[mlir][llvm] Support nusw and nuw in GEP #137272
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Thank you for submitting a Pull Request (PR) to the LLVM Project! This PR will be automatically labeled and the relevant teams will be notified. If you wish to, you can add reviewers by using the "Reviewers" section on this page. If this is not working for you, it is probably because you do not have write permissions for the repository. In which case you can instead tag reviewers by name in a comment by using If you have received no comments on your PR for a week, you can request a review by "ping"ing the PR by adding a comment “Ping”. The common courtesy "ping" rate is once a week. Please remember that you are asking for valuable time from other developers. If you have further questions, they may be answered by the LLVM GitHub User Guide. You can also ask questions in a comment on this PR, on the LLVM Discord or on the forums. |
|
@llvm/pr-subscribers-mlir @llvm/pr-subscribers-mlir-llvm Author: Peiyong Lin (lpy) Changesnusw and nuw were introduced in getelementptr, this patch plumbs them in MLIR. Since inbounds implies nusw, this patch also adds an inboundsFlag to represent the concept of raw inbounds with no nusw implication, and have the inbounds literal captured as the combination of inboundsFlag and nusw. Fixes: iree#20482 Full diff: https://github.com/llvm/llvm-project/pull/137272.diff 6 Files Affected:
diff --git a/mlir/include/mlir/Dialect/LLVMIR/LLVMEnums.td b/mlir/include/mlir/Dialect/LLVMIR/LLVMEnums.td
index 45ccf30644920..6c0fe363d5551 100644
--- a/mlir/include/mlir/Dialect/LLVMIR/LLVMEnums.td
+++ b/mlir/include/mlir/Dialect/LLVMIR/LLVMEnums.td
@@ -876,4 +876,32 @@ def UWTableKindEnum : LLVM_EnumAttr<
let cppNamespace = "::mlir::LLVM::uwtable";
}
+//===----------------------------------------------------------------------===//
+// GEPNoWrapFlags
+//===----------------------------------------------------------------------===//
+
+// These values must match llvm::GEPNoWrapFlags ones.
+// See llvm/include/llvm/IR/GEPNoWrapFlags.h.
+// Since inbounds implies nusw, create an inboundsFlag that represents the
+// concept of raw inbounds with no nusw implication and the actual inbounds
+// literal will be captured as the combination of inboundsFlag and nusw.
+
+def GEPNone : I32BitEnumCaseNone<"none">;
+def GEPInboundsFlag : I32BitEnumCaseBit<"inboundsFlag", 0, "inbounds_flag">;
+def GEPNusw : I32BitEnumCaseBit<"nusw", 1>;
+def GEPNuw : I32BitEnumCaseBit<"nuw", 2>;
+def GEPInbounds : BitEnumCaseGroup<"inbounds", [GEPInboundsFlag, GEPNusw]>;
+
+def GEPNoWrapFlags : I32BitEnum<
+ "GEPNoWrapFlags",
+ "::mlir::LLVM::GEPNoWrapFlags",
+ [GEPNone, GEPInboundsFlag, GEPNusw, GEPNuw, GEPInbounds]> {
+ let cppNamespace = "::mlir::LLVM";
+ let printBitEnumPrimaryGroups = 1;
+}
+
+def GEPNoWrapFlagsProp : EnumProp<GEPNoWrapFlags> {
+ let defaultValue = interfaceType # "::none";
+}
+
#endif // LLVMIR_ENUMS
diff --git a/mlir/include/mlir/Dialect/LLVMIR/LLVMOps.td b/mlir/include/mlir/Dialect/LLVMIR/LLVMOps.td
index 5745d370f7268..4584eee5a27ab 100644
--- a/mlir/include/mlir/Dialect/LLVMIR/LLVMOps.td
+++ b/mlir/include/mlir/Dialect/LLVMIR/LLVMOps.td
@@ -291,7 +291,7 @@ def LLVM_GEPOp : LLVM_Op<"getelementptr", [Pure,
Variadic<LLVM_ScalarOrVectorOf<AnySignlessInteger>>:$dynamicIndices,
DenseI32ArrayAttr:$rawConstantIndices,
TypeAttr:$elem_type,
- UnitAttr:$inbounds);
+ GEPNoWrapFlagsProp:$gepNoWrapFlags);
let results = (outs LLVM_ScalarOrVectorOf<LLVM_AnyPointer>:$res);
let skipDefaultBuilders = 1;
@@ -303,8 +303,10 @@ def LLVM_GEPOp : LLVM_Op<"getelementptr", [Pure,
as indices. In the case of indexing within a structure, it is required to
either use constant indices directly, or supply a constant SSA value.
