llvm-project/llvm/test/Transforms/InstCombine/AMDGPU
Roman Lebedev 2fea5d5d4a
[InstCombine] tmp alloca bypass: ensure that the replacement dominates all alloca uses
After 077bff39d4,
isDereferenceableForAllocaSize() can recurse into selects,
which is causing a problem for the new test case,
reduced from https://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20210412/904154.html
because the replacement (the select) is defined after the first use
of an alloca, so we'd end up with a verifier error.

Now, this new check is too restrictive.
We likely can handle *some* cases, by trying to sink all uses of an alloca
to after the the def.
2021-04-14 13:04:12 +03:00
..
amdgcn-demanded-vector-elts-inseltpoison.ll Use unary CreateShuffleVector if possible 2020-12-30 22:36:08 +09:00
amdgcn-demanded-vector-elts.ll Use unary CreateShuffleVector if possible 2020-12-30 22:36:08 +09:00
amdgcn-intrinsics-gfx8.ll
amdgcn-intrinsics.ll
fma_legacy.ll [AMDGPU] Add simplification/combines for llvm.amdgcn.fma.legacy 2020-10-23 16:16:13 +01:00
fmul_legacy.ll [AMDGPU] Add simplification/combines for llvm.amdgcn.fma.legacy 2020-10-23 16:16:13 +01:00
ldexp.ll [InstCombine] regenerate test checks; NFC 2020-09-24 09:34:17 -04:00
lit.local.cfg
memcpy-from-constant.ll [InstCombine] tmp alloca bypass: ensure that the replacement dominates all alloca uses 2021-04-14 13:04:12 +03:00
tan.ll