Allow weak/selectany external definitions in header units. #162713

akrieger · 2025-10-09T18:23:26Z

weak and selectany are mechanisms for allowing the linker to resolve ODR violations. [module.import/6] states

A header unit shall not contain a definition of a non-inline function or variable whose name has external linkage.

But this prevents compiling any headers with such weak symbols defined. These occur in eg. some Windows SDK headers like DirectXMath.h.

#ifndef XMGLOBALCONST
#if defined(__GNUC__) && !defined(__MINGW32__)
#define XMGLOBALCONST extern const __attribute__((weak))
#else
#define XMGLOBALCONST extern const __declspec(selectany)
#endif
#endif

    XMGLOBALCONST XMVECTORF32 g_XMSinCoefficients0 = { { { -0.16666667f, +0.0083333310f, -0.00019840874f, +2.7525562e-06f } } };

Proposed solution: Do not emit diag::err_extern_def_in_header_unit if the FD or VDecl have either SelectAnyAttr or WeakAttr.

github-actions · 2025-10-09T18:23:49Z

Thank you for submitting a Pull Request (PR) to the LLVM Project!

This PR will be automatically labeled and the relevant teams will be notified.

If you wish to, you can add reviewers by using the "Reviewers" section on this page.

If this is not working for you, it is probably because you do not have write permissions for the repository. In which case you can instead tag reviewers by name in a comment by using @ followed by their GitHub username.

If you have received no comments on your PR for a week, you can request a review by "ping"ing the PR by adding a comment “Ping”. The common courtesy "ping" rate is once a week. Please remember that you are asking for valuable time from other developers.

If you have further questions, they may be answered by the LLVM GitHub User Guide.

You can also ask questions in a comment on this PR, on the LLVM Discord or on the forums.

llvmbot · 2025-10-09T18:24:25Z

@llvm/pr-subscribers-clang-modules

@llvm/pr-subscribers-clang

Author: None (akrieger)

Changes

weak and selectany are mechanisms for allowing the linker to resolve ODR violations. [module.import/6] states

> A header unit shall not contain a definition of a non-inline function or variable whose name has external linkage.

But this prevents compiling any headers with such weak symbols defined. These occur in eg. some Windows SDK headers like DirectXMath.h.

#ifndef XMGLOBALCONST
#if defined(__GNUC__) &amp;&amp; !defined(__MINGW32__)
#define XMGLOBALCONST extern const __attribute__((weak))
#else
#define XMGLOBALCONST extern const __declspec(selectany)
#endif
#endif

    XMGLOBALCONST XMVECTORF32 g_XMSinCoefficients0 = { { { -0.16666667f, +0.0083333310f, -0.00019840874f, +2.7525562e-06f } } };

Proposed solution: Do not emit diag::err_extern_def_in_header_unit if the FD or VDecl have either SelectAnyAttr or WeakAttr.

Full diff: https://github.com/llvm/llvm-project/pull/162713.diff

2 Files Affected:

(modified) clang/lib/Sema/SemaDecl.cpp (+4-2)
(modified) clang/test/CXX/module/module.import/p6.cpp (+13)

