Skip to content

[llvm-dlltool] Use EXPORTAS name type for renamed imports on ARM64EC. #99346

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Jul 22, 2024
Merged

[llvm-dlltool] Use EXPORTAS name type for renamed imports on ARM64EC. #99346

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 4 additions & 1 deletion llvm/lib/Object/COFFImportFile.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -729,7 +729,10 @@ Error writeImportLibrary(StringRef ImportName, StringRef Path,
else if (Machine == IMAGE_FILE_MACHINE_I386 &&
applyNameType(IMPORT_NAME_NOPREFIX, Name) == E.ImportName)
NameType = IMPORT_NAME_NOPREFIX;
else if (Name == E.ImportName)
else if (isArm64EC(M)) {
NameType = IMPORT_NAME_EXPORTAS;
ExportName = E.ImportName;
} else if (Name == E.ImportName)
NameType = IMPORT_NAME;
else {
Deferred D;
Expand Down
29 changes: 29 additions & 0 deletions llvm/test/tools/llvm-dlltool/arm64ec.test
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -44,6 +44,29 @@ RUN: llvm-nm --print-armap test3.lib | FileCheck --check-prefix=ARMAP2 %s
RUN: not llvm-dlltool -m arm64 -d test.def -N test2.def -l test4.lib 2>&1 | FileCheck --check-prefix=ERR %s
ERR: native .def file is supported only on arm64ec target

RUN: llvm-dlltool -m arm64ec -d test3.def -l test3.lib
RUN: llvm-readobj test3.lib | FileCheck --check-prefix=ALIAS %s

ALIAS: File: test.dll
ALIAS-NEXT: Format: COFF-import-file-ARM64EC
ALIAS-NEXT: Type: code
ALIAS-NEXT: Name type: export as
ALIAS-NEXT: Export name: efunc
ALIAS-NEXT: Symbol: __imp_func
ALIAS-NEXT: Symbol: func
ALIAS-NEXT: Symbol: __imp_aux_func
ALIAS-NEXT: Symbol: #func
ALIAS-EMPTY:
ALIAS-NEXT: File: test.dll
ALIAS-NEXT: Format: COFF-import-file-ARM64EC
ALIAS-NEXT: Type: code
ALIAS-NEXT: Name type: export as
ALIAS-NEXT: Export name: efunc
ALIAS-NEXT: Symbol: __imp_efunc
ALIAS-NEXT: Symbol: efunc
ALIAS-NEXT: Symbol: __imp_aux_efunc
ALIAS-NEXT: Symbol: #efunc

#--- test.def
LIBRARY test.dll
EXPORTS
Expand All @@ -53,3 +76,9 @@ EXPORTS
LIBRARY test.dll
EXPORTS
otherfunc

#--- test3.def
LIBRARY test.dll
EXPORTS
func == efunc
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hmm, isn't this the case that end up using EXPORTAS already, since #98229? Would it be clearer to test a case here, where you'd have a separate efunc entry, so that it shows that even if we have the target exported, we'd still use the EXPORTAS type instead of weak aliases?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good point, that's a more interesting case. I pushed more tests.

FWIW, we were not using weak alias path in that test, but for a wrong reason: we were comparing mangled name to demangled one (and generally skipping mangling fixup). That's no longer a problem with this PR as there are no deferred on EC target. Looking at it again we should probably take ExportName into account when filling RegularImports map anyway for corner cases with a mixture of mingw-style aliases and explicit EXPORTAS .

efunc
Loading