BCLKIN frequency
baud-rate generator output frequency = ----------------------------------------------------
divisor
bit rate (shifting rate) = baud-rate generator output frequency/16
NOTE: The divisor latch registers share address ports with other SIO registers. Bit 7 (DLAB) of
LCRn must be set in order to access the divisor latch registers.
If DLL = DLH = 00H, baud-rate generator ouput frequency = 0 (stops clock).
D-12
SYSTEM REGISTER QUICK REFERENCE
Expanded Addr: F018H
D.8 DMABSR
DMA Bus Size
DMABSR
ISA Addr:
—
(write only)
Reset State:
X1X10000B
7
0
—
RBS
—
TBS
CS
—
—
0
Bit
Number
Bit
Mnemonic
Function
7
—
Reserved; for compatibility with future devices, write zero to this bit.
Requester Bus Size:
6
RBS
Specifies the requester’s data bus width for the channel specified by bit
0.
0 = 16-bit bus
1 = 8-bit bus
5
4
—
Reserved; for compatibility with future devices, write zero to this bit.
TBS
Target Bus Size:
Specifies the target’s data bus width for the channel specified by bit 0.
0 = 16-bit bus
1 = 8-bit bus
3–1
0
Must be 0 for correct operation.
0
CS
Channel Select:
0 = The selections for bits 7–4 affect channel 0.
1 = The selections for bits 7–4 affect channel 1.
D-13
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.9 DMACFG
DMA Configuration
DMACFG
Expanded Addr: F830H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
D1MSK
D1REQ2
D1REQ1
D1REQ0
D0MSK
D0REQ2
D0REQ1
D0REQ0
Bit
Number
Bit
Mnemonic
Function
7
D1MSK
DMA Acknowledge 1 Mask:
0 = DMA channel 1’s acknowledge (DMAACK1#) signal is not masked.
1 = Masks DMA channel 1’s acknowledge (DMAACK1#) signal. Useful
when channel 1’s request (DREQ1) input is connected to an internal
peripheral.
6–4
D1REQ2:0
DMA Channel 1 Request Connection:
Connects one of the eight possible hardware sources to channel 1’s
request input (DREQ1).
000 = DRQ1 pin (external peripheral)
001 = SIO channel 1’s receive buffer full signal (RBFDMA1)
010 = SIO channel 0’s transmit buffer empty signal (TXEDMA0)
011 = SSIO receive holding buffer full signal (SSRBF)
100 = TCU counter 2’s output signal (OUT2)
101 = SIO channel 0’s receive buffer full signal (RBFDMA0)
110 = SIO channel 1’s transmit buffer empty signal (TXEDMA1)
111 = SSIO transmit holding buffer empty signal (SSTBE)
3
D0MSK
DMA Acknowledge 0 Mask:
0 = DMA channel 0’s acknowledge (DMAACK0#) signal is not masked.
1 = Masks DMA channel 0’s acknowledge (DMAACK0#) signal. Useful
when channel 0’s request (DREQ0) input is connected to an internal
peripheral.
2–0
D0REQ2:0
DMA Channel 0 Request Connection:
Connects one of the eight possible hardware sources to channel 0’s
request input (DREQ0).
000 = DRQ0 pin (external peripheral)
001 = SIO channel 0’s receive buffer full signal (RBFDMA0)
010 = SIO channel 1’s transmit buffer empty signal (TXEDMA1)
011 = SSIO transmit holding buffer empty signal (SSTBE)
100 = TCU counter 1’s output signal (OUT1)
101 = SIO channel 1’s receive buffer full signal (RBFDMA1)
110 = SIO channel 0’s transmit buffer empty signal (TXEDMA0)
111 = SSIO receive holding buffer full signal (SSRBF)
D-14
SYSTEM REGISTER QUICK REFERENCE
Expanded Addr: F019H
D.10 DMACHR
DMA Chaining
DMACHR
ISA Addr:
—
(write only)
Reset State:
00H
7
0
—
—
—
—
—
CE
0
CS
Bit
Number
Bit
Mnemonic
Function
7–3
—
Reserved; for compatibility with future devices, write zeros to these bits.
Chaining Enable:
2
CE
0 = Disables the chaining buffer-transfer mode for the channel specified
by bit 0.
1 = Enables the chaining buffer-transfer mode for the channel specified
by bit 0.
1
0
0
Must be 0 for correct operation.
Channel Select:
CS
0 = The selection for bit 2 affects channel 0.
1 = The selection for bit 2 affects channel 1.
D-15
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.11 DMACMD1
DMA Command 1
DMACMD1
Expanded Addr: F008H
ISA Addr:
0008H
(write only)
Reset State:
00H
7
0
—
—
—
PRE
—
CE
—
—
Bit
Number
Bit
Mnemonic
Function
7–5
—
Reserved; for compatibility with future devices, write zeros to these bits.
Priority Rotation Enable:
4
PRE
0 = Priority is fixed based on value in DMACMD2.
1 = Enables the rotation method for changing the bus control priority
structure. That is, after the external bus master or one of the DMA
channels is given bus control, it is assigned to the lowest priority
level.
3
2
—
Reserved; for compatibility with future devices, write zero to this bit.
Channel Enable:
CE
0 = Enables channel 0 and 1.
1 = Disables the channels.
1–0
—
Reserved; for compatibility with future devices, write zeros to these bits.
D-16
SYSTEM REGISTER QUICK REFERENCE
Expanded Addr: F01AH
D.12 DMACMD2
DMA Command 2
DMACMD2
ISA Addr:
—
(write only)
Reset State:
08H
7
0
—
—
—
—
PL1
PL0
ES
DS
Bit
Number
Bit
Mnemonic
Function
7–4
—
Reserved; for compatibility with future devices, write zeros to these bits.
Low Priority Level Set:
3–2
PL1:0
Use these bits to assign a particular bus request to the lowest priority
level in fixed priority mode.
00 = Assigns channel 0’s request (DREQ0) to the lowest priority level
01 = Assigns channel 1’s request (DREQ1) to the lowest priority level
10 = Assigns the external bus master request (HOLD) to the lowest
priority level
11 = Reserved
1
0
ES
DS
EOP# Sampling:
0 = Causes the DMA to sample the EOP# input asynchronously.
1 = Causes the DMA to sample the end-of-process (EOP#) input
synchronously.
DREQn Sampling:
0 = Causes the DMA to sample the DREQn inputs asynchronously.
1 = Causes the DMA to sample the channel request (DREQn) inputs
synchronously.
D-17
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.13 DMAGRPMSK
DMA Group Channel Mask
DMAGRPMSK
Expanded Addr: F00FH
ISA Addr:
000FH
(read/write)
Reset State:
03H
7
0
—
—
—
—
—
—
HRM1
HRM0
Bit
Number
Bit
Mnemonic
Function
7–2
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
1
0
HRM1
HRM0
Hardware Request Mask 1:
0 = Channel 1’s hardware requests are not masked.
1 = Masks (disables) channel 1’s hardware requests. When this bit is
set, channel 1 can still receive software requests.
Hardware Request Mask 0:
0 = Channel 0’s hardware requests are not masked.
1 = Masks (disables) channel 0’s hardware requests. When this bit is
set, channel 0 can still receive software requests.
D-18
SYSTEM REGISTER QUICK REFERENCE
Expanded Addr: F01CH
D.14 DMAIEN
DMA Interrupt Enable
DMAIEN
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
—
—
—
—
—
—
TC1
TC0
Bit
Number
Bit
Mnemonic
Function
7–2
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
1
TC1
Transfer Complete 1:
0 = Disables Transfer Complete interrupts.
1 = Connects channel 1’s transfer complete signal to the interrupt
control unit’s DMAINT input.
Note: When channel 1 is in chaining mode (DMACHR.2=1 and
DMACHR.0=1), this bit is a don’t care.
0
TC0
Transfer Complete 0:
0 = Disables Transfer Complete interrupts.
1 = Connects channel 0’s transfer complete signal to the interrupt
control unit’s DMAINT input.
Note: When channel 0 is in chaining mode (DMACHR.2=1 and
DMACHR.0=0), this bit is a don’t care.
D-19
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.15 DMAIS
DMA Interrupt Status
DMAIS
Expanded Addr: F019H
ISA Addr:
—
(read only)
Reset State:
00H
7
0
TC1
TC0
CI1
CI0
—
—
—
—
Bit
Number
Bit
Mnemonic
Function
7–6
Reserved. These bits are undefined.
Transfer Complete 1:
—
5
TC1
When set, this bit indicates that channel 1 has completed a buffer
transfer (either its byte count expired or it received an EOP# input). This
bit is set only if bit 1 of the interrupt enable register is set. Reading the
DMA status register (DMASTS) clears this bit.
Note: In chaining mode, this bit becomes a don’t care.
Transfer Complete 0:
4
TC0
When set, this bit indicates that channel 0 has completed a buffer
transfer (either its byte count expired or it received an EOP# input). This
bit is set only if bit 0 of the interrupt enable register is set. Reading the
DMA status register (DMASTS) clears this bit.
Note: In chaining mode, this bit becomes a don’t care.
Reserved. These bits are undefined.
3–2
1
—
CI1
Chaining Interrupt 1:
When set, this bit indicates that new requester and target addresses and
a new byte count should be written to channel 1. This bit is cleared when
new transfer information is written to the channel. (Writing to the most-
significant byte of the target address clears this bit.)
Note: Outside chaining mode, this bit becomes a don’t care.
Chaining Interrupt 0:
0
CI0
When set, this bit indicates that new requester and target addresses and
a new byte count should be written to channel 0. This bit is cleared when
new transfer information is written to the channel. (Writing to the most-
significant byte of the target address clears this bit.)
Note: Outside chaining mode, this bit becomes a don’t care.
D-20
SYSTEM REGISTER QUICK REFERENCE
Expanded Addr: F00BH
D.16 DMAMOD1
DMA Mode 1
DMAMOD1
(write only)
ISA Addr:
000BH
Reset State:
00H
7
0
DTM1
DTM0
TI
AI
TD1
TD0
0
CS
Bit
Number
Bit
Mnemonic
Function
7–6
DTM1:0
Data-transfer Mode:
00 = Demand
01 = Single
10 = Block
11 = Cascade
5
TI
Target Increment/Decrement:
0 = Causes the target address to be incremented after each data
transfer in a buffer transfer.
1 = Causes the target address for the channel specified by bit 0 to be
decremented after each data transfer in a buffer transfer. Note that it
does not decrement words. When decrementing it will do two byte
transfers for a word.
Note: When the target address is programmed to remain constant
(DMAMOD2.2 = 1), this bit is a don’t care.
4
AI
Autoinitialize:
0 = Disables the autoinitialize buffer-transfer mode for the channel
specified by bit 0.
1 = Enables the autoinitialize buffer-transfer mode for the channel
specified by bit 0.
3–2
TD1:0
Transfer Direction:
Determines the transfer direction for the channel specified by bit 0.
00 = Target is read; nothing is written (used for testing)
01 = Data is transferred from the requester to the target
10 = Data is transferred from the target to the requester
11 = Reserved
Note: In cascade mode, these bits become don’t cares.
Must be 0 for correct operation.
Channel Select:
1
0
0
CS
0 = The selections for bits 7–2 affect channel 0.
1 = The selections for bits 7–2 affect channel 1.
D-21
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.17 DMAMOD2
DMA Mode 2
DMAMOD2
(write only)
Expanded Addr: F01BH
ISA Addr:
—
Reset State:
00H
7
0
BCO
RD
TD
RH
RI
TH
0
CS
Bit
Number
Bit
Mnemonic
Function
7
BCO
Bus Cycle Option:
0 = Selects the fly-by data transfer bus cycle option for the channel specified
by bit 0.
1 = Selects the two-cycle data transfer bus cycle option for the channel
specified by bit 0.
6
RD
Requester Device Type:
0 = Clear this bit when the requester for the channel specified by bit 0 is in
memory space.
1 = Set this bit when the requester for the channel specified by bit 0 is in I/O
space.
This bit is ignored if BCO is cleared.
Target Device Type:
5
4
3
TD
RH
RI
0 = Clear this bit when the target for the channel specified by bit 0 is in
memory space.
1 = Set this bit when the target for the channel specified by bit 0 is in I/O
space.
Requester Address Hold:
0 = Causes the address to be modified (incremented or decremented,
depending on DMAMOD2.3).
1 = Causes the requester’s address for the channel specified by bit 0 to
remain constant during a buffer transfer.
Requester Address Increment/Decrement:
0 = Causes the requester address to be incremented after each data transfer
in a buffer transfer.
1 = Causes the requester address for the channel specified by bit 0 to be
decremented after each data transfer in a buffer transfer. Note that it
does not decrement words. When decrementing it will do two byte
transfers for a word.
Note: When the target address is programmed to remain constant
(DMAMOD2.4 = 1), this bit is a don’t care.
2
TH
Target Address Hold:
0 = Causes the address to be modified (incremented or decremented,
depending on DMAMOD1.5).
1 = Causes the target’s address for the channel specified by bit 0 to remain
constant during a buffer transfer.
1
0
0
Must be 0 for correct operation.
Channel Select:
CS
0 = The selections for bits 7–2 affect channel 0.
1 = The selections for bits 7–2 affect channel 1.
D-22
SYSTEM REGISTER QUICK REFERENCE
D.18 DMAMSK
DMA Individual Channel Mask
DMAMSK
Expanded Addr: F00AH
ISA Addr:
000AH
(write only)
Reset State:
04H
7
0
—
—
—
—
—
HRM
0
CS
Bit
Number
Bit
Mnemonic
Function
7–3
—
Reserved; for compatibility with future devices, write zeros to these bits.
Hardware Request Mask:
2
HRM
0 = Unmasks (enables) hardware requests for the channel specified by
bit 0.
1 = Masks (disables) hardware requests for the channel specified by
bit 0.
NOTE: When this bit is set, the channel can still receive software
requests.
1
0
0
Must be 0 for correct operation.
Channel Select:
CS
0 = The selection for bit 2 affects channel 0.
1 = The selection for bit 2 affects channel 1.
D-23
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.19 DMAnBYCn, DMAnREQn AND DMAnTARn
DMA Channel 0
24
DMA0REQ3
16
DMA0REQ2
8
0
Requester Address
Target Address
Byte Count
DMA0REQ1
DMA0REQ0
F011H (BP=1)
F011H (BP=0)
F010H (BP=1)
F010H (BP=0)
24
DMA0TAR3
F086H
16
DMA0TAR2
F087H
8
DMA0TAR1
F000H (BP=1)
0
DMA0TAR0
F000H (BP=0)
16
DMA0BYC2
F098H
8
DMA0BYC1
F001H (BP=1)
0
DMA0BYC0
F001H (BP=0)
DMA Channel 1
24
DMA1REQ3
16
DMA1REQ2
F013H (BP=0)
8
DMA1REQ1
F012H (BP=1)
0
DMA1REQ0
F012H (BP=0)
Requester Address
Target Address
Byte Count
F013H (BP=1)
24
DMA1TAR3
F085H
16
DMA1TAR2
F083H
8
DMA1TAR1
F002H (BP=1)
0
DMA1TAR0
F002H (BP=0)
16
DMA1BYC2
F099H
8
DMA1BYC1
F003H (BP=1)
0
DMA1BYC0
F003H (BP=0)
D-24
SYSTEM REGISTER QUICK REFERENCE
Expanded Addr: F01DH
D.20 DMAOVFE
DMA Overflow Enable
DMAOVFE
ISA Addr:
—
(read/write)
Reset State:
0AH
7
0
—
—
—
—
ROV1
TOV1
ROV0
TOV0
Bit
Number
Bit
Mnemonic
Function
7–4
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
3
2
ROV1
TOV1
Channel 1 Requester Overflow Enable:
0 = lowest 16 bits of requester address increment/decrement
1 = all bits of requester address increment/decrement
Channel 1 Target & Byte Counter Overflow Enable:
0 = lowest 16 bits of target address and byte count
increment/decrement
1 = all bits of target address and byte count increment/decrement
1
0
ROV0
TOV0
Channel 0 Requester Overflow Enable:
0 = lowest 16 bits of requester address increment/decrement
1 = all bits of requester address increment/decrement
Channel 0 Target & Byte Counter Overflow Enable:
0 = lowest 16 bits of target address and byte count
increment/decrement
1 = all bits of target address and byte count increment/decrement
D-25
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.21 DMASRR
DMA Software Request (read format)
DMASRR
Expanded Addr: F009H
ISA Addr:
0009H
Reset State:
00H
7
0
—
—
—
—
—
—
SR1
SR0
Bit
Number
Bit
Mnemonic
Function
7–2
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
1
0
SR1
SR0
Software Request 1:
When set, this bit indicates that channel 1 has a software request
pending.
Software Request 0:
When set, this bit indicates that channel 0 has a software request
pending.
DMA Software Request (write format)
DMASRR
Expanded Addr: F009H
ISA Addr:
0009H
Reset State:
00H
7
0
—
—
—
—
—
SR
0
CS
Bit
Number
Bit
Mnemonic
Function
7–3
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
2
SR
Software Request:
Setting this bit generates a software request for the channel specified by
bit 0. When the channel’s buffer transfer completes, this bit is cleared.
1
0
0
Must be 0 for correct operation.
Channel Select:
CS
0 = The selection for bit 2 affects channel 0.
1 = The selection for bit 2 affects channel 1.
D-26
SYSTEM REGISTER QUICK REFERENCE
Expanded Addr: F008H
D.22 DMASTS
DMA Status
DMASTS
ISA Addr:
0008H
(read only)
Reset State:
00H
7
0
—
—
R1
R0
—
—
TC1
TC0
Bit Number
Bit Mnemonic
Function
7–6
—
Reserved. These bits are undefined.
Request 1:
5
R1
When set, this bit indicates that channel 1 has a hardware request
pending. When the request is removed, this bit is cleared.
4
R0
Request 0:
When set, this bit indicates that channel 0 has a hardware request
pending. When the request is removed, this bit is cleared.
3–2
1
—
Reserved. These bits are undefined.
Transfer Complete 1:
TC1
When set, this bit indicates that channel 1 has completed a buffer
transfer (either its byte count expired or it received an EOP# input).
Reading this register clears this bit and clears TC1 in DMAIS.
0
TC0
Transfer Complete 0:
When set, this bit indicates that channel 0 has completed a buffer
transfer (either its byte count expired or it received an EOP# input).
Reading this register clears this bit and clears TC0 in DMAIS.
D-27
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.23 ICW1 (MASTER AND SLAVE)
Initialization Command Word 1
master slave
ICW1 (master and slave)
(write only)
Expanded Addr:
ISA Addr:
Reset State:
F020H
0020H
XXH
F0A0H
00A0H
XXH
7
0
0
0
0
RSEL1
LS
0
0
1
Bit
Number
Bit
Mnemonic
Function
7–5
—
Clear these bits to guarantee device operation.
Register Select 1 (Also see OCW2 and OCW3):
4
RSEL1
ICW1, OCW2, and OCW3 are accessed through the same addresses.
0 = OCW2 or OCW3 is accessed (Figure 9-13 and Figure 9-15).
1 = ICW1 register is accessed.
3
LS
Level/Edge Sensitive:
0 = Selects edge-triggered IR input signals.
1 = Selects level-sensitive IR input signals.
All internal peripherals interface with the 82C59As in edge-triggered
mode only. This is compatible with the PC/AT bus specification. Each
source signal initiates an interrupt request by making a low-to-high
transition. External peripherals interface with the 8259As in edge-
triggered or level-sensitive mode. The modes are selected for the
device, not for individual interrupts.
NOTE: If an internal peripheral interrupt is used, the 8259A that the
interrupt is connected to must be programmed for edge-triggered
interrupts.
2–1
0
—
—
Clear these bits to guarantee device operation.
Set this bit to guarantee device operation.
NOTE: The 82C59A must be initialized before it can be used. After reset, the 82C59A register states are
undefined. The 82C59A modules must be initialized before the IF flag in the core FLAG register is
set. All peripherals that use interrupts connected to the ICU must be initialized before initializing
the ICU.
D-28
SYSTEM REGISTER QUICK REFERENCE
D.24 ICW2 (MASTER AND SLAVE)
Initialization Command Word 2
ICW2 (master and slave)
(write only)
master slave
Expanded Addr:
ISA Addr:
Reset State:
F021H
0021H
XXH
F0A1H
00A1H
XXH
7
0
T7
T6
T5
T4
T3
0
0
0
Bit
Number
Bit
Mnemonic
Function
7–3
T7:3
Base Interrupt Type:
Write the base interrupt vector’s five most-significant bits to these bits.
Clear these bits to guarantee device operation.
2–0
T2:0
D.25 ICW3 (MASTER)
Initialization Command Word 3
ICW3 (master)
Expanded Addr: F021H
ISA Addr:
0021H
(write only)
Reset State:
XXH
7
0
S7
S6
S5
S4
S3
S2
S1
0
Bit
Number
Bit
Mnemonic
Function
7–3
S7:3
Slave IRs
0 = No slave 8259A is attached to the corresponding IR signal of the
master.
1 = A slave 82C59A is attached to the corresponding IR signal of the
master.
2
1
S2
S1
0 = Internal slave not used
1 = Internal slave is cascaded from the master’s IR2 signal.
Slave IRs
0 = No slave 8259A is attached to the master through the IR1 signal of
the master.
1 = A slave 82C59A is attached to the IR1 signal of the master.
0
—
Clear this bit to guarantee device operation.
D-29
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.26 ICW3 (SLAVE)
Initialization Command Word 3
ICW3 (slave)
Expanded Addr: F0A1H
ISA Addr:
00A1H
(write only)
Reset State:
XXH
7
0
0
0
0
0
0
0
1
0
Bit
Number
Bit
Mnemonic
Function
7–2
—
Clear these bits to guarantee device operation.
Set this bit to guarantee device operation.
Clear this bit to guarantee device operation.
1
0
—
—
D.27 ICW4 (MASTER AND SLAVE)
Initialization Command Word 4
ICW4 (master and slave)
(write only)
master slave
Expanded Addr:
ISA Addr:
Reset State:
F021H
0021H
XXH
F0A1H
00A1H
XXH
7
0
0
0
0
SFNM
0
0
AEOI
1
Bit
Number
Bit
Mnemonic
Function
7–5
—
Write zero to these bits to guarantee device operation.
4
SFNM
Special-fully Nested Mode:
0 = Selects fully nested mode.
1 = Selects special-fully nested mode. Only the master 82C59A can
operate in special-fully nested mode.
3–2
1
—
Write zero to these bits to guarantee device operation.
AEOI
Automatic EOI Mode:
0 = Disables automatic EOI mode.
1 = Enables automatic EOI mode. Only the master 82C59A can operate
in automatic EOI mode.
0
—
Write one to this bit to guarantee device operation.
