Commit graph

140 commits

Author SHA1 Message Date
MerryMage
a59c335b05 A64: Add options for detecting misaligned loads and stores 2020-04-22 21:04:23 +01:00
Lioncash
b50114ae70 A64/fuzz_with_unicorn: Avoid repeated unnecessary library calls in RunTestInstance()
Repeatedly retrieving the vectors and registers from unicorn involves
copying the entire set of registers and vectors by value instead of
simply retrieving a reference to them. Instead, we can just do the work
once and print out the values.

While we're at it, also make our bracing consistent.
2020-04-22 21:04:22 +01:00
MerryMage
06b31448aa emit_x64_vector: Changes to VectorSignedSaturatedDoublingMultiply
* Return both the upper and lower parts of the multiply if required
* SSE2 does not support the pmuldq instruction, do sign correction to an unsigned result instead
* Improve port utilisation where possible (punpck instructions were a bottleneck)
2020-04-22 20:55:06 +01:00
MerryMage
3caeb62ef1 emit_x64_floating_point: Hardware FMA implementation of FPRSqrtStepFused 2020-04-22 20:55:06 +01:00
MerryMage
b8daa4feac simd_two_register_misc: FNEG (vector) with Q == 0 had dirty upper 2020-04-22 20:55:05 +01:00
Lioncash
134bb02e19 ir/value: Replace includes with forward declarations
enum classes are still considered complete types when forward declared
(as the compiler knows the exact size of the type from the declaration
alone). The only difference in this case being that the members of the
enum class aren't visible. Given we don't use the members within this
header in any way, we can simply forward declare them here and remove
the inclusions.
2020-04-22 20:55:05 +01:00
MerryMage
9b65100660 A64: Implement FastDispatchHint 2020-04-22 20:53:46 +01:00
MerryMage
656ceff225 emit_x64_floating_point: Fix smallest normal check in EmitFPMulAdd 2020-04-22 20:53:45 +01:00
MerryMage
d1d6f4feb5 system: Implement MRS CNTFRQ_EL0 2020-04-22 20:53:45 +01:00
MerryMage
c96c534615 fuzz_with_unicorn: Split utility functions into fuzz_util 2020-04-22 20:46:23 +01:00
MerryMage
35c758f211 tests: Refactor unicorn_emu to allow for A32 unicorn 2020-04-22 20:46:23 +01:00
MerryMage
cd12a1862e fuzz_with_unicorn: Remove FCVT_float from ignore list 2020-04-22 20:46:23 +01:00
MerryMage
a8b938ef32 fuzz_with_unicorn: Randomize SP 2020-04-22 20:46:23 +01:00
MerryMage
9a7d75be3b fuzz_with_unicorn: Randomize PC 2020-04-22 20:46:23 +01:00
MerryMage
ff84740ea6 testenv: Make code_mem mobile 2020-04-22 20:46:23 +01:00
MerryMage
63eb4e0f31 fuzz_with_unicorn: Randomize FPCR.AHP and FPCR.FZ16 2020-04-22 20:46:23 +01:00
MerryMage
bb93353f94 emit_x64_vector_floating_point: Correct FMA in FTZ mode
x64 rounds before flushing to zero
AArch64 rounds after flushing to zero

