Skip to content

[mlir][linalg] NFC: Use tablegen macro for pass constructors #82892

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

Jump to bottom
Merged
merged 1 commit into from
Feb 24, 2024
Merged

[mlir][linalg] NFC: Use tablegen macro for pass constructors #82892

merged 1 commit into from
Feb 24, 2024

Conversation

qedawkins
Copy link
Contributor

@qedawkins qedawkins commented Feb 24, 2024 •
edited
Loading

This uses the tablegen macros for generating pass constructors, exposing pass options for fold-unit-extent-dims and linalg-detensorize.

Additionally aligns some of the pass namings to their text counterpart. This includes an API change:

createLinalgGeneralizationPass -> createLinalgGeneralizeNamedOpsPass

@qedawkins
[mlir][linalg] NFC: Use tablegen macro for pass constructors
cde1996 
This uses the tablegen macros for generating pass constructors,
including exposing pass options for fold-unit-extent-dims and
linalg-detensorize.

Additionally aligns some of the pass namings to their text
counterpart. This includes an API change:

createLinalgGeneralizationPass -> createLinalgGeneralizeNamedOpsPass
@llvmbot llvmbot added mlir:linalg mlir:sparse Sparse compiler in MLIR mlir labels Feb 24, 2024
@llvmbot
Copy link
Member

llvmbot commented Feb 24, 2024 •
edited
Loading

@llvm/pr-subscribers-mlir-sparse
@llvm/pr-subscribers-mlir

@llvm/pr-subscribers-mlir-linalg

Author: Quinn Dawkins (qedawkins)

Changes

This uses the tablegen macros for generating pass constructors, including exposing pass options for fold-unit-extent-dims and linalg-detensorize.

Additionally aligns some of the pass namings to their text counterpart. This includes an API change:

createLinalgGeneralizationPass -> createLinalgGeneralizeNamedOpsPass

Patch is 22.00 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/82892.diff

12 Files Affected:

  • (modified) mlir/include/mlir/Dialect/Linalg/Passes.h (+1-37)
  • (modified) mlir/include/mlir/Dialect/Linalg/Passes.td (+38-49)
  • (modified) mlir/lib/Dialect/Linalg/Transforms/Bufferize.cpp (+4-6)
  • (modified) mlir/lib/Dialect/Linalg/Transforms/Detensorize.cpp (+4-6)
  • (modified) mlir/lib/Dialect/Linalg/Transforms/DropUnitDims.cpp (+5-6)
  • (modified) mlir/lib/Dialect/Linalg/Transforms/ElementwiseOpFusion.cpp (+4-7)
  • (modified) mlir/lib/Dialect/Linalg/Transforms/ElementwiseToLinalg.cpp (+4-6)
  • (modified) mlir/lib/Dialect/Linalg/Transforms/Generalization.cpp (+7-8)
  • (modified) mlir/lib/Dialect/Linalg/Transforms/InlineScalarOperands.cpp (+4-6)
  • (modified) mlir/lib/Dialect/Linalg/Transforms/Loops.cpp (+12-18)
  • (modified) mlir/lib/Dialect/Linalg/Transforms/NamedOpConversions.cpp (+5-8)
  • (modified) mlir/lib/Dialect/SparseTensor/Pipelines/SparseTensorPipelines.cpp (+1-1)
diff --git a/mlir/include/mlir/Dialect/Linalg/Passes.h b/mlir/include/mlir/Dialect/Linalg/Passes.h
index 5f46affe592a2d..d36d1e70f0b14d 100644
--- a/mlir/include/mlir/Dialect/Linalg/Passes.h
+++ b/mlir/include/mlir/Dialect/Linalg/Passes.h
@@ -27,43 +27,7 @@ struct OneShotBufferizationOptions;
 } // namespace bufferization
 