D-30
SYSTEM REGISTER QUICK REFERENCE
D.28 IDCODE
Identification Code Register
IDCODE
2027 0013H (3V)
2827 0013H (5V)
Reset State:
31
24
0
0
1
0
0 (3V)
0
0
0
1 (5V)
23
15
7
16
8
0
0
0
0
0
0
1
0
0
0
0
1
0
0
0
1
0
0
1
0
1
1
0
1
0
Bit
Number
Bit
Mnemonic
Function
V3:0
PN15:0
MFR10:0
IDP
31–28
27–12
11–1
0
Device version number.
Device part number.
Manufacturer identification (compressed JEDEC106-A code).
Identification Present. Always true for this device.
This is the first data bit shifted out of the device during a data scan
immediately following an exit from the test-logic-reset state. A one
indicates that an IDCODE register is present. (A zero originates from the
BYPASS register and indicates that the device being interrogated has no
IDCODE register.)
D-31
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.29 IERn
Interrupt Enable
IER0
IER1
IER0, IER1
(read/write)
Expanded Addr:
ISA Addr:
Reset State:
F4F9H
03F9H
00H
F8F9H
02F9H
00H
7
0
—
—
—
—
MS
RLS
TBE
RBF
Bit
Number
Bit
Mnemonic
Function
7–4
3
—
Reserved; for compatibility with future devices, write zeros to these bits.
Modem Status Interrupt Enable:
MS
0 = Modem input signal changes do not cause interrupts.
1 = Connects the modem status signal to the interrupt control unit’s
SIOINTn output. A change on one or more of the modem input
signals activates the modem status signal.
2
RLS
Receiver Line Status Interrupt Enable:
0 = LSR error conditions do not cause interrupts.
1 = Connects the receiver line status signal to the interrupt control unit’s
SIOINTn output. Sources for this interrupt include overrun error,
parity error, framing error, and break interrupt.
1
0
TBE
RBF
Transmit Buffer Empty Interrupt Enable:
0 = Transmit Buffer Empty signal does not cause interrupts.
1 = Connects the transmit buffer empty signal to the interrupt control
unit’s SIOINTn output.
Receive Buffer Full Interrupt Enable:
0 = Receive buffer full signal does not cause interrupts.
1 = Connects the receive buffer full signal to the interrupt control unit’s
SIOINTn output.
NOTE: The interrupt enable register is multiplexed with the divisor latch high register. You must clear
bit 7 (DLAB) of the serial line control register (LCRn) before you can access the interrupt
control register.
D-32
SYSTEM REGISTER QUICK REFERENCE
D.30 IIRn
Interrupt ID
IIR0, IIR1
(read only)
IIR0
IIR1
Expanded Addr:
ISA Addr:
Reset State:
F4FAH F8FAH
03FAH
01H
02FAH
01H
7
0
—
—
—
—
—
IS2
IS1
IP#
Bit
Number
Bit
Mnemonic
Function
7–3
2
—
Reserved. These bits are undefined.
Interrupt Source:
IS2:1
If an interrupt is pending (bit 0 = 0), these bits specify which status signal
caused the pending interrupt.
IS2
IS1
Interrupt Source
0
0
1
1
0
1
0
1
modem status signal*
transmitter buffer empty signal
receive buffer full signal
receiver line status signal**
* When one of the modem input signals (CTSn#, DSRn#, RIn#, and
DCDn#) changes state, the modem status signal is activated.
** A framing error, overrun error, parity error, or break interrupt activates
the receiver line status signal.
Reading the modem status register clears the modem status signal.
Reading the IIRn register or writing to the transmit buffer register clears
the transmit buffer empty signal. Reading the receive buffer register
clears the receive buffer full signal. Reading the receive buffer register or
the serial line status register clears the LSRn error bits, which clears the
receiver line status signal.
0
IP#
Interrupt Pending:
This bit indicates whether an interrupt is pending.
0 = Interrupt is pending
1 = No interrupt is pending
D-33
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.31 INTCFG
Interrupt Configuration
INTCFG
Expanded Addr: F832H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
CE
IR3
IR4
SWAP
IR6
IR5/IR4
IR1
IR0
Bit
Number
Bit
Mnemonic
Function
7
CE
Cascade Enable:
0 = Disables the cascade signals CAS2:0 from appearing on the A18:16
address lines during interrupt acknowledge cycles.
1 = Enables the cascade signals CAS2:0, providing access to external
slave 82C59A devices. The cascade signals are used to address
specific slaves. If enabled, slave IDs appear on the A18:16 address
lines during interrupt acknowledge cycles, but are high during idle
cycles.
6
5
4
IR3
Internal Master IR3 Connection:
See Table 5-1 on page 5-8 for all the IR3 configuration options.
IR4
Internal Master IR4 Connection:
See Table 5-2 on page 5-8 for all the IR4 configuration options.
SWAP
INT6/DMAINT Connection:
0 = Connects DMAINT to the slave IR4. Connects INT6 to the slave IR5.
1 = Connects the INT6 pin to the slave IR4. Connects DMAINT to the slave
IR5.
3
2
IR6
Internal Slave IR6 Connection:
0 = Connects VSS to the slave IR6 signal.
1 = Connects the INT7 pin to the slave IR6 signal.
IR5/IR4
Internal Slave IR4 or IR5 Connection:
These depend on whether INTCFG.4 is set or clear.
0 = Connects VSS to the slave IR5 signal.
1 = Connects either the INT6 pin or DMAINT to the slave IR5 signal.
1
0
IR1
IR0
Internal Slave IR1 Connection:
0 = Connects the SSIO interrupt signal (SSIOINT) to the slave IR1 signal.
1 = Connects the INT5 pin to the slave IR1 signal.
Internal Slave IR0 Connection:
0 = Connects VSS to the slave IR0 signal.
1 = Connects the INT4 pin to the slave IR0 signal.
D-34
SYSTEM REGISTER QUICK REFERENCE
D.32 IR
Instruction Register
IR
Reset State
(Using TRST#):
02H
3
0
INST3
INST2
INST1
INST0
Bit
Number
Bit
Mnemonic
Function
Instruction opcode. At reset (using TRST#, or after 5 TCK cycles with
3–0
INST3:0
TMS held low), this field is loaded with 0010, the opcode for the IDCODE
instruction. Instructions are shifted into this field serially through the TDI
pin. (Table 18-4 lists the valid instruction opcodes.)
D-35
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.33 LCRn
Serial Line Control
LCR0
LCR1
LCR0, LCR1
(read/write)
Expanded Addr:
ISA Addr:
Reset State:
F4FBH F8FBH
03FBH 02FBH
00H
00H
7
0
DLAB
SB
Bit
SP
EPS
PEN
STB
WLS1
WLS0
Bit
Number
Function
Mnemonic
7
DLAB
Divisor Latch Access Bit:
This bit determines which of the multiplexed registers is accessed.
0 = Allows access to the receiver and transmit buffer registers (RBRn and
TBRn) and the interrupt enable register (IERn).
1 = Allows access to the divisor latch registers (DLLn and DLHn).
6
SB
Set Break:
0 = No effect on TXDn.
1 = Forces the TXDn pin to the spacing (logic 0) state for as long as bit is
set.
5
4
3
SP
Sticky Parity, Even Parity Select, and Parity Enable:
These bits determine whether the control logic produces (during
transmission) or checks for (during reception) even, odd, no, or forced
parity.
EPS
PEN
SP
X
0
0
1
EPS PEN Function
X
0
1
0
1
0
1
1
1
1
parity disabled (no parity option)
produce or check for odd parity
produce or check for even parity
produce or check for forced parity (parity bit = 1)
produce or check for forced parity (parity bit = 0)
1
2
STB
Stop Bits:
This bit specifies the number of stop bits transmitted and received in each
serial character.
0 = 1 stop bit
1 = 2 stop bits (1.5 stop bits for 5-bit characters)
1–0
WLS1:0
Word Length Select:
These bits specify the number of data bits in each transmitted or received
serial character.
00 = 5-bit character
01 = 6-bit character
10 = 7-bit character
11 = 8-bit character
D-36
SYSTEM REGISTER QUICK REFERENCE
D.34 LSRn
Serial Line Status
LSR0, LSR1
(read only)
LSR0
F4FDH F8FDH
03FDH 02FDH
LSR1
Expanded Addr:
ISA Addr:
Reset State:
60H
60H
7
0
—
TE
TBE
BI
FE
PE
OE
RBF
Bit
Number
Bit
Mnemonic
Function
7
—
Reserved. This bit is undefined.
Transmitter Empty:
6
5
4
TE
TBE
BI
The transmitter sets this bit to indicate that the transmit shift register and
transmit buffer register are both empty. Writing to the transmit buffer
register clears this bit.
Transmit Buffer Empty:
The transmitter sets this bit after it transfers data from the transmit buffer
to the transmit shift register. Writing to the transmit buffer register clears
this bit.
Break Interrupt:
The receiver sets this bit whenever the received data input is held in the
spacing (logic 0) state for longer than a full word transmission time.
Reading the receive buffer register or the serial line status register clears
this bit.
3
FE
Framing Error
The receiver sets this bit to indicate that the received character did not
have a valid stop bit. Reading the receive buffer register or the serial line
status register clears this bit. If data frame is set for two stop bits the
second stop bit is ignored.
2
1
PE
OE
Parity Error:
The receiver sets this bit to indicate that the received data character did
not have the correct parity. Reading the receive buffer register or the
serial line status register clears this bit.
Overrun Error:
The receiver sets this bit to indicate an overrun error. An overrun occurs
when the receiver transfers a received character to the receive buffer
register before the CPU reads the buffer’s old character. Reading the
serial line status register clears this bit.
0
RBF
Receive Buffer Full:
The receiver sets this bit after it transfers a received character from the
receive shift register to the receive buffer register. Reading the receive
buffer register clears this bit.
D-37
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.35 MCRn
Modem Control
MCR0, MCR1
(read/write)
MCR0
F4FCH
03FCH
00H
MCR1
F8FCH
02FCH
00H
Expanded Addr:
ISA Addr:
Reset State:
7
0
—
—
—
LOOP
OUT2
OUT1
RTS
DTR
Bit
Number
Bit
Mnemonic
Function
7–5
4
—
Reserved; for compatibility with future devices, write zeros to these bits.
LOOP
Loop Back Test Mode:
0 = Normal mode
1 = Setting this bit puts the SIOn into diagnostic (or loop back test) mode. This causes
the SIO channel to:
•
•
•
•
set its transmit serial output (TXDn)
disconnect its receive serial input (RXDn) from the package pin
loop back the transmitter shift register’s output to the receive shift register’s input
disconnect the modem control inputs (CTSn#, DSRn#, RIn#, and DCDn#) from the
package pins
•
•
force modem control outputs (RTSn# and DTRn#) to their inactive states
connects MCRn bits to MSRn bits
3–2
OUT2:1
Test Bits:
In diagnostic mode (bit 4=1), these bits control the ring indicator (RIn) and data carrier
detect (DCDn#) modem inputs. Setting OUT1 activates the internal RIn bit; clearing
OUT1 deactivates the internal RIn bit. Setting OUT2 activates the internal DCDn bit;
clear OUT2 deactivates the internal DCDn bit.
In normal user mode (bit 4=0) OUT1 has no effect and OUT2 in conjunction with
INTCFG.5/6 selects internal SIO interrupt or external interrupt. See Table 5-1 on page
5-8 for the configuration options.
1
RTS
Ready to Send:
The function of this bit depends on whether the SIOn is in diagnostic mode
(MCRn.4=1), internal connection mode, or standard mode.
In diagnostic mode, setting this bit activates the internal CTSn bit; clearing this bit
deactivates the internal CTSn bit.
In internal connection mode, setting this bit activates the internal CTSn# signal and the
RTSn# pin; clearing this bit deactivates the internal CTSn# signal and the RTSn# pin.
In standard mode, setting this bit activates the RTSn# pin; clearing this bit deactivates
the RTSn# pin. Note that pin is inverted from bit.
0
DTR
Data Terminal Ready:
The function of this bit depends on whether the SIOn is in diagnostic mode
(MCRn.4=1), internal connection mode, or standard mode.
In diagnostic mode, setting this bit activates the internal DSRn# signal; clearing this bit
deactivates the internal DSRn# signal.
In internal connection mode, setting this bit activates the internal DSRn# and DCDn#
signals and the DTRn# pin; clearing this bit deactivates the internal DSRn# and DCDn#
signals and the DTRn# pin. Note that pin is inverted from bit.
In standard mode, setting this bit activates the DTRn# pin; clearing this bit deactivates
the DTRn# pin. Note that pin is inverted from bit.
D-38
SYSTEM REGISTER QUICK REFERENCE
D.36 MSRn
Modem Status
MSR0, MSR1
(read only)
MSR0
F4FEH F8FEH
03FEH
X0H
MSR1
Expanded Addr:
ISA Addr:
Reset State:
02FEH
X0H
7
0
DCD
RI
DSR
CTS
DDCD
TERI
DDSR
DCTS
Bit
Number
Bit
Mnemonic
Function
7
DCD
Data Carrier Detect:
This bit is the complement of the data carrier detect (DCDn#) input. In
diagnostic test mode, this bit is equivalent to MCRn.3 (OUT2).
6
5
4
3
RI
Ring Indicator:
This bit is the complement of the ring indicator (RIn#) input. In diagnostic
test mode, this bit is equivalent to MCRn.2 (OUT1).
DSR
CTS
DDCD
Data Set Ready:
This bit is the complement of the data set ready (DSRn#) input. In
diagnostic test mode, this bit is equivalent to MCRn.0 (DTR).
Clear to Send:
This bit is the complement of the clear to send (CTSn#) input. In
diagnostic test mode, this bit is equivalent to MCRn.1 (RTS).
Delta Data Carrier Detect:
When set, this bit indicates that the DCDn# input has changed state
since the last time this register was read. Reading this register clears
this bit.
2
1
0
TERI
Trailing Edge Ring Indicator:
When set, this bit indicates that the RIn# input has changed from a low
to a high state since the last time this register was read. Reading this
register clears this bit.
DDSR
DCTS
Delta Data Set Ready:
When set, this bit indicates that the DSRn# input has changed state
since the last time this register was read. Reading this register clears
this bit.
Delta Clear to Send:
When set, this bit indicates that the CTSn# input has changed state
since the last time this register was read. Reading this register clears
this bit.
D-39
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.37 OCW1 (MASTER AND SLAVE)
Operation Command Word 1
master slave
OCW1 (master and slave)
(read/write)
Expanded Addr:
ISA Addr:
Reset State:
F021H
0021H
XXH
F0A1H
00A1H
XXH
7
0
M7
M6
M5
M4
M3
M2
M1
M0
Bit
Number
Bit
Mnemonic
Function
7–0
M7:0
Mask IR:
0 = Enables interrupts on the corresponding IR signal.
1 = Disables interrupts on the corresponding IR signal.
NOTE: Setting the mask bit does not clear the respective interrupt
pending bit.
NOTE: The 8259A must be initialized before it can be used. After reset, the 8259A register states are
undefined. The 8259A modules must be initialized before the IF flag in the core FLAG register is
set.
D-40
SYSTEM REGISTER QUICK REFERENCE
D.38 OCW2 (MASTER AND SLAVE)
Operation Command Word 2
OCW2 (master and slave)
(write only)
master slave
Expanded Addr:
ISA Addr:
Reset State:
F020H
0020H
XXH
F0A0H
00A0H
XXH
7
0
R
SL
EOI
RSEL1
RSEL0
L2
L1
L0
Bit
Number
Bit
Mnemonic
Function
7
R
The Rotate (R), Specific Level (SL), and End-of-Interrupt (EOI) Bits:
These bits change the priority structure and/or send an EOI command.
6
5
SL
R
SL EOI Command
EOI
0
0
0
0
1
1
1
1
*
0
0
1
1
0
0
1
1
0
1
0
1
0
1
0
1
Cancel automatic rotation*
Send a nonspecific EOI command
No operation
Send a specific EOI command**
Enable automatic rotation*
Enable automatic rotation and send a nonspecific EOI
Initiate specific rotation**
Initiate specific rotation and send a specific EOI**
These cases allow you to change the priority structure while the
82C59A is operating in the automatic EOI mode.
** The L2:0 bits (see below) specify the specific level for these cases.
Register Select Bits:
4–3
RSEL1:0
ICW1, OCW2 and OCW3 are accessed through the same addresses.
The states of RSEL1:0 determine which register is accessed. Write 00
to these bits to access OCW2.
RSEL1 RSEL0
0
0
1
0
1
X
OCW2
OCW3
ICW1
2–0
L2:0
IR Level:
When you program bits 7–5 to initiate specific rotation, these bits specify
the IR signal that is assigned the lowest level.
When you program bits 7–5 to send a specific EOI command, these bits
specify the IR signal that receives the EOI command.
If SL=0, then these bits have no effect.
D-41
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.39 OCW3 (MASTER AND SLAVE)
Operation Command Word 3
master slave
OCW3 (master and slave)
(write only)
Expanded Addr:
ISA Addr:
Reset State:
F020H
0020H
XXH
F0A0H
00A0H
XXH
7
0
0
ESMM
SMM
RSEL1
RSEL0
POLL
ENRR
RDSEL
Bit
Number
Bit
Mnemonic
Function
Clear this bit to guarantee device operation.
7
—
6
5
ESMM
SMM
Enable Special Mask Mode (ESMM) and Special Mask Mode (SMM):
Use these bits to enable or disable special mask mode.
ESMM SMM
0
0
1
1
0
1
0
1
No action
No action
Disable special mask mode
Enable special mask mode
4–3
RSEL1:0
Register Select:
ICW1, OCW2 and OCW3 are accessed through the same addresses.
The states of RSEL1:0 determine which register is accessed. Write 01
to these bits to access OCW3.
RSEL1 RSEL0
0
0
1
0
1
X
OCW2
OCW3
ICW1
2
POLL
Poll Command:
Set this bit to issue a poll command, thus initiating the polling process.
1
0
ENRR
Enable Register Read Select (ENRR) and Read Register Select
(RDSEL):
RDSEL
These bits select which register is read during the next F020H and
F0A0H (or PC/AT address 0020H, 00A0H) read access.
ENRR
RDSEL Register Read on Next Read Pulse
0
0
1
1
0
1
0
1
No action
No action
Interrupt Request Register
In-service Register
D-42
SYSTEM REGISTER QUICK REFERENCE
Expanded Addr: F820H
D.40 P1CFG
Port 1 Configuration
P1CFG
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
PM7
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7
PM7
Pin Mode:
0 = Selects P1.7 at the package pin.
1 = Selects HLDA at the package pin.
6
5
4
3
2
1
0
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Pin Mode:
0 = Selects P1.6 at the package pin.
1 = Selects HOLD at the package pin.
Pin Mode:
0 = Selects P1.5 at the package pin.
1 = Selects LOCK# at the package pin.
Pin Mode:
0 = Selects P1.4 at the package pin.
1 = Selects RI0# at the package pin.
Pin Mode:
0 = Selects P1.3 at the package pin.
1 = Selects DSR0# at the package pin.
Pin Mode:
0 = Selects P1.2 at the package pin.
1 = Selects DTR0# at the package pin.
Pin Mode:
0 = Selects P1.1 at the package pin.
1 = Selects RTS0# at the package pin.
Pin Mode:
0 = Selects P1.0 at the package pin.
1 = Selects DCD0# at the package pin.
D-43
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.41 P2CFG
Port 2 Configuration
P2CFG
Expanded Addr: F822H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
PM7
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7
PM7
Pin Mode:
0 = Selects P2.7 at the package pin.
1 = Selects CTS0# at the package pin.
6
5
4
3
2
1
0
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Pin Mode:
0 = Selects P2.6 at the package pin.
1 = Selects TXD0 at the package pin.
Pin Mode:
0 = Selects P2.5 at the package pin.
1 = Selects RXD0 at the package pin.
Pin Mode:
0 = Selects P2.4 at the package pin.
1 = Selects CS4# at the package pin.
Pin Mode:
0 = Selects P2.3 at the package pin.
1 = Selects CS3# at the package pin.
Pin Mode:
0 = Selects P2.2 at the package pin.
1 = Selects CS2# at the package pin.
Pin Mode:
0 = Selects P2.1 at the package pin.
1 = Selects CS1# at the package pin.
Pin Mode:
0 = Selects P2.0 at the package pin.
1 = Selects CS0# at the package pin.
D-44
SYSTEM REGISTER QUICK REFERENCE
Expanded Addr: F824H
D.42 P3CFG
Port 3 Configuration
P3CFG
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
PM7
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7
PM7
Pin Mode:
0 = Selects P3.7 at the package pin.
1 = Selects COMCLK at the package pin.
6
5
4
3
2
PM6
PM5
PM4
PM3
PM2
Pin Mode:
0 = Selects P3.6 at the package pin.
1 = Selects PWRDOWN at the package pin.
Pin Mode:
0 = Selects P3.5 at the package pin.
1 = Connects master IR7 to the package pin (INT3).
Pin Mode:
0 = Selects P3.4 at the package pin.
1 = Connects master IR6 to the package pin (INT2).
Pin Mode:
0 = Selects P3.3 at the package pin.
1 = Connects master IR5 to the package pin (INT1).
Pin Mode:
0 = Selects P3.2 at the package pin.
1 = Connects master IR1 to the package pin (INT0).
1
0
PM1
PM0
Pin Mode:
See Table 5-1 on page 5-8 for all the PM1 configuration options.
Pin Mode:
See Table 5-1 on page 5-8 for all the PM0 configuration options.
D-45
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.43 PINCFG
Pin Configuration
PINCFG
Expanded Addr: F826H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
—
PM6
PM5
PM4
PM3
PM2
PM1
PM0
Bit
Number
Bit
Mnemonic
Function
7
—
Reserved. This bit is undefined; for compatibility with future devices, do
not modify this bit.
6
5
PM6
PM5
Pin Mode:
0 = Selects CS6# at the package pin.
1 = Selects REFRESH# at the package pin.
Pin Mode:
0 = Selects the coprocessor signals, PEREQ, BUSY#, and ERROR#, at
the package pins.
1 = Selects the timer control unit signals, TMROUT2, TMRCLK2, and
TMRGATE2, at the package pins.
4
3
2
1
0
PM4
PM3
PM2
PM1
PM0
Pin Mode:
0 = Selects DACK0# at the package pin.
1 = Selects CS5# at the package pin.
Pin Mode:
0 = Selects EOP# at the package pin.
1 = Selects CTS1# at the package pin.
Pin Mode:
0 = Selects DACK1# at the package pin.
1 = Selects TXD1 at the package pin.
Pin Mode:
0 = Selects SRXCLK at the package pin.
1 = Selects DTR1# at the package pin.
Pin Mode:
0 = Selects SSIOTX at the package pin.
1 = Selects RTS1# at the package pin.
D-46
SYSTEM REGISTER QUICK REFERENCE
Expanded Addr: F864H, F86CH, F874H
D.44 PnDIR
Port DIrection
PnDIR (n=1–3)
(read/write)
ISA Addr:
—
Reset State:
FFH
7
0
PD7
PD6
PD5
PD4
PD3
PD2
PD1
PD0
Bit
Number
Bit
Mnemonic
Function
7–0
PD7:0
Pin Direction:
0 = Configures the pin as a complementary output.
