projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 70e2d63) | patch
[AMDGPU] Do not fix implicit vcc operand on INLINEASM
authorJay Foad <jay.foad@amd.com>
Wed, 29 Mar 2023 14:06:53 +0000 (15:06 +0100)
committerJay Foad <jay.foad@amd.com>
Wed, 29 Mar 2023 19:23:36 +0000 (20:23 +0100)
commit3e3594c771e65b48ceeb89a7186d5e3b119ec3b9
tree8db8998d4de83fda05e06d08374f7eb0b347ddd3tree | snapshot
parent70e2d631e178fbf5cfcd45d8af43bfcad24db744commit | diff
[AMDGPU] Do not fix implicit vcc operand on INLINEASM

An INLINEASM can have an implicit def of vcc. It is not appropriate for
fixImplicitOperands to change this to vcc_lo on wave32.

Differential Revision: https://reviews.llvm.org/D147157
llvm/lib/Target/AMDGPU/SIInstrInfo.cpp diff | blob | history
Domain: System / Toolchain;