1 |
2 |
jamieiles |
= s80x86 Core Development Guide
|
2 |
|
|
Jamie Iles
|
3 |
|
|
|
4 |
|
|
:source-highlighter: coderay
|
5 |
|
|
|
6 |
|
|
= Developer's Guide
|
7 |
|
|
|
8 |
|
|
== Building
|
9 |
|
|
|
10 |
|
|
=== System Requirements
|
11 |
|
|
|
12 |
|
|
- Docker, tested with version 1.12.1.
|
13 |
|
|
|
14 |
|
|
=== Quick Start Build
|
15 |
|
|
|
16 |
|
|
The `scripts/build` script provides everything that is needed to quickly build
|
17 |
|
|
and test the project. On the first run, the script will build the required
|
18 |
|
|
Docker images from `docker/build/Dockerfile`, configure and build the project
|
19 |
|
|
and run all of the built-in tests.
|
20 |
|
|
|
21 |
|
|
[source,bash]
|
22 |
|
|
----
|
23 |
|
|
./scripts/build
|
24 |
|
|
----
|
25 |
|
|
|
26 |
|
|
=== Build Environments
|
27 |
|
|
|
28 |
|
|
The s80x86 project uses Docker to provide convenient build environments. This
|
29 |
|
|
means that it is possible to build and test the design on any Linux system
|
30 |
|
|
with Docker, regardless of distribution. The Docker images used internally
|
31 |
|
|
are all based on Ubuntu 16.04 LTS.
|
32 |
|
|
|
33 |
|
|
There are two primary build environments:
|
34 |
|
|
|
35 |
|
|
- s80x86-build
|
36 |
|
|
- s80x86-dev
|
37 |
|
|
|
38 |
|
|
Each build environment has a script in `docker` to enter the container and
|
39 |
|
|
build it if not already built. A 'ccache' cache is created in
|
40 |
|
|
`_build/.ccache` that persists across container runs to increase build
|
41 |
|
|
performance.
|
42 |
|
|
|
43 |
|
|
's80x86-build' is a minimal Ubuntu 16.04 LTS container with the dependencies for
|
44 |
|
|
building the project and running the tests. This environment runs everything
|
45 |
|
|
as the current user's uid/gid to preserve file permissions outside of the
|
46 |
|
|
container and bind-mounts the project directory into `/build`. The
|
47 |
|
|
's80x86-dev' container is intended for developing the project itself and
|
48 |
|
|
includes extra packages and convenience scripts to make developing easier.
|
49 |
|
|
Unlike the 's80x86-build' container, this container bind mounts `/home` from
|
50 |
|
|
the host into the container for convenience and includes packages like GTKWave
|
51 |
|
|
for viewing waveforms. It is recommended to use the Docker build environments
|
52 |
|
|
for all builds and developments as those are used for the baseline development
|
53 |
|
|
and verification.
|
54 |
|
|
|
55 |
|
|
=== CI Build Scripts
|
56 |
|
|
|
57 |
|
|
Several build scripts are included suitable for use in a continuous
|
58 |
|
|
integration environment.
|
59 |
|
|
|
60 |
|
|
*scripts/ci/unittest* builds the project from scratch and then runs all of the
|
61 |
|
|
unit tests, producing JUnit XML suitable for importing into the CI test runner
|
62 |
|
|
history.
|
63 |
|
|
|
64 |
|
|
*scripts/ci/gcov-coverage* performs the same steps as `scripts/ci/coverage`
|
65 |
|
|
but produces Cobertura compatible coverage information that can be read into
|
66 |
|
|
Jenkins or other CI systems supporting this format.
|
67 |
|
|
|
68 |
|
|
=== Build Configurations
|
69 |
|
|
|
70 |
|
|
The CMake based build system supports the following build configurations and
|
71 |
|
|
can be selected by passing `-DCMAKE_BUILD_TYPE=`'CONFIGURATION' to `cmake`.
|
72 |
|
|
|
73 |
|
|
*Release* is optimized for performance, no debug information.
|
74 |
|
|
|
75 |
|
|
*Debug* enables debug information in all C/{cpp} executables, and tracing of
|
76 |
|
|
Verilog models which will write VCD files for each test run.
|
77 |
|
|
|
78 |
|
|
*Coverage* builds everything for coverage including {cpp} and Verilog.
|
79 |
|
|
|
80 |
|
|
=== Configuration Options
|
81 |
|
|
|
82 |
|
|
There are some CMake configuration options that control features:
|
83 |
|
|
|
84 |
|
|
- *-DS80X86_TRAP_ESCAPE* controls whether the escape fault is taken on an
|
85 |
|
|
escape opcode. This defaults to off for DOS compatibility in which case
|
86 |
|
|
escape opcodes act like a NOP.
|
87 |
|
|
|
88 |
|
|
== Simulator
|
89 |
|
|
|
90 |
|
|
The simulator can run either the software simulation (SoftwareCPU) or the RTL
|
91 |
|
|
model (RTLCPU). Typing `^]` will cause the simulator to exit.
|
92 |
|
|
|
93 |
|
|
----
|
94 |
|
|
Options:
|
95 |
|
|
-h [ --help ] print this usage information and exit
|
96 |
|
|
-b [ --backend ] arg the simulator backend module to use, either SoftwareCPU
|
97 |
|
|
or RTLCPU, default SoftwareCPU
|
98 |
|
|
-r [ --restore ] arg restore file to load from
|
99 |
|
|
-s [ --save ] arg save file to write from
|
100 |
|
|
--bios arg the bios image to use
|
101 |
|
|
--diskimage arg the boot disk image
|
102 |
|
|
----
|
103 |
|
|
|
104 |
|
|
The save and restore functionality are a useful tool for debugging - the
|
105 |
|
|
entire simulation state can be saved at exit and restored later making it
|
106 |
|
|
possible to checkpoint once the system is in a known state and then repeatedly
|
107 |
|
|
debug without waiting for the system to return to the same state.
