[Polly][Fix] Dead memory accesses should not confuse RTC generation

[jdoerfert]

Summary

This fixes http://llvm.org/bugs/show_bug.cgi?id=21166 .

[simbuerg]

lib/Analysis/ScopInfo.cpp

1278

You are iterating over all memory accesses of one statement, so I assume MA->getStatement() is always equal to Stmt, right?
With that in mind, you can do the empty check outside this loop and skip the whole loop, if true.

[jdoerfert]

Thanks, will be updated.

lib/Analysis/ScopInfo.cpp

1278

100% correct... my bad.

[jdoerfert]

Closed by commit rL219131 (authored by @jdoerfert).

lib/Analysis/ScopInfo.cpp

bool Scop::buildAliasGroups(AliasAnalysis &AA) {
   AliasSetTracker AST(AA);
 
   DenseMap<Value *, MemoryAccess *> PtrToAcc;
   DenseSet<Value *> HasWriteAccess;
   for (ScopStmt *Stmt : *this) {
     for (MemoryAccess *MA : *Stmt) {
       if (MA->isScalar())
         continue;
+
+      // Skip accesses with an empty domain as they will never be executed.
+      isl_set *MADomain = MA->getStatement()->getDomain();
+      bool MADomainEmpty = isl_set_is_empty(MADomain);
+      isl_set_free(MADomain);
+      if (MADomainEmpty)
+        continue;
+
       if (!MA->isRead())
         HasWriteAccess.insert(MA->getBaseAddr());
       Instruction *Acc = MA->getAccessInstruction();
       PtrToAcc[getPointerOperand(*Acc)] = MA;
       AST.add(Acc);
     }
   }
 

test/ScopInfo/aliasing_dead_access.ll

+; RUN: opt %loadPolly -polly-codegen-scev -polly-code-generator=isl -analyze -polly-scops < %s | FileCheck %s
+;
+; Check that RTC generation does not die when accesses are dead.
+;
+; CHECK: Alias Groups (0):
+;
+;    void jd(int *A, int *B) {
+;      for (int i = 0; i < 1024; i++)
+;        for (int j = i; j < 0; j++)
+;          A[i] = B[i];
+;    }
+;
+target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
+
+define void @jd(i32* %A, i32* %B) {
+entry:
+  br label %for.cond
+
+for.cond:                                         ; preds = %for.inc6, %entry
+  %indvars.iv = phi i64 [ %indvars.iv.next, %for.inc6 ], [ 0, %entry ]
+  %exitcond = icmp ne i64 %indvars.iv, 1024
+  br i1 %exitcond, label %for.body, label %for.end8
+
+for.body:                                         ; preds = %for.cond
+  %tmp = trunc i64 %indvars.iv to i32
+  br label %for.cond1
+
+for.cond1:                                        ; preds = %for.inc, %for.body
+  %j.0 = phi i32 [ %tmp, %for.body ], [ %inc, %for.inc ]
+  %cmp2 = icmp slt i32 %j.0, 0
+  br i1 %cmp2, label %for.body3, label %for.end
+
+for.body3:                                        ; preds = %for.cond1
+  %arrayidx = getelementptr inbounds i32* %B, i64 %indvars.iv
+  %tmp1 = load i32* %arrayidx, align 4
+  %arrayidx5 = getelementptr inbounds i32* %A, i64 %indvars.iv
+  store i32 %tmp1, i32* %arrayidx5, align 4
+  br label %for.inc
+
+for.inc:                                          ; preds = %for.body3
+  %inc = add nsw i32 %j.0, 1
+  br label %for.cond1
+
+for.end:                                          ; preds = %for.cond1
+  br label %for.inc6
+
+for.inc6:                                         ; preds = %for.end
+  %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
+  br label %for.cond
+
+for.end8:                                         ; preds = %for.cond
+  ret void
+}