1 = Configures the pin as an open-drain output or high-impedance input.
D-47
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.45 PnLTC
Port Data Latch
PnLTC (n=1–3)
(read/write)
Expanded Addr: F862H, F86AH, F872H
ISA Addr:
—
Reset State:
FFH
7
0
PL7
PL6
PL5
PL4
PL3
PL2
PL1
PL0
Bit
Number
Bit
Mnemonic
Function
7–0
PL7:0
Port Data Latch:
Writing a value to a PL bit causes that value to be driven onto the
corresponding pin.
For a complementary output, write the desired pin value to its PL bit.
This value is strongly driven onto the pin.
For an open-drain output, a one results in a high-impedance (input) state
at the pin.
For a high-impedance input, write a one to the corresponding PL bit. A
one results in a high-impedance state at the pin, allowing external
hardware to drive it.
D.46 PnPIN
Port Pin State
PnPIN (n=1–3)
(read only)
Expanded Addr: F860H, F868H, F870H
ISA Addr:
—
Reset State:
XXH
7
0
PS7
PS6
PS5
PS4
PS3
PS2
PS1
PS0
Bit
Number
Bit
Mnemonic
Function
7–0
PS7:0
Pin State:
Reading a PS bit returns the logic state present on the associated port
pin.
D-48
SYSTEM REGISTER QUICK REFERENCE
D.47 POLL (MASTER AND SLAVE)
Poll Status Byte
POLL (master and slave)
(read only)
master slave
Expanded Addr:
ISA Addr:
Reset State:
F020H
0020H
XXH
F0A0H
00A0H
XXH
7
0
INT
—
—
—
—
L2
L1
L0
Bit
Number
Bit
Mnemonic
Function
7
INT
Interrupt Pending:
0 = No request pending.
1 = Indicates that a device attached to the 82C59A requires servicing.
6–3
2–0
—
Reserved. These bits are undefined.
L2:0
Interrupt Request Level:
When bit 7 is set, these bits indicate the highest-priority IR signal that
requires servicing. When bit 7 is clear, i.e., no request is pending, these
bits are indeterminate.
D-49
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.48 PORT92
Port 92 Configuration
PORT92
Expanded Addr: F092H
ISA Addr:
0092H
(read/write)
Reset State:
XXXXXX10B
7
0
—
—
—
—
—
—
A20G
CPURST
Bit
Number
Bit
Mnemonic
Function
7–2
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
1
0
A20G
A20 Grounded:
0 = Clearing this bit forces address line A20 to 0. This bit affects
addresses generated only by the core. Addresses generated by the
DMA and the Refresh Unit are not affected by this bit.
1 = Setting this bit leaves core-generated addresses unmodified.
CPURST
CPU Reset:
0 = Clearing this bit performs no operation.
1 = Setting this bit resets the core without resetting the peripherals.
This bit must be cleared before issuing another reset.
D-50
SYSTEM REGISTER QUICK REFERENCE
D.49 PWRCON
Power Control Register
PWRCON
Expanded Addr:
ISA Addr:
F800H
—
(read/write)
Reset State:
00H
7
0
—
—
—
—
WDTRDY HSREADY
PC1
PC0
Bit
Number
Bit
Mnemonic
Function
7–4
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
3
WDTRDY
HSREADY
PC1:0
Watch Dog Timer Ready:
0 = An external READY must be generated to terminate the cycle when
the WDT times out in Bus Monitor Mode.
1 = Internal logic generates READY# to terminate the cycle when the
WDT times out in Bus Monitor Mode.
2
Halt/Shutdown Ready:
0 = An external ready must be generated to terminate a HALT/Shutdown
cycle.
1 = Internal logic generates READY# to terminate a HALT/Shutdown
cycle.
1–0
Power Control:
Program these bits, then execute a HALT instruction. The device enters
the programmed mode when READY# (internal or external) terminates
the halt bus cycle. When these bits have equal values, the HALT
instruction causes a normal halt and the device remains in active mode.
PC1 PC0
0
1
0
1
0
0
1
1
active mode
idle mode
powerdown mode
active mode
D-51
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.50 RBRn
Receive Buffer
RBR0, RBR1
(read only)
RBR0
F4F8H
03F8H
XXH
RBR1
F8F8H
02F8H
XXH
Expanded Addr:
ISA Addr:
Reset State:
7
0
RB7
RB6
RB5
RB4
RB3
RB2
RB1
RB0
Bit
Number
Bit
Mnemonic
Function
7–0
RB7:0
Receive Buffer Bits:
These bits make up the last word received. The receiver shifts bits in,
starting with the least-significant-bit. The receiver then strips off the
asynchronous bits (start, parity, and stop) and transfers the received
data bits from the receive shift register to the receive buffer.
NOTE: The receive buffer register shares an address port with other SIO registers. Bit 7 (DLAB) of
the LCRn must be cleared in order to read the receive buffer register.
D-52
SYSTEM REGISTER QUICK REFERENCE
D.51 REMAPCFG
Address Configuration Register
REMAPCFG
Expanded Addr: 0022H
PC/AT Address: 0022H
Reset State:
0000H
15
8
0
ESE
—
—
—
—
—
—
—
—
—
7
S1R
S0R
ISR
IMR
DR
TR
Bit
Number
Bit
Mnemonic
Function
15
ESE
0 = Disables expanded I/O space
1 = Enables expanded I/O space
14–7
6
—
Reserved.
S1R
0 = Makes serial channel 1 (COM2) accessible in both DOS I/O space
and expanded I/O space
1 = Remaps serial channel 1 (COM2) address into expanded I/O space
5
4
S0R
ISR
0 = Makes serial channel 0 (COM1) accessible in both DOS I/O space
and expanded I/O space
1 = Remaps serial channel 0 (COM1) address into expanded I/O space
0 = Makes the slave 82C59A interrupt controller accessible in both DOS
I/O space and expanded I/O space
1 = Remaps slave 82C59A interrupt controller address into expanded
I/O space
3
2
IMR
DR
0 = Makes the master 82C59A interrupt controller accessible in both
DOS I/O space and expanded I/O space
1 = Remaps master 82C59A interrupt controller address into expanded
I/O space
0 = Makes the DMA address accessible in both DOS I/O space and
expanded I/O space
1 = Remaps DMA address into expanded I/O space
1
0
—
Reserved.
TR
0 = Makes the timer control unit accessible in both DOS I/O space and
expanded I/O space
1 = Remaps timer control unit address into expanded I/O space
D-53
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.52 RFSADD
Refresh Address
RFSADD
Expanded Addr: F4A6H
ISA Addr:
—
(read/write)
Reset State:
00FFH
15
8
0
—
—
RA13
RA5
RA12
RA4
RA11
RA10
RA2
RA9
RA1
RA8
1
7
RA7
RA6
RA3
Bit
Number
Bit
Mnemonic
Function
15–14
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
13–1
0
RA13:1
—
Refresh Address Bits:
These bits comprise A13:1 of the refresh address.
Refresh Bit 0:
A0 of the refresh address. This bit is always 1 and is read-only.
D.53 RFSBAD
Refresh Base Address
RFSBAD
Expanded Addr: F4A0H
ISA Addr:
—
(read/write)
Reset State:
0000H
15
8
0
—
—
—
—
RA25
RA24
RA16
RA23
RA15
RA22
RA14
7
RA21
RA20
RA19
RA18
RA17
Bit
Number
Bit
Mnemonic
Function
15–12
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
11–0
RA25:14
Refresh Base:
These bits make up the A25:14 address bits of the refresh address. This
establishes a memory region for refreshing.
D-54
SYSTEM REGISTER QUICK REFERENCE
Expanded Addr: F4A2H
D.54 RFSCIR
Refresh Clock Interval
RFSCIR
ISA Addr:
—
(read/write)
Reset State:
0000H
15
8
0
—
—
—
—
—
—
RC9
RC1
RC8
RC0
7
RC7
RC6
RC5
RC4
RC3
RC2
Bit
Number
Bit
Mnemonic
Function
15–10
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
9–0
RC9:0
Refresh Counter Value:
Write the counter value to these ten bits. The interval counter counts
down from this value. When the interval counter reaches one, the control
unit initiates a refresh request (provided it does not have a request
pending). The counter value is a function of DRAM specifications and
processor frequency (see the equation above).
D.55 RFSCON
Refresh Control
RFSCON
Expanded Addr: F4A4H
ISA Addr:
—
(read/write)
Reset State:
0000H
15
8
0
REN
CV7
—
—
—
—
—
CV9
CV1
CV8
CV0
7
CV6
CV5
CV4
CV3
CV2
Bit
Number
Bit
Mnemonic
Function
15
REN
Refresh Control Unit Enable:
This bit enables or disables the refresh control unit.
0 = Disables refresh control unit
1 = Enables refresh control unit
14–10
9–0
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
CV9:0
Counter Value:
These read-only bits represent the current value of the interval counter.
Write operations to these bits have no effect.
D-55
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.56 SCRn
Scratch Pad
SCR0, SCR1
(read/write)
SCR0
F4FFH
03FFH
XXH
SCR1
F8FFH
02FFH
XXH
Expanded Addr:
ISA Addr:
Reset State:
7
0
SP7
SP6
SP5
SP4
SP3
SP2
SP1
SP0
Bit
Number
Bit
Mnemonic
Function
Writing and reading this register has no effect on SIOn operation.
7–0
SP7:0
D-56
SYSTEM REGISTER QUICK REFERENCE
Expanded Addr: F836H
D.57 SIOCFG
SIO and SSIO Configuration
SIOCFG
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
S1M
S0M
—
—
—
SSBSRC
S1BSRC
S0BSRC
Bit
Number
Bit
Mnemonic
Function
SIO1 Modem Signal Connections:
7
S1M
0 = Connects the SIO1 modem input signals to the package pins.
1 = Connects the SIO1 modem input signals internally.
6
S0M
SIO0 Modem Signal Connections:
0 = Connects the SIO0 modem input signals to the package pins.
1 = Connects the SIO0 modem input signals internally.
5–3
2
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
SSBSRC
SSIO Baud-rate Generator Clock Source:
0 = Connects the internal PSCLK signal to the SSIO baud-rate
generator.
1 = Connects the internal SERCLK signal to the SSIO baud-rate
generator.
1
0
S1BSRC
S0BSRC
SIO1 Baud-rate Generator Clock Source:
0 = Connects the COMCLK pin to the SIO1 baud-rate generator.
1 = Connects the internal SERCLK signal to the SIO1 baud-rate
generator.
SIO0 Baud-rate Generator Clock Source:
0 = Connects the COMCLK pin to the SIO0 baud-rate generator.
1 = Connects the internal SERCLK signal to the SIO0 baud-rate
generator.
D-57
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.58 SSIOBAUD
SSIO Baud-rate Control
SSIOBAUD
Expanded Addr: F484H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
BEN
BV6
BV5
BV4
BV3
BV2
BV1
BV0
Bit
Number
Bit
Mnemonic
Function
7
BEN
Baud-rate Generator Enable:
Setting this bit enables the baud-rate generator. Clearing this bit disables
the baud-rate generator, clears the baud-rate count value, and forces the
baud rate clock to zero.
6–0
BV6:0
Baud-rate Value:
The baud-rate value (BV) is the reload value for the baud-rate
generator’s seven-bit down counter. The baud-rate generator’s output is
a function of BV and the baud-rate generator’s input (BCLKIN), as
follows.
BCLKIN
baud-rate output frequency (Hz) = --------------------- (Hz)
2BV + 2
If you know the desired output baud-rate frequency, you can determine
BV as follows.
BCLKIN
---------------------------------------------------------------------------------
BV =
– 1
2 × baud-rate output frequency
D-58
SYSTEM REGISTER QUICK REFERENCE
D.59 SSIOCON1
SSIO Control 1
SSIOCON1
Expanded Addr: F486H
ISA Addr:
—
(read/write)
Reset State:
C0H
7
0
TUE
THBE
TIE
TEN
ROE
RHBF
RIE
REN
Bit
Number
Bit
Mnemonic
Function
7
6
TUE
Transmit Underrun Error:
The transmitter sets this bit to indicate a transmit underrun error in the
TEN transfer mode. Clear this bit to clear the error flag. If a one is written
to TUE, it is ignored and TUE retains its previous value.
THBE
Transmit Holding Buffer Empty:
(read only bit)
The transmitter sets this bit when the transmit buffer contents have been
transferred to the transmit shift register, indicating that the buffer is now
ready to accept new data. Writing data to the transmit buffer clears THBE.
When this bit is clear, the buffer is not ready to accept any new data.
5
TIE
Transmitter Interrupt Enable:
0 = Clearing this bit prevents the Interrupt Control Unit from sensing
when the transmit buffer is empty.
1 = Setting this bit connects the transmit buffer empty internal signal to
the Interrupt Control Unit.
4
3
TEN
ROE
Transmitter Enable:
0 = Disables the transmitter.
1 = Enables the transmitter.
Receive Overflow Error:
The receiver sets this bit to indicate a receiver overflow error. Write zero
to this bit to clear the flag.
If a one is written to ROE, the one is ignored and ROE retains its previous
value.
2
1
RHBF
(read only bit)
Receive Holding Buffer Full:
The receiver sets this bit when the receive shift register contents have
been transferred to the receive buffer.
Reading the buffer clears this bit.
Receive Interrupt Enable:
RIE
0 = Clearing this bit prevents the Interrupt Control Unit from sensing
when the receive buffer is full.
1 = Setting this bit connects the receiver buffer full internal signal to the
Interrupt Control Unit.
0
REN
Receiver Enable:
0 = Clearing this bit disables the receiver.
1 = Setting this bit enables the receiver.
D-59
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.60 SSIOCON2
SSIO Control 2
SSIOCON2
Expanded Addr: F488H
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
—
—
—
—
—
AUTOTXM
TXMM
RXMM
Bit
Number
Bit
Mnemonic
Function
7–3
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
2
1
AUTOTXM
TXMM
Automatic Transmit off mode for master mode
0 = Clearing this bit puts the TEN bit into normal operation
1 = Setting this bit and the TXMM bit causes TEN to be ignored. Every
time a word is loaded into the transmit shift register from the transmit
holding buffer it is transmitted out and then stops.
Transmit Master Mode:
0 = Clearing this bit puts the transmitter in slave mode. In slave mode, an
external device controls the transmit serial communications. An input
on the STXCLK pin clocks the transmitter.
1 = Setting this bit puts the transmitter in master mode. In master mode,
the internal baud-rate generator controls the transmit serial
communications. The baud-rate generator’s output clocks the
internal transmitter and appears on the STXCLK pin.
0
RXMM
Receive Master Mode:
0 = Clearing this bit puts the receiver in slave mode. In slave mode, an
external device controls the receive serial communications. An input
on the SRXCLK pin clocks the receiver.
1 = Setting this bit puts the receiver in master mode. In master mode, the
internal baud-rate generator controls the receive serial
communications. The baud-rate generator’s output clocks the
internal receiver and appears on the SRXCLK pin.
D-60
SYSTEM REGISTER QUICK REFERENCE
Expanded Addr: F48AH
D.61 SSIOCTR
Baud-rate Count Down
SSIOCTR
ISA Addr:
—
(read only)
Reset State:
00H
7
0
BSTAT
CV6
CV5
CV4
CV3
CV2
CV1
CV0
Bit
Number
Bit
Mnemonic
Function
7
BSTAT
Baud-rate Generator Status:
0 = The baud-rate generator is disabled.
1 = The baud-rate generator is enabled.
6–0
CV6:0
Current Value:
These bits indicate the current value of the baud-rate down counter.
D.62 SSIORBUF
Receive Holding Buffer
SSIORBUF
Expanded Addr: F482H
ISA Addr:
—
(read only)
Reset State:
0000H
15
8
0
RB15
RB14
RB6
RB13
RB5
RB12
RB4
RB11
RB10
RB2
RB9
RB1
RB8
RB0
7
RB7
RB3
Bit
Number
Bit
Mnemonic
Function
15–0
RB15:0
Receive Buffer Bits:
This register contains the last word received. The receive shift register
shifts bits in with the rising edge of the receiver clock pin. Data is shifted
in starting with the most-significant bit. The control logic then transfers
the received word from the receive shift register to SSIORBUF.
D-61
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.63 SSIOTBUF
Transmit Holding Buffer
SSIOTBUF
Expanded Addr: F480H
ISA Addr:
—
(read/write)
Reset State:
0000H
15
8
0
TB15
TB7
TB14
TB6
TB13
TB5
TB12
TB4
TB11
TB10
TB2
TB9
TB1
TB8
TB0
7
TB3
Bit
Number
Bit
Mnemonic
Function
15–0
TB15:0
Transmit Buffer Bits:
These bits make up the next data word to be transmitted. The control
logic loads this word into the transmit shift register. The transmit shift
register shifts the bits out on the falling edge of the tranmitter clock pin.
The word is transmitted out starting with the most-significant bit (TB15).
D.64 TBRn
Transmit Buffer
TBR0, TBR1
(write only)
TBR0
TBR1
F8F8H
02F8H
XXH
Expanded Addr: F4F8H
ISA Addr:
Reset State:
03F8H
XXH
7
0
TB7
TB6
TB5
TB4
TB3
TB2
TB1
TB0
Bit
Number
Bit
Mnemonic
Function
7–0
TB7:0
Transmit Buffer Bits:
These bits make up the next data word to be transmitted. The transmitter
loads this word into the transmit shift register. The transmit shift register
then shifts the bits out, along with the asynchronous communication bits
(start, stop, and parity). The data bits are shifted out least-significant bit
(TB0) first.
NOTE: The transmit buffer register shares an address port with other SIO registers. You must clear
bit 7 (DLAB) of LCRn before you can write to the transmit buffer register.
D-62
SYSTEM REGISTER QUICK REFERENCE
Expanded Addr: F834H
D.65 TMRCFG
.
Timer Configuration
TMRCFG
ISA Addr:
—
(read/write)
Reset State:
00H
7
0
TMRDIS
SWGTEN GT2CON CK2CON
GT1CON CK1CON GT0CON CK0CON
Bit
Number
Bit
Mnemonic
Function
7
TMRDIS
Timer Disable:
0 = Enables the CLKINn signals.
1 = Disables the CLKINn signals.
6
5
SWGTEN
Software GATEn Enable
0 = Connects GATEn to either the VCC pin or the TMRGATEn pin.
1 = Enables GT2CON, GT1CON, and GT0CON to control the
connections to GATE2, GATE1 and GATE0 respectively.
GT2CON
Gate 2 Connection:
SWGTEN
GT2CON
0
0
1
1
0
1
0
1
Connects GATE2 to VCC.
Connects GATE2 to the TMRGATE2 pin.
Turns GATE2 off.
Turns GATE2 on.
4
3
CK2CON
GT1CON
Clock 2 Connection:
0 = Connects CLKIN2 to the internal PSCLK signal.
1 = Connects CLKIN2 to the TMRCLK2 pin.
Gate 1 Connection:
SWGTEN
GT1CON
0
0
1
1
0
1
0
1
Connects GATE1 to VCC.
Connects GATE1 to the TMRGATE1 pin.
Turns GATE1 off.
Turns GATE1 on.
2
1
CK1CON
GT0CON
Clock 1 Connection:
0 = Connects CLKIN1 to the internal PSCLK signal.
1 = Connects CLKIN1 to the TMRCLK1 pin.
Gate 0 Connection:
SWGTEN
GT0CON
0
0
1
1
0
1
0
1
Connects GATE0 to VCC.
Connects GATE0 to the TMRGATE1 pin.
Turns GATE0 off.
Turns GATE0 on.
0
CK0CON
Clock 0 Connection:
0 = Connects CLKIN0 to the internal PSCLK signal.
1 = Connects CLKIN0 to the TMRCLK0 pin.
D-63
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.66 TMRCON
Timer Control (Control Word Format)
TMRCON
Expanded Addr: F043H
ISA Addr:
0043H
Reset State:
XXH
7
0
SC1
SC0
RW1
RW0
M2
M1
M0
CNTFMT
Bit
Number
Bit
Mnemonic
Function
7–6
SC1:0
Select Counter:
Use these bits to specify a particular counter. The selections you make for
bits 5–0 define this counter’s operation.
00 = counter 0
01 = counter 1
10 = counter 2
11 is not an option for TMRCON’s control word format. Selecting 11
accesses TMRCON’s read-back format, which is shown in Figure 10-29.
5–4
RW1:0
Read/Write Select:
These bits select a read/write option for the counter specified by bits 7–6.
01 = read/write least-significant byte only
10 = read/write most-significant byte only
11 = read/write least-significant byte first, then most-significant byte
00 is not an option for TMRCON’s control word format. Selecting 00
accesses TMRCON’s counter-latch format, which is shown in Figure
10-27.
3–1
M2:0
Mode Select:
These bits select an operating mode for the counter specified by bits 7–6.
000 = mode 0
001 = mode 1
X10 = mode 2
X11 = mode 3
100 = mode 4
101 = mode 5
X is a don’t care.
0
CNTFMT
Count Format:
This bit selects the count format for the counter specified by bits 7–6.
0 = binary (16 bits)
1 = binary coded decimal (4 decades)
D-64
SYSTEM REGISTER QUICK REFERENCE
D.67 TMRn
Timer n (Read Format)
TMRn (n = 0–2)
Expanded Addr: F040H, F041H
F042H
ISA Addr:
0040H, 0041H
0042H
Reset State:
XXH
7
0
CV7
CV6
CV5
CV4
CV3
CV2
CV1
CV0
Bit
Number
Bit
Mnemonic
Function
7–0
CV7:0
Count Value:
These bits contain the counter’s count value. When reading the
counter’s count value, follow the read selection specified in the counter’s
control word.
D-65
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Timer n (Status Format)
TMRn (n = 0–2)
Expanded Addr: F040H, F041H
F042H
ISA Addr:
0040H, 0041H
0042H
Reset State:
XXH
7
0
OUTPUT
NULCNT
RW1
RW0
M2
M1
M0
CNTFMT
Bit
Number
Bit
Mnemonic
Function
7
OUTPUT
Output Status:
This bit indicates the current state of the counter’s output signal.
0 = OUTn is low
1 = OUTn is high
6
NULCNT
Count Status:
This bit indicates whether the latest count written to the counter has
been loaded. Some modes require a gate-trigger before the counter
loads new count values.
0 = the latest count written to the counter has been loaded
1 = a count has been written to the counter but has not yet been loaded
5–4
3–1
RW1:0
M2:0
Read/Write Select Status:
These bits indicate the counter’s programmed read/write selection.
00 = Never occurs
01 = read/write least-significant byte only
10 = read/write most-significant byte only
11 = read/write least-significant byte first, then most-significant byte
Mode Status:
These bits indicate the counter’s programmed operating mode.
000 = mode 0
001 = mode 1
X10 = mode 2
X11 = mode 3
100 = mode 4
101 = mode 5
X is a don’t care.
