afonso360 requested jameysharp for a review on PR #5994.
afonso360 opened PR #5994 from fuzz-simd-arith
to main
:
:wave: Hey,
This is part 2 of #5991. In that PR I added some arithmetic ops to the interpreter, in this one we add those to the fuzzer.
Some notes:
- SIMD
rotl
/rotr
/ishl
/ushr
/sshr
were affected by #5991 but since the RHS for these ops is scalar, the interpreter is returning the wrong result so I didn't add them here.bnot
is a unary op, but was being inserted as a binary op. We were ignoring the second argument, I've now fixed that in the opcode list.This PR is built on top of #5991, but the only relevant commit is the last one.
afonso360 edited PR #5994 from fuzz-simd-arith
to main
:
:wave: Hey,
This is part 2 of #5991. In that PR I added some arithmetic ops to the interpreter, in this one we add those to the fuzzer.
Some notes:
- SIMD
rotl
/rotr
/ishl
/ushr
/sshr
were affected by #5991 but since the RHS for these ops is scalar, the interpreter is returning the wrong result so I didn't add them here.bnot
is a unary op, but was being inserted as a binary op. We were ignoring the second argument, I've now fixed that in the opcode list.This PR is built on top of #5991, but the only relevant commit is the last one.
I've been running the fuzzer for about an hour on x86 and AArch64 and it hasn't complained yet, I'll update this if it finds something.
jameysharp submitted PR review.
afonso360 updated PR #5994 from fuzz-simd-arith
to main
.
afonso360 has enabled auto merge for PR #5994.
afonso360 merged PR #5994.
Last updated: Dec 23 2024 at 12:05 UTC