[msan] More precise instrumentation of select IR.

Some bits of select result may be initialized even if select condition
is not.

https://code.google.com/p/memory-sanitizer/issues/detail?id=50


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204716 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Evgeniy Stepanov
2014-03-25 13:08:34 +00:00
parent 3527e5fcb3
commit 800399636a
2 changed files with 74 additions and 34 deletions

View File

@ -1324,6 +1324,17 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {
// TODO: handle struct types. // TODO: handle struct types.
} }
/// \brief Cast an application value to the type of its own shadow.
Value *CreateAppToShadowCast(IRBuilder<> &IRB, Value *V) {
Type *ShadowTy = getShadowTy(V);
if (V->getType() == ShadowTy)
return V;
if (V->getType()->isPtrOrPtrVectorTy())
return IRB.CreatePtrToInt(V, ShadowTy);
else
return IRB.CreateBitCast(V, ShadowTy);
}
/// \brief Propagate shadow for arbitrary operation. /// \brief Propagate shadow for arbitrary operation.
void handleShadowOr(Instruction &I) { void handleShadowOr(Instruction &I) {
IRBuilder<> IRB(&I); IRBuilder<> IRB(&I);
@ -2180,40 +2191,51 @@ struct MemorySanitizerVisitor : public InstVisitor<MemorySanitizerVisitor> {
void visitSelectInst(SelectInst& I) { void visitSelectInst(SelectInst& I) {
IRBuilder<> IRB(&I); IRBuilder<> IRB(&I);
// a = select b, c, d // a = select b, c, d
Value *S = IRB.CreateSelect(I.getCondition(), getShadow(I.getTrueValue()), Value *B = I.getCondition();
getShadow(I.getFalseValue())); Value *C = I.getTrueValue();
Value *D = I.getFalseValue();
Value *Sb = getShadow(B);
Value *Sc = getShadow(C);
Value *Sd = getShadow(D);
// Result shadow if condition shadow is 0.
Value *Sa0 = IRB.CreateSelect(B, Sc, Sd);
Value *Sa1;
if (I.getType()->isAggregateType()) { if (I.getType()->isAggregateType()) {
// To avoid "sign extending" i1 to an arbitrary aggregate type, we just do // To avoid "sign extending" i1 to an arbitrary aggregate type, we just do
// an extra "select". This results in much more compact IR. // an extra "select". This results in much more compact IR.
// Sa = select Sb, poisoned, (select b, Sc, Sd) // Sa = select Sb, poisoned, (select b, Sc, Sd)
S = IRB.CreateSelect(getShadow(I.getCondition()), Sa1 = getPoisonedShadow(getShadowTy(I.getType()));
getPoisonedShadow(getShadowTy(I.getType())), S,
"_msprop_select_agg");
} else { } else {
// Sa = (sext Sb) | (select b, Sc, Sd) // Sa = select Sb, [ (c^d) | Sc | Sd ], [ b ? Sc : Sd ]
S = IRB.CreateOr(S, CreateShadowCast(IRB, getShadow(I.getCondition()), // If Sb (condition is poisoned), look for bits in c and d that are equal
S->getType(), true), // and both unpoisoned.
"_msprop_select"); // If !Sb (condition is unpoisoned), simply pick one of Sc and Sd.
// Cast arguments to shadow-compatible type.
C = CreateAppToShadowCast(IRB, C);
D = CreateAppToShadowCast(IRB, D);
// Result shadow if condition shadow is 1.
Sa1 = IRB.CreateOr(IRB.CreateXor(C, D), IRB.CreateOr(Sc, Sd));
} }
setShadow(&I, S); Value *Sa = IRB.CreateSelect(Sb, Sa1, Sa0, "_msprop_select");
setShadow(&I, Sa);
if (MS.TrackOrigins) { if (MS.TrackOrigins) {
// Origins are always i32, so any vector conditions must be flattened. // Origins are always i32, so any vector conditions must be flattened.
// FIXME: consider tracking vector origins for app vectors? // FIXME: consider tracking vector origins for app vectors?
Value *Cond = I.getCondition(); if (B->getType()->isVectorTy()) {
Value *CondShadow = getShadow(Cond); Type *FlatTy = getShadowTyNoVec(B->getType());
if (Cond->getType()->isVectorTy()) { B = IRB.CreateICmpNE(IRB.CreateBitCast(B, FlatTy),
Type *FlatTy = getShadowTyNoVec(Cond->getType());
Cond = IRB.CreateICmpNE(IRB.CreateBitCast(Cond, FlatTy),
ConstantInt::getNullValue(FlatTy)); ConstantInt::getNullValue(FlatTy));
CondShadow = IRB.CreateICmpNE(IRB.CreateBitCast(CondShadow, FlatTy), Sb = IRB.CreateICmpNE(IRB.CreateBitCast(Sb, FlatTy),
ConstantInt::getNullValue(FlatTy)); ConstantInt::getNullValue(FlatTy));
} }
// a = select b, c, d // a = select b, c, d
// Oa = Sb ? Ob : (b ? Oc : Od) // Oa = Sb ? Ob : (b ? Oc : Od)
setOrigin(&I, IRB.CreateSelect( setOrigin(&I, IRB.CreateSelect(
CondShadow, getOrigin(I.getCondition()), Sb, getOrigin(I.getCondition()),
IRB.CreateSelect(Cond, getOrigin(I.getTrueValue()), IRB.CreateSelect(B, getOrigin(C), getOrigin(D))));
getOrigin(I.getFalseValue()))));
} }
} }

