diff options
| author | Matt Arsenault <Matthew.Arsenault@amd.com> | 2017-03-21 21:39:51 +0000 |
|---|---|---|
| committer | Matt Arsenault <Matthew.Arsenault@amd.com> | 2017-03-21 21:39:51 +0000 |
| commit | 3dbeefa978fb7e7b231b249f9cd90c67b9e83277 (patch) | |
| tree | d74bf7fe30e44588d573919f3625edacb2586112 /llvm/test/CodeGen/AMDGPU/llvm.r600.dot4.ll | |
| parent | f6021ecddc73d14c94ad70938250d58f330795be (diff) | |
| download | bcm5719-llvm-3dbeefa978fb7e7b231b249f9cd90c67b9e83277.tar.gz bcm5719-llvm-3dbeefa978fb7e7b231b249f9cd90c67b9e83277.zip | |
AMDGPU: Mark all unspecified CC functions in tests as amdgpu_kernel
Currently the default C calling convention functions are treated
the same as compute kernels. Make this explicit so the default
calling convention can be changed to a non-kernel.
Converted with perl -pi -e 's/define void/define amdgpu_kernel void/'
on the relevant test directories (and undoing in one place that actually
wanted a non-kernel).
llvm-svn: 298444
Diffstat (limited to 'llvm/test/CodeGen/AMDGPU/llvm.r600.dot4.ll')
| -rw-r--r-- | llvm/test/CodeGen/AMDGPU/llvm.r600.dot4.ll | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/llvm/test/CodeGen/AMDGPU/llvm.r600.dot4.ll b/llvm/test/CodeGen/AMDGPU/llvm.r600.dot4.ll index 4db29c58385..de8a47741c9 100644 --- a/llvm/test/CodeGen/AMDGPU/llvm.r600.dot4.ll +++ b/llvm/test/CodeGen/AMDGPU/llvm.r600.dot4.ll @@ -2,7 +2,7 @@ declare float @llvm.r600.dot4(<4 x float>, <4 x float>) nounwind readnone -define void @test_dp4(float addrspace(1)* %out, <4 x float> addrspace(1)* %a, <4 x float> addrspace(1)* %b) nounwind { +define amdgpu_kernel void @test_dp4(float addrspace(1)* %out, <4 x float> addrspace(1)* %a, <4 x float> addrspace(1)* %b) nounwind { %src0 = load <4 x float>, <4 x float> addrspace(1)* %a, align 16 %src1 = load <4 x float>, <4 x float> addrspace(1)* %b, align 16 %dp4 = call float @llvm.r600.dot4(<4 x float> %src0, <4 x float> %src1) nounwind readnone |
