[SimplifyCFG] propagate branch metadata when creating select (retry r268550 with possible fix)

Retrying r268550 which was reverted at r268577 due a memory sanitizer failure.
I have not been able to reproduce that failure, but I've taken a guess at fixing
the problem in this version of the patch and will watch for another failure.

Original commit message:
Unlike earlier similar fixes, we need to recalculate the branch weights
in this case.

Differential Revision: http://reviews.llvm.org/D19674



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@268751 91177308-0d34-0410-b5e6-96231b3b80d8

Author: rotateright

lib/Transforms/Utils/SimplifyCFG.cpp

@@ -848,7 +848,7 @@ static void GetBranchWeights(TerminatorInst *TI,
   }
 }
 
-/// Keep halving the weights until all can fit in uint32_t.
+/// Scale each weight so they all fit in uint32_t.
 static void FitWeights(MutableArrayRef<uint64_t> Weights) {
   uint64_t Max = *std::max_element(Weights.begin(), Weights.end());
   if (Max > UINT_MAX) {
@@ -2820,28 +2820,28 @@ static bool SimplifyCondBranchToCondBranch(BranchInst *PBI, BranchInst *BI,
   PBI->setSuccessor(1, OtherDest);
 
   // Update branch weight for PBI.
+  MDBuilder MDB(BI->getContext());
   uint64_t PredTrueWeight, PredFalseWeight, SuccTrueWeight, SuccFalseWeight;
+  uint64_t PredCommon, PredOther, SuccCommon, SuccOther;
   bool PredHasWeights =
       PBI->extractProfMetadata(PredTrueWeight, PredFalseWeight);
   bool SuccHasWeights =
       BI->extractProfMetadata(SuccTrueWeight, SuccFalseWeight);
   if (PredHasWeights && SuccHasWeights) {
-    uint64_t PredCommon = PBIOp ? PredFalseWeight : PredTrueWeight;
-    uint64_t PredOther = PBIOp ?PredTrueWeight : PredFalseWeight;
-    uint64_t SuccCommon = BIOp ? SuccFalseWeight : SuccTrueWeight;
-    uint64_t SuccOther = BIOp ? SuccTrueWeight : SuccFalseWeight;
+    PredCommon = PBIOp ? PredFalseWeight : PredTrueWeight;
+    PredOther = PBIOp ? PredTrueWeight : PredFalseWeight;
+    SuccCommon = BIOp ? SuccFalseWeight : SuccTrueWeight;
+    SuccOther = BIOp ? SuccTrueWeight : SuccFalseWeight;
     // The weight to CommonDest should be PredCommon * SuccTotal +
     //                                    PredOther * SuccCommon.
     // The weight to OtherDest should be PredOther * SuccOther.
     uint64_t NewWeights[2] = {PredCommon * (SuccCommon + SuccOther) +
                                   PredOther * SuccCommon,
                               PredOther * SuccOther};
-    // Halve the weights if any of them cannot fit in an uint32_t
     FitWeights(NewWeights);
 
     PBI->setMetadata(LLVMContext::MD_prof,
-                     MDBuilder(BI->getContext())
-                         .createBranchWeights(NewWeights[0], NewWeights[1]));
+                     MDB.createBranchWeights(NewWeights[0], NewWeights[1]));
   }
 
   // OtherDest may have phi nodes.  If so, add an entry from PBI's
@@ -2860,9 +2860,24 @@ static bool SimplifyCondBranchToCondBranch(BranchInst *PBI, BranchInst *BI,
     Value *PBIV = PN->getIncomingValue(PBBIdx);
     if (BIV != PBIV) {
       // Insert a select in PBI to pick the right value.
-      Value *NV = cast<SelectInst>
+      SelectInst *NV = cast<SelectInst>
         (Builder.CreateSelect(PBICond, PBIV, BIV, PBIV->getName() + ".mux"));
       PN->setIncomingValue(PBBIdx, NV);
+      // Although the select has the same condition as PBI, the original branch
+      // weights for PBI do not apply to the new select because the select's
+      // 'logical' edges are incoming edges of the phi that is eliminated, not
+      // the outgoing edges of PBI.
+      if (PredHasWeights && SuccHasWeights) {
+        // The weight to PredCommonDest should be PredCommon * SuccTotal.
+        // The weight to PredOtherDest should be PredOther * SuccCommon.
+        uint64_t NewWeights[2] = {PredCommon * (SuccCommon + SuccOther),
+                                  PredOther * SuccCommon};
+
+        FitWeights(NewWeights);
+
+        NV->setMetadata(LLVMContext::MD_prof,
+                        MDB.createBranchWeights(NewWeights[0], NewWeights[1]));
+      }
     }
   }
 

test/Transforms/SimplifyCFG/preserve-branchweights.ll

@@ -412,22 +412,48 @@ return:
   ret i32 %retval.0
 }
 
-; The 1st select should have branch weights equal to the 1st branch.
-; The 2nd select should have freshly calculated branch weights.
+; The selects should have freshly calculated branch weights.
 