- An optional 'inbounds' attribute specifies the low-level pointer arithmetic
+ Optional attributes can be used to specify the low-level pointer arithmetic
overflow behavior that LLVM uses after lowering the operation to LLVM IR.
+ The acceptable attributes could be one of or the combination of 'inbounds',
+ 'nusw' or 'nuw'.
Examples:
@@ -323,10 +325,12 @@ def LLVM_GEPOp : LLVM_Op<"getelementptr", [Pure,
let builders = [
OpBuilder<(ins "Type":$resultType, "Type":$elementType, "Value":$basePtr,
- "ValueRange":$indices, CArg<"bool", "false">:$inbounds,
+ "ValueRange":$indices,
+ CArg<"GEPNoWrapFlags", "GEPNoWrapFlags::none">:$gepNoWrapFlags,
CArg<"ArrayRef<NamedAttribute>", "{}">:$attributes)>,
OpBuilder<(ins "Type":$resultType, "Type":$elementType, "Value":$basePtr,
- "ArrayRef<GEPArg>":$indices, CArg<"bool", "false">:$inbounds,
+ "ArrayRef<GEPArg>":$indices,
+ CArg<"GEPNoWrapFlags", "GEPNoWrapFlags::none">:$gepNoWrapFlags,
CArg<"ArrayRef<NamedAttribute>", "{}">:$attributes)>,
];
let llvmBuilder = [{
@@ -343,10 +347,13 @@ def LLVM_GEPOp : LLVM_Op<"getelementptr", [Pure,
}
Type baseElementType = op.getElemType();
llvm::Type *elementType = moduleTranslation.convertType(baseElementType);
- $res = builder.CreateGEP(elementType, $base, indices, "", $inbounds);
+ $res = builder.CreateGEP(elementType, $base, indices, "",
+ llvm::GEPNoWrapFlags::fromRaw(
+ static_cast<unsigned>(
+ op.getGepNoWrapFlags())));
}];
let assemblyFormat = [{
- (`inbounds` $inbounds^)?
+ ($gepNoWrapFlags^)?
