diff --git a/mlir/include/mlir/Transforms/Passes.td b/mlir/include/mlir/Transforms/Passes.td
--- a/mlir/include/mlir/Transforms/Passes.td
+++ b/mlir/include/mlir/Transforms/Passes.td
@@ -39,7 +39,9 @@
            /*default=*/"10",
            "Max. iterations between applying patterns / simplifying regions">,
     Option<"maxNumRewrites", "max-num-rewrites", "int64_t", /*default=*/"-1",
-           "Max. number of pattern rewrites within an iteration">
+           "Max. number of pattern rewrites within an iteration">,
+    Option<"testConvergence", "test-convergence", "bool", /*default=*/"false",
+           "Test only: Fail pass on non-convergence to detect cyclic pattern">
   ] # RewritePassUtils.options;
 }
 
diff --git a/mlir/lib/Dialect/GPU/IR/GPUDialect.cpp b/mlir/lib/Dialect/GPU/IR/GPUDialect.cpp
--- a/mlir/lib/Dialect/GPU/IR/GPUDialect.cpp
+++ b/mlir/lib/Dialect/GPU/IR/GPUDialect.cpp
@@ -686,6 +686,8 @@
       // Check if size is trivially one.
       if (!matchPattern(size, m_One()))
         return;
+      if (id.getUses().empty())
+        return;
       if (!simplified) {
         // Create a zero value the first time.
         OpBuilder::InsertionGuard guard(rewriter);
@@ -693,7 +695,7 @@
         zero =
             rewriter.create<arith::ConstantIndexOp>(op.getLoc(), /*value=*/0);
       }
-      id.replaceAllUsesWith(zero);
+      rewriter.replaceAllUsesWith(id, zero);
       simplified = true;
     };
     constPropIdUses(op.getBlockIds().x, op.getGridSizeX());
diff --git a/mlir/lib/Dialect/OpenACC/IR/OpenACC.cpp b/mlir/lib/Dialect/OpenACC/IR/OpenACC.cpp
--- a/mlir/lib/Dialect/OpenACC/IR/OpenACC.cpp
+++ b/mlir/lib/Dialect/OpenACC/IR/OpenACC.cpp
@@ -178,16 +178,15 @@
     // Early return if there is no condition.
     Value ifCond = op.getIfCond();
     if (!ifCond)
-      return success();
+      return failure();
 
     IntegerAttr constAttr;
-    if (matchPattern(ifCond, m_Constant(&constAttr))) {
-      if (constAttr.getInt())
-        rewriter.updateRootInPlace(op,
-                                   [&]() { op.getIfCondMutable().erase(0); });
-      else
-        rewriter.eraseOp(op);
-    }
+    if (!matchPattern(ifCond, m_Constant(&constAttr)))
+      return failure();
+    if (constAttr.getInt())
+      rewriter.updateRootInPlace(op, [&]() { op.getIfCondMutable().erase(0); });
+    else
+      rewriter.eraseOp(op);
 
     return success();
   }
diff --git a/mlir/lib/Transforms/Canonicalizer.cpp b/mlir/lib/Transforms/Canonicalizer.cpp
--- a/mlir/lib/Transforms/Canonicalizer.cpp
+++ b/mlir/lib/Transforms/Canonicalizer.cpp
@@ -57,8 +57,11 @@
     config.enableRegionSimplification = enableRegionSimplification;
     config.maxIterations = maxIterations;
     config.maxNumRewrites = maxNumRewrites;
+    LogicalResult converged =
+        applyPatternsAndFoldGreedily(getOperation(), patterns, config);
     // Canonicalization is best-effort. Non-convergence is not a pass failure.
-    (void)applyPatternsAndFoldGreedily(getOperation(), patterns, config);
+    if (testConvergence && failed(converged))
+      signalPassFailure();
   }
 
   FrozenRewritePatternSet patterns;
diff --git a/mlir/test/Dialect/AMDGPU/canonicalize.mlir b/mlir/test/Dialect/AMDGPU/canonicalize.mlir
--- a/mlir/test/Dialect/AMDGPU/canonicalize.mlir
+++ b/mlir/test/Dialect/AMDGPU/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s -split-input-file -canonicalize  | FileCheck %s
+// RUN: mlir-opt %s -split-input-file -canonicalize="test-convergence" | FileCheck %s
 
