sidorovd created this revision.
sidorovd added reviewers: yaxunl, Anastasia.
Herald added a subscriber: cfe-commits.
If a function argument is byval and RV is located in default or alloca address
space
an optimization of creating addrspacecast instead of memcpy is performed. That
is
not correct for OpenCL, where that can lead to a situation of address space
casting
from __private * to __global *. See an example below:
typedef struct {
int x;
} MyStruct;
void foo(MyStruct val) {}
kernel void KernelOneMember(__global MyStruct* x) {
foo (*x);
}
for this code clang generated following IR:
...
%0 = load %struct.MyStruct addrspace(1)*, %struct.MyStruct addrspace(1)**
%x.addr, align 4
%1 = addrspacecast %struct.MyStruct addrspace(1)* %0 to %struct.MyStruct*
...
So the optimization was disallowed for OpenCL if RV is located in an address
space
different than that of the argument (0).
Repository:
rC Clang
https://reviews.llvm.org/D54947
Files:
lib/CodeGen/CGCall.cpp
test/CodeGenOpenCL/addr-space-struct-arg.cl
Index: test/CodeGenOpenCL/addr-space-struct-arg.cl
===================================================================
--- test/CodeGenOpenCL/addr-space-struct-arg.cl
+++ test/CodeGenOpenCL/addr-space-struct-arg.cl
@@ -1,6 +1,7 @@
// RUN: %clang_cc1 %s -emit-llvm -o - -O0 -finclude-default-header
-ffake-address-space-map -triple i686-pc-darwin | FileCheck -enable-var-scope
-check-prefixes=COM,X86 %s
// RUN: %clang_cc1 %s -emit-llvm -o - -O0 -finclude-default-header -triple
amdgcn | FileCheck -enable-var-scope -check-prefixes=COM,AMDGCN %s
// RUN: %clang_cc1 %s -emit-llvm -o - -cl-std=CL2.0 -O0
-finclude-default-header -triple amdgcn | FileCheck -enable-var-scope
-check-prefixes=COM,AMDGCN,AMDGCN20 %s
+// RUN: %clang_cc1 %s -emit-llvm -o - -cl-std=CL1.2 -O0
-finclude-default-header -triple spir-unknown-unknown-unknown | FileCheck
-enable-var-scope -check-prefixes=SPIR %s
typedef struct {
int cells[9];
@@ -130,6 +131,12 @@
FuncOneMember(u);
}
+// SPIR: call void @llvm.memcpy.p0i8.p1i8.i32
+// SPIR-NOT: addrspacecast
+kernel void KernelOneMemberSpir(global struct StructOneMember* u) {
+ FuncOneMember(*u);
+}
+
// AMDGCN-LABEL: define amdgpu_kernel void @KernelLargeOneMember(
// AMDGCN: %[[U:.*]] = alloca %struct.LargeStructOneMember, align 8,
addrspace(5)
// AMDGCN: store %struct.LargeStructOneMember %u.coerce,
%struct.LargeStructOneMember addrspace(5)* %[[U]], align 8
Index: lib/CodeGen/CGCall.cpp
===================================================================
--- lib/CodeGen/CGCall.cpp
+++ lib/CodeGen/CGCall.cpp
@@ -3954,15 +3954,30 @@
} else if (I->hasLValue()) {
auto LV = I->getKnownLValue();
auto AS = LV.getAddressSpace();
+
if ((!ArgInfo.getIndirectByVal() &&
(LV.getAlignment() >=
- getContext().getTypeAlignInChars(I->Ty))) ||
- (ArgInfo.getIndirectByVal() &&
- ((AS != LangAS::Default && AS != LangAS::opencl_private &&
- AS != CGM.getASTAllocaAddressSpace())))) {
+ getContext().getTypeAlignInChars(I->Ty)))) {
+ NeedCopy = true;
+ }
+ if (!getLangOpts().OpenCL) {
+ if ((ArgInfo.getIndirectByVal() &&
+ (AS != LangAS::Default &&
+ AS != CGM.getASTAllocaAddressSpace()))) {
+ NeedCopy = true;
+ }
+ }
+ // For OpenCL even if RV is located in default or alloca address
space
+ // we don't want to perform address space cast for it, since that
+ // leads to casting __private * (default addr space in OpenCL) to
+ // __global * which is not valid. Create memcpy call instead.
