[llvm-commits] [llvm] r106631 - /llvm/trunk/lib/Target/ARM/ARMISelLowering.cpp

[Jim Grosbach]

Author: grosbach
Date: Wed Jun 23 11:08:49 2010
New Revision: 106631

URL: http://llvm.org/viewvc/llvm-project?rev=106631&view=rev
Log:
When using libcall expansions for the atomic intrinsics, the explicit
MEMBARRIER fences aren't necessary for ARM. Tell the combiner to fold them
away.

Modified:
    llvm/trunk/lib/Target/ARM/ARMISelLowering.cpp

Modified: llvm/trunk/lib/Target/ARM/ARMISelLowering.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/ARM/ARMISelLowering.cpp?rev=106631&r1=106630&r2=106631&view=diff
==============================================================================
--- llvm/trunk/lib/Target/ARM/ARMISelLowering.cpp (original)
+++ llvm/trunk/lib/Target/ARM/ARMISelLowering.cpp Wed Jun 23 11:08:49 2010
@@ -441,6 +441,8 @@
     setOperationAction(ISD::ATOMIC_LOAD_NAND, MVT::i8,  Expand);
     setOperationAction(ISD::ATOMIC_LOAD_NAND, MVT::i16, Expand);
     setOperationAction(ISD::ATOMIC_LOAD_NAND, MVT::i32, Expand);
+    // Since the libcalls include locking, fold in the fences
+    setShouldFoldAtomicFences(true);
   }
   // 64-bit versions are always libcalls (for now)
   setOperationAction(ISD::ATOMIC_CMP_SWAP,  MVT::i64, Expand);