[clang] Add nuw attribute to GEPs #105496
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
Add nuw attribute to inbounds GEPs where the expression used to form the GEP is an addition of unsigned indices.
llvmbot
added
clang
|
@llvm/pr-subscribers-backend-powerpc @llvm/pr-subscribers-hlsl Author: Hari Limaye (hazzlim) ChangesAdd nuw attribute to inbounds GEPs where the expression used to form the GEP is an addition of unsigned indices. Patch is 576.17 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/105496.diff 91 Files Affected:
diff --git a/clang/lib/CodeGen/CGBuilder.h b/clang/lib/CodeGen/CGBuilder.h
index 08730a6a6672a1..b8036cf6e6a306 100644
--- a/clang/lib/CodeGen/CGBuilder.h
+++ b/clang/lib/CodeGen/CGBuilder.h
@@ -14,6 +14,7 @@
#include "CodeGenTypeCache.h"
#include "llvm/Analysis/Utils/Local.h"
#include "llvm/IR/DataLayout.h"
+#include "llvm/IR/GEPNoWrapFlags.h"
#include "llvm/IR/IRBuilder.h"
#include "llvm/IR/Type.h"
@@ -334,9 +335,10 @@ class CGBuilderTy : public CGBuilderBaseTy {
Address CreateGEP(Address Addr, ArrayRef<llvm::Value *> IdxList,
llvm::Type *ElementType, CharUnits Align,
- const Twine &Name = "") {
+ const Twine &Name = "",
+ llvm::GEPNoWrapFlags NW = llvm::GEPNoWrapFlags::none()) {
llvm::Value *Ptr = emitRawPointerFromAddress(Addr);
- return RawAddress(CreateGEP(Addr.getElementType(), Ptr, IdxList, Name),
+ return RawAddress(CreateGEP(Addr.getElementType(), Ptr, IdxList, Name, NW),
ElementType, Align);
}
diff --git a/clang/lib/CodeGen/CGExprScalar.cpp b/clang/lib/CodeGen/CGExprScalar.cpp
index 3bda254c86adf6..41a91aa435f999 100644
--- a/clang/lib/CodeGen/CGExprScalar.cpp
+++ b/clang/lib/CodeGen/CGExprScalar.cpp
@@ -36,6 +36,7 @@
#include "llvm/IR/DerivedTypes.h"
#include "llvm/IR/FixedPointBuilder.h"
#include "llvm/IR/Function.h"
+#include "llvm/IR/GEPNoWrapFlags.h"
#include "llvm/IR/GetElementPtrTypeIterator.h"
#include "llvm/IR/GlobalVariable.h"
#include "llvm/IR/Intrinsics.h"
@@ -5756,7 +5757,12 @@ CodeGenFunction::EmitCheckedInBoundsGEP(llvm::Type *ElemTy, Value *Ptr,
bool SignedIndices, bool IsSubtraction,
SourceLocation Loc, const Twine &Name) {
llvm::Type *PtrTy = Ptr->getType();
- Value *GEPVal = Builder.CreateInBoundsGEP(ElemTy, Ptr, IdxList, Name);
+
+ llvm::GEPNoWrapFlags NWFlags = llvm::GEPNoWrapFlags::inBounds();
+ if (!SignedIndices && !IsSubtraction)
+ NWFlags |= llvm::GEPNoWrapFlags::noUnsignedWrap();
+
+ Value *GEPVal = Builder.CreateGEP(ElemTy, Ptr, IdxList, Name, NWFlags);
// If the pointer overflow sanitizer isn't enabled, do nothing.
if (!SanOpts.has(SanitizerKind::PointerOverflow))
@@ -5871,8 +5877,13 @@ Address CodeGenFunction::EmitCheckedInBoundsGEP(
Address Addr, ArrayRef<Value *> IdxList, llvm::Type *elementType,
bool SignedIndices, bool IsSubtraction, SourceLocation Loc, CharUnits Align,
const Twine &Name) {
- if (!SanOpts.has(SanitizerKind::PointerOverflow))
- return Builder.CreateInBoundsGEP(Addr, IdxList, elementType, Align, Name);
+ if (!SanOpts.has(SanitizerKind::PointerOverflow)) {
+ llvm::GEPNoWrapFlags NWFlags = llvm::GEPNoWrapFlags::inBounds();
+ if (!SignedIndices && !IsSubtraction)
+ NWFlags |= llvm::GEPNoWrapFlags::noUnsignedWrap();
+
+ return Builder.CreateGEP(Addr, IdxList, elementType, Align, Name, NWFlags);
+ }
return RawAddress(
EmitCheckedInBoundsGEP(Addr.getElementType(), Addr.emitRawPointer(*this),
diff --git a/clang/test/CodeGen/2005-01-02-ConstantInits.c b/clang/test/CodeGen/2005-01-02-ConstantInits.c
index 7772a64331ffb7..1878c00fa89914 100644
--- a/clang/test/CodeGen/2005-01-02-ConstantInits.c
+++ b/clang/test/CodeGen/2005-01-02-ConstantInits.c
@@ -7,13 +7,6 @@
struct X { int a[2]; };
extern int bar();
-//.
-// CHECK: @test.i23 = internal global i32 4, align 4
-// CHECK: @i = global i32 4, align 4
-// CHECK: @Arr = global [100 x i32] zeroinitializer, align 16
-// CHECK: @foo2.X = internal global ptr getelementptr (i8, ptr @Arr, i64 196), align 8
-// CHECK: @foo2.i23 = internal global i32 0, align 4
-//.
// CHECK-LABEL: define {{[^@]+}}@test
// CHECK-SAME: () #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: entry:
@@ -51,7 +44,7 @@ int foo(int i) { return bar(&Arr[49])+bar(&Arr[i]); }
// CHECK-NEXT: store i32 [[I]], ptr [[I_ADDR]], align 4
// CHECK-NEXT: store ptr @Arr, ptr [[P]], align 8
// CHECK-NEXT: [[TMP0:%.*]] = load ptr, ptr [[P]], align 8
-// CHECK-NEXT: [[INCDEC_PTR:%.*]] = getelementptr inbounds i32, ptr [[TMP0]], i32 1
+// CHECK-NEXT: [[INCDEC_PTR:%.*]] = getelementptr inbounds nuw i32, ptr [[TMP0]], i32 1
// CHECK-NEXT: store ptr [[INCDEC_PTR]], ptr [[P]], align 8
// CHECK-NEXT: [[TMP1:%.*]] = load i32, ptr [[I_ADDR]], align 4
// CHECK-NEXT: [[IDX_EXT:%.*]] = sext i32 [[TMP1]] to i64
diff --git a/clang/test/CodeGen/PowerPC/ppc-emmintrin.c b/clang/test/CodeGen/PowerPC/ppc-emmintrin.c
index a3650beec625f2..4c4d0dfce05eaf 100644
--- a/clang/test/CodeGen/PowerPC/ppc-emmintrin.c
+++ b/clang/test/CodeGen/PowerPC/ppc-emmintrin.c
@@ -1012,14 +1012,14 @@ test_shuffle() {
// CHECK: %[[SHR:[0-9a-zA-Z_.]+]] = ashr i32 %{{[0-9a-zA-Z_.]+}}, 6
// CHECK: %[[AND4:[0-9a-zA-Z_.]+]] = and i32 %[[SHR]], 3
// CHECK: sext i32 %[[AND4]] to i64
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %{{[0-9a-zA-Z_.]+}}, i32 0
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %{{[0-9a-zA-Z_.]+}}, i32 1
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK: %[[ADD:[0-9a-zA-Z_.]+]] = add i32 %{{[0-9a-zA-Z_.]+}}, 269488144
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %[[ADD]], i32 2
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK: add i32 %{{[0-9a-zA-Z_.]+}}, 269488144
// CHECK: call <4 x i32> @vec_perm(int vector[4], int vector[4], unsigned char vector[16])
@@ -1050,7 +1050,7 @@ test_shuffle() {
// CHECK: sext i32 %[[AND4]] to i64
// CHECK-LE: store <2 x i64> <i64 1663540288323457296, i64 0>, ptr %{{[0-9a-zA-Z_.]+}}, align 16
// CHECK-BE: store <2 x i64> <i64 1157726452361532951, i64 0>, ptr %{{[0-9a-zA-Z_.]+}}, align 16
-// CHECK-COUNT-4: getelementptr inbounds [4 x i16], ptr @_mm_shufflehi_epi16.__permute_selectors, i64 0, i64 {{[0-9a-zA-Z_%.]+}}
+// CHECK-COUNT-4: getelementptr inbounds nuw [4 x i16], ptr @_mm_shufflehi_epi16.__permute_selectors, i64 0, i64 {{[0-9a-zA-Z_%.]+}}
// CHECK: call <2 x i64> @vec_perm(unsigned long long vector[2], unsigned long long vector[2], unsigned char vector[16])
// CHECK-LABEL: define available_externally <2 x i64> @_mm_shufflelo_epi16
@@ -1067,7 +1067,7 @@ test_shuffle() {
// CHECK: sext i32 %[[AND4]] to i64
