llvmbot wrote:
<!--LLVM PR SUMMARY COMMENT--> @llvm/pr-subscribers-clang Author: Alex Voicu (AlexVlx) <details> <summary>Changes</summary> Turns out these tests are a bit unwieldy to hand-update, so switch them over to being generated, as requested in #<!-- -->112442. --- Patch is 111.84 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/112554.diff 2 Files Affected: - (modified) clang/test/CodeGenOpenCL/addr-space-struct-arg.cl (+1171-81) - (modified) clang/test/CodeGenOpenCL/amdgcn-automatic-variable.cl (+81-41) ``````````diff diff --git a/clang/test/CodeGenOpenCL/addr-space-struct-arg.cl b/clang/test/CodeGenOpenCL/addr-space-struct-arg.cl index bab0e21067eeae..7377b5bcbc347a 100644 --- a/clang/test/CodeGenOpenCL/addr-space-struct-arg.cl +++ b/clang/test/CodeGenOpenCL/addr-space-struct-arg.cl @@ -1,9 +1,10 @@ -// RUN: %clang_cc1 %s -emit-llvm -o - -O0 -ffake-address-space-map -triple i686-pc-darwin | FileCheck -enable-var-scope -check-prefixes=ALL,X86 %s -// RUN: %clang_cc1 %s -emit-llvm -o - -O0 -triple amdgcn | FileCheck -enable-var-scope -check-prefixes=ALL,AMDGCN %s -// RUN: %clang_cc1 %s -emit-llvm -o - -cl-std=CL2.0 -O0 -triple amdgcn | FileCheck -enable-var-scope -check-prefixes=ALL,AMDGCN,AMDGCN20 %s -// RUN: %clang_cc1 %s -emit-llvm -o - -cl-std=CL1.2 -O0 -triple spir-unknown-unknown-unknown | FileCheck -enable-var-scope -check-prefixes=SPIR %s -// RUN: %clang_cc1 %s -emit-llvm -o - -cl-std=CL3.0 -O0 -triple amdgcn -cl-ext=+__opencl_c_program_scope_global_variables | FileCheck -enable-var-scope -check-prefixes=ALL,AMDGCN,AMDGCN20 %s -// RUN: %clang_cc1 %s -emit-llvm -o - -cl-std=CL3.0 -O0 -triple amdgcn | FileCheck -enable-var-scope -check-prefixes=ALL,AMDGCN %s +// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5 +// RUN: %clang_cc1 %s -emit-llvm -o - -O0 -ffake-address-space-map -triple i686-pc-darwin | FileCheck -check-prefixes=X86 %s +// RUN: %clang_cc1 %s -emit-llvm -o - -O0 -triple amdgcn | FileCheck -check-prefixes=AMDGCN %s +// RUN: %clang_cc1 %s -emit-llvm -o - -cl-std=CL2.0 -O0 -triple amdgcn | FileCheck -check-prefixes=AMDGCN20 %s +// RUN: %clang_cc1 %s -emit-llvm -o - -cl-std=CL1.2 -O0 -triple spir-unknown-unknown-unknown | FileCheck -check-prefixes=SPIR %s +// RUN: %clang_cc1 %s -emit-llvm -o - -cl-std=CL3.0 -O0 -triple amdgcn -cl-ext=+__opencl_c_program_scope_global_variables | FileCheck -check-prefixes=AMDGCN30-GVAR %s +// RUN: %clang_cc1 %s -emit-llvm -o - -cl-std=CL3.0 -O0 -triple amdgcn | FileCheck -check-prefixes=AMDGCN30 %s typedef int int2 __attribute__((ext_vector_type(2))); @@ -45,147 +46,1236 @@ struct LargeStructTwoMember { struct LargeStructOneMember g_s; #endif -// X86-LABEL: define{{.