[LLD][COFF] Add support for x86_64 archives on ARM64X #128241
Merged
+84
−1
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
@llvm/pr-subscribers-lld Author: Jacek Caban (cjacek) ChangesIf the ECSYMBOLS section is missing in the archive, the archive could be either a native-only ARM64 or x86_64 archive. Check the machine type of the object containing a symbol to determine which symbol table to use. Full diff: https://github.com/llvm/llvm-project/pull/128241.diff 2 Files Affected:
diff --git a/lld/COFF/InputFiles.cpp b/lld/COFF/InputFiles.cpp
index 7b105fb4c17a2..bb9f1407b51f7 100644
--- a/lld/COFF/InputFiles.cpp
+++ b/lld/COFF/InputFiles.cpp
@@ -29,6 +29,7 @@
#include "llvm/LTO/LTO.h"
#include "llvm/Object/Binary.h"
#include "llvm/Object/COFF.h"
+#include "llvm/Object/COFFImportFile.h"
#include "llvm/Support/Casting.h"
#include "llvm/Support/Endian.h"
#include "llvm/Support/Error.h"
@@ -122,6 +123,8 @@ ArchiveFile::ArchiveFile(COFFLinkerContext &ctx, MemoryBufferRef m)
void ArchiveFile::parse() {
COFFLinkerContext &ctx = symtab.ctx;
+ SymbolTable *archiveSymtab = &symtab;
+
// Parse a MemoryBufferRef as an archive file.
file = CHECK(Archive::create(mb), this);
@@ -136,12 +139,50 @@ void ArchiveFile::parse() {
// Read both EC and native symbols on ARM64X.
if (!ctx.hybridSymtab)
return;
+ } else if (ctx.hybridSymtab) {
+ // If the ECSYMBOLS section is missing in the archive, the archive could
+ // be either a native-only ARM64 or x86_64 archive. Check the machine type
+ // of the object containing a symbol to determine which symbol table to
+ // use.
+ Archive::symbol_iterator sym = file->symbol_begin();
+ if (sym != file->symbol_end()) {
+ MachineTypes machine = IMAGE_FILE_MACHINE_UNKNOWN;
+ Archive::Child child =
+ CHECK(sym->getMember(),
+ file->getFileName() +
+ ": could not get the buffer for a child of the archive");
+ MemoryBufferRef mb = CHECK(
+ child.getMemoryBufferRef(),
+ file->getFileName() +
+ ": could not get the buffer for a child buffer of the archive");
+ switch (identify_magic(mb.getBuffer())) {
+ case file_magic::coff_object: {
+ std::unique_ptr<COFFObjectFile> obj =
+ CHECK(COFFObjectFile::create(mb),
+ check(child.getName()) + ":" + ": not a valid COFF file");
+ machine = MachineTypes(obj->getMachine());
+ break;
+ }
+ case file_magic::coff_import_library:
+ machine = MachineTypes(COFFImportFile(mb).getMachine());
+ break;
+ case file_magic::bitcode: {
+ std::unique_ptr<lto::InputFile> obj =
+ check(lto::InputFile::create(mb));
+ machine = BitcodeFile::getMachineType(obj.get());
+ break;
+ }
+ default:
+ break;
+ }
+ archiveSymtab = &ctx.getSymtab(machine);
+ }
}
}
// Read the symbol table to construct Lazy objects.
for (const Archive::Symbol &sym : file->symbols())
- ctx.symtab.addLazyArchive(this, sym);
+ archiveSymtab->addLazyArchive(this, sym);
}
// Returns a buffer pointing to a member file containing a given symbol.
diff --git a/lld/test/COFF/arm64x-symtab.s b/lld/test/COFF/arm64x-symtab.s
index 2b269dde61f61..5f06db150a06d 100644
--- a/lld/test/COFF/arm64x-symtab.s
+++ b/lld/test/COFF/arm64x-symtab.s
@@ -7,7 +7,11 @@
// RUN: llvm-mc -filetype=obj -triple=aarch64-windows symref.s -o symref-aarch64.obj
// RUN: llvm-mc -filetype=obj -triple=arm64ec-windows symref.s -o symref-arm64ec.obj
// RUN: llvm-mc -filetype=obj -triple=x86_64-windows symref.s -o symref-x86_64.obj
+// RUN: llvm-as sym.ll -o sym.ll.obj
// RUN: llvm-lib -machine:arm64x -out:sym.lib sym-aarch64.obj sym-arm64ec.obj
+// RUN: llvm-lib -machine:amd64 -out:sym-x86_64.lib sym-x86_64.obj
+// RUN: llvm-lib -machine:amd64 -out:sym-ll.lib sym.ll.obj
+// RUN: llvm-lib -machine:amd64 -out:sym-imp.lib -def:sym.def
// Check that native object files can't reference EC symbols.
