[llvm] [IndVarSimplify] Fix Masking Issue by Adding nsw/nuw Flags to Trunc Instruction (PR #150179)
via llvm-commits
llvm-commits at lists.llvm.org
Tue Sep 2 04:05:39 PDT 2025
https://github.com/buggfg updated https://github.com/llvm/llvm-project/pull/150179
>From c6a39d83ea7c90dd00bb1a6832816d5d69be02f7 Mon Sep 17 00:00:00 2001
From: buggfg <3171290993 at qq.com>
Date: Wed, 23 Jul 2025 15:20:40 +0800
Subject: [PATCH 1/5] [IndVarSimplify] Fix Masking Issue by Adding nsw/nuw
Flags to Trunc Instruction
---
llvm/lib/Transforms/Scalar/IndVarSimplify.cpp | 21 ++++++++++++++++++-
.../IndVarSimplify/X86/eliminate-trunc.ll | 8 +++----
.../Transforms/IndVarSimplify/lftr-pr41998.ll | 2 +-
llvm/test/Transforms/IndVarSimplify/lftr.ll | 2 +-
.../constraint-elimination-placement.ll | 3 +--
5 files changed, 27 insertions(+), 9 deletions(-)
diff --git a/llvm/lib/Transforms/Scalar/IndVarSimplify.cpp b/llvm/lib/Transforms/Scalar/IndVarSimplify.cpp
index 334c911191cb8..04a1f4831b8d8 100644
--- a/llvm/lib/Transforms/Scalar/IndVarSimplify.cpp
+++ b/llvm/lib/Transforms/Scalar/IndVarSimplify.cpp
@@ -1049,9 +1049,28 @@ linearFunctionTestReplace(Loop *L, BasicBlock *ExitingBB,
if (Extended) {
bool Discard;
L->makeLoopInvariant(ExitCnt, Discard);
- } else
+ } else{
CmpIndVar = Builder.CreateTrunc(CmpIndVar, ExitCnt->getType(),
"lftr.wideiv");
+
+ // Set the correct wrap flag to avoid the masking issue.
+ Instruction *TruncInst = dyn_cast<Instruction>(CmpIndVar);
+
+ // The TruncatedIV is incrementing.
+ if (const SCEVAddRecExpr *TruncAR =
+ dyn_cast<SCEVAddRecExpr>(TruncatedIV)) {
+ // If TruncIV does not cause self-wrap, explicitly add the nsw and nuw
+ // flags to TruncInst.
+ if (TruncAR->hasNoSelfWrap()) {
+ TruncInst->setHasNoSignedWrap();
+ TruncInst->setHasNoUnsignedWrap();
+ } else if (TruncAR->hasNoSignedWrap()) {
+ TruncInst->setHasNoSignedWrap();
+ } else if (TruncAR->hasNoUnsignedWrap()) {
+ TruncInst->setHasNoUnsignedWrap();
+ }
+ }
+ }
}
LLVM_DEBUG(dbgs() << "INDVARS: Rewriting loop exit condition to:\n"
<< " LHS:" << *CmpIndVar << '\n'
diff --git a/llvm/test/Transforms/IndVarSimplify/X86/eliminate-trunc.ll b/llvm/test/Transforms/IndVarSimplify/X86/eliminate-trunc.ll
index 565ac5c8743d4..7e7a3f192f998 100644
--- a/llvm/test/Transforms/IndVarSimplify/X86/eliminate-trunc.ll
+++ b/llvm/test/Transforms/IndVarSimplify/X86/eliminate-trunc.ll
@@ -227,7 +227,7 @@ define void @test_01_unsigned(i32 %n) {
; CHECK: loop:
; CHECK-NEXT: [[IV:%.*]] = phi i64 [ 0, [[ENTRY:%.*]] ], [ [[IV_NEXT:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
-; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[IV_NEXT]] to i32
+; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc nuw nsw i64 [[IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i32 [[LFTR_WIDEIV]], [[TMP0]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[LOOP]], label [[EXIT:%.*]]
; CHECK: exit:
@@ -255,7 +255,7 @@ define void @test_02_unsigned(i32 %n) {
; CHECK: loop:
; CHECK-NEXT: [[IV:%.*]] = phi i64 [ 4294967294, [[ENTRY:%.*]] ], [ [[IV_NEXT:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
-; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[IV_NEXT]] to i32
+; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc nuw nsw i64 [[IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i32 [[LFTR_WIDEIV]], [[TMP0]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[LOOP]], label [[EXIT:%.*]]
; CHECK: exit:
@@ -304,7 +304,7 @@ define void @test_04_unsigned(i32 %n) {
; CHECK: loop:
; CHECK-NEXT: [[IV:%.*]] = phi i64 [ 0, [[ENTRY:%.*]] ], [ [[IV_NEXT:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
-; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[IV_NEXT]] to i32
+; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc nuw nsw i64 [[IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i32 [[LFTR_WIDEIV]], [[TMP0]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[LOOP]], label [[EXIT:%.*]]
; CHECK: exit:
@@ -332,7 +332,7 @@ define void @test_05_unsigned(i32 %n) {
; CHECK: loop:
; CHECK-NEXT: [[IV:%.*]] = phi i64 [ 1, [[ENTRY:%.*]] ], [ [[IV_NEXT:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
-; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[IV_NEXT]] to i32
+; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc nuw nsw i64 [[IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i32 [[LFTR_WIDEIV]], [[TMP0]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[LOOP]], label [[EXIT:%.*]]
; CHECK: exit:
diff --git a/llvm/test/Transforms/IndVarSimplify/lftr-pr41998.ll b/llvm/test/Transforms/IndVarSimplify/lftr-pr41998.ll
index b7f4756b2757f..376ef1ac5ffac 100644
--- a/llvm/test/Transforms/IndVarSimplify/lftr-pr41998.ll
+++ b/llvm/test/Transforms/IndVarSimplify/lftr-pr41998.ll
@@ -13,7 +13,7 @@ define void @test_int(i32 %start, ptr %p) {
; CHECK-NEXT: [[I2:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[I2_INC:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[I2_INC]] = add nuw nsw i32 [[I2]], 1
; CHECK-NEXT: store volatile i32 [[I2_INC]], ptr [[P:%.*]], align 4
-; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i32 [[I2_INC]] to i3
+; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc nuw nsw i32 [[I2_INC]] to i3
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i3 [[LFTR_WIDEIV]], [[TMP1]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[END:%.*]], label [[LOOP]]
; CHECK: end:
diff --git a/llvm/test/Transforms/IndVarSimplify/lftr.ll b/llvm/test/Transforms/IndVarSimplify/lftr.ll
index 5ee62ba357ab6..cfa4baa2d3b11 100644
--- a/llvm/test/Transforms/IndVarSimplify/lftr.ll
+++ b/llvm/test/Transforms/IndVarSimplify/lftr.ll
@@ -415,7 +415,7 @@ define void @wide_trip_count_test1(ptr %autoc,
; CHECK-NEXT: [[ADD3:%.*]] = fadd float [[TEMP2]], [[MUL]]
; CHECK-NEXT: store float [[ADD3]], ptr [[ARRAYIDX2]], align 4
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
-; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[INDVARS_IV_NEXT]] to i32
+; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc nuw nsw i64 [[INDVARS_IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i32 [[LFTR_WIDEIV]], [[SUB]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_BODY]], label [[FOR_END_LOOPEXIT:%.*]]
; CHECK: for.end.loopexit:
diff --git a/llvm/test/Transforms/PhaseOrdering/AArch64/constraint-elimination-placement.ll b/llvm/test/Transforms/PhaseOrdering/AArch64/constraint-elimination-placement.ll
index bbdbd95c6017a..ddd98f21c36a4 100644
--- a/llvm/test/Transforms/PhaseOrdering/AArch64/constraint-elimination-placement.ll