 #define GEN_PASS_DECL
-#include "mlir/Dialect/Linalg/Passes.h.inc"
-
-std::unique_ptr<Pass> createConvertElementwiseToLinalgPass();
-
-std::unique_ptr<Pass> createLinalgFoldUnitExtentDimsPass();
-
-std::unique_ptr<Pass> createLinalgElementwiseOpFusionPass();
-std::unique_ptr<Pass> createFoldReshapeOpsByLinearizationPass();
-
-std::unique_ptr<Pass> createLinalgNamedOpConversionPass();
-
-std::unique_ptr<Pass> createLinalgInlineScalarOperandsPass();
-
-/// Create a pass to convert Linalg operations to scf.for loops and
-/// memref.load/memref.store accesses.
-std::unique_ptr<Pass> createConvertLinalgToLoopsPass();
-
-/// Create a pass to convert Linalg operations to scf.parallel loops and
-/// memref.load/memref.store accesses.
-std::unique_ptr<Pass> createConvertLinalgToParallelLoopsPass();
-
-/// Create a pass to convert Linalg operations to affine.for loops and
-/// affine_load/affine_store accesses.
-/// Placeholder for now, this is NYI.
-std::unique_ptr<Pass> createConvertLinalgToAffineLoopsPass();
-
-/// Create a pass to convert Linalg operations which work on tensors to use
-/// buffers instead.
-std::unique_ptr<Pass> createLinalgBufferizePass();
-
-/// Create a pass to convert named Linalg operations to Linalg generic
-/// operations.
-std::unique_ptr<Pass> createLinalgGeneralizationPass();
-
-/// Create a pass to convert Linalg operations to equivalent operations that
-/// work on primitive types, if possible.
-std::unique_ptr<Pass> createLinalgDetensorizePass();
+#include "mlir/Dialect/Linalg/Passes.h.inc" // IWYU pragma: keep
 
 //===----------------------------------------------------------------------===//
 // Registration
diff --git a/mlir/include/mlir/Dialect/Linalg/Passes.td b/mlir/include/mlir/Dialect/Linalg/Passes.td
index cca50e21d5ce0b..85f11c66d29a73 100644
--- a/mlir/include/mlir/Dialect/Linalg/Passes.td
+++ b/mlir/include/mlir/Dialect/Linalg/Passes.td
@@ -11,7 +11,7 @@
 
 include "mlir/Pass/PassBase.td"
 
-def ConvertElementwiseToLinalg : Pass<"convert-elementwise-to-linalg", ""> {
+def ConvertElementwiseToLinalgPass : Pass<"convert-elementwise-to-linalg", ""> {
   let summary = "Convert ElementwiseMappable ops to linalg";
   let description = [{
     Convert ops with the `ElementwiseMappable` trait to linalg parallel loops.
@@ -20,54 +20,17 @@ def ConvertElementwiseToLinalg : Pass<"convert-elementwise-to-linalg", ""> {
     run on op which contains linalg ops (most commonly a
     FunctionOpInterface op).
   }];
-  let constructor = "mlir::createConvertElementwiseToLinalgPass()";
   let dependentDialects = ["linalg::LinalgDialect", "memref::MemRefDialect"];
 }
 