@@ -40,12 +44,40 @@
// RUN: lld-link -machine:arm64x -dll -noentry -out:out2.dll symref-aarch64.obj symref-arm64ec.obj sym.lib
+// Check that EC object files can reference x86_64 library symbols.
+
+// RUN: lld-link -machine:arm64x -dll -noentry -out:out3.dll symref-arm64ec.obj sym-x86_64.lib
+// RUN: lld-link -machine:arm64x -dll -noentry -out:out4.dll symref-arm64ec.obj sym-ll.lib
+// RUN: lld-link -machine:arm64x -dll -noentry -out:out5.dll symref-arm64ec.obj sym-imp.lib
+
+// Check that native object files can't reference x86_64 library symbols.
+
+// RUN: not lld-link -machine:arm64x -dll -noentry -out:err3.dll symref-aarch64.obj sym-arm64ec.obj \
+// RUN: 2>&1 | FileCheck --check-prefix=UNDEF %s
+
#--- symref.s
.data
.rva sym
+ .text
+ .globl __icall_helper_arm64ec
+__icall_helper_arm64ec:
+ ret
+
#--- sym.s
.data
.globl sym
sym:
.word 0
+
+#--- sym.ll
+target datalayout = "e-m:w-p270:32:32-p271:32:32-p272:64:64-i64:64-i128:128-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-unknown-windows-msvc19.33.0"
+
+@sym = dso_local global i32 0, align 4
+
+#--- sym.def
+LIBRARY test.dll
+EXPORTS
+ Func
+ sym
|
|
@llvm/pr-subscribers-platform-windows Author: Jacek Caban (cjacek) ChangesIf the ECSYMBOLS section is missing in the archive, the archive could be either a native-only ARM64 or x86_64 archive. Check the machine type of the object containing a symbol to determine which symbol table to use. Full diff: https://github.com/llvm/llvm-project/pull/128241.diff 2 Files Affected:
diff --git a/lld/COFF/InputFiles.cpp b/lld/COFF/InputFiles.cpp
index 7b105fb4c17a2..bb9f1407b51f7 100644
--- a/lld/COFF/InputFiles.cpp
+++ b/lld/COFF/InputFiles.cpp
@@ -29,6 +29,7 @@
#include "llvm/LTO/LTO.h"
#include "llvm/Object/Binary.h"
#include "llvm/Object/COFF.h"
+#include "llvm/Object/COFFImportFile.h"
#include "llvm/Support/Casting.h"
#include "llvm/Support/Endian.h"
#include "llvm/Support/Error.h"
@@ -122,6 +123,8 @@ ArchiveFile::ArchiveFile(COFFLinkerContext &ctx, MemoryBufferRef m)
void ArchiveFile::parse() {
COFFLinkerContext &ctx = symtab.ctx;
+ SymbolTable *archiveSymtab = &symtab;
+
// Parse a MemoryBufferRef as an archive file.
file = CHECK(Archive::create(mb), this);
@@ -136,12 +139,50 @@ void ArchiveFile::parse() {
// Read both EC and native symbols on ARM64X.
if (!ctx.hybridSymtab)
return;
+ } else if (ctx.hybridSymtab) {
+ // If the ECSYMBOLS section is missing in the archive, the archive could
+ // be either a native-only ARM64 or x86_64 archive. Check the machine type
+ // of the object containing a symbol to determine which symbol table to
+ // use.
