[PATCH] D87538: [VectorCombine] Don't vectorize scalar load under asan/hwasan/memtag/tsan

[Fangrui Song via Phabricator via llvm-commits]

MaskRay marked 2 inline comments as done.
MaskRay added inline comments.


================
Comment at: llvm/lib/Transforms/Vectorize/VectorCombine.cpp:147
   // use an identity shuffle to shrink/grow the vector.
   if (Ty != MinVecTy) {
     unsigned OutputNumElts = Ty->getNumElements();
----------------
vitalybuka wrote:
> Shouldn't we disable only this case?
> 
This branch performs  (1) load widening followed by (2) widening/narrowing the result with shufflevector. Due to the first step (load widening), the whole optimization (D86160) is unsafe. 

 Added a specific test for it.


================
Comment at: llvm/lib/Transforms/Vectorize/VectorCombine.cpp:693
+  // concurrently modified.
+  bool CanWiden = !F.hasFnAttribute(Attribute::SanitizeAddress) &&
+                  !F.hasFnAttribute(Attribute::SanitizeThread);
----------------
vitalybuka wrote:
> It should apply to SanitizeHWAddress and SanitizeMemTag as well
> 
> 
Thanks! (I know little about hwasan/memtag. I will learn about them)


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D87538/new/

https://reviews.llvm.org/D87538