DAGCombiner: Fix broken size check in isAlias

[arsenm]

Summary

This should have been converting the size to bytes, but wasn't really.
These should probably all be using getStoreSize instead.

I haven't been able to come up with a meaningful testcase for this.
I can trigger it using combinations of struct loads and stores,
but can't observe a difference in non-broken testcases.

isAlias is only really used during store merging, so I'm not sure how
to get into the vector splitting situation the comment describes
since store merging is only done before type legalization.

[hfinkel]

LGTM

Can you open a PR against Clang here? We should warn on this, the term (the boolean, 1 or 0, shifted right by three) is always zero. I assume the general warning here is considering the boolean to be a byte in size for some reason.

[hfinkel]

In D20845#445129, @hfinkel wrote:

LGTM

Can you open a PR against Clang here? We should warn on this, the term (the boolean, 1 or 0, shifted right by three) is always zero. I assume the general warning here is considering the boolean to be a byte in size for some reason.

Specifically, I would have expected this warning:

/tmp/f.c:2:12: warning: shift count >= width of type [-Wshift-count-overflow]
  return a >> 48;
           ^  ~~
1 warning generated.

$ cat /tmp/f.c 
int test(int a) {
  return a >> 48;
}

[arsenm]

r271356

Filed https://llvm.org/bugs/show_bug.cgi?id=27960

lib/CodeGen/SelectionDAG/DAGCombiner.cpp

bool DAGCombiner::isAlias(LSBaseSDNode *Op0, LSBaseSDNode *Op1) const {
   // If we know required SrcValue1 and SrcValue2 have relatively large alignment
   // compared to the size and offset of the access, we may be able to prove they
   // do not alias.  This check is conservative for now to catch cases created by
   // splitting vector types.
   if ((Op0->getOriginalAlignment() == Op1->getOriginalAlignment()) &&
       (Op0->getSrcValueOffset() != Op1->getSrcValueOffset()) &&
       (Op0->getMemoryVT().getSizeInBits() >> 3 ==
        Op1->getMemoryVT().getSizeInBits() >> 3) &&
-      (Op0->getOriginalAlignment() > Op0->getMemoryVT().getSizeInBits()) >> 3) {
+      (Op0->getOriginalAlignment() > (Op0->getMemoryVT().getSizeInBits() >> 3))) {
     int64_t OffAlign1 = Op0->getSrcValueOffset() % Op0->getOriginalAlignment();
     int64_t OffAlign2 = Op1->getSrcValueOffset() % Op1->getOriginalAlignment();
 
     // There is no overlap between these relatively aligned accesses of similar
     // size, return no alias.
     if ((OffAlign1 + (Op0->getMemoryVT().getSizeInBits() >> 3)) <= OffAlign2 ||
         (OffAlign2 + (Op1->getMemoryVT().getSizeInBits() >> 3)) <= OffAlign1)
       return false;