-def LinalgFoldUnitExtentDims : Pass<"linalg-fold-unit-extent-dims", ""> {
-  let summary = "Remove unit-extent dimension in Linalg ops on tensors";
-  let constructor = "mlir::createLinalgFoldUnitExtentDimsPass()";
-  let options = [
-    Option<"useRankReducingSlices", "use-rank-reducing-slices", "bool",
-           /*default=*/"false",
-           "Generate rank-reducing slices instead of reassociative reshapes">
-  ];
-  let dependentDialects = [
-    "linalg::LinalgDialect", "affine::AffineDialect", "memref::MemRefDialect"
-  ];
-}
-
-def LinalgElementwiseOpFusion : Pass<"linalg-fuse-elementwise-ops"> {
-  let summary = "Fuse elementwise operations on tensors";
-  let constructor = "mlir::createLinalgElementwiseOpFusionPass()";
-  let dependentDialects = [
-    "affine::AffineDialect", "linalg::LinalgDialect", "memref::MemRefDialect"
-  ];
-}
-
-def LinalgNamedOpConversion: Pass<"linalg-named-op-conversion"> {
-  let summary = "Convert from one named linalg op to another.";
-  let constructor = "mlir::createLinalgNamedOpConversionPass()";
-  let dependentDialects = ["linalg::LinalgDialect", "tensor::TensorDialect"];
-}
-
-def LinalgInlineScalarOperands : Pass<"linalg-inline-scalar-operands"> {
-  let summary = "Inline scalar operands into linalg generic ops";
-  let constructor = "mlir::createLinalgInlineScalarOperandsPass()";
-  let dependentDialects = [
-    "linalg::LinalgDialect"
-  ];
-}
-
-def LinalgLowerToAffineLoops : Pass<"convert-linalg-to-affine-loops"> {
+def ConvertLinalgToAffineLoopsPass : Pass<"convert-linalg-to-affine-loops"> {
   let summary = "Lower the operations from the linalg dialect into affine "
                 "loops";
-  let constructor = "mlir::createConvertLinalgToAffineLoopsPass()";
   let dependentDialects = [
     "affine::AffineDialect", "linalg::LinalgDialect", "memref::MemRefDialect"];
 }
 
-def LinalgLowerToLoops : Pass<"convert-linalg-to-loops"> {
+def ConvertLinalgToLoopsPass : Pass<"convert-linalg-to-loops"> {
   let summary = "Lower the operations from the linalg dialect into loops";
   let description = [{
     Lowers the `linalg` ops to loop nests using `scf.for`.
@@ -76,7 +39,6 @@ def LinalgLowerToLoops : Pass<"convert-linalg-to-loops"> {
     i.e., tensor operands and results must be converted to memrefs via
     bufferization.
   }];
-  let constructor = "mlir::createConvertLinalgToLoopsPass()";
   let dependentDialects = [
     "linalg::LinalgDialect",
     "scf::SCFDialect",
@@ -84,11 +46,10 @@ def LinalgLowerToLoops : Pass<"convert-linalg-to-loops"> {
   ];
 }
 
-def LinalgLowerToParallelLoops
+def ConvertLinalgToParallelLoopsPass
     : Pass<"convert-linalg-to-parallel-loops"> {
   let summary = "Lower the operations from the linalg dialect into parallel "
                 "loops";
-  let constructor = "mlir::createConvertLinalgToParallelLoopsPass()";
   let dependentDialects = [
     "affine::AffineDialect",
     "linalg::LinalgDialect",
@@ -97,9 +58,39 @@ def LinalgLowerToParallelLoops
   ];
 }
 
-def LinalgBufferize : Pass<"linalg-bufferize"> {
+def LinalgFoldUnitExtentDimsPass : Pass<"linalg-fold-unit-extent-dims", ""> {
+  let summary = "Remove unit-extent dimension in Linalg ops on tensors";
+  let options = [
+    Option<"useRankReducingSlices", "use-rank-reducing-slices", "bool",
+           /*default=*/"false",
+           "Generate rank-reducing slices instead of reassociative reshapes">
+  ];
+  let dependentDialects = [
+    "linalg::LinalgDialect", "affine::AffineDialect", "memref::MemRefDialect"
+  ];
+}
+
+def LinalgElementwiseOpFusionPass : Pass<"linalg-fuse-elementwise-ops"> {
+  let summary = "Fuse elementwise operations on tensors";
+  let dependentDialects = [
+    "affine::AffineDialect", "linalg::LinalgDialect", "memref::MemRefDialect"
+  ];
+}
+
+def LinalgNamedOpConversionPass: Pass<"linalg-named-op-conversion"> {
+  let summary = "Convert from one named linalg op to another.";
+  let dependentDialects = ["linalg::LinalgDialect", "tensor::TensorDialect"];
+}
+
+def LinalgInlineScalarOperandsPass : Pass<"linalg-inline-scalar-operands"> {
+  let summary = "Inline scalar operands into linalg generic ops";
+  let dependentDialects = [
+    "linalg::LinalgDialect"
+  ];
+}
+
+def LinalgBufferizePass : Pass<"linalg-bufferize"> {
   let summary = "Bufferize the linalg dialect";
-  let constructor = "mlir::createLinalgBufferizePass()";
   let dependentDialects = [
     "affine::AffineDialect",
     "bufferization::BufferizationDialect",
@@ -108,15 +99,13 @@ def LinalgBufferize : Pass<"linalg-bufferize"> {
   ];
 }
 
