[CodeGenPrepare] Filter out unrecreatable addresses from memory optimization (#143566)

Follow up on #139303

Author: e-kud

llvm/lib/CodeGen/CodeGenPrepare.cpp

@@ -6092,6 +6092,13 @@ bool CodeGenPrepare::optimizeMemoryInst(Instruction *MemoryInst, Value *Addr,
       }
 
       if (!ResultIndex) {
+        auto PtrInst = dyn_cast<Instruction>(ResultPtr);
+        // We know that we have a pointer without any offsets. If this pointer
+        // originates from a different basic block than the current one, we
+        // must be able to recreate it in the current basic block.
+        // We do not support the recreation of any instructions yet.
+        if (PtrInst && PtrInst->getParent() != MemoryInst->getParent())
+          return Modified;
         SunkAddr = ResultPtr;
       } else {
         if (ResultPtr->getType() != I8PtrTy)

llvm/test/Transforms/CodeGenPrepare/X86/sink-addr-recreate.ll

@@ -0,0 +1,106 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
+; RUN: opt -S -p 'require<profile-summary>,codegenprepare' -cgpp-huge-func=0 < %s | FileCheck %s
+
+target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-i128:128-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-grtev4-linux-gnu"
+
+@globalptr = external global ptr
+declare ptr @get_ptr(i64)
+
+; Can't recreate invoke instruction
+
+define void @addr_from_invoke() personality ptr null {
+; CHECK-LABEL: define void @addr_from_invoke() personality ptr null {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    [[PTR:%.*]] = invoke ptr @get_ptr(i64 0)
+; CHECK-NEXT:            to label %[[BODY_1:.*]] unwind label %[[EHCLEANUP:.*]]
+; CHECK:       [[EHCLEANUP]]:
+; CHECK-NEXT:    [[PAD:%.*]] = cleanuppad within none []
+; CHECK-NEXT:    cleanupret from [[PAD]] unwind to caller
+; CHECK:       [[BODY_1]]:
+; CHECK-NEXT:    [[GEP1:%.*]] = bitcast ptr [[PTR]] to ptr
+; CHECK-NEXT:    store <4 x i32> zeroinitializer, ptr [[GEP1]], align 4
+; CHECK-NEXT:    [[TMP0:%.*]] = bitcast ptr [[PTR]] to ptr
+; CHECK-NEXT:    [[UNUSED:%.*]] = load <4 x i32>, ptr [[TMP0]], align 4
+; CHECK-NEXT:    store <4 x i32> zeroinitializer, ptr [[TMP0]], align 4
+; CHECK-NEXT:    ret void
+;
+entry:
+  %ptr = invoke ptr @get_ptr(i64 0) to label %body.1 unwind label %ehcleanup
+
+body.2:
+  %unused = load <4 x i32>, ptr %gep, align 4
+  store <4 x i32> zeroinitializer, ptr %gep, align 4
+  ret void
+
+ehcleanup:
+  %pad = cleanuppad within none []
+  cleanupret from %pad unwind to caller
+
+body.1:
+  %gep = getelementptr { i32 }, ptr %ptr, i64 0, i32 0
+  store <4 x i32> zeroinitializer, ptr %gep, align 4
+  br label %body.2
+}
+
+define void @addr_from_arg(ptr %ptr, i1 %p) {
+; CHECK-LABEL: define void @addr_from_arg(
+; CHECK-SAME: ptr [[PTR:%.*]], i1 [[P:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    br i1 [[P]], label %[[BODY_1:.*]], label %[[EXIT:.*]]
+; CHECK:       [[EXIT]]:
+; CHECK-NEXT:    ret void
+; CHECK:       [[BODY_1]]:
+; CHECK-NEXT:    [[TMP0:%.*]] = bitcast ptr [[PTR]] to ptr
+; CHECK-NEXT:    store <4 x i32> zeroinitializer, ptr [[TMP0]], align 4
+; CHECK-NEXT:    [[UNUSED:%.*]] = load <4 x i32>, ptr [[PTR]], align 4
+; CHECK-NEXT:    store <4 x i32> zeroinitializer, ptr [[PTR]], align 4
+; CHECK-NEXT:    ret void
+;
+entry:
+  br i1 %p, label %body.1, label %exit
+
+body.2:
+  %unused = load <4 x i32>, ptr %gep, align 4
+  store <4 x i32> zeroinitializer, ptr %gep, align 4
+  ret void
+
+exit:
+  ret void
+
+body.1:
+  %gep = getelementptr { i32 }, ptr %ptr, i64 0, i32 0
+  store <4 x i32> zeroinitializer, ptr %gep, align 4
+  br label %body.2
+}
+
+define void @addr_from_global(i1 %p) {
+; CHECK-LABEL: define void @addr_from_global(
+; CHECK-SAME: i1 [[P:%.*]]) {
+; CHECK-NEXT:  [[ENTRY:.*:]]
+; CHECK-NEXT:    br i1 [[P]], label %[[BODY_1:.*]], label %[[EXIT:.*]]
+; CHECK:       [[EXIT]]:
+; CHECK-NEXT:    ret void
+; CHECK:       [[BODY_1]]:
+; CHECK-NEXT:    [[GEP1:%.*]] = bitcast ptr @globalptr to ptr
+; CHECK-NEXT:    store <4 x i32> zeroinitializer, ptr [[GEP1]], align 4
+; CHECK-NEXT:    [[UNUSED:%.*]] = load <4 x i32>, ptr @globalptr, align 4
+; CHECK-NEXT:    store <4 x i32> zeroinitializer, ptr @globalptr, align 4
+; CHECK-NEXT:    ret void
+;
+entry:
+  br i1 %p, label %body.1, label %exit
+
+body.2:
+  %unused = load <4 x i32>, ptr %gep, align 4
+  store <4 x i32> zeroinitializer, ptr %gep, align 4
+  ret void
+
+exit:
+  ret void
+
+body.1:
+  %gep = getelementptr { i32 }, ptr @globalptr, i64 0, i32 0
+  store <4 x i32> zeroinitializer, ptr %gep, align 4
+  br label %body.2
+}