Jacob Lifshay [Wed, 25 Oct 2023 03:18:48 +0000 (20:18 -0700)]
install pytest-subtests==0.11.0
Jacob Lifshay [Wed, 25 Oct 2023 03:18:27 +0000 (20:18 -0700)]
generate syscalls.json
Luke Kenneth Casson Leighton [Tue, 24 Oct 2023 17:16:06 +0000 (18:16 +0100)]
whitespace cleanup
Jacob Lifshay [Fri, 20 Oct 2023 01:57:00 +0000 (18:57 -0700)]
Revert "skip broken test"
requested by luke:
https://bugs.libre-soc.org/show_bug.cgi?id=1193#c1
This reverts commit
e0a4f19b2c90be84a77a4aa584c6d60e508d92f5.
Jacob Lifshay [Fri, 20 Oct 2023 01:18:51 +0000 (18:18 -0700)]
Revert "Revert "fix bug where pseudo-code assignments modify more than just the variable being assigned to""
we need copy_assign_rhs
See https://bugs.libre-soc.org/show_bug.cgi?id=1066
This reverts commit
bd3b54e83101217dc32da09083c6a3858fd7c600.
Jacob Lifshay [Fri, 20 Oct 2023 01:17:20 +0000 (18:17 -0700)]
Revert "fix bug introduced by having to revert unauthorized addition of"
we need copy_assign_rhs
See https://bugs.libre-soc.org/show_bug.cgi?id=1066
This reverts commit
9dab88318a2938f14873804d83bf85ef9ae2fb93.
Jacob Lifshay [Fri, 20 Oct 2023 01:00:55 +0000 (18:00 -0700)]
skip broken test
it wasn't obvious how to fix it, see https://bugs.libre-soc.org/show_bug.cgi?id=1193
Jacob Lifshay [Fri, 20 Oct 2023 00:49:04 +0000 (17:49 -0700)]
fill in manually verified expected state for TrapTestCase.case_2_kaivb_test
based on the Programming Note on left side of PowerISA v3.1B page 1289 (1315)
Jacob Lifshay [Fri, 20 Oct 2023 00:37:34 +0000 (17:37 -0700)]
format code
Jacob Lifshay [Mon, 23 Oct 2023 23:09:51 +0000 (16:09 -0700)]
reduce mmap BLOCK_SIZE to 1 << 28 so it works on armv7a
Dmitry Selyutin [Mon, 23 Oct 2023 20:32:57 +0000 (23:32 +0300)]
syscall: handle architecture aliases
Dmitry Selyutin [Mon, 23 Oct 2023 20:23:08 +0000 (23:23 +0300)]
syscall: handle arm and aarch64 architectures
Dmitry Selyutin [Mon, 23 Oct 2023 06:17:55 +0000 (09:17 +0300)]
test_syscall: hardcode MSR validation
Dmitry Selyutin [Sun, 22 Oct 2023 13:14:55 +0000 (16:14 +0300)]
test_syscall: check MSR; update expected PC
Dmitry Selyutin [Sun, 22 Oct 2023 06:44:37 +0000 (09:44 +0300)]
isa/test_runner: support initial_msr parameter
Dmitry Selyutin [Sun, 22 Oct 2023 06:29:50 +0000 (09:29 +0300)]
isa/caller: return from interrupt upon syscall emulation
Dmitry Selyutin [Fri, 20 Oct 2023 17:16:22 +0000 (20:16 +0300)]
test_syscall: provide code for future SPR checks
Dmitry Selyutin [Fri, 20 Oct 2023 17:15:12 +0000 (20:15 +0300)]
isa/caller: refactor sc logic
Dmitry Selyutin [Wed, 18 Oct 2023 17:11:51 +0000 (20:11 +0300)]
test_caller: introduce syscall tests
Dmitry Selyutin [Wed, 18 Oct 2023 15:32:37 +0000 (18:32 +0300)]
isa/caller: enable host-backed memory for scemu
Dmitry Selyutin [Wed, 18 Oct 2023 15:26:14 +0000 (18:26 +0300)]
test/runner: introduce use_syscall_emu parameter
Dmitry Selyutin [Wed, 18 Oct 2023 15:24:31 +0000 (18:24 +0300)]
isa/caller: introduce use_syscall_emu parameter
Dmitry Selyutin [Wed, 18 Oct 2023 17:06:17 +0000 (20:06 +0300)]
isa/test_runner: support additional parameters
Dmitry Selyutin [Wed, 18 Oct 2023 15:33:02 +0000 (18:33 +0300)]
isa/caller: remove redundant check
Dmitry Selyutin [Fri, 22 Sep 2023 19:08:10 +0000 (22:08 +0300)]
isa/caller: provide sc and scv instructions wrapper
Luke Kenneth Casson Leighton [Sat, 21 Oct 2023 18:10:23 +0000 (18:10 +0000)]
add extra comments to sc-rfid test
Luke Kenneth Casson Leighton [Fri, 20 Oct 2023 21:09:22 +0000 (22:09 +0100)]
add a test which does both sc and rfid, and does rudimentary
checking that they are executed in expected order by setting some GPRs.
a whole bunch of NOPs were added to get the assembler to start at 0xc00
with a jump right at the start. terrible hack but does the job.