*}} void @foo(ptr dead_on_unwind noalias writable sret(%struct.Mat4X4) align 4 %agg.result, ptr noundef byval(%struct.Mat3X3) align 4 %in) -// AMDGCN-LABEL: define{{.*}} %struct.Mat4X4 @foo([9 x i32] %in.coerce) +// +// X86-LABEL: define void @foo( +// X86-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_MAT4X4:%.*]]) align 4 [[AGG_RESULT:%.*]], ptr noundef byval([[STRUCT_MAT3X3:%.*]]) align 4 [[IN:%.*]]) #[[ATTR0:[0-9]+]] { +// X86-NEXT: [[ENTRY:.*:]] +// X86-NEXT: [[RESULT_PTR:%.*]] = alloca ptr, align 4 +// X86-NEXT: store ptr [[AGG_RESULT]], ptr [[RESULT_PTR]], align 4 +// X86-NEXT: ret void +// +// AMDGCN-LABEL: define dso_local %struct.Mat4X4 @foo( +// AMDGCN-SAME: [9 x i32] [[IN_COERCE:%.*]]) #[[ATTR0:[0-9]+]] { +// AMDGCN-NEXT: [[ENTRY:.*:]] +// AMDGCN-NEXT: [[RETVAL:%.*]] = alloca [[STRUCT_MAT4X4:%.*]], align 4, addrspace(5) +// AMDGCN-NEXT: [[IN:%.*]] = alloca [[STRUCT_MAT3X3:%.*]], align 4, addrspace(5) +// AMDGCN-NEXT: [[COERCE_DIVE:%.*]] = getelementptr inbounds nuw [[STRUCT_MAT3X3]], ptr addrspace(5) [[IN]], i32 0, i32 0 +// AMDGCN-NEXT: store [9 x i32] [[IN_COERCE]], ptr addrspace(5) [[COERCE_DIVE]], align 4 +// AMDGCN-NEXT: [[TMP0:%.*]] = load [[STRUCT_MAT4X4]], ptr addrspace(5) [[RETVAL]], align 4 +// AMDGCN-NEXT: ret [[STRUCT_MAT4X4]] [[TMP0]] +// +// AMDGCN20-LABEL: define dso_local %struct.Mat4X4 @foo( +// AMDGCN20-SAME: [9 x i32] [[IN_COERCE:%.*]]) #[[ATTR0:[0-9]+]] { +// AMDGCN20-NEXT: [[ENTRY:.*:]] +// AMDGCN20-NEXT: [[RETVAL:%.*]] = alloca [[STRUCT_MAT4X4:%.*]], align 4, addrspace(5) +// AMDGCN20-NEXT: [[IN:%.*]] = alloca [[STRUCT_MAT3X3:%.*]], align 4, addrspace(5) +// AMDGCN20-NEXT: [[COERCE_DIVE:%.*]] = getelementptr inbounds nuw [[STRUCT_MAT3X3]], ptr addrspace(5) [[IN]], i32 0, i32 0 +// AMDGCN20-NEXT: store [9 x i32] [[IN_COERCE]], ptr addrspace(5) [[COERCE_DIVE]], align 4 +// AMDGCN20-NEXT: [[TMP0:%.*]] = load [[STRUCT_MAT4X4]], ptr addrspace(5) [[RETVAL]], align 4 +// AMDGCN20-NEXT: ret [[STRUCT_MAT4X4]] [[TMP0]] +// +// SPIR-LABEL: define dso_local spir_func void @foo( +// SPIR-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_MAT4X4:%.*]]) align 4 [[AGG_RESULT:%.*]], ptr noundef byval([[STRUCT_MAT3X3:%.*]]) align 4 [[IN:%.*]]) #[[ATTR0:[0-9]+]] { +// SPIR-NEXT: [[ENTRY:.*:]] +// SPIR-NEXT: ret void +// +// AMDGCN30-GVAR-LABEL: define dso_local %struct.Mat4X4 @foo( +// AMDGCN30-GVAR-SAME: [9 x i32] [[IN_COERCE:%.*]]) #[[ATTR0:[0-9]+]] { +// AMDGCN30-GVAR-NEXT: [[ENTRY:.*:]] +// AMDGCN30-GVAR-NEXT: [[RETVAL:%.