Skip to content

[clang][OpenCL][CodeGen][AMDGPU] Do not use private as the default AS for when generic is available #112442

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 11 commits into from
Oct 22, 2024
Merged

[clang][OpenCL][CodeGen][AMDGPU] Do not use private as the default AS for when generic is available #112442

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
11 commits
Select commit Hold shift + click to select a range
ca91962
`private` is a terrible default AS, and is terrible and wrong for OCL…
AlexVlx Oct 15, 2024
a742e92
Fix formatting.
AlexVlx Oct 15, 2024
2b2c236
Merge branch 'main' of https://github.com/llvm/llvm-project into priv…
AlexVlx Oct 16, 2024
667fd20
Merge branch 'main' of https://github.com/llvm/llvm-project into priv…
AlexVlx Oct 16, 2024
7fcb506
Strip less.
AlexVlx Oct 16, 2024
5696c42
Merge branch 'main' of https://github.com/llvm/llvm-project into priv…
AlexVlx Oct 16, 2024
c4e603d
Fix borked merge.
AlexVlx Oct 16, 2024
e7fdefd
Update tests.
AlexVlx Oct 16, 2024
428fdd2
Merge branch 'main' of https://github.com/llvm/llvm-project into priv…
AlexVlx Oct 21, 2024
988a1b8
Fix mangled test.
AlexVlx Oct 21, 2024
85bba76
Re-add missing `CHECK`.
AlexVlx Oct 21, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 3 additions & 3 deletions clang/lib/Basic/Targets/AMDGPU.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -260,9 +260,9 @@ AMDGPUTargetInfo::AMDGPUTargetInfo(const llvm::Triple &Triple,
void AMDGPUTargetInfo::adjust(DiagnosticsEngine &Diags, LangOptions &Opts) {
TargetInfo::adjust(Diags, Opts);
// ToDo: There are still a few places using default address space as private
// address space in OpenCL, which needs to be cleaned up, then Opts.OpenCL
// can be removed from the following line.
setAddressSpaceMap(/*DefaultIsPrivate=*/Opts.OpenCL ||
// address space in OpenCL, which needs to be cleaned up, then the references
// to OpenCL can be removed from the following line.
setAddressSpaceMap((Opts.OpenCL && !Opts.OpenCLGenericAddressSpace) ||
!isAMDGCN(getTriple()));
}

Expand Down
3 changes: 2 additions & 1 deletion clang/lib/CodeGen/CGBlocks.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1397,7 +1397,8 @@ void CodeGenFunction::setBlockContextParameter(const ImplicitParamDecl *D,
DI->setLocation(D->getLocation());
DI->EmitDeclareOfBlockLiteralArgVariable(
*BlockInfo, D->getName(), argNum,
cast<llvm::AllocaInst>(alloc.getPointer()), Builder);
cast<llvm::AllocaInst>(alloc.getPointer()->stripPointerCasts()),
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I would hope alloc would be the direct alloca reference to begin with

Builder);
}
}

Expand Down
5 changes: 3 additions & 2 deletions clang/lib/CodeGen/CGBuiltin.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5858,7 +5858,8 @@ RValue CodeGenFunction::EmitBuiltinExpr(const GlobalDecl GD, unsigned BuiltinID,
auto Tmp = CreateMemTemp(SizeArrayTy, "block_sizes");
llvm::Value *TmpPtr = Tmp.getPointer();
llvm::Value *TmpSize = EmitLifetimeStart(
CGM.getDataLayout().getTypeAllocSize(Tmp.getElementType()), TmpPtr);
CGM.getDataLayout().getTypeAllocSize(Tmp.getElementType()),
TmpPtr->stripPointerCasts());
llvm::Value *ElemPtr;
// Each of the following arguments specifies the size of the corresponding
// argument passed to the enqueued block.
Expand Down Expand Up @@ -5903,7 +5904,7 @@ RValue CodeGenFunction::EmitBuiltinExpr(const GlobalDecl GD, unsigned BuiltinID,
auto Call = RValue::get(
EmitRuntimeCall(CGM.CreateRuntimeFunction(FTy, Name), Args));
if (TmpSize)
EmitLifetimeEnd(TmpSize, TmpPtr);
EmitLifetimeEnd(TmpSize, TmpPtr->stripPointerCasts());
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm surprised CreateMemTemp isn't returning the raw alloca. I would expect it to return the direct reference, and not the addrspacecasted one.