-def LinalgGeneralization : Pass<"linalg-generalize-named-ops"> {
+def LinalgGeneralizeNamedOpsPass : Pass<"linalg-generalize-named-ops"> {
   let summary = "Convert named ops into generic ops";
-  let constructor = "mlir::createLinalgGeneralizationPass()";
   let dependentDialects = ["linalg::LinalgDialect"];
 }
 
-def LinalgDetensorize : InterfacePass<"linalg-detensorize", "FunctionOpInterface"> {
+def LinalgDetensorizePass : InterfacePass<"linalg-detensorize", "FunctionOpInterface"> {
   let summary = "Detensorize linalg ops";
-  let constructor = "mlir::createLinalgDetensorizePass()";
   let dependentDialects = [];
 
   let description = [{
diff --git a/mlir/lib/Dialect/Linalg/Transforms/Bufferize.cpp b/mlir/lib/Dialect/Linalg/Transforms/Bufferize.cpp
index 73c4d4779750d9..8812ca14ba6109 100644
--- a/mlir/lib/Dialect/Linalg/Transforms/Bufferize.cpp
+++ b/mlir/lib/Dialect/Linalg/Transforms/Bufferize.cpp
@@ -21,7 +21,7 @@
 #include "mlir/Pass/Pass.h"
 
 namespace mlir {
-#define GEN_PASS_DEF_LINALGBUFFERIZE
+#define GEN_PASS_DEF_LINALGBUFFERIZEPASS
 #include "mlir/Dialect/Linalg/Passes.h.inc"
 } // namespace mlir
 
@@ -32,7 +32,9 @@ namespace {
 /// Converts Linalg operations that work on tensor-type operands or results to
 /// work on buffers.
 struct LinalgBufferizePass
-    : public impl::LinalgBufferizeBase<LinalgBufferizePass> {
+    : public impl::LinalgBufferizePassBase<LinalgBufferizePass> {
+  using impl::LinalgBufferizePassBase<
+      LinalgBufferizePass>::LinalgBufferizePassBase;
   void runOnOperation() override {
     BufferizationOptions options = getPartialBufferizationOptions();
     options.opFilter.allowDialect<linalg::LinalgDialect>();
@@ -48,7 +50,3 @@ struct LinalgBufferizePass
   }
 };
 } // namespace
-
-std::unique_ptr<Pass> mlir::createLinalgBufferizePass() {
-  return std::make_unique<LinalgBufferizePass>();
-}
diff --git a/mlir/lib/Dialect/Linalg/Transforms/Detensorize.cpp b/mlir/lib/Dialect/Linalg/Transforms/Detensorize.cpp
index 98cd0444760ece..22968096a68913 100644
--- a/mlir/lib/Dialect/Linalg/Transforms/Detensorize.cpp
+++ b/mlir/lib/Dialect/Linalg/Transforms/Detensorize.cpp
@@ -21,7 +21,7 @@
 #include <utility>
 
 namespace mlir {
-#define GEN_PASS_DEF_LINALGDETENSORIZE
+#define GEN_PASS_DEF_LINALGDETENSORIZEPASS
 #include "mlir/Dialect/Linalg/Passes.h.inc"
 } // namespace mlir
 
