63 lines
4.1 KiB
LLVM
63 lines
4.1 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-globals all --version 5
|
|
; RUN: opt -S -mtriple=amdgcn-amd-amdhsa -mcpu=gfx803 -passes=amdgpu-attributor %s | FileCheck %s
|
|
|
|
@buf_shared = internal addrspace(3) global [2080 x i8] poison, align 16
|
|
|
|
; Constant expression element may not have a pointer type and the
|
|
; addrspacecast may not be the toplevel operation.
|
|
|
|
|
|
; This should infer "amdgpu-no-flat-scratch-init". It should not infer "amdgpu-no-queue-ptr"
|
|
;.
|
|
; CHECK: @buf_shared = internal addrspace(3) global [2080 x i8] poison, align 16
|
|
; CHECK: @buf_private = internal addrspace(5) global [2080 x i8] poison, align 16
|
|
;.
|
|
define amdgpu_kernel void @issue120256(ptr addrspace(1) %out) {
|
|
; CHECK-LABEL: define amdgpu_kernel void @issue120256(
|
|
; CHECK-SAME: ptr addrspace(1) [[OUT:%.*]]) #[[ATTR0:[0-9]+]] {
|
|
; CHECK-NEXT: [[CONV_I:%.*]] = and i32 trunc (i64 sub (i64 16, i64 ptrtoint (ptr addrspacecast (ptr addrspace(3) @buf_shared to ptr) to i64)) to i32), 15
|
|
; CHECK-NEXT: [[ADD_PTR:%.*]] = getelementptr inbounds nuw i8, ptr addrspace(3) @buf_shared, i32 [[CONV_I]]
|
|
; CHECK-NEXT: [[LD:%.*]] = load i8, ptr addrspace(3) [[ADD_PTR]], align 1
|
|
; CHECK-NEXT: store i8 [[LD]], ptr addrspace(1) [[OUT]], align 1
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
%conv.i = and i32 trunc (i64 sub (i64 16, i64 ptrtoint (ptr addrspacecast (ptr addrspace(3) @buf_shared to ptr) to i64)) to i32), 15
|
|
%add.ptr = getelementptr inbounds nuw i8, ptr addrspace(3) @buf_shared, i32 %conv.i
|
|
%ld = load i8, ptr addrspace(3) %add.ptr, align 1
|
|
store i8 %ld, ptr addrspace(1) %out, align 1
|
|
ret void
|
|
}
|
|
|
|
@buf_private = internal addrspace(5) global [2080 x i8] poison, align 16
|
|
|
|
; Constant expression element may not have a pointer type and the
|
|
; addrspacecast may not be the toplevel operation.
|
|
|
|
; This should not infer "amdgpu-no-flat-scratch-init" nor "amdgpu-no-queue-ptr"
|
|
define amdgpu_kernel void @issue120256_private(ptr addrspace(1) %out) {
|
|
; CHECK-LABEL: define amdgpu_kernel void @issue120256_private(
|
|
; CHECK-SAME: ptr addrspace(1) [[OUT:%.*]]) #[[ATTR1:[0-9]+]] {
|
|
; CHECK-NEXT: [[CONV_I:%.*]] = and i32 trunc (i64 sub (i64 16, i64 ptrtoint (ptr addrspacecast (ptr addrspace(5) @buf_private to ptr) to i64)) to i32), 15
|
|
; CHECK-NEXT: [[ADD_PTR:%.*]] = getelementptr inbounds nuw i8, ptr addrspace(5) @buf_private, i32 [[CONV_I]]
|
|
; CHECK-NEXT: [[LD:%.*]] = load i8, ptr addrspace(5) [[ADD_PTR]], align 1
|
|
; CHECK-NEXT: store i8 [[LD]], ptr addrspace(1) [[OUT]], align 1
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
%conv.i = and i32 trunc (i64 sub (i64 16, i64 ptrtoint (ptr addrspacecast (ptr addrspace(5) @buf_private to ptr) to i64)) to i32), 15
|
|
%add.ptr = getelementptr inbounds nuw i8, ptr addrspace(5) @buf_private, i32 %conv.i
|
|
%ld = load i8, ptr addrspace(5) %add.ptr, align 1
|
|
store i8 %ld, ptr addrspace(1) %out, align 1
|
|
ret void
|
|
}
|
|
|
|
!llvm.module.flags = !{!0}
|
|
|
|
; FIXME: Inference of amdgpu-no-queue-ptr should not depend on code object version.
|
|
!0 = !{i32 1, !"amdhsa_code_object_version", i32 400}
|
|
;.
|
|
; CHECK: attributes #[[ATTR0]] = { "amdgpu-no-agpr" "amdgpu-no-completion-action" "amdgpu-no-default-queue" "amdgpu-no-dispatch-id" "amdgpu-no-dispatch-ptr" "amdgpu-no-flat-scratch-init" "amdgpu-no-heap-ptr" "amdgpu-no-hostcall-ptr" "amdgpu-no-implicitarg-ptr" "amdgpu-no-lds-kernel-id" "amdgpu-no-multigrid-sync-arg" "amdgpu-no-workgroup-id-x" "amdgpu-no-workgroup-id-y" "amdgpu-no-workgroup-id-z" "amdgpu-no-workitem-id-x" "amdgpu-no-workitem-id-y" "amdgpu-no-workitem-id-z" "target-cpu"="gfx803" "uniform-work-group-size"="false" }
|
|
; CHECK: attributes #[[ATTR1]] = { "amdgpu-no-agpr" "amdgpu-no-completion-action" "amdgpu-no-default-queue" "amdgpu-no-dispatch-id" "amdgpu-no-dispatch-ptr" "amdgpu-no-heap-ptr" "amdgpu-no-hostcall-ptr" "amdgpu-no-implicitarg-ptr" "amdgpu-no-lds-kernel-id" "amdgpu-no-multigrid-sync-arg" "amdgpu-no-workgroup-id-x" "amdgpu-no-workgroup-id-y" "amdgpu-no-workgroup-id-z" "amdgpu-no-workitem-id-x" "amdgpu-no-workitem-id-y" "amdgpu-no-workitem-id-z" "target-cpu"="gfx803" "uniform-work-group-size"="false" }
|
|
;.
|
|
; CHECK: [[META0:![0-9]+]] = !{i32 1, !"amdhsa_code_object_version", i32 400}
|
|
;.
|