[llvm] r218755 - [x86] Delete some extraneous logic from the new vector shuffle lowering.
Chandler Carruth
chandlerc at gmail.com
Wed Oct 1 04:13:57 PDT 2014
Author: chandlerc
Date: Wed Oct 1 06:13:57 2014
New Revision: 218755
URL: http://llvm.org/viewvc/llvm-project?rev=218755&view=rev
Log:
[x86] Delete some extraneous logic from the new vector shuffle lowering.
Nothing was relying on this and there are potentially some edge cases
that it would not be correct under. Removing it seems better than trying
to "fix" it as nothing was relying on it.
Modified:
llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
Modified: llvm/trunk/lib/Target/X86/X86ISelLowering.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/X86/X86ISelLowering.cpp?rev=218755&r1=218754&r2=218755&view=diff
==============================================================================
--- llvm/trunk/lib/Target/X86/X86ISelLowering.cpp (original)
+++ llvm/trunk/lib/Target/X86/X86ISelLowering.cpp Wed Oct 1 06:13:57 2014
@@ -7768,13 +7768,6 @@ static SDValue lowerVectorShuffleAsEleme
return SDValue(); // Not inserting into a zero vector.
}
- // Step over any bitcasts on either input so we can scan the actual
- // BUILD_VECTOR nodes.
- while (V1.getOpcode() == ISD::BITCAST)
- V1 = V1.getOperand(0);
- while (V2.getOpcode() == ISD::BITCAST)
- V2 = V2.getOperand(0);
-
// Check for a single input from a SCALAR_TO_VECTOR node.
// FIXME: All of this should be canonicalized into INSERT_VECTOR_ELT and
// all the smarts here sunk into that routine. However, the current
More information about the llvm-commits
mailing list