Assuming fixing that is hard, should just do the strip once and use it consistently

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I would extremely strongly prefer not to mess with this ball of yarn or other pieces of functionality (specifically, CreateMemTemp), as part of what is essentially a point fix of a long standing bug. I'll leave a TODO in place for someone that is braver and more skilled.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can still just do the stripPointerCasts once above

return Call;
}
// Any calls now have event arguments passed.
Expand Down
1,521 changes: 1,440 additions & 81 deletions clang/test/CodeGenOpenCL/addr-space-struct-arg.cl
View file
Open in desktop

Large diffs are not rendered by default.

126 changes: 85 additions & 41 deletions clang/test/CodeGenOpenCL/amdgcn-automatic-variable.cl
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,67 +1,111 @@
// RUN: %clang_cc1 -O0 -cl-std=CL1.2 -triple amdgcn---amdgizcl -emit-llvm %s -o - | FileCheck -check-prefixes=CHECK,CL12 %s
// RUN: %clang_cc1 -O0 -cl-std=CL2.0 -triple amdgcn---amdgizcl -emit-llvm %s -o - | FileCheck -check-prefixes=CHECK,CL20 %s
// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
// RUN: %clang_cc1 -O0 -cl-std=CL1.2 -triple amdgcn---amdgizcl -emit-llvm %s -o - | FileCheck -check-prefixes=CL12 %s
// RUN: %clang_cc1 -O0 -cl-std=CL2.0 -triple amdgcn---amdgizcl -emit-llvm %s -o - | FileCheck -check-prefixes=CL20 %s

// CL12-LABEL: define{{.*}} void @func1(ptr addrspace(5) noundef %x)
// CL20-LABEL: define{{.*}} void @func1(ptr noundef %x)
// CL12-LABEL: define dso_local void @func1(
// CL12-SAME: ptr addrspace(5) noundef [[X:%.*]]) #[[ATTR0:[0-9]+]] {
// CL12-NEXT: [[ENTRY:.*:]]
// CL12-NEXT: [[X_ADDR:%.*]] = alloca ptr addrspace(5), align 4, addrspace(5)
// CL12-NEXT: store ptr addrspace(5) [[X]], ptr addrspace(5) [[X_ADDR]], align 4
// CL12-NEXT: [[TMP0:%.*]] = load ptr addrspace(5), ptr addrspace(5) [[X_ADDR]], align 4
// CL12-NEXT: store i32 1, ptr addrspace(5) [[TMP0]], align 4
// CL12-NEXT: ret void
//
// CL20-LABEL: define dso_local void @func1(
// CL20-SAME: ptr noundef [[X:%.*]]) #[[ATTR0:[0-9]+]] {
// CL20-NEXT: [[ENTRY:.*:]]
// CL20-NEXT: [[X_ADDR:%.*]] = alloca ptr, align 8, addrspace(5)
// CL20-NEXT: [[X_ADDR_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[X_ADDR]] to ptr
// CL20-NEXT: store ptr [[X]], ptr [[X_ADDR_ASCAST]], align 8
// CL20-NEXT: [[TMP0:%.*]] = load ptr, ptr [[X_ADDR_ASCAST]], align 8
// CL20-NEXT: store i32 1, ptr [[TMP0]], align 4
// CL20-NEXT: ret void
//
void func1(int *x) {
// CL12: %[[x_addr:.*]] = alloca ptr addrspace(5){{.*}}addrspace(5)
// CL12: store ptr addrspace(5) %x, ptr addrspace(5) %[[x_addr]]
// CL12: %[[r0:.*]] = load ptr addrspace(5), ptr addrspace(5) %[[x_addr]]
// CL12: store i32 1, ptr addrspace(5) %[[r0]]
// CL20: %[[x_addr:.*]] = alloca ptr{{.*}}addrspace(5)
// CL20: store ptr %x, ptr addrspace(5) %[[x_addr]]
// CL20: %[[r0:.*]] = load ptr, ptr addrspace(5) %[[x_addr]]
// CL20: store i32 1, ptr %[[r0]]
*x = 1;
}

