mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-10-22 12:33:33 +02:00
dd9ab77318
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
51 lines
2.2 KiB
LLVM
51 lines
2.2 KiB
LLVM
; RUN: opt -S -mtriple=amdgcn-amd-amdhsa -amdgpu-promote-alloca < %s | FileCheck %s
|
|
|
|
; Make sure this allocates the correct size if the alloca has a non-0
|
|
; number of elements.
|
|
|
|
; CHECK-LABEL: @array_alloca(
|
|
; CHECK: %stack = alloca i32, i32 5, align 4
|
|
define amdgpu_kernel void @array_alloca(i32 addrspace(1)* nocapture %out, i32 addrspace(1)* nocapture %in) #0 {
|
|
entry:
|
|
%stack = alloca i32, i32 5, align 4
|
|
%ld0 = load i32, i32 addrspace(1)* %in, align 4
|
|
%arrayidx1 = getelementptr inbounds i32, i32* %stack, i32 %ld0
|
|
store i32 4, i32* %arrayidx1, align 4
|
|
%arrayidx2 = getelementptr inbounds i32, i32 addrspace(1)* %in, i32 1
|
|
%ld1 = load i32, i32 addrspace(1)* %arrayidx2, align 4
|
|
%arrayidx3 = getelementptr inbounds i32, i32* %stack, i32 %ld1
|
|
store i32 5, i32* %arrayidx3, align 4
|
|
%arrayidx10 = getelementptr inbounds i32, i32* %stack, i32 0
|
|
%ld2 = load i32, i32* %arrayidx10, align 4
|
|
store i32 %ld2, i32 addrspace(1)* %out, align 4
|
|
%arrayidx12 = getelementptr inbounds i32, i32* %stack, i32 1
|
|
%ld3 = load i32, i32* %arrayidx12
|
|
%arrayidx13 = getelementptr inbounds i32, i32 addrspace(1)* %out, i32 1
|
|
store i32 %ld3, i32 addrspace(1)* %arrayidx13
|
|
ret void
|
|
}
|
|
|
|
; CHECK-LABEL: @array_alloca_dynamic(
|
|
; CHECK: %stack = alloca i32, i32 %size, align 4
|
|
define amdgpu_kernel void @array_alloca_dynamic(i32 addrspace(1)* nocapture %out, i32 addrspace(1)* nocapture %in, i32 %size) #0 {
|
|
entry:
|
|
%stack = alloca i32, i32 %size, align 4
|
|
%ld0 = load i32, i32 addrspace(1)* %in, align 4
|
|
%arrayidx1 = getelementptr inbounds i32, i32* %stack, i32 %ld0
|
|
store i32 4, i32* %arrayidx1, align 4
|
|
%arrayidx2 = getelementptr inbounds i32, i32 addrspace(1)* %in, i32 1
|
|
%ld1 = load i32, i32 addrspace(1)* %arrayidx2, align 4
|
|
%arrayidx3 = getelementptr inbounds i32, i32* %stack, i32 %ld1
|
|
store i32 5, i32* %arrayidx3, align 4
|
|
%arrayidx10 = getelementptr inbounds i32, i32* %stack, i32 0
|
|
%ld2 = load i32, i32* %arrayidx10, align 4
|
|
store i32 %ld2, i32 addrspace(1)* %out, align 4
|
|
%arrayidx12 = getelementptr inbounds i32, i32* %stack, i32 1
|
|
%ld3 = load i32, i32* %arrayidx12
|
|
%arrayidx13 = getelementptr inbounds i32, i32 addrspace(1)* %out, i32 1
|
|
store i32 %ld3, i32 addrspace(1)* %arrayidx13
|
|
ret void
|
|
}
|
|
|
|
attributes #0 = { nounwind }
|