@@ -164,7 +164,9 @@ class DetensorizeTypeConverter : public TypeConverter {
 
 /// @see LinalgDetensorize in Linalg/Passes.td for more details.
 struct LinalgDetensorize
-    : public impl::LinalgDetensorizeBase<LinalgDetensorize> {
+    : public impl::LinalgDetensorizePassBase<LinalgDetensorize> {
+  using impl::LinalgDetensorizePassBase<
+      LinalgDetensorize>::LinalgDetensorizePassBase;
   LinalgDetensorize() = default;
 
   class CostModel {
@@ -576,7 +578,3 @@ struct LinalgDetensorize
   }
 };
 } // namespace
-
-std::unique_ptr<Pass> mlir::createLinalgDetensorizePass() {
-  return std::make_unique<LinalgDetensorize>();
-}
diff --git a/mlir/lib/Dialect/Linalg/Transforms/DropUnitDims.cpp b/mlir/lib/Dialect/Linalg/Transforms/DropUnitDims.cpp
index c46e3694b70ecd..45cab81be4f5ff 100644
--- a/mlir/lib/Dialect/Linalg/Transforms/DropUnitDims.cpp
+++ b/mlir/lib/Dialect/Linalg/Transforms/DropUnitDims.cpp
@@ -33,7 +33,7 @@
 #include "llvm/Support/Debug.h"
 
 namespace mlir {
-#define GEN_PASS_DEF_LINALGFOLDUNITEXTENTDIMS
+#define GEN_PASS_DEF_LINALGFOLDUNITEXTENTDIMSPASS
 #include "mlir/Dialect/Linalg/Passes.h.inc"
 } // namespace mlir
 
@@ -689,7 +689,10 @@ void mlir::linalg::populateMoveInitOperandsToInputPattern(
 namespace {
 /// Pass that removes unit-extent dims within generic ops.
 struct LinalgFoldUnitExtentDimsPass
-    : public impl::LinalgFoldUnitExtentDimsBase<LinalgFoldUnitExtentDimsPass> {
+    : public impl::LinalgFoldUnitExtentDimsPassBase<
+          LinalgFoldUnitExtentDimsPass> {
+  using impl::LinalgFoldUnitExtentDimsPassBase<
+      LinalgFoldUnitExtentDimsPass>::LinalgFoldUnitExtentDimsPassBase;
   void runOnOperation() override {
     Operation *op = getOperation();
     MLIRContext *context = op->getContext();
@@ -705,7 +708,3 @@ struct LinalgFoldUnitExtentDimsPass
   }
 };
 } // namespace
-
-std::unique_ptr<Pass> mlir::createLinalgFoldUnitExtentDimsPass() {
-  return std::make_unique<LinalgFoldUnitExtentDimsPass>();
-}
diff --git a/mlir/lib/Dialect/Linalg/Transforms/ElementwiseOpFusion.cpp b/mlir/lib/Dialect/Linalg/Transforms/ElementwiseOpFusion.cpp
index 0d8d670904f2a8..4977940cfbd797 100644
--- a/mlir/lib/Dialect/Linalg/Transforms/ElementwiseOpFusion.cpp
+++ b/mlir/lib/Dialect/Linalg/Transforms/ElementwiseOpFusion.cpp
@@ -27,8 +27,7 @@
 #include <utility>
 
 namespace mlir {
-#define GEN_PASS_DEF_LINALGFOLDUNITEXTENTDIMS
-#define GEN_PASS_DEF_LINALGELEMENTWISEOPFUSION
+#define GEN_PASS_DEF_LINALGELEMENTWISEOPFUSIONPASS
 #include "mlir/Dialect/Linalg/Passes.h.inc"
 } // namespace mlir
 
@@ -1927,8 +1926,10 @@ namespace {
 // favor of test passes that check the functionality of each of the patterns
 // added here individually.
 struct LinalgElementwiseOpFusionPass
-    : public impl::LinalgElementwiseOpFusionBase<
+    : public impl::LinalgElementwiseOpFusionPassBase<
           LinalgElementwiseOpFusionPass> {
+  using impl::LinalgElementwiseOpFusionPassBase<
+      LinalgElementwiseOpFusionPass>::LinalgElementwiseOpFusionPassBase;
   void runOnOperation() override {
     Operation *op = getOperation();
     MLIRContext *context = op->getContext();
@@ -1963,7 +1964,3 @@ struct LinalgElementwiseOpFusionPass
 };
 