|
108 |
|
|
Importantly, this is transferable between backends, so it is possible to boot
|
109 |
|
|
the system and run an application with the `SoftwareCPU` and then exit and
|
110 |
|
|
restart with the `RTLCPU` to greatly reduce time taken to get to the
|
111 |
|
|
interesting debug point.
|
112 |
|
|
|
113 |
|
|
== Microcode
|
114 |
|
|
|
115 |
|
|
The microcode is stored in `rtl/microcode` where microcode files have the
|
116 |
|
|
`.us` suffix. The microassembler first passes these files through the C
|
117 |
|
|
preprocessor to allow inclusion of other files and creating macros.
|
118 |
|
|
|
119 |
|
|
=== Directives
|
120 |
|
|
|
121 |
|
|
Directives are used to provide information to the microassembler about
|
122 |
|
|
microcode layout without actually generating microinstructions.
|
123 |
|
|
|
124 |
|
|
.Microassembler Directives
|
125 |
|
|
[cols="3,7"]
|
126 |
|
|
|===
|
127 |
|
|
| Name | Description
|
128 |
|
|
| .opcode NUM
|
129 |
|
|
| The `.opcode NUM` directive tells the microassembler to insert the next
|
130 |
|
|
microinstruction at address `NUM` in the microprogram. This is used for the
|
131 |
|
|
first 256 opcodes so that efficient dispatch can be performed by jumping to
|
132 |
|
|
the address corresponding to the value of the opcode.
|
133 |
|
|
| .auto_address
|
134 |
|
|
| Returns the address assignment to automatically assigned addresses after
|
135 |
|
|
using the `.opcode` directive.
|
136 |
|
|
|===
|
137 |
|
|
|
138 |
|
|
=== Microinstruction Fields
|
139 |
|
|
|
140 |
|
|
==== Jumps
|
141 |
|
|
|
142 |
|
|
.Microcode Jump Types
|
143 |
|
|
[cols="3,7"]
|
144 |
|
|
|===
|
145 |
|
|
| Name | Description
|
146 |
|
|
| jmp_rm_reg_mem LABEL a| tells the microassembler to generate a jump that will
|
147 |
|
|
jump to the label if the Mod R/M decoder indicates a register operand in the
|
148 |
|
|
R/M field and the label + 1 if the R/M field encodes a memory operand. For
|
149 |
|
|
example:
|
150 |
|
|
|
151 |
|
|
[source,asm]
|
152 |
|
|
----
|
153 |
|
|
jmp_rm_reg_mem foo_reg;
|
154 |
|
|
foo_reg:
|
155 |
|
|
next_instruction;
|
156 |
|
|
foo_mem:
|
157 |
|
|
next_instruction;
|
158 |
|
|
----
|
159 |
|
|
|
160 |
|
|
will jump to `foo_reg` if the R/M operand is a register operand and `foo_mem`
|
161 |
|
|
if the R/M operand is a memory operand. The two microinstructions must be
|
162 |
|
|
adjacent with the register based instruction appearing first.
|
163 |
|
|
|
164 |
|
|
| jmp_opcode | Takes the value of the opcode that was fetched and jumps to
|
165 |
|
|
that address as an absolute value, used in combination with the `.at`
|
166 |
|
|
directive to implement opcodes.
|
167 |
|
|
|
168 |
|
|
| jmp_dispatch_reg LABEL | Using the reg field of the mod r/m byte, jump to
|
169 |
|
|
the target address + the value of the reg field. Used for implementing
|
170 |
|
|
different instructions that share the same opcode.
|
171 |
|
|
|
172 |
|
|
| jmp_if_not_rep LABEL | Jump to the target address if the string instruction
|
173 |
|
|
does not have a rep prefix, otherwise continue execution at the next
|
174 |
|
|
incremented address. Only valid on string instructions that may be combined
|
175 |
|
|
with a rep prefix.
|
176 |
|
|
|
177 |
|
|
| jmp_if_zero LABEL | Jump to the target address if the Z flag is set,
|
178 |
|
|
otherwise continue with the adjacent instruction. Note that this uses the
|
179 |
|
|
flags register and not the combinational flags output of the current ALU
|
180 |
|
|
operation.
|
181 |
|
|
|
182 |
|
|
| jmp_rb_zero LABEL | Jump to the target address if RB value is zero,
|
183 |
|
|
otherwise continue with the adjacent instruction.
|
184 |
|
|
|
185 |
|
|
| jmp_if_rep_not_taken LABEL | Check the condition for the current rep prefix
|
186 |
|
|
and jump to the target if the termination condition is not met, otherwise
|
187 |
|
|
execute the adjacent instruction. Only valid when there is a rep prefix
|
188 |
|
|
present.
|
189 |
|
|
|
190 |
|
|
| jmp_if_taken LABEL | Jump to the target if the jump instruction has the
|
191 |
|
|
condition met, otherwise continue with the adjacent instruction. This is
|
192 |
|
|
only valid for jump instructions, and INTO.
|
193 |
|
|
|
194 |
|
|
| jmp_loop_done LABEL | Jump to the target if loop has completed. The loop
|
195 |
|
|
counter is loaded from the 5 MSB's of the immediate and decremented on each
|
196 |
|
|
iteration. This is only used for the `enter` instruction.
|
197 |
|
|
|
198 |
|
|
| jmp LABEL | An unconditional jump, will always transfer control to LABEL.
|
199 |
|
|
|===
|
200 |
|
|
|
201 |
|
|
=== Data Sources
|
202 |
|
|
|
203 |
|
|
.Microcode Data Sources
|
204 |
|
|
[cols="3,7"]
|
205 |
|
|
|===
|
206 |
|
|
| Name | Description
|
207 |
|
|
| ra_sel | Which general purpose register to fetch for RA. Note that register
|
208 |
|
|
fetches have a single cycle latency. Only valid when `ra_modrm_rm_reg` is
|
209 |
|
|
not set.