// CHECK-LE: store <2 x i64> <i64 0, i64 2242261671028070680>, ptr %{{[0-9a-zA-Z_.]+}}, align 16
// CHECK-BE: store <2 x i64> <i64 0, i64 1736447835066146335>, ptr %{{[0-9a-zA-Z_.]+}}, align 16
-// CHECK-COUNT-4: getelementptr inbounds [4 x i16], ptr @_mm_shufflelo_epi16.__permute_selectors, i64 0, i64 {{[0-9a-zA-Z_%.]+}}
+// CHECK-COUNT-4: getelementptr inbounds nuw [4 x i16], ptr @_mm_shufflelo_epi16.__permute_selectors, i64 0, i64 {{[0-9a-zA-Z_%.]+}}
// CHECK: call <2 x i64> @vec_perm(unsigned long long vector[2], unsigned long long vector[2], unsigned char vector[16])
void __attribute__((noinline))
diff --git a/clang/test/CodeGen/PowerPC/ppc-xmmintrin.c b/clang/test/CodeGen/PowerPC/ppc-xmmintrin.c
index 95dfd1202f1575..4a15fa9f76ceea 100644
--- a/clang/test/CodeGen/PowerPC/ppc-xmmintrin.c
+++ b/clang/test/CodeGen/PowerPC/ppc-xmmintrin.c
@@ -894,16 +894,16 @@ test_shuffle() {
// CHECK: %[[SHR3:[0-9a-zA-Z_.]+]] = ashr i32 %{{[0-9a-zA-Z_.]+}}, 6
// CHECK: %[[AND4:[0-9a-zA-Z_.]+]] = and i32 %[[SHR3]], 3
// CHECK: sext i32 %[[AND4]] to i64
-// CHECK: getelementptr inbounds [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK-LE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 0
// CHECK-BE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 3
-// CHECK: getelementptr inbounds [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK-LE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 1
// CHECK-BE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 2
-// CHECK: getelementptr inbounds [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK-LE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 2
// CHECK-BE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 1
-// CHECK: getelementptr inbounds [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK-LE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 3
// CHECK-BE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 0
// CHECK: call <2 x i64> @vec_splats(unsigned long long)
@@ -923,14 +923,14 @@ test_shuffle() {
// CHECK: %[[SHR3:[0-9a-zA-Z_.]+]] = ashr i32 %{{[0-9a-zA-Z_.]+}}, 6
// CHECK: %[[AND4:[0-9a-zA-Z_.]+]] = and i32 %[[SHR3]], 3
// CHECK: sext i32 %[[AND4]] to i64
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %{{[0-9a-zA-Z_.]+}}, i32 0
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %{{[0-9a-zA-Z_.]+}}, i32 1
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
// CHECK: %[[ADD:[0-9a-zA-Z_.]+]] = add i32 %{{[0-9a-zA-Z_.]+}}, 269488144
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %[[ADD]], i32 2
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
// CHECK: %[[ADD2:[0-9a-zA-Z_.]+]] = add i32 %{{[0-9a-zA-Z_.]+}}, 269488144
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %[[ADD2]], i32 3
// CHECK: call <4 x float> @vec_perm(float vector[4], float vector[4], unsigned char vector[16])
diff --git a/clang/test/CodeGen/attr-counted-by.c b/clang/test/CodeGen/attr-counted-by.c
index 9acc896c0f0e9d..a1182cdd5ceef9 100644
--- a/clang/test/CodeGen/attr-counted-by.c
+++ b/clang/test/CodeGen/attr-counted-by.c
@@ -118,7 +118,7 @@ void test1(struct annotated *p, int index, int val) {
// SANITIZE-WITH-ATTR-NEXT: unreachable, !nosanitize [[META2]]
// SANITIZE-WITH-ATTR: cont3:
// SANITIZE-WITH-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// SANITIZE-WITH-ATTR-NEXT: [[TMP2:%.*]] = shl i32 [[DOT_COUNTED_BY_LOAD]], 2
// SANITIZE-WITH-ATTR-NEXT: [[DOTINV:%.*]] = icmp slt i32 [[DOT_COUNTED_BY_LOAD]], 0
// SANITIZE-WITH-ATTR-NEXT: [[CONV:%.*]] = select i1 [[DOTINV]], i32 0, i32 [[TMP2]]
@@ -134,7 +134,7 @@ void test1(struct annotated *p, int index, int val) {
// NO-SANITIZE-WITH-ATTR-NEXT: [[DOTINV:%.*]] = icmp slt i32 [[DOT_COUNTED_BY_LOAD]], 0
// NO-SANITIZE-WITH-ATTR-NEXT: [[CONV:%.*]] = select i1 [[DOTINV]], i32 0, i32 [[TMP0]]
// NO-SANITIZE-WITH-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// NO-SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// NO-SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// NO-SANITIZE-WITH-ATTR-NEXT: store i32 [[CONV]], ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA2]]
// NO-SANITIZE-WITH-ATTR-NEXT: ret void
//
@@ -142,7 +142,7 @@ void test1(struct annotated *p, int index, int val) {
// SANITIZE-WITHOUT-ATTR-SAME: ptr noundef [[P:%.*]], i64 noundef [[INDEX:%.*]]) local_unnamed_addr #[[ATTR0]] {
// SANITIZE-WITHOUT-ATTR-NEXT: entry:
// SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// SANITIZE-WITHOUT-ATTR-NEXT: store i32 -1, ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA2]]
// SANITIZE-WITHOUT-ATTR-NEXT: ret void
//
@@ -150,7 +150,7 @@ void test1(struct annotated *p, int index, int val) {
// NO-SANITIZE-WITHOUT-ATTR-SAME: ptr noundef [[P:%.*]], i64 noundef [[INDEX:%.*]]) local_unnamed_addr #[[ATTR0]] {
// NO-SANITIZE-WITHOUT-ATTR-NEXT: entry:
// NO-SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// NO-SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// NO-SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// NO-SANITIZE-WITHOUT-ATTR-NEXT: store i32 -1, ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA2]]
// NO-SANITIZE-WITHOUT-ATTR-NEXT: ret void
//
@@ -207,7 +207,7 @@ size_t test2_bdos(struct annotated *p) {
// SANITIZE-WITH-ATTR-NEXT: unreachable, !nosanitize [[META2]]
// SANITIZE-WITH-ATTR: cont3:
// SANITIZE-WITH-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// SANITIZE-WITH-ATTR-NEXT: [[TMP2:%.*]] = sext i32 [[DOT_COUNTED_BY_LOAD]] to i64
// SANITIZE-WITH-ATTR-NEXT: [[TMP3:%.*]] = shl nsw i64 [[TMP2]], 2
// SANITIZE-WITH-ATTR-NEXT: [[TMP4:%.*]] = tail call i64 @llvm.smax.i64(i64 [[TMP3]], i64 4)
@@ -231,7 +231,7 @@ size_t test2_bdos(struct annotated *p) {
// NO-SANITIZE-WITH-ATTR-NEXT: [[DOTINV:%.*]] = icmp slt i32 [[DOT_COUNTED_BY_LOAD]], 0
// NO-SANITIZE-WITH-ATTR-NEXT: [[CONV:%.*]] = select i1 [[DOTINV]], i32 0, i32 [[TMP4]]
// NO-SANITIZE-WITH-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// NO-SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// NO-SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// NO-SANITIZE-WITH-ATTR-NEXT: store i32 [[CONV]], ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA2]]
// NO-SANITIZE-WITH-ATTR-NEXT: ret void
//
@@ -239,7 +239,7 @@ size_t test2_bdos(struct annotated *p) {
// SANITIZE-WITHOUT-ATTR-SAME: ptr noundef [[P:%.*]], i64 noundef [[INDEX:%.*]]) local_unnamed_addr #[[ATTR0]] {
// SANITIZE-WITHOUT-ATTR-NEXT: entry:
// SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// SANITIZE-WITHOUT-ATTR-NEXT: store i32 -1, ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA2]]
