[PATCH] D39374: CodeGen: Fix insertion position of addrspace cast for alloca

[Yaxun Liu via Phabricator via cfe-commits]

yaxunl created this revision.

For non-zero alloca addr space, alloca is usually casted to default addr
space immediately.

For non-vla, alloca is inserted at AllocaInsertPt, therefore the addr
space cast should also be insterted at AllocaInsertPt. However,
for vla, alloca is inserted at the current insertion point of IRBuilder,
therefore the addr space cast should also inserted at the current
insertion point of IRBuilder.

Currently clang always insert addr space cast at AllocaInsertPt, which
causes invalid IR.

This patch fixes that.


https://reviews.llvm.org/D39374

Files:
  lib/CodeGen/CGExpr.cpp
  test/CodeGenCXX/vla.cpp


Index: test/CodeGenCXX/vla.cpp
===================================================================
--- test/CodeGenCXX/vla.cpp
+++ test/CodeGenCXX/vla.cpp
@@ -1,4 +1,5 @@
-// RUN: %clang_cc1 -std=c++11 -triple x86_64-apple-darwin %s -emit-llvm -o - | FileCheck %s
+// RUN: %clang_cc1 -std=c++11 -triple x86_64-apple-darwin %s -emit-llvm -o - | FileCheck -check-prefixes=X64,CHECK %s
+// RUN: %clang_cc1 -std=c++11 -triple amdgcn---amdgiz %s -emit-llvm -o - | FileCheck -check-prefixes=AMD,CHECK %s
 
 template<typename T>
 struct S {
@@ -9,18 +10,26 @@
 int f() {
   // Make sure that the reference here is enough to trigger the instantiation of
   // the static data member.
-  // CHECK: @_ZN1SIiE1nE = linkonce_odr global i32 5
+  // CHECK: @_ZN1SIiE1nE = linkonce_odr{{.*}} global i32 5
   int a[S<int>::n];
   return sizeof a;
 }
 
 // rdar://problem/9506377
 void test0(void *array, int n) {
   // CHECK-LABEL: define void @_Z5test0Pvi(
-  // CHECK:      [[ARRAY:%.*]] = alloca i8*, align 8
-  // CHECK-NEXT: [[N:%.*]] = alloca i32, align 4
-  // CHECK-NEXT: [[REF:%.*]] = alloca i16*, align 8
-  // CHECK-NEXT: [[S:%.*]] = alloca i16, align 2
+  // X64:        [[ARRAY:%.*]] = alloca i8*, align 8
+  // AMD:        [[ARRAY0:%.*]] = alloca i8*, align 8, addrspace(5)
+  // AMD-NEXT:   [[ARRAY:%.*]] = addrspacecast i8* addrspace(5)* [[ARRAY0]] to i8**
+  // X64-NEXT:   [[N:%.*]] = alloca i32, align 4
+  // AMD:        [[N0:%.*]] = alloca i32, align 4, addrspace(5)
+  // AMD-NEXT:   [[N:%.*]] = addrspacecast i32 addrspace(5)* [[N0]] to i32*
+  // X64-NEXT:   [[REF:%.*]] = alloca i16*, align 8
+  // AMD:        [[REF0:%.*]] = alloca i16*, align 8, addrspace(5)
+  // AMD-NEXT:   [[REF:%.*]] = addrspacecast i16* addrspace(5)* [[REF0]] to i16**
+  // X64-NEXT:   [[S:%.*]] = alloca i16, align 2
+  // AMD:        [[S0:%.*]] = alloca i16, align 2, addrspace(5)
+  // AMD-NEXT:   [[S:%.*]] = addrspacecast i16 addrspace(5)* [[S0]] to i16*
   // CHECK-NEXT: store i8* 
   // CHECK-NEXT: store i32
 
@@ -59,6 +68,8 @@
 void test2(int b) {
   // CHECK-LABEL: define void {{.*}}test2{{.*}}(i32 %b)
   int varr[b];
+  // AMD: %__end = alloca i32*, align 8, addrspace(5)
+  // AMD: [[END:%.*]] = addrspacecast i32* addrspace(5)* %__end to i32**
   // get the address of %b by checking the first store that stores it 
   //CHECK: store i32 %b, i32* [[PTR_B:%.*]]
 
@@ -75,13 +86,16 @@
   //CHECK: [[VLA_SIZEOF:%.*]] = mul nuw i64 4, [[VLA_NUM_ELEMENTS_PRE]]
   //CHECK-NEXT: [[VLA_NUM_ELEMENTS_POST:%.*]] = udiv i64 [[VLA_SIZEOF]], 4
   //CHECK-NEXT: [[VLA_END_PTR:%.*]] = getelementptr inbounds i32, i32* {{%.*}}, i64 [[VLA_NUM_ELEMENTS_POST]]
-  //CHECK-NEXT: store i32* [[VLA_END_PTR]], i32** %__end
+  //X64-NEXT: store i32* [[VLA_END_PTR]], i32** %__end
+  //AMD-NEXT: store i32* [[VLA_END_PTR]], i32** [[END]]
   for (int d : varr) 0;
 }
 
 void test3(int b, int c) {
   // CHECK-LABEL: define void {{.*}}test3{{.*}}(i32 %b, i32 %c)
   int varr[b][c];
+  // AMD: %__end = alloca i32*, align 8, addrspace(5)
+  // AMD: [[END:%.*]] = addrspacecast i32* addrspace(5)* %__end to i32**
   // get the address of %b by checking the first store that stores it 
   //CHECK: store i32 %b, i32* [[PTR_B:%.*]]
   //CHECK-NEXT: store i32 %c, i32* [[PTR_C:%.*]]
@@ -105,7 +119,8 @@
   //CHECK-NEXT: [[VLA_NUM_ELEMENTS:%.*]] = udiv i64 [[VLA_SIZEOF]], [[VLA_SIZEOF_DIM2]]
   //CHECK-NEXT: [[VLA_END_INDEX:%.*]] = mul nsw i64 [[VLA_NUM_ELEMENTS]], [[VLA_DIM2_PRE]]
   //CHECK-NEXT: [[VLA_END_PTR:%.*]] = getelementptr inbounds i32, i32* {{%.*}}, i64 [[VLA_END_INDEX]]
-  //CHECK-NEXT: store i32* [[VLA_END_PTR]], i32** %__end
+  //X64-NEXT: store i32* [[VLA_END_PTR]], i32** %__end
+  //AMD-NEXT: store i32* [[VLA_END_PTR]], i32** [[END]]
  
   for (auto &d : varr) 0;
 }
Index: lib/CodeGen/CGExpr.cpp
===================================================================
--- lib/CodeGen/CGExpr.cpp
+++ lib/CodeGen/CGExpr.cpp
@@ -75,7 +75,11 @@
   if (CastToDefaultAddrSpace && getASTAllocaAddressSpace() != LangAS::Default) {
     auto DestAddrSpace = getContext().getTargetAddressSpace(LangAS::Default);
     llvm::IRBuilderBase::InsertPointGuard IPG(Builder);
-    Builder.SetInsertPoint(AllocaInsertPt);
+    // When ArraySize is nullptr, alloca is inserted at AllocaInsertPt,
+    // otherwise alloca is inserted at the current insertion point of the
+    // builder.
+    if (!ArraySize)
+      Builder.SetInsertPoint(AllocaInsertPt);
     V = getTargetHooks().performAddrSpaceCast(
         *this, V, getASTAllocaAddressSpace(), LangAS::Default,
         Ty->getPointerTo(DestAddrSpace), /*non-null*/ true);


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D39374.120609.patch
Type: text/x-patch
Size: 4602 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20171027/69f0f8d2/attachment-0001.bin>