projects / oota-llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 85b811e) | patch
Revert "Remove unnecessary call to getAllocatableRegClass"
authorTom Stellard <thomas.stellard@amd.com>
Thu, 12 Nov 2015 21:43:25 +0000 (21:43 +0000)
committerTom Stellard <thomas.stellard@amd.com>
Thu, 12 Nov 2015 21:43:25 +0000 (21:43 +0000)
commitb6cc1fddcde927fbad11bd833eb8382f02fc244a
tree4b63e32d5dfcaea0b502afe0408046dd9f4487d2tree | snapshot
parent85b811eedd5ba15cee44bef5ec05d846c36dbed1commit | diff
Revert "Remove unnecessary call to getAllocatableRegClass"

This reverts commit r252565.

This also includes the revert of the commit mentioned below in order to
avoid breaking tests in AMDGPU:

Revert "AMDGPU: Set isAllocatable = 0 on VS_32/VS_64"

This reverts commit r252674.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@252956 91177308-0d34-0410-b5e6-96231b3b80d8
lib/CodeGen/SelectionDAG/InstrEmitter.cpp diff | blob | history
lib/Target/AMDGPU/SIRegisterInfo.cpp diff | blob | history
lib/Target/AMDGPU/SIRegisterInfo.h diff | blob | history
lib/Target/AMDGPU/SIRegisterInfo.td diff | blob | history
test/CodeGen/AMDGPU/ftrunc.f64.ll diff | blob | history
test/CodeGen/AMDGPU/llvm.round.f64.ll diff | blob | history
C/C++ LLVM-based compilers that forbids OOTA behaviors