|
210 |
|
|
| rb_cl | Set to use the value of `CL` for RB after a single cycle of latency,
|
211 |
|
|
used primarily for shifts.
|
212 |
|
|
| segment | Set the default segment for the memory operation or segment
|
213 |
|
|
register read. This is the default segment and may be overriden with a
|
214 |
|
|
segment override prefix unless `segment_force` is also set.
|
215 |
|
|
| a_sel a|
|
216 |
|
|
Selects which operand source to use for the internal A bus:
|
217 |
|
|
|
218 |
|
|
- RA: the fetched RA GPR value.
|
219 |
|
|
- IP: the instruction pointer of the next instruction.
|
220 |
|
|
- MAR: the contents of the memory address register.
|
221 |
|
|
- MDR: the contents of the memory data register.
|
222 |
|
|
| b_sel a|
|
223 |
|
|
Selects which operand source to use for the internal B bus:
|
224 |
|
|
|
225 |
|
|
- RB: the fetched RB GPR value.
|
226 |
|
|
- IMMEDIATE: an immediate value, either from the immediate reader or from
|
227 |
|
|
the constant pool if a microinstruction defined constant is being used.
|
228 |
|
|
- SR: the fetched segment register value.
|
229 |
|
|
- TEMP: the contents of the temporary register.
|
230 |
|
|
| immediate | The immediate constant to use. This forms a constant pool in
|
231 |
|
|
the microcode and can be used for operations such as fetching exception
|
232 |
|
|
handler addresses, incrementing/decrementing pointers etc.
|
233 |
|
|
| mar_wr_sel a| Selects the source of the value to be written to the memory
|
234 |
|
|
address register:
|
235 |
|
|
|
236 |
|
|
- EA: the effective address calculated by the mod r/m decoder.
|
237 |
|
|
- Q: the Q bus driven by the ALU.
|
238 |
|
|
|===
|
239 |
|
|
|
240 |
|
|
=== Control Signals
|
241 |
|
|
|
242 |
|
|
.Microcode Control Signals
|
243 |
|
|
[cols="3,7"]
|
244 |
|
|
|===
|
245 |
|
|
| Name | Description
|
246 |
|
|
| next_instruction | Ends processing of the current instruction, will check
|
247 |
|
|
for pending interrupts, jump to the instruction dispatch address, update
|
248 |
|
|
CS:IP and reset any intermediate state.
|
249 |
|
|
| mar_write | Write the value of the `mar_wr_sel` source into the memory
|
250 |
|
|
address register.
|
251 |
|
|
| mdr_write | Write the value of the ALU output into the memory data register.
|
252 |
|
|
| mem_read | Perform a memory access with the specified segment and memory
|
253 |
|
|
address register value, reading into the memory data register. Note that
|
254 |
|
|
the segment register must have had the fetch initiated in the previous
|
255 |
|
|
instruction and should be held for the duration of the access. This field
|
256 |
|
|
will cause the microsequencer to stall until the access is complete. The
|
257 |
|
|
`width` field will specify the size of the access.
|
258 |
|
|
| mem_write | Perform a memory write, writing the contents of the memory data
|
259 |
|
|
register to the address specified by the fetched segment and the memory
|
260 |
|
|
address register. As with `mem_read`, the segment must have had the fetch
|
261 |
|
|
initiated in the previous instruction and held for the duration of this
|
262 |
|
|
instruction.
|
263 |
|
|
| segment_force | When used in combination with the `segment` field, this will
|
264 |
|
|
force that segment to be used unconditionally, ignoring any segment override
|
265 |
|
|
prefix.
|
266 |
|
|
| alu_op | The ALU operation to execute, see
|
267 |
|
|
"scripts/microassembler/microasm/types.py" for a full list of operations.
|
268 |
|
|
| update_flags a| A list of flags that should be written when performing an ALU
|
269 |
|
|
operation. If not specified, no flags will be update. For example:
|
270 |
|
|
|
271 |
|
|
[source,asm]
|
272 |
|
|
----
|
273 |
|
|
alu_op ADD, update_flags CF OF ZF AF
|
274 |
|
|
----
|
275 |
|
|
|
276 |
|
|
will update the carry, overflow, zero and adjust flags to the result of the
|
277 |
|
|
ALU operation.
|
278 |
|
|
| modrm_start | Trigger the mod r/m decoding. This will stall until complete
|
279 |
|
|
and calculate any effective addresses required.
|
280 |
|
|
| rd_sel_source a| The source of the destination register number:
|
281 |
|
|
|
282 |
|
|
- MODRM_REG: use the reg field of the mod r/m byte as the destination
|
283 |
|
|
register.
|
284 |
|
|
- MODRM_RM_REG: use the rm field of the mod r/m byte as the destination
|
285 |
|
|
register.
|
286 |
|
|
- MICROCODE_RD_SEL: use the rd_sel field of the instruction to select the
|
287 |
|
|
destination register.
|
288 |
|
|
|
289 |
|
|
| reg_wr_source a| Selects which result should be written to the destination
|
290 |
|
|
register:
|
291 |
|
|
|
292 |
|
|
- Q: the result of the ALU operation.
|
293 |
|
|
- QUOTIENT: the quotient of a division operation.
|
294 |
|
|
- REMAINDER: the remainder of a division operation.
|
295 |
|
|
|
296 |
|
|
| tmp_wr_en | Set to write the output of the ALU into the temporary register.
|
297 |
|
|
| tmp_wr_sel a| Select the source of the temporary register write:
|
298 |
|
|
|
299 |
|
|
- Q_LOW: the low 16-bits of the ALU output by default.
|
300 |
|
|
- Q_HIGH: the high 16-bits of the ALU output, only used for 16x16
|
301 |
|
|
multiplications.