*]] = alloca [[STRUCT_MAT4X4:%.*]], align 4, addrspace(5) +// AMDGCN30-GVAR-NEXT: [[IN:%.*]] = alloca [[STRUCT_MAT3X3:%.*]], align 4, addrspace(5) +// AMDGCN30-GVAR-NEXT: [[COERCE_DIVE:%.*]] = getelementptr inbounds nuw [[STRUCT_MAT3X3]], ptr addrspace(5) [[IN]], i32 0, i32 0 +// AMDGCN30-GVAR-NEXT: store [9 x i32] [[IN_COERCE]], ptr addrspace(5) [[COERCE_DIVE]], align 4 +// AMDGCN30-GVAR-NEXT: [[TMP0:%.*]] = load [[STRUCT_MAT4X4]], ptr addrspace(5) [[RETVAL]], align 4 +// AMDGCN30-GVAR-NEXT: ret [[STRUCT_MAT4X4]] [[TMP0]] +// +// AMDGCN30-LABEL: define dso_local %struct.Mat4X4 @foo( +// AMDGCN30-SAME: [9 x i32] [[IN_COERCE:%.*]]) #[[ATTR0:[0-9]+]] { +// AMDGCN30-NEXT: [[ENTRY:.*:]] +// AMDGCN30-NEXT: [[RETVAL:%.*]] = alloca [[STRUCT_MAT4X4:%.*]], align 4, addrspace(5) +// AMDGCN30-NEXT: [[IN:%.*]] = alloca [[STRUCT_MAT3X3:%.*]], align 4, addrspace(5) +// AMDGCN30-NEXT: [[COERCE_DIVE:%.*]] = getelementptr inbounds nuw [[STRUCT_MAT3X3]], ptr addrspace(5) [[IN]], i32 0, i32 0 +// AMDGCN30-NEXT: store [9 x i32] [[IN_COERCE]], ptr addrspace(5) [[COERCE_DIVE]], align 4 +// AMDGCN30-NEXT: [[TMP0:%.*]] = load [[STRUCT_MAT4X4]], ptr addrspace(5) [[RETVAL]], align 4 +// AMDGCN30-NEXT: ret [[STRUCT_MAT4X4]] [[TMP0]] +// Mat4X4 __attribute__((noinline)) foo(Mat3X3 in) { Mat4X4 out; return out; } -// ALL-LABEL: define {{.*}} void @ker -// Expect two mem copies: one for the argument "in", and one for -// the return value. -// X86: call void @llvm.memcpy.p0.p1.i32(ptr -// X86: call void @llvm.memcpy.p1.p0.i32(ptr addrspace(1) - -// AMDGCN: load [9 x i32], ptr addrspace(1) -// AMDGCN: call %struct.Mat4X4 @foo([9 x i32] -// AMDGCN: call void @llvm.memcpy.p1.p5.i64(ptr addrspace(1) +// +// X86-LABEL: define spir_kernel void @ker( +// X86-SAME: ptr addrspace(1) noundef align 4 [[IN:%.*]], ptr addrspace(1) noundef align 4 [[OUT:%.*]]) #[[ATTR1:[0-9]+]] !kernel_arg_addr_space [[META4:![0-9]+]] !kernel_arg_access_qual [[META5:![0-9]+]] !kernel_arg_type [[META6:![0-9]+]] !kernel_arg_base_type [[META6]] !kernel_arg_type_qual [[META7:![0-9]+]] { +// X86-NEXT: [[ENTRY:.*:]] +// X86-NEXT: [[IN_ADDR:%.*]] = alloca ptr addrspace(1), align 4 +// X86-NEXT: [[OUT_ADDR:%.*]] = alloca ptr addrspace(1), align 4 +// X86-NEXT: [[TMP:%.*]] = alloca [[STRUCT_MAT4X4:%.*]], align 4 +// X86-NEXT: [[BYVAL_TEMP:%.*]] = alloca [[STRUCT_MAT3X3:%.*]], align 4 +// X86-NEXT: store ptr addrspace(1) [[IN]], ptr [[IN_ADDR]], align 4 +// X86-NEXT: store ptr addrspace(1) [[OUT]], ptr [[OUT_ADDR]], align 4 +// X86-NEXT: [[TMP0:%.