projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 4481eef) | patch
AMDGPU: Fix interaction of tfe and d16
authorMatt Arsenault <Matthew.Arsenault@amd.com>
Fri, 17 Jan 2020 20:40:15 +0000 (15:40 -0500)
committerMatt Arsenault <arsenm2@gmail.com>
Wed, 22 Jan 2020 14:26:17 +0000 (09:26 -0500)
commit9c928649a085646c4c779bac095643b50b464d83
tree9dd51c9e5fab735e82617a682c6c56351b59c19dtree | snapshot
parent4481eefbe8425c63289186dd13319aaa7043e67fcommit | diff
AMDGPU: Fix interaction of tfe and d16

This using the wrong result register, and dropping the result entirely
for v2f16. This would fail to select on the scalar case. I believe it
was also mishandling packed/unpacked subtargets.
llvm/lib/Target/AMDGPU/SIISelLowering.cpp diff | blob | history
llvm/test/CodeGen/AMDGPU/image-load-d16-tfe.ll [new file with mode: 0644] blob
Domain: System / Toolchain;