From: Tom Stellard <thomas.stellard@amd.com>
Date: Thu, 13 Jun 2013 20:14:00 +0000 (+0000)
Subject: R600: Don't try to fix reg class when copying IMPLICIT_DEF to a register
X-Git-Url: http://demsky.eecs.uci.edu/git/?a=commitdiff_plain;h=2948e693cb9ed0f99e8ba2448459d8a34f091075;p=oota-llvm.git

R600: Don't try to fix reg class when copying IMPLICIT_DEF to a register

The test case for this is way too complex to be useful as a lit test,
and I was unable to reduce it.

https://bugs.freedesktop.org/show_bug.cgi?id=65438

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@183937 91177308-0d34-0410-b5e6-96231b3b80d8
---

diff --git a/lib/Target/R600/AMDILISelDAGToDAG.cpp b/lib/Target/R600/AMDILISelDAGToDAG.cpp
index 93432a2c97a..9a5e9e6147a 100644
--- a/lib/Target/R600/AMDILISelDAGToDAG.cpp
+++ b/lib/Target/R600/AMDILISelDAGToDAG.cpp
@@ -766,7 +766,8 @@ void AMDGPUDAGToDAGISel::PostprocessISelDAG() {
         continue;
       }
 
-      if (!Val.getNode()->isMachineOpcode()) {
+      if (!Val.getNode()->isMachineOpcode() ||
+          Val.getNode()->getMachineOpcode() == AMDGPU::IMPLICIT_DEF) {
         continue;
       }