// SANITIZE-WITHOUT-ATTR-NEXT: ret void
//
@@ -247,7 +247,7 @@ size_t test2_bdos(struct annotated *p) {
// NO-SANITIZE-WITHOUT-ATTR-SAME: ptr noundef [[P:%.*]], i64 noundef [[INDEX:%.*]]) local_unnamed_addr #[[ATTR0]] {
// NO-SANITIZE-WITHOUT-ATTR-NEXT: entry:
// NO-SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// NO-SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// NO-SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// NO-SANITIZE-WITHOUT-ATTR-NEXT: store i32 -1, ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA2]]
// NO-SANITIZE-WITHOUT-ATTR-NEXT: ret void
//
diff --git a/clang/test/CodeGen/catch-nullptr-and-nonzero-offset-when-nullptr-is-defined.c b/clang/test/CodeGen/catch-nullptr-and-nonzero-offset-when-nullptr-is-defined.c
index 39ede01d6e3b83..8a560a47ad1e10 100644
--- a/clang/test/CodeGen/catch-nullptr-and-nonzero-offset-when-nullptr-is-defined.c
+++ b/clang/test/CodeGen/catch-nullptr-and-nonzero-offset-when-nullptr-is-defined.c
@@ -33,7 +33,7 @@ char *add_unsigned(char *base, unsigned long offset) {
// CHECK-NEXT: store i64 %[[OFFSET]], ptr %[[OFFSET_ADDR]], align 8
// CHECK-NEXT: %[[BASE_RELOADED:.*]] = load ptr, ptr %[[BASE_ADDR]], align 8
// CHECK-NEXT: %[[OFFSET_RELOADED:.*]] = load i64, ptr %[[OFFSET_ADDR]], align 8
- // CHECK-NEXT: %[[ADD_PTR:.*]] = getelementptr inbounds i8, ptr %[[BASE_RELOADED]], i64 %[[OFFSET_RELOADED]]
+ // CHECK-NEXT: %[[ADD_PTR:.*]] = getelementptr inbounds nuw i8, ptr %[[BASE_RELOADED]], i64 %[[OFFSET_RELOADED]]
// CHECK-SANITIZE-NEXT: %[[COMPUTED_OFFSET_AGGREGATE:.*]] = call { i64, i1 } @llvm.smul.with.overflow.i64(i64 1, i64 %[[OFFSET_RELOADED]]), !nosanitize
// CHECK-SANITIZE-NEXT: %[[COMPUTED_OFFSET_OVERFLOWED:.*]] = extractvalue { i64, i1 } %[[COMPUTED_OFFSET_AGGREGATE]], 1, !nosanitize
// CHECK-SANITIZE-NEXT: %[[OR_OV:.+]] = or i1 %[[COMPUTED_OFFSET_OVERFLOWED]], false, !nosanitize
diff --git a/clang/test/CodeGen/catch-nullptr-and-nonzero-offset.c b/clang/test/CodeGen/catch-nullptr-and-nonzero-offset.c
index e93dbcb9f647bf..d884993ffb2b30 100644
--- a/clang/test/CodeGen/catch-nullptr-and-nonzero-offset.c
+++ b/clang/test/CodeGen/catch-nullptr-and-nonzero-offset.c
@@ -50,7 +50,7 @@ char *var_var(char *base, unsigned long offset) {
// CHECK-NEXT: store i64 %[[OFFSET]], ptr %[[OFFSET_ADDR]], align 8
// CHECK-NEXT: %[[BASE_RELOADED:.*]] = load ptr, ptr %[[BASE_ADDR]], align 8
// CHECK-NEXT: %[[OFFSET_RELOADED:.*]] = load i64, ptr %[[OFFSET_ADDR]], align 8
- // CHECK-NEXT: %[[ADD_PTR:.*]] = getelementptr inbounds i8, ptr %[[BASE_RELOADED]], i64 %[[OFFSET_RELOADED]]
+ // CHECK-NEXT: %[[ADD_PTR:.*]] = getelementptr inbounds nuw i8, ptr %[[BASE_RELOADED]], i64 %[[OFFSET_RELOADED]]
// CHECK-SANITIZE-NEXT: %[[COMPUTED_OFFSET_AGGREGATE:.*]] = call { i64, i1 } @llvm.smul.with.overflow.i64(i64 1, i64 %[[OFFSET_RELOADED]]), !nosanitize
// CHECK-SANITIZE-NEXT: %[[COMPUTED_OFFSET_OVERFLOWED:.*]] = extractvalue { i64, i1 } %[[COMPUTED_OFFSET_AGGREGATE]], 1, !nosanitize
// CHECK-SANITIZE-NEXT: %[[OR_OV:.+]] = or i1 %[[COMPUTED_OFFSET_OVERFLOWED]], false, !nosanitize
@...
[truncated]
|
|
@llvm/pr-subscribers-clang Author: Hari Limaye (hazzlim) ChangesAdd nuw attribute to inbounds GEPs where the expression used to form the GEP is an addition of unsigned indices. Patch is 576.17 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/105496.diff 91 Files Affected:
diff --git a/clang/lib/CodeGen/CGBuilder.h b/clang/lib/CodeGen/CGBuilder.h
index 08730a6a6672a1..b8036cf6e6a306 100644
--- a/clang/lib/CodeGen/CGBuilder.h
+++ b/clang/lib/CodeGen/CGBuilder.h
@@ -14,6 +14,7 @@
#include "CodeGenTypeCache.h"
#include "llvm/Analysis/Utils/Local.h"
#include "llvm/IR/DataLayout.h"
+#include "llvm/IR/GEPNoWrapFlags.h"
#include "llvm/IR/IRBuilder.h"
#include "llvm/IR/Type.h"
@@ -334,9 +335,10 @@ class CGBuilderTy : public CGBuilderBaseTy {
Address CreateGEP(Address Addr, ArrayRef<llvm::Value *> IdxList,
llvm::Type *ElementType, CharUnits Align,
- const Twine &Name = "") {
+ const Twine &Name = "",
+ llvm::GEPNoWrapFlags NW = llvm::GEPNoWrapFlags::none()) {
llvm::Value *Ptr = emitRawPointerFromAddress(Addr);
- return RawAddress(CreateGEP(Addr.getElementType(), Ptr, IdxList, Name),
+ return RawAddress(CreateGEP(Addr.getElementType(), Ptr, IdxList, Name, NW),
ElementType, Align);
}
diff --git a/clang/lib/CodeGen/CGExprScalar.cpp b/clang/lib/CodeGen/CGExprScalar.cpp
index 3bda254c86adf6..41a91aa435f999 100644
--- a/clang/lib/CodeGen/CGExprScalar.cpp
+++ b/clang/lib/CodeGen/CGExprScalar.cpp
@@ -36,6 +36,7 @@
#include "llvm/IR/DerivedTypes.h"
#include "llvm/IR/FixedPointBuilder.h"
#include "llvm/IR/Function.h"
+#include "llvm/IR/GEPNoWrapFlags.h"
#include "llvm/IR/GetElementPtrTypeIterator.h"
#include "llvm/IR/GlobalVariable.h"
#include "llvm/IR/Intrinsics.h"
@@ -5756,7 +5757,12 @@ CodeGenFunction::EmitCheckedInBoundsGEP(llvm::Type *ElemTy, Value *Ptr,
bool SignedIndices, bool IsSubtraction,
SourceLocation Loc, const Twine &Name) {
llvm::Type *PtrTy = Ptr->getType();
- Value *GEPVal = Builder.CreateInBoundsGEP(ElemTy, Ptr, IdxList, Name);
+
+ llvm::GEPNoWrapFlags NWFlags = llvm::GEPNoWrapFlags::inBounds();
+ if (!SignedIndices && !IsSubtraction)
+ NWFlags |= llvm::GEPNoWrapFlags::noUnsignedWrap();
+
+ Value *GEPVal = Builder.CreateGEP(ElemTy, Ptr, IdxList, Name, NWFlags);
// If the pointer overflow sanitizer isn't enabled, do nothing.
if (!SanOpts.has(SanitizerKind::PointerOverflow))
@@ -5871,8 +5877,13 @@ Address CodeGenFunction::EmitCheckedInBoundsGEP(
Address Addr, ArrayRef<Value *> IdxList, llvm::Type *elementType,
bool SignedIndices, bool IsSubtraction, SourceLocation Loc, CharUnits Align,
const Twine &Name) {
- if (!SanOpts.has(SanitizerKind::PointerOverflow))
- return Builder.CreateInBoundsGEP(Addr, IdxList, elementType, Align, Name);
+ if (!SanOpts.has(SanitizerKind::PointerOverflow)) {
+ llvm::GEPNoWrapFlags NWFlags = llvm::GEPNoWrapFlags::inBounds();
+ if (!SignedIndices && !IsSubtraction)
+ NWFlags |= llvm::GEPNoWrapFlags::noUnsignedWrap();
+
+ return Builder.CreateGEP(Addr, IdxList, elementType, Align, Name, NWFlags);
+ }
return RawAddress(
EmitCheckedInBoundsGEP(Addr.getElementType(), Addr.emitRawPointer(*this),
diff --git a/clang/test/CodeGen/2005-01-02-ConstantInits.c b/clang/test/CodeGen/2005-01-02-ConstantInits.c
index 7772a64331ffb7..1878c00fa89914 100644
--- a/clang/test/CodeGen/2005-01-02-ConstantInits.c
+++ b/clang/test/CodeGen/2005-01-02-ConstantInits.c
@@ -7,13 +7,6 @@
struct X { int a[2]; };
extern int bar();
-//.