 define i32 @SimplifyCondBranchToCondBranch(i1 %cmpa, i1 %cmpb) {
 ; CHECK-LABEL: @SimplifyCondBranchToCondBranch(
 ; CHECK-NEXT:  block1:
-; CHECK-NEXT:    [[BRMERGE:%.*]] = or i1 %cmpb, %cmpa
-; CHECK-NEXT:    [[DOTMUX:%.*]] = select i1 %cmpb, i32 0, i32 2
-; CHECK-NEXT:    [[OUTVAL:%.*]] = select i1 [[BRMERGE]], i32 [[DOTMUX]], i32 1, !prof !12
+; CHECK-NEXT:    [[BRMERGE:%.*]] = or i1 %cmpa, %cmpb
+; CHECK-NEXT:    [[DOTMUX:%.*]] = select i1 %cmpa, i32 0, i32 2, !prof !12
+; CHECK-NEXT:    [[OUTVAL:%.*]] = select i1 [[BRMERGE]], i32 [[DOTMUX]], i32 1, !prof !13
 ; CHECK-NEXT:    ret i32 [[OUTVAL]]
 ;
 block1:
-  br i1 %cmpb, label %block3, label %block2, !prof !0
+  br i1 %cmpa, label %block3, label %block2, !prof !13
 
 block2:
-  br i1 %cmpa, label %block3, label %exit, !prof !2
+  br i1 %cmpb, label %block3, label %exit, !prof !14
+
+block3:
+  %cowval = phi i32 [ 2, %block2 ], [ 0, %block1 ]
+  br label %exit
+
+exit:
+  %outval = phi i32 [ %cowval, %block3 ], [ 1, %block2 ]
+  ret i32 %outval
+}
+
+; Swap the operands of the compares to verify that the weights update correctly.
+
+define i32 @SimplifyCondBranchToCondBranchSwap(i1 %cmpa, i1 %cmpb) {
+; CHECK-LABEL: @SimplifyCondBranchToCondBranchSwap(
+; CHECK-NEXT:  block1:
+; CHECK-NEXT:    [[CMPA_NOT:%.*]] = xor i1 %cmpa, true
+; CHECK-NEXT:    [[CMPB_NOT:%.*]] = xor i1 %cmpb, true
+; CHECK-NEXT:    [[BRMERGE:%.*]] = or i1 [[CMPA_NOT]], [[CMPB_NOT]]
+; CHECK-NEXT:    [[DOTMUX:%.*]] = select i1 [[CMPA_NOT]], i32 0, i32 2, !prof !14
+; CHECK-NEXT:    [[OUTVAL:%.*]] = select i1 [[BRMERGE]], i32 [[DOTMUX]], i32 1, !prof !15
+; CHECK-NEXT:    ret i32 [[OUTVAL]]
+;
+block1:
+  br i1 %cmpa, label %block2, label %block3, !prof !13
+
+block2:
+  br i1 %cmpb, label %exit, label %block3, !prof !14
 
 block3:
   %cowval = phi i32 [ 2, %block2 ], [ 0, %block1 ]
@@ -452,6 +478,8 @@ exit:
 !10 = !{!"branch_weights", i32 672646, i32 21604207}
 !11 = !{!"branch_weights", i32 6960, i32 21597248}
 !12 = !{!"these_are_not_the_branch_weights_you_are_looking_for", i32 3, i32 5}
+!13 = !{!"branch_weights", i32 2, i32 3}
+!14 = !{!"branch_weights", i32 4, i32 7}
 
 ; CHECK: !0 = !{!"branch_weights", i32 5, i32 11}
 ; CHECK: !1 = !{!"branch_weights", i32 1, i32 5}
@@ -467,5 +495,8 @@ exit:
 ;; treat the weight as an unsigned integer.
 ; CHECK: !10 = !{!"branch_weights", i32 112017436, i32 -735157296}
 ; CHECK: !11 = !{!"branch_weights", i32 3, i32 5}
-; CHECK: !12 = !{!"branch_weights", i32 14, i32 10}
+; CHECK: !12 = !{!"branch_weights", i32 22, i32 12}
+; CHECK: !13 = !{!"branch_weights", i32 34, i32 21}
+; CHECK: !14 = !{!"branch_weights", i32 33, i32 14}
+; CHECK: !15 = !{!"branch_weights", i32 47, i32 8}