0
CNTFMT
Counter Format Status:
This bit indicates the counter’s programmed count format.
0 = binary (16 bits)
1 = binary coded decimal (4 decades)
D-66
SYSTEM REGISTER QUICK REFERENCE
D.68 UCSADH
See “CSnADH (UCSADH)” on page D-8.
D.69 UCSADL
See “CSnADL (UCSADL)” on page D-9.
D.70 UCSMSKH
See “CSnMSKH (UCSMSKH)” on page D-10.
D.71 UCSMSKL
See “CSnMSKL (UCSMSKL)” on page D-11.
D-67
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.72 WDTCNTH AND WDTCNTL
WDT Counter Value (High)
WDTCNTH
Expanded Addr: F4C4H
ISA Addr:
—
(read only)
Reset State:
003FH
15
8
0
WC31
WC23
WC30
WC22
WC29
WC21
WC28
WC20
WC27
WC26
WC18
WC25
WC24
WC16
7
WC19
WC17
WDT Counter Value (Low)
WDTCNTL
Expanded Addr: F4C6H
ISA Addr:
—
(read only)
Reset State:
FFFFH
15
8
0
WC15
WC7
WC14
WC6
WC13
WC5
WC12
WC4
WC11
WC10
WC2
WC9
WC1
WC8
WC0
7
WC3
Bit
Mnemonic
Bit Number
Function
WDT Counter Value High Word and Low Word:
High 15–0
Low 15–0
WC31:16
WC15:0
Read the high word of the counter value from WDTCNTH and the low
word from WDTCNTL.
D-68
SYSTEM REGISTER QUICK REFERENCE
Expanded Addr: F4C0H
D.73 WDTRLDH AND WDTRLDL
WDT Reload Value (High)
WDTRLDH
ISA Addr:
—
(read/write)
Reset State:
003FH
15
8
0
WR31
WR30
WR22
WR29
WR21
WR28
WR20
WR27
WR26
WR18
WR25
WR24
WR16
7
WR23
WR19
WR17
WDT Reload Value (Low)
WDTRLDL
Expanded Addr: F4C2H
ISA Addr:
—
(read/write)
Reset State:
FFFFH
15
8
0
WR15
WR14
WR6
WR13
WR5
WR12
WR4
WR11
WR10
WR2
WR9
WR1
WR8
WR0
7
WR7
WR3
Bit
Number
Bit
Mnemonic
Function
WDT Reload Value (High Word and Low Word):
High 15–0 WR31:16
Low 15–0 WR15:0
Write the high word of the reload value to WDTRLDH and the low word
to the WDTRLDL.
D-69
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
D.74 WDTSTATUS
WDT Status
WDTSTATUS
(read/write)
Expanded Addr: F4CAH
ISA Addr:
—
Reset State:
00H
7
0
WDTEN
—
—
—
—
—
BUSMON
CLKDIS
Bit
Number
Bit
Mnemonic
Function
7
WDTEN
Watchdog Mode Enabled:
This read-only bit indicates whether watchdog mode is enabled. Only a
lockout sequence can set this bit and only a device reset can clear it.
0 = Watchdog mode disabled
1 = Watchdog mode enabled
6–2
1
—
Reserved. These bits are undefined; for compatibility with future devices,
do not modify these bits.
BUSMON
Bus Monitor Enable:
0 = Disables bus monitor mode
1 = Enables bus monitor mode
Read this bit to determine the current status. A lockout sequence clears
BUSMON and prevents writes to the WDTSTATUS register.
0
CLKDIS
Clock Disable:
Write to this bit to stop or restart the clock to the WDT; read it to
determine the current clock status. A lockout sequence clears CLKDIS
and prevents writing to this register.
0 = Clock enabled
1 = Processor clock (frequency=CLK2/2) disabled (stopped)
.
D-70
E
INSTRUCTION SET
SUMMARY
APPENDIX E
This appendix provides reference information for the Intel386™ processor family instruction set.
The appendix is organized as follows:
• Instruction Encoding (page E-22)
E.1 INSTRUCTION ENCODING AND CLOCK COUNT SUMMARY
To calculate elapsed time for an instruction, multiply the instruction clock count, as listed in Table
E-1, by the processor clock period (e.g., 62.5 ns for 16 MHz).
Instruction clock count assumptions:
• The instruction has been prefetched, decoded, and is ready for execution.
• Bus cycles do not require wait states.
• There are no local bus HOLD requests delaying processor access to the bus.
• No exceptions are detected during instruction execution.
• When an effective address is calculated, it does not use two general register components.
One register, scaling and displacement can be used within the clock counts shown.
However, when the effective address calculation uses two general register components, add
1 clock to the clock count shown.
Instruction clock count notation:
• When two clock counts are given, the smaller refers to a register operand and the larger
refers to a memory operand.
• n = number of times repeated.
• m = number of components in the next instruction executed, where the entire displacement
(if any) counts as one component, the entire immediate data (if any) counts as one
component, and all other bytes of the instruction and prefix(es) of each count as one
component.
Misaligned or 32-bit operand accesses:
• When instructions access a misaligned 16-bit operand or 32-bit operand on even address:
— add 2* clocks for read or write
— add 4** clocks for read and write
• When instructions access a 32-bit operand on odd address:
— add 4* clocks for read or write
— add 8** clocks for read and write
E-1
Wait states:
Wait states add 1 clock per wait state to instruction execution for each data access.
Table E-1 lists the instructions with their formats and execution times. The description of the
notes for Table E-1 begins on page E-20. See “Instruction Encoding” on page E-22 for the defi-
nition of the terms used in this table.
Table E-1. Instruction Set Summary (Sheet 1 of 19)
Clock Count
Notes
Real
Ad-
Pro-
tected
Real
Ad-
Pro-
tected
dress Virtual dress Virtual
Instruction
Format
Mode
or
Ad-
dress
Mode
or
Ad-
dress
Virtual Mode Virtual Mode
8086
8086
Mode
Mode
GENERAL DATA TRANSFER
MOV = Move
register to
register/memory
2/2
2/4
2/2
2
2/2*
2/4*
2/2*
2
b
b
b
h
h
h
1 0 0 0 1 0 0 w
mod reg r/m
register/memory to
register
1 0 0 0 1 0 1 w
1 1 0 0 0 1 1 w
1 0 1 1 w reg
1 0 1 0 0 0 0 w
1 0 1 0 0 0 1 w
1 0 0 0 1 1 1 0
1 0 0 0 1 1 0 0
mod reg r/m
immediate to
register/memory
mod 0 0 0 r/m
immediate data
full displacement
full displacement
mod sreg3 r/m
mod sreg3 r/m
immediate data
immediate to register
(short form)
memory to accumulator
(short form)
4*
4*
b
b
b
b
h
h
accumulator to memory
(short form)
2*
2*
register memory to
segment register
2/5
2/2
22/23
2/2
h, i, j
h
segment register to
register/memory
MOVSX = Move with sign extension
register from
register/memory
3/6*
3/6*
3/6*
3/6*
b
b
h
h
0 0 0 0 1 1 1 1
1 0 1 1 1 1 1 w
mod reg r/m
mod reg r/m
MOVZX = Move with zero extension
register from
0 0 0 0 1 1 1 1
1 0 1 1 0 1 1 w
mod 1 1 0 r/m
register/memory
PUSH = Push
register/memory
register (short form)
1 1 1 1 1 1 1 1
0 1 0 1 0 reg
5/7*
2
7/9*
4
b
b
b
h
h
h
segment register (ES,
CS, SS, or DS)
(short form)
2
4
000 sreg2 110
0 0 0 0 1 1 1 1
segment register (ES,
CS, SS, or DS, FS or
GS)
2
4
b
h
10 sreg3 000
E-2
INSTRUCTION SET SUMMARY
Table E-1. Instruction Set Summary (Sheet 2 of 19)
Clock Count
Notes
Real
Ad-
Pro-
tected
Real
Ad-
Pro-
tected
dress Virtual dress Virtual
Instruction
Format
Mode
or
Ad-
dress
Mode
or
Ad-
dress
Virtual Mode Virtual Mode
8086
8086
Mode
Mode
immediate
0 1 1 0 1 0 s 0
0 1 1 0 0 0 0 0
immediate data
2
4
b
b
h
h
PUSHA = Push All
POP = Pop
18
34
register/memory
register (short form)
1 0 0 0 1 1 1 1
0 1 0 1 1 reg
mod 0 0 0 r/m
5/7
6
7/9
6
b
b
b
h
h
segment register (ES,
CS, SS, or DS)
(short form)
7
25
h, i, j
000 sreg2 111
segment register (ES,
CS, SS, or DS) FS or
GS
7
25
35
b
h, i, j
0 0 0 0 1 1 1 1
0 1 1 0 0 0 0 1
10 sreg3 001
mod reg r/m
POPA = Pop all
29
b
h
XCHG = Exchange
register/memory
with register
3/5**
3
3/5**
3
b, f
f, h
1 0 0 0 0 1 1 w
1 0 0 1 0 reg
register with accumula-
tor (short form)
Clk Count Virtual
8086 Mode
IN = Input from
fixed port
1 1 1 0 0 1 0 w
1 1 1 0 1 1 0 w
port number
†27
†28
14*
15*
8*/29*
9*/29*
s/t, m
s/t, m
variable port
OUT = Output to
fixed port
1 1 1 0 0 1 1 w
1 1 1 0 1 1 1 w
port number
mod reg r/m
†27
†28
14*
15*
2
8*/29*
9*/29*
2
s/t, m
s/t, m
variable port
LEA = Load EA to regis-
ter
1 0 0 0 1 1 0 1
SEGMENT CONTROL
LDS = Load pointer to
DS
7*
7*
7*
7*
7*
26*/28*
26*/28*
29*/31*
26*/28*
26*/28*
b
b
b
b
b
h, i, j
h, i, j
h, i, j
h, i, j
h, i, j
1 1 0 0 0 1 0 1
1 1 0 0 0 1 0 0
0 0 0 0 1 1 1 1
0 0 0 0 1 1 1 1
0 0 0 0 1 1 1 1
mod reg r/m
LES = Load pointer to
ES
mod reg r/m
LFS = Load pointer to
FS
1 0 1 1 0 1 0 0
1 0 1 1 0 1 0 1
1 0 1 1 0 0 1 0
mod reg r/m
LGS = Load pointer to
GS
mod reg r/m
mod reg r/m
LSS = Load pointer to
SS
FLAG CONTROL
CLC = Clear carry flag
1 1 1 1 1 0 0 0
2
2
E-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table E-1. Instruction Set Summary (Sheet 3 of 19)
Clock Count
Notes
Real
Ad-
Pro-
tected
Real
Ad-
Pro-
tected
dress Virtual dress Virtual
Instruction
Format
Mode
or
Ad-
dress
Mode
or
Ad-
dress
Virtual Mode Virtual Mode
8086
8086
Mode
Mode
CLD = Clear direction
flag
2
8
5
2
2
2
8
5
2
2
1 1 1 1 1 1 0 0
1 1 1 1 1 0 1 0
0 0 0 0 1 1 1 1
1 1 1 1 0 1 0 1
1 0 0 1 1 1 1 1
CLI = Clear interrupt
enable flag
m
l
CLTS = Clear task
switched flag
c
0 0 0 0 0 1 1 0
CMC = Complement
carry flag
LAHF = Load AH into
flag
POPF = Pop flags
1 0 0 1 1 1 0 1
1 0 0 1 1 1 0 0
5
4
3
5
4
3
b
b
h, n
h
PUSHF = Push flags
SAHF = Store AH into
flags
1 0 0 1 1 1 1 0
1 1 1 1 1 0 0 1
STC = Set carry flag
2
8
2
8
STD = Set direction flag 1 1 1 1 1 1 0 1
STI = Set interrupt
1 1 1 1 1 0 1 1
enable flag
m
ARITHMETIC INSTRUCTIONS
ADD = Add
register to register
register to memory
memory to register
0 0 0 0 0 0 d w
0 0 0 0 0 0 0 w
0 0 0 0 0 0 1 w
mod reg r/m
mod reg r/m
mod reg r/m
2
7**
2
7**
b
b
b
h
h
h
6*
6*
immediate to
register/memory
2/7**
2/7**
1 0 0 0 0 0 s w
0 0 0 0 0 1 0 w
mod 0 0 0 r/m
immediate data
immediate data
immediate to accumu-
lator (short form)
2
2
ADC = Add with carry
register to register
register to memory
memory to register
0 0 0 1 0 0 d w
0 0 0 1 0 0 0 w
0 0 0 1 0 0 1 w
mod reg r/m
mod reg r/m
mod reg r/m
2
7**
2
7**
b
b
b
h
h
h
6*
6*
immediate to
register/memory
2/7**
2/7**
1 0 0 0 0 0 s w
0 0 0 1 0 1 0 w
mod 0 1 0 r/m
immediate data
immediate data
immediate to accumu-
lator (short form)
2
2
INC = Increment
register/memory
register (short form)
1 1 1 1 1 1 1 w
0 1 0 0 0 reg
mod 0 0 0 r/m
2/6**
2
2/6**
2
b
h
E-4
INSTRUCTION SET SUMMARY
Table E-1. Instruction Set Summary (Sheet 4 of 19)
Clock Count
Notes
Real
Ad-
Pro-
tected
Real
Ad-
Pro-
tected
dress Virtual dress Virtual
Instruction
Format
Mode
or
Ad-
dress
Mode
or
Ad-
dress
Virtual Mode Virtual Mode
8086
8086
Mode
Mode
SUB = Subtract
register from register
register from memory
memory from register
0 0 1 0 1 0 d w
0 0 1 0 1 0 0 w
0 0 1 0 1 0 1 w
mod reg r/m
2
7**
2
7**
mod reg r/m
mod reg r/m
b
b
b
h
h
h
6*
6*
immediate from
register/memory
2/7**
2/7**
1 0 0 0 0 0 s w
0 0 1 0 1 1 0 w
mod 101 r/m
immediate data
immediate from accu-
mulator (short form)
2
2
immediate data
SBB = Subtract with borrow
register from register
register from memory
memory from register
0 0 0 1 1 0 d w
0 0 0 1 1 0 0 w
0 0 0 1 1 0 1 w
mod reg r/m
mod reg r/m
mod reg r/m
2
7**
2
7**
b
b
b
h
h
h
6*
6*
immediate from
register/memory
2/7**
2/7**
1 0 0 0 0 0 s w
0 0 0 1 1 1 0 w
mod 0 1 1 r/m
immediate data
immediate data
immediate from accu-
mulator (short form)
2
2
DEC = Decrement
register/memory
1 1 1 1 1 1 1 w
0 1 0 0 1 reg
reg 0 0 1 r/m
2/6
2
2/6
2
b
h
register (short form)
CMP = Compare
register with register
memory with register
register with memory
0 0 1 1 1 0 d w
0 0 1 1 1 0 0 w
0 0 1 1 1 0 1 w
mod reg r/m
mod reg r/m
mod reg r/m
2
5*
2
5*
b
b
b
h
h
h
6*
6*
immediate with
register/memory
immediate
data
2/5*
2/5*
1 0 0 0 0 0 s w
mod 1 1 1 r/m
immediate with accu-
mulator (short form)
immediate
data
2
2
0 0 1 1 1 1 0 w
1 1 1 1 0 1 1 w
0 0 1 1 0 1 1 1
NEG = Change sign
mod 0 1 1 r/m
2/6*
4
2/6*
4
b
h
AAA = ASCII adjust for
addition
AAS = ASCII adjust for
subtraction
4
4
4
4
4
4
0 0 1 1 1 1 1 1
0 0 1 0 0 1 1 1
0 0 1 0 1 1 1 1
DAA = Decimal adjust
for addition
DAS = Decimal adjust for
subtraction
E-5
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table E-1. Instruction Set Summary (Sheet 5 of 19)
Clock Count
Notes
Real
Ad-
Pro-
tected
Real
Ad-
Pro-
tected
dress Virtual dress Virtual
Instruction
Format
Mode
or
Ad-
dress
Mode
or
Ad-
dress
Virtual Mode Virtual Mode
8086
8086
Mode
Mode
MUL = multiply (unsigned)
accumulator with
register/memory
1 1 1 1 0 1 1 w
mod 1 0 0 r/m
multiplier
— byte
12-17/
15-20*
12-25/
15-28*
12-41/
17-46*
12-17/
15-20*
12-25/
15-28*
12-41/
17-46*
b, d
b, d
b, d
d, h
d, h
d, h
— word
— doubleword
IMUL = Integer multiply (signed)
accumulator with regis-
ter/memory
1 1 1 1 0 1 1 w
mod 1 0 1 r/m
multiplier
— byte
12-17/
15-20*
12-25/
15-28*
12-41/
17-46*
12-17/
15-20*
12-25/
15-28*
12-41/
17-46*
b, d
b, d
b, d
d, h
d, h
d, h
— word
— doubleword
register with regis-
ter/memory
0 0 0 0 1 1 1 1
1 0 1 0 1 1 1 1
mod reg r/m
multiplier
— byte
12-17/
15-20*
12-25/
15-28*
12-41/
17-46*
12-17/
15-20*
12-25/
15-28*
12-41/
17-46*
b, d
b, d
b, d
d, h
d, h
d, h
— word
— doubleword
register/memory with
immediate to register
0 1 1 0 1 0 s 1
mod reg r/m
immediate data
— word
13-26
13-42
13-26/
14-27
13-42/
16-45
b, d
b, d
d, h
d, h
— doubleword
DIV = Divide (unsigned)
Accumulator by
register/memory
1 1 1 1 0 1 1 w
mod 1 1 0 r/m
divisor
— byte
— word
— doubleword
14/17
22/25
38/43
14/17
22/25
38/43
b, e
b, e
b, e
e, h
e, h
e, h
E-6
INSTRUCTION SET SUMMARY
Table E-1. Instruction Set Summary (Sheet 6 of 19)
Clock Count
Notes
Real
Ad-
Pro-
tected
Real
Ad-
Pro-
tected
dress Virtual dress Virtual
Instruction
Format
Mode
or
Ad-
dress
Mode
or
Ad-
dress
Virtual Mode Virtual Mode
8086
8086
Mode
Mode
IDIV = Integer divide (signed)
Accumulator by
register/memory
1 1 1 1 0 1 1 w
mod 111 r/m
divisor
— byte
— word
— doubleword
19/22
27/30
43/48
19/22
27/30
43/48
b, e
b, e
b, e
e, h
e, h
e, h
AAD = ASCII adjust for
divide
19
17
3
19
17
3
1 1 0 1 0 1 0 1
1 1 0 1 0 1 0 0
1 0 0 1 1 0 0 0
1 0 0 1 1 0 0 1
0 0 0 0 1 0 1 0
0 0 0 0 1 0 1 0
AAM = ASCII adjust for
multiply
CBW = Convert byte to
word
CWD = Convert word to
double-word
2
2
LOGIC
shift rotate instructions
not through carry (ROL, ROR, SAL, SAR, SHL, and SHR)
register/memory by 1
register/memory by CL
1 1 0 1 0 0 0 w
1 1 0 1 0 0 1 w
mod TTT r/m
mod TTT r/m
3/7**
3/7*
3/7*
3/7**
3/7*
3/7*
b
b
b
h
h
h
register/memory by
immediate count
1 1 0 0 0 0 0 w
mod TTT r/m
immed 8-bit data
immed 8-bit data
through carry (RCL and RCR)
register/memory by 1
register/memory by CL
1 1 0 1 0 0 0 w
mod TTT r/m
mod TTT r/m
9/10*
9/10*
9/10*
9/10*
9/10*
9/10*
b
b
b
h
h
h
1 1 0 1 0 0 1 w
1 1 0 0 0 0 0 w
register/memory by
immediate count
mod TTT r/m
T T T
0 0 0
0 0 1
0 1 0
0 1 1
1 0 0
1 0 1
1 1 1
Instruction
ROL
ROR
RCL
RCR
SHL/SAL
SHR
SAR
SHLD = Shift left double
register/memory by
immediate
immed
8-bit data
3/7**
3/7**
3/7**
3/7**
0 0 0 0 1 1 1 1
0 0 0 0 1 1 1 1
1 0 1 0 0 1 0 0
1 0 1 0 0 1 0 1
mod reg r/m
mod reg r/m
register/memory by CL
E-7
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table E-1. Instruction Set Summary (Sheet 7 of 19)
Clock Count
Notes
Real
Ad-
Pro-
tected
Real
Ad-
Pro-
tected
dress Virtual dress Virtual
Instruction
Format
Mode
or
Ad-
dress
Mode
or
Ad-
dress
Virtual Mode Virtual Mode
8086
8086
Mode
Mode
SHRD = Shift right double
register/memory by
immediate
immed
8-bit data
3/7**
3/7**
3/7**
3/7**
0 0 0 0 1 1 1 1
0 0 0 0 1 1 1 1
1 0 1 0 1 1 0 0
mod reg r/m
mod reg r/m
register/memory by CL
AND = And
1 0 1 0 1 1 0 1
register to register
register to memory
memory to register
0 0 1 0 0 0 d w
0 0 1 0 0 0 0 w
0 0 1 0 0 0 1 w
mod reg r/m
mod reg r/m
mod reg r/m
2
2
7**
6*
7**
6*
b
b
b
h
h
h
immediate to
register/memory
2/7*
2/7*
1 0 0 0 0 0 0 w
0 0 1 0 0 1 0 w
mod 1 0 0 r/m
immediate data
immediate data
immediate to accumu-
lator (short form)
2
2
TEST= And function to flags, no result
register/memory
and register
2/5*
2/5*
2
2/5*
2/5*
2
b
b
h
h
1 0 0 0 0 1 0 w
mod reg r/m
immediate data and
register/memory
1 1 1 1 0 1 1 w
1 0 1 0 1 0 0 w
mod 0 0 0 r/m
immediate data
immediate data and
accumulator (short
form)
immediate
data
OR = Or
register to register
register to memory
memory to register
0 0 0 0 1 0 d w
0 0 0 0 1 0 0 w
0 0 0 0 1 0 1 w
mod reg r/m
mod reg r/m
mod reg r/m
2
7**
2
7**
b
b
b
h
h
h
6*
6*
immediate to
register/memory
2/7**
2/7**
1 0 0 0 0 0 0 w
0 0 0 0 1 1 0 w
mod 0 0 1 r/m
immediate data
immediate data
immediate to accumu-
lator (short form)
2
2
XOR = Exclusive or
register to register
register to memory
memory to register
0 0 1 1 0 0 d w
0 0 1 1 0 0 0 w
0 0 1 1 0 0 1 w
mod reg r/m
mod reg r/m
mod reg r/m
2
7**
2
7**
b
b
b
h
h
h
6*
6*
immediate to
register/memory
2/7**
2/7**
1 0 0 0 0 0 0 w
0 0 1 1 0 1 0 w
1 1 1 1 0 1 1 w
mod 1 1 0 r/m
immediate data
mod 0 1 0 r/m
immediate data
immediate to accumu-
lator (short form)
2
2
NOT= Invert
register/memory
2/6**
2/6**
b
h
E-8
INSTRUCTION SET SUMMARY
Table E-1. Instruction Set Summary (Sheet 8 of 19)
Clock Count
Notes
Real
Ad-
Pro-
tected
Real
Ad-
Pro-
tected
dress Virtual dress Virtual
Instruction
Format
Mode
or
Ad-
dress
Mode
or
Ad-
dress
Virtual Mode Virtual Mode
8086
8086
Mode
Mode
STRING MANIPULATION INSTRUCTIONS
Clk Count
Virtual
8086 Mode
CMPS = Compare
1 0 1 0 0 1 1 w
byte word
10*
17
5
10*
10*/32**
5*
b
b
b
h
s/t, h, m
h
INS = Input byte/word
0 1 1 0 1 1 0 w
†30
from DX port
LODS = Load byte/word
1 0 1 0 1 1 0 w
to AL/AX/EAX
MOVS = Move byte word 1 0 1 0 0 1 0 w
7
7**
b
b
h
OUTS = Output
0 1 1 0 1 1 1 w
†31
18
11*/33*
s/t, h, m
byte/word to DX port
SCAS = Scan byte word
1 0 1 0 1 1 1 w
1 0 1 0 1 0 1 w
1 1 0 1 0 1 1 1
7*
4*
7*
4*
b
b
h
h
STOS = Store byte/word
from AL/AX/EX
XLAT = Translate String
5*
5*
h
REPEATED STRING MANIPULATION
Repeated by count in CX or ECX:
REPE CMPS = Compare string
find non-match
find match
1 1 1 1 0 0 1 1
1 1 1 1 0 0 1 0
1 0 1 0 0 1 1 w
Clk Count