Andrey Miroshnikov [Thu, 19 Oct 2023 10:05:05 +0000 (10:05 +0000)]
Added assert to check inner/outer results match
Andrey Miroshnikov [Thu, 19 Oct 2023 09:25:17 +0000 (09:25 +0000)]
Replace flatten func with builtin reduce()
Andrey Miroshnikov [Thu, 19 Oct 2023 06:49:14 +0000 (06:49 +0000)]
Readded the flatten func (removed accidentally)
Luke Kenneth Casson Leighton [Wed, 18 Oct 2023 21:05:52 +0000 (22:05 +0100)]
add expected results to "sc" instruction in TrapTestCase.
this demonstrates how "sc" is meant to work in standard (system) mode.
this *may* be exactly what TestIssuer does, it will have to be checked
Luke Kenneth Casson Leighton [Wed, 18 Oct 2023 21:04:33 +0000 (22:04 +0100)]
add SRR0 and SRR1 to list of special_regs in parser
which are not treated as "create on assign".
Luke Kenneth Casson Leighton [Wed, 18 Oct 2023 21:03:38 +0000 (22:03 +0100)]
annoying - call the TRAP() function in system.mdwn "sc" instruction.
setting NIA and MSR is tricky, it involves reading some english text
that is very unclear (Book III section 4.3.1 which then in turn says
"go to section 7.5 page 1076").
given that we are not implementing hypervisor or LEV=1/2/3 it is just
simpler to call TRAP(0xc00)
Luke Kenneth Casson Leighton [Wed, 18 Oct 2023 20:30:43 +0000 (21:30 +0100)]
fix bug introduced by having to revert unauthorized addition of
copy_assign_rhs
Luke Kenneth Casson Leighton [Wed, 18 Oct 2023 20:24:04 +0000 (21:24 +0100)]
Revert "fix bug where pseudo-code assignments modify more than just the variable being assigned to"
This reverts commit
4e701a851536bba6648779c183293ba75e7ea7b8.
adding copy_assign_rhs was added without authorization or discussion and
is damaging the simulator
Luke Kenneth Casson Leighton [Wed, 18 Oct 2023 20:21:48 +0000 (21:21 +0100)]
add sc test to TestTrapCases
Luke Kenneth Casson Leighton [Wed, 18 Oct 2023 19:56:40 +0000 (20:56 +0100)]
add test_caller_trap.py which stunningly actually works reasonably well
(TrapTestCase has only previously been run on TestIssuer)
Dmitry Selyutin [Fri, 22 Sep 2023 19:07:40 +0000 (22:07 +0300)]
power_enums: mention sc and scv instructions
Luke Kenneth Casson Leighton [Wed, 18 Oct 2023 14:52:30 +0000 (15:52 +0100)]
add "is_idle" capability to inorder.py so that after adding
instructions the pipeline continues to propagate
Shriya Sharma [Wed, 18 Oct 2023 14:47:51 +0000 (15:47 +0100)]
settting pushed_to_decode true
Shriya Sharma [Wed, 18 Oct 2023 14:32:36 +0000 (15:32 +0100)]
Added an extra unit test test_trace1
Andrey Miroshnikov [Tue, 17 Oct 2023 13:06:31 +0000 (13:06 +0000)]
Added pure python mat multiply (outer and inner product versions). Made result printing parametrisable.