*]] = load ptr addrspace(1), ptr [[OUT_ADDR]], align 4 +// X86-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [[STRUCT_MAT4X4]], ptr addrspace(1) [[TMP0]], i32 0 +// X86-NEXT: [[TMP1:%.*]] = load ptr addrspace(1), ptr [[IN_ADDR]], align 4 +// X86-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds [[STRUCT_MAT3X3]], ptr addrspace(1) [[TMP1]], i32 1 +// X86-NEXT: call void @llvm.memcpy.p0.p1.i32(ptr align 4 [[BYVAL_TEMP]], ptr addrspace(1) align 4 [[ARRAYIDX1]], i32 36, i1 false) +// X86-NEXT: call void @foo(ptr dead_on_unwind writable sret([[STRUCT_MAT4X4]]) align 4 [[TMP]], ptr noundef byval([[STRUCT_MAT3X3]]) align 4 [[BYVAL_TEMP]]) #[[ATTR3:[0-9]+]] +// X86-NEXT: call void @llvm.memcpy.p1.p0.i32(ptr addrspace(1) align 4 [[ARRAYIDX]], ptr align 4 [[TMP]], i32 64, i1 false) +// X86-NEXT: ret void +// +// AMDGCN-LABEL: define dso_local amdgpu_kernel void @ker( +// AMDGCN-SAME: ptr addrspace(1) noundef align 4 [[IN:%.*]], ptr addrspace(1) noundef align 4 [[OUT:%.*]]) #[[ATTR1:[0-9]+]] !kernel_arg_addr_space [[META4:![0-9]+]] !kernel_arg_access_qual [[META5:![0-9]+]] !kernel_arg_type [[META6:![0-9]+]] !kernel_arg_base_type [[META6]] !kernel_arg_type_qual [[META7:![0-9]+]] { +// AMDGCN-NEXT: [[ENTRY:.*:]] +// AMDGCN-NEXT: [[IN_ADDR:%.*]] = alloca ptr addrspace(1), align 8, addrspace(5) +// AMDGCN-NEXT: [[OUT_ADDR:%.*]] = alloca ptr addrspace(1), align 8, addrspace(5) +// AMDGCN-NEXT: [[TMP:%.*]] = alloca [[STRUCT_MAT4X4:%.*]], align 4, addrspace(5) +// AMDGCN-NEXT: store ptr addrspace(1) [[IN]], ptr addrspace(5) [[IN_ADDR]], align 8 +// AMDGCN-NEXT: store ptr addrspace(1) [[OUT]], ptr addrspace(5) [[OUT_ADDR]], align 8 +// AMDGCN-NEXT: [[TMP0:%.*]] = load ptr addrspace(1), ptr addrspace(5) [[OUT_ADDR]], align 8 +// AMDGCN-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [[STRUCT_MAT4X4]], ptr addrspace(1) [[TMP0]], i64 0 +// AMDGCN-NEXT: [[TMP1:%.*]] = load ptr addrspace(1), ptr addrspace(5) [[IN_ADDR]], align 8 +// AMDGCN-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds [[STRUCT_MAT3X3:%.*]], ptr addrspace(1) [[TMP1]], i64 1 +// AMDGCN-NEXT: [[TMP2:%.*]] = getelementptr inbounds nuw [[STRUCT_MAT3X3]], ptr addrspace(1) [[ARRAYIDX1]], i32 0, i32 0 +// AMDGCN-NEXT: [[TMP3:%.*]] = load [9 x i32], ptr addrspace(1) [[TMP2]], align 4 +// AMDGCN-NEXT: [[CALL:%.*]] = call [[STRUCT_MAT4X4]] @[[FOO:[a-zA-Z0-9_$\"\\.-]*[a-zA-Z_$\"\\.-][a-zA-Z0-9_$\"\\.-]*]]([9 x i32] [[TMP3]]) #[[ATTR3:[0-9]+]] +// AMDGCN-NEXT: [[TMP4:%.*]] = getelementptr inbounds nuw [[STRUCT_MAT4X4]], ptr addrspace(5) [[TMP]], i32 0, i32 0 +// AMDGCN-NEXT: [[TMP5:%.