Virtual
8086 Mode
5 + 9n**
5 + 9n**
5 + 9n**
5 + 9n**
b
b
h
h
1 0 1 0 0 1 1 w
†31+6n
17 + 7n*
11 + 7n*/
32+ 6n*
b
s/t, h, m
REP INS = Input string
1 1 1 1 0 0 1 0
0 1 1 0 1 1 0 w
1 0 1 0 1 1 0 w
1 0 1 0 0 1 0 w
REP LODS = Load string 1 1 1 1 0 0 1 0
5 + 6n*
7 + 4n*
5 + 6n*
7 + 4n**
b
b
h
h
REP MOVS = Move
1 1 1 1 0 0 1 0
string
REP OUTS = Output
1 1 1 1 0 0 1 0
string
†30+8n
16 + 8n* 10 + 8n*/
31+ 8n*
b
b
s/t, h, m
h
0 1 1 0 1 1 1 w
1 0 1 0 1 1 1 w
REPE SCAS = Scan
5 + 8n*
5 + 8n*
5 + 5n*
5 + 8n*
5 + 8n*
5 + 5n*
string
1 1 1 1 0 0 1 1
Find non-AL/AX/EAX
REPNE SCAS = Scan
string
Find AL/AX/EAX
b
b
h
h
1 1 1 1 0 0 1 0
1 1 1 1 0 0 1 0
1 0 1 0 1 1 1 w
1 0 1 0 1 0 1 w
REP STOS = Store
string
BIT MANIPULATION
BSF = scan bit forward
BSR = scan bit reverse
BT = test bit
0 0 0 0 1 1 1 1
0 0 0 0 1 1 1 1
1 0 1 1 1 1 0 0
1 0 1 1 1 1 0 1
mod reg r/m
mod reg r/m
10 + 3n*
10 + 3n*
10 + 3n*
10 + 3n*
b
b
h
h
register/memory,
immediate
immed
8-bit data
3/6*
3/6*
b
h
0 0 0 0 1 1 1 1
1 0 1 1 1 0 1 0
mod 1 0 0 r/m
E-9
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table E-1. Instruction Set Summary (Sheet 9 of 19)
Clock Count
Notes
Real
Ad-
Pro-
tected
Real
Ad-
Pro-
tected
dress Virtual dress Virtual
Instruction
Format
Mode
or
Ad-
dress
Mode
or
Ad-
dress
Virtual Mode Virtual Mode
8086
8086
Mode
Mode
register/memory,
register
3/12*
3/12*
b
h
0 0 0 0 1 1 1 1
1 0 1 0 0 0 1 1
mod reg r/m
BTC = test bit and complement
register/memory,
immediate
immed
8-bit data
6/8*
6/8*
b
b
h
h
0 0 0 0 1 1 1 1
1 0 1 1 1 0 1 0
1 0 1 1 1 0 1 1
mod 1 1 1 r/m
mod reg r/m
register/memory,
register
6/13*
6/13*
0 0 0 0 1 1 1 1
BTR = test bit and reset
register/memory,
immediate
immed
8-bit data
6/8*
6/8*
b
b
h
h
0 0 0 0 1 1 1 1
0 0 0 0 1 1 1 1
1 0 1 1 1 0 1 0
1 0 1 1 0 0 1 1
mod 1 1 0 r/m
mod reg r/m
register/memory,
register
6/13*
6/13*
BTS = test bit and set
register/memory,
immediate
immed
8-bit data
6/8*
6/8*
b
b
h
h
0 0 0 0 1 1 1 1
0 0 0 0 1 1 1 1
1 0 1 1 1 0 1 0
1 0 1 0 1 0 1 1
mod 1 0 1r/m
mod reg r/m
register/memory,
register
6/13*
6/13*
CONTROL TRANSFER
CALL = Call
direct within segment
1 1 1 0 1 0 0 0
1 1 1 1 1 1 1 1
1 0 0 1 1 0 1 0
full displacement
mod 0 1 0 r/m
7 + m*
9 + m*
b
b
r
reg/memory indirect
within segment
7 + m*/
10 + m*
9 + m*/
12 + m*
h, r
direct intersegment
unsigned full offset, selector
17 + m*
42 + m*
b
j, k, r
Protected mode only (direct intersegment)
Via call gate to same privilege level
64 + m
98 + m
h, j, k, r
h, j, k, r
h, j, k, r
Via call gate to different privilege level (no parameters)
Via call gate to different privilege level (x parameters)
106+8x+
m
From 286 task to 286 TSS
285
310
h, j, k, r
h, j, k, r
h, j, k, r
h, j, k, r
h, j, k, r
h, j, k, r
h, j, k, r
From 286 task to Intel386 SX CPU TSS
From 286 task to virtual 8086 task (Intel386 SX CPU TSS)
From Intel386 SX CPU task to 286 TSS
229
285
From Intel386 SX CPU task to Intel386 SX CPU TSS
From Intel386 SX CPU task to Virtual 8086 task (Intel386 SX CPU TSS)
392
309
indirect intersegment
1 1 1 1 1 1 1 1
mod 0 1 1 r/m
30 + m
46 + m
b
Protected mode only (indirect intersegment)
Via call gate to same privilege level
68 + m
h, j, k, r
E-10
INSTRUCTION SET SUMMARY
Table E-1. Instruction Set Summary (Sheet 10 of 19)
Clock Count
Notes
Real
Ad-
Pro-
tected
Real
Ad-
Pro-
tected
dress Virtual dress Virtual
Instruction
Format
Mode
or
Ad-
dress
Mode
or
Ad-
dress
Virtual Mode Virtual Mode
8086
8086
Mode
Mode
Via call gate to different privilege level (no parameters)
Via call gate to different privilege level (x parameters)
102 + m
h, j, k, r
h, j, k, r
110+8x+
m
From 286 task to 286 TSS
h, j, k, r
h, j, k, r
h, j, k, r
h, j, k, r
h, j, k, r
h, j, k, r
From 286 task to Intel386 SX CPU TSS
From 286 task to virtual 8086 task (Intel386 SX CPU TSS)
From Intel386 SX CPU task to 286 TSS
From Intel386 SX CPU task to Intel386 SX CPU TSS
From Intel386 SX CPU task to Virtual 8086 task (Intel386 SX CPU TSS)
JMP = Unconditional jump
399
short
1 1 1 0 1 0 1 1
1 1 1 0 1 0 0 1
8-bit displacement
full displacement
7 + m
7 + m
7 + m
7 + m
r
r
direct within segment
reg/memory indirect
within segment
9 + m/
14 + m
9 + m/
14 + m
b
h, r
1 1 1 1 1 1 1 1
1 1 1 0 1 0 1 0
mod 1 0 0 r/m
direct intersegment
unsigned full offset, selector
16 + m
31 + m
53 + m
j, k, r
Protected mode only (direct intersegment)
Via call gate to same privilege level
From 286 task to 286 TSS
h, j, k, r
h, j, k, r
h, j, k, r
h, j, k, r
h, j, k, r
h, j, k, r
h, j, k, r
h, j, k, r
From 286 task to Intel386 SX CPU TSS
From 286 task to virtual 8086 task (Intel386 SX CPU TSS)
From Intel386 SX CPU task to 286 TSS
From Intel386 SX CPU task to Intel386 SX CPU TSS
From Intel386 SX CPU task to Virtual 8086 task (Intel386 SX CPU TSS)
395
indirect intersegment
1 1 1 1 1 1 1 1
mod 1 0 1 r/m
17 + m
31 + m
b
Protected mode only (indirect intersegment)
Via call gate to same privilege level
From 286 task to 286 TSS
49 + m
h, j, k, r
h, j, k, r
h, j, k, r
h, j, k, r
h, j, k, r
h, j, k, r
h, j, k, r
From 286 task to Intel386 SX CPU TSS
From 286 task to virtual 8086 task (Intel386 SX CPU TSS)
From Intel386 SX CPU task to 286 TSS
From Intel386 SX CPU task to Intel386 SX CPU TSS
From Intel386 SX CPU task to Virtual 8086 task (Intel386 SX CPU TSS)
RET = Return from CALL
328
within segment
1 1 0 0 0 0 1 1
12 + m
b
g, h, r
E-11
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table E-1. Instruction Set Summary (Sheet 11 of 19)
Clock Count
Notes
Real
Ad-
Pro-
tected
Real
Ad-
Pro-
tected
dress Virtual dress Virtual
Instruction
Format
Mode
or
Ad-
dress
Mode
or
Ad-
dress
Virtual Mode Virtual Mode
8086
8086
Mode
Mode
within segment adding
immed to SP
12 + m
36 + m
36 + m
b
b
b
g, h, r
1 1 0 0 0 0 1 0
1 1 0 0 1 0 1 1
1 1 0 0 1 0 1 0
16-bit displacement
16-bit displacement
intersegment
g, h, j, k,
r
intersegment adding
immed to SP
g, h, j, k,
r
Protected mode only (RET):
to different privilege level
Intersegment
72
72
h, j, k, r
h, j, k, r
Intersegment adding immed to SP
CONDITIONAL JUMPS
(times are jump “Taken or not Taken”)
JO = jump on overflow
8-bit displacement
0 1 1 1 0 0 0 0
7 + m
or 3
7 + m
or 3
r
r
8-bit displacement
Full displacement
0 0 0 0 1 1 1 1
7 + m
or 3
7 + m
or 3
1 0 0 0 0 0 0 0
Full displacement
JNO = Jump on not overflow
8-bit displacement
0 1 1 1 0 0 0 1
7 + m
or 3
7 + m
or 3
r
r
8-bit displacement
1 0 0 0 0 0 0 1
Full displacement
0 0 0 0 1 1 1 1
7 + m
or 3
7 + m
or 3
Full displacement
Full displacement
Full displacement
Full displacement
JB/JNAE = jump on below/not above or equal
8-bit displacement
0 1 1 1 0 0 1 0
7 + m
or 3
7 + m
or 3
r
r
8-bit displacement
1 0 0 0 0 0 1 0
Full displacement
0 0 0 0 1 1 1 1
7 + m
or 3
7 + m
or 3
JNB/JAE = jump on not below/above or equal
8-bit displacement
0 1 1 1 0 0 1 1
7 + m
or 3
7 + m
or 3
r
r
8-bit displacement
1 0 0 0 0 0 1 1
Full displacement
0 0 0 0 1 1 1 1
7 + m
or 3
7 + m
or 3
JE/JZ= jump on equal/zero
8-bit displacement
0 1 1 1 0 1 0 0
7 + m
or 3
7 + m
or 3
r
r
8-bit displacement
1 0 0 0 0 1 0 0
Full displacement
0 0 0 0 1 1 1 1
7 + m
or 3
7 + m
or 3
JNE/JNZ = jump on not equal/not zero
8-bit displacement
0 1 1 1 0 1 0 1
7 + m
or 3
7 + m
or 3
r
8-bit displacement
E-12
INSTRUCTION SET SUMMARY
Table E-1. Instruction Set Summary (Sheet 12 of 19)
Clock Count
Notes
Real
Ad-
Pro-
tected
Real
Ad-
Pro-
tected
dress Virtual dress Virtual
Instruction
Format
Mode
or
Ad-
dress
Mode
or
Ad-
dress
Virtual Mode Virtual Mode
8086
8086
Mode
Mode
Full displacement
7 + m
or 3
7 + m
or 3
r
0 0 0 0 1 1 1 1
1 0 0 0 0 1 0 1
Full displacement
JBE/JNA = jump on below or equal/not above
8-bit displacement
0 1 1 1 0 1 1 0
7 + m
or 3
7 + m
or 3
r
r
8-bit displacement
1 0 0 0 0 1 1 0
Full displacement
0 0 0 0 1 1 1 1
7 + m
or 3
7 + m
or 3
Full displacement
Full displacement
Full displacement
Full displacement
Full displacement
Full displacement
JNBE/JA = jump on not below or equal/above
8-bit displacement
0 1 1 1 0 1 1 1
7 + m or
3
7 + m or
3
r
r
8-bit displacement
1 0 0 0 0 1 1 1
Full displacement
0 0 0 0 1 1 1 1
7 + m or
3
7 + m or
3
JS = jump on sign
8-bit displacement
0 1 1 1 1 0 0 0
7 + m or
3
7 + m or
3
r
r
8-bit displacement
1 0 0 0 1 0 0 0
Full displacement
0 0 0 0 1 1 1 1
7 + m or
3
7 + m or
3
JNS = jump on not sign
8-bit displacement
0 1 1 1 1 0 0 1
7 + m or
3
7 + m or
3
r
r
8-bit displacement
1 0 0 0 1 0 0 1
Full displacement
0 0 0 0 1 1 1 1
7 + m or
3
7 + m or
3
JP/JPE = jump on parity/parity even
8-bit displacement
0 1 1 1 1 0 1 0
7 + m or
3
7 + m or
3
r
r
8-bit displacement
1 0 0 0 1 0 1 0
Full displacement
0 0 0 0 1 1 1 1
7 + m or
3
7 + m or
3
JNP/JPO = jump on not parity/parity odd
8-bit displacement
0 1 1 1 1 0 1 1
7 + m or
3
7 + m or
3
r
r
8-bit displacement
1 0 0 0 1 0 1 1
Full displacement
0 0 0 0 1 1 1 1
7 + m or
3
7 + m or
3
JL/JNGE = jump on less/not greater or
equal
8-bit displacement
0 1 1 1 1 1 0 0
7 + m or
3
7 + m or
3
r
r
8-bit displacement
1 0 0 0 1 1 0 0
Full displacement
0 0 0 0 1 1 1 1
7 + m or
3
7 + m or
3
Full displacement
JNL/JGE = jump on not less/greater or equal
8-bit displacement
0 1 1 1 1 1 0 1
7 + m or
3
7 + m or
3
r
8-bit displacement
E-13
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table E-1. Instruction Set Summary (Sheet 13 of 19)
Clock Count
Notes
Real
Ad-
Pro-
tected
Real
Ad-
Pro-
tected
dress Virtual dress Virtual
Instruction
Format
Mode
or
Ad-
dress
Mode
or
Ad-
dress
Virtual Mode Virtual Mode
8086
8086
Mode
Mode
Full displacement
7 + m or
3
7 + m or
3
r
0 0 0 0 1 1 1 1
1 0 0 0 1 1 0 1
Full displacement
JLE/JNG = jump on less or equal/not greater
8-bit displacement
0 1 1 1 1 1 1 0
7 + m or
3
7 + m or
3
r
r
8-bit displacement
1 0 0 0 1 1 1 0
Full displacement
0 0 0 0 1 1 1 1
7 + m or
3
7 + m or
3
Full displacement
JNLE/JG = jump on not less or equal/greater
8-bit displacement
0 1 1 1 1 1 1 1
7 + m or
3
7 + m or
3
r
r
r
r
8-bit displacement
1 0 0 0 1 1 1 1
Full displacement
0 0 0 0 1 1 1 1
7 + m or
3
7 + m or
3
Full displacement
9 + m or
5
9 + m or
5
JCXZ = jump on CX zero 1 1 1 0 0 0 1 1
8-bit displacement
8-bit displacement
JECXZ = jump on ECX
1 1 1 0 0 0 1 1
zero
9 + m or
5
9 + m or
5
(Address size prefix differentiates JCXZ from JECXZ)
LOOP = loop CX times
1 1 1 0 0 0 1 0
1 1 1 0 0 0 0 1
8-bit displacement
8-bit displacement
11 + m
11 + m
11 + m
11 + m
r
r
LOOPZ/LOOPE = loop
with zero/equal
LOOPNZ/LOOPNE =
loop while not zero
11 + m
11 + m
r
1 1 1 0 0 0 0 0
8-bit displacement
CONDITIONAL BYTE SET
(Note: Times are register/memory)
SETO = set byte on overflow
to register/memory
0 0 0 0 1 1 1 1
1 0 0 1 0 0 0 0
1 0 0 1 0 0 0 1
mod 0 0 0 r/m
mod 0 0 0 r/m
mod 0 0 0 r/m
mod 0 0 0 r/m
mod 0 0 0 r/m
mod 0 0 0 r/m
4/5*
4/5*
4/5*
4/5*
4/5*
4/5*
4/5*
4/5*
4/5*
4/5*
4/5*
4/5*
h
h
h
h
h
h
SETNO = set byte on not overflow
to register/memory
0 0 0 0 1 1 1 1
SETB/SETNAE = set byte on below/not above or equal
to register/memory
SETNB = set byte on below/above or equal
to register/memory 0 0 0 0 1 1 1 1
SETE/SETZ = set byte on equal/zero
to register/memory 0 0 0 0 1 1 1 1
SETNE/SETNZ = set byte on not equal/not zero
to register/memory 0 0 0 0 1 1 1 1 1 0 0 1 0 1 0 1
0 0 0 0 1 1 1 1
1 0 0 1 0 0 1 0
1 0 0 1 0 0 1 1
1 0 0 1 0 1 0 0
E-14
INSTRUCTION SET SUMMARY
Table E-1. Instruction Set Summary (Sheet 14 of 19)
Clock Count
Notes
Real
Ad-
Pro-
tected
Real
Ad-
Pro-
tected
dress Virtual dress Virtual
Instruction
Format
Mode
or
Ad-
dress
Mode
or
Ad-
dress
Virtual Mode Virtual Mode
8086
8086
Mode
Mode
SETBE/SETNA = set byte on below or equal/not above
to register/memory
0 0 0 0 1 1 1 1
1 0 0 1 0 1 1 0
mod 0 0 0 r/m
4/5*
4/5*
4/5*
4/5*
4/5*
4/5*
4/5*
4/5*
h
h
h
h
SETNBE/SETA = set byte on not below or equal/above
to register/memory
SETS = set byte on sign
to register/memory
0 0 0 0 1 1 1 1
0 0 0 0 1 1 1 1
1 0 0 1 0 1 1 1
1 0 0 1 1 0 0 0
1 0 0 1 1 0 0 1
mod 0 0 0 r/m
mod 0 0 0 r/m
mod 0 0 0 r/m
SETNS = set byte on not sign
to register/memory
0 0 0 0 1 1 1 1
SETP/SETPE = set byte on parity/parity
even
to register/memory
0 0 0 0 1 1 1 1
1 0 0 1 1 01 0
mod 0 0 0 r/m
mod 0 0 0 r/m
mod 0 0 0 r/m
mod 0 0 0 r/m
mod 0 0 0 r/m
mod 0 0 0 r/m
4/5*
4/5*
4/5*
4/5*
4/5*
4/5*
4/5*
4/5*
4/5*
4/5*
4/5*
4/5*
h
h
h
h
h
h
SETNP/SETPO = set byte on not parity/parity odd
to register/memory 0 0 0 0 1 1 1 1 1 0 0 1 1 0 1 1
SETL/SETNGE = set byte on less/not greater or equal
to register/memory 0 0 0 0 1 1 1 1 1 0 0 1 1 1 0 0
SETNL/SETGE = set byte on not less/greater or equal
to register/memory 0 0 0 0 1 1 1 1 0 1 1 1 1 1 0 1
SETLE/SETNG = set byte on less or equal/not greater
to register/memory 0 0 0 0 1 1 1 1 1 0 0 1 1 1 1 0
SETNLE/SETG = set byte on not less or equal/greater
to register/memory
0 0 0 0 1 1 1 1
1 1 0 0 1 0 0 0
1 0 0 1 1 1 1 1
16-bit displacement, 8-bit level
ENTER = enter proce-
dure
L = 0
L = 1
L > 1
10
14
17+8(n -
1)
10
14
17+8(n -
1)
b
b
h
h
LEAVE = leave proce-
dure
4
4
b
h
1 1 0 0 1 0 0 1
INTERRUPT INSTRUCTIONS
INT = Interrupt:
Type specified
Type 3
1 1 0 0 1 1 0 1
type
37
33
b
b
1 1 0 0 1 1 0 0
1 1 0 0 1 1 1 0
INTO = Interrupt 4 if
overflow flag set
If OF = 1
If OF = 0
35
3
b, e
b, e
3
E-15
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table E-1. Instruction Set Summary (Sheet 15 of 19)
Clock Count
Notes
Real
Ad-
Pro-
tected
Real
Ad-
Pro-
tected
dress Virtual dress Virtual
Instruction
Format
Mode
or
Ad-
dress
Mode
or
Ad-
dress
Virtual Mode Virtual Mode
8086
8086
Mode
Mode
BOUND = Interrupt 5 if
Detect value out of range
0 1 1 0 0 0 1 0
mod reg r/m
If out of range
If in range
44
10
b, e
b, e
e, g, h, j,
k, r
10
e, g, h, j,
k, r
Protected Mode Only (INT)
INT: Type Specified
Via interrupt or Trap Gate
To same privilege level
71
g, j, k, r
g, j, k, r
Via Interrupt or Trap Gate
111
To different privilege level
From 286 task to 286 TSS via Task Gate
438
465
382
440
467
384
445
472
275
g, j, k, r
g, j, k, r
g, j, k, r
g, j, k, r
g, j, k, r
g, j, k, r
g, j, k, r
g, j, k, r
g, j, k, r
From 286 task to Intel 386 SX CPU TSS via Task Gate
From 286 task to virtual 8086 mode via Task Gate
From Intel386 SX CPU task to 286 TSS via Task Gate
From Intel386 SX CPU task to Intel 386 SX CPU TSS via task gate
From Intel386 SX CPU task to virtual 8086 mode via Task Gate
From virtual 8086 mode to 286 TSS via Task Gate
From virtual 8086 mode to Intel 386 SX CPU TSS via task gate
From virtual 8086 mode to privilege level 0 via trap gate or interrupt gate
INT: TYPE 3
Via interrupt or Trap Gate
To same privilege level
71
g, j, k, r
g, j, k, r
Via Interrupt or Trap Gate
To different privilege level
111
From 286 task to 286 TSS via Task Gate
382
409
326
384
411
328
389
416
223
g, j, k, r
g, j, k, r
g, j, k, r
g, j, k, r
g, j, k, r
g, j, k, r
g, j, k, r
g, j, k, r
From 286 task to Intel 386 SX CPU TSS via Task Gate
From 286 task to virtual 8086 mode via Task Gate
From Intel386 SX CPU task to 286 TSS via Task Gate
From Intel386 SX CPU task to Intel 386 SX CPU TSS via task gate
From Intel386 SX CPU task to virtual 8086 mode via Task Gate
From virtual 8086 mode to 286 TSS via Task Gate
From virtual 8086 mode to Intel 386 SX CPU TSS via task gate
From virtual 8086 mode to privilege level 0 via trap gate or interrupt gate
E-16
INSTRUCTION SET SUMMARY
Table E-1. Instruction Set Summary (Sheet 16 of 19)
Clock Count
Notes
Real
Ad-
Pro-
tected
Real
Ad-
Pro-
tected
dress Virtual dress Virtual
Instruction
Format
Mode
or
Ad-
dress
Mode
or
Ad-
dress
Virtual Mode Virtual Mode
8086
8086
Mode
Mode
INTO:
Via interrupt or Trap Gate
To same privilege level
71
g, j, k, r
g, j, k, r
Via Interrupt or Trap Gate
To different privilege level
111
From 286 task to 286 TSS via Task Gate
384
411
328
g, j, k, r
g, j, k, r
g, j, k, r
g, j, k, r
From 286 task to Intel 386 SX CPU TSS via Task Gate
From 286 task to virtual 8086 mode via Task Gate
From Intel386 SX CPU task to 286 TSS via Task Gate
Intel386
DX
From Intel386 SX CPU task to Intel 386 SX CPU TSS via task gate
From Intel386 SX CPU task to virtual 8086 mode via Task Gate
From virtual 8086 mode to 286 TSS via Task Gate
From virtual 8086 mode to Intel 386 SX CPU TSS via task gate
From virtual 8086 mode to privilege level 0 via trap gate or interrupt gate
BOUND:
413
329
391
418
223
g, j, k, r
g, j, k, r
g, j, k, r
g, j, k, r
Via interrupt or Trap Gate
To same privilege level
71
g, j, k, r
g, j, k, r
Via Interrupt or Trap Gate
To different privilege level
111
From 286 task to 286 TSS via Task Gate
358
388
335
368
398
347
368
398
223
g, j, k, r
g, j, k, r
g, j, k, r
g, j, k, r
g, j, k, r
g, j, k, r
g, j, k, r
g, j, k, r
From 286 task to Intel 386 SX CPU TSS via Task Gate
From 286 task to virtual 8086 mode via Task Gate
From Intel386 SX CPU task to 286 TSS via Task Gate
From Intel386 SX CPU task to Intel 386 SX CPU TSS via task gate
From Intel386 SX CPU task to virtual 8086 mode via Task Gate
From virtual 8086 mode to 286 TSS via Task Gate
From virtual 8086 mode to Intel 386 SX CPU TSS via task gate
From virtual 8086 mode to privilege level 0 via trap gate or interrupt gate
INTERRUPT RETURN
24
g, h, j, k,
r
IRET = Interrupt return
1 1 0 0 1 1 1 1
Protected Mode Only
(IRET)
To same privilege level (within task)
42
g, h, j, k,
r
E-17
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table E-1. Instruction Set Summary (Sheet 17 of 19)
Clock Count
Notes
Real
Ad-