$base `[` custom<GEPIndices>($dynamicIndices, $rawConstantIndices) `]` attr-dict
`:` functional-type(operands, results) `,` $elem_type
}];
diff --git a/mlir/lib/Dialect/LLVMIR/IR/LLVMDialect.cpp b/mlir/lib/Dialect/LLVMIR/IR/LLVMDialect.cpp
index 26c3ef1e8b8bf..9f9e31185e07c 100644
--- a/mlir/lib/Dialect/LLVMIR/IR/LLVMDialect.cpp
+++ b/mlir/lib/Dialect/LLVMIR/IR/LLVMDialect.cpp
@@ -673,29 +673,29 @@ static void destructureIndices(Type currType, ArrayRef<GEPArg> indices,
void GEPOp::build(OpBuilder &builder, OperationState &result, Type resultType,
Type elementType, Value basePtr, ArrayRef<GEPArg> indices,
- bool inbounds, ArrayRef<NamedAttribute> attributes) {
+ GEPNoWrapFlags noWrapFlags,
+ ArrayRef<NamedAttribute> attributes) {
SmallVector<int32_t> rawConstantIndices;
SmallVector<Value> dynamicIndices;
destructureIndices(elementType, indices, rawConstantIndices, dynamicIndices);
result.addTypes(resultType);
result.addAttributes(attributes);
- result.addAttribute(getRawConstantIndicesAttrName(result.name),
- builder.getDenseI32ArrayAttr(rawConstantIndices));
- if (inbounds) {
- result.addAttribute(getInboundsAttrName(result.name),
- builder.getUnitAttr());
- }
- result.addAttribute(kElemTypeAttrName, TypeAttr::get(elementType));
+ result.getOrAddProperties<Properties>().rawConstantIndices =
+ builder.getDenseI32ArrayAttr(rawConstantIndices);
+ result.getOrAddProperties<Properties>().gepNoWrapFlags = noWrapFlags;
+ result.getOrAddProperties<Properties>().elem_type =
+ TypeAttr::get(elementType);
result.addOperands(basePtr);
result.addOperands(dynamicIndices);
}
void GEPOp::build(OpBuilder &builder, OperationState &result, Type resultType,
Type elementType, Value basePtr, ValueRange indices,
- bool inbounds, ArrayRef<NamedAttribute> attributes) {
+ GEPNoWrapFlags noWrapFlags,
+ ArrayRef<NamedAttribute> attributes) {
build(builder, result, resultType, elementType, basePtr,
- SmallVector<GEPArg>(indices), inbounds, attributes);
+ SmallVector<GEPArg>(indices), noWrapFlags, attributes);
}
static ParseResult
diff --git a/mlir/lib/Dialect/LLVMIR/IR/LLVMMemorySlot.cpp b/mlir/lib/Dialect/LLVMIR/IR/LLVMMemorySlot.cpp
index 8640ef28a9e56..e463daf69ad57 100644
--- a/mlir/lib/Dialect/LLVMIR/IR/LLVMMemorySlot.cpp
+++ b/mlir/lib/Dialect/LLVMIR/IR/LLVMMemorySlot.cpp
@@ -891,7 +891,7 @@ DeletionKind LLVM::GEPOp::rewire(const DestructurableMemorySlot &slot,
auto byteType = IntegerType::get(builder.getContext(), 8);
auto newPtr = builder.createOrFold<LLVM::GEPOp>(
getLoc(), getResult().getType(), byteType, newSlot.ptr,
- ArrayRef<GEPArg>(accessInfo->subslotOffset), getInbounds());
+ ArrayRef<GEPArg>(accessInfo->subslotOffset), getGepNoWrapFlags());
getResult().replaceAllUsesWith(newPtr);
return DeletionKind::Delete;
}
diff --git a/mlir/lib/Target/LLVMIR/ModuleImport.cpp b/mlir/lib/Target/LLVMIR/ModuleImport.cpp
index 3f80002c15ebb..8ba768d86db0f 100644
--- a/mlir/lib/Target/LLVMIR/ModuleImport.cpp
+++ b/mlir/lib/Target/LLVMIR/ModuleImport.cpp
@@ -1984,8 +1984,9 @@ LogicalResult ModuleImport::convertInstruction(llvm::Instruction *inst) {
}
Type type = convertType(inst->getType());
- auto gepOp = builder.create<GEPOp>(loc, type, sourceElementType, *basePtr,
- indices, gepInst->isInBounds());
+ auto gepOp = builder.create<GEPOp>(
+ loc, type, sourceElementType, *basePtr, indices,
+ gepInst->isInBounds() ? GEPNoWrapFlags::inbounds : GEPNoWrapFlags());
mapValue(inst, gepOp);
return success();
}
diff --git a/mlir/test/Target/LLVMIR/llvmir.mlir b/mlir/test/Target/LLVMIR/llvmir.mlir