// CHECK-LABEL: define{{.*}} void @func2()
// CL12-LABEL: define dso_local void @func2(
// CL12-SAME: ) #[[ATTR0]] {
// CL12-NEXT: [[ENTRY:.*:]]
// CL12-NEXT: [[LV1:%.*]] = alloca i32, align 4, addrspace(5)
// CL12-NEXT: [[LV2:%.*]] = alloca i32, align 4, addrspace(5)
// CL12-NEXT: [[LA:%.*]] = alloca [100 x i32], align 4, addrspace(5)
// CL12-NEXT: [[LP1:%.*]] = alloca ptr addrspace(5), align 4, addrspace(5)
// CL12-NEXT: [[LP2:%.*]] = alloca ptr addrspace(5), align 4, addrspace(5)
// CL12-NEXT: [[LVC:%.*]] = alloca i32, align 4, addrspace(5)
// CL12-NEXT: store i32 1, ptr addrspace(5) [[LV1]], align 4
// CL12-NEXT: store i32 2, ptr addrspace(5) [[LV2]], align 4
// CL12-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [100 x i32], ptr addrspace(5) [[LA]], i64 0, i64 0
// CL12-NEXT: store i32 3, ptr addrspace(5) [[ARRAYIDX]], align 4
// CL12-NEXT: store ptr addrspace(5) [[LV1]], ptr addrspace(5) [[LP1]], align 4
// CL12-NEXT: [[ARRAYDECAY:%.*]] = getelementptr inbounds [100 x i32], ptr addrspace(5) [[LA]], i64 0, i64 0
// CL12-NEXT: store ptr addrspace(5) [[ARRAYDECAY]], ptr addrspace(5) [[LP2]], align 4
// CL12-NEXT: call void @func1(ptr addrspace(5) noundef [[LV1]]) #[[ATTR2:[0-9]+]]
// CL12-NEXT: store i32 4, ptr addrspace(5) [[LVC]], align 4
// CL12-NEXT: store i32 4, ptr addrspace(5) [[LV1]], align 4
// CL12-NEXT: ret void
//
// CL20-LABEL: define dso_local void @func2(
// CL20-SAME: ) #[[ATTR0]] {
// CL20-NEXT: [[ENTRY:.*:]]
// CL20-NEXT: [[LV1:%.*]] = alloca i32, align 4, addrspace(5)
// CL20-NEXT: [[LV2:%.*]] = alloca i32, align 4, addrspace(5)
// CL20-NEXT: [[LA:%.*]] = alloca [100 x i32], align 4, addrspace(5)
// CL20-NEXT: [[LP1:%.*]] = alloca ptr, align 8, addrspace(5)
// CL20-NEXT: [[LP2:%.*]] = alloca ptr, align 8, addrspace(5)
// CL20-NEXT: [[LVC:%.*]] = alloca i32, align 4, addrspace(5)
// CL20-NEXT: [[LV1_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[LV1]] to ptr
// CL20-NEXT: [[LV2_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[LV2]] to ptr
// CL20-NEXT: [[LA_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[LA]] to ptr
// CL20-NEXT: [[LP1_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[LP1]] to ptr
// CL20-NEXT: [[LP2_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[LP2]] to ptr
// CL20-NEXT: [[LVC_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[LVC]] to ptr
// CL20-NEXT: store i32 1, ptr [[LV1_ASCAST]], align 4
// CL20-NEXT: store i32 2, ptr [[LV2_ASCAST]], align 4
// CL20-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [100 x i32], ptr [[LA_ASCAST]], i64 0, i64 0
// CL20-NEXT: store i32 3, ptr [[ARRAYIDX]], align 4
// CL20-NEXT: store ptr [[LV1_ASCAST]], ptr [[LP1_ASCAST]], align 8
// CL20-NEXT: [[ARRAYDECAY:%.*]] = getelementptr inbounds [100 x i32], ptr [[LA_ASCAST]], i64 0, i64 0
// CL20-NEXT: store ptr [[ARRAYDECAY]], ptr [[LP2_ASCAST]], align 8
// CL20-NEXT: call void @func1(ptr noundef [[LV1_ASCAST]]) #[[ATTR2:[0-9]+]]
// CL20-NEXT: store i32 4, ptr [[LVC_ASCAST]], align 4
// CL20-NEXT: store i32 4, ptr [[LV1_ASCAST]], align 4
// CL20-NEXT: ret void
//
void func2(void) {
// CHECK: %lv1 = alloca i32, align 4, addrspace(5)
// CHECK: %lv2 = alloca i32, align 4, addrspace(5)
// CHECK: %la = alloca [100 x i32], align 4, addrspace(5)
// CL12: %lp1 = alloca ptr addrspace(5), align 4, addrspace(5)
// CL12: %lp2 = alloca ptr addrspace(5), align 4, addrspace(5)
// CL20: %lp1 = alloca ptr, align 8, addrspace(5)
// CL20: %lp2 = alloca ptr, align 8, addrspace(5)
// CHECK: %lvc = alloca i32, align 4, addrspace(5)

