maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity

Fix PR18165: LSR must avoid scaling factors that exceed the limit on truncated use. 

Patch by Michael Zolotukhin!

llvm-svn: 202273
This commit is contained in:
Andrew Trick 2014-02-26 16:31:56 +00:00
parent f4733b159d
commit 429e9edd08
2 changed files with 100 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
llvm/lib/Transforms/Scalar/LoopStrengthReduce.cpp
View File

@ -3389,6 +3389,10 @@ void LSRInstance::GenerateICmpZeroScales(LSRUse &LU, unsigned LUIdx,
int64_t NewBaseOffset = (uint64_t)Base.BaseOffset * Factor;
if (NewBaseOffset / Factor != Base.BaseOffset)
continue;
// If the offset will be truncated at this use, check that it is in bounds.
if (!IntTy->isPointerTy() &&
!ConstantInt::isValueValidForType(IntTy, NewBaseOffset))
continue;
// Check that multiplying with the use offset doesn't overflow.
int64_t Offset = LU.MinOffset;
@ -3397,6 +3401,10 @@ void LSRInstance::GenerateICmpZeroScales(LSRUse &LU, unsigned LUIdx,
Offset = (uint64_t)Offset * Factor;
if (Offset / Factor != LU.MinOffset)
continue;
// If the offset will be truncated at this use, check that it is in bounds.
if (!IntTy->isPointerTy() &&
!ConstantInt::isValueValidForType(IntTy, Offset))
continue;
Formula F = Base;
F.BaseOffset = NewBaseOffset;
@ -3431,6 +3439,10 @@ void LSRInstance::GenerateICmpZeroScales(LSRUse &LU, unsigned LUIdx,
F.UnfoldedOffset = (uint64_t)F.UnfoldedOffset * Factor;
if (F.UnfoldedOffset / Factor != Base.UnfoldedOffset)
continue;
// If the offset will be truncated, check that it is in bounds.
if (!IntTy->isPointerTy() &&
!ConstantInt::isValueValidForType(IntTy, F.UnfoldedOffset))
continue;
}
// If we make it here and it's legal, add it.

88
llvm/test/Transforms/LoopStrengthReduce/pr18165.ll Normal file
View File

@ -0,0 +1,88 @@
; RUN: opt < %s -loop-reduce -S | FileCheck %s
target datalayout = "e-m:o-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-apple-macosx10.9.0"
; LSR shouldn't reuse IV if the resultant offset is not valid for the operand type.
; CHECK-NOT: trunc i32 %.ph to i8
%struct.anon = type { i32, i32, i32 }
@c = global i32 1, align 4
@.str = private unnamed_addr constant [4 x i8] c"%d\0A\00", align 1
@b = common global i32 0, align 4
@a = common global %struct.anon zeroinitializer, align 4
@e = common global %struct.anon zeroinitializer, align 4
@d = common global i32 0, align 4
@f = common global i32 0, align 4
@g = common global i32 0, align 4
@h = common global i32 0, align 4
; Function Attrs: nounwind optsize ssp uwtable
define i32 @main() #0 {
entry:
%0 = load i32* getelementptr inbounds (%struct.anon* @a, i64 0, i32 0), align 4, !tbaa !1
%tobool7.i = icmp eq i32 %0, 0
%.promoted.i = load i32* getelementptr inbounds (%struct.anon* @a, i64 0, i32 2), align 4, !tbaa !6
%f.promoted.i = load i32* @f, align 4, !tbaa !7
br label %for.body6.i.outer
for.body6.i.outer: ; preds = %entry, %lor.end.i
%.ph = phi i32 [ %add.i, %lor.end.i ], [ 0, %entry ]
%or1512.i.ph = phi i32 [ %or15.i, %lor.end.i ], [ %f.promoted.i, %entry ]
%or1410.i.ph = phi i32 [ %or14.i, %lor.end.i ], [ %.promoted.i, %entry ]
%p.addr.16.i.ph = phi i8 [ %inc10.i, %lor.end.i ], [ -128, %entry ]
br i1 %tobool7.i, label %if.end9.i, label %lbl.loopexit.i
lbl.loopexit.i: ; preds = %for.body6.i.outer, %lbl.loopexit.i
br label %lbl.loopexit.i
if.end9.i: ; preds = %for.body6.i.outer
%inc10.i = add i8 %p.addr.16.i.ph, 1
%tobool12.i = icmp eq i8 %p.addr.16.i.ph, 0
br i1 %tobool12.i, label %lor.rhs.i, label %lor.end.i
lor.rhs.i: ; preds = %if.end9.i
%1 = load i32* @b, align 4, !tbaa !7
%dec.i = add nsw i32 %1, -1
store i32 %dec.i, i32* @b, align 4, !tbaa !7
%tobool13.i = icmp ne i32 %1, 0
br label %lor.end.i
lor.end.i: ; preds = %lor.rhs.i, %if.end9.i
%2 = phi i1 [ true, %if.end9.i ], [ %tobool13.i, %lor.rhs.i ]
%lor.ext.i = zext i1 %2 to i32
%or14.i = or i32 %lor.ext.i, %or1410.i.ph
%or15.i = or i32 %or14.i, %or1512.i.ph
%add.i = add nsw i32 %.ph, 2
%cmp.i = icmp slt i32 %add.i, 21
br i1 %cmp.i, label %for.body6.i.outer, label %fn1.exit
fn1.exit: ; preds = %lor.end.i
store i32 0, i32* @g, align 4, !tbaa !7
store i32 %or14.i, i32* getelementptr inbounds (%struct.anon* @a, i64 0, i32 2), align 4, !tbaa !6
store i32 %or15.i, i32* @f, align 4, !tbaa !7
store i32 %add.i, i32* getelementptr inbounds (%struct.anon* @e, i64 0, i32 1), align 4, !tbaa !8
store i32 0, i32* @h, align 4, !tbaa !7
%3 = load i32* @b, align 4, !tbaa !7
%call1 = tail call i32 (i8*, ...)* @printf(i8* getelementptr inbounds ([4 x i8]* @.str, i64 0, i64 0), i32 %3) #2
ret i32 0
}
; Function Attrs: nounwind optsize
declare i32 @printf(i8* nocapture readonly, ...) #1
attributes #0 = { nounwind optsize ssp uwtable "less-precise-fpmad"="false" "no-frame-pointer-elim"="true" "no-frame-pointer-elim-non-leaf" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
attributes #1 = { nounwind optsize "less-precise-fpmad"="false" "no-frame-pointer-elim"="true" "no-frame-pointer-elim-non-leaf" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
attributes #2 = { nounwind optsize }
!llvm.ident = !{!0}
!0 = metadata !{metadata !"clang version 3.5 "}
!1 = metadata !{metadata !2, metadata !3, i64 0}
!2 = metadata !{metadata !"", metadata !3, i64 0, metadata !3, i64 4, metadata !3, i64 8}
!3 = metadata !{metadata !"int", metadata !4, i64 0}
!4 = metadata !{metadata !"omnipotent char", metadata !5, i64 0}
!5 = metadata !{metadata !"Simple C/C++ TBAA"}
!6 = metadata !{metadata !2, metadata !3, i64 8}
!7 = metadata !{metadata !3, metadata !3, i64 0}
!8 = metadata !{metadata !2, metadata !3, i64 4}