OSDN Git Service

[AMDGPU] Fix illegal shrink of V_SUBB_U32 and V_ADDC_U32
authorStanislav Mekhanoshin <Stanislav.Mekhanoshin@amd.com>
Tue, 20 Jun 2017 20:33:44 +0000 (20:33 +0000)
committerStanislav Mekhanoshin <Stanislav.Mekhanoshin@amd.com>
Tue, 20 Jun 2017 20:33:44 +0000 (20:33 +0000)
commit64373efcabd901d75cd153b8a4bf2621ec0049db
treeccfba4493a9ef2b1ac295b9060e451a57eabce86
parent864e349715f5e32049bef560a23777b465f6c8ca
[AMDGPU] Fix illegal shrink of V_SUBB_U32 and V_ADDC_U32

If there is an immediate operand we shall not shrink V_SUBB_U32
and V_ADDC_U32, it does not fit e32 encoding.

Differential Revison: https://reviews.llvm.org/D34291

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@305840 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Target/AMDGPU/SIShrinkInstructions.cpp
test/CodeGen/AMDGPU/shrink-carry.mir [new file with mode: 0644]