|
302 |
|
|
|
303 |
|
|
| width | Selects the width of the operation. Defaults to 16-bit, but "width
|
304 |
|
|
W8" will perform byte operations for register read/write, memory read/write,
|
305 |
|
|
immediate fetch and ALU operations. "width WAUTO" will infer the width from
|
306 |
|
|
the opcode, where bit 0 being set indicates a 16-bit operation.
|
307 |
|
|
| load_ip | Causes the ALU result to be used as the new IP to be taken when
|
308 |
|
|
the next instruction is executed.
|
309 |
|
|
| read_immed | Triggers the immediate reader to read an immediate from the
|
310 |
|
|
instruction stream with the specified width.
|
311 |
|
|
| io | Combined with `mem_read`/`mem_write` to indicate that the operation
|
312 |
|
|
should use the I/O address space. This will cause the segment to be ignored
|
313 |
|
|
and the io pin to be asserted for the duration of this microinstruction.
|
314 |
|
|
| ext_int_inhibit | Used at the end of a microprogram, this flag indicates
|
315 |
|
|
that the microsequencer should not check for interrupts after this
|
316 |
|
|
instruction. This is used for instructions like `mov ss, bx` where the
|
317 |
|
|
following instruction would set `sp`.
|
318 |
|
|
| ext_int_inhibit | Used in string instructions to indicate that interrupts
|
319 |
|
|
may be serviced at this point.
|
320 |
|
|
|===
|
321 |
|
|
|
322 |
|
|
== Debug
|
323 |
|
|
|
324 |
|
|
The microsequencer provides a very simple way to implement on-chip debug. The
|
325 |
|
|
core has a number of signals to interface between a debug controller
|
326 |
|
|
(typically JTAG) and the microsequencer. These signals are all in the core
|
327 |
|
|
clock domain and will require synchronization with a debug controller in a
|
328 |
|
|
different clock domain.
|
329 |
|
|
|
330 |
|
|
The debug mechanism works by putting the core into a halt mode where it will
|
331 |
|
|
perform a tight loop in the microsequencer at which point other debug
|
332 |
|
|
operations can be issued. Operations are issued by running a microprogram at
|
333 |
|
|
a known address allowing more debug procedures to be added easily. To perform
|
334 |
|
|
a debug operation, the debug controller first halts the core by raising
|
335 |
|
|
`debug_seize` and waits for the core to enter the halted state with
|
336 |
|
|
`debug_stopped` asserted which will be at the end of the current microprogram.
|
337 |
|
|
Once stopped, the controller can write data to the temporary register if
|
338 |
|
|
required with `debug_wr_val` and `debug_wr_en` and then run the debug procedure
|
339 |
|
|
by writing the procedure address to `debug_addr` and asserting `debug_run` for
|
340 |
|
|
a single clock cycle.
|
341 |
|
|
|
342 |
|
|
=== Debug Signals
|
343 |
|
|
|
344 |
|
|
.Debug Interface Signals
|
345 |
|
|
[cols="2,1,1,3",options="header"]
|
346 |
|
|
|===
|
347 |
|
|
| Name | Width | Direction | Description
|
348 |
|
|
|
349 |
|
|
| debug_stopped | 1 | output | Asserted when the core is in a debug halt and
|
350 |
|
|
is ready for debug operations. The debug controller must not issue any
|
351 |
|
|
operations when `debug_stopped` is not asserted.
|
352 |
|
|
| debug_seize | 1 | input | Asserted by the controller to request that the core
|
353 |
|
|
enters debug mode. This may be deasserted once `debug_stopped` has been
|
354 |
|
|
asserted and then the run procedure executed to continue normal operation.
|
355 |
|
|
| debug_addr | 8 | input | The address of the debug procedure to execute, must
|
356 |
|
|
be written at the same time as `debug_run`. The core will run the procedure
|
357 |
|
|
at 100h + `debug_addr`.
|
358 |
|
|
| debug_run | 1 | input | Asserted by the debug controller to begin the debug
|
359 |
|
|
procedure specified in `debug_addr`.
|
360 |
|
|
| debug_wr_val | 16 | input | Asserted by the debug controller to write the
|
361 |
|
|
value in `debug_wr_val` into the temporary register.
|
362 |
|
|
| debug_wr_en | 1 | input | Asserted by the debug controller to write
|
363 |
|
|
`debug_wr_val` into the temporary register.
|
364 |
|
|
|===
|
365 |
|
|
|
366 |
|
|
=== Control and Reserved Debug Procedures
|
367 |
|
|
|
368 |
|
|
- *0x00*: resume execution. If `debug_seize` is held high then this will
|
369 |
|
|
single-step one instruction, otherwise run indefinitely until seized.
|
370 |
|
|
- *0x01 - 0x02: reserved for internal use, execution will yield undefined
|
371 |
|
|
behaviour.*
|
372 |
|
|
|
373 |
|
|
=== Data Transfer Debug Procedures
|
374 |
|
|
|
375 |
|
|
These debug procedures are used to transfer data between the debug controller
|
376 |
|
|
and the core.
