[llvm-commits] [llvm] r63201 - /llvm/branches/release_25/lib/Analysis/MemoryDependenceAnalysis.cpp

Tanya Lattner tonic at nondot.org
Wed Jan 28 07:22:28 PST 2009


Author: tbrethou
Date: Wed Jan 28 09:22:24 2009
New Revision: 63201

URL: http://llvm.org/viewvc/llvm-project?rev=63201&view=rev
Log:
a minor tweak to my previous patch, handle the invalidation case
when there are multiple iterations of the loop.  This fixes PR3375.


Modified:
    llvm/branches/release_25/lib/Analysis/MemoryDependenceAnalysis.cpp

Modified: llvm/branches/release_25/lib/Analysis/MemoryDependenceAnalysis.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/branches/release_25/lib/Analysis/MemoryDependenceAnalysis.cpp?rev=63201&r1=63200&r2=63201&view=diff

==============================================================================
--- llvm/branches/release_25/lib/Analysis/MemoryDependenceAnalysis.cpp (original)
+++ llvm/branches/release_25/lib/Analysis/MemoryDependenceAnalysis.cpp Wed Jan 28 09:22:24 2009
@@ -751,9 +751,10 @@
         // cache remains sorted.  Sort it now (if needed) so that recursive
         // invocations of getNonLocalPointerDepFromBB that could reuse the cache
         // value will only see properly sorted cache arrays.
-        if (NumSortedEntries != Cache->size()) {
+        if (Cache && NumSortedEntries != Cache->size()) {
           std::sort(Cache->begin(), Cache->end());
           NumSortedEntries = Cache->size();
+          Cache = 0;
         }
         
         // FIXME: it is entirely possible that PHI translating will end up with
@@ -767,12 +768,12 @@
                                         Result, Visited))
           goto PredTranslationFailure;
       }
-      
+
       // Refresh the CacheInfo/Cache pointer so that it isn't invalidated.
       CacheInfo = &NonLocalPointerDeps[CacheKey];
       Cache = &CacheInfo->second;
       NumSortedEntries = Cache->size();
-
+      
       // Since we did phi translation, the "Cache" set won't contain all of the
       // results for the query.  This is ok (we can still use it to accelerate
       // specific block queries) but we can't do the fastpath "return all





More information about the llvm-commits mailing list