diff --git a/clang/lib/Sema/SemaDecl.cpp b/clang/lib/Sema/SemaDecl.cpp
index 6eaf7b9435491..0063bd51aa612 100644
--- a/clang/lib/Sema/SemaDecl.cpp
+++ b/clang/lib/Sema/SemaDecl.cpp
@@ -13822,7 +13822,8 @@ void Sema::AddInitializerToDecl(Decl *RealDecl, Expr *Init, bool DirectInit) {
       !VDecl->isInvalidDecl() && VDecl->isThisDeclarationADefinition() &&
       VDecl->getFormalLinkage() == Linkage::External && !VDecl->isInline() &&
       !VDecl->isTemplated() && !isa<VarTemplateSpecializationDecl>(VDecl) &&
-      !VDecl->getInstantiatedFromStaticDataMember()) {
+      !VDecl->getInstantiatedFromStaticDataMember() &&
+      !(VDecl->hasAttr<SelectAnyAttr>() || VDecl->hasAttr<WeakAttr>())) {
     Diag(VDecl->getLocation(), diag::err_extern_def_in_header_unit);
     VDecl->setInvalidDecl();
   }
@@ -16162,7 +16163,8 @@ Decl *Sema::ActOnStartOfFunctionDef(Scope *FnBodyScope, Decl *D,
       !FD->isInvalidDecl() && !FD->isInlined() &&
       BodyKind != FnBodyKind::Delete && BodyKind != FnBodyKind::Default &&
       FD->getFormalLinkage() == Linkage::External && !FD->isTemplated() &&
-      !FD->isTemplateInstantiation()) {
+      !FD->isTemplateInstantiation() &&
+      !(FD->hasAttr<SelectAnyAttr>() || FD->hasAttr<WeakAttr>())) {
     assert(FD->isThisDeclarationADefinition());
     Diag(FD->getLocation(), diag::err_extern_def_in_header_unit);
     FD->setInvalidDecl();
diff --git a/clang/test/CXX/module/module.import/p6.cpp b/clang/test/CXX/module/module.import/p6.cpp
index cb2d799e5b565..9e378a5fe7759 100644
--- a/clang/test/CXX/module/module.import/p6.cpp
+++ b/clang/test/CXX/module/module.import/p6.cpp
@@ -3,6 +3,9 @@
 
 // RUN: %clang_cc1 -std=c++20 -x c++-header %t/bad-header-unit.h \
 // RUN:  -emit-header-unit -o %t/bad-header-unit.pcm -verify
+// RUN: %clang_cc1 -std=c++20 -x c++-header %t/bad-header-unit-declspec.h \
+// RUN:  -emit-header-unit -o %t/bad-header-unit.pcm -verify \
+// RUN:  -fdeclspec
 
 //--- bad-header-unit.h
 
@@ -77,3 +80,13 @@ template <typename T> bool b() {
 }
 
 inline bool B = b<int>();
+
+__attribute__((weak)) int weak_fun_definition() { return 42; }
+
+__attribute__((weak)) int weak_var_definition = 42;
+
+//--- bad-header-unit-declspec.h
+
+/* The cases below should compile without diagnostics.  */
+
+__declspec(selectany) int selectany_var_definition = 42; // expected-no-diagnostics

github-actions · 2025-10-09T20:25:16Z

⚠️ We detected that you are using a GitHub private e-mail address to contribute to the repo.
Please turn off Keep my email addresses private setting in your account.
See LLVM Developer Policy and LLVM Discourse for more information.

ChuanqiXu9

Looks not bad. Let's add a comment for it.

akrieger · 2025-10-14T18:37:47Z

Added a comment, in the correct style I hope.

ChuanqiXu9 · 2025-10-15T01:46:40Z

clang/lib/Sema/SemaDecl.cpp

+      // The intent of the above section seems to be to prevent ODR issues.
+      // Weak symbols do the same but are not part of the c++ spec.


We should cite the wording in the spec first and explain why we do here as you said in summary of the PR.

The spec is cited already at lines 13819-13820 which is what I am referencing.

https://eel.is/c++draft/module.import#6
I could import the exact phrasing A header unit shall not contain a definition of a non-inline function or variable whose name has external linkage. I guess? I will iterate on it and see if I find something I like.

I feel better to have "we choose to permit entities with weak and selectany attribute as they are common in headers in practice."

Really they are the same thing. It's not that they are common, they just have approximately the same semantics as inline.

I don't like the wording of "The intent of the above section seems to be" because it makes it sound like we don't actually know what the compiler is doing. We know it's disallowing multiple definitions of entities that are not allowed to have multiple definitions, and weak/selectany are just another way to allow multiple definitions.

"the above section" was meant to refer to the section of the c++ spec, I see how that's ambiguous here.

