[PATCH] D76975: Make GVN aware of aligned_alloc

[Uday Bondhugula via Phabricator via llvm-commits]

bondhugula created this revision.
Herald added subscribers: llvm-commits, hiraditya.
Herald added a project: LLVM.

Make the GVN pass aware of aligned_alloc. There don't appear to be any
GVN test cases for malloc/alloca - this change is trivial and along the
same lines malloc/alloca are treated; a test case could be added when
it's added for the former.

Signed-off-by: Uday Bondhugula <uday at polymagelabs.com>

Depends on D76974 <https://reviews.llvm.org/D76974>.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D76975

Files:
  llvm/lib/Transforms/Scalar/GVN.cpp


Index: llvm/lib/Transforms/Scalar/GVN.cpp
===================================================================
--- llvm/lib/Transforms/Scalar/GVN.cpp
+++ llvm/lib/Transforms/Scalar/GVN.cpp
@@ -946,6 +946,7 @@
 
   // Loading the allocation -> undef.
   if (isa<AllocaInst>(DepInst) || isMallocLikeFn(DepInst, TLI) ||
+      isAlignedAllocLikeFn(DepInst, TLI) ||
       // Loading immediately after lifetime begin -> undef.
       isLifetimeStart(DepInst)) {
     Res = AvailableValue::get(UndefValue::get(LI->getType()));
@@ -1450,7 +1451,7 @@
       Value *LHS = Cmp->getOperand(0);
       Value *RHS = Cmp->getOperand(1);
       // If we can prove either side non-zero, then equality must imply
-      // equivalence. 
+      // equivalence.
       // FIXME: We should do this optimization if 'no signed zeros' is
       // applicable via an instruction-level fast-math-flag or some other
       // indicator that relaxed FP semantics are being used.
@@ -1515,10 +1516,10 @@
   // If we find an equality fact, canonicalize all dominated uses in this block
   // to one of the two values.  We heuristically choice the "oldest" of the
   // two where age is determined by value number. (Note that propagateEquality
-  // above handles the cross block case.) 
-  // 
+  // above handles the cross block case.)
+  //
   // Key case to cover are:
-  // 1) 
+  // 1)
   // %cmp = fcmp oeq float 3.000000e+00, %0 ; const on lhs could happen
   // call void @llvm.assume(i1 %cmp)
   // ret float %0 ; will change it to ret float 3.000000e+00
@@ -1559,7 +1560,7 @@
                  << *CmpLHS << " with "
                  << *CmpRHS << " in block "
                  << IntrinsicI->getParent()->getName() << "\n");
-      
+
 
       // Setup the replacement map - this handles uses within the same block
       if (hasUsersIn(CmpLHS, IntrinsicI->getParent()))
@@ -1824,7 +1825,7 @@
 bool GVN::replaceOperandsForInBlockEquality(Instruction *Instr) const {
   bool Changed = false;
   for (unsigned OpNum = 0; OpNum < Instr->getNumOperands(); ++OpNum) {
-    Value *Operand = Instr->getOperand(OpNum); 
+    Value *Operand = Instr->getOperand(OpNum);
     auto it = ReplaceOperandsWithMap.find(Operand);
     if (it != ReplaceOperandsWithMap.end()) {
       LLVM_DEBUG(dbgs() << "GVN replacing: " << *Operand << " with "
@@ -1944,7 +1945,7 @@
 
       // If "A == B" is known true, or "A != B" is known false, then replace
       // A with B everywhere in the scope.  For floating point operations, we
-      // have to be careful since equality does not always imply equivalance.  
+      // have to be careful since equality does not always imply equivalance.
       if ((isKnownTrue && impliesEquivalanceIfTrue(Cmp)) ||
           (isKnownFalse && impliesEquivalanceIfFalse(Cmp)))
         Worklist.push_back(std::make_pair(Op0, Op1));


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D76975.253304.patch
Type: text/x-patch
Size: 2832 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200328/6974b341/attachment.bin>