Pro-
tected
Real
Ad-
Pro-
tected
dress Virtual dress Virtual
Instruction
Format
Mode
or
Ad-
dress
Mode
or
Ad-
dress
Virtual Mode Virtual Mode
8086
8086
Mode
Mode
To different privilege level (within task)
From 286 task to 286 TSS
86
g, h, j, k,
r
285
318
267
g, h, j, k,
r
From 286 task to Intel386 SX CPU TSS
From 286 task to virtual 8086 task
g, h, j, k,
r
g, h, j, k,
r
From 286 task to virtual 8086 mode (within task)
From Intel386 SX CPU task to 286 TSS
113
324
g, h, j, k,
r
From Intel386 SX CPU task to Intel386 SX CPU TSS
From Intel386 SX CPU task to virtual 8086 task
328
377
113
g, h, j, k,
r
g, h, j, k,
r
From Intel386 SX CPU task to virtual 8086 mode (within task)
PROCESSOR CONTROL INSTRUCTIONS
HLT = Halt
1 1 1 1 0 1 0 0
7
7
l
l
MOV = Move to and from control/debug/test registers
CR0/CR2/CR3 from
register
10/4/5
10/4/5
0 0 0 0 1 1 1 1
0 0 1 0 0 0 1 0
1 1 e e e reg
register from CR0-3
DR0-3 from register
DR6-7 from register
register from DR6-7
register from DR0-3
TR6-7 from register
register from TR6-7
NOP = No operation
0 0 0 0 1 1 1 1
0 0 0 0 1 1 1 1
0 0 0 0 1 1 1 1
0 0 0 0 1 1 1 1
0 0 0 0 1 1 1 1
0 0 0 0 1 1 1 1
0 0 0 0 1 1 1 1
1 0 0 1 0 0 0 0
0 0 1 0 0 0 0 0
0 0 1 0 0 0 1 1
0 0 1 0 0 0 1 1
0 0 1 0 0 0 0 1
0 0 1 0 0 0 0 1
0 0 1 0 0 1 1 0
0 0 1 0 0 1 0 0
1 1 e e e reg
1 1 e e e reg
1 1 e e e reg
1 1 e e e reg
1 1 e e e reg
1 1 e e e reg
1 1 e e e reg
6
6
l
l
l
l
l
l
l
22
16
14
22
12
12
3
22
16
14
22
12
12
3
WAIT = Wait until BUSY#
pin is negated
6
6
1 0 0 1 1 0 1 1
PROCESSOR EXTENSION INSTRUCTIONS
Processor Extension
Escape
1 1 0 1 1 T T T
mod L L L r/m
See
Intel387
SX
datashe
et for
h
TTT and LLL bits are opcode
information for coprocessor
clock
counts
E-18
INSTRUCTION SET SUMMARY
Table E-1. Instruction Set Summary (Sheet 18 of 19)
Clock Count
Notes
Real
Ad-
Pro-
tected
Real
Ad-
Pro-
tected
dress Virtual dress Virtual
Instruction
Format
Mode
or
Ad-
dress
Mode
or
Ad-
dress
Virtual Mode Virtual Mode
8086
8086
Mode
Mode
PREFIX BYTES
Address size prefix
0 1 1 0 0 1 1 1
1 1 1 1 0 0 0 0
0 1 1 0 0 1 1 0
0
0
0
0
0
0
LOCK = Bus lock prefix
m
Operand size prefix
Segment override prefix
CS:
0 0 1 0 1 1 1 0
0 0 1 1 1 1 1 0
0 0 1 0 0 1 1 0
0 1 1 0 0 1 0 0
0 1 1 0 0 1 0 1
0 0 1 1 0 1 1 0
0
0
0
0
0
0
0
0
0
0
0
0
DS:
ES:
FS:
GS:
SS:
PROTECTION CONTROL
ARPL = adjust requested privilege level
from register/memory
LAR = load access rights
from register/memory
0 1 1 0 0 0 1 1
mod reg r/m
N/A
N/A
11*
11*
N/A
20/21**
15/16*
11*
a
a
h
0 0 0 0 1 1 1 1
0 0 0 0 0 0 1 0
0 0 0 0 0 0 0 1
0 0 0 0 0 0 0 1
mod reg r/m
mod 0 1 0 r/m
mod 0 1 1 r/m
g, h, j, p
h, l
LGDT = load global descriptor
table register
0 0 0 0 1 1 1 1
b, c
b, c
a
LIDT = load interrupt descriptor
table register
0 0 0 0 1 1 1 1
11*
h, l
LLDT = load local descriptor
table register to
register/memory
20/24*
g, h, j, l
0 0 0 0 1 1 1 1
0 0 0 0 0 0 0 0
mod 0 1 0 r/m
LMSW = load machine status word
from register/memory
LSL = load segment limit
from register memory
0 0 0 0 1 1 1 1
0 0 0 0 0 0 0 1
0 0 0 0 0 0 1 1
mod 1 1 0 r/m
mod reg r/m
10/13
10/13*
b, c
h, l
0 0 0 0 1 1 1 1
Byte-Granular limit
Page-Granular limit
N/A
N/A
20/21*
25/26*
a
a
g, h, j, p
g, h, j, p
LTR = load task register
from register/memory
0 0 0 0 1 1 1 1
0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 1
mod 0 0 1 r/m
mod 0 0 0 r/m
N/A
9*
23/27*
9*
a
g, h, j, l
SGDT = store global descriptor
table register
0 0 0 0 1 1 1 1
b, c
h
SIDT = store interrupt descriptor
E-19
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table E-1. Instruction Set Summary (Sheet 19 of 19)
Clock Count
Notes
Real
Ad-
Pro-
tected
Real
Ad-
Pro-
tected
dress Virtual dress Virtual
Instruction
Format
Mode
or
Ad-
dress
Mode
or
Ad-
dress
Virtual Mode Virtual Mode
8086
8086
Mode
Mode
table register
0 0 0 0 1 1 1 1
0 0 0 0 0 0 0 1
mod 0 0 1 r/m
9*
9*
b, c
h
SLDT = store local descriptor table
to register/memory
0 0 0 0 1 1 1 1
0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 1
mod 0 0 0 r/m
mod 1 0 0 r/m
N/A
2/2*
2/2*
a
h
SMSW = store machine
status word
2/2*
b, c
h, l
0 0 0 0 1 1 1 1
STR = store task register
to register/memory
0 0 0 0 1 1 1 1
0 0 0 0 0 0 0 0
mod 0 0 1 r/m
N/A
2/2*
a
h
VERR = verify read access
register/memory
0 0 0 0 1 1 1 1
0 0 0 0 1 1 1 1
0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0
mod 1 0 0 r/m
mod 1 0 1 r/m
N/A
N/A
10/11*
15/16*
a
a
g, h, j, p
g, h, j, p
VERW = verify write
access
NOTES:
Notes a through c apply to Real Address Mode only:
a.
b.
This is a Protected Mode instruction. Attempted execution in Real Mode will result in exception 6
(invalid opcode).
Exception 13 fault (general protection) will occur in Real Mode if an operand reference is made that
partially or fully extends beyond the maximum CS, DS, ES, FS, or GS limit, FFFFH. Exception 12 fault
(stack segment limit violation or not present) will occur in Real Mode if an operand reference is made
that partially or fully extends beyond the maximum SS limit.
c.
This instruction may be executed in Real Mode, its purpose is primarily to initialize the CPU for
Protected Mode.
Notes d through g apply to Real Address Mode and Protected Virtual Address Mode:
d.
The Intel386 SX CPU uses an early-out multiply algorithm. The actual number of clocks depends on
the position of the most significant bit in the operand (multiplier).
Clock counts given are minimum to maximum. To calculate actual clocks use this formula:
Actual Clock = if m < > 0 then max ([log2 |m| ], 3) + b clocks:
if m = 0 then 3 + b clocks
In this formula, m is the multiplier, and
b = 9 for register to register
b = 12 for memory to register
E-20
INSTRUCTION SET SUMMARY
b = 10 for register with immediate to register
b = 11 for memory with immediate to register.
e.
f.
An exception may occur, depending on the value of the operand.
LOCK# is automatically asserted, regardless of the presence or absence of the LOCK# prefix.
LOCK# is asserted during descriptor table accesses.
g.
Notes h through r apply to Protected Virtual Address Mode only:
h.
i.
Exception 13 fault (general protection violation) will occur if the memory operand in CS, DS, ES, FS, or
GS cannot be used due to either a segment limit violation or access rights violation. If a stack limit is
violated, an exception 12 (stack segment limit violation or not present) occurs.
For segment load operations, the CPL, RPL, and DPL must agree with the privilege rules to avoid an
exception 13 fault (general protection violation). The segment’s descriptor must indicate “present” or
exception 11 (CS, DS, ES, FS, GS not present.) If the SS register is loaded and a stack segment not
present is detected, an exception 12 (stack segment limit violation or not present) occurs.
j.
All segment descriptor accesses in the GDT or LDT made by this instruction will automatically assert
LOCK# to maintain descriptor integrity in multiprocessor systems.
k.
JMP, CALL, INT, RET, and IRET instructions referring to another code segment will cause an
exception 13 (general protection violation) if an applicable privilege rule is violated.
l.
An exception 13 fault occurs if CPL is greater than 0 (0 is the most privileged level)
An exception 13 fault occurs if CPL is greater than IOPL.
m.
n.
The IF bit of the flag register is not updated if CPL is greater than IOPL. The IOPL and VM fields of the
flag register are updated only if CPL = 0.
o.
p.
q.
The PE bit of the MSW (CR0) cannot be reset by this instruction. Use MOV into CR0 when resetting
the PE bit.
Any violation of privilege rules as applied to the selector operand does not cause a protection
exception; rather the zero flag is cleared.
If the coprocessor’s memory operand violates a segment limit or segment access rights, an exception
13 fault (general protection exception) will occur before the ESC instruction is executed. An exception
12 fault (stack segment limit violation or not present) will occur if the stack limit is violated by the
operand’s starting address.
r.
The destination of a JMP, CALL, INT, RET, or IRET must be in the defined limit of a code segment or
an exception 13 fault (general protection violation) will occur.
s.
†
The instruction will execute in s clocks if CPL ≤ IOPL. If CPL IOPL, the instruction will take t clocks.
>
Clock count shown applies if I/O permission allows I/O to the port in virtual 8086 mode. If I/O bit map
denies permission, exception 13 (general protection fault occurs; refer to clock counts for INT 3
instruction.
E-21
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
E.2 INSTRUCTION ENCODING
All instruction encodings are subsets of the general instruction format shown in Figure E-1. In-
structions consist of one or two primary opcode bytes, possibly an address specifier consisting of
the “mod r/m” byte and “scaled index” byte, a displacement if required, and an immediate data
field if required.
Within the primary opcode or opcodes, smaller encoding fields may be defined. These fields vary
according to the class of operation. The fields define such information as direction of the opera-
tion, size of the displacements, register encoding, or sign extension.
Almost all instructions referring to an operand in memory have an addressing mode byte follow-
ing the primary opcode byte(s). This byte, the mod r/m byte, specifies the address mode to be
used. Certain encodings of the mod r/m byte indicate a second addressing byte, the scale-index-
Addressing modes can include a displacement immediately following the mod r/m byte, or scaled
index byte. If a displacement is present, the possible sizes are 8, 16, or 32 bits.
If the instruction specifies an immediate operand, the immediate operand follows any displace-
ment bytes. The immediate operand, if specified, is always the last field of the instruction.
Figure E-1 illustrates several of the fields that can appear in an instruction, such as the mod field
and the r/m field, but the figure does not show all fields. Several smaller fields also appear in cer-
tain instructions, sometimes within the opcode bytes themselves. Table E-2 is a complete list of
all fields appearing in the instruction set. Following Table E-2 are detailed tables for each field.
T T T T T T T T T T T T T T T T mod T T T r/m
ss index base d32 16
7 6 5 3 2 0
8
none data32 16
8
none
7
0
7
0
7 6 5 3 2 0
opcode
(one or two bytes)
(T represents an opcode bit)
address
displacement
(4,2,1 bytes
or none)
immediate
data
(4, 2, 1 bytes
or none)
“mod r/m”
byte
“s-i-b”
byte
register and address
mode specifier
Figure E-1. General Instruction Format
E-22
INSTRUCTION SET SUMMARY
Table E-2. Fields Within Instructions
Number of
Bits
Field Name
Description
w
d
s
Specifies if data is byte of full size (full size is either 16 or 32 bits)
Specifies Direction of Data Operation
1
1
1
3
Specifies if an Immediate Data Field must be Sign-Extended
General Register Specifier
reg
mod r/m
Address Mode Specifier (Effective Address can be a General Register)
2 for mod:
3 for r/m
ss
Scale Factor for Scaled Index Address Mode
General Register to be used as Base Register
Segment Register Specifier for CS, SS, DS, ES
Segment Register Specifier for CS, SS, DS, ES, FS, GS
2
3
3
2
3
index
base
sreg2
sreg3
tttn
For Conditional Instructions, specifies a condition asserted
or a condition negated
4
NOTE: Figure E-1 shows encoding of individual instructions.
E.2.1 32-bit Extensions of the Instruction Set
With the Intel386 EX processor the 8086/80186/80286 instruction set is extended in two orthog-
onal directions: 32-bit forms of all 16-bit instructions are added to support the 32-bit data types,
and 32-bit addressing modes are made available for all instructions referencing memory. This or-
thogonal instruction set extension is accomplished having a Default (D) bit in the code segment
descriptor, and by having 2 prefixes to the instruction set.
The instruction defaults to operations of 16 bits or 32 bits, depending on the setting of the D bit
in the code segment descriptor, which gives the default length (either 32 bites or 16 bits) for both
operands and effective addresses when executing that code segment. In the Real Address Mode
or Virtual 8086 Mode, no code segment descriptors are used, but a D value of 0 is assumed inter-
nally by the Intel386 EX processor when operating in those modes (for 16-bit default sizes com-
patible with the 8086/80186/80286).
Two prefixes, the Operand Size Prefix and the Effective Address Size Prefix, allow overriding
individually the Default selection of operand size and effective address size. These prefixes may
precede any opcode bytes and affect only the instruction they precede. If necessary, one or both
prefixes may be placed before the opcode bytes. The presence of the Operand Size Prefix and the
Effective Address Prefix toggles the operand size or the effective address size, respectively, to the
value “opposite” from the default setting. For example, if the default operand size is for 32-bit
data operations, then presence of the Operand Size Prefix toggles the instruction to use 32-bit ef-
fective address computations.
These 32-bit extensions are available in all modes, including the Real Address Mode or the Vir-
tual 8086 Mode. In these modes the default is always 16 bits, so prefixes are needed to specify
32-bit operands or addresses. For instructions with more than one prefix the order of prefixes is
unimportant.
E-23
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Unless specified otherwise, instructions with 8-bit and 16-bit operands do not affect the contents
of the high-order bits of the extended registers.
E.2.2 Encoding of Instruction Fields
Within the instruction are several fields indicating register selection, addressing mode, and so on.
The exact encodings of these fields are defined in the next several section.
E.2.2.1
Encoding of Operand Length (w) Field
For any given instruction performing a data operation, the instruction is executing as a 32-bit op-
eration or a 16-bit operation. Within the constraints of the operation size, the w field encodes the
operand size as either one byte or the full operation size, as shown in Table E-3.
Table E-3. Encoding of Operand Length (w) Field
Operand Size During 16-bit Operand Size During 32-bit
w Field
Data Operations
Data Operations
0
1
8 bits
8 bits
16 bits
32 bits
E.2.2.2
Encoding of the General Register (reg) Field
The general register is specified by the reg field, which may appear in the primary opcode bytes,
or as the reg field of the “mod r/m” byte, or as the r/m field of the “mod r/m” byte.
Table E-4. Encoding of reg Field When w Field is not Present in Instruction
Register Selected During
16-bit Data Operations
Register Selected During
32-bit Data Operations
reg Field
000
001
010
011
100
101
110
111
AX
CX
DX
BX
SP
BP
SI
EAX
ECX
EDX
EBX
ESP
EBP
ESI
DI
EDI
E-24
INSTRUCTION SET SUMMARY
Table E-5. Encoding of reg Field When w Field is Present in Instruction
Register Specified by reg Field During 16-bit Data Operations
Function of w Field
reg
(when w = 0)
(when w = 1)
000
001
010
011
100
101
110
111
AL
CL
DL
BL
AH
CH
DH
BH
AX
CX
DX
BX
SP
BP
SI
DI
Register Specified by reg Field During 32-bit Data Operations
Function of w Field
reg
(when w = 0)
(when w = 1)
000
001
010
011
100
101
110
111
AL
CL
DL
BL
AH
CH
DH
BH
EAX
ECX
EDX
EBX
ESP
EBP
ESI
EDI
E.2.2.3
Encoding of the Segment Register (sreg) Field
The sreg field in certain instructions is a 2-bit field allowing one of the four 80286 segment reg-
isters to be specified. The sreg field in other instructions is a 3-bit field, allowing the FS and GS
segment registers to be specified.
Table E-6. Encoding of the Segment Register (sreg) Field
2-Bit sreg2
Field
3-Bit sreg3
Field
Segment Register Selected
Segment Register Selected
00
01
10
11
ES
CS
SS
DS
000
001
010
011
100
101
110
111
ES
CS
SS
DS
FS
GS
do not use
do not use
E-25
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
E.2.2.4
Encoding of Address Mode
Except for special instructions, such as PUSH or POP, where the addressing mode is pre-deter-
mined, the addressing mode for the current instruction is specified by addressing bytes following
the primary opcode. The primary addressing byte is the “mod r/m” byte, and a second byte of ad-
The s-i-b byte is specified when using 32-bit addressing mode and the “mod r/m” byte has r/m =
100 and mod = 00, 01, or 10. When the s-i-b byte is present, the 32-bit addressing mode is a func-
tion of the mod, ss, index, and base fields.
The primary addressing byte, the “mod r/m” byte, also contains three bits (shown as TTT in Fig-
ure E-1) sometimes used as an extension of the primary opcode. The three bits, however, may
also be used as a register field (reg).
When calculating an effective address, either 16-bit addressing or 32-bit addressing is used. 16-
bit addressing uses 16-bit address components to calculate the effective address while 32-bit ad-
dressing uses 32-bit address components to calculate the effective address. When 16-bit address-
ing is used, the “mod r/m” byte is interpreted as a 32-bit addressing mode specifier.
The following tables define all encodings of all 16-bit addressing modes and 32-bit addressing
modes.