+++ b/llvm/test/Transforms/PhaseOrdering/AArch64/constraint-elimination-placement.ll
@@ -33,8 +33,7 @@ define i1 @test_order_1(ptr %this, ptr noalias %other, i1 %tobool9.not, i32 %cal
; CHECK-NEXT: br i1 [[CMP44]], label [[FOR_BODY45]], label [[FOR_COND]]
; CHECK: for.inc57:
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nsw i64 [[INDVARS_IV]], 1
-; CHECK-NEXT: [[TMP1:%.*]] = and i64 [[INDVARS_IV_NEXT]], 4294967295
-; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i64 [[TMP1]], 1
+; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i64 [[INDVARS_IV]], 0
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_COND41_PREHEADER_PREHEADER]], label [[FOR_COND41_PREHEADER]]
; CHECK: exit:
; CHECK-NEXT: ret i1 false
>From ac8981b9d1d7f053d324f22ac63c00720a5ab914 Mon Sep 17 00:00:00 2001
From: bernadate <3171290993 at qq.com>
Date: Thu, 24 Jul 2025 17:28:29 +0800
Subject: [PATCH 2/5] Update llvm/lib/Transforms/Scalar/IndVarSimplify.cpp
Co-authored-by: Luke Lau <luke_lau at icloud.com>
---
llvm/lib/Transforms/Scalar/IndVarSimplify.cpp | 7 ++-----
1 file changed, 2 insertions(+), 5 deletions(-)
diff --git a/llvm/lib/Transforms/Scalar/IndVarSimplify.cpp b/llvm/lib/Transforms/Scalar/IndVarSimplify.cpp
index 04a1f4831b8d8..bd04981f8b0d4 100644
--- a/llvm/lib/Transforms/Scalar/IndVarSimplify.cpp
+++ b/llvm/lib/Transforms/Scalar/IndVarSimplify.cpp
@@ -1050,11 +1050,8 @@ linearFunctionTestReplace(Loop *L, BasicBlock *ExitingBB,
bool Discard;
L->makeLoopInvariant(ExitCnt, Discard);
} else{
- CmpIndVar = Builder.CreateTrunc(CmpIndVar, ExitCnt->getType(),
- "lftr.wideiv");
-
- // Set the correct wrap flag to avoid the masking issue.
- Instruction *TruncInst = dyn_cast<Instruction>(CmpIndVar);
+CmpIndVar = Builder.CreateTrunc(CmpIndVar, ExitCnt->getType(),
+"lftr.wideiv", TruncAR->hasNoUnsignedWrap(), TruncAR->hasSignedWrap());
// The TruncatedIV is incrementing.
if (const SCEVAddRecExpr *TruncAR =
>From 2e146339ad528e29186d4e0dc1b1cc1d6f10f702 Mon Sep 17 00:00:00 2001
From: buggfg <3171290993 at qq.com>
Date: Tue, 12 Aug 2025 14:27:26 +0800
Subject: [PATCH 3/5] Revert "[IndVarSimplify] Fix Masking Issue by Adding
nsw/nuw Flags to Trunc Instruction"
This reverts commit c6a39d83ea7c90dd00bb1a6832816d5d69be02f7.
# Conflicts:
# llvm/lib/Transforms/Scalar/IndVarSimplify.cpp
---
llvm/lib/Transforms/Scalar/IndVarSimplify.cpp | 22 +++----------------
.../IndVarSimplify/X86/eliminate-trunc.ll | 8 +++----
.../Transforms/IndVarSimplify/lftr-pr41998.ll | 2 +-
llvm/test/Transforms/IndVarSimplify/lftr.ll | 2 +-
.../constraint-elimination-placement.ll | 3 ++-
5 files changed, 11 insertions(+), 26 deletions(-)
diff --git a/llvm/lib/Transforms/Scalar/IndVarSimplify.cpp b/llvm/lib/Transforms/Scalar/IndVarSimplify.cpp
index bd04981f8b0d4..334c911191cb8 100644
--- a/llvm/lib/Transforms/Scalar/IndVarSimplify.cpp
+++ b/llvm/lib/Transforms/Scalar/IndVarSimplify.cpp
@@ -1049,25 +1049,9 @@ linearFunctionTestReplace(Loop *L, BasicBlock *ExitingBB,
if (Extended) {
bool Discard;
L->makeLoopInvariant(ExitCnt, Discard);
- } else{
-CmpIndVar = Builder.CreateTrunc(CmpIndVar, ExitCnt->getType(),
-"lftr.wideiv", TruncAR->hasNoUnsignedWrap(), TruncAR->hasSignedWrap());
-
- // The TruncatedIV is incrementing.