 } // namespace
-
-std::unique_ptr<Pass> mlir::createLinalgElementwiseOpFusionPass() {
-  return std::make_unique<LinalgElementwiseOpFusionPass>();
-}
diff --git a/mlir/lib/Dialect/Linalg/Transforms/ElementwiseToLinalg.cpp b/mlir/lib/Dialect/Linalg/Transforms/ElementwiseToLinalg.cpp
index 9088a1de837a79..5508aaf9d87537 100644
--- a/mlir/lib/Dialect/Linalg/Transforms/ElementwiseToLinalg.cpp
+++ b/mlir/lib/Dialect/Linalg/Transforms/ElementwiseToLinalg.cpp
@@ -15,7 +15,7 @@
 #include "mlir/Transforms/DialectConversion.h"
 
 namespace mlir {
-#define GEN_PASS_DEF_CONVERTELEMENTWISETOLINALG
+#define GEN_PASS_DEF_CONVERTELEMENTWISETOLINALGPASS
 #include "mlir/Dialect/Linalg/Passes.h.inc"
 } // namespace mlir
 
@@ -121,8 +121,10 @@ void mlir::linalg::populateElementwiseToLinalgConversionPatterns(
 
 namespace {
 class ConvertElementwiseToLinalgPass
-    : public impl::ConvertElementwiseToLinalgBase<
+    : public impl::ConvertElementwiseToLinalgPassBase<
           ConvertElementwiseToLinalgPass> {
+  using impl::ConvertElementwiseToLinalgPassBase<
+      ConvertElementwiseToLinalgPass>::ConvertElementwiseToLinalgPassBase;
 
   void runOnOperation() final {
     auto *func = getOperation();
@@ -140,7 +142,3 @@ class ConvertElementwiseToLinalgPass
   }
 };
 } // namespace
-
-std::unique_ptr<Pass> mlir::createConvertElementwiseToLinalgPass() {
-  return std::make_unique<ConvertElementwiseToLinalgPass>();
-}
diff --git a/mlir/lib/Dialect/Linalg/Transforms/Generalization.cpp b/mlir/lib/Dialect/Linalg/Transforms/Generalization.cpp
index d03d1f3a163c32..7ab3fef5dd039d 100644
--- a/mlir/lib/Dialect/Linalg/Transforms/Generalization.cpp
+++ b/mlir/lib/Dialect/Linalg/Transforms/Generalization.cpp
@@ -26,7 +26,7 @@
 #include "llvm/Support/Debug.h"
 
 namespace mlir {
-#define GEN_PASS_DEF_LINALGGENERALIZATION
+#define GEN_PASS_DEF_LINALGGENERALIZENAMEDOPSPASS
 #include "mlir/Dialect/Linalg/Passes.h.inc"
 } // namespace mlir
 
@@ -76,14 +76,17 @@ FailureOr<GenericOp> mlir::linalg::generalizeNamedOp(RewriterBase &rewriter,
 
 namespace {
 
-struct LinalgGeneralizationPass
-    : public impl::LinalgGeneralizationBase<LinalgGeneralizationPass> {
+struct LinalgGeneralizeNamedOpsPass
+    : public impl::LinalgGeneralizeNamedOpsPassBase<
+          LinalgGeneralizeNamedOpsPass> {
+  using impl::LinalgGeneralizeNamedOpsPassBase<
+      LinalgGeneralizeNamedOpsPass>::LinalgGeneralizeNamedOpsPassBase;
   void runOnOperation() override;
 };
 
