Skip to content
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.

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

[DeviceSanitizer] Ignore target extension type #15484

Merged
merged 1 commit into from
Oct 14, 2024
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
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
41 changes: 37 additions & 4 deletions llvm/lib/Transforms/Instrumentation/AddressSanitizer.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -1506,19 +1506,49 @@ static bool isUnsupportedAMDGPUAddrspace(Value *Addr) {
return false;
}

static bool isUnsupportedSPIRAccess(Value *Addr, Function *Func) {
static bool containsTargetExtType(const Type *Ty) {
if (isa<TargetExtType>(Ty))
return true;

if (Ty->isVectorTy())
return containsTargetExtType(Ty->getScalarType());

if (Ty->isArrayTy())
return containsTargetExtType(Ty->getArrayElementType());

if (auto *STy = dyn_cast<StructType>(Ty)) {
for (unsigned int i = 0; i < STy->getNumElements(); i++)
if (containsTargetExtType(STy->getElementType(i)))
return true;
return false;
}

return false;
}

static bool isUnsupportedSPIRAccess(Value *Addr, Instruction *Inst) {
// Skip SPIR-V built-in varibles
auto *OrigValue = Addr->stripInBoundsOffsets();
if (OrigValue->getName().starts_with("__spirv_BuiltIn"))
return true;

// Ignore load/store for target ext type since we can't know exactly what size
// it is.
if (isa<StoreInst>(Inst) &&
containsTargetExtType(
cast<StoreInst>(Inst)->getValueOperand()->getType()))
return true;

if (isa<LoadInst>(Inst) && containsTargetExtType(Inst->getType()))
return true;

Type *PtrTy = cast<PointerType>(Addr->getType()->getScalarType());
switch (PtrTy->getPointerAddressSpace()) {
case kSpirOffloadPrivateAS: {
if (!ClSpirOffloadPrivates)
return true;
// Skip kernel arguments
return Func->getCallingConv() == CallingConv::SPIR_KERNEL &&
return Inst->getFunction()->getCallingConv() == CallingConv::SPIR_KERNEL &&
isa<Argument>(Addr);
}
case kSpirOffloadGlobalAS: {
Expand Down Expand Up @@ -1756,7 +1786,10 @@ bool AddressSanitizer::isInterestingAlloca(const AllocaInst &AI) {
// swifterror allocas are register promoted by ISel
!AI.isSwiftError() &&
// safe allocas are not interesting
!(SSGI && SSGI->isSafe(AI)));
!(SSGI && SSGI->isSafe(AI)) &&
// ignore alloc contains target ext type since we can't know exactly what
// size it is.
!containsTargetExtType(AI.getAllocatedType()));

ProcessedAllocas[&AI] = IsInteresting;
return IsInteresting;
Expand All @@ -1765,7 +1798,7 @@ bool AddressSanitizer::isInterestingAlloca(const AllocaInst &AI) {
bool AddressSanitizer::ignoreAccess(Instruction *Inst, Value *Ptr) {
// SPIR has its own rules to filter the instrument accesses
if (TargetTriple.isSPIROrSPIRV()) {
if (isUnsupportedSPIRAccess(Ptr, Inst->getFunction()))
if (isUnsupportedSPIRAccess(Ptr, Inst))
return true;
} else {
// Instrument accesses from different address spaces only for AMDGPU.
Expand Down
Original file line number Diff line number Diff line change
@@ -0,0 +1,20 @@
; RUN: opt < %s -passes=asan -asan-instrumentation-with-call-threshold=0 -asan-stack=0 -asan-globals=0 -asan-use-after-return=never -asan-stack-dynamic-alloca=0 -asan-mapping-scale=4 -S | FileCheck %s

target datalayout = "e-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-n8:16:32:64-G1"
target triple = "spir64-unknown-unknown"

%"struct.sycl::_V1::ext::oneapi::experimental::matrix::joint_matrix" = type { target("spirv.JointMatrixINTEL", i16, 16, 32, 0, 3, 0, 1) }

define spir_kernel void @_ZTS4multIN4sycl3_V13ext6oneapi8bfloat16ELm16ELm16ELm32EE() {
entry:
; CHECK-NOT: MyAlloc
%sub_a.i = alloca [2 x %"struct.sycl::_V1::ext::oneapi::experimental::matrix::joint_matrix"], i32 0, align 8
br label %for.cond10.i

for.cond10.i: ; preds = %for.cond10.i, %entry
%0 = load target("spirv.JointMatrixINTEL", i16, 16, 32, 0, 3, 0, 1), ptr null, align 8
store target("spirv.JointMatrixINTEL", float, 16, 16, 3, 3, 2) zeroinitializer, ptr null, align 8
; CHECK-NOT: asan_load
; CHECK-NOT: asan_store
br label %for.cond10.i
}
Loading