index 74fa327809864..4a2447263ae68 100644
--- a/mlir/test/Target/LLVMIR/llvmir.mlir
+++ b/mlir/test/Target/LLVMIR/llvmir.mlir
@@ -1057,6 +1057,14 @@ llvm.func @gep(%ptr: !llvm.ptr, %idx: i64,
llvm.getelementptr %ptr[%idx, 1, 0] : (!llvm.ptr, i64) -> !llvm.ptr, !llvm.struct<(i32, struct<(i32, f32)>)>
// CHECK: = getelementptr inbounds { [10 x float] }, ptr %{{.*}}, i64 %{{.*}}, i32 0, i64 %{{.*}}
llvm.getelementptr inbounds %ptr2[%idx, 0, %idx] : (!llvm.ptr, i64, i64) -> !llvm.ptr, !llvm.struct<(array<10 x f32>)>
+ // CHECK: = getelementptr inbounds nuw { [10 x float] }, ptr %{{.*}}, i64 %{{.*}}, i32 0, i64 %{{.*}}
+ llvm.getelementptr inbounds | nuw %ptr2[%idx, 0, %idx] : (!llvm.ptr, i64, i64) -> !llvm.ptr, !llvm.struct<(array<10 x f32>)>
+ // CHECK: = getelementptr nusw { [10 x float] }, ptr %{{.*}}, i64 %{{.*}}, i32 0, i64 %{{.*}}
+ llvm.getelementptr nusw %ptr2[%idx, 0, %idx] : (!llvm.ptr, i64, i64) -> !llvm.ptr, !llvm.struct<(array<10 x f32>)>
+ // CHECK: = getelementptr nusw nuw { [10 x float] }, ptr %{{.*}}, i64 %{{.*}}, i32 0, i64 %{{.*}}
+ llvm.getelementptr nusw | nuw %ptr2[%idx, 0, %idx] : (!llvm.ptr, i64, i64) -> !llvm.ptr, !llvm.struct<(array<10 x f32>)>
+ // CHECK: = getelementptr nuw { [10 x float] }, ptr %{{.*}}, i64 %{{.*}}, i32 0, i64 %{{.*}}
+ llvm.getelementptr nuw %ptr2[%idx, 0, %idx] : (!llvm.ptr, i64, i64) -> !llvm.ptr, !llvm.struct<(array<10 x f32>)>
llvm.return
}
|
gysit
reviewed
Apr 25, 2025
|
Thanks for the comments. I updated the patch, please take another look :) |
Dinistro
approved these changes
Apr 28, 2025
gysit
reviewed
Apr 28, 2025
|
Thanks for the feedback. I have added a few lines to the roundtrip test and updated the |
joker-eph
reviewed
Apr 28, 2025
|
Thanks for the comments, I have updated the patch, please take a look :) |
joker-eph
reviewed
Apr 28, 2025
|
Thanks, added a case to cover that. Please take a look :) |
gysit
approved these changes
Apr 29, 2025
nusw and nuw were introduced in getelementptr, this patch plumbs them in MLIR. Since inbounds implies nusw, this patch also adds an inboundsFlag to represent the concept of raw inbounds with no nusw implication, and have the inbounds literal captured as the combination of inboundsFlag and nusw. Fixes: iree-org/iree#20482 Signed-off-by: Lin, Peiyong <[email protected]>
krzysz00
approved these changes
Apr 29, 2025
|
@lpy Congratulations on having your first Pull Request (PR) merged into the LLVM Project! Your changes will be combined with recent changes from other authors, then tested by our build bots. If there is a problem with a build, you may receive a report in an email or a comment on this PR. Please check whether problems have been caused by your change specifically, as the builds can include changes from many authors. It is not uncommon for your change to be included in a build that fails due to someone else's changes, or infrastructure issues. How to do this, and the rest of the post-merge process, is covered in detail here. If your change does cause a problem, it may be reverted, or you can revert it yourself. This is a normal part of LLVM development. You can fix your changes and open a new PR to merge them again. If you don't get any reports, no action is required from you. Your changes are working as expected, well done! |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/169/builds/10987 Here is the relevant piece of the build log for the reference |
IanWood1
added a commit
to IanWood1/iree
that referenced
this pull request
May 5, 2025
