[llvm] 0aed0db - [LCSSA] Don't invalidate entire loop in SCEV

[Nikita Popov via llvm-commits]

Author: Nikita Popov
Date: 2023-04-28T12:17:26+02:00
New Revision: 0aed0dbec2434d9d1a41266be009e56cc964633b

URL: https://github.com/llvm/llvm-project/commit/0aed0dbec2434d9d1a41266be009e56cc964633b
DIFF: https://github.com/llvm/llvm-project/commit/0aed0dbec2434d9d1a41266be009e56cc964633b.diff

LOG: [LCSSA] Don't invalidate entire loop in SCEV

We already invalidate each individual instruction for which LCSSA
is formed in formLCSSAForInstructions(), so I don't see a reason
why we would need to invalidate the entire loop on top of that.

I believe we also no longer need the instruction-level invalidation
now that SCEV looks through LCSSA phis, but I'll leave that for a
separate patch, as it's less obvious.

Differential Revision: https://reviews.llvm.org/D149331

Added: 
    

Modified: 
    llvm/lib/Transforms/Utils/LCSSA.cpp

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Transforms/Utils/LCSSA.cpp b/llvm/lib/Transforms/Utils/LCSSA.cpp
index af79dc456ea6..4b006f5288dc 100644
--- a/llvm/lib/Transforms/Utils/LCSSA.cpp
+++ b/llvm/lib/Transforms/Utils/LCSSA.cpp
@@ -395,12 +395,6 @@ bool llvm::formLCSSA(Loop &L, const DominatorTree &DT, const LoopInfo *LI,
   IRBuilder<> Builder(L.getHeader()->getContext());
   Changed = formLCSSAForInstructions(Worklist, DT, *LI, SE, Builder);
 
-  // If we modified the code, remove any caches about the loop from SCEV to
-  // avoid dangling entries.
-  // FIXME: This is a big hammer, can we clear the cache more selectively?
-  if (SE && Changed)
-    SE->forgetLoop(&L);
-
   assert(L.isLCSSAForm(DT));
 
   return Changed;