 // CHECK-LABEL: func @known_oob_load
 func.func @known_oob_load(%arg0: memref<4xf32>) -> f32 {
diff --git a/mlir/test/Dialect/Affine/canonicalize.mlir b/mlir/test/Dialect/Affine/canonicalize.mlir
--- a/mlir/test/Dialect/Affine/canonicalize.mlir
+++ b/mlir/test/Dialect/Affine/canonicalize.mlir
@@ -1,5 +1,5 @@
-// RUN: mlir-opt -allow-unregistered-dialect %s -split-input-file -canonicalize | FileCheck %s
-// RUN: mlir-opt -allow-unregistered-dialect %s -split-input-file -canonicalize="top-down=0" | FileCheck %s --check-prefix=CHECK-BOTTOM-UP
+// RUN: mlir-opt -allow-unregistered-dialect %s -split-input-file -canonicalize="test-convergence" | FileCheck %s
+// RUN: mlir-opt -allow-unregistered-dialect %s -split-input-file -canonicalize="test-convergence top-down=0" | FileCheck %s --check-prefix=CHECK-BOTTOM-UP
 
 // -----
 
diff --git a/mlir/test/Dialect/Arith/canonicalize.mlir b/mlir/test/Dialect/Arith/canonicalize.mlir
--- a/mlir/test/Dialect/Arith/canonicalize.mlir
+++ b/mlir/test/Dialect/Arith/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s -canonicalize --split-input-file | FileCheck %s
+// RUN: mlir-opt %s -canonicalize="test-convergence" --split-input-file | FileCheck %s
 
 // CHECK-LABEL: @select_same_val
 //       CHECK:   return %arg1
diff --git a/mlir/test/Dialect/Bufferization/canonicalize.mlir b/mlir/test/Dialect/Bufferization/canonicalize.mlir
--- a/mlir/test/Dialect/Bufferization/canonicalize.mlir
+++ b/mlir/test/Dialect/Bufferization/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s -canonicalize --split-input-file \
+// RUN: mlir-opt %s -canonicalize="test-convergence" --split-input-file \
 // RUN:   -allow-unregistered-dialect |\
 // RUN: FileCheck %s
 
diff --git a/mlir/test/Dialect/Builtin/canonicalize.mlir b/mlir/test/Dialect/Builtin/canonicalize.mlir
--- a/mlir/test/Dialect/Builtin/canonicalize.mlir
+++ b/mlir/test/Dialect/Builtin/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s -canonicalize | FileCheck %s
+// RUN: mlir-opt %s -canonicalize="test-convergence" | FileCheck %s
 
 //===----------------------------------------------------------------------===//
 // UnrealizedConversionCastOp
diff --git a/mlir/test/Dialect/Complex/canonicalize.mlir b/mlir/test/Dialect/Complex/canonicalize.mlir
--- a/mlir/test/Dialect/Complex/canonicalize.mlir
+++ b/mlir/test/Dialect/Complex/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s -canonicalize | FileCheck %s
+// RUN: mlir-opt %s -canonicalize="test-convergence" | FileCheck %s
 
 // CHECK-LABEL: func @create_of_real_and_imag
 // CHECK-SAME: (%[[CPLX:.*]]: complex<f32>)
diff --git a/mlir/test/Dialect/ControlFlow/canonicalize.mlir b/mlir/test/Dialect/ControlFlow/canonicalize.mlir
--- a/mlir/test/Dialect/ControlFlow/canonicalize.mlir
+++ b/mlir/test/Dialect/ControlFlow/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s -allow-unregistered-dialect -pass-pipeline='builtin.module(func.func(canonicalize))' -split-input-file | FileCheck --dump-input-context 20 %s
+// RUN: mlir-opt %s -allow-unregistered-dialect -canonicalize="test-convergence" -split-input-file | FileCheck --dump-input-context 20 %s
 
 /// Test the folding of BranchOp.
 
diff --git a/mlir/test/Dialect/GPU/canonicalize.mlir b/mlir/test/Dialect/GPU/canonicalize.mlir
--- a/mlir/test/Dialect/GPU/canonicalize.mlir
+++ b/mlir/test/Dialect/GPU/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s -canonicalize --split-input-file -allow-unregistered-dialect | FileCheck %s
+// RUN: mlir-opt %s -canonicalize="test-convergence" --split-input-file -allow-unregistered-dialect | FileCheck %s
 
 // Fold all the gpu.wait ops as they are redundant.
 // CHECK-LABEL: func @fold_wait_op_test1
diff --git a/mlir/test/Dialect/LLVMIR/canonicalize.mlir b/mlir/test/Dialect/LLVMIR/canonicalize.mlir
--- a/mlir/test/Dialect/LLVMIR/canonicalize.mlir
+++ b/mlir/test/Dialect/LLVMIR/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt -canonicalize %s -split-input-file | FileCheck %s
+// RUN: mlir-opt -canonicalize="test-convergence" %s -split-input-file | FileCheck %s
 
