[llvm-branch-commits] [cfe-branch] r181919 - Merging r181728:

Bill Wendling isanbard at gmail.com
Wed May 15 14:06:54 PDT 2013


Author: void
Date: Wed May 15 16:06:54 2013
New Revision: 181919

URL: http://llvm.org/viewvc/llvm-project?rev=181919&view=rev
Log:
Merging r181728:
------------------------------------------------------------------------
r181728 | rafael | 2013-05-13 13:09:47 -0700 (Mon, 13 May 2013) | 6 lines

Use atomic instructions on ARM linux.

This is safe given how the pre-v6 atomic ops funcions in libgcc are
implemented.

This fixes pr15429.
------------------------------------------------------------------------

Added:
    cfe/branches/release_33/test/CodeGen/linux-arm-atomic.c
      - copied unchanged from r181728, cfe/trunk/test/CodeGen/linux-arm-atomic.c
Modified:
    cfe/branches/release_33/   (props changed)
    cfe/branches/release_33/lib/Basic/Targets.cpp

Propchange: cfe/branches/release_33/
------------------------------------------------------------------------------
--- svn:mergeinfo (original)
+++ svn:mergeinfo Wed May 15 16:06:54 2013
@@ -1,4 +1,4 @@
 /cfe/branches/type-system-rewrite:134693-134817
-/cfe/trunk:181286,181299,181368
+/cfe/trunk:181286,181299,181368,181728
 /cfe/trunk/test:170344
 /cfe/trunk/test/SemaTemplate:126920

Modified: cfe/branches/release_33/lib/Basic/Targets.cpp
URL: http://llvm.org/viewvc/llvm-project/cfe/branches/release_33/lib/Basic/Targets.cpp?rev=181919&r1=181918&r2=181919&view=diff
==============================================================================
--- cfe/branches/release_33/lib/Basic/Targets.cpp (original)
+++ cfe/branches/release_33/lib/Basic/Targets.cpp Wed May 15 16:06:54 2013
@@ -3527,6 +3527,20 @@ class ARMTargetInfo : public TargetInfo
 
   static const Builtin::Info BuiltinInfo[];
 
+  static bool shouldUseInlineAtomic(const llvm::Triple &T) {
+    // On linux, binaries targeting old cpus call functions in libgcc to
+    // perform atomic operations. The implementation in libgcc then calls into
+    // the kernel which on armv6 and newer uses ldrex and strex. The net result
+    // is that if we assume the kernel is at least as recent as the hardware,
+    // it is safe to use atomic instructions on armv6 and newer.
+    if (T.getOS() != llvm::Triple::Linux)
+     return false;
+    StringRef ArchName = T.getArchName();
+    if (ArchName.startswith("armv6") || ArchName.startswith("armv7"))
+      return true;
+    return false;
+  }
+
 public:
   ARMTargetInfo(const std::string &TripleStr)
     : TargetInfo(TripleStr), ABI("aapcs-linux"), CPU("arm1136j-s"), IsAAPCS(true)
@@ -3559,8 +3573,9 @@ public:
     TheCXXABI.set(TargetCXXABI::GenericARM);
 
     // ARM has atomics up to 8 bytes
-    // FIXME: Set MaxAtomicInlineWidth if we have the feature v6e
     MaxAtomicPromoteWidth = 64;
+    if (shouldUseInlineAtomic(getTriple()))
+      MaxAtomicInlineWidth = 64;
 
     // Do force alignment of members that follow zero length bitfields.  If
     // the alignment of the zero-length bitfield is greater than the member 





More information about the llvm-branch-commits mailing list