*]] = extractvalue [[STRUCT_MAT4X4]] [[CALL]], 0 +// AMDGCN-NEXT: store [16 x i32] [[TMP5]], ptr addrspace(5) [[TMP4]], align 4 +// AMDGCN-NEXT: call void @llvm.memcpy.p1.p5.i64(ptr addrspace(1) align 4 [[ARRAYIDX]], ptr addrspace(5) align 4 [[TMP]], i64 64, i1 false) +// AMDGCN-NEXT: ret void +// +// AMDGCN20-LABEL: define dso_local amdgpu_kernel void @ker( +// AMDGCN20-SAME: ptr addrspace(1) noundef align 4 [[IN:%.*]], ptr addrspace(1) noundef align 4 [[OUT:%.*]]) #[[ATTR1:[0-9]+]] !kernel_arg_addr_space [[META4:![0-9]+]] !kernel_arg_access_qual [[META5:![0-9]+]] !kernel_arg_type [[META6:![0-9]+]] !kernel_arg_base_type [[META6]] !kernel_arg_type_qual [[META7:![0-9]+]] { +// AMDGCN20-NEXT: [[ENTRY:.*:]] +// AMDGCN20-NEXT: [[IN_ADDR:%.*]] = alloca ptr addrspace(1), align 8, addrspace(5) +// AMDGCN20-NEXT: [[OUT_ADDR:%.*]] = alloca ptr addrspace(1), align 8, addrspace(5) +// AMDGCN20-NEXT: [[TMP:%.*]] = alloca [[STRUCT_MAT4X4:%.*]], align 4, addrspace(5) +// AMDGCN20-NEXT: store ptr addrspace(1) [[IN]], ptr addrspace(5) [[IN_ADDR]], align 8 +// AMDGCN20-NEXT: store ptr addrspace(1) [[OUT]], ptr addrspace(5) [[OUT_ADDR]], align 8 +// AMDGCN20-NEXT: [[TMP0:%.*]] = load ptr addrspace(1), ptr addrspace(5) [[OUT_ADDR]], align 8 +// AMDGCN20-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [[STRUCT_MAT4X4]], ptr addrspace(1) [[TMP0]], i64 0 +// AMDGCN20-NEXT: [[TMP1:%.*]] = load ptr addrspace(1), ptr addrspace(5) [[IN_ADDR]], align 8 +// AMDGCN20-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds [[STRUCT_MAT3X3:%.*]], ptr addrspace(1) [[TMP1]], i64 1 +// AMDGCN20-NEXT: [[TMP2:%.*]] = getelementptr inbounds nuw [[STRUCT_MAT3X3]], ptr addrspace(1) [[ARRAYIDX1]], i32 0, i32 0 +// AMDGCN20-NEXT: [[TMP3:%.*]] = load [9 x i32], ptr addrspace(1) [[TMP2]], align 4 +// AMDGCN20-NEXT: [[CALL:%.*]] = call [[STRUCT_MAT4X4]] @[[FOO:[a-zA-Z0-9_$\"\\.-]*[a-zA-Z_$\"\\.-][a-zA-Z0-9_$\"\\.-]*]]([9 x i32] [[TMP3]]) #[[ATTR3:[0-9]+]] +// AMDGCN20-NEXT: [[TMP4:%.*]] = getelementptr inbounds nuw [[STRUCT_MAT4X4]], ptr addrspace(5) [[TMP]], i32 0, i32 0 +// AMDGCN20-NEXT: [[TMP5:%.*]] = extractvalue [[STRUCT_MAT4X4]] [[CALL]], 0 +// AMDGCN20-NEXT: store [16 x i32] [[TMP5]], ptr addrspace(5) [[TMP4]], align 4 +// AMDGCN20-NEXT: call void @llvm.memcpy.p1.p5.i64(ptr addrspace(1) align 4 [[ARRAYIDX]], ptr addrspace(5) align 4 [[TMP]], i64 64, i1 false) +// AMDGCN20-NEXT: ret void +// +// SPIR-LABEL: define dso_local spir_kernel void @ker( +// SPIR-SAME: ptr addrspace(1) noundef align 4 [[IN:%.*]], ptr addrspace(1) noundef align 4 [[OUT:%.*]]) #[[ATTR1:[0-9]+]] !kernel_arg_addr_space [[META3:![0-9]+]] !kernel_arg_access_qual [[META4:![0-9]+]] !kernel_arg_type [[META5:![0-9]+]] !kernel_arg_base_type [[META5]] !kernel_arg_type_qual [[META6:![0-9]+]] { +// SPIR-NEXT: [[ENTRY:.*:]] +// SPIR-NEXT: [[IN_ADDR:%.*]] = alloca ptr addrspace(1), align 4 +// SPIR-NEXT: [[OUT_ADDR:%.*]] = alloca ptr addrspace(1), align 4 +// SPIR-NEXT: [[TMP:%.*]] = alloca [[STRUCT_MAT4X4:%.*]], align 4 +// SPIR-NEXT: [[BYVAL_TEMP:%.*]] = alloca [[STRUCT_MAT3X3:%.*]], align 4 +// SPIR-NEXT: store ptr addrspace(1) [[IN]], ptr [[IN_ADDR]], align 4 +// SPIR-NEXT: store ptr addrspace(1) [[OUT]], ptr [[OUT_ADDR]], align 4 +// SPIR-NEXT: [[TMP0:%.*]] = load ptr addrspace(1), ptr [[OUT_ADDR]], align 4 +// SPIR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [[STRUCT_MAT4X4]], ptr addrspace(1) [[TMP0]], i32 0 +// SPIR-NEXT: [[TMP1:%.*]] = load ptr addrspace(1), ptr [[IN_ADDR]], align 4 +// SPIR-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds [[STRUCT_MAT3X3]], ptr addrspace(1) [[TMP1]], i32 1 +// SPIR-NEXT: call void @llvm.memcpy.p0.p1.i32(ptr align 4 [[BYVAL_TEMP]], ptr addrspace(1) align 4 [[ARRAYIDX1]], i32 36, i1 false) +// SPIR-NEXT: call spir_func void @foo(ptr dead_on_unwind writable sret([[STRUCT_MAT4X4]]) align 4 [[TMP]], ptr noundef byval([[STRUCT_MAT3X3]]) align 4 [[BYVAL_TEMP]]) #[[ATTR3:[0-9]+]] +// SPIR-NEXT: call void @llvm.memcpy.p1.p0.i32(ptr addrspace(1) align 4 [[ARRAYIDX]], ptr align 4 [[TMP]], i32 64, i1 false) +// SPIR-NEXT: ret void +// +// AMDGCN30-GVAR-LABEL: define dso_local amdgpu_kernel void @ker( +// AMDGCN30-GVAR-SAME: ptr addrspace(1) noundef align 4 [[IN:%.*]], ptr addrspace(1) noundef align 4 [[OUT:%.*]]) #[[ATTR1:[0-9]+]] !kernel_arg_addr_space [[META4:![0-9]+]] !kernel_arg_access_qual [[META5:![0-9]+]] !kernel_arg_type [[META6:![0-9]+]] !kernel_arg_base_type [[META6]] !kernel_arg_type_qual [[META7:![0-9]+]] { +// AMDGCN30-GVAR-NEXT: [[ENTRY:.*:]] +// AMDGCN30-GVAR-NEXT: [[IN_ADDR:%.*]] = alloca ptr addrspace(1), align 8, addrspace(5) +// AMDGCN30-GVAR-NEXT: [[OUT_ADDR:%.*]] = alloca ptr addrspace(1), align 8, addrspace(5) +// AMDGCN30-GVAR-NEXT: [[TMP:%.*]] = alloca [[STRUCT_MAT4X4:%.*]], align 4, addrspace(5) +// AMDGCN30-GVAR-NEXT: store ptr addrspace(1) [[IN]], ptr addrspace(5) [[IN_ADDR]], align 8 +// AMDGCN30-GVAR-NEXT: store ptr addrspace(1) [[OUT]], ptr addrspace(5) [[OUT_ADDR]], align 8 +// AMDGCN30-GVAR-NEXT: [[TMP0:%.*]] = load ptr addrspace(1), ptr addrspace(5) [[OUT_ADDR]], align 8 +// AMDGCN30-GVAR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [[STRUCT_MAT4X4]], ptr addrspace(1) [[TMP0]], i64 0 +// AMDGCN30-GVAR-NEXT: [[TMP1:%.