[llvm] r219868 - R600: Remove unnecessary part of computeKnownBitsForTargetNode

[Matt Arsenault]

Author: arsenm
Date: Wed Oct 15 18:37:49 2014
New Revision: 219868

URL: http://llvm.org/viewvc/llvm-project?rev=219868&view=rev
Log:
R600: Remove unnecessary part of computeKnownBitsForTargetNode

Zero-width BFEs are combined away already, so there's no point in
handling them.

Modified:
    llvm/trunk/lib/Target/R600/AMDGPUISelLowering.cpp

Modified: llvm/trunk/lib/Target/R600/AMDGPUISelLowering.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/R600/AMDGPUISelLowering.cpp?rev=219868&r1=219867&r2=219868&view=diff
==============================================================================
--- llvm/trunk/lib/Target/R600/AMDGPUISelLowering.cpp (original)
+++ llvm/trunk/lib/Target/R600/AMDGPUISelLowering.cpp Wed Oct 15 18:37:49 2014
@@ -2388,11 +2388,6 @@ void AMDGPUTargetLowering::computeKnownB
 
     unsigned BitWidth = 32;
     uint32_t Width = CWidth->getZExtValue() & 0x1f;
-    if (Width == 0) {
-      KnownZero = APInt::getAllOnesValue(BitWidth);
-      KnownOne = APInt::getNullValue(BitWidth);
-      return;
-    }
 
     // FIXME: This could do a lot more. If offset is 0, should be the same as
     // sign_extend_inreg implementation, but that involves duplicating it.