|
377 |
|
|
|
378 |
|
|
.Data Transfer Debug Procedures
|
379 |
|
|
[cols=3*,options="header"]
|
380 |
|
|
|===
|
381 |
|
|
| Program Number
|
382 |
|
|
| Source
|
383 |
|
|
| Destination
|
384 |
|
|
| 0x03 | `AX` | `debug_val`
|
385 |
|
|
| 0x04 | `CX` | `debug_val`
|
386 |
|
|
| 0x05 | `DX` | `debug_val`
|
387 |
|
|
| 0x06 | `BX` | `debug_val`
|
388 |
|
|
| 0x07 | `SP` | `debug_val`
|
389 |
|
|
| 0x08 | `BP` | `debug_val`
|
390 |
|
|
| 0x09 | `SI` | `debug_val`
|
391 |
|
|
| 0x0a | `DI` | `debug_val`
|
392 |
|
|
| 0x0b | `ES` | `debug_val`
|
393 |
|
|
| 0x0c | `CS` | `debug_val`
|
394 |
|
|
| 0x0d | `SS` | `debug_val`
|
395 |
|
|
| 0x0e | `DS` | `debug_val`
|
396 |
|
|
| 0x0f | `IP` | `debug_val`
|
397 |
|
|
| 0x10 | `FLAGS` | `debug_val`
|
398 |
|
|
| 0x11 | `debug_val` | `IP`
|
399 |
|
|
| 0x12 | `debug_val` | `FLAGS`
|
400 |
|
|
| 0x13 | `debug_val` | `AX`
|
401 |
|
|
| 0x14 | `debug_val` | `CX`
|
402 |
|
|
| 0x15 | `debug_val` | `DX`
|
403 |
|
|
| 0x16 | `debug_val` | `BX`
|
404 |
|
|
| 0x17 | `debug_val` | `SP`
|
405 |
|
|
| 0x18 | `debug_val` | `BP`
|
406 |
|
|
| 0x19 | `debug_val` | `SI`
|
407 |
|
|
| 0x1a | `debug_val` | `DI`
|
408 |
|
|
| 0x1b | `debug_val` | `ES`
|
409 |
|
|
| 0x1c | `debug_val` | `CS`
|
410 |
|
|
| 0x1d | `debug_val` | `SS`
|
411 |
|
|
| 0x1e | `debug_val` | `DS`
|
412 |
|
|
| 0x1f | `debug_val` | `MAR`
|
413 |
|
|
| 0x20 | `debug_val` | `MDR`
|
414 |
|
|
| 0x21 | mem8[DS:MAR] | `debug_val`
|
415 |
|
|
| 0x22 | mem16[DS:MAR] | `debug_val`
|
416 |
|
|
| 0x23 | MDR | mem8[DS:MAR]
|
417 |
|
|
| 0x24 | MDR | mem16[DS:MAR]
|
418 |
|
|
| 0x25 | io8[MAR] | `debug_val`
|
419 |
|
|
| 0x26 | io16[MAR] | `debug_val`
|
420 |
|
|
| 0x27 | MDR | io8[MAR]
|
421 |
|
|
| 0x28 | MDR | io16[MAR]
|
422 |
|
|
|===
|
423 |
|
|
|
424 |
|
|
[NOTE]
|
425 |
|
|
====
|
426 |
|
|
All memory transfers implicitly use DS as the segment. To write outside of
|
427 |
|
|
the current data segment, save the value of DS, write it with the new value,
|
428 |
|
|
perform the access and then restore DS.
|
429 |
|
|
====
|
430 |
|
|
|
431 |
|
|
== FPGA JTAG
|
432 |
|
|
|
433 |
|
|
The DE0-Nano and DE0-CV boards use the Altera Virtual JTAG to implement a
|
434 |
|
|
debug bridge between the development machine and the FPGA design. This is not
|
435 |
|
|
a compliant JTAG TAP, but provides a reference implementation of implementing
|
436 |
|
|
a debug interface for the core.
|
437 |
|
|
|
438 |
|
|
The implementation uses a 2 bit instruction register and variable length data
|
439 |
|
|
register.
|
440 |
|
|
|
441 |
|
|
.JTAG Instruction Register Definitions
|
442 |
|
|
[cols=2*,options="header"]
|
443 |
|
|
|===
|
444 |
|
|
| Register
|
445 |
|
|
| Name
|
446 |
|
|
| 2'b00 | IDCODE
|
447 |
|
|
| 2'b01 | STATUS_CONTROL
|
448 |
|
|
| 2'b10 | DEBUG_VALUE
|
449 |
|
|
| 2'b11 | RUN_PROCEDURE
|
450 |
|
|
|===
|
451 |
|
|
|
452 |
|
|
=== IDCODE
|
453 |
|
|
|
454 |
|
|
The IDCODE register is a 32-bit register containing the device ID code. This
|
455 |
|
|
register is read-only, values shifted in are ignored.
|
456 |
|
|
|
457 |
|
|
=== STATUS_CONTROL
|
458 |
|
|
|
459 |
|
|
.JTAG STATUS_CONTROL Data Register
|
460 |
|
|
[cols="2,1,1,3",options="header"]
|
461 |
|
|
|===
|
462 |
|
|
| Field
|
463 |
|
|
| Bits
|
464 |
|
|
| Access
|
465 |
|
|
| Description
|
466 |
|
|
| RUN | [0:0] | R/W | Returns the current execution state of the CPU, "1"
|
467 |
|
|
indicates that the core is executing in normal mode. Write a "0" to enter
|
468 |
|
|
debug mode, polling this bit until it reflects that the core has stopped. To
|
469 |
|
|
restart the core, write a "1", and then run debug procedure "0".
|
470 |
|
|
| RESET | [1:1] | R/W | Core reset control, write "1" to start a reset, write
|
471 |
|
|
"0" to clear.
|
472 |
|
|
| RESERVED | [15:2] | RO | Reserved for future use.
|
473 |
|
|
| WRITE_ENABLE | [16] | WO | Write as "1" to write the value shifted in into
|
474 |
|
|
the core, otherwise the shifted value will be discarded.
|
475 |
|
|
|===
|
476 |
|
|
|
477 |
|
|
=== DEBUG_VALUE
|
478 |
|
|
|
479 |
|
|
.JTAG DEBUG_VALUE Data Register
|
480 |
|
|
[cols="2,1,1,3",options="header"]
|
481 |
|
|
|===
|
482 |
|
|
| Field
|
483 |
|
|
| Bits
|
484 |
|
|
| Access
|
485 |
|
|
| Description
|
486 |
|
|
| VALUE | [15:0] | RW | The value to be written to/read from the debug
|
487 |
|
|
controller.
|
488 |
|
|
| WRITE_ENABLE | [16] | R/W | For the value shifted in, if this is set to "1",
|
489 |
|
|
then the VALUE will be written into the debug controller, otherwise discarded.
