[mlir][affine] Fix a crash when cast incompatible type #145162
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
This PR fixes a crash in `getSemiAffineExprFromFlatForm` when localExpr is not `AffineBinaryOpExpr`.
Member
|
@llvm/pr-subscribers-mlir Author: Longsheng Mou (CoTinker) ChangesThis PR fixes a crash in Full diff: https://github.com/llvm/llvm-project/pull/145162.diff 2 Files Affected:
diff --git a/mlir/lib/IR/AffineExpr.cpp b/mlir/lib/IR/AffineExpr.cpp
index c8d9761511bec..cc81f9d19aca7 100644
--- a/mlir/lib/IR/AffineExpr.cpp
+++ b/mlir/lib/IR/AffineExpr.cpp
@@ -1174,11 +1174,15 @@ static AffineExpr getSemiAffineExprFromFlatForm(ArrayRef<int64_t> flatExprs,
// the indices in `coefficients` map, and affine expression corresponding to
// in indices in `indexToExprMap` map.
for (const auto &it : llvm::enumerate(localExprs)) {
- AffineExpr expr = it.value();
if (flatExprs[numDims + numSymbols + it.index()] == 0)
continue;
- AffineExpr lhs = cast<AffineBinaryOpExpr>(expr).getLHS();
- AffineExpr rhs = cast<AffineBinaryOpExpr>(expr).getRHS();
+ AffineExpr expr = it.value();
+ auto binaryExpr = dyn_cast<AffineBinaryOpExpr>(expr);
+ if (!binaryExpr)
+ continue;
+
+ AffineExpr lhs = binaryExpr.getLHS();
+ AffineExpr rhs = binaryExpr.getRHS();
if (!((isa<AffineDimExpr>(lhs) || isa<AffineSymbolExpr>(lhs)) &&
(isa<AffineDimExpr>(rhs) || isa<AffineSymbolExpr>(rhs) ||
isa<AffineConstantExpr>(rhs)))) {
diff --git a/mlir/test/Dialect/Affine/simplify-structures.mlir b/mlir/test/Dialect/Affine/simplify-structures.mlir
index e4a8512b002ee..6f2737a982752 100644
--- a/mlir/test/Dialect/Affine/simplify-structures.mlir
+++ b/mlir/test/Dialect/Affine/simplify-structures.mlir
@@ -592,3 +592,19 @@ func.func @semiaffine_modulo_dim(%arg0: index, %arg1: index, %arg2: index) -> in
//CHECK: affine.apply #[[$MAP]]()[%{{.*}}, %{{.*}}, %{{.*}}]
return %a : index
}
+
+// -----
+
+// CHECK-LABEL: func @semiaffine_simplification_floordiv_and_ceildiv_const
+func.func @semiaffine_simplification_floordiv_and_ceildiv_const(%arg0: tensor<?xf32>) -> (index, index) {
+ %c0 = arith.constant 0 : index
+ %c1 = arith.constant 1 : index
+ %c13 = arith.constant 13 : index
+ %dim = tensor.dim %arg0, %c0 : tensor<?xf32>
+ %a = affine.apply affine_map<()[s0, s1, s2] -> (s0 floordiv (s1 + (-s1 + 2) * (-s1 + s1 * s2 + 1)))>()[%c13, %dim, %c1]
+ %b = affine.apply affine_map<()[s0, s1, s2] -> (s0 ceildiv (s1 + (-s1 + 2) * (-s1 + s1 * s2 + 1)))>()[%c13, %dim, %c1]
+ // CHECK: %[[C6:.*]] = arith.constant 6 : index
+ // CHECK-NEXT: %[[C7:.*]] = arith.constant 7 : index
+ // CHECK-NEXT: return %[[C6]], %[[C7]]
+ return %a, %b : index, index
+}
|
Member
|
@llvm/pr-subscribers-mlir-core Author: Longsheng Mou (CoTinker) ChangesThis PR fixes a crash in Full diff: https://github.com/llvm/llvm-project/pull/145162.diff 2 Files Affected:
diff --git a/mlir/lib/IR/AffineExpr.cpp b/mlir/lib/IR/AffineExpr.cpp
index c8d9761511bec..cc81f9d19aca7 100644
--- a/mlir/lib/IR/AffineExpr.cpp
+++ b/mlir/lib/IR/AffineExpr.cpp
@@ -1174,11 +1174,15 @@ static AffineExpr getSemiAffineExprFromFlatForm(ArrayRef<int64_t> flatExprs,
// the indices in `coefficients` map, and affine expression corresponding to
// in indices in `indexToExprMap` map.