-// CHECK: @test.i23 = internal global i32 4, align 4
-// CHECK: @i = global i32 4, align 4
-// CHECK: @Arr = global [100 x i32] zeroinitializer, align 16
-// CHECK: @foo2.X = internal global ptr getelementptr (i8, ptr @Arr, i64 196), align 8
-// CHECK: @foo2.i23 = internal global i32 0, align 4
-//.
// CHECK-LABEL: define {{[^@]+}}@test
// CHECK-SAME: () #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: entry:
@@ -51,7 +44,7 @@ int foo(int i) { return bar(&Arr[49])+bar(&Arr[i]); }
// CHECK-NEXT: store i32 [[I]], ptr [[I_ADDR]], align 4
// CHECK-NEXT: store ptr @Arr, ptr [[P]], align 8
// CHECK-NEXT: [[TMP0:%.*]] = load ptr, ptr [[P]], align 8
-// CHECK-NEXT: [[INCDEC_PTR:%.*]] = getelementptr inbounds i32, ptr [[TMP0]], i32 1
+// CHECK-NEXT: [[INCDEC_PTR:%.*]] = getelementptr inbounds nuw i32, ptr [[TMP0]], i32 1
// CHECK-NEXT: store ptr [[INCDEC_PTR]], ptr [[P]], align 8
// CHECK-NEXT: [[TMP1:%.*]] = load i32, ptr [[I_ADDR]], align 4
// CHECK-NEXT: [[IDX_EXT:%.*]] = sext i32 [[TMP1]] to i64
diff --git a/clang/test/CodeGen/PowerPC/ppc-emmintrin.c b/clang/test/CodeGen/PowerPC/ppc-emmintrin.c
index a3650beec625f2..4c4d0dfce05eaf 100644
--- a/clang/test/CodeGen/PowerPC/ppc-emmintrin.c
+++ b/clang/test/CodeGen/PowerPC/ppc-emmintrin.c
@@ -1012,14 +1012,14 @@ test_shuffle() {
// CHECK: %[[SHR:[0-9a-zA-Z_.]+]] = ashr i32 %{{[0-9a-zA-Z_.]+}}, 6
// CHECK: %[[AND4:[0-9a-zA-Z_.]+]] = and i32 %[[SHR]], 3
// CHECK: sext i32 %[[AND4]] to i64
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %{{[0-9a-zA-Z_.]+}}, i32 0
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %{{[0-9a-zA-Z_.]+}}, i32 1
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK: %[[ADD:[0-9a-zA-Z_.]+]] = add i32 %{{[0-9a-zA-Z_.]+}}, 269488144
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %[[ADD]], i32 2
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK: add i32 %{{[0-9a-zA-Z_.]+}}, 269488144
// CHECK: call <4 x i32> @vec_perm(int vector[4], int vector[4], unsigned char vector[16])
@@ -1050,7 +1050,7 @@ test_shuffle() {
// CHECK: sext i32 %[[AND4]] to i64
// CHECK-LE: store <2 x i64> <i64 1663540288323457296, i64 0>, ptr %{{[0-9a-zA-Z_.]+}}, align 16
// CHECK-BE: store <2 x i64> <i64 1157726452361532951, i64 0>, ptr %{{[0-9a-zA-Z_.]+}}, align 16
-// CHECK-COUNT-4: getelementptr inbounds [4 x i16], ptr @_mm_shufflehi_epi16.__permute_selectors, i64 0, i64 {{[0-9a-zA-Z_%.]+}}
+// CHECK-COUNT-4: getelementptr inbounds nuw [4 x i16], ptr @_mm_shufflehi_epi16.__permute_selectors, i64 0, i64 {{[0-9a-zA-Z_%.]+}}
// CHECK: call <2 x i64> @vec_perm(unsigned long long vector[2], unsigned long long vector[2], unsigned char vector[16])
// CHECK-LABEL: define available_externally <2 x i64> @_mm_shufflelo_epi16
@@ -1067,7 +1067,7 @@ test_shuffle() {
// CHECK: sext i32 %[[AND4]] to i64
// CHECK-LE: store <2 x i64> <i64 0, i64 2242261671028070680>, ptr %{{[0-9a-zA-Z_.]+}}, align 16
// CHECK-BE: store <2 x i64> <i64 0, i64 1736447835066146335>, ptr %{{[0-9a-zA-Z_.]+}}, align 16
-// CHECK-COUNT-4: getelementptr inbounds [4 x i16], ptr @_mm_shufflelo_epi16.__permute_selectors, i64 0, i64 {{[0-9a-zA-Z_%.]+}}
+// CHECK-COUNT-4: getelementptr inbounds nuw [4 x i16], ptr @_mm_shufflelo_epi16.__permute_selectors, i64 0, i64 {{[0-9a-zA-Z_%.]+}}
// CHECK: call <2 x i64> @vec_perm(unsigned long long vector[2], unsigned long long vector[2], unsigned char vector[16])
void __attribute__((noinline))
diff --git a/clang/test/CodeGen/PowerPC/ppc-xmmintrin.c b/clang/test/CodeGen/PowerPC/ppc-xmmintrin.c
index 95dfd1202f1575..4a15fa9f76ceea 100644
--- a/clang/test/CodeGen/PowerPC/ppc-xmmintrin.c
+++ b/clang/test/CodeGen/PowerPC/ppc-xmmintrin.c
@@ -894,16 +894,16 @@ test_shuffle() {
// CHECK: %[[SHR3:[0-9a-zA-Z_.]+]] = ashr i32 %{{[0-9a-zA-Z_.]+}}, 6
// CHECK: %[[AND4:[0-9a-zA-Z_.]+]] = and i32 %[[SHR3]], 3
// CHECK: sext i32 %[[AND4]] to i64
-// CHECK: getelementptr inbounds [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK-LE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 0
// CHECK-BE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 3
-// CHECK: getelementptr inbounds [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK-LE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 1
// CHECK-BE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 2
-// CHECK: getelementptr inbounds [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK-LE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 2
// CHECK-BE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 1
-// CHECK: getelementptr inbounds [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK-LE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 3
// CHECK-BE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 0
// CHECK: call <2 x i64> @vec_splats(unsigned long long)
@@ -923,14 +923,14 @@ test_shuffle() {
// CHECK: %[[SHR3:[0-9a-zA-Z_.]+]] = ashr i32 %{{[0-9a-zA-Z_.]+}}, 6
// CHECK: %[[AND4:[0-9a-zA-Z_.]+]] = and i32 %[[SHR3]], 3
// CHECK: sext i32 %[[AND4]] to i64
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %{{[0-9a-zA-Z_.]+}}, i32 0
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %{{[0-9a-zA-Z_.]+}}, i32 1
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
// CHECK: %[[ADD:[0-9a-zA-Z_.]+]] = add i32 %{{[0-9a-zA-Z_.]+}}, 269488144
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %[[ADD]], i32 2
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
// CHECK: %[[ADD2:[0-9a-zA-Z_.]+]] = add i32 %{{[0-9a-zA-Z_.]+}}, 269488144
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %[[ADD2]], i32 3
// CHECK: call <4 x float> @vec_perm(float vector[4], float vector[4], unsigned char vector[16])
diff --git a/clang/test/CodeGen/attr-counted-by.c b/clang/test/CodeGen/attr-counted-by.c
index 9acc896c0f0e9d..a1182cdd5ceef9 100644
--- a/clang/test/CodeGen/attr-counted-by.c
+++ b/clang/test/CodeGen/attr-counted-by.c
@@ -118,7 +118,7 @@ void test1(struct annotated *p, int index, int val) {
// SANITIZE-WITH-ATTR-NEXT: unreachable, !nosanitize [[META2]]
// SANITIZE-WITH-ATTR: cont3:
// SANITIZE-WITH-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// SANITIZE-WITH-ATTR-NEXT: [[TMP2:%.*]] = shl i32 [[DOT_COUNTED_BY_LOAD]], 2
// SANITIZE-WITH-ATTR-NEXT: [[DOTINV:%.*]] = icmp slt i32 [[DOT_COUNTED_BY_LOAD]], 0