+ Archive::symbol_iterator sym = file->symbol_begin();
+ if (sym != file->symbol_end()) {
+ MachineTypes machine = IMAGE_FILE_MACHINE_UNKNOWN;
+ Archive::Child child =
+ CHECK(sym->getMember(),
+ file->getFileName() +
+ ": could not get the buffer for a child of the archive");
+ MemoryBufferRef mb = CHECK(
+ child.getMemoryBufferRef(),
+ file->getFileName() +
+ ": could not get the buffer for a child buffer of the archive");
+ switch (identify_magic(mb.getBuffer())) {
+ case file_magic::coff_object: {
+ std::unique_ptr<COFFObjectFile> obj =
+ CHECK(COFFObjectFile::create(mb),
+ check(child.getName()) + ":" + ": not a valid COFF file");
+ machine = MachineTypes(obj->getMachine());
+ break;
+ }
+ case file_magic::coff_import_library:
+ machine = MachineTypes(COFFImportFile(mb).getMachine());
+ break;
+ case file_magic::bitcode: {
+ std::unique_ptr<lto::InputFile> obj =
+ check(lto::InputFile::create(mb));
+ machine = BitcodeFile::getMachineType(obj.get());
+ break;
+ }
+ default:
+ break;
+ }
+ archiveSymtab = &ctx.getSymtab(machine);
+ }
}
}
// Read the symbol table to construct Lazy objects.
for (const Archive::Symbol &sym : file->symbols())
- ctx.symtab.addLazyArchive(this, sym);
+ archiveSymtab->addLazyArchive(this, sym);
}
// Returns a buffer pointing to a member file containing a given symbol.
diff --git a/lld/test/COFF/arm64x-symtab.s b/lld/test/COFF/arm64x-symtab.s
index 2b269dde61f61..5f06db150a06d 100644
--- a/lld/test/COFF/arm64x-symtab.s
+++ b/lld/test/COFF/arm64x-symtab.s
@@ -7,7 +7,11 @@
// RUN: llvm-mc -filetype=obj -triple=aarch64-windows symref.s -o symref-aarch64.obj
// RUN: llvm-mc -filetype=obj -triple=arm64ec-windows symref.s -o symref-arm64ec.obj
// RUN: llvm-mc -filetype=obj -triple=x86_64-windows symref.s -o symref-x86_64.obj
+// RUN: llvm-as sym.ll -o sym.ll.obj
// RUN: llvm-lib -machine:arm64x -out:sym.lib sym-aarch64.obj sym-arm64ec.obj
+// RUN: llvm-lib -machine:amd64 -out:sym-x86_64.lib sym-x86_64.obj
+// RUN: llvm-lib -machine:amd64 -out:sym-ll.lib sym.ll.obj
+// RUN: llvm-lib -machine:amd64 -out:sym-imp.lib -def:sym.def
// Check that native object files can't reference EC symbols.
@@ -40,12 +44,40 @@
// RUN: lld-link -machine:arm64x -dll -noentry -out:out2.dll symref-aarch64.obj symref-arm64ec.obj sym.lib
+// Check that EC object files can reference x86_64 library symbols.
+
+// RUN: lld-link -machine:arm64x -dll -noentry -out:out3.dll symref-arm64ec.obj sym-x86_64.lib
+// RUN: lld-link -machine:arm64x -dll -noentry -out:out4.dll symref-arm64ec.obj sym-ll.lib
+// RUN: lld-link -machine:arm64x -dll -noentry -out:out5.dll symref-arm64ec.obj sym-imp.lib
+
+// Check that native object files can't reference x86_64 library symbols.
+
+// RUN: not lld-link -machine:arm64x -dll -noentry -out:err3.dll symref-aarch64.obj sym-arm64ec.obj \
+// RUN: 2>&1 | FileCheck --check-prefix=UNDEF %s
+
#--- symref.s
.data
.rva sym
+ .text
+ .globl __icall_helper_arm64ec
+__icall_helper_arm64ec:
+ ret
+
#--- sym.s
.data
.globl sym
sym:
.word 0
+
+#--- sym.ll
+target datalayout = "e-m:w-p270:32:32-p271:32:32-p272:64:64-i64:64-i128:128-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-unknown-windows-msvc19.33.0"
+
+@sym = dso_local global i32 0, align 4
+
+#--- sym.def
+LIBRARY test.dll
+EXPORTS
+ Func
+ sym
|
|
@llvm/pr-subscribers-lld-coff Author: Jacek Caban (cjacek) ChangesIf the ECSYMBOLS section is missing in the archive, the archive could be either a native-only ARM64 or x86_64 archive. Check the machine type of the object containing a symbol to determine which symbol table to use. Full diff: https://github.com/llvm/llvm-project/pull/128241.diff 2 Files Affected:
diff --git a/lld/COFF/InputFiles.cpp b/lld/COFF/InputFiles.cpp
index 7b105fb4c17a2..bb9f1407b51f7 100644
--- a/lld/COFF/InputFiles.cpp
+++ b/lld/COFF/InputFiles.cpp
@@ -29,6 +29,7 @@
#include "llvm/LTO/LTO.h"
#include "llvm/Object/Binary.h"
#include "llvm/Object/COFF.h"
+#include "llvm/Object/COFFImportFile.h"
#include "llvm/Support/Casting.h"
#include "llvm/Support/Endian.h"
#include "llvm/Support/Error.h"
@@ -122,6 +123,8 @@ ArchiveFile::ArchiveFile(COFFLinkerContext &ctx, MemoryBufferRef m)
void ArchiveFile::parse() {
COFFLinkerContext &ctx = symtab.ctx;
+ SymbolTable *archiveSymtab = &symtab;
+
// Parse a MemoryBufferRef as an archive file.