This difference of behaviour is noticable if something would round to a smallest normalized number
2020-04-22 20:46:23 +01:00
MerryMage
822fd4a875 backend_x64: Fix FPVectorMulAdd and FPMulAdd NaN handling with denormals
Denormals should be treated as zero in NaN handler
2020-04-22 20:46:23 +01:00
MerryMage
868ec44f30 fuzz_with_unicorn: Randomize FPCR.FZ 2020-04-22 20:46:23 +01:00
MerryMage
2b538b471f fuzz_with_unicorn: Extract RandomFpcr function
Deduplicate randomization of fpcr and make use of FP::FPCR
2020-04-22 20:46:23 +01:00
MerryMage
f2344f4c87 fuzz_with_unicorn: Randomize FPCR.DN 2020-04-22 20:46:23 +01:00
MerryMage
680395a803 fuzz_with_unicorn: Disable testing of FDIV 2020-04-22 20:46:22 +01:00
MerryMage
f020dbe4ed A64: Implement SQXTUN 2020-04-22 20:46:22 +01:00
MerryMage
76b07d6646 u128: StickyLogicalShiftRight requires special-casing for amount == 64
In this case (128 - amount) == 64, and this invokes undefined behaviour
2020-04-22 20:46:21 +01:00
MerryMage
0ce11b7b15 emit_x64_floating_point: Implement accurate fallback for FPMulAdd{32,64} 2020-04-22 20:46:21 +01:00
MerryMage
55eaa16615 a64_emit_x64: Ensure host has updated ticks in EmitA64GetCNTPCT
Discovered by @Subv.
Fixes incomplete fix begun in 5a91c94dca47c9702dee20fbd5ae1f4c07eef9df.
That fix fails to take into account that LinkBlock doesn't update ticks until there
are no remaining ticks to be executed.

Test added to confirm fix.
2020-04-22 20:46:21 +01:00
MerryMage
edd795e991 a64_emit_x64: Fix stack misalignment on Windows for 128-bit exclusive writes
Discovered by @Subv.
Includes a test to ensure this codepath is exercised on Windows.
2020-04-22 20:46:21 +01:00
MerryMage
4525209bab tests: Add FABD test 2020-04-22 20:46:21 +01:00
MerryMage
9c4f234417 fuzz_with_unicorn: Avoid self-modifying code
* Don't immediately terminate when unicorn raises an interrupt
* Detect self-modifying code
2020-04-22 20:46:19 +01:00
MerryMage
9f8c6f60f5 fuzz_with_unicorn: Configure as per qemu max configuration 2020-04-22 20:46:19 +01:00
Lioncash
a25bacc436 tests/unicorn: Add type aliases to the Unicorn class
Centralizes all register and vector array definitions to a single set of
aliases, so if these are ever changed, then the rest of the testing code
will follow suit without the need to manually change them.
2020-04-22 20:46:19 +01:00
MerryMage
8c65d58cb8 tests/A64: Randomize FPCR.RMode for single random instruction 2020-04-22 20:46:18 +01:00
Lioncash
7d52d7bef8 inst_gen: Compress loop into std::any_of in IsInvalidInstruction()
Same behavior, but using a more self-documenting function.
2020-04-22 20:46:18 +01:00
Lioncash
d909b0919e fuzz_with_unicorn: Move std::vector outside loop in small random block test case
Avoids constructing and destructing the vector repeatedly, we can just
alter the contents of the vector on each iteration instead. Also move
out the std::array instances as well, like with the floating-point test
case and the single random instruction test case.

We can also use the regular form of std::generate and avoid hardcoding
size values twice.
2020-04-22 20:46:18 +01:00
MerryMage
ef57d75b32 fuzz_with_unicorn: Temporarily disable FDIV 2020-04-22 20:46:18 +01:00
MerryMage
ec0a91d6ee tests/A64: Test small blocks 2020-04-22 20:46:18 +01:00
MerryMage
330e6111fa fuzz_with_unicorn: Randomize FPCR.RMode 2020-04-22 20:46:18 +01:00
MerryMage
8c90fcf58e IR: Implement FPMulAdd 2020-04-22 20:46:18 +01:00
Lioncash
e64978ed89 fuzz_with_unicorn: Make float_numbers in floating-point tests constexpr
Given this is just a lookup table, this can be made immutable.
2020-04-22 20:46:18 +01:00
Lioncash
329137a277 fuzz_with_unicorn: Remove exclusion of FMOV (imm) for FP-16 floats
Qemu, or rather, Unicorn now supports FP-16, since I backported support
for the recent changes to mainline Qemu relating to FP-16 support.
2020-04-22 20:46:18 +01:00
Lioncash
0c63e8f396 fuzz_with_unicorn: Silence unused variable warning
Currently, structured bindings don't provide a way to ignore unused variables.
2020-04-22 20:46:17 +01:00
Lioncash
ab58bbddc8 unicorn: Be explicit about casting away const to const-incorrect APIs
Uses C++ casts which silence relevant warnings in Xcode 9.3