|
490 |
|
|
For the value shifted out, if "1", then the VALUE field is valid. When
|
491 |
|
|
reading, this bit should be polled until it returns "1".
|
492 |
|
|
|===
|
493 |
|
|
|
494 |
|
|
=== RUN_PROCEDURE
|
495 |
|
|
|
496 |
|
|
.JTAG RUN_PROCEDURE Data Register
|
497 |
|
|
[cols="2,1,1,3",options="header"]
|
498 |
|
|
|===
|
499 |
|
|
| Field
|
500 |
|
|
| Bits
|
501 |
|
|
| Access
|
502 |
|
|
| Description
|
503 |
|
|
| VALUE | [7:0] | WO | The debug procedure to run. This is a write-only
|
504 |
|
|
field.
|
505 |
|
|
|===
|
506 |
|
|
|
507 |
|
|
== FPGA Reference Designs
|
508 |
|
|
|
509 |
|
|
=== DE0-Nano
|
510 |
|
|
|
511 |
|
|
To build the DE0-Nano design, configure the build with "-DBUILD_DE0_NANO=ON".
|
512 |
|
|
The build target "de0-nano" will build the FPGA, and "de0-nano-program" will
|
513 |
|
|
load the design into the FPGA via the Altera USB Blaster.
|
514 |
|
|
|
515 |
|
|
.DE0-Nano Memory Map
|
516 |
|
|
[cols="1,1,4",options="header"]
|
517 |
|
|
|===
|
518 |
|
|
| Start | End | Description
|
519 |
|
|
| 20'h0000 | 20'hfffff | SDRAM
|
520 |
|
|
|===
|
521 |
|
|
|
522 |
|
|
.DE0-Nano IO Port Map
|
523 |
|
|
[cols="1,1,4",options="header"]
|
524 |
|
|
|===
|
525 |
|
|
| Address | Width (bits) | Description
|
526 |
|
|
| 16'h0020 | 8 | PIC command register
|
527 |
|
|
| 16'h0021 | 8 | PIC data register
|
528 |
|
|
| 16'h0040 | 8 | PIT channel 0 data register
|
529 |
|
|
| 16'h0043 | 8 | PIT control register
|
530 |
|
|
| 16'hffec | 16 a| BIOS control register:
|
531 |
|
|
|
532 |
|
|
- [0]: BIOS ROM enabled.
|
533 |
|
|
|
534 |
|
|
| 16'hfff0 | 16 a| SPI control register:
|
535 |
|
|
|
536 |
|
|
- [15:10]: reserved.
|
537 |
|
|
- [9]: CS activate.
|
538 |
|
|
- [8:0]: clock divider.
|
539 |
|
|
| 16'hfff2 | 16 a| SPI transfer register:
|
540 |
|
|
|
541 |
|
|
- [15:9]: reserved.
|
542 |
|
|
- [8]: transfer busy.
|
543 |
|
|
- [7:0]: transfer data.
|
544 |
|
|
|
545 |
|
|
Writing to this register will initiate a one byte transfer. The CPU should
|
546 |
|
|
then poll until bit 8 is clear at which point [7:0] will contain the received
|
547 |
|
|
data.
|
548 |
|
|
|
549 |
|
|
| 16'ffff6 | 8 a| IRQ test register:
|
550 |
|
|
- [7]: write a 1 to raise NMI, 0 to clear NMI.
|
551 |
|
|
- [6:0]: write a 1 to raise interrupt N.
|
552 |
|
|
|
553 |
|
|
| 16'hfffa | 8 a| UART data register, write to transmit data, read to fetch
|
554 |
|
|
the received data.
|
555 |
|
|
| 16'hfffb | 8 a| UART status register:
|
556 |
|
|
|
557 |
|
|
- [7:2]: reserved.
|
558 |
|
|
- [1]: transmitter busy.
|
559 |
|
|
- [0]: receive data ready, cleared once the data register is read.
|
560 |
|
|
|
561 |
|
|
| 16'hfffc | 16 a| SDRAM configuration register:
|
562 |
|
|
|
563 |
|
|
- [15:1]: reserved.
|
564 |
|
|
- [0]: SDRAM configuration complete. The SDRAM should not be accessed until
|
565 |
|
|
this bit is set.
|
566 |
|
|
|
567 |
|
|
| 16'hfffe | 16 | LED register, writing will set the LED registers on the
|
568 |
|
|
board, a 1 is enabled, 0 is disabled.
|
569 |
|
|
|===
|
570 |
|
|
|
571 |
|
|
=== DE0-CV
|
572 |
|
|
|
573 |
|
|
To build the DE0-CV design, configure the build with "-DBUILD_DE0_CV=ON".
|
574 |
|
|
The build target "de0-cv" will build the FPGA, and "de0-cv-program" will
|
575 |
|
|
load the design into the FPGA via the Altera USB Blaster.
|
576 |
|
|
|
577 |
|
|
.DE0-CV Memory Map
|
578 |
|
|
[cols="1,1,4",options="header"]
|
579 |
|
|
|===
|
580 |
|
|
| Start | End | Description
|
581 |
|
|
| 20'h0000 | 20'hfffff | SDRAM
|
582 |
|
|
|===
|
583 |
|
|
|
584 |
|
|
.DE0-CV IO Port Map
|
585 |
|
|
[cols="1,1,4",options="header"]
|
586 |
|
|
|===
|
587 |
|
|
| Address | Width (bits) | Description
|
588 |
|
|
| 16'h0020 | 8 | PIC command register
|
589 |
|
|
| 16'h0021 | 8 | PIC data register
|
590 |
|
|
| 16'h0040 | 8 | PIT channel 0 data register
|
591 |
|
|
| 16'h0043 | 8 | PIT control register
|
592 |
|
|
| 16'h0060 | 8 a| PS/2 data register:
|
593 |
|
|
|
594 |
|
|
- [7:0]: read the head of the FIFO, 0 if no bytes available. Write to
|
595 |
|
|
transmit a byte, must only be written when bit 2 of the status register is
|
596 |
|
|
clear.