Luke Kenneth Casson Leighton [Wed, 11 Oct 2023 11:49:11 +0000 (12:49 +0100)]
doh, use reduce on operator.add already
Luke Kenneth Casson Leighton [Wed, 11 Oct 2023 11:47:39 +0000 (12:47 +0100)]
print expected flattened matrix results
Andrey Miroshnikov [Wed, 11 Oct 2023 11:46:39 +0000 (11:46 +0000)]
Add assert
Luke Kenneth Casson Leighton [Wed, 11 Oct 2023 11:43:48 +0000 (12:43 +0100)]
remove cruft, make comments clearer
Andrey Miroshnikov [Wed, 11 Oct 2023 11:43:26 +0000 (11:43 +0000)]
Use flatten
Andrey Miroshnikov [Wed, 11 Oct 2023 11:41:16 +0000 (11:41 +0000)]
Add flatten function, print expected
Luke Kenneth Casson Leighton [Wed, 11 Oct 2023 11:23:02 +0000 (12:23 +0100)]
rename expected to results (actual results)
Luke Kenneth Casson Leighton [Wed, 11 Oct 2023 11:20:58 +0000 (12:20 +0100)]
store integer results in expected array
Luke Kenneth Casson Leighton [Wed, 11 Oct 2023 11:19:32 +0000 (12:19 +0100)]
simplify matmult test code
Luke Kenneth Casson Leighton [Wed, 11 Oct 2023 11:18:09 +0000 (12:18 +0100)]
whitespace
Luke Kenneth Casson Leighton [Wed, 11 Oct 2023 11:17:29 +0000 (12:17 +0100)]
whitespace
Andrey Miroshnikov [Wed, 11 Oct 2023 11:16:34 +0000 (11:16 +0000)]
Add pure python matrix mul function
Luke Kenneth Casson Leighton [Wed, 11 Oct 2023 11:11:57 +0000 (12:11 +0100)]
clarify prints and no conversion of integer input in maddld matrix test
Luke Kenneth Casson Leighton [Wed, 11 Oct 2023 11:05:59 +0000 (12:05 +0100)]
add stub (non-working) matrix multiply using maddld
Sadoon Albader [Wed, 11 Oct 2023 19:49:50 +0000 (22:49 +0300)]
add basic isacaller inlining to poly1305-donna.py
Luke Kenneth Casson Leighton [Wed, 11 Oct 2023 10:57:58 +0000 (11:57 +0100)]
accidentally commented-out matrix tests
Luke Kenneth Casson Leighton [Sun, 8 Oct 2023 13:57:48 +0000 (14:57 +0100)]
whitespace
Luke Kenneth Casson Leighton [Tue, 3 Oct 2023 14:37:37 +0000 (15:37 +0100)]
add error message showing which instructions have been barfed
Sadoon Albader [Tue, 3 Oct 2023 18:29:38 +0000 (21:29 +0300)]
fixed another serious bug, C should output to CSV
Sadoon Albader [Tue, 3 Oct 2023 18:22:49 +0000 (21:22 +0300)]
add rudementary test script
Sadoon Albader [Tue, 3 Oct 2023 18:19:52 +0000 (21:19 +0300)]
add python poly1305 test that uses random data input
Sadoon Albader [Tue, 3 Oct 2023 18:18:42 +0000 (21:18 +0300)]
fix huge bug with C poly1305 function call
Sadoon Albader [Tue, 3 Oct 2023 16:33:17 +0000 (19:33 +0300)]
added poly1305 random message tester
Luke Kenneth Casson Leighton [Tue, 3 Oct 2023 14:34:35 +0000 (15:34 +0100)]
optional read of "Description" in pagereader.py
Luke Kenneth Casson Leighton [Tue, 3 Oct 2023 14:29:55 +0000 (15:29 +0100)]
move repeated code block to mini function for reading indented lines
Luke Kenneth Casson Leighton [Sun, 1 Oct 2023 15:05:30 +0000 (16:05 +0100)]
whoops should be 5x3 comment not 5x3
Luke Kenneth Casson Leighton [Sun, 1 Oct 2023 10:12:55 +0000 (11:12 +0100)]
manually revert damaged caused by jacob to pseudocode parser
the purpose of the parser database is to preserve precisely and exactly
the data that is read in, such that it is possible to re-write it
precisely and exactly
jacob had destroyed that extremely important requirement by making
unauthorized modifications to this fundamental low-level code.
jacobs task is now to review the reversions and re-implement the
otherwise extremely valuable enhancements, but this time in a
way that listens to the project leader and administrators
Luke Kenneth Casson Leighton [Sun, 1 Oct 2023 09:45:43 +0000 (10:45 +0100)]
Revert "demo moving pseudocode to separate file"
This reverts commit
b5d9084971dd761683a3a164af24c673a608aa23.
Luke Kenneth Casson Leighton [Sun, 1 Oct 2023 09:40:43 +0000 (10:40 +0100)]
Revert "add support for pseudocode being a [[!inline]] directive"
This reverts commit
43152e91f4530ddaef5cef2614b41e022c57fced.
Luke Kenneth Casson Leighton [Sun, 1 Oct 2023 09:40:34 +0000 (10:40 +0100)]
Revert "ignore indented comments too"
This reverts commit
60f9f523f78cae9e357b61e6bc55ca1b323dfa14.
Luke Kenneth Casson Leighton [Sat, 30 Sep 2023 13:30:02 +0000 (14:30 +0100)]
skip blank lines in pagereader.py pprint_ops()
Luke Kenneth Casson Leighton [Fri, 29 Sep 2023 18:31:38 +0000 (19:31 +0100)]
code-comments
Luke Kenneth Casson Leighton [Fri, 29 Sep 2023 18:23:59 +0000 (19:23 +0100)]
moving the temp array (t) along, so that adding to y is the same size
in bigmul python-based code. idea is to make everything line up
and be as uniform as possible, reduce number of instructions to bare min.
