projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 3cb6307) | patch
[AMDGPU] Preserve vcc_lo when shrinking V_CNDMASK
authorPiotr Sobczak <Piotr.Sobczak@amd.com>
Tue, 25 Aug 2020 09:39:01 +0000 (11:39 +0200)
committerPiotr Sobczak <Piotr.Sobczak@amd.com>
Thu, 27 Aug 2020 08:22:50 +0000 (10:22 +0200)
commit4e9d207117f684a7d74132ef10d1c2a2a043b161
tree536444083fc530cfb0cd118b98f6f0599b813375tree | snapshot
parent3cb63073ea0f9f0d5efdd22ed08d329d96fb5ae5commit | diff
[AMDGPU] Preserve vcc_lo when shrinking V_CNDMASK

There is no justification for changing vcc_lo to vcc
when shrinking V_CNDMASK, and such a change could
later confuse live variable analysis.

Make sure the original register is preserved.

Differential Revision: https://reviews.llvm.org/D86541
llvm/lib/Target/AMDGPU/SIInstrInfo.cpp diff | blob | history
llvm/test/CodeGen/AMDGPU/shrink-instructions-implicit-vcclo.mir [new file with mode: 0644] blob
Domain: System / Toolchain;