-
Notifications
You must be signed in to change notification settings - Fork 14.3k
[mlir] [mem2reg] Fix Mem2Reg attempting to promote in graph regions #104910
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.
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
Conversation
@llvm/pr-subscribers-mlir Author: Théo Degioanni (Moxinilian) ChangesMem2Reg assumes SSA dependencies but did not check for graph regions. This fixes it. Full diff: https://github.com/llvm/llvm-project/pull/104910.diff 4 Files Affected:
diff --git a/mlir/lib/Transforms/Mem2Reg.cpp b/mlir/lib/Transforms/Mem2Reg.cpp
index 1f6998709ae02e..406787e2f591ca 100644
--- a/mlir/lib/Transforms/Mem2Reg.cpp
+++ b/mlir/lib/Transforms/Mem2Reg.cpp
@@ -13,6 +13,7 @@
#include "mlir/IR/Builders.h"
#include "mlir/IR/Dominance.h"
#include "mlir/IR/PatternMatch.h"
+#include "mlir/IR/RegionKindInterface.h"
#include "mlir/IR/Value.h"
#include "mlir/Interfaces/ControlFlowInterfaces.h"
#include "mlir/Interfaces/MemorySlotInterfaces.h"
@@ -255,6 +256,18 @@ LogicalResult MemorySlotPromotionAnalyzer::computeBlockingUses(
// delete itself). We thus need to start from the use of the slot pointer and
// propagate further requests through the forward slice.
+ // Because this pass currently only supports analysing the parent region of
+ // the slot pointer, if a promotable memory op that needs promotion is within
+ // a graph region, the slot may only be used in a graph region and should
+ // therefore be ignored.
+ Region *slotPtrRegion = slot.ptr.getParentRegion();
+ auto slotPtrRegionOp =
+ llvm::dyn_cast<RegionKindInterface>(slotPtrRegion->getParentOp());
+ if (slotPtrRegionOp &&
+ slotPtrRegionOp.getRegionKind(slotPtrRegion->getRegionNumber()) ==
+ RegionKind::Graph)
+ return failure();
+
// First insert that all immediate users of the slot pointer must no longer
// use it.
for (OpOperand &use : slot.ptr.getUses()) {
diff --git a/mlir/test/Transforms/mem2reg.mlir b/mlir/test/Transforms/mem2reg.mlir
index daeaa2da076341..db11e2d83249a7 100644
--- a/mlir/test/Transforms/mem2reg.mlir
+++ b/mlir/test/Transforms/mem2reg.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s --pass-pipeline='builtin.module(func.func(mem2reg))' --split-input-file | FileCheck %s
+// RUN: mlir-opt %s --pass-pipeline='builtin.module(func.func(mem2reg),test.isolated_graph_region(mem2reg))' --split-input-file | FileCheck %s
// Verifies that allocators with mutliple slots are handled properly.
@@ -26,3 +26,16 @@ func.func @multi_slot_alloca_only_second() -> (i32, i32) {
%4 = memref.load %2[] : memref<i32>
return %3, %4 : i32, i32
}
+
+// -----
+
+// Checks that slots are not promoted if used in a graph region.
+
+// CHECK-LABEL: test.isolated_graph_region
+test.isolated_graph_region {
+ // CHECK: %{{[[:alnum:]]+}} = test.multi_slot_alloca
+ %slot = test.multi_slot_alloca : () -> (memref<i32>)
+ memref.store %a, %slot[] : memref<i32>
+ %a = memref.load %slot[] : memref<i32>
+ "test.foo"() : () -> ()
+}
diff --git a/mlir/test/lib/Dialect/Test/TestOpDefs.cpp b/mlir/test/lib/Dialect/Test/TestOpDefs.cpp
index fbaa102d3e33cc..69091fb893fad6 100644
--- a/mlir/test/lib/Dialect/Test/TestOpDefs.cpp
+++ b/mlir/test/lib/Dialect/Test/TestOpDefs.cpp
@@ -126,6 +126,14 @@ RegionKind GraphRegionOp::getRegionKind(unsigned index) {
return RegionKind::Graph;
}
+//===----------------------------------------------------------------------===//
+// IsolatedGraphRegionOp
+//===----------------------------------------------------------------------===//
+
+RegionKind IsolatedGraphRegionOp::getRegionKind(unsigned index) {
+ return RegionKind::Graph;
+}
+
//===----------------------------------------------------------------------===//
// AffineScopeOp
//===----------------------------------------------------------------------===//
diff --git a/mlir/test/lib/Dialect/Test/TestOps.td b/mlir/test/lib/Dialect/Test/TestOps.td
index 2b55bff3538d39..9e19966414d1d7 100644
--- a/mlir/test/lib/Dialect/Test/TestOps.td
+++ b/mlir/test/lib/Dialect/Test/TestOps.td
@@ -2048,6 +2048,18 @@ def GraphRegionOp : TEST_Op<"graph_region", [
let assemblyFormat = "attr-dict-with-keyword $region";
}
+def IsolatedGraphRegionOp : TEST_Op<"isolated_graph_region", [
+ DeclareOpInterfaceMethods<RegionKindInterface>,
+ IsolatedFromAbove]> {
+ let summary = "isolated from above operation with a graph region";
+ let description = [{
+ Test op that defines a graph region which is isolated from above.