- if (const SCEVAddRecExpr *TruncAR =
- dyn_cast<SCEVAddRecExpr>(TruncatedIV)) {
- // If TruncIV does not cause self-wrap, explicitly add the nsw and nuw
- // flags to TruncInst.
- if (TruncAR->hasNoSelfWrap()) {
- TruncInst->setHasNoSignedWrap();
- TruncInst->setHasNoUnsignedWrap();
- } else if (TruncAR->hasNoSignedWrap()) {
- TruncInst->setHasNoSignedWrap();
- } else if (TruncAR->hasNoUnsignedWrap()) {
- TruncInst->setHasNoUnsignedWrap();
- }
- }
- }
+ } else
+ CmpIndVar = Builder.CreateTrunc(CmpIndVar, ExitCnt->getType(),
+ "lftr.wideiv");
}
LLVM_DEBUG(dbgs() << "INDVARS: Rewriting loop exit condition to:\n"
<< " LHS:" << *CmpIndVar << '\n'
diff --git a/llvm/test/Transforms/IndVarSimplify/X86/eliminate-trunc.ll b/llvm/test/Transforms/IndVarSimplify/X86/eliminate-trunc.ll
index 7e7a3f192f998..565ac5c8743d4 100644
--- a/llvm/test/Transforms/IndVarSimplify/X86/eliminate-trunc.ll
+++ b/llvm/test/Transforms/IndVarSimplify/X86/eliminate-trunc.ll
@@ -227,7 +227,7 @@ define void @test_01_unsigned(i32 %n) {
; CHECK: loop:
; CHECK-NEXT: [[IV:%.*]] = phi i64 [ 0, [[ENTRY:%.*]] ], [ [[IV_NEXT:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
-; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc nuw nsw i64 [[IV_NEXT]] to i32
+; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i32 [[LFTR_WIDEIV]], [[TMP0]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[LOOP]], label [[EXIT:%.*]]
; CHECK: exit:
@@ -255,7 +255,7 @@ define void @test_02_unsigned(i32 %n) {
; CHECK: loop:
; CHECK-NEXT: [[IV:%.*]] = phi i64 [ 4294967294, [[ENTRY:%.*]] ], [ [[IV_NEXT:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
-; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc nuw nsw i64 [[IV_NEXT]] to i32
+; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i32 [[LFTR_WIDEIV]], [[TMP0]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[LOOP]], label [[EXIT:%.*]]
; CHECK: exit:
@@ -304,7 +304,7 @@ define void @test_04_unsigned(i32 %n) {
; CHECK: loop:
; CHECK-NEXT: [[IV:%.*]] = phi i64 [ 0, [[ENTRY:%.*]] ], [ [[IV_NEXT:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
-; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc nuw nsw i64 [[IV_NEXT]] to i32
+; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i32 [[LFTR_WIDEIV]], [[TMP0]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[LOOP]], label [[EXIT:%.*]]
; CHECK: exit:
@@ -332,7 +332,7 @@ define void @test_05_unsigned(i32 %n) {
; CHECK: loop:
; CHECK-NEXT: [[IV:%.*]] = phi i64 [ 1, [[ENTRY:%.*]] ], [ [[IV_NEXT:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
-; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc nuw nsw i64 [[IV_NEXT]] to i32
+; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i32 [[LFTR_WIDEIV]], [[TMP0]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[LOOP]], label [[EXIT:%.*]]
; CHECK: exit:
diff --git a/llvm/test/Transforms/IndVarSimplify/lftr-pr41998.ll b/llvm/test/Transforms/IndVarSimplify/lftr-pr41998.ll
index 376ef1ac5ffac..b7f4756b2757f 100644
--- a/llvm/test/Transforms/IndVarSimplify/lftr-pr41998.ll
+++ b/llvm/test/Transforms/IndVarSimplify/lftr-pr41998.ll
@@ -13,7 +13,7 @@ define void @test_int(i32 %start, ptr %p) {
; CHECK-NEXT: [[I2:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[I2_INC:%.*]], [[LOOP]] ]
; CHECK-NEXT: [[I2_INC]] = add nuw nsw i32 [[I2]], 1
; CHECK-NEXT: store volatile i32 [[I2_INC]], ptr [[P:%.*]], align 4
-; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc nuw nsw i32 [[I2_INC]] to i3
+; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i32 [[I2_INC]] to i3
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i3 [[LFTR_WIDEIV]], [[TMP1]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[END:%.*]], label [[LOOP]]
; CHECK: end:
diff --git a/llvm/test/Transforms/IndVarSimplify/lftr.ll b/llvm/test/Transforms/IndVarSimplify/lftr.ll
index cfa4baa2d3b11..5ee62ba357ab6 100644
--- a/llvm/test/Transforms/IndVarSimplify/lftr.ll
+++ b/llvm/test/Transforms/IndVarSimplify/lftr.ll
@@ -415,7 +415,7 @@ define void @wide_trip_count_test1(ptr %autoc,
; CHECK-NEXT: [[ADD3:%.*]] = fadd float [[TEMP2]], [[MUL]]
; CHECK-NEXT: store float [[ADD3]], ptr [[ARRAYIDX2]], align 4
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
-; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc nuw nsw i64 [[INDVARS_IV_NEXT]] to i32
+; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[INDVARS_IV_NEXT]] to i32
; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i32 [[LFTR_WIDEIV]], [[SUB]]
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_BODY]], label [[FOR_END_LOOPEXIT:%.*]]
; CHECK: for.end.loopexit:
diff --git a/llvm/test/Transforms/PhaseOrdering/AArch64/constraint-elimination-placement.ll b/llvm/test/Transforms/PhaseOrdering/AArch64/constraint-elimination-placement.ll
index ddd98f21c36a4..bbdbd95c6017a 100644
--- a/llvm/test/Transforms/PhaseOrdering/AArch64/constraint-elimination-placement.ll
+++ b/llvm/test/Transforms/PhaseOrdering/AArch64/constraint-elimination-placement.ll
@@ -33,7 +33,8 @@ define i1 @test_order_1(ptr %this, ptr noalias %other, i1 %tobool9.not, i32 %cal
; CHECK-NEXT: br i1 [[CMP44]], label [[FOR_BODY45]], label [[FOR_COND]]
; CHECK: for.inc57:
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nsw i64 [[INDVARS_IV]], 1
-; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i64 [[INDVARS_IV]], 0
+; CHECK-NEXT: [[TMP1:%.*]] = and i64 [[INDVARS_IV_NEXT]], 4294967295
+; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i64 [[TMP1]], 1
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_COND41_PREHEADER_PREHEADER]], label [[FOR_COND41_PREHEADER]]
; CHECK: exit:
; CHECK-NEXT: ret i1 false
>From 4578509dc7f49328664829f1ce227de470e47fe6 Mon Sep 17 00:00:00 2001
From: buggfg <3171290993 at qq.com>
Date: Tue, 12 Aug 2025 14:36:40 +0800
Subject: [PATCH 4/5] Solve the unnecessary masking issue
---
llvm/lib/Transforms/Scalar/IndVarSimplify.cpp | 57 ++++++++++++++++++-
.../constraint-elimination-placement.ll | 3 +-
2 files changed, 55 insertions(+), 5 deletions(-)
diff --git a/llvm/lib/Transforms/Scalar/IndVarSimplify.cpp b/llvm/lib/Transforms/Scalar/IndVarSimplify.cpp
index 334c911191cb8..df527ea2fd162 100644
--- a/llvm/lib/Transforms/Scalar/IndVarSimplify.cpp
+++ b/llvm/lib/Transforms/Scalar/IndVarSimplify.cpp
@@ -160,6 +160,8 @@ class IndVarSimplify {
bool sinkUnusedInvariants(Loop *L);
+ bool isIVInitValTargetType(Value *InitVal, unsigned TargetTypeSize);
+
public:
IndVarSimplify(LoopInfo *LI, ScalarEvolution *SE, DominatorTree *DT,
const DataLayout &DL, TargetLibraryInfo *TLI,
@@ -817,6 +819,37 @@ static bool isLoopCounter(PHINode* Phi, Loop *L,
isa<SCEVAddRecExpr>(SE->getSCEV(IncV)));
}
+/// Determine if the initial value of IV is within the range of TargetTypeSize.
+/// If it is a constant, check if it falls within the representable range. If
+/// not, verify whether the width of the initial type exceeds TargetTypeSize.