+ else if ((ArgInfo.getIndirectByVal() &&
+ Addr.getType()->getAddressSpace() != IRFuncTy->
+ getParamType(FirstIRArg)->getPointerAddressSpace())) {
NeedCopy = true;
}
}
+
if (NeedCopy) {
// Create an aligned temporary, and copy to it.
Address AI = CreateMemTempWithoutCast(
Index: test/CodeGenOpenCL/addr-space-struct-arg.cl
===================================================================
--- test/CodeGenOpenCL/addr-space-struct-arg.cl
+++ test/CodeGenOpenCL/addr-space-struct-arg.cl
@@ -1,6 +1,7 @@
// RUN: %clang_cc1 %s -emit-llvm -o - -O0 -finclude-default-header -ffake-address-space-map -triple i686-pc-darwin | FileCheck -enable-var-scope -check-prefixes=COM,X86 %s
// RUN: %clang_cc1 %s -emit-llvm -o - -O0 -finclude-default-header -triple amdgcn | FileCheck -enable-var-scope -check-prefixes=COM,AMDGCN %s
// RUN: %clang_cc1 %s -emit-llvm -o - -cl-std=CL2.0 -O0 -finclude-default-header -triple amdgcn | FileCheck -enable-var-scope -check-prefixes=COM,AMDGCN,AMDGCN20 %s
+// RUN: %clang_cc1 %s -emit-llvm -o - -cl-std=CL1.2 -O0 -finclude-default-header -triple spir-unknown-unknown-unknown | FileCheck -enable-var-scope -check-prefixes=SPIR %s
typedef struct {
int cells[9];
@@ -130,6 +131,12 @@
FuncOneMember(u);
}
+// SPIR: call void @llvm.memcpy.p0i8.p1i8.i32
+// SPIR-NOT: addrspacecast
+kernel void KernelOneMemberSpir(global struct StructOneMember* u) {
+ FuncOneMember(*u);
+}
+
// AMDGCN-LABEL: define amdgpu_kernel void @KernelLargeOneMember(
// AMDGCN: %[[U:.*]] = alloca %struct.LargeStructOneMember, align 8, addrspace(5)
// AMDGCN: store %struct.LargeStructOneMember %u.coerce, %struct.LargeStructOneMember addrspace(5)* %[[U]], align 8
Index: lib/CodeGen/CGCall.cpp
===================================================================
--- lib/CodeGen/CGCall.cpp
+++ lib/CodeGen/CGCall.cpp
@@ -3954,15 +3954,30 @@
} else if (I->hasLValue()) {
auto LV = I->getKnownLValue();
auto AS = LV.getAddressSpace();
+
if ((!ArgInfo.getIndirectByVal() &&
(LV.getAlignment() >=
- getContext().getTypeAlignInChars(I->Ty))) ||
- (ArgInfo.getIndirectByVal() &&
- ((AS != LangAS::Default && AS != LangAS::opencl_private &&
- AS != CGM.getASTAllocaAddressSpace())))) {
+ getContext().getTypeAlignInChars(I->Ty)))) {
+ NeedCopy = true;
+ }
+ if (!getLangOpts().OpenCL) {
+ if ((ArgInfo.getIndirectByVal() &&
+ (AS != LangAS::Default &&
+ AS != CGM.getASTAllocaAddressSpace()))) {
+ NeedCopy = true;
+ }
+ }
+ // For OpenCL even if RV is located in default or alloca address space
+ // we don't want to perform address space cast for it, since that
+ // leads to casting __private * (default addr space in OpenCL) to
+ // __global * which is not valid. Create memcpy call instead.
+ else if ((ArgInfo.getIndirectByVal() &&
+ Addr.getType()->getAddressSpace() != IRFuncTy->
+ getParamType(FirstIRArg)->getPointerAddressSpace())) {
NeedCopy = true;
}
}
+
if (NeedCopy) {
// Create an aligned temporary, and copy to it.
Address AI = CreateMemTempWithoutCast(
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