file = CHECK(Archive::create(mb), this);
@@ -136,12 +139,50 @@ void ArchiveFile::parse() {
// Read both EC and native symbols on ARM64X.
if (!ctx.hybridSymtab)
return;
+ } else if (ctx.hybridSymtab) {
+ // If the ECSYMBOLS section is missing in the archive, the archive could
+ // be either a native-only ARM64 or x86_64 archive. Check the machine type
+ // of the object containing a symbol to determine which symbol table to
+ // use.
+ Archive::symbol_iterator sym = file->symbol_begin();
+ if (sym != file->symbol_end()) {
+ MachineTypes machine = IMAGE_FILE_MACHINE_UNKNOWN;
+ Archive::Child child =
+ CHECK(sym->getMember(),
+ file->getFileName() +
+ ": could not get the buffer for a child of the archive");
+ MemoryBufferRef mb = CHECK(
+ child.getMemoryBufferRef(),
+ file->getFileName() +
+ ": could not get the buffer for a child buffer of the archive");
+ switch (identify_magic(mb.getBuffer())) {
+ case file_magic::coff_object: {
+ std::unique_ptr<COFFObjectFile> obj =
+ CHECK(COFFObjectFile::create(mb),
+ check(child.getName()) + ":" + ": not a valid COFF file");
+ machine = MachineTypes(obj->getMachine());
+ break;
+ }
+ case file_magic::coff_import_library:
+ machine = MachineTypes(COFFImportFile(mb).getMachine());
+ break;
+ case file_magic::bitcode: {
+ std::unique_ptr<lto::InputFile> obj =
+ check(lto::InputFile::create(mb));
+ machine = BitcodeFile::getMachineType(obj.get());
+ break;
+ }
+ default:
+ break;
+ }
+ archiveSymtab = &ctx.getSymtab(machine);
+ }
}
}
// Read the symbol table to construct Lazy objects.
for (const Archive::Symbol &sym : file->symbols())
- ctx.symtab.addLazyArchive(this, sym);
+ archiveSymtab->addLazyArchive(this, sym);
}
// Returns a buffer pointing to a member file containing a given symbol.
diff --git a/lld/test/COFF/arm64x-symtab.s b/lld/test/COFF/arm64x-symtab.s
index 2b269dde61f61..5f06db150a06d 100644
--- a/lld/test/COFF/arm64x-symtab.s
+++ b/lld/test/COFF/arm64x-symtab.s
@@ -7,7 +7,11 @@
// RUN: llvm-mc -filetype=obj -triple=aarch64-windows symref.s -o symref-aarch64.obj
// RUN: llvm-mc -filetype=obj -triple=arm64ec-windows symref.s -o symref-arm64ec.obj
// RUN: llvm-mc -filetype=obj -triple=x86_64-windows symref.s -o symref-x86_64.obj
+// RUN: llvm-as sym.ll -o sym.ll.obj
// RUN: llvm-lib -machine:arm64x -out:sym.lib sym-aarch64.obj sym-arm64ec.obj
+// RUN: llvm-lib -machine:amd64 -out:sym-x86_64.lib sym-x86_64.obj
+// RUN: llvm-lib -machine:amd64 -out:sym-ll.lib sym.ll.obj
+// RUN: llvm-lib -machine:amd64 -out:sym-imp.lib -def:sym.def
// Check that native object files can't reference EC symbols.