// SANITIZE-WITH-ATTR-NEXT: [[CONV:%.*]] = select i1 [[DOTINV]], i32 0, i32 [[TMP2]]
@@ -134,7 +134,7 @@ void test1(struct annotated *p, int index, int val) {
// NO-SANITIZE-WITH-ATTR-NEXT: [[DOTINV:%.*]] = icmp slt i32 [[DOT_COUNTED_BY_LOAD]], 0
// NO-SANITIZE-WITH-ATTR-NEXT: [[CONV:%.*]] = select i1 [[DOTINV]], i32 0, i32 [[TMP0]]
// NO-SANITIZE-WITH-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// NO-SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// NO-SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// NO-SANITIZE-WITH-ATTR-NEXT: store i32 [[CONV]], ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA2]]
// NO-SANITIZE-WITH-ATTR-NEXT: ret void
//
@@ -142,7 +142,7 @@ void test1(struct annotated *p, int index, int val) {
// SANITIZE-WITHOUT-ATTR-SAME: ptr noundef [[P:%.*]], i64 noundef [[INDEX:%.*]]) local_unnamed_addr #[[ATTR0]] {
// SANITIZE-WITHOUT-ATTR-NEXT: entry:
// SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// SANITIZE-WITHOUT-ATTR-NEXT: store i32 -1, ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA2]]
// SANITIZE-WITHOUT-ATTR-NEXT: ret void
//
@@ -150,7 +150,7 @@ void test1(struct annotated *p, int index, int val) {
// NO-SANITIZE-WITHOUT-ATTR-SAME: ptr noundef [[P:%.*]], i64 noundef [[INDEX:%.*]]) local_unnamed_addr #[[ATTR0]] {
// NO-SANITIZE-WITHOUT-ATTR-NEXT: entry:
// NO-SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// NO-SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// NO-SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// NO-SANITIZE-WITHOUT-ATTR-NEXT: store i32 -1, ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA2]]
// NO-SANITIZE-WITHOUT-ATTR-NEXT: ret void
//
@@ -207,7 +207,7 @@ size_t test2_bdos(struct annotated *p) {
// SANITIZE-WITH-ATTR-NEXT: unreachable, !nosanitize [[META2]]
// SANITIZE-WITH-ATTR: cont3:
// SANITIZE-WITH-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// SANITIZE-WITH-ATTR-NEXT: [[TMP2:%.*]] = sext i32 [[DOT_COUNTED_BY_LOAD]] to i64
// SANITIZE-WITH-ATTR-NEXT: [[TMP3:%.*]] = shl nsw i64 [[TMP2]], 2
// SANITIZE-WITH-ATTR-NEXT: [[TMP4:%.*]] = tail call i64 @llvm.smax.i64(i64 [[TMP3]], i64 4)
@@ -231,7 +231,7 @@ size_t test2_bdos(struct annotated *p) {
// NO-SANITIZE-WITH-ATTR-NEXT: [[DOTINV:%.*]] = icmp slt i32 [[DOT_COUNTED_BY_LOAD]], 0
// NO-SANITIZE-WITH-ATTR-NEXT: [[CONV:%.*]] = select i1 [[DOTINV]], i32 0, i32 [[TMP4]]
// NO-SANITIZE-WITH-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// NO-SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// NO-SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// NO-SANITIZE-WITH-ATTR-NEXT: store i32 [[CONV]], ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA2]]
// NO-SANITIZE-WITH-ATTR-NEXT: ret void
//
@@ -239,7 +239,7 @@ size_t test2_bdos(struct annotated *p) {
// SANITIZE-WITHOUT-ATTR-SAME: ptr noundef [[P:%.*]], i64 noundef [[INDEX:%.*]]) local_unnamed_addr #[[ATTR0]] {
// SANITIZE-WITHOUT-ATTR-NEXT: entry:
// SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// SANITIZE-WITHOUT-ATTR-NEXT: store i32 -1, ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA2]]
// SANITIZE-WITHOUT-ATTR-NEXT: ret void
//
@@ -247,7 +247,7 @@ size_t test2_bdos(struct annotated *p) {
// NO-SANITIZE-WITHOUT-ATTR-SAME: ptr noundef [[P:%.*]], i64 noundef [[INDEX:%.*]]) local_unnamed_addr #[[ATTR0]] {
// NO-SANITIZE-WITHOUT-ATTR-NEXT: entry:
// NO-SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// NO-SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// NO-SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// NO-SANITIZE-WITHOUT-ATTR-NEXT: store i32 -1, ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA2]]
// NO-SANITIZE-WITHOUT-ATTR-NEXT: ret void
//
diff --git a/clang/test/CodeGen/catch-nullptr-and-nonzero-offset-when-nullptr-is-defined.c b/clang/test/CodeGen/catch-nullptr-and-nonzero-offset-when-nullptr-is-defined.c
index 39ede01d6e3b83..8a560a47ad1e10 100644
--- a/clang/test/CodeGen/catch-nullptr-and-nonzero-offset-when-nullptr-is-defined.c
+++ b/clang/test/CodeGen/catch-nullptr-and-nonzero-offset-when-nullptr-is-defined.c
@@ -33,7 +33,7 @@ char *add_unsigned(char *base, unsigned long offset) {
// CHECK-NEXT: store i64 %[[OFFSET]], ptr %[[OFFSET_ADDR]], align 8
// CHECK-NEXT: %[[BASE_RELOADED:.*]] = load ptr, ptr %[[BASE_ADDR]], align 8
// CHECK-NEXT: %[[OFFSET_RELOADED:.*]] = load i64, ptr %[[OFFSET_ADDR]], align 8
- // CHECK-NEXT: %[[ADD_PTR:.*]] = getelementptr inbounds i8, ptr %[[BASE_RELOADED]], i64 %[[OFFSET_RELOADED]]
+ // CHECK-NEXT: %[[ADD_PTR:.*]] = getelementptr inbounds nuw i8, ptr %[[BASE_RELOADED]], i64 %[[OFFSET_RELOADED]]
// CHECK-SANITIZE-NEXT: %[[COMPUTED_OFFSET_AGGREGATE:.*]] = call { i64, i1 } @llvm.smul.with.overflow.i64(i64 1, i64 %[[OFFSET_RELOADED]]), !nosanitize
// CHECK-SANITIZE-NEXT: %[[COMPUTED_OFFSET_OVERFLOWED:.*]] = extractvalue { i64, i1 } %[[COMPUTED_OFFSET_AGGREGATE]], 1, !nosanitize
// CHECK-SANITIZE-NEXT: %[[OR_OV:.+]] = or i1 %[[COMPUTED_OFFSET_OVERFLOWED]], false, !nosanitize
diff --git a/clang/test/CodeGen/catch-nullptr-and-nonzero-offset.c b/clang/test/CodeGen/catch-nullptr-and-nonzero-offset.c
index e93dbcb9f647bf..d884993ffb2b30 100644
--- a/clang/test/CodeGen/catch-nullptr-and-nonzero-offset.c
+++ b/clang/test/CodeGen/catch-nullptr-and-nonzero-offset.c
@@ -50,7 +50,7 @@ char *var_var(char *base, unsigned long offset) {
// CHECK-NEXT: store i64 %[[OFFSET]], ptr %[[OFFSET_ADDR]], align 8
// CHECK-NEXT: %[[BASE_RELOADED:.*]] = load ptr, ptr %[[BASE_ADDR]], align 8
// CHECK-NEXT: %[[OFFSET_RELOADED:.*]] = load i64, ptr %[[OFFSET_ADDR]], align 8
- // CHECK-NEXT: %[[ADD_PTR:.*]] = getelementptr inbounds i8, ptr %[[BASE_RELOADED]], i64 %[[OFFSET_RELOADED]]
+ // CHECK-NEXT: %[[ADD_PTR:.*]] = getelementptr inbounds nuw i8, ptr %[[BASE_RELOADED]], i64 %[[OFFSET_RELOADED]]
// CHECK-SANITIZE-NEXT: %[[COMPUTED_OFFSET_AGGREGATE:.*]] = call { i64, i1 } @llvm.smul.with.overflow.i64(i64 1, i64 %[[OFFSET_RELOADED]]), !nosanitize
// CHECK-SANITIZE-NEXT: %[[COMPUTED_OFFSET_OVERFLOWED:.*]] = extractvalue { i64, i1 } %[[COMPUTED_OFFSET_AGGREGATE]], 1, !nosanitize
// CHECK-SANITIZE-NEXT: %[[OR_OV:.+]] = or i1 %[[COMPUTED_OFFSET_OVERFLOWED]], false, !nosanitize
@...