|
597 |
|
|
|
598 |
|
|
| 16'h0061 | 8 a| PS/2 control/status register:
|
599 |
|
|
|
600 |
|
|
- [7:3]: reserved
|
601 |
|
|
- [2]: transmit in progress, cleared when the transmitter is idle.
|
602 |
|
|
- [1]: a byte with a parity error was received and discarded. Cleared on
|
603 |
|
|
read.
|
604 |
|
|
- [0]: receive FIFO not empty/acknowledge. Write to acknowledge the last
|
605 |
|
|
received byte and pop it from the FIFO.
|
606 |
|
|
|
607 |
|
|
| 16'hffec | 16 a| BIOS control register:
|
608 |
|
|
|
609 |
|
|
- [0]: BIOS writable.
|
610 |
|
|
|
611 |
|
|
| 16'hfff0 | 16 a| SPI control register:
|
612 |
|
|
|
613 |
|
|
- [15:10]: reserved.
|
614 |
|
|
- [9]: CS activate.
|
615 |
|
|
- [8:0]: clock divider.
|
616 |
|
|
| 16'hfff2 | 16 a| SPI transfer register:
|
617 |
|
|
|
618 |
|
|
- [15:9]: reserved.
|
619 |
|
|
- [8]: transfer busy.
|
620 |
|
|
- [7:0]: transfer data.
|
621 |
|
|
|
622 |
|
|
Writing to this register will initiate a one byte transfer. The CPU should
|
623 |
|
|
then poll until bit 8 is clear at which point [7:0] will contain the received
|
624 |
|
|
data.
|
625 |
|
|
|
626 |
|
|
| 16'ffff6 | 8 a| IRQ test register:
|
627 |
|
|
- [7]: write a 1 to raise NMI, 0 to clear NMI.
|
628 |
|
|
- [6:0]: write a 1 to raise interrupt N.
|
629 |
|
|
|
630 |
|
|
| 16'hfffa | 8 a| UART data register, write to transmit data, read to fetch
|
631 |
|
|
the received data.
|
632 |
|
|
| 16'hfffb | 8 a| UART status register:
|
633 |
|
|
|
634 |
|
|
- [7:2]: reserved.
|
635 |
|
|
- [1]: transmitter busy.
|
636 |
|
|
- [0]: receive data ready, cleared once the data register is read.
|
637 |
|
|
|
638 |
|
|
| 16'hfffc | 16 a| SDRAM configuration register:
|
639 |
|
|
|
640 |
|
|
- [15:1]: reserved.
|
641 |
|
|
- [0]: SDRAM configuration complete. The SDRAM should not be accessed until
|
642 |
|
|
this bit is set.
|
643 |
|
|
|===
|
644 |
|
|
|
645 |
|
|
=== BIOS
|
646 |
|
|
|
647 |
|
|
The reference BIOS is a non-compliant BIOS for demonstration purposes. The
|
648 |
|
|
bios is built with the Mentor Graphics `ia16-elf` toolchain which is installed
|
649 |
|
|
in the standard docker images. The BIOS uses an SD card to emulate the floppy
|
650 |
|
|
drive and will boot the first sector of whatever is written to that SD card.
|
651 |
|
|
|
652 |
|
|
The BIOS is loaded at "f000:e000", and the BIOS stack grows down from
|
653 |
|
|
"f000:dffe". The UART is used for video and keyboard services.
|
654 |
|
|
|
655 |
|
|
== RTL Tests
|
656 |
|
|
|
657 |
|
|
The RTL tests are written in {cpp}, using Verilator to create {cpp} models of the
|
658 |
|
|
Verilog. For example, given a synchronous Fifo, the Verilator model can be
|
659 |
|
|
created using the Verilator CMake package:
|
660 |
|
|
|
661 |
|
|
[source,cmake]
|
662 |
|
|
----
|
663 |
|
|
include(Verilator)
|
664 |
|
|
add_library(verilator STATIC ${VERILATOR_LIB_SOURCES})
|
665 |
|
|
verilate(Fifo ${CMAKE_CURRENT_SOURCE_DIR}/Fifo.v)
|
666 |
|
|
----
|
667 |
|
|
|
668 |
|
|
This will generate a `verilator` library containing the common Verilator
|
669 |
|
|
support functions, run Verilator on `Fifo.v` and generate a `VFifo` library
|
670 |
|
|
and `VFifo.h` header for inclusion in the test code. A templated wrapper
|
671 |
|
|
'VerilogTestbench' in `VerilogTestbench.h` provides convenient methods for
|
672 |
|
|
resetting and clocking the device under test along with running deferred and
|
673 |
|
|
clock edge events, tracing and coverage.