@@ -40,12 +44,40 @@
// RUN: lld-link -machine:arm64x -dll -noentry -out:out2.dll symref-aarch64.obj symref-arm64ec.obj sym.lib
+// Check that EC object files can reference x86_64 library symbols.
+
+// RUN: lld-link -machine:arm64x -dll -noentry -out:out3.dll symref-arm64ec.obj sym-x86_64.lib
+// RUN: lld-link -machine:arm64x -dll -noentry -out:out4.dll symref-arm64ec.obj sym-ll.lib
+// RUN: lld-link -machine:arm64x -dll -noentry -out:out5.dll symref-arm64ec.obj sym-imp.lib
+
+// Check that native object files can't reference x86_64 library symbols.
+
+// RUN: not lld-link -machine:arm64x -dll -noentry -out:err3.dll symref-aarch64.obj sym-arm64ec.obj \
+// RUN: 2>&1 | FileCheck --check-prefix=UNDEF %s
+
#--- symref.s
.data
.rva sym
+ .text
+ .globl __icall_helper_arm64ec
+__icall_helper_arm64ec:
+ ret
+
#--- sym.s
.data
.globl sym
sym:
.word 0
+
+#--- sym.ll
+target datalayout = "e-m:w-p270:32:32-p271:32:32-p272:64:64-i64:64-i128:128-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-unknown-windows-msvc19.33.0"
+
+@sym = dso_local global i32 0, align 4
+
+#--- sym.def
+LIBRARY test.dll
+EXPORTS
+ Func
+ sym
|
If the ECSYMBOLS section is missing in the archive, the archive could be either a native-only ARM64 or x86_64 archive. Check the machine type of the object containing a symbol to determine which symbol table to use.
cjacek
force-pushed
the
arm64x-x86-lib
branch
from
d233bc7 to
9fcd027
Compare
|
Thanks! I added one more test (ensuring that we don't break native ARM64 archives) and merged. |
SquallATF
pushed a commit
to SquallATF/llvm-project
that referenced
this pull request
Mar 10, 2025
If the ECSYMBOLS section is missing in the archive, the archive could be either a native-only ARM64 or x86_64 archive. Check the machine type of the object containing a symbol to determine which symbol table to use.
SquallATF
pushed a commit
to SquallATF/llvm-project
that referenced
this pull request
Mar 20, 2025
If the ECSYMBOLS section is missing in the archive, the archive could be either a native-only ARM64 or x86_64 archive. Check the machine type of the object containing a symbol to determine which symbol table to use.
SquallATF
pushed a commit
to SquallATF/llvm-project
that referenced
this pull request
Apr 2, 2025
If the ECSYMBOLS section is missing in the archive, the archive could be either a native-only ARM64 or x86_64 archive. Check the machine type of the object containing a symbol to determine which symbol table to use.
SquallATF
pushed a commit
to SquallATF/llvm-project
that referenced
this pull request
Apr 17, 2025
If the ECSYMBOLS section is missing in the archive, the archive could be either a native-only ARM64 or x86_64 archive. Check the machine type of the object containing a symbol to determine which symbol table to use.
SquallATF
pushed a commit
to SquallATF/llvm-project
that referenced
this pull request
Apr 30, 2025
If the ECSYMBOLS section is missing in the archive, the archive could be either a native-only ARM64 or x86_64 archive. Check the machine type of the object containing a symbol to determine which symbol table to use.
SquallATF
pushed a commit
to SquallATF/llvm-project
that referenced
this pull request
May 15, 2025
If the ECSYMBOLS section is missing in the archive, the archive could be either a native-only ARM64 or x86_64 archive. Check the machine type of the object containing a symbol to determine which symbol table to use.
SquallATF
pushed a commit
to SquallATF/llvm-project
that referenced
this pull request
May 29, 2025
If the ECSYMBOLS section is missing in the archive, the archive could be either a native-only ARM64 or x86_64 archive. Check the machine type of the object containing a symbol to determine which symbol table to use.
SquallATF
pushed a commit
to SquallATF/llvm-project
that referenced
this pull request
Jun 13, 2025
If the ECSYMBOLS section is missing in the archive, the archive could be either a native-only ARM64 or x86_64 archive. Check the machine type of the object containing a symbol to determine which symbol table to use.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
If the ECSYMBOLS section is missing in the archive, the archive could be either a native-only ARM64 or x86_64 archive. Check the machine type of the object containing a symbol to determine which symbol table to use.