[llvm-objcopy] [COFF] Ignore associative sections in executables #151143
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
COFF associative sections is a feature where relocatable object files can have section snippets marked as related to another section snippet, so they are kept or discarded in relation to that other section snippet. When llvm-objcopy removes sections, it also removes sections that are marked as associative to the removed section (as the associative sections otherwise would end up orphaned). In a linked executable module (EXE or DLL), section associativity is meaningless - thus, we should ignore those fields from the input. After linking, GNU ld keeps the SectionDefinition auxillary part of symbols intact as it was in the source object file, which means that it references section numbers in the source object files. This fixes llvm#53433.
Member
Author
|
CC @kikairoya |
Member
|
@llvm/pr-subscribers-llvm-binary-utilities Author: Martin Storsjö (mstorsjo) ChangesCOFF associative sections is a feature where relocatable object files can have section snippets marked as related to another section snippet, so they are kept or discarded in relation to that other section snippet. When llvm-objcopy removes sections, it also removes sections that are marked as associative to the removed section (as the associative sections otherwise would end up orphaned). In a linked executable module (EXE or DLL), section associativity is meaningless - thus, we should ignore those fields from the input. After linking, GNU ld keeps the SectionDefinition auxillary part of symbols intact as it was in the source object file, which means that it references section numbers in the source object files. This fixes #53433. Full diff: https://github.com/llvm/llvm-project/pull/151143.diff 2 Files Affected:
diff --git a/llvm/lib/ObjCopy/COFF/COFFReader.cpp b/llvm/lib/ObjCopy/COFF/COFFReader.cpp
index 62a71d41ded5f..9b55f76e58404 100644
--- a/llvm/lib/ObjCopy/COFF/COFFReader.cpp
+++ b/llvm/lib/ObjCopy/COFF/COFFReader.cpp
@@ -135,7 +135,7 @@ Error COFFReader::readSymbols(Object &Obj, bool IsBigObj) const {
// it is, find the target section unique id.
const coff_aux_section_definition *SD = SymRef.getSectionDefinition();
const coff_aux_weak_external *WE = SymRef.getWeakExternal();
- if (SD && SD->Selection == IMAGE_COMDAT_SELECT_ASSOCIATIVE) {
+ if (SD && SD->Selection == IMAGE_COMDAT_SELECT_ASSOCIATIVE && !Obj.IsPE) {
int32_t Index = SD->getNumber(IsBigObj);
if (Index <= 0 || static_cast<uint32_t>(Index - 1) >= Sections.size())
return createStringError(object_error::parse_failed,
diff --git a/llvm/test/tools/llvm-objcopy/COFF/exe-bogus-assoc.test b/llvm/test/tools/llvm-objcopy/COFF/exe-bogus-assoc.test
new file mode 100644
index 0000000000000..12f14b5d58e1c
--- /dev/null
+++ b/llvm/test/tools/llvm-objcopy/COFF/exe-bogus-assoc.test
@@ -0,0 +1,134 @@
+## Test that bogus associative section symbols in executables are ignored.
+##
+## The executable contains two (bogus) associative section symbols, both for
+## (parts of) the .rdata section; one pointing at the .debug_info section
+## (which will be stripped out) and one pointing at a nonexistent section.
+##
+## Check that stripping does succeed, and that it doesn't end up removing
+## the .rdata section.
+
+# RUN: yaml2obj %s -o %t.in.exe
+
+# RUN: llvm-strip --strip-debug %t.in.exe -o %t.out.exe
+# RUN: llvm-readobj --sections %t.out.exe | FileCheck %s
+
+# CHECK: Name: .rdata
+
+--- !COFF
+OptionalHeader:
+ AddressOfEntryPoint: 4096
+ ImageBase: 5368709120
+ SectionAlignment: 4096
+ FileAlignment: 512
+ MajorOperatingSystemVersion: 4
+ MinorOperatingSystemVersion: 0
+ MajorImageVersion: 0
+ MinorImageVersion: 0
+ MajorSubsystemVersion: 5
+ MinorSubsystemVersion: 2
+ Subsystem: IMAGE_SUBSYSTEM_WINDOWS_CUI
+ DLLCharacteristics: [ ]
+ SizeOfStackReserve: 2097152
+ SizeOfStackCommit: 4096
+ SizeOfHeapReserve: 1048576
+ SizeOfHeapCommit: 4096
+header:
+ Machine: IMAGE_FILE_MACHINE_AMD64
+ Characteristics: [ ]
+sections:
+ - Name: .text
+ Characteristics: [ IMAGE_SCN_CNT_CODE, IMAGE_SCN_MEM_EXECUTE, IMAGE_SCN_MEM_READ ]
+ VirtualAddress: 4096
+ VirtualSize: 48
+ SectionData: E806000000E802000000C3C3C30F1F00FFFFFFFFFFFFFFFF0000000000000000FFFFFFFFFFFFFFFF0000000000000000
+ SizeOfRawData: 512
+ - Name: .rdata
+ Characteristics: [ IMAGE_SCN_CNT_INITIALIZED_DATA, IMAGE_SCN_MEM_READ ]
+ VirtualAddress: 8192
+ VirtualSize: 4
+ SectionData: '00000000'
+ SizeOfRawData: 512
+ - Name: .debug_info
+ Characteristics: [ IMAGE_SCN_CNT_INITIALIZED_DATA, IMAGE_SCN_MEM_DISCARDABLE, IMAGE_SCN_MEM_READ ]
+ VirtualAddress: 16384
+ VirtualSize: 4
+ SectionData: '00000000'
+ SizeOfRawData: 512
+symbols:
+ - Name: .text
+ Value: 0
+ SectionNumber: 1
+ SimpleType: IMAGE_SYM_TYPE_NULL
+ ComplexType: IMAGE_SYM_DTYPE_NULL
+ StorageClass: IMAGE_SYM_CLASS_STATIC
+ SectionDefinition:
+ Length: 11
+ NumberOfRelocations: 2
+ NumberOfLinenumbers: 0
+ CheckSum: 1703692295
+ Number: 1
+ - Name: '.text$func1'
+ Value: 11