[truncated]
|
|
@llvm/pr-subscribers-clang-codegen Author: Hari Limaye (hazzlim) ChangesAdd nuw attribute to inbounds GEPs where the expression used to form the GEP is an addition of unsigned indices. Patch is 576.17 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/105496.diff 91 Files Affected:
diff --git a/clang/lib/CodeGen/CGBuilder.h b/clang/lib/CodeGen/CGBuilder.h
index 08730a6a6672a1..b8036cf6e6a306 100644
--- a/clang/lib/CodeGen/CGBuilder.h
+++ b/clang/lib/CodeGen/CGBuilder.h
@@ -14,6 +14,7 @@
#include "CodeGenTypeCache.h"
#include "llvm/Analysis/Utils/Local.h"
#include "llvm/IR/DataLayout.h"
+#include "llvm/IR/GEPNoWrapFlags.h"
#include "llvm/IR/IRBuilder.h"
#include "llvm/IR/Type.h"
@@ -334,9 +335,10 @@ class CGBuilderTy : public CGBuilderBaseTy {
Address CreateGEP(Address Addr, ArrayRef<llvm::Value *> IdxList,
llvm::Type *ElementType, CharUnits Align,
- const Twine &Name = "") {
+ const Twine &Name = "",
+ llvm::GEPNoWrapFlags NW = llvm::GEPNoWrapFlags::none()) {
llvm::Value *Ptr = emitRawPointerFromAddress(Addr);
- return RawAddress(CreateGEP(Addr.getElementType(), Ptr, IdxList, Name),
+ return RawAddress(CreateGEP(Addr.getElementType(), Ptr, IdxList, Name, NW),
ElementType, Align);
}
diff --git a/clang/lib/CodeGen/CGExprScalar.cpp b/clang/lib/CodeGen/CGExprScalar.cpp
index 3bda254c86adf6..41a91aa435f999 100644
--- a/clang/lib/CodeGen/CGExprScalar.cpp
+++ b/clang/lib/CodeGen/CGExprScalar.cpp
@@ -36,6 +36,7 @@
#include "llvm/IR/DerivedTypes.h"
#include "llvm/IR/FixedPointBuilder.h"
#include "llvm/IR/Function.h"
+#include "llvm/IR/GEPNoWrapFlags.h"
#include "llvm/IR/GetElementPtrTypeIterator.h"
#include "llvm/IR/GlobalVariable.h"
#include "llvm/IR/Intrinsics.h"
@@ -5756,7 +5757,12 @@ CodeGenFunction::EmitCheckedInBoundsGEP(llvm::Type *ElemTy, Value *Ptr,
bool SignedIndices, bool IsSubtraction,
SourceLocation Loc, const Twine &Name) {
llvm::Type *PtrTy = Ptr->getType();
- Value *GEPVal = Builder.CreateInBoundsGEP(ElemTy, Ptr, IdxList, Name);
+
+ llvm::GEPNoWrapFlags NWFlags = llvm::GEPNoWrapFlags::inBounds();
+ if (!SignedIndices && !IsSubtraction)
+ NWFlags |= llvm::GEPNoWrapFlags::noUnsignedWrap();
+
+ Value *GEPVal = Builder.CreateGEP(ElemTy, Ptr, IdxList, Name, NWFlags);
// If the pointer overflow sanitizer isn't enabled, do nothing.
if (!SanOpts.has(SanitizerKind::PointerOverflow))
@@ -5871,8 +5877,13 @@ Address CodeGenFunction::EmitCheckedInBoundsGEP(
Address Addr, ArrayRef<Value *> IdxList, llvm::Type *elementType,
bool SignedIndices, bool IsSubtraction, SourceLocation Loc, CharUnits Align,
const Twine &Name) {
- if (!SanOpts.has(SanitizerKind::PointerOverflow))
- return Builder.CreateInBoundsGEP(Addr, IdxList, elementType, Align, Name);
+ if (!SanOpts.has(SanitizerKind::PointerOverflow)) {
+ llvm::GEPNoWrapFlags NWFlags = llvm::GEPNoWrapFlags::inBounds();
+ if (!SignedIndices && !IsSubtraction)
+ NWFlags |= llvm::GEPNoWrapFlags::noUnsignedWrap();
+
+ return Builder.CreateGEP(Addr, IdxList, elementType, Align, Name, NWFlags);
+ }
return RawAddress(
EmitCheckedInBoundsGEP(Addr.getElementType(), Addr.emitRawPointer(*this),
diff --git a/clang/test/CodeGen/2005-01-02-ConstantInits.c b/clang/test/CodeGen/2005-01-02-ConstantInits.c
index 7772a64331ffb7..1878c00fa89914 100644
--- a/clang/test/CodeGen/2005-01-02-ConstantInits.c
+++ b/clang/test/CodeGen/2005-01-02-ConstantInits.c
@@ -7,13 +7,6 @@
struct X { int a[2]; };
extern int bar();
-//.
-// CHECK: @test.i23 = internal global i32 4, align 4
-// CHECK: @i = global i32 4, align 4
-// CHECK: @Arr = global [100 x i32] zeroinitializer, align 16
-// CHECK: @foo2.X = internal global ptr getelementptr (i8, ptr @Arr, i64 196), align 8
-// CHECK: @foo2.i23 = internal global i32 0, align 4
-//.
// CHECK-LABEL: define {{[^@]+}}@test
// CHECK-SAME: () #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: entry:
@@ -51,7 +44,7 @@ int foo(int i) { return bar(&Arr[49])+bar(&Arr[i]); }
// CHECK-NEXT: store i32 [[I]], ptr [[I_ADDR]], align 4
// CHECK-NEXT: store ptr @Arr, ptr [[P]], align 8
// CHECK-NEXT: [[TMP0:%.*]] = load ptr, ptr [[P]], align 8
-// CHECK-NEXT: [[INCDEC_PTR:%.*]] = getelementptr inbounds i32, ptr [[TMP0]], i32 1
+// CHECK-NEXT: [[INCDEC_PTR:%.*]] = getelementptr inbounds nuw i32, ptr [[TMP0]], i32 1
// CHECK-NEXT: store ptr [[INCDEC_PTR]], ptr [[P]], align 8
// CHECK-NEXT: [[TMP1:%.*]] = load i32, ptr [[I_ADDR]], align 4
// CHECK-NEXT: [[IDX_EXT:%.*]] = sext i32 [[TMP1]] to i64
diff --git a/clang/test/CodeGen/PowerPC/ppc-emmintrin.c b/clang/test/CodeGen/PowerPC/ppc-emmintrin.c
index a3650beec625f2..4c4d0dfce05eaf 100644
--- a/clang/test/CodeGen/PowerPC/ppc-emmintrin.c
+++ b/clang/test/CodeGen/PowerPC/ppc-emmintrin.c
@@ -1012,14 +1012,14 @@ test_shuffle() {
// CHECK: %[[SHR:[0-9a-zA-Z_.]+]] = ashr i32 %{{[0-9a-zA-Z_.]+}}, 6
// CHECK: %[[AND4:[0-9a-zA-Z_.]+]] = and i32 %[[SHR]], 3
// CHECK: sext i32 %[[AND4]] to i64
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %{{[0-9a-zA-Z_.]+}}, i32 0
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %{{[0-9a-zA-Z_.]+}}, i32 1
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK: %[[ADD:[0-9a-zA-Z_.]+]] = add i32 %{{[0-9a-zA-Z_.]+}}, 269488144
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %[[ADD]], i32 2
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_epi32.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK: add i32 %{{[0-9a-zA-Z_.]+}}, 269488144
// CHECK: call <4 x i32> @vec_perm(int vector[4], int vector[4], unsigned char vector[16])
@@ -1050,7 +1050,7 @@ test_shuffle() {
// CHECK: sext i32 %[[AND4]] to i64
// CHECK-LE: store <2 x i64> <i64 1663540288323457296, i64 0>, ptr %{{[0-9a-zA-Z_.]+}}, align 16
// CHECK-BE: store <2 x i64> <i64 1157726452361532951, i64 0>, ptr %{{[0-9a-zA-Z_.]+}}, align 16
-// CHECK-COUNT-4: getelementptr inbounds [4 x i16], ptr @_mm_shufflehi_epi16.__permute_selectors, i64 0, i64 {{[0-9a-zA-Z_%.]+}}
+// CHECK-COUNT-4: getelementptr inbounds nuw [4 x i16], ptr @_mm_shufflehi_epi16.__permute_selectors, i64 0, i64 {{[0-9a-zA-Z_%.]+}}
// CHECK: call <2 x i64> @vec_perm(unsigned long long vector[2], unsigned long long vector[2], unsigned char vector[16])
// CHECK-LABEL: define available_externally <2 x i64> @_mm_shufflelo_epi16
@@ -1067,7 +1067,7 @@ test_shuffle() {
// CHECK: sext i32 %[[AND4]] to i64
// CHECK-LE: store <2 x i64> <i64 0, i64 2242261671028070680>, ptr %{{[0-9a-zA-Z_.]+}}, align 16
// CHECK-BE: store <2 x i64> <i64 0, i64 1736447835066146335>, ptr %{{[0-9a-zA-Z_.]+}}, align 16
-// CHECK-COUNT-4: getelementptr inbounds [4 x i16], ptr @_mm_shufflelo_epi16.__permute_selectors, i64 0, i64 {{[0-9a-zA-Z_%.]+}}
+// CHECK-COUNT-4: getelementptr inbounds nuw [4 x i16], ptr @_mm_shufflelo_epi16.__permute_selectors, i64 0, i64 {{[0-9a-zA-Z_%.]+}}
// CHECK: call <2 x i64> @vec_perm(unsigned long long vector[2], unsigned long long vector[2], unsigned char vector[16])
void __attribute__((noinline))
diff --git a/clang/test/CodeGen/PowerPC/ppc-xmmintrin.c b/clang/test/CodeGen/PowerPC/ppc-xmmintrin.c
index 95dfd1202f1575..4a15fa9f76ceea 100644
--- a/clang/test/CodeGen/PowerPC/ppc-xmmintrin.c
+++ b/clang/test/CodeGen/PowerPC/ppc-xmmintrin.c
@@ -894,16 +894,16 @@ test_shuffle() {
// CHECK: %[[SHR3:[0-9a-zA-Z_.]+]] = ashr i32 %{{[0-9a-zA-Z_.]+}}, 6