*]] = load ptr addrspace(1), ptr addrspace(5) [[IN_ADDR]], align 8 +// AMDGCN30-GVAR-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds [[STRUCT_MAT3X3:%.*]], ptr addrspace(1) [[TMP1]], i64 1 +// AMDGCN30-GVAR-NEXT: [[TMP2:%.*]] = getelementptr inbounds nuw [[STRUCT_MAT3X3]], ptr addrspace(1) [[ARRAYIDX1]], i32 0, i32 0 +// AMDGCN30-GVAR-NEXT: [[TMP3:%.*]] = load [9 x i32], ptr addrspace(1) [[TMP2]], align 4 +// AMDGCN30-GVAR-NEXT: [[CALL:%.*]] = call [[STRUCT_MAT4X4]] @[[FOO:[a-zA-Z0-9_$\"\\.-]*[a-zA-Z_$\"\\.-][a-zA-Z0-9_$\"\\.-]*]]([9 x i32] [[TMP3]]) #[[ATTR3:[0-9]+]] +// AMDGCN30-GVAR-NEXT: [[TMP4:%.*]] = getelementptr inbounds nuw [[STRUCT_MAT4X4]], ptr addrspace(5) [[TMP]], i32 0, i32 0 +// AMDGCN30-GVAR-NEXT: [[TMP5:%.*]] = extractvalue [[STRUCT_MAT4X4]] [[CALL]], 0 +// AMDGCN30-GVAR-NEXT: store [16 x i32] [[TMP5]], ptr addrspace(5) [[TMP4]], align 4 +// AMDGCN30-GVAR-NEXT: call void @llvm.memcpy.p1.p5.i64(ptr addrspace(1) align 4 [[ARRAYIDX]], ptr addrspace(5) align 4 [[TMP]], i64 64, i1 false) +// AMDGCN30-GVAR-NEXT: ret void +// +// AMDGCN30-LABEL: define dso_local amdgpu_kernel void @ker( +// AMDGCN30-SAME: ptr addrspace(1) noundef align 4 [[IN:%.*]], ptr addrspace(1) noundef align 4 [[OUT:%.*]]) #[[ATTR1:[0-9]+]] !kernel_arg_addr_space [[META4:![0-9]+]] !kernel_arg_access_qual [[META5:![0-9]+]] !kernel_arg_type [[META6:![0-9]+]] !kernel_arg_base_type [[META6]] !kernel_arg_type_qual [[META7:![0-9]+]] { +// AMDGCN30-NEXT: [[ENTRY:.*:]] +// AMDGCN30-NEXT: [[IN_ADDR:%.*]] = alloca ptr addrspace(1), align 8, addrspace(5) +// AMDGCN30-NEXT: [[OUT_ADDR:%.*]] = alloca ptr addrspace(1), align 8, addrspace(5) +// AMDGCN30-NEXT: [[TMP:%.*]] = alloca [[STRUCT_MAT4X4:%.*]], align 4, addrspace(5) +// AMDGCN30-NEXT: store ptr addrspace(1) [[IN]], ptr addrspace(5) [[IN_ADDR]], align 8 +// AMDGCN30-NEXT: store ptr addrspace(1) [[OUT]], ptr addrspace(5) [[OUT_ADDR]], align 8 +// AMDGCN30-NEXT: [[TMP0:%.*]] = load ptr addrspace(1), ptr addrspace(5) [[OUT_ADDR]], align 8 +// AMDGCN30-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [[STRUCT_MAT4X4]], ptr addrspace(1) [[TMP0]], i64 0 +// AMDGCN30-NEXT: [[TMP1:%.*]] = load ptr addrspace(1), ptr addrspace(5) [[IN_ADDR]], align 8 +// AMDGCN30-NEXT: [[ARRAYIDX1:%.*]] = getelementptr inbounds [[STRUCT_MAT3X3:%.*]], ptr addrspace(1) [[TMP1]], i64 1 +// AMDGCN30-NEXT: [[TMP2:%.*]] = getelementptr inbounds nuw [[STRUCT_MAT3X3]], ptr addrspace(1) [[ARRAYIDX1]], i32 0, i32 0 +// AMDGCN30-NEXT: [[TMP3:%.*]] = load [9 x i32], ptr addrspace(1) [[TMP2]], align 4 +// AMDGCN30-NEXT: [[CALL:%.*]] = call [[STRUCT_MAT4X4]] @[[FOO:[a-zA-Z0-9_$\"\\.-]*[a-zA-Z_$\"\\.