 // CHECK-LABEL: fold_extractvalue
 llvm.func @fold_extractvalue() -> i32 {
diff --git a/mlir/test/Dialect/Linalg/canonicalize.mlir b/mlir/test/Dialect/Linalg/canonicalize.mlir
--- a/mlir/test/Dialect/Linalg/canonicalize.mlir
+++ b/mlir/test/Dialect/Linalg/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s -canonicalize -split-input-file | FileCheck %s
+// RUN: mlir-opt %s -canonicalize="test-convergence" -split-input-file | FileCheck %s
 
 // CHECK-LABEL: func @memref_cast(
 func.func @memref_cast(%a: index, %b: index) -> memref<?x?xf32> {
diff --git a/mlir/test/Dialect/Math/canonicalize.mlir b/mlir/test/Dialect/Math/canonicalize.mlir
--- a/mlir/test/Dialect/Math/canonicalize.mlir
+++ b/mlir/test/Dialect/Math/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s -canonicalize | FileCheck %s
+// RUN: mlir-opt %s -canonicalize="test-convergence" | FileCheck %s
 
 // CHECK-LABEL: @ceil_fold
 // CHECK: %[[cst:.+]] = arith.constant 1.000000e+00 : f32
diff --git a/mlir/test/Dialect/MemRef/canonicalize.mlir b/mlir/test/Dialect/MemRef/canonicalize.mlir
--- a/mlir/test/Dialect/MemRef/canonicalize.mlir
+++ b/mlir/test/Dialect/MemRef/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s -canonicalize --split-input-file -allow-unregistered-dialect | FileCheck %s
+// RUN: mlir-opt %s -canonicalize="test-convergence" --split-input-file -allow-unregistered-dialect | FileCheck %s
 
 // CHECK-LABEL: func @subview_of_size_memcast
 //  CHECK-SAME:   %[[ARG0:.[a-z0-9A-Z_]+]]: memref<4x6x16x32xi8>
diff --git a/mlir/test/Dialect/OpenACC/canonicalize.mlir b/mlir/test/Dialect/OpenACC/canonicalize.mlir
--- a/mlir/test/Dialect/OpenACC/canonicalize.mlir
+++ b/mlir/test/Dialect/OpenACC/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s -canonicalize -split-input-file | FileCheck %s
+// RUN: mlir-opt %s -canonicalize="test-convergence" -split-input-file | FileCheck %s
 
 func.func @testenterdataop(%a: memref<10xf32>) -> () {
   %ifCond = arith.constant true
diff --git a/mlir/test/Dialect/OpenMP/canonicalize.mlir b/mlir/test/Dialect/OpenMP/canonicalize.mlir
--- a/mlir/test/Dialect/OpenMP/canonicalize.mlir
+++ b/mlir/test/Dialect/OpenMP/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s -canonicalize -split-input-file | FileCheck %s
+// RUN: mlir-opt %s -canonicalize="test-convergence" -split-input-file | FileCheck %s
 
 func.func @update_no_op(%x : memref<i32>) {
   omp.atomic.update %x : memref<i32> {
diff --git a/mlir/test/Dialect/PDL/canonicalize.mlir b/mlir/test/Dialect/PDL/canonicalize.mlir
--- a/mlir/test/Dialect/PDL/canonicalize.mlir
+++ b/mlir/test/Dialect/PDL/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt -canonicalize %s | FileCheck %s
+// RUN: mlir-opt -canonicalize="test-convergence" %s | FileCheck %s
 
 pdl.pattern @operation_op : benefit(1) {
   %root = operation "foo.op"
diff --git a/mlir/test/Dialect/Quant/canonicalize.mlir b/mlir/test/Dialect/Quant/canonicalize.mlir
--- a/mlir/test/Dialect/Quant/canonicalize.mlir
+++ b/mlir/test/Dialect/Quant/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s -split-input-file -pass-pipeline='builtin.module(func.func(canonicalize))' | FileCheck %s
+// RUN: mlir-opt %s -split-input-file -canonicalize="test-convergence" | FileCheck %s
 
 // -----
 // CHECK-LABEL: redundant_scast
diff --git a/mlir/test/Dialect/SCF/canonicalize.mlir b/mlir/test/Dialect/SCF/canonicalize.mlir
--- a/mlir/test/Dialect/SCF/canonicalize.mlir
+++ b/mlir/test/Dialect/SCF/canonicalize.mlir
@@ -1,7 +1,4 @@
-// RUN: mlir-opt %s -pass-pipeline='builtin.module(func.func(canonicalize))' -split-input-file | FileCheck %s
-
-
-// -----
+// RUN: mlir-opt %s -canonicalize="test-convergence" -split-input-file | FileCheck %s
 
 func.func @single_iteration_some(%A: memref<?x?x?xi32>) {
   %c0 = arith.constant 0 : index
diff --git a/mlir/test/Dialect/SPIRV/Transforms/canonicalize.mlir b/mlir/test/Dialect/SPIRV/Transforms/canonicalize.mlir
--- a/mlir/test/Dialect/SPIRV/Transforms/canonicalize.mlir
+++ b/mlir/test/Dialect/SPIRV/Transforms/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s -split-input-file -pass-pipeline='builtin.module(func.func(canonicalize))' | FileCheck %s
+// RUN: mlir-opt %s -split-input-file -canonicalize="test-convergence" | FileCheck %s
 
