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

Fix for PR20059 (instcombine reorders shufflevector after instruction that may trap) 

In PR20059 ( http://llvm.org/pr20059 ), instcombine eliminates shuffles that are necessary before performing an operation that can trap (srem).

This patch calls isSafeToSpeculativelyExecute() and bails out of the optimization in SimplifyVectorOp() if needed.

Differential Revision: http://reviews.llvm.org/D4424

llvm-svn: 212629
This commit is contained in:
Sanjay Patel 2014-07-09 16:34:54 +00:00
parent df734cdd39
commit 58814445d4
2 changed files with 38 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
llvm/lib/Transforms/InstCombine/InstructionCombining.cpp
View File

@ -42,6 +42,7 @@
#include "llvm/Analysis/ConstantFolding.h"
#include "llvm/Analysis/InstructionSimplify.h"
#include "llvm/Analysis/MemoryBuiltins.h"
#include "llvm/Analysis/ValueTracking.h"
#include "llvm/IR/CFG.h"
#include "llvm/IR/DataLayout.h"
#include "llvm/IR/GetElementPtrTypeIterator.h"
@ -1195,6 +1196,11 @@ static Value *CreateBinOpAsGiven(BinaryOperator &Inst, Value *LHS, Value *RHS,
Value *InstCombiner::SimplifyVectorOp(BinaryOperator &Inst) {
if (!Inst.getType()->isVectorTy()) return nullptr;
// It may not be safe to reorder shuffles and things like div, urem, etc.
// because we may trap when executing those ops on unknown vector elements.
// See PR20059.
if (!isSafeToSpeculativelyExecute(&Inst)) return nullptr;
unsigned VWidth = cast<VectorType>(Inst.getType())->getNumElements();
Value *LHS = Inst.getOperand(0), *RHS = Inst.getOperand(1);
assert(cast<VectorType>(LHS->getType())->getNumElements() == VWidth);

32
llvm/test/Transforms/InstCombine/pr20059.ll Normal file
View File

@ -0,0 +1,32 @@
; RUN: opt -S -instcombine < %s | FileCheck %s
; In PR20059 ( http://llvm.org/pr20059 ), shufflevector operations are reordered/removed
; for an srem operation. This is not a valid optimization because it may cause a trap
; on div-by-zero.
; CHECK-LABEL: @do_not_reorder
; CHECK: %splat1 = shufflevector <4 x i32> %p1, <4 x i32> undef, <4 x i32> zeroinitializer
; CHECK-NEXT: %splat2 = shufflevector <4 x i32> %p2, <4 x i32> undef, <4 x i32> zeroinitializer
; CHECK-NEXT: %retval = srem <4 x i32> %splat1, %splat2
define <4 x i32> @do_not_reorder(<4 x i32> %p1, <4 x i32> %p2) {
%splat1 = shufflevector <4 x i32> %p1, <4 x i32> undef, <4 x i32> zeroinitializer
%splat2 = shufflevector <4 x i32> %p2, <4 x i32> undef, <4 x i32> zeroinitializer
%retval = srem <4 x i32> %splat1, %splat2
ret <4 x i32> %retval
}
; RUN: opt -S -instcombine < %s | FileCheck %s
; In PR20059 ( http://llvm.org/pr20059 ), shufflevector operations are reordered/removed
; for an srem operation. This is not a valid optimization because it may cause a trap
; on div-by-zero.
; CHECK-LABEL: @do_not_reorder
; CHECK: %splat1 = shufflevector <4 x i32> %p1, <4 x i32> undef, <4 x i32> zeroinitializer
; CHECK-NEXT: %splat2 = shufflevector <4 x i32> %p2, <4 x i32> undef, <4 x i32> zeroinitializer
; CHECK-NEXT: %retval = srem <4 x i32> %splat1, %splat2
define <4 x i32> @do_not_reorder(<4 x i32> %p1, <4 x i32> %p2) {
%splat1 = shufflevector <4 x i32> %p1, <4 x i32> undef, <4 x i32> zeroinitializer
%splat2 = shufflevector <4 x i32> %p2, <4 x i32> undef, <4 x i32> zeroinitializer
%retval = srem <4 x i32> %splat1, %splat2
ret <4 x i32> %retval
}