[LLD] [COFF] Fix symbol names for import thunks #160694
Merged
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
9cc9efc changed LLD to use a StringTableBuilder for optimizing the string table, used for long section and symbol names. That commit had a bug, where the symbol table entry for an import thunk with a long symbol name wouldn't get fetched from the StringTableBuilder, if the base symbol name (without the "__imp_" prefix) wasn't over 8 chars. This should fix issues with Go, which errors out on reading the executables with a broken symbol table, as noted in mstorsjo/llvm-mingw#518 and golang/go#75219.
|
@llvm/pr-subscribers-lld-coff @llvm/pr-subscribers-lld Author: Martin Storsjö (mstorsjo) Changes9cc9efc changed LLD to use a StringTableBuilder for optimizing the string table, used for long section and symbol names. That commit had a bug, where the symbol table entry for an import thunk with a long symbol name wouldn't get fetched from the StringTableBuilder, if the base symbol name (without the "_imp" prefix) wasn't over 8 chars. This should fix issues with Go, which errors out on reading the executables with a broken symbol table, as noted in mstorsjo/llvm-mingw#518 and golang/go#75219. Full diff: https://github.com/llvm/llvm-project/pull/160694.diff 2 Files Affected:
diff --git a/lld/COFF/Writer.cpp b/lld/COFF/Writer.cpp
index b4f00996319b1..258a82e371f3a 100644
--- a/lld/COFF/Writer.cpp
+++ b/lld/COFF/Writer.cpp
@@ -1620,7 +1620,7 @@ void Writer::createSymbolAndStringTable() {
dthunk->wrappedSym->writtenToSymtab = true;
if (std::optional<coff_symbol16> sym =
createSymbol(dthunk->wrappedSym)) {
- if (d->getName().size() > COFF::NameSize)
+ if (dthunk->wrappedSym->getName().size() > COFF::NameSize)
longNameSymbols.emplace_back(outputSymtab.size(),
dthunk->wrappedSym->getName());
outputSymtab.push_back(*sym);
diff --git a/lld/test/COFF/strtab.s b/lld/test/COFF/strtab.s
index fbdd8df52d540..9edc13e19e825 100644
--- a/lld/test/COFF/strtab.s
+++ b/lld/test/COFF/strtab.s
@@ -1,17 +1,32 @@
# REQUIRES: x86
# RUN: llvm-mc -triple=x86_64-windows-msvc %s -filetype=obj -o %t.obj
-# RUN: lld-link -out:%t.exe -entry:main %t.obj -debug:dwarf
+# RUN: lld-link -machine:x64 -def:%S/Inputs/library.def -implib:%t.lib
+# RUN: lld-link -out:%t.exe -entry:main %t.obj %t.lib -debug:dwarf
# RUN: llvm-readobj --string-table %t.exe | FileCheck %s
+# RUN: llvm-nm %t.exe | FileCheck %s --check-prefix=SYMBOLS
+
+# Note, for this test to have the intended test coverage, the imported symbol
+# "function" needs to be such that the symbol name itself is <= 8 chars, while
+# "__imp_"+name is >8 chars.
# CHECK: StringTable {
-# CHECK-NEXT: Length: 87
+# CHECK-NEXT: Length: 102
# CHECK-NEXT: [ 4] .debug_abbrev
# CHECK-NEXT: [ 12] .debug_line
# CHECK-NEXT: [ 1e] long_name_symbolz
# CHECK-NEXT: [ 30] .debug_abbrez
-# CHECK-NEXT: [ 3e] __impl_long_name_symbolA
+# CHECK-NEXT: [ 3e] __imp_function
+# CHECK-NEXT: [ 4d] __impl_long_name_symbolA
# CHECK-NEXT: }
+# SYMBOLS: 140001000 N .debug_abbrez
+# SYMBOLS-NEXT: 140002070 R __imp_function
+# SYMBOLS-NEXT: 140001000 t __impl_long_name_symbolA
+# SYMBOLS-NEXT: 140001010 T function
+# SYMBOLS-NEXT: 140001000 t long_name_symbolA
+# SYMBOLS-NEXT: 140001000 t long_name_symbolz
+# SYMBOLS-NEXT: 140001000 T main
+# SYMBOLS-NEXT: 140001000 t name_symbolA
.global main
.text
@@ -21,6 +36,7 @@ long_name_symbolA:
__impl_long_name_symbolA:
name_symbolA:
.debug_abbrez:
+ call function
ret
.section .debug_abbrev,"dr"
|
|
@llvm/pr-subscribers-platform-windows Author: Martin Storsjö (mstorsjo) Changes9cc9efc changed LLD to use a StringTableBuilder for optimizing the string table, used for long section and symbol names. That commit had a bug, where the symbol table entry for an import thunk with a long symbol name wouldn't get fetched from the StringTableBuilder, if the base symbol name (without the "_imp" prefix) wasn't over 8 chars. This should fix issues with Go, which errors out on reading the executables with a broken symbol table, as noted in mstorsjo/llvm-mingw#518 and golang/go#75219. Full diff: https://github.com/llvm/llvm-project/pull/160694.diff 2 Files Affected:
diff --git a/lld/COFF/Writer.cpp b/lld/COFF/Writer.cpp
index b4f00996319b1..258a82e371f3a 100644
--- a/lld/COFF/Writer.cpp
+++ b/lld/COFF/Writer.cpp
@@ -1620,7 +1620,7 @@ void Writer::createSymbolAndStringTable() {
dthunk->wrappedSym->writtenToSymtab = true;
if (std::optional<coff_symbol16> sym =
createSymbol(dthunk->wrappedSym)) {
- if (d->getName().size() > COFF::NameSize)
+ if (dthunk->wrappedSym->getName().size() > COFF::NameSize)
longNameSymbols.emplace_back(outputSymtab.size(),
dthunk->wrappedSym->getName());
outputSymtab.push_back(*sym);
diff --git a/lld/test/COFF/strtab.s b/lld/test/COFF/strtab.s
index fbdd8df52d540..9edc13e19e825 100644
--- a/lld/test/COFF/strtab.s
+++ b/lld/test/COFF/strtab.s
@@ -1,17 +1,32 @@
# REQUIRES: x86
# RUN: llvm-mc -triple=x86_64-windows-msvc %s -filetype=obj -o %t.obj
-# RUN: lld-link -out:%t.exe -entry:main %t.obj -debug:dwarf
+# RUN: lld-link -machine:x64 -def:%S/Inputs/library.def -implib:%t.lib
+# RUN: lld-link -out:%t.exe -entry:main %t.obj %t.lib -debug:dwarf
# RUN: llvm-readobj --string-table %t.exe | FileCheck %s
+# RUN: llvm-nm %t.exe | FileCheck %s --check-prefix=SYMBOLS
+
+# Note, for this test to have the intended test coverage, the imported symbol
+# "function" needs to be such that the symbol name itself is <= 8 chars, while
+# "__imp_"+name is >8 chars.
# CHECK: StringTable {
-# CHECK-NEXT: Length: 87
+# CHECK-NEXT: Length: 102
# CHECK-NEXT: [ 4] .debug_abbrev
# CHECK-NEXT: [ 12] .debug_line
# CHECK-NEXT: [ 1e] long_name_symbolz
# CHECK-NEXT: [ 30] .debug_abbrez
-# CHECK-NEXT: [ 3e] __impl_long_name_symbolA
+# CHECK-NEXT: [ 3e] __imp_function
+# CHECK-NEXT: [ 4d] __impl_long_name_symbolA
# CHECK-NEXT: }
+# SYMBOLS: 140001000 N .debug_abbrez
+# SYMBOLS-NEXT: 140002070 R __imp_function
+# SYMBOLS-NEXT: 140001000 t __impl_long_name_symbolA
+# SYMBOLS-NEXT: 140001010 T function
+# SYMBOLS-NEXT: 140001000 t long_name_symbolA
+# SYMBOLS-NEXT: 140001000 t long_name_symbolz
+# SYMBOLS-NEXT: 140001000 T main
+# SYMBOLS-NEXT: 140001000 t name_symbolA
.global main
.text
@@ -21,6 +36,7 @@ long_name_symbolA:
__impl_long_name_symbolA:
name_symbolA:
.debug_abbrez:
+ call function
ret
.section .debug_abbrev,"dr"
|
aganea
approved these changes
Sep 25, 2025
|
/cherry-pick 0d6af2d |
|
/pull-request #160770 |
dyung
pushed a commit
to llvmbot/llvm-project
that referenced
this pull request
Oct 1, 2025
9cc9efc changed LLD to use a StringTableBuilder for optimizing the string table, used for long section and symbol names. That commit had a bug, where the symbol table entry for an import thunk with a long symbol name wouldn't get fetched from the StringTableBuilder, if the base symbol name (without the "__imp_" prefix) wasn't over 8 chars. This should fix issues with Go, which errors out on reading the executables with a broken symbol table, as noted in mstorsjo/llvm-mingw#518 and golang/go#75219. (cherry picked from commit 0d6af2d)
mahesh-attarde
pushed a commit
to mahesh-attarde/llvm-project
that referenced
this pull request
Oct 3, 2025
9cc9efc changed LLD to use a StringTableBuilder for optimizing the string table, used for long section and symbol names. That commit had a bug, where the symbol table entry for an import thunk with a long symbol name wouldn't get fetched from the StringTableBuilder, if the base symbol name (without the "__imp_" prefix) wasn't over 8 chars. This should fix issues with Go, which errors out on reading the executables with a broken symbol table, as noted in mstorsjo/llvm-mingw#518 and golang/go#75219.
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.
9cc9efc changed LLD to use a StringTableBuilder for optimizing the string table, used for long section and symbol names.
That commit had a bug, where the symbol table entry for an import thunk with a long symbol name wouldn't get fetched from the StringTableBuilder, if the base symbol name (without the "_imp" prefix) wasn't over 8 chars.
This should fix issues with Go, which errors out on reading the executables with a broken symbol table, as noted in mstorsjo/llvm-mingw#518 and golang/go#75219.