projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 2e9c663) | patch
AMDGPU: Stop trying to specially handle vector stores in printf lowering
authorMatt Arsenault <Matthew.Arsenault@amd.com>
Fri, 6 Jan 2023 23:23:53 +0000 (18:23 -0500)
committerMatt Arsenault <Matthew.Arsenault@amd.com>
Tue, 10 Jan 2023 18:07:01 +0000 (13:07 -0500)
commit777b905bae1ba0c219143f02d952d9cd5c242f50
treef61509271b7caa0cff101bab47a444711a49bd21tree | snapshot
parent2e9c663ab4a0385e619a64bc0726be0bb178c0b1commit | diff
AMDGPU: Stop trying to specially handle vector stores in printf lowering

This was broken for 1 element vectors and trying to create invalid
casts. We can directly store any type just fine, so don't bother with
this buggy conversion logic.
llvm/lib/Target/AMDGPU/AMDGPUPrintfRuntimeBinding.cpp diff | blob | history
llvm/test/CodeGen/AMDGPU/opencl-printf.ll diff | blob | history
Domain: System / Toolchain;