+ SectionNumber: 1
+ SimpleType: IMAGE_SYM_TYPE_NULL
+ ComplexType: IMAGE_SYM_DTYPE_NULL
+ StorageClass: IMAGE_SYM_CLASS_STATIC
+ SectionDefinition:
+ Length: 1
+ NumberOfRelocations: 0
+ NumberOfLinenumbers: 0
+ CheckSum: 40735498
+ Number: 3
+ Selection: IMAGE_COMDAT_SELECT_ANY
+ - Name: .rdata
+ Value: 0
+ SectionNumber: 2
+ SimpleType: IMAGE_SYM_TYPE_NULL
+ ComplexType: IMAGE_SYM_DTYPE_NULL
+ StorageClass: IMAGE_SYM_CLASS_STATIC
+ SectionDefinition:
+ Length: 1
+ NumberOfRelocations: 0
+ NumberOfLinenumbers: 0
+ CheckSum: 0
+ Number: 3
+ Selection: IMAGE_COMDAT_SELECT_ASSOCIATIVE
+ - Name: '.text$func2'
+ Value: 12
+ SectionNumber: 1
+ SimpleType: IMAGE_SYM_TYPE_NULL
+ ComplexType: IMAGE_SYM_DTYPE_NULL
+ StorageClass: IMAGE_SYM_CLASS_STATIC
+ SectionDefinition:
+ Length: 1
+ NumberOfRelocations: 0
+ NumberOfLinenumbers: 0
+ CheckSum: 40735498
+ Number: 4
+ Selection: IMAGE_COMDAT_SELECT_ANY
+ - Name: .rdata
+ Value: 1
+ SectionNumber: 2
+ SimpleType: IMAGE_SYM_TYPE_NULL
+ ComplexType: IMAGE_SYM_DTYPE_NULL
+ StorageClass: IMAGE_SYM_CLASS_STATIC
+ SectionDefinition:
+ Length: 1
+ NumberOfRelocations: 0
+ NumberOfLinenumbers: 0
+ CheckSum: 0
+ Number: 4
+ Selection: IMAGE_COMDAT_SELECT_ASSOCIATIVE
+ - Name: .debug_info
+ Value: 0
+ SectionNumber: 3
+ SimpleType: IMAGE_SYM_TYPE_NULL
+ ComplexType: IMAGE_SYM_DTYPE_NULL
+ StorageClass: IMAGE_SYM_CLASS_STATIC
+ SectionDefinition:
+ Length: 4
+ NumberOfRelocations: 0
+ NumberOfLinenumbers: 0
+ CheckSum: 0
+ Number: 0
+...
|
Member
Author
|
/cherry-pick fcbbcff |
Member
|
/pull-request #151336 |
tru
pushed a commit
to llvmbot/llvm-project
that referenced
this pull request
Aug 1, 2025
…m#151143) COFF associative sections is a feature where relocatable object files can have section snippets marked as related to another section snippet, so they are kept or discarded in relation to that other section snippet. When llvm-objcopy removes sections, it also removes sections that are marked as associative to the removed section (as the associative sections otherwise would end up orphaned). In a linked executable module (EXE or DLL), section associativity is meaningless - thus, we should ignore those fields from the input. After linking, GNU ld keeps the SectionDefinition auxillary part of symbols intact as it was in the source object file, which means that it references section numbers in the source object files. This fixes llvm#53433. (cherry picked from commit fcbbcff)
jeremyd2019
pushed a commit
to jeremyd2019/llvm-project
that referenced
this pull request
Aug 6, 2025
…m#151143) COFF associative sections is a feature where relocatable object files can have section snippets marked as related to another section snippet, so they are kept or discarded in relation to that other section snippet. When llvm-objcopy removes sections, it also removes sections that are marked as associative to the removed section (as the associative sections otherwise would end up orphaned). In a linked executable module (EXE or DLL), section associativity is meaningless - thus, we should ignore those fields from the input. After linking, GNU ld keeps the SectionDefinition auxillary part of symbols intact as it was in the source object file, which means that it references section numbers in the source object files. This fixes llvm#53433.
jeremyd2019
pushed a commit
to jeremyd2019/llvm-project
that referenced
this pull request
Aug 9, 2025
…m#151143) COFF associative sections is a feature where relocatable object files can have section snippets marked as related to another section snippet, so they are kept or discarded in relation to that other section snippet. When llvm-objcopy removes sections, it also removes sections that are marked as associative to the removed section (as the associative sections otherwise would end up orphaned). In a linked executable module (EXE or DLL), section associativity is meaningless - thus, we should ignore those fields from the input. After linking, GNU ld keeps the SectionDefinition auxillary part of symbols intact as it was in the source object file, which means that it references section numbers in the source object files. This fixes llvm#53433.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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.
COFF associative sections is a feature where relocatable object files can have section snippets marked as related to another section snippet, so they are kept or discarded in relation to that other section snippet.
When llvm-objcopy removes sections, it also removes sections that are marked as associative to the removed section (as the associative sections otherwise would end up orphaned).
In a linked executable module (EXE or DLL), section associativity is meaningless - thus, we should ignore those fields from the input.
After linking, GNU ld keeps the SectionDefinition auxillary part of symbols intact as it was in the source object file, which means that it references section numbers in the source object files.
This fixes #53433.