|
674 |
|
|
|
675 |
|
|
The device under test can then be encapsulated inside a class and used for
|
676 |
|
|
writing tests with Google Test. For example, wrapping the Verilog model:
|
677 |
|
|
|
678 |
|
|
[source,c++]
|
679 |
|
|
----
|
680 |
|
|
#include
|
681 |
|
|
|
682 |
|
|
#include "VerilogTestbench.h"
|
683 |
|
|
|
684 |
|
|
class FifoTestbench : public VerilogTestbench {
|
685 |
|
|
public:
|
686 |
|
|
FifoTestbench(VFifo *dut);
|
687 |
|
|
void push(uint32_t val);
|
688 |
|
|
uint32_t pop();
|
689 |
|
|
};
|
690 |
|
|
|
691 |
|
|
FifoTestbench::FifoTestbench(VFifo *dut)
|
692 |
|
|
: VerilogTestbench(dut)
|
693 |
|
|
{
|
694 |
|
|
dut->wr_en = 0;
|
695 |
|
|
dut->wr_data = 0LU;
|
696 |
|
|
dut->rd_en = 0;
|
697 |
|
|
}
|
698 |
|
|
|
699 |
|
|
void FifoTestbench::push(uint32_t val)
|
700 |
|
|
{
|
701 |
|
|
dut->wr_data = val;
|
702 |
|
|
dut->wr_en = 1;
|
703 |
|
|
cycle();
|
704 |
|
|
dut->wr_en = 0;
|
705 |
|
|
}
|
706 |
|
|
|
707 |
|
|
uint32_t FifoTestbench::pop()
|
708 |
|
|
{
|
709 |
|
|
dut->rd_en = 1;
|
710 |
|
|
cycle();
|
711 |
|
|
dut->rd_en = 0;
|
712 |
|
|
|
713 |
|
|
return dut->rd_data;
|
714 |
|
|
}
|
715 |
|
|
----
|
716 |
|
|
|
717 |
|
|
Then a test can be written to exercise it:
|
718 |
|
|
|
719 |
|
|
[source,c++]
|
720 |
|
|
----
|
721 |
|
|
TEST(Fifo, ResetClears)
|
722 |
|
|
{
|
723 |
|
|
FifoTestbench tb;
|
724 |
|
|
|
725 |
|
|
for (uint32_t m = 0; m < 4; ++m)
|
726 |
|
|
tb.push(m);
|
727 |
|
|
|
728 |
|
|
ASSERT_FALSE(tb.dut->empty);
|
729 |
|
|
tb.reset();
|
730 |
|
|
ASSERT_TRUE(tb.dut->empty);
|
731 |
|
|
}
|
732 |
|
|
----
|
733 |
|
|
|
734 |
|
|
More complex tests that have deferred events such as reading from memory can
|
735 |
|
|
be written by adding events on positive+negative clock edges and running after
|
736 |
|
|
a number of cycles. `tests/rtl/TestPrefetch.cpp` uses a number of these
|
737 |
|
|
concepts. With the right abstractions it can be possible to type-parameterize
|
738 |
|
|
these test cases to run against pure software simulations and Verilog models.
|
739 |
|
|
|
740 |
|
|
= Programmer's Reference
|
741 |
|
|
|
742 |
|
|
== Interrupts
|
743 |
|
|
|
744 |
|
|
The CPU core implements the following exceptions. Traps are handled after the
|
745 |
|
|
instruction and the saved CS:IP points to the next instruction, faults are
|
746 |
|
|
restartable and the saved CS:IP points to the instruction that caused the
|
747 |
|
|
fault and so can be restarted. The core correctly handles multiple prefix
|
748 |
|
|
bytes during an interrupted string instruction.
|
749 |
|
|
|
750 |
|
|
.Exceptions
|
751 |
|
|
[cols="2,2,1,6",options="header"]
|
752 |
|
|
|===
|
753 |
|
|
| Name | Type | Number | Description
|
754 |
|
|
| Divide-by-zero | Trap | 0 | Raised when division by zero occurs or the result of
|
755 |
|
|
the division operation does not fit in the range of the destination register.
|
756 |
|
|
| Single-step | Trap | 1 | Raised when `TF` is set and the core steps an instruction.
|
757 |
|
|
| NMI | Interrupt | 2 | Non-maskable interrupt, raised when the `nmi` signal
|
758 |
|
|
has a negative to positive edge.
|
759 |
|
|
| INT | Interrupt | 3 | Normal interrupt, raised by the `int3` instruction.
|
760 |
|
|
| Overflow | Trap | 4 | Overflow, raised by an `into` instruction if `OF` is
|
761 |
|
|
set.
|
762 |
|
|
| Bound | Trap | 5 | Bounds check, raised when the `bound` instruction detects
|
763 |
|
|
an out-of-bounds address.
|
764 |
|
|
| Invalid Opcode | Trap | 6 a| Raised when an invalid opcode is executed.
|
765 |
|
|
Invalid opcodes do not include unimplemented opcodes or undocumented opcodes.
|
766 |
|
|
The opcodes that raise this trap are:
|
767 |
|
|
|
768 |
|
|
* 8'h0f
|
769 |
|
|
* 8'h63
|
770 |
|
|
* 8'h64
|
771 |
|
|
* 8'h65
|
772 |
|
|
* 8'h66
|
773 |
|
|
* 8'h67
|
774 |
|
|
* 8'hf1
|
775 |
|
|
* 8'hff or 8'hfe with /reg=7
|
776 |
|
|
* 8'h62 (bound) with a register operand
|
777 |
|
|
|
778 |
|
|
| Escape | Fault | 7 | Escape opcode, this will always be raised on an `esc`
|
779 |
|
|
instruction as no coprocessors are supported.
|
780 |
|
|
|===
|
781 |
|
|
|
782 |
|
|
== Instructions
|
783 |
|
|
|
784 |
|
|
The following tables list all of the supported instructions along with sizes
|
785 |
|
|
and timings for each. The timings are measured using the RTL simulation under
|
786 |
|
|
the following conditions:
|
787 |
|
|
|
788 |
|
|
* The prefetch FIFO contains the instruction and is padded with other bytes
|
789 |
|
|
to be full.
|
790 |
|
|
* The instruction and data busses are connected to an arbiter that gives
|
791 |
|
|
priority to data accesses.
|
792 |
|
|
* Memory accesses take a single cycle to complete.
|
793 |
|
|
|
794 |
|
|
Instructions with a MOD R/M byte have a variable length and execution time
|
795 |
|
|
depending on whether the addressing mode has a displacement. In the
|
796 |
|
|
instruction tables, the length reflects this, and if the addressing mode uses
|
797 |
|
|
a displacement, add one cycle per displacement byte for the execution timing.
|
798 |
|
|
|
799 |
|
|
Prefix bytes add one byte to the instruction length and one cycle to the
|
800 |
|
|
execution time, and for conflicting prefixes such as multiple segment
|
801 |
|
|
overrides, the last prefix is used.
|
802 |
|
|
|
803 |
|
|
<<<
|
804 |
|
|
|
805 |
|
|
include::instructions.asciidoc[]
|