+bool IndVarSimplify::isIVInitValTargetType(Value *InitVal,
+ unsigned TargetTypeSize) {
+
+ unsigned InitTypeSize = SE->getTypeSizeInBits(InitVal->getType());
+ if (InitTypeSize <= TargetTypeSize)
+ return true;
+
+ // If the initial value of the IV is a constant, check whether it is within
+ // the range.
+ if (auto *CI = dyn_cast<ConstantInt>(InitVal)) {
+
+ APInt Value = CI->getValue();
+ APInt MinRange = APInt(TargetTypeSize, 1ULL << (TargetTypeSize - 1));
+ APInt MaxRange = APInt(TargetTypeSize, (1ULL << (TargetTypeSize - 1)) - 1);
+ // Check if the constant is within a valid range
+ return (Value.sle(MaxRange) && Value.sge(-MinRange));
+ }
+
+ // Check the initial type of the initial value.
+ if (auto *SextInst = dyn_cast<SExtInst>(InitVal)) {
+ return isIVInitValTargetType(SextInst->getOperand(0), TargetTypeSize);
+ }
+ if (auto *ZextInst = dyn_cast<ZExtInst>(InitVal)) {
+ return isIVInitValTargetType(ZextInst->getOperand(0), TargetTypeSize);
+ }
+ return false;
+}
+
/// Search the loop header for a loop counter (anadd rec w/step of one)
/// suitable for use by LFTR. If multiple counters are available, select the
/// "best" one based profitable heuristics.
@@ -1049,9 +1082,27 @@ linearFunctionTestReplace(Loop *L, BasicBlock *ExitingBB,
if (Extended) {
bool Discard;
L->makeLoopInvariant(ExitCnt, Discard);
- } else
- CmpIndVar = Builder.CreateTrunc(CmpIndVar, ExitCnt->getType(),
- "lftr.wideiv");
+ } else {
+ // The design of LFTR requires:
+ // 1. The IV to be a LoopCounter, ensuring a step of 1.
+ // 2. The ICmpInst::Predicate can only be eq or ne, meaning that
+ // ExitCnt must represent the final value of IV.
+ // Given these conditions, if the initial value 'start' of IV does not
+ // exceed ExitCntSize, then the range [start, end) of IV will stay
+ // within ExitCntSize. Consequently, the truncation will not result in
+ // signed or unsigned overflow.
+ // By legally adding the nsw/nuw flag to the Trunc instruction, avoid the
+ // unnecessary masking issue.
+ Value *InitVal = IndVar->getIncomingValueForBlock(L->getLoopPreheader());
+ if (isIVInitValTargetType(InitVal, ExitCntSize))
+ CmpIndVar = Builder.CreateTrunc(
+ CmpIndVar, ExitCnt->getType(), "lftr.wideiv",
+ cast<ICmpInst>(BI->getCondition())->isUnsigned(),
+ cast<ICmpInst>(BI->getCondition())->isSigned());
+ else
+ CmpIndVar =
+ Builder.CreateTrunc(CmpIndVar, ExitCnt->getType(), "lftr.wideiv");
+ }
}
LLVM_DEBUG(dbgs() << "INDVARS: Rewriting loop exit condition to:\n"
<< " LHS:" << *CmpIndVar << '\n'
diff --git a/llvm/test/Transforms/PhaseOrdering/AArch64/constraint-elimination-placement.ll b/llvm/test/Transforms/PhaseOrdering/AArch64/constraint-elimination-placement.ll
index bbdbd95c6017a..ddd98f21c36a4 100644
--- a/llvm/test/Transforms/PhaseOrdering/AArch64/constraint-elimination-placement.ll
+++ b/llvm/test/Transforms/PhaseOrdering/AArch64/constraint-elimination-placement.ll
@@ -33,8 +33,7 @@ define i1 @test_order_1(ptr %this, ptr noalias %other, i1 %tobool9.not, i32 %cal
; CHECK-NEXT: br i1 [[CMP44]], label [[FOR_BODY45]], label [[FOR_COND]]
; CHECK: for.inc57:
