[llvm] 8ef3e89 - [X86] combineX86ShufflesRecursively - add TODO not to generate temporary nodes

Simon Pilgrim via llvm-commits llvm-commits at lists.llvm.org
Sun Feb 20 07:59:38 PST 2022


Author: Simon Pilgrim
Date: 2022-02-20T15:59:23Z
New Revision: 8ef3e895ad8ab1724e2b87cabad1dacdc7a397a3

URL: https://github.com/llvm/llvm-project/commit/8ef3e895ad8ab1724e2b87cabad1dacdc7a397a3
DIFF: https://github.com/llvm/llvm-project/commit/8ef3e895ad8ab1724e2b87cabad1dacdc7a397a3.diff

LOG: [X86] combineX86ShufflesRecursively - add TODO not to generate temporary nodes

Extension to PR45974, unless we actual combine the target shuffles we shouldn't be generating temporary nodes as they may interfere with the one use checks in the shuffle recursions

Added: 
    

Modified: 
    llvm/lib/Target/X86/X86ISelLowering.cpp

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Target/X86/X86ISelLowering.cpp b/llvm/lib/Target/X86/X86ISelLowering.cpp
index e2d37ee917f5e..3e9b51dbcaa0f 100644
--- a/llvm/lib/Target/X86/X86ISelLowering.cpp
+++ b/llvm/lib/Target/X86/X86ISelLowering.cpp
@@ -38704,6 +38704,8 @@ static SDValue combineX86ShufflesRecursively(
   // FIXME: should we rerun resolveTargetShuffleInputsAndMask() now?
 
   // Widen any subvector shuffle inputs we've collected.
+  // TODO: Remove this to avoid generating temporary nodes, we should only
+  // widen once combineX86ShuffleChain has found a match.
   if (any_of(Ops, [RootSizeInBits](SDValue Op) {
         return Op.getValueSizeInBits() < RootSizeInBits;
       })) {


        


More information about the llvm-commits mailing list