,
Luke Kenneth Casson Leighton [Fri, 29 Sep 2023 17:46:51 +0000 (18:46 +0100)]
first attempt to create an Indexed Schedule, for bigmul powmod,
but it is not perfect. needs thought
Jacob Lifshay [Thu, 28 Sep 2023 02:51:35 +0000 (19:51 -0700)]
fix divmod
Jacob Lifshay [Thu, 28 Sep 2023 02:50:47 +0000 (19:50 -0700)]
in divmod algorithm log regexes that match against expected register values
Jacob Lifshay [Thu, 28 Sep 2023 02:48:50 +0000 (19:48 -0700)]
test python_divmod_algorithm
Jacob Lifshay [Thu, 28 Sep 2023 02:45:34 +0000 (19:45 -0700)]
format code
Jacob Lifshay [Thu, 28 Sep 2023 02:25:57 +0000 (19:25 -0700)]
log asmop to LogKind.InstrInOuts too since only printing `.long 0xFOOBAR` isn't very useful
Luke Kenneth Casson Leighton [Wed, 27 Sep 2023 19:13:16 +0000 (20:13 +0100)]
remove use of addc, use adde instead setting ca to zero.
eliminates one more unnecessary instruction.
Luke Kenneth Casson Leighton [Wed, 27 Sep 2023 18:44:43 +0000 (19:44 +0100)]
reduce 4-repeats of identical code down to 1 copy with indices in powmod.py
Luke Kenneth Casson Leighton [Wed, 27 Sep 2023 18:13:47 +0000 (19:13 +0100)]
add seeming-redundant addc/adde (actually part of big-mul-*add*)
which completes the pattern for REMAP transformation
Luke Kenneth Casson Leighton [Wed, 27 Sep 2023 15:19:32 +0000 (16:19 +0100)]
convert basic_pypowersim to hex rather than broken octal (?)
Luke Kenneth Casson Leighton [Wed, 27 Sep 2023 10:25:43 +0000 (11:25 +0100)]
code-cleanup, bit of comments, copyright, blah blah, link to bugreport
all preparation before doing code-morph on simple-demo to work out how
to demonstrate REMAP Indexed (then BigMul) viability
Luke Kenneth Casson Leighton [Wed, 27 Sep 2023 10:18:53 +0000 (11:18 +0100)]
add what is currently a duplicate of python_mul_algorithm, plan is to
morph python_mul_algorithm2 to be "REMAP"-friendly
Jacob Lifshay [Wed, 27 Sep 2023 04:41:58 +0000 (21:41 -0700)]
working on adding divmod 512x256 to 256x256
Jacob Lifshay [Wed, 27 Sep 2023 04:39:31 +0000 (21:39 -0700)]
log writing CA[32]/OV[32] for OP_ADD
Jacob Lifshay [Wed, 27 Sep 2023 03:34:39 +0000 (20:34 -0700)]
add unit test for mcrxrx
Jacob Lifshay [Wed, 27 Sep 2023 03:34:25 +0000 (20:34 -0700)]
fix mcrxrx
Jacob Lifshay [Wed, 27 Sep 2023 03:28:16 +0000 (20:28 -0700)]
fix concat when the first argument is a FieldSelectableInt
Jacob Lifshay [Wed, 27 Sep 2023 01:56:03 +0000 (18:56 -0700)]
fix wrong register in docs
Jacob Lifshay [Wed, 27 Sep 2023 00:23:59 +0000 (17:23 -0700)]
256x256-bit mul no longer broken since bug #1161 was fixed
Jacob Lifshay [Mon, 25 Sep 2023 22:57:23 +0000 (15:57 -0700)]
add MemMMap tests
Jacob Lifshay [Mon, 25 Sep 2023 22:29:25 +0000 (15:29 -0700)]
skip zero words when iterating words in MemMMap
Jacob Lifshay [Mon, 25 Sep 2023 21:41:49 +0000 (14:41 -0700)]
format src/openpower/decoder/isa/test_mem.py
Luke Kenneth Casson Leighton [Mon, 25 Sep 2023 13:59:33 +0000 (14:59 +0100)]
add basis of Context Manager for capturing which inputs and outputsa
are involved in a carry-roll-over math primitive.
also very useful to generate (automated) unit tests
Luke Kenneth Casson Leighton [Sun, 24 Sep 2023 18:07:15 +0000 (19:07 +0100)]
minor alteration of reporting hash in mini-test of poly1305-donna.py
Luke Kenneth Casson Leighton [Sun, 24 Sep 2023 18:06:28 +0000 (19:06 +0100)]
detect if add arg2 is greater than 7 and ignore it for poly1305 tracking.
this allows narrowing down of some data for test purposes