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nsw i64 [[INDVARS_IV]], 1
-; CHECK-NEXT: [[TMP1:%.*]] = and i64 [[INDVARS_IV_NEXT]], 4294967295
-; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i64 [[TMP1]], 1
+; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i64 [[INDVARS_IV]], 0
; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_COND41_PREHEADER_PREHEADER]], label [[FOR_COND41_PREHEADER]]
; CHECK: exit:
; CHECK-NEXT: ret i1 false
>From 23bd524ba9c436c145b8501e859e9fe07782130a Mon Sep 17 00:00:00 2001
From: buggfg <3171290993 at qq.com>
Date: Tue, 2 Sep 2025 19:05:17 +0800
Subject: [PATCH 5/5] Add a test
---
.../Transforms/IndVarSimplify/lftr-masking.ll | 45 +++++++++++++++++++
1 file changed, 45 insertions(+)
create mode 100644 llvm/test/Transforms/IndVarSimplify/lftr-masking.ll
diff --git a/llvm/test/Transforms/IndVarSimplify/lftr-masking.ll b/llvm/test/Transforms/IndVarSimplify/lftr-masking.ll
new file mode 100644
index 0000000000000..af66002f02dc6
--- /dev/null
+++ b/llvm/test/Transforms/IndVarSimplify/lftr-masking.ll
@@ -0,0 +1,45 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt < %s -passes='indvars' -S | FileCheck %s
+
+target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-i128:128-f80:128-n8:16:32:64-S128"
+
+define dso_local void @func(ptr noundef captures(none) %0, i32 noundef %1) local_unnamed_addr {
+; CHECK-LABEL: define dso_local void @func(ptr noundef captures(none) %0, i32 noundef %1) local_unnamed_addr {
+; CHECK-NEXT: [[CHECK:%.*]] = icmp slt i32 %1, 100
+; CHECK-NEXT: br i1 [[CHECK]], label [[LOOP_PREHEADER:%.*]], label [[LOOP_EXIT:%.*]]
+; CHECK: loop.preheader:
+; CHECK-NEXT: [[SEXT_START:%.*]] = sext i32 %1 to i64
+; CHECK-NEXT: br label [[LOOP_BODY:%.*]]
+; CHECK: loop.exit.loopexit:
+; CHECK-NEXT: br label [[LOOPEXIT:%.*]]
+; CHECK: loop.exit:
+; CHECK-NEXT: ret void
+; CHECK: loop.body:
+; CHECK-NEXT: [[IV:%.*]] = phi i64 [ [[POSTINC:%.*]], [[LOOP_BODY]] ], [ [[SEXT_START:%.*]], [[LOOP_PREHEADER]] ]
+; CHECK-NEXT: [[GEP:%.*]] = getelementptr inbounds i32, ptr %0, i64 [[IV]]
+; CHECK-NEXT: [[LOAD:%.*]] = load i32, ptr [[GEP]], align 4
+; CHECK-NEXT: [[DATA:%.*]] = add nsw i32 [[LOAD]], 1
+; CHECK-NEXT: store i32 [[DATA]], ptr [[GEP]], align 4
+; CHECK-NEXT: [[POSTINC:%.*]] = add nsw i64 [[IV]], 1
+; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc nsw i64 [[POSTINC]] to i32
+; CHECK-NEXT: [[EXITCOND:%.*]] = icmp ne i32 [[LFTR_WIDEIV]], 100
+; CHECK-NEXT: br i1 [[EXITCOND]], label [[LOOP_BODY]], label [[LOOP_EXIT_LOOPEXIT:%.*]]
+; CHECK-NEXT: }
+ %3 = icmp slt i32 %1, 100
+ br i1 %3, label %loop.preheader, label %loop.exit
+loop.preheader:
+ br label %loop.body
+
+loop.exit:
+ ret void
+loop.body:
+ %8 = phi i32 [ %13, %loop.body ], [ %1, %loop.preheader ]
+ %9 = sext i32 %8 to i64
+ %10 = getelementptr inbounds i32, ptr %0, i64 %9
+ %11 = load i32, ptr %10, align 4
+ %12 = add nsw i32 %11, 1
+ store i32 %12, ptr %10, align 4
+ %13 = add nsw i32 %8, 1
+ %14 = icmp slt i32 %8, 99
+ br i1 %14, label %loop.body, label %loop.exit
+}
More information about the llvm-commits
mailing list