New approach to r136737: insert the necessary fences for atomic ops in platform-independent code, since a bunch of platforms (ARM, Mips, PPC, Alpha are the relevant targets here) need to do essentially the same thing.
I think this completes the basic CodeGen for atomicrmw and cmpxchg.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@136813 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/lib/Target/Alpha/AlphaISelLowering.cpp b/lib/Target/Alpha/AlphaISelLowering.cpp
index de003fb..8fb6340 100644
--- a/lib/Target/Alpha/AlphaISelLowering.cpp
+++ b/lib/Target/Alpha/AlphaISelLowering.cpp
@@ -160,6 +160,8 @@
setMinFunctionAlignment(4);
+ setInsertFencesForAtomic(true);
+
computeRegisterProperties();
}