 //===----------------------------------------------------------------------===//
 // spirv.AccessChain
diff --git a/mlir/test/Dialect/Shape/canonicalize.mlir b/mlir/test/Dialect/Shape/canonicalize.mlir
--- a/mlir/test/Dialect/Shape/canonicalize.mlir
+++ b/mlir/test/Dialect/Shape/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt -split-input-file -allow-unregistered-dialect -canonicalize %s | FileCheck %s
+// RUN: mlir-opt -split-input-file -allow-unregistered-dialect -canonicalize="test-convergence" %s | FileCheck %s
 
 // CHECK-LABEL: func @f
 func.func @f(%arg0: tensor<2x3x4xf32>) -> tensor<3xindex> {
diff --git a/mlir/test/Dialect/Tensor/canonicalize.mlir b/mlir/test/Dialect/Tensor/canonicalize.mlir
--- a/mlir/test/Dialect/Tensor/canonicalize.mlir
+++ b/mlir/test/Dialect/Tensor/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s -split-input-file -canonicalize | FileCheck %s
+// RUN: mlir-opt %s -split-input-file -canonicalize="test-convergence" | FileCheck %s
 
 // Checks that NOP casts are removed.
 // CHECK-LABEL: cast_values
diff --git a/mlir/test/Dialect/Tosa/canonicalize.mlir b/mlir/test/Dialect/Tosa/canonicalize.mlir
--- a/mlir/test/Dialect/Tosa/canonicalize.mlir
+++ b/mlir/test/Dialect/Tosa/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt --canonicalize %s | FileCheck %s
+// RUN: mlir-opt --canonicalize="test-convergence" %s | FileCheck %s
 
 // CHECK-LABEL: @argmax_nofold
 func.func @argmax_nofold(%arg0: tensor<?x1xf32>) -> tensor<?x1xf32> {
diff --git a/mlir/test/Dialect/Vector/canonicalize.mlir b/mlir/test/Dialect/Vector/canonicalize.mlir
--- a/mlir/test/Dialect/Vector/canonicalize.mlir
+++ b/mlir/test/Dialect/Vector/canonicalize.mlir
@@ -1,6 +1,4 @@
-// RUN: mlir-opt %s -pass-pipeline='builtin.module(func.func(canonicalize))' -split-input-file -allow-unregistered-dialect | FileCheck %s
-
-// -----
+// RUN: mlir-opt %s -canonicalize="test-convergence" -split-input-file -allow-unregistered-dialect | FileCheck %s
 
 // CHECK-LABEL: create_vector_mask_to_constant_mask
 func.func @create_vector_mask_to_constant_mask() -> (vector<4x3xi1>) {
diff --git a/mlir/test/Pass/run-reproducer.mlir b/mlir/test/Pass/run-reproducer.mlir
--- a/mlir/test/Pass/run-reproducer.mlir
+++ b/mlir/test/Pass/run-reproducer.mlir
@@ -14,7 +14,7 @@
   external_resources: {
     mlir_reproducer: {
       verify_each: true,
-      // CHECK:  builtin.module(func.func(cse,canonicalize{ max-iterations=1 max-num-rewrites=-1 region-simplify=false top-down=false}))
+      // CHECK:  builtin.module(func.func(cse,canonicalize{ max-iterations=1 max-num-rewrites=-1 region-simplify=false test-convergence=false top-down=false}))
       pipeline: "builtin.module(func.func(cse,canonicalize{max-iterations=1 max-num-rewrites=-1 region-simplify=false top-down=false}))",
       disable_threading: true
     }
diff --git a/mlir/test/Transforms/canonicalize.mlir b/mlir/test/Transforms/canonicalize.mlir
--- a/mlir/test/Transforms/canonicalize.mlir
+++ b/mlir/test/Transforms/canonicalize.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt -allow-unregistered-dialect %s -pass-pipeline='builtin.module(func.func(canonicalize))' -split-input-file | FileCheck %s
+// RUN: mlir-opt -allow-unregistered-dialect %s -canonicalize="test-convergence" -split-input-file | FileCheck %s
 
 // CHECK-LABEL: func @test_subi_zero
 func.func @test_subi_zero(%arg0: i32) -> i32 {