+ }];
+
+ let regions = (region AnyRegion:$region);
+ let assemblyFormat = "attr-dict-with-keyword $region";
+}
+
def AffineScopeOp : TEST_Op<"affine_scope", [AffineScope]> {
let summary = "affine scope operation";
let description = [{
|
@llvm/pr-subscribers-mlir-core Author: Théo Degioanni (Moxinilian) ChangesMem2Reg assumes SSA dependencies but did not check for graph regions. This fixes it. Full diff: https://github.com/llvm/llvm-project/pull/104910.diff 4 Files Affected:
diff --git a/mlir/lib/Transforms/Mem2Reg.cpp b/mlir/lib/Transforms/Mem2Reg.cpp
index 1f6998709ae02e..406787e2f591ca 100644
--- a/mlir/lib/Transforms/Mem2Reg.cpp
+++ b/mlir/lib/Transforms/Mem2Reg.cpp
@@ -13,6 +13,7 @@
#include "mlir/IR/Builders.h"
#include "mlir/IR/Dominance.h"
#include "mlir/IR/PatternMatch.h"
+#include "mlir/IR/RegionKindInterface.h"
#include "mlir/IR/Value.h"
#include "mlir/Interfaces/ControlFlowInterfaces.h"
#include "mlir/Interfaces/MemorySlotInterfaces.h"
@@ -255,6 +256,18 @@ LogicalResult MemorySlotPromotionAnalyzer::computeBlockingUses(
// delete itself). We thus need to start from the use of the slot pointer and
// propagate further requests through the forward slice.
+ // Because this pass currently only supports analysing the parent region of
+ // the slot pointer, if a promotable memory op that needs promotion is within
+ // a graph region, the slot may only be used in a graph region and should
+ // therefore be ignored.
+ Region *slotPtrRegion = slot.ptr.getParentRegion();
+ auto slotPtrRegionOp =
+ llvm::dyn_cast<RegionKindInterface>(slotPtrRegion->getParentOp());
+ if (slotPtrRegionOp &&
+ slotPtrRegionOp.getRegionKind(slotPtrRegion->getRegionNumber()) ==
+ RegionKind::Graph)
+ return failure();
+
// First insert that all immediate users of the slot pointer must no longer
// use it.
for (OpOperand &use : slot.ptr.getUses()) {
diff --git a/mlir/test/Transforms/mem2reg.mlir b/mlir/test/Transforms/mem2reg.mlir
index daeaa2da076341..db11e2d83249a7 100644
--- a/mlir/test/Transforms/mem2reg.mlir
+++ b/mlir/test/Transforms/mem2reg.mlir
@@ -1,4 +1,4 @@
-// RUN: mlir-opt %s --pass-pipeline='builtin.module(func.func(mem2reg))' --split-input-file | FileCheck %s
+// RUN: mlir-opt %s --pass-pipeline='builtin.module(func.func(mem2reg),test.isolated_graph_region(mem2reg))' --split-input-file | FileCheck %s
// Verifies that allocators with mutliple slots are handled properly.
@@ -26,3 +26,16 @@ func.func @multi_slot_alloca_only_second() -> (i32, i32) {
%4 = memref.load %2[] : memref<i32>
return %3, %4 : i32, i32
}
+
+// -----
+
+// Checks that slots are not promoted if used in a graph region.
+
+// CHECK-LABEL: test.isolated_graph_region
+test.isolated_graph_region {
+ // CHECK: %{{[[:alnum:]]+}} = test.multi_slot_alloca
+ %slot = test.multi_slot_alloca : () -> (memref<i32>)
+ memref.store %a, %slot[] : memref<i32>
+ %a = memref.load %slot[] : memref<i32>
+ "test.foo"() : () -> ()
+}
diff --git a/mlir/test/lib/Dialect/Test/TestOpDefs.cpp b/mlir/test/lib/Dialect/Test/TestOpDefs.cpp
index fbaa102d3e33cc..69091fb893fad6 100644
--- a/mlir/test/lib/Dialect/Test/TestOpDefs.cpp
+++ b/mlir/test/lib/Dialect/Test/TestOpDefs.cpp
@@ -126,6 +126,14 @@ RegionKind GraphRegionOp::getRegionKind(unsigned index) {
return RegionKind::Graph;
}
+//===----------------------------------------------------------------------===//
+// IsolatedGraphRegionOp
+//===----------------------------------------------------------------------===//
+
+RegionKind IsolatedGraphRegionOp::getRegionKind(unsigned index) {
+ return RegionKind::Graph;
+}
+
//===----------------------------------------------------------------------===//
// AffineScopeOp
//===----------------------------------------------------------------------===//
diff --git a/mlir/test/lib/Dialect/Test/TestOps.td b/mlir/test/lib/Dialect/Test/TestOps.td
index 2b55bff3538d39..9e19966414d1d7 100644
--- a/mlir/test/lib/Dialect/Test/TestOps.td
+++ b/mlir/test/lib/Dialect/Test/TestOps.td
@@ -2048,6 +2048,18 @@ def GraphRegionOp : TEST_Op<"graph_region", [
let assemblyFormat = "attr-dict-with-keyword $region";
}
+def IsolatedGraphRegionOp : TEST_Op<"isolated_graph_region", [
+ DeclareOpInterfaceMethods<RegionKindInterface>,
+ IsolatedFromAbove]> {
+ let summary = "isolated from above operation with a graph region";
+ let description = [{
+ Test op that defines a graph region which is isolated from above.
+ }];
+
+ let regions = (region AnyRegion:$region);
+ let assemblyFormat = "attr-dict-with-keyword $region";
+}
+
def AffineScopeOp : TEST_Op<"affine_scope", [AffineScope]> {
let summary = "affine scope operation";
let description = [{
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM % a nits. Thanks for the improvement.
Co-authored-by: Christian Ulmann <[email protected]>
Co-authored-by: Christian Ulmann <[email protected]>
✅ With the latest revision this PR passed the C/C++ code formatter. |
…lvm#104910) Mem2Reg assumes SSA dependencies but did not check for graph regions. This fixes it. --------- Co-authored-by: Christian Ulmann <[email protected]>
Mem2Reg assumes SSA dependencies but did not check for graph regions. This fixes it.