 } // namespace
 
-void LinalgGeneralizationPass::runOnOperation() {
+void LinalgGeneralizeNamedOpsPass::runOnOperation() {
   RewritePatternSet patterns(&getContext());
   populateLinalgNamedOpsGeneralizationPatterns(patterns);
   (void)applyPatternsAndFoldGreedily(getOperation(), std::move(patterns));
@@ -93,7 +96,3 @@ void mlir::linalg::populateLinalgNamedOpsGeneralizationPatterns(
     RewritePatternSet &patterns) {
   patterns.add<LinalgGeneralizationPattern>(patterns.getContext());
 }
-
-std::unique_ptr<Pass> mlir::createLinalgGeneralizationPass() {
-  return std::make_unique<LinalgGeneralizationPass>();
-}
diff --git a/mlir/lib/Dialect/Linalg/Transforms/InlineScalarOperands.cpp b/mlir/lib/Dialect/Linalg/Transforms/InlineScalarOperands.cpp
index 34db710b1721d6..6db51f4b84d112 100644
--- a/mlir/lib/Dialect/Linalg/Transforms/InlineScalarOperands.cpp
+++ b/mlir/lib/Dialect/Linalg/Transforms/InlineScalarOperands.cpp
@@ -23,7 +23,7 @@
 #include "mlir/Transforms/GreedyPatternRewriteDriver.h"
 
 namespace mlir {
-#define GEN_PASS_DEF_LINALGINLINESCALAROPERANDS
+#define GEN_PASS_DEF_LINALGINLINESCALAROPERANDSPASS
 #include "mlir/Dialect/Linalg/Passes.h.inc"
 } // namespace mlir
 
@@ -101,8 +101,10 @@ void mlir::linalg::populateInlineConstantOperandsPatterns(
 namespace {
 /// Pass that removes unit-extent dims within generic ops.
 struct LinalgInlineScalarOperandsPass
-    : public impl::LinalgInlineScalarOperandsBase<
+    : public impl::LinalgInlineScalarOperandsPassBase<
           LinalgInlineScalarOperandsPass> {
+  using impl::LinalgInlineScalarOperandsPassBase<
+      LinalgInlineScalarOperandsPass>::LinalgInlineScalarOperandsPassBase;
   void runOnOperation() override {
     Operation *op = getOperation();
     MLIRContext &ctx = getContext();
@@ -112,7 +114,3 @@ struct LinalgInlineScalarOperandsPass
   }
 };
 } // namespace
-
-std::unique_ptr<Pass> mlir::createLinalgInlineScalarOperandsPass() {
-  return std::make_unique<LinalgInlineScalarOperandsPass>();
-}
diff --git a/mlir/lib/Dialect/Linalg/Transforms/Loops.cpp b/mlir/lib/Dialect/Linalg/Transforms/Loops.cpp
index 4c93da6fe9253f..b0a4de2da1e869 100644
--- a/mlir/lib/Dialect/Linalg/Transforms/Loops.cpp
+++ b/mlir/lib/Dialect/Linalg/Transforms/Loops.cpp
@@ -27,9 +27,9 @@
 #include "llvm/ADT/TypeSwitch.h"
 
 namespace mlir {
-#define GEN_PASS_DEF_LINALGLOWERTOAFFINELOOPS
-#define GEN_PASS_DEF_LINALGLOWERTOLOOPS
-#define GEN_PASS_DEF_LINALGLOWERTOPARALLELLOOPS
+#define GEN_PASS_DEF_CONVERTLINALGTOAFFINELOOPSPASS
+#define GEN_PASS_DEF_CONVERTLINALGTOLOOPSPASS
+#define GEN_PASS_DEF_CONVERTLINALGTOPARALLELLOOPSPASS
 #include "mlir/Dialect/Linalg/Passes.h.inc"
 } // namespace mlir
 
@@ -326,7 +326,9 @@ static void lowerLinalgToLoopsImpl(Operation *enclosingOp) {
 }
 
 struct LowerToAffineLoops
-    : public impl::LinalgLowerToAffineLoopsBase<LowerToAffineLoops> {
+    : public impl::ConvertLinalgToAffineLoopsPassBase<LowerToAffineLoops> {
+  using impl::ConvertLinalgToAffineLoopsPassBase<
+      LowerToAffineLoops>::ConvertLinalgToAffineLoopsPassBase;
   void getDependentDialects(DialectRegistry &registry) const override {
     registry.insert<memref::MemRefDialect>();
   }
@@ -335,7 +337,9 @@ struct LowerToAffineLoops
   }
 };
 
