fix an overly conservative caching issue that caused memdep to

cache a pointer as being unavailable due to phi trans in the
wrong place.  This would cause later queries to fail even when
they didn't involve phi trans.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@91787 91177308-0d34-0410-b5e6-96231b3b80d8

2 files changed, 40 insertions, 32 deletions

diff --git a/lib/Analysis/MemoryDependenceAnalysis.cpp b/lib/Analysis/MemoryDependenceAnalysis.cpp
index a0c77063d9..743420b5b8 100644
--- a/lib/Analysis/MemoryDependenceAnalysis.cpp
+++ b/lib/Analysis/MemoryDependenceAnalysis.cpp
@@ -894,36 +894,12 @@ getNonLocalPointerDepFromBB(const PHITransAddr &Pointer, uint64_t PointeeSize,
                                PredPtrVal);
         Result.push_back(Entry);
 
-        // Add it to the cache for this CacheKey so that subsequent queries get
-        // this result.
-        Cache = &NonLocalPointerDeps[CacheKey].second;
-        MemoryDependenceAnalysis::NonLocalDepInfo::iterator It =
-          std::upper_bound(Cache->begin(), Cache->end(), Entry);
-        
-        if (It != Cache->begin() && (It-1)->getBB() == Pred)
-          --It;
-
-        if (It == Cache->end() || It->getBB() != Pred) {
-          Cache->insert(It, Entry);
-          // Add it to the reverse map.
-          ReverseNonLocalPtrDeps[Pred->getTerminator()].insert(CacheKey);
-        } else if (!It->getResult().isDirty()) {
-          // noop
-        } else if (It->getResult().getInst() == Pred->getTerminator()) {
-          // Same instruction, clear the dirty marker.
-          It->setResult(Entry.getResult(), PredPtrVal);
-        } else if (It->getResult().getInst() == 0) {
-          // Dirty, with no instruction, just add this.
-          It->setResult(Entry.getResult(), PredPtrVal);
-          ReverseNonLocalPtrDeps[Pred->getTerminator()].insert(CacheKey);
-        } else {
-          // Otherwise, dirty with a different instruction.
-          RemoveFromReverseMap(ReverseNonLocalPtrDeps,
-                               It->getResult().getInst(), CacheKey);
-          It->setResult(Entry.getResult(),PredPtrVal);
-          ReverseNonLocalPtrDeps[Pred->getTerminator()].insert(CacheKey);
-        }
-        Cache = 0;
+        // Since we had a phi translation failure, the cache for CacheKey won't
+        // include all of the entries that we need to immediately satisfy future
+        // queries.  Mark this in NonLocalPointerDeps by setting the
+        // BBSkipFirstBlockPair pointer to null.  This requires reuse of the
+        // cached value to do more work but not miss the phi trans failure.
+        NonLocalPointerDeps[CacheKey].first = BBSkipFirstBlockPair();
         continue;
       }
 
@@ -961,10 +937,10 @@ getNonLocalPointerDepFromBB(const PHITransAddr &Pointer, uint64_t PointeeSize,
       NumSortedEntries = Cache->size();
     }
     
-    // Since we did phi translation, the "Cache" set won't contain all of the
+    // Since we failed 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
-    // results from the set"  Clear out the indicator for this.
+    // results from the set".  Clear out the indicator for this.
     CacheInfo->first = BBSkipFirstBlockPair();
     
     // If *nothing* works, mark the pointer as being clobbered by the first
diff --git a/test/Transforms/GVN/rle-phi-translate.ll b/test/Transforms/GVN/rle-phi-translate.ll
index 0a060fa0cf..6731f43c0d 100644
--- a/test/Transforms/GVN/rle-phi-translate.ll
+++ b/test/Transforms/GVN/rle-phi-translate.ll
@@ -112,3 +112,35 @@ bb2:
   ret i32 %dv
 }
 
+
+
+; void test5(int N, double* G) {
+;   for (long j = 1; j < 1000; j++)
+;     G[j] = G[j] + G[j-1];
+; }
+;
+; Should compile into one load in the loop.
+define void @test5(i32 %N, double* nocapture %G) nounwind ssp {
+; CHECK: @test5
+bb.nph:
+  br label %for.body
+
+for.body:
+  %indvar = phi i64 [ 0, %bb.nph ], [ %tmp, %for.body ]
+  %arrayidx6 = getelementptr double* %G, i64 %indvar
+  %tmp = add i64 %indvar, 1
+  %arrayidx = getelementptr double* %G, i64 %tmp
+  %tmp3 = load double* %arrayidx
+  %tmp7 = load double* %arrayidx6
+  %add = fadd double %tmp3, %tmp7
+  store double %add, double* %arrayidx
+  %exitcond = icmp eq i64 %tmp, 999
+  br i1 %exitcond, label %for.end, label %for.body
+; CHECK: for.body:
+; CHECK: phi double
+; CHECK: load double
+; CHECK-NOT: load double
+; CHECK: br i1
+for.end:
+  ret void
+}