Also migrates relevant Read function equivalents over for consistency.
2020-04-22 20:46:16 +01:00
MerryMage
c490a45af6 fuzz_with_unicorn: Correct GenRandomInst
UnallocatedEncoding is now handled in ShouldTestInst
2020-04-22 20:46:15 +01:00
MerryMage
be57608353 A64: Partially implement FCVTZU (scalar, fixed-point) and FCVTZS (scalar, fixed-point) 2020-04-22 20:46:15 +01:00
MerryMage
9e4e4e9c1d A64: Implement system register CNTPCT_EL0 2020-04-22 20:46:15 +01:00
MerryMage
44a5b57f2a fuzz_with_unicorn: QEMU's implementation of FCVT is incorrect 2020-04-22 20:46:14 +01:00
MerryMage
1311f67b4a fuzz_with_unicorn: Explicitly test floating point instructions 2020-04-22 20:46:14 +01:00
MerryMage
e97581d063 fuzz_with_unicorn: Print AArch64 disassembly 2020-04-22 20:46:14 +01:00
MerryMage
1f5b3bca43 Exclusive fixups
* Incorrect size of exclusive_address
* Disable tests on exclusive memory instructions for now
2020-04-22 20:46:14 +01:00
MerryMage
f6a2104ab3 fuzz_with_unicorn: Speed up tests by not initializing/tearing down constantly 2020-04-22 20:46:14 +01:00
MerryMage
ef02658049 fuzz_with_unicorn: Fix read-past-end access via jit_iter 2020-04-22 20:46:14 +01:00
MerryMage
5edd623b9d Implement DC instructions 2020-04-22 20:46:14 +01:00
MerryMage
f378d2ef1b Forward declare IR::Opcode and IR::Type where possible 2020-04-22 20:46:14 +01:00
MerryMage
cc9efd13c9 A64: Implement STLLRB, STLLRH, STLLR, LDLARB, LDLARH, LDLAR 2020-04-22 20:46:13 +01:00
Lioncash
12c6f841c2 inst_gen: Make invalid_instructions a static inline variable 2020-04-22 20:46:13 +01:00
Lioncash
f96e83c486 fuzz_with_unicorn: Move instruction generator vector into GenRandomInst
Keeps scope localized and prevents potential static initialization issues.
2020-04-22 20:46:13 +01:00
MerryMage
47661b746b basic_block: Fix bogus GCC maybe-uninitialized warning 2020-04-22 20:46:13 +01:00
MerryMage
ca38225e08 fuzz_with_unicorn: Skip instructions that need to be interpreted 2020-04-22 20:46:13 +01:00
MerryMage
4be55b8b84 A64: Implement FMOV (scalar, immediate) 2020-04-22 20:46:13 +01:00
MerryMage
7e0f14303c fuzz_with_unicorn: Print information on test failure 2020-04-22 20:46:12 +01:00
MerryMage
d7044bc751 assert: Use fmt in ASSERT_MSG 2020-04-22 20:46:12 +01:00
Lioncash
71422c2b48 fuzz_with_unicorn: Move data outside loop
Given we know we're only ever going to use one instruction, we can just presize the vector and reassign to it
instead of potentially reallocating the same memory 100000 times
2020-04-22 20:46:12 +01:00
Lioncash
cd69c6a17c fuzz_with_unicorn: Dehardcode some constants 2020-04-22 20:46:12 +01:00
Lioncash
d3b1a72bca unicorn: Display EC and ISS separately beside the full ESR value
Makes it a little nicer to pick out the exception class details at a glance
