llvm-6502/test/CodeGen/R600/array-ptr-calc-i32.ll
Matt Arsenault 2baa7c53c9 R600/SI: Fix verifier error with pseudo store instructions.
Use i32 instead of specifying SReg_32. When this is
the pseudo INDIRECT_BASE_ADDR, this would give a bogus
verifier error.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207770 91177308-0d34-0410-b5e6-96231b3b80d8
2014-05-01 16:37:52 +00:00

32 lines
1.3 KiB
LLVM

; RUN: llc -verify-machineinstrs -march=r600 -mcpu=SI < %s | FileCheck -check-prefix=SI %s
declare i32 @llvm.SI.tid() nounwind readnone
declare void @llvm.AMDGPU.barrier.local() nounwind noduplicate
; The required pointer calculations for the alloca'd actually requires
; an add and won't be folded into the addressing, which fails with a
; 64-bit pointer add. This should work since private pointers should
; be 32-bits.
; SI-LABEL: @test_private_array_ptr_calc:
; SI: V_ADD_I32_e32 [[PTRREG:v[0-9]+]]
; SI: V_MOVRELD_B32_e32 {{v[0-9]+}}, [[PTRREG]]
define void @test_private_array_ptr_calc(i32 addrspace(1)* noalias %out, i32 addrspace(1)* noalias %inA, i32 addrspace(1)* noalias %inB) {
%alloca = alloca [4 x i32], i32 4, align 16
%tid = call i32 @llvm.SI.tid() readnone
%a_ptr = getelementptr i32 addrspace(1)* %inA, i32 %tid
%b_ptr = getelementptr i32 addrspace(1)* %inB, i32 %tid
%a = load i32 addrspace(1)* %a_ptr
%b = load i32 addrspace(1)* %b_ptr
%result = add i32 %a, %b
%alloca_ptr = getelementptr [4 x i32]* %alloca, i32 1, i32 %b
store i32 %result, i32* %alloca_ptr, align 4
; Dummy call
call void @llvm.AMDGPU.barrier.local() nounwind noduplicate
%reload = load i32* %alloca_ptr, align 4
%out_ptr = getelementptr i32 addrspace(1)* %out, i32 %tid
store i32 %reload, i32 addrspace(1)* %out_ptr, align 4
ret void
}