View File

@ -249,13 +249,16 @@ entry:
} }
; CHECK: @Select ; CHECK: @Select
; CHECK: select ; CHECK: select i1
; CHECK-NEXT: sext i1 {{.*}} to i32
; CHECK-NEXT: or i32 ; CHECK-NEXT: or i32
; CHECK-NEXT: select ; CHECK-NEXT: xor i32
; CHECK-NEXT: or i32
; CHECK-NEXT: select i1
; CHECK-ORIGINS: select ; CHECK-ORIGINS: select
; CHECK-ORIGINS: select ; CHECK-ORIGINS: select
; CHECK: select ; CHECK-NEXT: select i1
; CHECK: store i32{{.*}}@__msan_retval_tls
; CHECK-ORIGINS: store i32{{.*}}@__msan_retval_origin_tls
; CHECK: ret i32 ; CHECK: ret i32
@ -271,19 +274,18 @@ entry:
; CHECK: @SelectVector ; CHECK: @SelectVector
; CHECK: select <8 x i1> ; CHECK: select <8 x i1>
; CHECK-NEXT: sext <8 x i1> {{.*}} to <8 x i16>
; CHECK-NEXT: or <8 x i16> ; CHECK-NEXT: or <8 x i16>
; CHECK-ORIGINS: bitcast <8 x i1> {{.*}} to i8 ; CHECK-NEXT: xor <8 x i16>
; CHECK-ORIGINS: icmp ne i8 {{.*}}, 0 ; CHECK-NEXT: or <8 x i16>
; CHECK-ORIGINS: bitcast <8 x i1> {{.*}} to i8 ; CHECK-NEXT: select <8 x i1>
; CHECK-ORIGINS: icmp ne i8 {{.*}}, 0 ; CHECK-ORIGINS: select
; CHECK-ORIGINS: select i1 ; CHECK-ORIGINS: select
; CHECK-ORIGINS: select i1 ; CHECK-NEXT: select <8 x i1>
; CHECK: select <8 x i1> ; CHECK: store <8 x i16>{{.*}}@__msan_retval_tls
; CHECK-ORIGINS: store i32{{.*}}@__msan_retval_origin_tls
; CHECK: ret <8 x i16> ; CHECK: ret <8 x i16>
; Check that we propagate origin for "select" with scalar condition and vector ; Check that we propagate origin for "select" with scalar condition and vector
; arguments. Select condition shadow is sign-extended to the vector type and ; arguments. Select condition shadow is sign-extended to the vector type and
; mixed into the result shadow. ; mixed into the result shadow.
@ -296,9 +298,10 @@ entry:
; CHECK: @SelectVector2 ; CHECK: @SelectVector2
; CHECK: select i1 ; CHECK: select i1
; CHECK: sext i1 {{.*}} to i128
; CHECK: bitcast i128 {{.*}} to <8 x i16>
; CHECK: or <8 x i16> ; CHECK: or <8 x i16>
; CHECK: xor <8 x i16>
; CHECK: or <8 x i16>
; CHECK: select i1
; CHECK-ORIGINS: select i1 ; CHECK-ORIGINS: select i1
; CHECK-ORIGINS: select i1 ; CHECK-ORIGINS: select i1
; CHECK: select i1 ; CHECK: select i1
@ -320,6 +323,21 @@ entry:
; CHECK: ret { i64, i64 } ; CHECK: ret { i64, i64 }
define { i64*, double } @SelectStruct2(i1 zeroext %x, { i64*, double } %a, { i64*, double } %b) readnone sanitize_memory {
entry:
%c = select i1 %x, { i64*, double } %a, { i64*, double } %b
ret { i64*, double } %c
}
; CHECK: @SelectStruct2
; CHECK: select i1 {{.*}}, { i64, i64 }
; CHECK-NEXT: select i1 {{.*}}, { i64, i64 } { i64 -1, i64 -1 }, { i64, i64 }
; CHECK-ORIGINS: select i1
; CHECK-ORIGINS: select i1
; CHECK-NEXT: select i1 {{.*}}, { i64*, double }
; CHECK: ret { i64*, double }
define i8* @IntToPtr(i64 %x) nounwind uwtable readnone sanitize_memory { define i8* @IntToPtr(i64 %x) nounwind uwtable readnone sanitize_memory {
entry: entry:
%0 = inttoptr i64 %x to i8* %0 = inttoptr i64 %x to i8*