2020-04-22 20:44:38 +01:00
Lioncash
8fa9849c25 unicorn: Use static_cast instead of reinterpret_cast
It's well-defined to cast from void* back to the original pointer type.
2020-04-22 20:44:38 +01:00
Lioncash
cf824fb2b2 unicorn_load: Minor Windows-related changes
- Add missing include
- Fix a potential compilation issue where the constructor wouldn't be able to execute, as it would be private.
2020-04-22 20:44:38 +01:00
MerryMage
a8ed248a13 tests/A64: Test memory writes 2020-04-22 20:44:38 +01:00
Lioncash
f1057aa362 tests: Fix truncation in GetFpcr() 2020-04-22 20:44:38 +01:00
MerryMage
e1df7ae621 IR: Add IR instructions A64Memory{Read,Write}128
This implementation only works on macOS and Linux.
2020-04-22 20:44:37 +01:00
MerryMage
1749780929 interface: Move Vector typedef to config.h 2020-04-22 20:44:37 +01:00
MerryMage
2a493f8b50 fuzz_with_unicorn: Compare vectors 2020-04-22 20:42:46 +01:00
MerryMage
35aaee6cc6 tests/A64: Randomize vectors 2020-04-22 20:42:46 +01:00
MerryMage
33a02ed91a tests/A64/unicorn: Print interrupt number when InterruptHook is hit 2020-04-22 20:42:46 +01:00
MerryMage
e9f6e7c82c tests/A64: Allow RunTestInstance to start from an arbitrary offset 2020-04-22 20:42:46 +01:00
Thomas Guillemard
896cf44f96 A64: Implement REV, REV32, and REV16 (#126) 2020-04-22 20:42:46 +01:00
MerryMage
7992a319ba A64/tests: Split unicorn sanity checking from other tests 2020-04-22 20:42:45 +01:00
MerryMage
9d42bc3228 tests/A64: Single random instruction: Test branch instructions as well 2020-04-22 20:42:45 +01:00
MerryMage
9f57283a30 A64: Implement SBFM, BFM, UBFM 2020-04-22 20:42:45 +01:00
MerryMage
a59e9ad9c6 travis: Run A64 tests 2020-04-22 20:42:45 +01:00
MerryMage
e99db8e745 tests/A64: Randomize PSTATE.<NZCV> 2020-04-22 20:42:45 +01:00
MerryMage
99b7516c8c testenv: Use format constants 2020-04-22 20:42:45 +01:00
MerryMage
39d083aa87 tests/A64: Unicorn interface fixes
- Use a std::unique_ptr instead of new/delete.
- UnmappedMemoryHook: Correct range when wraparound of the address space occurs
- UnmappedMemoryHook: Handle case when we attempt to map the same page twice
2020-04-22 20:42:45 +01:00
MerryMage
d5725de26a tests/A64: Fuzz against unicorn 2020-04-22 20:42:45 +01:00
MerryMage
a0ef6eda19 tests/A64: Move TestEnvironment to own header 2020-04-22 20:42:45 +01:00
MerryMage
2f8d7ae86f tests/a64: Use format constants 2020-04-22 20:42:45 +01:00
MerryMage
cb481a3a48 A64: Implement compare and branch 2020-04-22 20:42:45 +01:00
MerryMage
e8bcf72ee5 A64: PSTATE access and tests 2020-04-22 20:42:45 +01:00
MerryMage
0641445e51 A64: Implement logical 2020-04-22 20:42:45 +01:00
MerryMage
d1cef6ffb0 A64: Implement ADD_shifted 2020-04-22 20:42:44 +01:00
MerryMage
d1eb757f93 A64: Backend framework 2020-04-22 20:42:44 +01:00