[RS4GC] Use one DVCache for both inlineGetBaseAndOffset() and insertParsePoints()

[yrouban]

Summary

Improve the RewriteStatepointsForGC pass by reusing the same defining value cache for both inlineGetBaseAndOffset() and insertParsePoints().

[reames]

LGTM

Sorry for delay here.

llvm/lib/Transforms/Scalar/RewriteStatepointsForGC.cpp

 }
 
 static bool inlineGetBaseAndOffset(Function &F,
-                                   SmallVectorImpl<CallInst *> &Intrinsics) {
-  DefiningValueMapTy DVCache;
+                                   SmallVectorImpl<CallInst *> &Intrinsics,
+                                   DefiningValueMapTy &DVCache) {
   auto &Context = F.getContext();
   auto &DL = F.getParent()->getDataLayout();
   bool Changed = false;

 
 static bool insertParsePoints(Function &F, DominatorTree &DT,
                               TargetTransformInfo &TTI,
-                              SmallVectorImpl<CallBase *> &ToUpdate) {
+                              SmallVectorImpl<CallBase *> &ToUpdate,
+                              DefiningValueMapTy &DVCache) {
 #ifndef NDEBUG
   // sanity check the input
   std::set<CallBase *> Uniqued;

   findLiveReferences(F, DT, ToUpdate, Records);
 
   // B) Find the base pointers for each live pointer
-  /* scope for caching */ {
-    // Cache the 'defining value' relation used in the computation and
-    // insertion of base phis and selects.  This ensures that we don't insert
-    // large numbers of duplicate base_phis.
-    DefiningValueMapTy DVCache;
-    for (size_t i = 0; i < Records.size(); i++) {
-      PartiallyConstructedSafepointRecord &info = Records[i];
-      findBasePointers(DT, DVCache, ToUpdate[i], info);
-    }
-  } // end of cache scope
+  for (size_t i = 0; i < Records.size(); i++) {
+    PartiallyConstructedSafepointRecord &info = Records[i];
+    findBasePointers(DT, DVCache, ToUpdate[i], info);
+  }
 
   // The base phi insertion logic (for any safepoint) may have inserted new
   // instructions which are now live at some safepoint.  The simplest such

     }
   }
 
+  // Cache the 'defining value' relation used in the computation and
+  // insertion of base phis and selects.  This ensures that we don't insert
+  // large numbers of duplicate base_phis. Use one cache for both
+  // inlineGetBaseAndOffset() and insertParsePoints().
+  DefiningValueMapTy DVCache;
+
   if (!Intrinsics.empty())
     // Inline @gc.get.pointer.base() and @gc.get.pointer.offset() before finding
     // live references.
-    MadeChange |= inlineGetBaseAndOffset(F, Intrinsics);
+    MadeChange |= inlineGetBaseAndOffset(F, Intrinsics, DVCache);
 
   if (!ParsePointNeeded.empty())
-    MadeChange |= insertParsePoints(F, DT, TTI, ParsePointNeeded);
+    MadeChange |= insertParsePoints(F, DT, TTI, ParsePointNeeded, DVCache);
+
   return MadeChange;
 }
 

llvm/test/Transforms/RewriteStatepointsForGC/intrinsics-bare.ll

 ; CHECK-NEXT:    [[OBJ1_12:%.*]] = getelementptr inbounds i8, i8 addrspace(1)* [[OBJ1:%.*]], i64 12
 ; CHECK-NEXT:    [[OBJ2_16:%.*]] = getelementptr inbounds i8, i8 addrspace(1)* [[OBJ2:%.*]], i64 16
 ; CHECK-NEXT:    [[SELECTED_BASE:%.*]] = select i1 [[C:%.*]], i8 addrspace(1)* [[OBJ2]], i8 addrspace(1)* [[OBJ1]], !is_base_value !0
-; CHECK-NEXT:    [[SELECTED_BASE1:%.*]] = select i1 [[C]], i8 addrspace(1)* [[OBJ2]], i8 addrspace(1)* [[OBJ1]], !is_base_value !0
 ; CHECK-NEXT:    [[SELECTED:%.*]] = select i1 [[C]], i8 addrspace(1)* [[OBJ2_16]], i8 addrspace(1)* [[OBJ1_12]]
-; CHECK-NEXT:    [[STATEPOINT_TOKEN:%.*]] = call token (i64, i32, void ()*, i32, i32, ...) @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 2882400000, i32 0, void ()* @foo, i32 0, i32 0, i32 0, i32 0) [ "gc-live"(i8 addrspace(1)* [[SELECTED]], i8 addrspace(1)* [[SELECTED_BASE1]]) ]
+; CHECK-NEXT:    [[STATEPOINT_TOKEN:%.*]] = call token (i64, i32, void ()*, i32, i32, ...) @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 2882400000, i32 0, void ()* @foo, i32 0, i32 0, i32 0, i32 0) [ "gc-live"(i8 addrspace(1)* [[SELECTED]], i8 addrspace(1)* [[SELECTED_BASE]]) ]
 ; CHECK-NEXT:    [[SELECTED_RELOCATED:%.*]] = call coldcc i8 addrspace(1)* @llvm.experimental.gc.relocate.p1i8(token [[STATEPOINT_TOKEN]], i32 1, i32 0)
-; CHECK-NEXT:    [[SELECTED_BASE1_RELOCATED:%.*]] = call coldcc i8 addrspace(1)* @llvm.experimental.gc.relocate.p1i8(token [[STATEPOINT_TOKEN]], i32 1, i32 1)
+; CHECK-NEXT:    [[SELECTED_BASE_RELOCATED:%.*]] = call coldcc i8 addrspace(1)* @llvm.experimental.gc.relocate.p1i8(token [[STATEPOINT_TOKEN]], i32 1, i32 1)
 ; CHECK-NEXT:    ret i8 addrspace(1)* [[SELECTED_RELOCATED]]
 ;
 entry: