[PATCH] D115661: [clang][amdgpu] - Choose when to promote VarDecl to address space 4.

[Ethan Stewart via Phabricator via cfe-commits]

estewart08 added a comment.

In D115661#3190477 <https://reviews.llvm.org/D115661#3190477>, @yaxunl wrote:

> This may cause perf regressions for HIP.

Do you have a test that would show such a regression? Emitting a store to address space (4) in a constructor seems the wrong thing to do.



================
Comment at: clang/test/CodeGenCXX/cxx11-extern-constexpr.cpp:10
 // X86: @_ZN1A3FooE ={{.*}} constant i32 123, align 4
-// AMD: @_ZN1A3FooE ={{.*}} addrspace(4) constant i32 123, align 4
+// AMD: @_ZN1A3FooE ={{.*}} addrspace(1) constant i32 123, align 4
 const int *p = &A::Foo; // emit available_externally
----------------
yaxunl wrote:
> Do you know why this is not treated as constant initialization?
> 
There is no initialization here:
```
const int A::Foo;
```

It seems the compiler ignores the 123 in the struct.

I see address space (4) when the test is written like this:
```
struct A {
  static const int Foo;
};
const int A::Foo = 123;
```


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D115661/new/

https://reviews.llvm.org/D115661