E-26
INSTRUCTION SET SUMMARY
Table E-7. Encoding of 16-bit Address Mode with “mod r/m” Byte
mod r/m
Effective Address
mod r/m
Effective Address
00 000
00 001
00 010
00 011
00 100
00 101
00 110
00 111
DS:[BX + SI]
DS:[BX + DI]
SS:[BP + SI]
SS:[BP + DI]
DS:[SI]
DS:[DI]
DS:d16
DS:[BX]
10 000
10 001
10 010
10 011
10 100
10 101
10 110
10 111
DS:[BX + SI + d16]
DS:[BX + DI + d16]
SS:[BP + SI + d16]
SS:[BX + DI + d16]
DS:[SI + d16]
DS:[DI + d16]
SS:[BP + d16]
DS:[BX + d16]
01 000
01 001
01 010
01 011
01 100
01 101
01 110
01 111
DS:[BX + SI + d8]
DS:[BX + DI + d8]
SS:[BP + SI + d8]
SS:[BP + DI + d8]
DS:[SI + d8]
DS:[DI + d8]
SS:[BP + d8]
11 000
11 001
11 010
11 011
11 100
11 101
11 110
11 111
register - see tables below
register - see tables below
register - see tables below
register - see tables below
register - see tables below
register - see tables below
register - see tables below
register - see tables below
DS:[BX + d8]
Register Specified by r/m
During 16-bit Data Operations
Function of w Field
mod r/m
(when w = 0) (when w = 1)
11 000
11 001
11 010
11 011
11 100
11 101
11 110
11 111
AL
CL
DL
BL
AH
CH
DH
BH
AX
CX
DX
BX
SP
BP
SI
DI
Register Specified by r/m
During 32-bit Data Operations
Function of w Field
mod r/m
(when w = 0) (when w = 1)
11 000
11 001
11 010
11 011
11 100
11 101
11 110
11 111
AL
CL
DL
BL
AH
CH
DH
BH
EAX
ECX
EDX
EBX
ESP
EBP
ESI
EDI
E-27
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
Table E-8. Encoding of 32-bit Address Mode with “mod r/m” Byte (No s-i-b Byte Present)
mod r/m
Effective Address
mod r/m
Effective Address
00 000
00 001
00 010
00 011
00 100
00 101
00 110
00 111
DS:[EAX]
DS:[ECX]
SS:[EDX]
SS:[EBX]
s-i-b is present
DS:[d32]
10 000
10 001
10 010
10 011
10 100
10 101
10 110
10 111
DS:[EAX + d32]
DS:[ECX + d32]
SS:[EDX + d32]
SS:[EBX + d32]
s-i-b is present
SS:[EBP + d32]
SS:[ESI + d32]
DS:[EDI + d32]
DS:[ESI]
DS:[EDI]
01 000
01 001
01 010
01 011
01 100
01 101
01 110
01 111
DS:[EAX + d8]
DS:[ECX + d8]
SS:[EDX + d8]
SS:[EBX + d8]
s-i-b is present
SS:[EBP + d8]
DS:[ESI + d8]
DS:[EDI + d8]
11 000
11 001
11 010
11 011
11 100
11 101
11 110
11 111
register - see tables below
register - see tables below
register - see tables below
register - see tables below
register - see tables below
register - see tables below
register - see tables below
register - see tables below
Register Specified by r/m
During 16-bit Data Operations
Function of w Field
mod r/m
(when w = 0) (when w = 1)
11 000
11 001
11 010
11 011
11 100
11 101
11 110
11 111
AL
CL
DL
BL
AH
CH
DH
BH
AX
CX
DX
BX
SP
BP
SI
DI
Register Specified by r/m
During 32-bit Data Operations
Function of w Field
mod r/m
(when w = 0) (when w = 1)
11 000
11 001
11 010
11 011
11 100
11 101
11 110
11 111
AL
CL
DL
BL
AH
CH
DH
BH
EAX
ECX
EDX
EBX
ESP
EBP
ESI
EDI
E-28
INSTRUCTION SET SUMMARY
Table E-9. Encoding of 32-bit Address Mode (“mod r/m” Byte and s-i-b Byte Present)
mod r/m
Effective Address
00 000
00 001
00 010
00 011
00 100
00 101
00 110
00 111
DS:[EAX + (scaled index)]
DS:[ECX + (scaled index)]
DS:[EDX + (scaled index)]
DS:[EBX + (scaled index)]
SS:[ESP + (scaled index)]
DS:[d32 + (scaled index)]
DS:[ESI + (scaled index)]
DS:[EDI + (scaled index)]
01 000
01 001
01 010
01 011
01 100
01 101
01 110
01 111
DS:[EAX + (scaled index) + d8]
DS:[ECX + (scaled index) + d8]
DS:[EDX + (scaled index) + d8]
DS:[EBX + (scaled index) + d8]
SS:[ESP + (scaled index) + d8]
SS:[EBP + (scaled index) + d8]
DS:[ESI + (scaled index) + d8]
DS:[EDI + (scaled index) + d8]
10 000
10 001
10 010
10 011
10 100
10 101
10 110
10 111
DS:[EAX + (scaled index) + d32]
DS:[ECX + (scaled index) + d32]
DS:[EDX + (scaled index) + d32]
DS:[EBX + (scaled index) + d32]
SS:[ESP + (scaled index) + d32]
SS:[EBP + (scaled index) + d32]
DS:[ESI + (scaled index) + d32]
DS:[EDI + (scaled index) + d32]
NOTE: Mod field in “mod r/m” byte; ss, index, base fields in “s-i-b” byte.
ss
Scale Factor
00
01
10
11
x1
x2
x4
x8
Index
Index Register
000
001
010
011
100
101
110
111
EAX
ECX
EDX
EBX
no index reg†
EBP
ESI
EDI
†
When index field is 100, indicating “no index register,”
the ss field must equal 00. If this is not true, the
effective address is undefined
E-29
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
E.2.2.5
Encoding of Operation Direction (d) Field
In many two-operand instructions the d field is present to indicate which operand is considered
the source and which is the destination.
Table E-10. Encoding of Operation Direction (d) Field
d
Direction of Operation
0
Register/Memory<--Register
“reg” field indicates source operand;
“mod r/m” or “mod ss index base” indicates destination operand.
1
Register<--Register/Memory
“reg” field indicates destination operand;
“mod r/m” or “mod ss index base” indicates source operand.
E.2.2.6
Encoding of Sign-Extend (s) Field
The s field occurs primarily to instructions with immediate data fields. The s field has an effect
only if the size of the immediate data is 8 bits and is being placed in a 16-bit or 32-bit destination.
Table E-11. Encoding of Sign-Extend (s) Field
s
Effect on Immediate Data8
Effect on Immediate Data 16|32
0
1
None
None
None
Sign-Extend Data8 to fill
16-bit or 32-bit destination
E.2.2.7
Encoding of Conditional Test (tttn) Field
For the conditional instructions (conditional jumps and set on condition), tttn is encoded with n
indicating to use the condition (n=0) or its negation (n=1), and ttt giving the condition to test.
Table E-12. Encoding of Conditional Test (tttn) Field
Mnemonic
Condition
tttn
O
NO
Overflow
No Overflow
Below/Not Above or Equal
Not Below/Above or Equal
Equal/Zero
0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111
B/NAE
NB/AE
E/Z
NE/NZ
BE/NA
NBE/A
S
Not Equal/Not Zero
Below or Equal/Not Above
Not Below or Equal/Above
Sign
NS
Not Sign
Parity/Parity Even
P/PE
NP/PO
L/NGE
NL/GE
LE/NG
NLE/G
Not Parity/Parity Odd
Less Than/Not Greater or Equal
Not Less Than/Greater or Equal
Less Than or Equal/Greater Than
Not Less or Equal/Greater Than
E-30
INSTRUCTION SET SUMMARY
E.2.2.8
Encoding of Control or Debug or Test Register (eee) Field
For the loading and storing of the Control, Debug and Test registers.
Table E-13. When Interpreted as Control Register Field
eee Code
Reg Name
000
010
011
CR0
CR2
CR3
NOTE: Do not use any other encoding
Table E-14. When Interpreted as Debug Register Field
eee Code
Reg Name
000
001
010
011
110
111
DR0
DR1
DR2
DR3
DR6
DR7
NOTE: Do not use any other encoding
Table E-15. When Interpreted as Test Register Field
eee Code
Reg Name
110
111
TR6
TR7
NOTE: Do not use any other encoding
E-31
GLOSSARY
GLOSSARY
This glossary defines acronyms, abbreviations, and terms that have special meaning in this man-
ual. (Chapter 1, GUIDE TO THIS MANUAL, discusses notational conventions.)
Assert
The act of making a signal active (enabled). The
polarity (high/low) is defined by the signal name.
Active-low signals are designated by a pound symbol
(#) suffix; active-high signals have no suffix. To
assert RD# is to drive it low; to assert HLDA is to
drive it high.
BIOS
Basic input/output system. The interface between the
hardware and the operating system.
BIU
Bus interface unit. The internal peripheral that
controls the external bus.
Boundary-scan
The term boundary-scan refers to the ability to scan
(observe) the signals at the boundary (the pins) of a
device. A major component of the JTAG standard.
CSU
Chip-select unit. The internal peripheral that selects
an external memory device during an external bus
cycle.
Clear
The term clear refers to the value of a bit or the act of
giving it a value. If a bit is clear, its value is “0”;
clearing a bit gives it a “0” value.
Deassert
The act of making a signal inactive (disabled). The
polarity (high/low) is defined by the signal name.
Active-low signals are designated by a pound symbol
(#) suffix; active-high signals have no suffix. To
deassert RD# is to drive it high; to deassert HLDA is
to drive it low.
DMA
Direct memory access controller. The internal
peripheral that allows external or internal peripherals
to transfer information directly to or from the system.
The two-channel DMA controller is an enhanced
version of the industry-standard 8237A DMA
peripheral.
DOS Address Space
Addresses 0H–03FFH. The internal timers, interrupt
controller, serial I/O ports, and DMA controller can
Glossary-1
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
be mapped into this space. In this manual, the terms
DOS address and PC/AT address are synonymous.
DOS-compatible Mode
Edge-triggered
The addressing mode in which the internal timer,
interrupt controller, serial I/O ports, and DMA
controller are mapped into the DOS address space.
This mode decodes only the lower 10 address bits, so
the expanded address space is inaccessible.
The mode in which the interrupt controller recognizes
a rising edge (low-to-high transition) on an interrupt
request signal as an interrupt request. The internal
peripherals use edge-triggered interrupt requests; this
is compatible with the PC/AT bus specification.
External peripherals can use either edge-triggered or
level-sensitive interrupt requests.
Enhanced DOS Mode
The addressing mode in which the internal timer,
interrupt controller, serial I/O ports, and DMA
controller are mapped into both the DOS address
space and the expanded address space. This mode
decodes all 16 address bits. All internal peripherals
can be accessed in the expanded address space; the
internal timer, interrupt controller, serial I/O ports,
and DMA controller can also be accessed in the DOS
address space.
Expanded Address Space
Addresses 0F000H–0F8FFH. All internal peripheral
registers reside in this space. The internal timer,
interrupt controller, serial I/O ports, and DMA
controller can also be mapped into DOS (or PC/AT)
address space.
ICU
Interrupt control unit. The internal peripheral that
receives interrupt requests from internal peripherals
and external pins, resolves priority, and presents the
requests to the CPU. The ICU is functionally identical
to two industry-standard 82C59A programmable
interrupt controllers connected in cascade.
Idle Mode
The power conservation mode that freezes the core
clocks but leaves the peripheral clocks running.
Interrupt Latency
The delay between the time that the master 82C59A
presents an interrupt request to the CPU and the time
that the interrupt acknowledge cycle begins.
Glossary-2
GLOSSARY
Interrupt Response Time
Interrupt Resolution
The amount of time required to complete an interrupt
acknowledge cycle and transfer program control to
the interrupt service routine.
The delay between the time that the interrupt
controller receives an interrupt request and the time
that the master 82C59A presents the request to the
CPU.
ISR
Interrupt service routine. A user-supplied software
routine designed to service specific interrupt requests.
JTAG
Joint Test Action Group. The IEEE technical
subcommittee that developed the testability standard
published as Standard 1149.1-1990, IEEE Standard
Test Access Port and Boundary-Scan Architecture,
and its supplement, Standard 1149.1a-1993. The test-
logic unit is fully compliant with this standard.
Level-sensitive
The mode in which the interrupt controller recognizes
a high level (logic one) on an interrupt request signal
as an interrupt request. Unlike an edge-triggered
interrupt request, a level-sensitive interrupt request
will continue to generate interrupts as long as it is
asserted.
LSB
Least-significant bit of a byte or least-significant byte
of a word.
NonDOS Mode
The addressing mode in which the internal timer,
interrupt controller, serial I/O ports, and DMA
controller are mapped into the expanded address
space. This mode decodes all 16 address bits. All
internal peripherals can be accessed only in the
expanded address space.
Nonintrusive DOS Mode
The addressing mode in which the internal timer,
interrupt controller, serial I/O ports, and DMA
controller can be individually mapped out of the DOS
address space and replaced by the corresponding
external peripherals. This mode decodes only the
lower 10 address bits, so the expanded address space
is inaccessible.
Normally not-ready
The term normally not-ready refers to a system in
which a bus cycle continues until the accessed device
asserts READY#.
Glossary-3
Intel386™ EX EMBEDDED MICROPROCESSOR USER’S MANUAL
PC/AT Address Space
Pipelining
Addresses 0H–03FFH. The internal timers, interrupt
controller, serial I/O ports, and DMA controller can
be mapped into this space. In this manual, the terms
DOS address and PC/AT address are synonymous.
A bus interface technique that controls the address
and status outputs so the outputs for the next bus cycle
become valid before the end of the current bus cycle,
allowing external bus cycles to overlap. By increasing
the amount of time available for external memory or
I/O devices to respond, pipelining allows systems to
achieve high bandwidth with relatively slow,
inexpensive components.
Powerdown Mode
RCU
The power conservation mode that freezes both the
core clocks and the peripheral clocks.
Refresh control unit. The module that simplifies the
interface between the processor and DRAM
components by providing the necessary bus control
and timing for refresh operations.
Reserved Bits
Set
Register bits that are not used in this device but may
be used in future implementations. Avoid any
software dependence on these bits.
The term set refers to the value of a bit or the act of
giving it a value. If a bit is set, its value is “1”; setting
a bit gives it a “1” value.
SIO Unit
SMM
Serial input/output unit. The internal peripheral that
allows the system to communicate with external
peripheral devices and modems.
System management mode. The hardware and
software enhancement that reduces system power
consumption by allowing the device to execute
specific routines for power management.
SMRAM
A 32-Kbyte memory partition (38000H–3FFFFH)
used for SMM. The upper 512 bytes (3FE00H–
3FFFFH) are reserved for the CPU and must reside in
RAM; the remainder of the partition is used for user-
supplied driver code and may reside in read-only
storage.
SSIO Unit
Synchronous serial input/output unit. The internal
peripheral that provides 16-bit bidirectional serial
Glossary-4
GLOSSARY
communications. The transmitter and receiver can
operate independently (with different clocks) to
provide full-duplex communication.
State Time (or State)
The basic time unit of the device; the combined
period of the two internal timing signals, PH1 and
PH2. With a 50 MHz external clock, one state time
equals 80 ns. Because the device can operate at many
frequencies, this manual defines time requirements in
terms of state times rather than in specific units of
time.
TAP
Test access port. The dedicated input and output pins
through which a tester communicates with the test-
logic unit. A major component of the JTAG standard.
TCU
Timer/counter unit. The internal peripheral that
provides three independent 16-bit down-counters.
Test-logic Unit
The module that facilitates testing of the device logic
and interconnections between the device and the
board. This module is fully compliant with IEEE
Standard 1149.1, commonly called the JTAG
standard.
UART
WDT
Universal asynchronous receiver and transmitter. A
part of the SIO unit.
Watchdog timer. An internal, 32-bit down-counter
that can operate as a general-purpose timer, a software
watchdog timer, or a bus monitor.
Glossary-5
Index
INDEX
#, defined, 1-3
82C59A, 9-1
SSIO unit, 13-2, 13-3
baud-rate generator clock, 13-5
timer/counter unit, 10-2
watchdog timer unit, 17-2
BOUND, 18-2
Boundary scan register, 18-1
Built-in self-test, 8-12
A
Address bus, 6-1
Address lines
new, 3-1
Address space
configuration register, 4-6
expanded I/O, 4-3
Bulletin board system (BBS), 1-7
Bus arbiter
register addresses, 4-15, D-1
Bus arbiter, configuration, 5-3
Bus control arbitration, 12-9
Bus cycle length adjustments for overlapping
chip-select regions, 14-11, 14-12
Bus interface pins, 6-3
Bus interface unit, 3-4, 6-1–6-37
address bus, 6-1
enabling/disabling, 4-8
I/O decoding techniques, 4-6
I/O for PC/AT systems, 4-2
peripheral registers, 4-15
Addressing modes, 4-9–4-14
DOS-compatible mode, 4-9–4-10
enhanced DOS mode, 4-11, 4-13
nonDOS mode, 4-11, 4-14
nonintrusive DOS mode, 4-11, 4-12
AEN signal, deriving, B-2–B-3
AEOI mode, 9-9
aligned data transfers, 6-9
Applications, typical, 2-1
Architectural overview, 2-1–2-4
Assert, defined, 1-4
Asynchronous serial I/O unit, See Serial I/O unit
Automatic end of interrupt (AEOI) mode, 9-9
bus control pins, 6-2
bus cycles, 6-13–6-33
BS8, 6-31–6-33
halt/shutdown, 6-26–6-27
interrupt acknowledge, 6-23–6-25
pipelined, 6-19–6-23
read, 6-13–6-14
refresh, 6-28–6-30
write, 6-16–6-18
bus lock, 6-34–6-35
LOCK# signal duration, 6-35
locked cycle activators, 6-34
locked cycle timing, 6-34
bus operation, 6-5–6-14
bus state diagram, 6-8, 6-20
bus states, 6-7–6-8
bus status
definitions, 6-5
data bus, 6-1
transfers and operand alignment, 6-9
HOLD/HLDA, 6-20, 6-35
departures from PC/AT architecture, B-4
HOLD signal latency, 6-37
timing, 6-36
B
Baud-rate generator, 11-4–11-5, 13-5–13-6
BIU, See Bus interface unit
Block diagram
clock and power management unit, 8-2
DMA unit, 12-2
I/O port, 16-2
JTAG test-logic unit, 18-2
SIO unit, 11-2
baud-rate generator clock, 11-4
modem control signals, 11-29
receiver, 11-9
transmitter, 11-7
Index-1
INTEL386™ EX MICROPROCESSOR USER’S MANUAL
operation during idle mode, 8-5
overview, 6-1–6-3
pipelining, 6-8
ready logic, 6-10
See also Bus control arbitration
signals, 6-3–6-4
Bus signals, departures from PC/AT
architecture, B-2–B-3
Bus size control for chip-selects, 14-11
BYPASS, 18-2
idle mode, 8-9
overview, 8-1–8-7
power management logic, 8-3–8-5
powerdown mode, 8-10
registers, 8-6
CLKPRS, 8-7
PWRCON, 8-8
reset considerations, 8-11
signals, 8-6
synchronization, 8-3
timing diagram, 8-9–8-11
Clock management
C
CAS#-before-RAS# refresh, 15-1, 15-12
Chip-select unit, 14-1–14-24
operation, 14-2–14-12
register addresses, 4-19, D-5
Clock synchronization, 8-3
Code Prefetch Unit, 3-4
CompuServe forums, 1-7
Configuration
bus arbiter, 5-3–5-5
core, 5-21–5-22
device, 5-1–5-37
DMA controller, 5-3
example, 5-28–5-33
I/O ports, 5-23, 5-25, 5-26, 5-27, 9-18, 10-22,
11-18, 11-19, 11-20, 14-16, D-43, D-44,
D-45
bus cycle length adjustments, 14-12
bus cycle length control, 14-11
bus size control, 14-11
defining a channel’s address block, 14-2–
14-9
overlapping regions, 14-11
system management mode support, 14-10
overview, 14-1
programming, 14-13–14-20
considerations, 14-22
CSnADH, 14-17, D-8
interrupt control unit, 5-7
pins, 5-23–5-27
Port92, 5-22
CSnADL, 14-18, D-9
CSnMSKH, 14-19, D-10
CSnMSKL, 14-20, D-11
P2CFG register, 14-16
PINCFG register, 14-15
UCSADH, 14-17, D-8
procedure, 5-28