Summary: weak and selectany are mechanisms for allowing the linker to resolve ODR violations. [module.import/6] states > A header unit shall not contain a definition of a non-inline function or variable whose name has external linkage. But this prevents compiling any headers with such weak symbols defined. These occur in eg. some Windows SDK headers like `DirectXMath.h`.

ChuanqiXu9

LGTM now.

github-actions · 2025-10-16T02:13:08Z

@akrieger Congratulations on having your first Pull Request (PR) merged into the LLVM Project!

Your changes will be combined with recent changes from other authors, then tested by our build bots. If there is a problem with a build, you may receive a report in an email or a comment on this PR.

Please check whether problems have been caused by your change specifically, as the builds can include changes from many authors. It is not uncommon for your change to be included in a build that fails due to someone else's changes, or infrastructure issues.

How to do this, and the rest of the post-merge process, is covered in detail here.

If your change does cause a problem, it may be reverted, or you can revert it yourself. This is a normal part of LLVM development. You can fix your changes and open a new PR to merge them again.

If you don't get any reports, no action is required from you. Your changes are working as expected, well done!

akrieger · 2025-10-18T02:16:41Z

Thank you everyone for the review.

llvmbot added clang Clang issues not falling into any other category clang:frontend Language frontend issues, e.g. anything involving "Sema" clang:modules C++20 modules and Clang Header Modules labels Oct 9, 2025

mpark requested review from ChuanqiXu9 and dmpolukhin October 9, 2025 18:27

akrieger force-pushed the selectany_in_header_units_gh branch from 832e71b to 9bc35dd Compare October 9, 2025 20:33

ChuanqiXu9 reviewed Oct 10, 2025

View reviewed changes

akrieger force-pushed the selectany_in_header_units_gh branch from 9bc35dd to 0134ee7 Compare October 14, 2025 18:37

ChuanqiXu9 reviewed Oct 15, 2025

View reviewed changes

akrieger force-pushed the selectany_in_header_units_gh branch from 0134ee7 to 21cf212 Compare October 16, 2025 00:16

akrieger force-pushed the selectany_in_header_units_gh branch from 21cf212 to d6a4502 Compare October 16, 2025 00:19

ChuanqiXu9 approved these changes Oct 16, 2025

View reviewed changes

ChuanqiXu9 enabled auto-merge (squash) October 16, 2025 01:47

ChuanqiXu9 merged commit 705b996 into llvm:main Oct 16, 2025
9 of 10 checks passed

akrieger deleted the selectany_in_header_units_gh branch October 16, 2025 20:46

		// The intent of the above section seems to be to prevent ODR issues.
		// Weak symbols do the same but are not part of the c++ spec.

Allow weak/selectany external definitions in header units. #162713

Allow weak/selectany external definitions in header units. #162713

Conversation

akrieger commented Oct 9, 2025

Uh oh!

github-actions bot commented Oct 9, 2025

Uh oh!

llvmbot commented Oct 9, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions bot commented Oct 9, 2025

Uh oh!

ChuanqiXu9 left a comment

Choose a reason for hiding this comment

Uh oh!

akrieger commented Oct 14, 2025

Uh oh!

ChuanqiXu9 Oct 15, 2025

Choose a reason for hiding this comment

Uh oh!

akrieger Oct 15, 2025

Choose a reason for hiding this comment

Uh oh!

akrieger Oct 15, 2025

Choose a reason for hiding this comment

Uh oh!

ChuanqiXu9 Oct 15, 2025

Choose a reason for hiding this comment

Uh oh!

Bigcheese Oct 15, 2025

Choose a reason for hiding this comment

Uh oh!

akrieger Oct 16, 2025

Choose a reason for hiding this comment

Uh oh!

ChuanqiXu9 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

github-actions bot commented Oct 16, 2025

Uh oh!

akrieger commented Oct 18, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

llvmbot commented Oct 9, 2025 •

edited

Loading