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

Fix a horrible infloop in value tracking in the face of dead code. 

Amazingly, we just never triggered this without:
1) Moving code around for MetadataTracking so that a certain *different*
   amount of inlining occurs in the per-TU compile step.
2) Then you LTO opt or clang with a bootstrap, and get inlining, loop
   opts, and GVN line up everything *just* right.

I don't really know how we didn't hit this before. We really need to be
fuzz testing stuff, it shouldn't be hard to trigger. I'm working on
crafting a reduced nice test case, and will submit that when I have it,
but I want to get LTO build bots going again.

llvm-svn: 256735
This commit is contained in:
Chandler Carruth 2016-01-04 07:23:12 +00:00
parent ec9a589598
commit 7664127f8c
1 changed files with 6 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
llvm/lib/Analysis/ValueTracking.cpp
View File

@ -2830,7 +2830,12 @@ Value *llvm::GetPointerBaseWithConstantOffset(Value *Ptr, int64_t &Offset,
const DataLayout &DL) { const DataLayout &DL) {
unsigned BitWidth = DL.getPointerTypeSizeInBits(Ptr->getType()); unsigned BitWidth = DL.getPointerTypeSizeInBits(Ptr->getType());
APInt ByteOffset(BitWidth, 0); APInt ByteOffset(BitWidth, 0);
while (1) {
// We walk up the defs but use a visited set to handle unreachable code. In
// that case, we stop after accumulating the cycle once (not that it
// matters).
SmallPtrSet<Value *, 16> Visited;
while (Visited.insert(Ptr).second) {
if (Ptr->getType()->isVectorTy()) if (Ptr->getType()->isVectorTy())
break; break;