This is an archive of the discontinued LLVM Phabricator instance.

Differential D87075

[mlir] Check for type conversion success in std->llvm function conversion
ClosedPublic

Authored by ftynse on Sep 3 2020, 1:05 AM.

Details

Reviewers
mehdi_amini
Commits
rG5cac85c931d9: [mlir] Check for type conversion success in std->llvm function conversion
Summary

Type converter may fail and return nullptr on unconvertible types. The function
conversion did not include a check and was attempting to use a nullptr type to
construct an LLVM function, leading to a crash. Add a check and return early.
The rest of the call stack propagates errors properly.

Fixes PR47403.

Diff Detail

Event Timeline

ftynse created this revision.Sep 3 2020, 1:05 AM
Herald added a project: Restricted Project. · View Herald TranscriptSep 3 2020, 1:05 AM
Herald added subscribers: msifontes, jurahul, Kayjukh and 11 others. · View Herald Transcript
ftynse requested review of this revision.Sep 3 2020, 1:05 AM
Herald added subscribers: stephenneuendorffer, nicolasvasilache. · View Herald TranscriptSep 3 2020, 1:05 AM
Harbormaster completed remote builds in B70517: Diff 289652.Sep 3 2020, 1:19 AM
mehdi_amini accepted this revision.Sep 13 2020, 9:40 PM
This revision is now accepted and ready to land.Sep 13 2020, 9:40 PM
Herald added a subscriber: tatianashp. · View Herald TranscriptSep 13 2020, 9:40 PM
Closed by commit rG5cac85c931d9: [mlir] Check for type conversion success in std->llvm function conversion (authored by ftynse). · Explain WhySep 14 2020, 4:16 AM
This revision was automatically updated to reflect the committed changes.
ftynse added a commit: rG5cac85c931d9: [mlir] Check for type conversion success in std->llvm function conversion.

Revision Contents

PathSize
mlir/
lib/
Conversion/
StandardToLLVM/
2 lines
test/
Conversion/
StandardToLLVM/
5 lines

Diff 291536

mlir/lib/Conversion/StandardToLLVM/StandardToLLVM.cpp

Show First 20 Lines • Show All 1,106 Lines • ▼ Show 20 Linesprotected:
convertFuncOpToLLVMFuncOp(FuncOp funcOp, convertFuncOpToLLVMFuncOp(FuncOp funcOp,
ConversionPatternRewriter &rewriter) const { ConversionPatternRewriter &rewriter) const {
// Convert the original function arguments. They are converted using the // Convert the original function arguments. They are converted using the
// LLVMTypeConverter provided to this legalization pattern. // LLVMTypeConverter provided to this legalization pattern.
auto varargsAttr = funcOp.getAttrOfType<BoolAttr>("std.varargs"); auto varargsAttr = funcOp.getAttrOfType<BoolAttr>("std.varargs");
TypeConverter::SignatureConversion result(funcOp.getNumArguments()); TypeConverter::SignatureConversion result(funcOp.getNumArguments());
auto llvmType = typeConverter.convertFunctionSignature( auto llvmType = typeConverter.convertFunctionSignature(
funcOp.getType(), varargsAttr && varargsAttr.getValue(), result); funcOp.getType(), varargsAttr && varargsAttr.getValue(), result);
if (!llvmType)
return nullptr;
// Propagate argument attributes to all converted arguments obtained after // Propagate argument attributes to all converted arguments obtained after
// converting a given original argument. // converting a given original argument.
SmallVector<NamedAttribute, 4> attributes; SmallVector<NamedAttribute, 4> attributes;
filterFuncAttributes(funcOp.getAttrs(), /*filterArgAttrs=*/true, filterFuncAttributes(funcOp.getAttrs(), /*filterArgAttrs=*/true,
attributes); attributes);
for (unsigned i = 0, e = funcOp.getNumArguments(); i < e; ++i) { for (unsigned i = 0, e = funcOp.getNumArguments(); i < e; ++i) {
auto attr = impl::getArgAttrDict(funcOp, i); auto attr = impl::getArgAttrDict(funcOp, i);
▲ Show 20 Lines • Show All 2,384 Lines • Show Last 20 Lines

mlir/test/Conversion/StandardToLLVM/invalid.mlir

Show All 23 Lines
// ----- // -----
func @mlir_cast_to_llvm_vec(%0 : vector<1x1xf32>) -> !llvm.vec<1 x float> { func @mlir_cast_to_llvm_vec(%0 : vector<1x1xf32>) -> !llvm.vec<1 x float> {
// expected-error@+1 {{'llvm.mlir.cast' op only 1-d vector is allowed}} // expected-error@+1 {{'llvm.mlir.cast' op only 1-d vector is allowed}}
%1 = llvm.mlir.cast %0 : vector<1x1xf32> to !llvm.vec<1 x float> %1 = llvm.mlir.cast %0 : vector<1x1xf32> to !llvm.vec<1 x float>
return %1 : !llvm.vec<1 x float> return %1 : !llvm.vec<1 x float>
} }
// -----
// Should not crash on unsupported types in function signatures.
func @unsupported_signature() -> tensor<10 x i32>