llvm/llvm-project#137272 plumbs through nusw and nuw by removing `inbounds` and replaces it with a enum. Signed-off-by: Ian Wood <[email protected]>
IanWood1
added a commit
to IanWood1/iree
that referenced
this pull request
May 5, 2025
llvm/llvm-project#137272 plumbs through nusw and nuw by removing `inbounds` and replaces it with a enum. Signed-off-by: Ian Wood <[email protected]>
IanWood1
pushed a commit
to IanWood1/llvm-project
that referenced
this pull request
May 6, 2025
nusw and nuw were introduced in getelementptr, this patch plumbs them in MLIR. Since inbounds implies nusw, this patch also adds an inboundsFlag to represent the concept of raw inbounds with no nusw implication, and have the inbounds literal captured as the combination of inboundsFlag and nusw. Fixes: iree#20482 Signed-off-by: Lin, Peiyong <[email protected]>
IanWood1
pushed a commit
to IanWood1/llvm-project
that referenced
this pull request
May 6, 2025
nusw and nuw were introduced in getelementptr, this patch plumbs them in MLIR. Since inbounds implies nusw, this patch also adds an inboundsFlag to represent the concept of raw inbounds with no nusw implication, and have the inbounds literal captured as the combination of inboundsFlag and nusw. Fixes: iree#20482 Signed-off-by: Lin, Peiyong <[email protected]>
IanWood1
pushed a commit
to IanWood1/llvm-project
that referenced
this pull request
May 6, 2025
nusw and nuw were introduced in getelementptr, this patch plumbs them in MLIR. Since inbounds implies nusw, this patch also adds an inboundsFlag to represent the concept of raw inbounds with no nusw implication, and have the inbounds literal captured as the combination of inboundsFlag and nusw. Fixes: iree#20482 Signed-off-by: Lin, Peiyong <[email protected]>
IanWood1
added a commit
to iree-org/iree
that referenced
this pull request
May 6, 2025
- Carries the same 4 reverts from #20674. - Uses enum when building `LLVM::GEPOp` (llvm/llvm-project#137272) - Reverts llvm/llvm-project@7318074 which can be undone after torch-mlir and stablehlo have been updated. - Reverts llvm/llvm-project#137930 and llvm/llvm-project@e1cff21 to fix correctness failure of e2e_matmul_cdna3_pad_i8_rocm_hip. --------- Signed-off-by: Ian Wood <[email protected]>
GeorgeARM
pushed a commit
to GeorgeARM/llvm-project
that referenced
this pull request
May 7, 2025
nusw and nuw were introduced in getelementptr, this patch plumbs them in MLIR. Since inbounds implies nusw, this patch also adds an inboundsFlag to represent the concept of raw inbounds with no nusw implication, and have the inbounds literal captured as the combination of inboundsFlag and nusw. Fixes: iree#20482 Signed-off-by: Lin, Peiyong <[email protected]>
KyleHerndon
pushed a commit
to KyleHerndon/iree
that referenced
this pull request
May 7, 2025
- Carries the same 4 reverts from iree-org#20674. - Uses enum when building `LLVM::GEPOp` (llvm/llvm-project#137272) - Reverts llvm/llvm-project@7318074 which can be undone after torch-mlir and stablehlo have been updated. - Reverts llvm/llvm-project#137930 and llvm/llvm-project@e1cff21 to fix correctness failure of e2e_matmul_cdna3_pad_i8_rocm_hip. --------- Signed-off-by: Ian Wood <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
nusw and nuw were introduced in getelementptr, this patch plumbs them in MLIR.
Since inbounds implies nusw, this patch also adds an inboundsFlag to represent the concept of raw inbounds with no nusw implication, and have the inbounds literal captured as the combination of inboundsFlag and nusw.
Fixes: iree#20482