-][a-zA-Z0-9_$\"\\.-]*]]([9 x i32] [[TMP3]]) #[[ATTR3:[0-9]+]] +// AMDGCN30-NEXT: [[TMP4:%.*]] = getelementptr inbounds nuw [[STRUCT_MAT4X4]], ptr addrspace(5) [[TMP]], i32 0, i32 0 +// AMDGCN30-NEXT: [[TMP5:%.*]] = extractvalue [[STRUCT_MAT4X4]] [[CALL]], 0 +// AMDGCN30-NEXT: store [16 x i32] [[TMP5]], ptr addrspace(5) [[TMP4]], align 4 +// AMDGCN30-NEXT: call void @llvm.memcpy.p1.p5.i64(ptr addrspace(1) align 4 [[ARRAYIDX]], ptr addrspace(5) align 4 [[TMP]], i64 64, i1 false) +// AMDGCN30-NEXT: ret void +// kernel void ker(global Mat3X3 *in, global Mat4X4 *out) { out[0] = foo(in[1]); } -// X86-LABEL: define{{.*}} void @foo_large(ptr dead_on_unwind noalias writable sret(%struct.Mat64X64) align 4 %agg.result, ptr noundef byval(%struct.Mat32X32) align 4 %in) -// AMDGCN-LABEL: define{{.*}} void @foo_large(ptr addrspace(5) dead_on_unwind noalias writable sret(%struct.Mat64X64) align 4 %agg.result, ptr addrspace(5) noundef byref(%struct.Mat32X32) align 4 %{{.*}} -// AMDGCN: %in = alloca %struct.Mat32X32, align 4, addrspace(5) -// AMDGCN-NEXT: call void @llvm.memcpy.p5.p5.i64(ptr addrspace(5) align 4 %in, ptr addrspace(5) align 4 %{{.*}}, i64 4096, i1 false) +// +// X86-LABEL: define void @foo_large( +// X86-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_MAT64X64:%.*]]) align 4 [[AGG_RESULT:%.*]], ptr noundef byval([[STRUCT_MAT32X32:%.*]]) align 4 [[IN:%.*]]) #[[ATTR0]] { +// X86-NEXT: [[ENTRY:.*:]] +// X86-NEXT: [[RESULT_PTR:%.*]] = alloca ptr, align 4 +// X86-NEXT: store ptr [[AGG_RESULT]], ptr [[RESULT_PTR]], align 4 +// X86-NEXT: ret void +// +// AMDGCN-LABEL: define dso_local void @foo_large( +// AMDGCN-SAME: ptr addrspace(5) dead_on_unwind noalias writable sret([[STRUCT_MAT64X64:%.*]]) align 4 [[AGG_RESULT:%.*]], ptr addrspace(5) noundef byref([[STRUCT_MAT32X32:%.*]]) align 4 [[TMP0:%.*]]) #[[ATTR0]] { +// AMDGCN-NEXT: [[ENTRY:.*:]] +// AMDGCN-NEXT: [[IN:%.*]] = alloca [[STRUCT_MAT32X32]], align 4, addrspace(5) +// AMDGCN-NEXT: call void @llvm.memcpy.p5.p5.i64(ptr addrspace(5) align 4 [[IN]], ptr addrspace(5) align 4 [[TMP0]], i64 4096, i1 false) +// AMDGCN-NEXT: ret void +// +// AMDGCN20-LABEL: define dso_local void @foo_large( +// AMDGCN20-SAME: ptr addrspace(5) dead_on_unwind noalias writable sret([[STRUCT_MAT64X64:%.*]]) align 4 [[AGG_RESULT:%.*]], ptr addrspace(5) noundef byref([[STRUCT_MAT32X32:%.*]]) align 4 [[TMP0:%.*]]) #[[ATTR0]] { +// AMDGCN20-NEXT: [[ENTRY:.*:]] +// AMDGCN20-NEXT:... [truncated] `````````` </details> https://github.com/llvm/llvm-project/pull/112554 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