refresh control unit, 5-3
serial I/O unit, 5-14
serial synchronous I/O unit, 5-18
timer/counter unit, 5-11
worksheets, 5-34–5-37
Core
configuring, 5-21–5-22
Core architecture, 2-1
Core overview
CX enhancements, 3-1
Internal architecture, 3-2
CPU-only reset, 5-22, B-4
CSU, See Chip-select unit
Customer service, 1-6
CX internal architecture, 3-2
UCSADL, 14-18, D-9
UCSMSKH, 14-19, D-10
UCSMSKL, 14-20, D-11
register addresses, 4-17, D-3
registers, 14-14–14-20
signals, 14-13
Clear, defined, 1-5
Clock and power management unit, 8-1–8-13
clock generation logic, 8-1–8-3
controlling power management modes, 8-8
controlling PSCLK frequency, 8-7
design considerations
powerdown considerations, 8-13
reset considerations, 8-11
Index-2
INDEX
programming, 12-28–12-51
D
address and byte count registers, 12-33
channel registers, 12-33
considerations, 12-50
Deassert, defined, 1-4
Decoding techniques, I/O address, 4-6
Design considerations
DMA0BYCn, 12-33, D-24
DMA0REQn, 12-33, D-24
DMA0TARn, 12-33, D-24
DMA1BYCn, 12-33, D-24
DMA1REQn, 12-33, D-24
DMA1TARn, 12-33, D-24
DMABSR register, 12-46
DMACFG, 5-6, 12-32, D-14
DMACFG register, 12-32
DMACHR, 12-47, D-15
DMACHR register, 12-47
DMACMD1, 12-35, D-16
DMACMD1 register, 12-35
DMACMD2, 12-37, D-17
DMACMD2 register, 12-37
DMAGRPMSK, 12-45, D-18
DMAIEN, 12-48, D-19
clock and power management unit, 8-11
input/output ports, 16-10
interrupt control unit, 9-29–9-30
JTAG test-logic unit, 18-14
refresh control unit, 15-11
synchronous serial I/O unit, 13-25
Device configuration, 5-1–5-37
procedure, 5-28
register addresses, 4-19, D-5
worksheets, 5-34–5-37
DMA controller, 12-1–12-61
block diagram, 12-2
configuring, 5-3
departures from PC/AT architecture, B-1–B-3
DMACLR command, 12-50
DMACLRBP command, 12-50
DMACLRMSK command, 12-50
DMACLRTC command, 12-50
interrupts, 12-26–12-27
DMAIEN register, 12-48
DMAIS, 12-49, D-20
DMAIS register, 12-49
DMAMOD1, 12-39, D-21
DMAMOD1 register, 12-38
DMAMOD2, 12-41, D-22
DMAMOD2 register, 12-40–12-41
DMAMSK, 12-44, D-23
DMAOVFE register, 12-34
DMASRR, 12-43, D-26
DMASRR register, 12-42, 12-43
DMASTS, 12-36, D-27
DMASTS register, 12-36
PINCFG register, 12-28, 12-31
register addresses, 4-15, D-1
registers, 12-28
operation, 12-5–12-27
8237A compatibility, 12-27
basic refresh cycle, 15-5
buffer-transfer modes, 12-12
bus control arbitration, 12-9
bus cycle options for data transfers, 12-5–
12-8
cascade mode, 12-25–12-26
changing priority of DMA channel and
external bus requests, 12-10
data-transfer modes
block, 12-18–12-20
demand, 12-21–12-24
signals, 12-4
using with external devices, 5-3
Documents, related, 1-5
single, 12-14–12-17
DMA transfers, 12-5
ending DMA transfers, 12-10
starting DMA transfers, 12-9
overview, 12-1–12-4
DOS Address, defined, 1-4
Index-3
INTEL386™ EX MICROPROCESSOR USER’S MANUAL
DOS compatibility
IDCODE, 18-2
departures from PC/AT architecture
bus signals, B-2
Identifier registers, 3-6, 7-15
Idle mode, 8-9
bus interface unit operation during, 8-5
SMM interaction with, 8-5
timing diagram, 8-9
watchdog timer unit operation during, 8-5
IEEE Standard Test Access Port and
Boundary-Scan Architecture, 18-1
Input/output ports, 16-1–16-10
block diagram, 16-2
design considerations, 16-10
overview, 16-1–16-5
pin multiplexing, 16-5
pin reset status, 16-5
programming
CPU-only reset, B-4
DMA unit, B-1
HOLD, HLDA pins, B-4, B-5
interrupt control unit, B-4
SIO units, B-4
DRAM, refreshing, 15-12
DRAM, See Refresh control unit
E
EISA compatibility, 4-3–4-5
ESE bit programming, 4-8
Exceptions and interrupts, relative priority, 7-7
Execution Unit, 3-4, 3-5
Expanded address, defined, 1-4
Expanded I/O address space, 4-3
enabling/disabling, 4-8
initialization sequence, 16-10
pin configuration, 16-7
PnCFG register, 16-7
PnDIR register, 16-8
PnLTC register, 16-8
PnPIN register, 16-9
F
FaxBack service, 1-6
Flow diagram
register addresses, 4-19, D-5
registers, 16-6
signals, 16-5
Instruction Decode Unit, 3-4
Instruction Queue, 3-5
Instruction Register (IR), 18-7
Instructions, notational conventions, 1-3
Interrupt control unit, 9-1–9-30
configuring, 5-7
departure from PC/AT architecture, B-4
design considerations, 9-29
interrupt acknowledge cycle, 9-29–9-30
interrupt detection, 9-29
interrupt polling, 9-14–9-15
interrupt priority, 9-6–9-8
assigning an interrupt level, 9-6
changing the default interrupt structure, 9-7
determining priority, 9-7–9-8
interrupt process, 9-9–9-14
interrupt sources, 9-4
CSU bus cycle length adjustment, 14-12
demand data-transfer mode, 12-22–12-24
DMA block data-transfer mode, 12-19–12-20
DMA cascade mode, 12-26
DMA demand data-transfer mode, 12-22
DMA single data-transfer mode, 12-15–12-17
interrupt process, 9-11, 9-12, 9-13
SIO reception, 11-11
SIO transmission, 11-8
H
HALT cycle
Ready generation, 8-10
HALT restart from SMM, 7-9
HOLD, HLDA
departures from PC/AT architecture, B-4, B-5
timing, 6-20, 6-35
I
I/O ports, See Input/output ports
I/O restart from SMM, 7-9
ICU, See Interrupt control unit
interrupt service routine, 6-23
interrupt vectors, 9-8
Index-4
INDEX
operation, 9-4–9-16
overview, 9-1
overview, 18-1–18-2
Resetting upon power-up, 18-3
programming, 9-15–9-32
considerations, 9-32
testing, 18-10–18-11
bypassing devices on a board, 18-10
disabling the output drivers, 18-11
identifying the device, 18-10
ICW1, 9-20, D-28
ICW1 register, 9-20
ICW2, 9-21, D-29
sampling device operation and preloading
ICW2 register, 9-21
data, 18-10
testing the interconnections, 18-10
timing information, 18-12–18-13
ICW3, 9-22, 9-23, D-29, D-30
ICW3 register, 9-22, 9-23
ICW4, 9-24, D-30
L
ICW4 register, 9-24
IERn, 11-27, D-32
Literature, 1-8
IIRn, 11-28, D-33
Literature, ordering, 1-5, 1-8
LOCK#, 6-34–6-35
lockout sequence, 17-4
INTCFG, 5-10, 9-19, D-34
INTCFG register, 9-19
OCW1, 9-25, D-40
OCW1 register, 9-25
OCW2, 9-26, D-41
OCW2 register, 9-26
OCW3, 9-27, D-42
OCW3 register, 9-27
P3CFG register, 9-18
POLL, 9-28, D-49
M
Manual contents, summary, 1-1–1-2
Measurements, defined, 1-3
Misaligned data transfers, 6-9
Mode, 12-22
N
POLL register, 9-28
Naming conventions, 1-3–1-4
Non-page mode, 15-13
Nonspecific EOI command, 9-14
Notational conventions, 1-3–1-4
Numbers, conventions, 1-3
register addresses, 4-16, 4-17, D-2, D-3
registers, 9-15–9-17
signals, 9-5
spurious interrupts, 9-30
Interrupt priority, 9-6–9-8
Interrupt service routine, 6-23
Interrupts and exceptions, relative priority, 7-7
O
Operand alignment
aligned, 6-9
misaligned, 6-9
Operating mode, 9-8
J
JTAG reset, 8-12
JTAG test-logic unit, 18-1–18-14
block diagram, 18-2
P
design considerations, 18-14
operation, 18-3–18-9
Page mode, 15-12
Paging Unit, 3-4, 3-5
PC/AT Address, defined, 1-4
PC/AT system architecture, departures from, B-1
Performance, 2-1
Peripherals, internal
configuring, 5-3–5-37
DOS compatible, 4-2
boundary-scan register, 18-9
bypass register, 18-8
identification code register, 18-8
instruction register, 18-7
test access port controller, 18-4–18-6
instructions, 18-7–18-8
state diagram, 18-6
embedded application-specific, 4-2
Index-5
INTEL386™ EX MICROPROCESSOR USER’S MANUAL
register locations, 4-5, 4-15
Peripherals, summary, 2-3
Physical address space, 3-1
Pin configuration, 5-23
PINCFG, 5-24, 10-23, 11-17, 12-31, 13-17,
14-15, D-46
Pin descriptions, A-1–A-10
Pin states after reset and during idle, powerdown,
and hold, A-9
interrupt control unit, 9-32
RCU, 15-6–15-10
REMAPCFG example, 4-8
serial I/O unit, 11-15–11-32
SSIO, 13-17–13-25
timer/counter unit, 10-20–10-33
watchdog timer unit, 17-7–17-12
Programming considerations
chip-select unit, 14-22
DMA controller, 12-50
serial I/O unit, 11-32
Pipelined instructions, defined, 3-2
Port configuration
P1CFG, 5-25, 11-18, D-43
P2CFG, 5-26, 11-19, 14-16, D-44
P3CFG, 5-27, 9-18, 10-22, 11-20, D-45
PnCFG, 16-7
timer/counter unit, 10-33
Protected mode, 9-8
Protection Test Unit, 3-5
PSCLK, 8-1–8-2, 8-7, 10-1, 10-3, 10-21, 13-1,
13-5, 13-18
PnDIR, 16-8, D-47
PnLTC, 16-8, D-48
PSCLK frequency
PnPIN, 16-9, D-48
Controlling, 8-7
PORT92, 5-22, D-50
PSRAM, 15-11
PORT92
register addresses, 4-17, D-3
Power management
R
RAS#-only refresh, 15-1, 15-12
RCU, See Refresh control unit
Ready logic, 6-10
controlling modes, 8-8, 17-4–17-6
logic, 8-3–8-7
programming
Real mode, 9-8
Refresh control unit, 15-1–15-16
bus arbitration, 15-5
configuring, 5-3
PWRCON, 8-8, 17-11, D-51
register addresses, 4-19, D-5
See also Idle mode, powerdown mode, system
management mode
connections, 15-3
design considerations, 15-11
dynamic memory control, 15-1
operation, 15-5
Powerdown mode
considerations, 8-13
SMM interaction with, 8-5
timing diagram, 8-11
Powerup
Built-in self-test, 8-12
JTAG reset, 8-12
Prefetch Queue, 3-4
Priority of exceptions and interrupts, 7-7
Programmed operating mode, 9-8
Programming
chip-select unit, 14-13–14-20
clock and power management unit, 8-7–8-10
DMA controller, 12-28–12-51
ESE bit, 4-8
overview, 15-2–15-5
programming, 15-6–15-10
RFSADD, 15-10, D-54
RFSADD register, 15-10
RFSBAD, 15-9, D-54
RFSBAD register, 15-9
RFSCIR, 15-7, D-55
RFSCIR register, 15-7
RFSCON, 15-8, D-55
RFSCON register, 15-8
refresh addresses, 15-4
refresh intervals, 15-4
refresh methods, 15-1
Index-6
INDEX
register addresses, 4-18, D-4
registers, 15-6
IDCODE, 18-2, D-31
Identifier, 7-15
signals, 15-4
Register
naming conventions, 1-4
organization, 4-1–4-20
IERn, 11-15, 11-27, D-32
IIRn, 11-16, 11-28, D-33
INTCFG, 5-10, 9-19, D-34
IR, 18-7, D-35
Register bits, notational conventions, 1-4
Register names, notational conventions, 1-4
Register, status during SMM, 7-3
Registers
LCRn, 11-15, 11-25, D-36
LSRn, 11-15, 11-26, D-37
MCRn, 11-16, 11-29, 11-30, D-38
MSRn, 11-16, 11-31, D-39
OCW1, 9-25, D-40
BOUND, 18-2
BYPASS, 18-2
OCW2, 9-26, D-41
CLKPRS, 8-6, 8-7, 13-16, 13-19, D-7
Component and revision ID, 7-15
CSnADH, 14-14, 14-17, D-8
CSnADL, 14-14, 14-18, D-9
CSnMSKH, 14-14, 14-19, D-10
CSnMSKL, 14-14, 14-20, D-11
DLHn, 11-15, 11-22, D-12
OCW3, 9-27, D-42
P1CFG, 5-25, 11-15, 11-18, D-43
P2CFG, 5-26, 11-15, 11-19, 14-14, 14-16,
D-44
P3CFG, 5-27, 9-18, 10-4, 10-22, 11-15,
11-20, D-45
PINCFG, 5-24, 10-4, 10-23, 11-15, 11-17,
12-28, 12-31, 13-16, 13-17, 14-14, 14-15,
D-46
DLLn, 11-15, 11-22, D-12
DMA0BYCn, 12-28, 12-33, D-24
DMA0REQn, 12-28, 12-33, D-24
DMA0TARn, 12-28, 12-33, D-24
DMA1BYCn, 12-28, 12-33, D-24
DMA1REQn, 12-28, 12-33, D-24
DMA1TARn, 12-28, 12-33, D-24
DMABSR, 12-29, 12-46, D-13
DMACFG, 5-6, 12-28, 12-32, D-14
DMACHR, 12-30, 12-47, D-15
DMACMD1, 12-28, 12-35, D-16
DMACMD2, 12-29, 12-37, D-17
DMAGRPMSK, 12-29, 12-45, D-18
DMAIEN, 12-30, 12-48, D-19
DMAIS, 12-30, 12-49, D-20
DMAMOD1, 12-29, 12-38, 12-39, D-21
DMAMOD2, 12-29, 12-40–12-41, D-22
DMAMSK, 12-29, 12-44, D-23
DMAOVFE, 12-30, 12-34
DMASRR, 12-29, 12-42, 12-43, D-26
DMASTS, 12-29, 12-36, D-27
ICW1, 9-20, D-28
PnCFG, 11-15, 16-6, 16-7
PnDIR, 16-6, 16-8, D-47
PnLTC, 16-6, 16-8, D-48
PnPIN, 16-6, 16-9, D-48
POLL, 9-28, D-49
PORT92, 5-22, D-50
Port92, 5-22
PWRCON, 8-6, 8-8, 17-11, D-51
RBRn, 11-15, 11-24, D-52
REMAPCFG, 4-6, 4-7, D-53
RFSADD, 15-10, D-54
RFSBAD, 15-9, D-54
RFSCIR, 15-7, D-55
RFSCON, 15-8, D-55
SCRn, 11-16, 11-32, D-56
SIOCFG, 5-17, 11-15, 11-21, 13-16, 13-18,
D-57
SMM revision ID, 7-15
SSIOBAUD, 13-16, 13-20, D-58
SSIOCON1, 13-16, 13-21, 13-22, D-59
SSIOCON2, 13-16, 13-23
SSIOCTR, 13-16, 13-21, D-60
SSIORBUF, 13-16, 13-25, D-60
ICW2, 9-21, D-29
ICW3, 9-22, 9-23, D-29, D-30
ICW4, 9-24, D-30
Index-7
INTEL386™ EX MICROPROCESSOR USER’S MANUAL
SSIOTBUF, 13-16, 13-24, D-61
TBRn, 11-15, 11-23, D-61
programming
accessing multiplexed registers, 11-16
considerations, 11-32
DLHn register, 11-22
DLLn register, 11-22
IERn register, 11-27
IIRn register, 11-28
LCRn, 11-25, D-36
LCRn register, 11-25
LSRn, 11-26, D-37
TMRCFG, 5-13, 10-4, 10-21, D-62
TMRCON, 10-4, 10-25, 10-28, 10-30, D-63
TMRn, 10-4, 10-26, 10-29, 10-32, D-64, D-65
UCSADH, 14-14, 14-17, D-8
UCSADL, 14-14, 14-18, D-9
UCSMSKH, 14-14, 14-19, D-10
UCSMSKL, 14-14, 14-20, D-11
WDTCLR, 17-7
LSRn register, 11-26
WDTCNTH, 17-7, 17-8, D-67
WDTCNTL, 17-7, 17-8, D-67
WDTRLDH, 17-7, 17-10, D-68
WDTRLDL, 17-7, 17-10, D-68
WDTSTATUS, 17-7, 17-9, D-69
reload event, 17-4
Reserved bits, defined, 1-5
Reset
considerations, 8-11
MCRn, 11-30, D-38
MCRn register, 11-29–11-30
modem control signals, 11-29–11-30
MSRn, 11-31, D-39
MSRn register, 11-31
P1CFG register, 11-18
P2CFG register, 11-19
P3CFG register, 11-20
PINCFG register, 11-17
RBRn, 11-24, D-52
RBRn register, 11-24
CPU-only, B-4
Resume instruction (RSM), 7-15
RSM, See Resume instruction
SCRn register, 11-32
SIOCFG, 5-17, 11-21, 13-18, D-57
SIOCFG register, 11-21
TBRn, 11-23, D-61
S
Scratch pad registers
SCRn, 11-32, D-56
TBRn register, 11-23
register addresses, 4-19, 4-20, D-5, D-6
registers, 11-15–11-16
signals, 11-3
Segment Descriptor Cache, 3-5
Segmentation Unit, 3-4, 3-5
SERCLK, 8-1–8-2, 11-1, 11-4, 11-21, 13-1,
13-5, 13-18
Serial I/O unit, 11-1–11-45
block diagram, 11-2
configuring, 5-14
departure from PC/AT architecture, B-3
DMA service, 5-3–5-5
Set, defined, 1-5
Signal descriptions, A-1–A-10
Signal names, notational conventions, 1-4
SIO, See Serial I/O unit
SMM, See System management mode
SMM, see System Management Mode, 7-3
SMRAM, 7-2
chip-select unit support for, 7-12
state dump area, 7-14
Specific EOI command, 9-14
SSIO, See Synchronous serial I/O unit
Synchronous serial I/O unit, 13-1–13-25
configuring, 5-18
operation, 11-4–11-14
baud-rate generator, 11-4–11-5
data transmission process flow, 11-8
diagnostic mode, 11-12
interrupt sources, 11-13
modem control, 11-12
receiver, 11-9–11-10
transmitter, 11-6–11-8
design considerations, 13-25
DMA service, 5-3
master/slave mode arrangements, 13-2–13-3
overview, 11-1–11-3
Index-8
INDEX
operation, 13-5–13-15
baud-rate generator, 13-5–13-6
receiver, 13-12–13-15
transmitter, 13-6
overview, 13-1–13-4
programming, 13-16–13-25
CLKPRS register, 13-19
PINCFG, 13-17
SIOCFG register, 13-18
SSIOBAUD, 13-20, D-58
SSIOBAUD register, 13-20
SSIOCON1, 13-22, D-59
SSIOCON1 register, 13-21, 13-22, D-59
SSIOCON2 register, 13-23
SSIOCTR, 13-21, D-60
SSIOCTR register, 13-21
SSIORBUF, 13-25, D-60
SSIORBUF register, 13-25
SSIOTBUF, 13-24, D-61
SSIOTBUF register, 13-24
register addresses, 4-18, D-4
registers, 13-16
address configuration register, 4-6
address space, I/O for PC/AT systems, 4-2
addressing modes, 4-9
DOS-compatible mode, 4-9
enhanced DOS mode, 4-11
nonDOS mode, 4-11
nonintrusive DOS mode, 4-11
enabling/disabling expanded I/O space, 4-8
expanded I/O address space, 4-3
I/O address decoding techniques, 4-6
organization of peripheral registers, 4-5
overview, 4-1
peripheral register addresses, 4-15
peripheral registers, 4-2
processor core architecture, 4-2
programming
ESE bit, 4-8
REMAPCFG example, 4-8
T
TAP controller, 18-4
TAP Test Access Port, 18-1
TCU, See Timer/counter unit
Technical support, 1-7
Terminology, 1-4–1-5, Glossary-1–Glossary-5
Test access port, 18-1
Test-logic unit, See JTAG test-logic unit
Timer/counter unit, 10-1–10-33
configuring, 5-11
signals, 13-4
SIOCFG, 5-17, 11-21, 13-18, D-57
System management mode, 2-1, 7-1–7-15
CSU support, 7-12, 14-10
HALT restart, 7-9
hardware interface, 3-1, 7-1
SMI#, 7-1
SMIACT#, 7-2
hardware triggerable one-shot, See Mode 1
hardware-triggered strobe, See Mode 5
initial count values, 10-26
interrupt on terminal count, See Mode 0
mode 0, 10-6–10-8
SMRAM state dump area, 7-14
I/O restart, 7-1
identifier registers, 3-6, 7-15
interaction with idle and powerdown, 8-5
overview, 7-1
basic operation, 10-7
disabling the count, 10-7
writing a new count, 10-8
priority, 7-7
resume instruction, 7-15
SMI# interrupt, 7-3, 7-11–7-15
during HALT cycle, 7-8
during I/O instruction, 7-9
during SMM handler, 7-10
HALT during SMM handler, 7-11
SMI# during SMM operation, 7-12
SMRAM, 7-2
mode 1, 10-8–10-10
basic operation, 10-9
retriggering the one-shot, 10-9
writing a new count, 10-10
mode 2, 10-10–10-12
basic operation, 10-11
disabling the count, 10-11
writing a new count, 10-12
state dump area, 7-14–7-15
System register organization, 4-1
Index-9
INTEL386™ EX MICROPROCESSOR USER’S MANUAL
mode 3, 10-12–10-15
counter mode 2, 10-11, 10-12
basic operation, 10-13–10-14
basic operation (odd count), 10-14
disabling the count, 10-14
writing a new count, 10-15
mode 4, 10-16–10-17
counter mode 3, 10-13, 10-14, 10-15
counter mode 4, 10-16, 10-17
counter mode 5, 10-18, 10-19
DMA transfer, 12-9, 12-11, 12-21
entering and leaving idle mode, 8-9
entering and leaving powerdown mode, 8-11
HALT cycle, 6-27
interrupt acknowledge cycle, 6-25, 9-29
JTAG test-logic unit, 18-12, 18-13
LOCK# signal during pipelining, 6-35
nonpipelined read cycle, 6-15
nonpipelined write cycle, 6-18
pipelined cycles, 6-21
basic operation, 10-16
disabling the count, 10-17
writing a new count, 10-17
mode 5, 10-18–10-19
basic operation, 10-18
retriggering the strobe, 10-19
writing a new count, 10-19
operation, 10-5–10-19
operations caused by GATEn, 10-6
refresh cycle during HOLD/HLDA, 6-30
SSIO receiver, 13-15
overview, 10-1–10-4
programming
considerations, 10-33
SSIO transmitter, 13-11
initializing the counters, 10-24–10-25,
D-63
input and output signals, 10-20–10-23
reading the counter, 10-27–10-33
U
Units of measure, defined, 1-3
V
counter-latch command, 10-27
read-back command, 10-30
Virtual-86 mode, 9-8
simple read, 10-27
TMRCFG, 5-13, 10-21, D-62
W
Watchdog timer unit, 17-1–17-16
block diagram, 17-2
design considerations, 17-12
disabling the WDT, 17-6
lockout sequence, 17-4
operation, 17-3–17-4
during idle mode, 8-5
overview, 17-1–17-2
TMRCON, 10-25, 10-28, D-63
TMRn, 10-29, 10-32, D-64, D-65
writing the counters, 10-26
rate generator, See Mode 2
read-back commands, multiple, 10-33
register addresses, 4-16, D-2
registers, 10-4
TMRCON, 10-30
TMRn, 10-26
signals, 10-3
software-triggered strobe, See Mode 4
square wave, See Mode 3
Timing, 8-9
programming, 17-5–17-6
bus monitor mode, 17-5
general-purpose timer mode, 17-4
software watchdog mode, 17-5
WDTCNTH, 17-8, D-67
WDTCNTL, 17-8, D-67
WDTRLDH, 17-10, D-68
WDTRLDL, 17-10, D-68
WDTSTATUS, 17-9, D-69
register addresses, 4-18, D-4
Timing diagram
basic external bus cycles, 6-6
basic internal and external bus cycles, 6-12
basic refresh cycle, 6-29
BS8 cycle, 6-33
counter mode 0, 10-7
counter mode 1, 10-9, 10-10
Index-10
INDEX
registers, 17-7
WDTCLR, 17-7
WDTCNTH, 17-7
WDTCNTL, 17-7
WDTRLDH, 17-7
WDTRLDL, 17-7
WDTSTATUS, 17-7
reload event, 17-4
signals, 17-3
WDT, See Watchdog timer unit
Worksheets
peripheral configuration, 5-34
pin configuration, 5-34
World Wide Web, 1-7
Index-11
|
IBM 21L4365 User Manual
IBM THINKCENTRE 8145 User Manual
JVC AV 28R57SK AV 32R57SK User Manual
JVC CU VD20AG User Manual
KitchenAid KECC568MBL00 User Manual
Lenovo 9417 HE2 User Manual
Lenovo P97 User Manual
Lenovo Thinkvision Lt1421 Widescreen Lcd Monitor 1452DS6 User Manual
LG Electronics LG Life Good Computer Monitor L227WTP User Manual
Motorola C331 User Manual