diff --git a/llvm/lib/Transforms/InstCombine/InstCombineLoadStoreAlloca.cpp b/llvm/lib/Transforms/InstCombine/InstCombineLoadStoreAlloca.cpp
--- a/llvm/lib/Transforms/InstCombine/InstCombineLoadStoreAlloca.cpp
+++ b/llvm/lib/Transforms/InstCombine/InstCombineLoadStoreAlloca.cpp
@@ -12,6 +12,7 @@
 
 #include "InstCombineInternal.h"
 #include "llvm/ADT/MapVector.h"
+#include "llvm/ADT/STLExtras.h"
 #include "llvm/ADT/SmallString.h"
 #include "llvm/ADT/Statistic.h"
 #include "llvm/Analysis/AliasAnalysis.h"
@@ -53,14 +54,13 @@
     const bool IsOffset = ValuePair.second;
     for (auto &U : ValuePair.first->uses()) {
       auto *I = cast<Instruction>(U.getUser());
-
       if (auto *LI = dyn_cast<LoadInst>(I)) {
         // Ignore non-volatile loads, they are always ok.
         if (!LI->isSimple()) return false;
         continue;
       }
 
-      if (isa<BitCastInst>(I) || isa<AddrSpaceCastInst>(I)) {
+      if (isa<BitCastInst>(I) || isa<AddrSpaceCastInst>(I) || isa<PHINode>(I)) {
         // If uses of the bitcast are ok, we are ok.
         ValuesToInspect.emplace_back(I, IsOffset);
         continue;
@@ -256,6 +256,12 @@
       if (Load->isVolatile())
         return false;
       Worklist.insert(Load);
+    } else if (auto *PHI = dyn_cast<PHINode>(Inst)) {
+      for (unsigned I = 0; I < PHI->getNumIncomingValues(); ++I)
+        Worklist.insert(cast<Instruction>(PHI->getIncomingValue(I)));
+      Worklist.insert(PHI);
+      if (!collectUsers(*PHI))
+        return false;
     } else if (isa<GetElementPtrInst>(Inst) || isa<BitCastInst>(Inst)) {
       Worklist.insert(Inst);
       if (!collectUsers(*Inst))
@@ -293,6 +299,15 @@
     IC.InsertNewInstWith(NewI, *LT);
     IC.replaceInstUsesWith(*LT, NewI);
     WorkMap[LT] = NewI;
+  } else if (auto *PHI = dyn_cast<PHINode>(I)) {
+    auto *NewTy = getReplacement(PHI->getIncomingValue(0))->getType();
+    auto *NewPHI = PHINode::Create(NewTy, PHI->getNumIncomingValues());
+    for (unsigned int I = 0; I < PHI->getNumIncomingValues(); ++I)
+      NewPHI->addIncoming(getReplacement(PHI->getIncomingValue(I)),
+                          PHI->getIncomingBlock(I));
+    IC.InsertNewInstWith(NewPHI, *PHI);
+    NewPHI->takeName(PHI);
+    WorkMap[PHI] = NewPHI;
   } else if (auto *GEP = dyn_cast<GetElementPtrInst>(I)) {
     auto *V = getReplacement(GEP->getPointerOperand());
     assert(V && "Operand not replaced");
@@ -430,7 +445,6 @@
       if (PtrReplacer.collectUsers(AI)) {
         for (Instruction *Delete : ToDelete)
           eraseInstFromFunction(*Delete);
-
         Value *Cast = Builder.CreateBitCast(TheSrc, DestTy);
         PtrReplacer.replacePointer(AI, Cast);
         ++NumGlobalCopies;
diff --git a/llvm/test/Transforms/InstCombine/replace-alloca-phi.ll b/llvm/test/Transforms/InstCombine/replace-alloca-phi.ll
new file mode 100644
--- /dev/null
+++ b/llvm/test/Transforms/InstCombine/replace-alloca-phi.ll
@@ -0,0 +1,110 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt -passes=instcombine -S -o - %s | FileCheck %s
+
+target triple="amdgcn-amd-amdhsa"
+
+define i8 @remove_alloca_use_arg([32 x i8] addrspace(4)* noalias readonly align 4 dereferenceable(32) %arg, i1 %cond) {
+; CHECK-LABEL: @remove_alloca_use_arg(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    br i1 [[COND:%.*]], label [[IF:%.*]], label [[ELSE:%.*]]
+; CHECK:       if:
+; CHECK-NEXT:    [[VAL_IF:%.*]] = getelementptr [32 x i8], [32 x i8] addrspace(4)* [[ARG:%.*]], i64 0, i64 1
+; CHECK-NEXT:    br label [[SINK:%.*]]
+; CHECK:       else:
+; CHECK-NEXT:    [[VAL_ELSE:%.*]] = getelementptr [32 x i8], [32 x i8] addrspace(4)* [[ARG]], i64 0, i64 0
+; CHECK-NEXT:    br label [[SINK]]
+; CHECK:       sink:
+; CHECK-NEXT:    [[PTR:%.*]] = phi i8 addrspace(4)* [ [[VAL_IF]], [[IF]] ], [ [[VAL_ELSE]], [[ELSE]] ]
+; CHECK-NEXT:    [[LOAD:%.*]] = load i8, i8 addrspace(4)* [[PTR]], align 1
+; CHECK-NEXT:    ret i8 [[LOAD]]
+;
+entry:
+  %alloca = alloca [32 x i8], align 4, addrspace(5)
+  call void @llvm.memcpy.p5i8.p4i8.i64([32 x i8] addrspace(5)* %alloca, [32 x i8] addrspace(4)* %arg, i64 256, i1 false)
+  br i1 %cond, label %if, label %else
+
+if:
+  %val.if = getelementptr inbounds [32 x i8], [32 x i8] addrspace(5)* %alloca, i32 0, i32 1
+  br label %sink
+
+else:
+  %val.else = getelementptr inbounds [32 x i8], [32 x i8] addrspace(5)* %alloca, i32 0, i32 0
+  br label %sink
+
+sink:
+  %ptr = phi i8 addrspace(5)* [ %val.if, %if ], [ %val.else, %else ]
+  %load = load i8, i8 addrspace(5)* %ptr
+  ret i8 %load
+}
+
+define i8 @volatile_load_keep_alloca([32 x i8] addrspace(4)* noalias readonly align 4 dereferenceable(32) %arg, i1 %cond) {
+; CHECK-LABEL: @volatile_load_keep_alloca(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    [[ALLOCA:%.*]] = alloca [32 x i8], align 4, addrspace(5)
+; CHECK-NEXT:    call void @llvm.memcpy.p5a32i8.p4a32i8.i64([32 x i8] addrspace(5)* noundef nonnull align 4 dereferenceable(256) [[ALLOCA]], [32 x i8] addrspace(4)* noundef align 4 dereferenceable(256) [[ARG:%.*]], i64 256, i1 false)
+; CHECK-NEXT:    br i1 [[COND:%.*]], label [[IF:%.*]], label [[ELSE:%.*]]
+; CHECK:       if:
+; CHECK-NEXT:    [[VAL_IF:%.*]] = getelementptr inbounds [32 x i8], [32 x i8] addrspace(5)* [[ALLOCA]], i64 0, i64 1
+; CHECK-NEXT:    br label [[SINK:%.*]]
+; CHECK:       else:
+; CHECK-NEXT:    [[VAL_ELSE:%.*]] = getelementptr inbounds [32 x i8], [32 x i8] addrspace(5)* [[ALLOCA]], i64 0, i64 0
+; CHECK-NEXT:    br label [[SINK]]
+; CHECK:       sink:
+; CHECK-NEXT:    [[PTR:%.*]] = phi i8 addrspace(5)* [ [[VAL_IF]], [[IF]] ], [ [[VAL_ELSE]], [[ELSE]] ]
+; CHECK-NEXT:    [[LOAD:%.*]] = load volatile i8, i8 addrspace(5)* [[PTR]], align 1
+; CHECK-NEXT:    ret i8 [[LOAD]]
+;
+entry:
+  %alloca = alloca [32 x i8], align 4, addrspace(5)
+  call void @llvm.memcpy.p5i8.p4i8.i64([32 x i8] addrspace(5)* %alloca, [32 x i8] addrspace(4)* %arg, i64 256, i1 false)
+  br i1 %cond, label %if, label %else
+
+if:
+  %val.if = getelementptr inbounds [32 x i8], [32 x i8] addrspace(5)* %alloca, i32 0, i32 1
+  br label %sink
+
+else:
+  %val.else = getelementptr inbounds [32 x i8], [32 x i8] addrspace(5)* %alloca, i32 0, i32 0
+  br label %sink
+
+sink:
+  %ptr = phi i8 addrspace(5)* [ %val.if, %if ], [ %val.else, %else ]
+  %load = load volatile i8, i8 addrspace(5)* %ptr
+  ret i8 %load
+}
+
+define i8 @no_memcpy_keep_alloca([32 x i8] addrspace(4)* noalias readonly align 4 dereferenceable(32) %arg, i1 %cond) {
+; CHECK-LABEL: @no_memcpy_keep_alloca(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    [[ALLOCA:%.*]] = alloca [32 x i8], align 4, addrspace(5)
+; CHECK-NEXT:    br i1 [[COND:%.*]], label [[IF:%.*]], label [[ELSE:%.*]]
+; CHECK:       if:
+; CHECK-NEXT:    [[VAL_IF:%.*]] = getelementptr inbounds [32 x i8], [32 x i8] addrspace(5)* [[ALLOCA]], i64 0, i64 1
+; CHECK-NEXT:    br label [[SINK:%.*]]
+; CHECK:       else:
+; CHECK-NEXT:    [[VAL_ELSE:%.*]] = getelementptr inbounds [32 x i8], [32 x i8] addrspace(5)* [[ALLOCA]], i64 0, i64 0
+; CHECK-NEXT:    br label [[SINK]]
+; CHECK:       sink:
+; CHECK-NEXT:    [[PTR:%.*]] = phi i8 addrspace(5)* [ [[VAL_IF]], [[IF]] ], [ [[VAL_ELSE]], [[ELSE]] ]
+; CHECK-NEXT:    [[LOAD:%.*]] = load volatile i8, i8 addrspace(5)* [[PTR]], align 1
+; CHECK-NEXT:    ret i8 [[LOAD]]
+;
+entry:
+  %alloca = alloca [32 x i8], align 4, addrspace(5)
+  br i1 %cond, label %if, label %else
+
+if:
+  %val.if = getelementptr inbounds [32 x i8], [32 x i8] addrspace(5)* %alloca, i32 0, i32 1
+  br label %sink
+
+else:
+  %val.else = getelementptr inbounds [32 x i8], [32 x i8] addrspace(5)* %alloca, i32 0, i32 0
+  br label %sink
+
+sink:
+  %ptr = phi i8 addrspace(5)* [ %val.if, %if ], [ %val.else, %else ]
+  %load = load volatile i8, i8 addrspace(5)* %ptr
+  ret i8 %load
+}
+
+declare void @llvm.memcpy.p5i8.p4i8.i64([32 x i8] addrspace(5)*, [32 x i8] addrspace(4)*, i64, i1)