-struct LowerToLoops : public impl::LinalgLowerToLoopsBase<LowerToLoops> {
+struct LowerToLoops : public impl::ConvertLinalgToLoopsPassBase<LowerToLoops> {
+  using impl::ConvertLinalgToLoopsPassBase<
+      LowerToLoops>::ConvertLinalgToLoopsPassBase;
   void getDependentDialects(DialectRegistry &registry) const override {
     registry.insert<memref::MemRefDialect, scf::SCFDialect>();
   }
@@ -345,7 +349,9 @@ struct LowerToLoops : public impl::LinalgLowerToLoopsBase<LowerToLoops> {
 };
 
 struct LowerToParallelLoops
-    : public impl::LinalgLowerToParallelLoopsBase<LowerToParallelLoops> {
+    : public impl::ConvertLinalgToParallelLoopsPassBase<LowerToParallelLoops> {
+  using impl::ConvertLinalgToParallelLoopsPassBase<
+      LowerToParallelLoops>::ConvertLinalgToParallelLoopsPassBase;
   void runOnOperation() override {
     lowerLinalgToLoopsImpl<scf::ParallelOp>(getOperation());
   }
@@ -353,18 +359,6 @@ struct LowerToParallelLoops
 
 } // namespace
 
-std::unique_ptr<Pass> mlir::createConvertLinalgToLoopsPass() {
-  return std::make_unique<LowerToLoops>();
-}
-
-std::unique_ptr<Pass> mlir::createConvertLinalgToParallelLoopsPass() {
-  return std::make_unique<LowerToParallelLoops>();
-}
-
-std::unique_ptr<Pass> mlir::createConvertLinalgToAffineLoopsPass() {
-  return std::make_unique<LowerToAffineLoops>();
-}
-
 /// Emits a loop nest of `affine.for` with the proper body for `linalgOp`.
 FailureOr<LinalgLoops>
 mlir::linalg::linalgOpToAffineLoops(RewriterBase &rewriter, LinalgOp linalgOp) {
diff --git a/mlir/lib/Dialect/Linalg/Transforms/NamedOpConversions.cpp b/mlir/lib/Dialect/Linalg/Transforms/NamedOpConversions.cpp
index 250360603fa5dd..84bde1bc0b8464 100644
--- a/...
[truncated]

@qedawkins qedawkins merged commit 1e98d48 into llvm:main Feb 24, 2024
@qedawkins qedawkins deleted the linalg_tablegen_pass_constructor branch February 24, 2024 19:35
joker-eph
joker-eph reviewed Feb 24, 2024
View reviewed changes
Copy link
Collaborator

@joker-eph joker-eph left a comment

Choose a reason for hiding this comment

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

Thanks so much!

I had filed a ticket for this years ago and it never got addressed (I can't seem to find it just now)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@joker-eph joker-eph joker-eph left review comments

@MaheshRavishankar MaheshRavishankar MaheshRavishankar approved these changes

@aartbik aartbik Awaiting requested review from aartbik aartbik is a code owner

@PeimingLiu PeimingLiu Awaiting requested review from PeimingLiu PeimingLiu is a code owner

@yinying-lisa-li yinying-lisa-li Awaiting requested review from yinying-lisa-li yinying-lisa-li is a code owner

@matthias-springer matthias-springer Awaiting requested review from matthias-springer matthias-springer is a code owner

@nicolasvasilache nicolasvasilache Awaiting requested review from nicolasvasilache nicolasvasilache is a code owner

@dcaballe dcaballe Awaiting requested review from dcaballe dcaballe is a code owner

Assignees
No one assigned
Labels
mlir:linalg mlir:sparse Sparse compiler in MLIR mlir
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@qedawkins @llvmbot @MaheshRavishankar @joker-eph