for (const auto &it : llvm::enumerate(localExprs)) {
- AffineExpr expr = it.value();
if (flatExprs[numDims + numSymbols + it.index()] == 0)
continue;
- AffineExpr lhs = cast<AffineBinaryOpExpr>(expr).getLHS();
- AffineExpr rhs = cast<AffineBinaryOpExpr>(expr).getRHS();
+ AffineExpr expr = it.value();
+ auto binaryExpr = dyn_cast<AffineBinaryOpExpr>(expr);
+ if (!binaryExpr)
+ continue;
+
+ AffineExpr lhs = binaryExpr.getLHS();
+ AffineExpr rhs = binaryExpr.getRHS();
if (!((isa<AffineDimExpr>(lhs) || isa<AffineSymbolExpr>(lhs)) &&
(isa<AffineDimExpr>(rhs) || isa<AffineSymbolExpr>(rhs) ||
isa<AffineConstantExpr>(rhs)))) {
diff --git a/mlir/test/Dialect/Affine/simplify-structures.mlir b/mlir/test/Dialect/Affine/simplify-structures.mlir
index e4a8512b002ee..6f2737a982752 100644
--- a/mlir/test/Dialect/Affine/simplify-structures.mlir
+++ b/mlir/test/Dialect/Affine/simplify-structures.mlir
@@ -592,3 +592,19 @@ func.func @semiaffine_modulo_dim(%arg0: index, %arg1: index, %arg2: index) -> in
//CHECK: affine.apply #[[$MAP]]()[%{{.*}}, %{{.*}}, %{{.*}}]
return %a : index
}
+
+// -----
+
+// CHECK-LABEL: func @semiaffine_simplification_floordiv_and_ceildiv_const
+func.func @semiaffine_simplification_floordiv_and_ceildiv_const(%arg0: tensor<?xf32>) -> (index, index) {
+ %c0 = arith.constant 0 : index
+ %c1 = arith.constant 1 : index
+ %c13 = arith.constant 13 : index
+ %dim = tensor.dim %arg0, %c0 : tensor<?xf32>
+ %a = affine.apply affine_map<()[s0, s1, s2] -> (s0 floordiv (s1 + (-s1 + 2) * (-s1 + s1 * s2 + 1)))>()[%c13, %dim, %c1]
+ %b = affine.apply affine_map<()[s0, s1, s2] -> (s0 ceildiv (s1 + (-s1 + 2) * (-s1 + s1 * s2 + 1)))>()[%c13, %dim, %c1]
+ // CHECK: %[[C6:.*]] = arith.constant 6 : index
+ // CHECK-NEXT: %[[C7:.*]] = arith.constant 7 : index
+ // CHECK-NEXT: return %[[C6]], %[[C7]]
+ return %a, %b : index, index
+}
|
Member
|
@llvm/pr-subscribers-mlir-affine Author: Longsheng Mou (CoTinker) ChangesThis PR fixes a crash in Full diff: https://github.com/llvm/llvm-project/pull/145162.diff 2 Files Affected:
diff --git a/mlir/lib/IR/AffineExpr.cpp b/mlir/lib/IR/AffineExpr.cpp
index c8d9761511bec..cc81f9d19aca7 100644
--- a/mlir/lib/IR/AffineExpr.cpp
+++ b/mlir/lib/IR/AffineExpr.cpp
@@ -1174,11 +1174,15 @@ static AffineExpr getSemiAffineExprFromFlatForm(ArrayRef<int64_t> flatExprs,
// the indices in `coefficients` map, and affine expression corresponding to
// in indices in `indexToExprMap` map.
for (const auto &it : llvm::enumerate(localExprs)) {
- AffineExpr expr = it.value();
if (flatExprs[numDims + numSymbols + it.index()] == 0)
continue;
- AffineExpr lhs = cast<AffineBinaryOpExpr>(expr).getLHS();
- AffineExpr rhs = cast<AffineBinaryOpExpr>(expr).getRHS();
+ AffineExpr expr = it.value();
+ auto binaryExpr = dyn_cast<AffineBinaryOpExpr>(expr);
+ if (!binaryExpr)
+ continue;
+
+ AffineExpr lhs = binaryExpr.getLHS();
+ AffineExpr rhs = binaryExpr.getRHS();
if (!((isa<AffineDimExpr>(lhs) || isa<AffineSymbolExpr>(lhs)) &&
(isa<AffineDimExpr>(rhs) || isa<AffineSymbolExpr>(rhs) ||
isa<AffineConstantExpr>(rhs)))) {
diff --git a/mlir/test/Dialect/Affine/simplify-structures.mlir b/mlir/test/Dialect/Affine/simplify-structures.mlir
index e4a8512b002ee..6f2737a982752 100644
--- a/mlir/test/Dialect/Affine/simplify-structures.mlir
+++ b/mlir/test/Dialect/Affine/simplify-structures.mlir
@@ -592,3 +592,19 @@ func.func @semiaffine_modulo_dim(%arg0: index, %arg1: index, %arg2: index) -> in
//CHECK: affine.apply #[[$MAP]]()[%{{.*}}, %{{.*}}, %{{.*}}]
return %a : index
}
+
+// -----
+
+// CHECK-LABEL: func @semiaffine_simplification_floordiv_and_ceildiv_const
+func.func @semiaffine_simplification_floordiv_and_ceildiv_const(%arg0: tensor<?xf32>) -> (index, index) {
+ %c0 = arith.constant 0 : index
+ %c1 = arith.constant 1 : index
+ %c13 = arith.constant 13 : index
+ %dim = tensor.dim %arg0, %c0 : tensor<?xf32>
+ %a = affine.apply affine_map<()[s0, s1, s2] -> (s0 floordiv (s1 + (-s1 + 2) * (-s1 + s1 * s2 + 1)))>()[%c13, %dim, %c1]
+ %b = affine.apply affine_map<()[s0, s1, s2] -> (s0 ceildiv (s1 + (-s1 + 2) * (-s1 + s1 * s2 + 1)))>()[%c13, %dim, %c1]
+ // CHECK: %[[C6:.*]] = arith.constant 6 : index
+ // CHECK-NEXT: %[[C7:.*]] = arith.constant 7 : index
+ // CHECK-NEXT: return %[[C6]], %[[C7]]
+ return %a, %b : index, index
+}
|
Contributor
|
The fix is not correct. It simply hides the original bug. Local expression is always supposed to be a binary expression -- it cannot be dimension, symbol or constant. This PR fixes this. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
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.
This PR fixes a crash in
getSemiAffineExprFromFlatFormwhen localExpr is notAffineBinaryOpExpr. Fixes #144091.