// CHECK: store i32 1, ptr addrspace(5) %lv1
int lv1;
lv1 = 1;
// CHECK: store i32 2, ptr addrspace(5) %lv2
int lv2 = 2;

// CHECK: %[[arrayidx:.*]] = getelementptr inbounds [100 x i32], ptr addrspace(5) %la, i64 0, i64 0
// CHECK: store i32 3, ptr addrspace(5) %[[arrayidx]], align 4
int la[100];
la[0] = 3;

// CL12: store ptr addrspace(5) %lv1, ptr addrspace(5) %lp1, align 4
// CL20: %[[r0:.*]] = addrspacecast ptr addrspace(5) %lv1 to ptr
// CL20: store ptr %[[r0]], ptr addrspace(5) %lp1, align 8
int *lp1 = &lv1;

// CHECK: %[[arraydecay:.*]] = getelementptr inbounds [100 x i32], ptr addrspace(5) %la, i64 0, i64 0
// CL12: store ptr addrspace(5) %[[arraydecay]], ptr addrspace(5) %lp2, align 4
// CL20: %[[r1:.*]] = addrspacecast ptr addrspace(5) %[[arraydecay]] to ptr
// CL20: store ptr %[[r1]], ptr addrspace(5) %lp2, align 8
int *lp2 = la;

// CL12: call void @func1(ptr addrspace(5) noundef %lv1)
// CL20: %[[r2:.*]] = addrspacecast ptr addrspace(5) %lv1 to ptr
// CL20: call void @func1(ptr noundef %[[r2]])
func1(&lv1);

// CHECK: store i32 4, ptr addrspace(5) %lvc
// CHECK: store i32 4, ptr addrspace(5) %lv1
const int lvc = 4;
lv1 = lvc;
}

// CHECK-LABEL: define{{.*}} void @func3()
// CHECK: %a = alloca [16 x [1 x float]], align 4, addrspace(5)
// CHECK: call void @llvm.memset.p5.i64(ptr addrspace(5) align 4 %a, i8 0, i64 64, i1 false)
// CL12-LABEL: define dso_local void @func3(
// CL12-SAME: ) #[[ATTR0]] {
// CL12-NEXT: [[ENTRY:.*:]]
// CL12-NEXT: [[A:%.*]] = alloca [16 x [1 x float]], align 4, addrspace(5)
// CL12-NEXT: call void @llvm.memset.p5.i64(ptr addrspace(5) align 4 [[A]], i8 0, i64 64, i1 false)
// CL12-NEXT: ret void
//
// CL20-LABEL: define dso_local void @func3(
// CL20-SAME: ) #[[ATTR0]] {
// CL20-NEXT: [[ENTRY:.*:]]
// CL20-NEXT: [[A:%.*]] = alloca [16 x [1 x float]], align 4, addrspace(5)
// CL20-NEXT: [[A_ASCAST:%.*]] = addrspacecast ptr addrspace(5) [[A]] to ptr
// CL20-NEXT: call void @llvm.memset.p0.i64(ptr align 4 [[A_ASCAST]], i8 0, i64 64, i1 false)
// CL20-NEXT: ret void
//
void func3(void) {
float a[16][1] = {{0.}};
}
Loading
Loading