// CHECK: %[[AND4:[0-9a-zA-Z_.]+]] = and i32 %[[SHR3]], 3
// CHECK: sext i32 %[[AND4]] to i64
-// CHECK: getelementptr inbounds [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK-LE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 0
// CHECK-BE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 3
-// CHECK: getelementptr inbounds [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK-LE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 1
// CHECK-BE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 2
-// CHECK: getelementptr inbounds [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK-LE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 2
// CHECK-BE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 1
-// CHECK: getelementptr inbounds [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
+// CHECK: getelementptr inbounds nuw [4 x i16], ptr @_mm_shuffle_pi16.__permute_selectors, i64 0, i64 %{{[0-9a-zA-Z_.]+}}
// CHECK-LE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 3
// CHECK-BE: getelementptr inbounds [4 x i16], ptr %{{[0-9a-zA-Z_.]+}}, i64 0, i64 0
// CHECK: call <2 x i64> @vec_splats(unsigned long long)
@@ -923,14 +923,14 @@ test_shuffle() {
// CHECK: %[[SHR3:[0-9a-zA-Z_.]+]] = ashr i32 %{{[0-9a-zA-Z_.]+}}, 6
// CHECK: %[[AND4:[0-9a-zA-Z_.]+]] = and i32 %[[SHR3]], 3
// CHECK: sext i32 %[[AND4]] to i64
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %{{[0-9a-zA-Z_.]+}}, i32 0
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %{{[0-9a-zA-Z_.]+}}, i32 1
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
// CHECK: %[[ADD:[0-9a-zA-Z_.]+]] = add i32 %{{[0-9a-zA-Z_.]+}}, 269488144
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %[[ADD]], i32 2
-// CHECK: getelementptr inbounds [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
+// CHECK: getelementptr inbounds nuw [4 x i32], ptr @_mm_shuffle_ps.__permute_selectors, i64 0, i64
// CHECK: %[[ADD2:[0-9a-zA-Z_.]+]] = add i32 %{{[0-9a-zA-Z_.]+}}, 269488144
// CHECK: insertelement <4 x i32> %{{[0-9a-zA-Z_.]+}}, i32 %[[ADD2]], i32 3
// CHECK: call <4 x float> @vec_perm(float vector[4], float vector[4], unsigned char vector[16])
diff --git a/clang/test/CodeGen/attr-counted-by.c b/clang/test/CodeGen/attr-counted-by.c
index 9acc896c0f0e9d..a1182cdd5ceef9 100644
--- a/clang/test/CodeGen/attr-counted-by.c
+++ b/clang/test/CodeGen/attr-counted-by.c
@@ -118,7 +118,7 @@ void test1(struct annotated *p, int index, int val) {
// SANITIZE-WITH-ATTR-NEXT: unreachable, !nosanitize [[META2]]
// SANITIZE-WITH-ATTR: cont3:
// SANITIZE-WITH-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// SANITIZE-WITH-ATTR-NEXT: [[TMP2:%.*]] = shl i32 [[DOT_COUNTED_BY_LOAD]], 2
// SANITIZE-WITH-ATTR-NEXT: [[DOTINV:%.*]] = icmp slt i32 [[DOT_COUNTED_BY_LOAD]], 0
// SANITIZE-WITH-ATTR-NEXT: [[CONV:%.*]] = select i1 [[DOTINV]], i32 0, i32 [[TMP2]]
@@ -134,7 +134,7 @@ void test1(struct annotated *p, int index, int val) {
// NO-SANITIZE-WITH-ATTR-NEXT: [[DOTINV:%.*]] = icmp slt i32 [[DOT_COUNTED_BY_LOAD]], 0
// NO-SANITIZE-WITH-ATTR-NEXT: [[CONV:%.*]] = select i1 [[DOTINV]], i32 0, i32 [[TMP0]]
// NO-SANITIZE-WITH-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// NO-SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// NO-SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// NO-SANITIZE-WITH-ATTR-NEXT: store i32 [[CONV]], ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA2]]
// NO-SANITIZE-WITH-ATTR-NEXT: ret void
//
@@ -142,7 +142,7 @@ void test1(struct annotated *p, int index, int val) {
// SANITIZE-WITHOUT-ATTR-SAME: ptr noundef [[P:%.*]], i64 noundef [[INDEX:%.*]]) local_unnamed_addr #[[ATTR0]] {
// SANITIZE-WITHOUT-ATTR-NEXT: entry:
// SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// SANITIZE-WITHOUT-ATTR-NEXT: store i32 -1, ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA2]]
// SANITIZE-WITHOUT-ATTR-NEXT: ret void
//
@@ -150,7 +150,7 @@ void test1(struct annotated *p, int index, int val) {
// NO-SANITIZE-WITHOUT-ATTR-SAME: ptr noundef [[P:%.*]], i64 noundef [[INDEX:%.*]]) local_unnamed_addr #[[ATTR0]] {
// NO-SANITIZE-WITHOUT-ATTR-NEXT: entry:
// NO-SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// NO-SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// NO-SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// NO-SANITIZE-WITHOUT-ATTR-NEXT: store i32 -1, ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA2]]
// NO-SANITIZE-WITHOUT-ATTR-NEXT: ret void
//
@@ -207,7 +207,7 @@ size_t test2_bdos(struct annotated *p) {
// SANITIZE-WITH-ATTR-NEXT: unreachable, !nosanitize [[META2]]
// SANITIZE-WITH-ATTR: cont3:
// SANITIZE-WITH-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// SANITIZE-WITH-ATTR-NEXT: [[TMP2:%.*]] = sext i32 [[DOT_COUNTED_BY_LOAD]] to i64
// SANITIZE-WITH-ATTR-NEXT: [[TMP3:%.*]] = shl nsw i64 [[TMP2]], 2
// SANITIZE-WITH-ATTR-NEXT: [[TMP4:%.*]] = tail call i64 @llvm.smax.i64(i64 [[TMP3]], i64 4)
@@ -231,7 +231,7 @@ size_t test2_bdos(struct annotated *p) {
// NO-SANITIZE-WITH-ATTR-NEXT: [[DOTINV:%.*]] = icmp slt i32 [[DOT_COUNTED_BY_LOAD]], 0
// NO-SANITIZE-WITH-ATTR-NEXT: [[CONV:%.*]] = select i1 [[DOTINV]], i32 0, i32 [[TMP4]]
// NO-SANITIZE-WITH-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// NO-SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// NO-SANITIZE-WITH-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// NO-SANITIZE-WITH-ATTR-NEXT: store i32 [[CONV]], ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA2]]
// NO-SANITIZE-WITH-ATTR-NEXT: ret void
//
@@ -239,7 +239,7 @@ size_t test2_bdos(struct annotated *p) {
// SANITIZE-WITHOUT-ATTR-SAME: ptr noundef [[P:%.*]], i64 noundef [[INDEX:%.*]]) local_unnamed_addr #[[ATTR0]] {
// SANITIZE-WITHOUT-ATTR-NEXT: entry:
// SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// SANITIZE-WITHOUT-ATTR-NEXT: store i32 -1, ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA2]]
// SANITIZE-WITHOUT-ATTR-NEXT: ret void
//
@@ -247,7 +247,7 @@ size_t test2_bdos(struct annotated *p) {
// NO-SANITIZE-WITHOUT-ATTR-SAME: ptr noundef [[P:%.*]], i64 noundef [[INDEX:%.*]]) local_unnamed_addr #[[ATTR0]] {
// NO-SANITIZE-WITHOUT-ATTR-NEXT: entry:
// NO-SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAY:%.*]] = getelementptr inbounds nuw i8, ptr [[P]], i64 12
-// NO-SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
+// NO-SANITIZE-WITHOUT-ATTR-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds nuw [0 x i32], ptr [[ARRAY]], i64 0, i64 [[INDEX]]
// NO-SANITIZE-WITHOUT-ATTR-NEXT: store i32 -1, ptr [[ARRAYIDX]], align 4, !tbaa [[TBAA2]]
// NO-SANITIZE-WITHOUT-ATTR-NEXT: ret void
//
diff --git a/clang/test/CodeGen/catch-nullptr-and-nonzero-offset-when-nullptr-is-defined.c b/clang/test/CodeGen/catch-nullptr-and-nonzero-offset-when-nullptr-is-defined.c
index 39ede01d6e3b83..8a560a47ad1e10 100644
--- a/clang/test/CodeGen/catch-nullptr-and-nonzero-offset-when-nullptr-is-defined.c
+++ b/clang/test/CodeGen/catch-nullptr-and-nonzero-offset-when-nullptr-is-defined.c
@@ -33,7 +33,7 @@ char *add_unsigned(char *base, unsigned long offset) {
// CHECK-NEXT: store i64 %[[OFFSET]], ptr %[[OFFSET_ADDR]], align 8
// CHECK-NEXT: %[[BASE_RELOADED:.*]] = load ptr, ptr %[[BASE_ADDR]], align 8
// CHECK-NEXT: %[[OFFSET_RELOADED:.*]] = load i64, ptr %[[OFFSET_ADDR]], align 8
- // CHECK-NEXT: %[[ADD_PTR:.*]] = getelementptr inbounds i8, ptr %[[BASE_RELOADED]], i64 %[[OFFSET_RELOADED]]
+ // CHECK-NEXT: %[[ADD_PTR:.*]] = getelementptr inbounds nuw i8, ptr %[[BASE_RELOADED]], i64 %[[OFFSET_RELOADED]]
// CHECK-SANITIZE-NEXT: %[[COMPUTED_OFFSET_AGGREGATE:.*]] = call { i64, i1 } @llvm.smul.with.overflow.i64(i64 1, i64 %[[OFFSET_RELOADED]]), !nosanitize
// CHECK-SANITIZE-NEXT: %[[COMPUTED_OFFSET_OVERFLOWED:.*]] = extractvalue { i64, i1 } %[[COMPUTED_OFFSET_AGGREGATE]], 1, !nosanitize
// CHECK-SANITIZE-NEXT: %[[OR_OV:.+]] = or i1 %[[COMPUTED_OFFSET_OVERFLOWED]], false, !nosanitize
diff --git a/clang/test/CodeGen/catch-nullptr-and-nonzero-offset.c b/clang/test/CodeGen/catch-nullptr-and-nonzero-offset.c
index e93dbcb9f647bf..d884993ffb2b30 100644
--- a/clang/test/CodeGen/catch-nullptr-and-nonzero-offset.c
+++ b/clang/test/CodeGen/catch-nullptr-and-nonzero-offset.c
@@ -50,7 +50,7 @@ char *var_var(char *base, unsigned long offset) {
// CHECK-NEXT: store i64 %[[OFFSET]], ptr %[[OFFSET_ADDR]], align 8
// CHECK-NEXT: %[[BASE_RELOADED:.*]] = load ptr, ptr %[[BASE_ADDR]], align 8
// CHECK-NEXT: %[[OFFSET_RELOADED:.*]] = load i64, ptr %[[OFFSET_ADDR]], align 8
- // CHECK-NEXT: %[[ADD_PTR:.*]] = getelementptr inbounds i8, ptr %[[BASE_RELOADED]], i64 %[[OFFSET_RELOADED]]
+ // CHECK-NEXT: %[[ADD_PTR:.*]] = getelementptr inbounds nuw i8, ptr %[[BASE_RELOADED]], i64 %[[OFFSET_RELOADED]]
// CHECK-SANITIZE-NEXT: %[[COMPUTED_OFFSET_AGGREGATE:.*]] = call { i64, i1 } @llvm.smul.with.overflow.i64(i64 1, i64 %[[OFFSET_RELOADED]]), !nosanitize
// CHECK-SANITIZE-NEXT: %[[COMPUTED_OFFSET_OVERFLOWED:.*]] = extractvalue { i64, i1 } %[[COMPUTED_OFFSET_AGGREGATE]], 1, !nosanitize
// CHECK-SANITIZE-NEXT: %[[OR_OV:.+]] = or i1 %[[COMPUTED_OFFSET_OVERFLOWED]], false, !nosanitize
@...
[truncated]
|
hazzlim
requested review from
nikic,
fhahn,
jdoerfert,
efriedma-quic and
david-arm
nikic
reviewed
Aug 21, 2024
| // CHECK: @i = global i32 4, align 4 | ||
| // CHECK: @Arr = global [100 x i32] zeroinitializer, align 16 | ||
| // CHECK: @foo2.X = internal global ptr getelementptr (i8, ptr @Arr, i64 196), align 8 | ||
| // CHECK: @foo2.i23 = internal global i32 0, align 4 |
There was a problem hiding this comment.
Something went wrong with update_cc_test_checks here? It has --check-globals, so this was supposed to be checked...
There was a problem hiding this comment.
Hmm good point - I think the issue is this file uses --global-value-regex "@.+", which prevents matching because the update script internals actually prefix "^@" to the regex (and obviously "^@@.+" doesn't match)
llvm-project/llvm/utils/UpdateTestChecks/common.py
Lines 2280 to 2281 in 0cff3e8
I'm not sure what's best to do here - we can remove --global-value-regex "@.+", which restores CHECK lines for these globals, but that will also add CHECK lines for some attributes because they match the regex on line 2281 there...
We could add --global-value-regex "[A-Za-z].*" to capture globals only, not sure if that's too hacky?
There was a problem hiding this comment.
Interesting. So the way this is designed you can't cleanly distinguish globals and attributes.
I think you solution sounds fine.
There was a problem hiding this comment.
Yeah, possibly something that could be addressed in the script if desired - not sure how much it comes up.
Ok great, I've updated the --global-value-regex argument and fixed up the missing global CHECKs.
update_cc_test_checks.py will prefix the --global-value-regex argument with "^@" and "^!" when trying to match globals, and so the current argument of "@.+" will not match globals. Replace "@.+" with "[A-Za-z].*" to correctly match globals and not attributes.
efriedma-quic
approved these changes
Aug 21, 2024
|
I plan to land this patch later today, unless there are any objections / further comments. Thanks for reviewing @nikic @efriedma-quic |
vitalybuka
added a commit
that referenced
this pull request
Aug 28, 2024
This reverts commit 3d2fd31.
From the buildbot run I can see 12 or 13 changes in the build that failed. Just out of curiosity how did you find out it was this patch that caused it? |
Bisected and reverted locally. |
Thanks! I was just wondering if it was something obvious from the patch or whether there was a bisecting tool to use in github. But bisecting and reverting locally is the same thing I'd do. |
|
Maybe the patch introduced some miscompile? Also there are branches checking sanitizers. |
|
We're seeing many test failures in Chromium after this change as well: https://crbug.com/362522336 |
zmodem
pushed a commit
that referenced
this pull request
Aug 28, 2024
Reverts #105496 This patch breaks: https://lab.llvm.org/buildbot/#/builders/25/builds/1952 https://lab.llvm.org/buildbot/#/builders/52/builds/1775 Somehow output is different with sanitizers. Maybe non-determinism in the code?
|
Hi, Downstream we also saw miscompiles with this patch. I've debugged one such case and I end up with GVN doing things I don't understand. on we get so it decides that no matter the value of input parameter %n, the return value is undef. I don't understand this. And this only happens when there is "nuw" on the gep. |
|
@mikaelholmen Thanks for the reproducer, this makes the issue clear. BasicAA is incorrectly returning NoAlias for the pointers due to #98608. The issue is that the |
Thanks for identifying this - I'm looking into a fix. Would you prefer I revert #98608 in the meantime @nikic ? |
|
I don't think we need to revert it. Without this patch it's mostly harmless. I've been working on a fix for this issue. I think we need to track NUW through LinearExpression using something along these lines: https://gist.github.com/nikic/d6986bb62539bba048ced15dd497a4fb |
Ah great thank you, that’s what I was thinking - if we then make the NoAlias condition added in #98608 conditional on Offset.IsNUW that should fix things, if I understand correctly. Sounds like you are intending to put up a fix for this, so I will leave this to you - thank you :) |
nikic
added a commit
that referenced
this pull request
Sep 2, 2024
When we decompose the GEP offset expression, and the arithmetic is not performed using nuw operations, we cannot retain the nuw flag on the decomposed GEP. For example, if we have `gep nuw p, (a-1)`, this is not at all the same as `gep nuw (gep nuw p, a), -1`. Fix this by tracking NUW through linear expression decomposition, similarly to what we already do for the NSW flag. This fixes the miscompilation reported in #105496 (comment).
|
This should be ready to re-apply now. |
hazzlim
added a commit
to hazzlim/llvm-project
that referenced
this pull request
Sep 4, 2024
Add nuw attribute to inbounds GEPs where the expression used to form the GEP is an addition of unsigned indices. Relands llvm#105496, which was reverted because it exposed a miscompilation arising from llvm#98608. This is now fixed by llvm#106512.
|
This breaks wasm codegen: #108770 |
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.
Add nuw attribute to inbounds GEPs where the expression used to form the GEP is an addition of unsigned indices.