caio/clang-p2996
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[lld-macho] Implement symbol string deduplication (#123874)

Browse Source 
The symbol string table does not have deduplication. 
Here we add code to deduplicate the symbol string table. 
This has a rather large size impact (20-30%) on unstripped binaries
(typically debug binaries) but no size impact on stripped
binaries(typically release binaries).

We enable deduplication by default and add a flag to disable it
(`-no-deduplicate-symbol-strings`).
This commit is contained in:
alx32
2025-01-23 15:48:11 -08:00
committed by GitHub
parent 28ad8978ee
commit c676104875
6 changed files with 27 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
lld/MachO/Config.h
View File

@@ -143,6 +143,7 @@ struct Configuration {
bool timeTraceEnabled = false;
bool dataConst = false;
bool dedupStrings = true;
bool dedupSymbolStrings = true;
bool deadStripDuplicates = false;
bool omitDebugInfo = false;
bool warnDylibInstallName = false;

1
lld/MachO/Driver.cpp
View File

@@ -1806,6 +1806,7 @@ bool link(ArrayRef<const char *> argsArr, llvm::raw_ostream &stdoutOS,
config->keepICFStabs = args.hasArg(OPT_keep_icf_stabs);
config->dedupStrings =
args.hasFlag(OPT_deduplicate_strings, OPT_no_deduplicate_strings, true);
config->dedupSymbolStrings = !args.hasArg(OPT_no_deduplicate_symbol_strings);
config->deadStripDuplicates = args.hasArg(OPT_dead_strip_duplicates);
config->warnDylibInstallName = args.hasFlag(
OPT_warn_dylib_install_name, OPT_no_warn_dylib_install_name, false);

5
lld/MachO/Options.td
View File

@@ -1476,3 +1476,8 @@ def no_warn_duplicate_libraries : Flag<["-"], "no_warn_duplicate_libraries">,
HelpText<"Do not warn if the input contains duplicate library options.">,
Flags<[HelpHidden]>,
Group<grp_ignored_silently>;
// Add this with the other flags in the rare options group
def no_deduplicate_symbol_strings : Flag<["-"], "no-deduplicate-symbol-strings">,
HelpText<"Do not deduplicate strings in the symbol string table. Might result in larger binaries but slightly faster link times.">,
Group<grp_rare>;

9
lld/MachO/SyntheticSections.cpp
View File

@@ -1541,7 +1541,14 @@ StringTableSection::StringTableSection()
uint32_t StringTableSection::addString(StringRef str) {
uint32_t strx = size;
strings.push_back(str); // TODO: consider deduplicating strings
if (config->dedupSymbolStrings) {
llvm::CachedHashStringRef hashedStr(str);
auto [it, inserted] = stringMap.try_emplace(hashedStr, strx);
if (!inserted)
return it->second;
}
strings.push_back(str);
size += str.size() + 1; // account for null terminator
return strx;
}

1
lld/MachO/SyntheticSections.h
View File

@@ -447,6 +447,7 @@ private:
// match its behavior here since some tools depend on it.
// Consequently, the empty string will be at index 1, not zero.
std::vector<StringRef> strings{" "};
llvm::DenseMap<llvm::CachedHashStringRef, uint32_t> stringMap;
size_t size = 2;
};

11
lld/test/MachO/cfstring-dedup.s
View File

@@ -7,6 +7,17 @@
# RUN: %lld -dylib -framework CoreFoundation %t/foo1.o %t/foo2.o -o %t/foo
# RUN: llvm-objdump --no-print-imm-hex --macho --rebase --bind --syms -d %t/foo | FileCheck %s --check-prefix=LITERALS
# Check that string deduplication for symbol names is working
# RUN: %lld -dylib -framework CoreFoundation %t/foo1.o %t/foo2.o -o %t/foo_no_dedup -no-deduplicate-symbol-strings
# RUN: llvm-strings %t/foo | FileCheck %s --check-prefix=CHECK-DEDUP
# RUN: llvm-strings %t/foo_no_dedup | FileCheck %s --check-prefix=CHECK-NO-DEDUP
# CHECK-DEDUP: _named_cfstring
# CHECK-DEDUP-NOT: _named_cfstring
# CHECK-NO-DEDUP: _named_cfstring
# CHECK-NO-DEDUP: _named_cfstring
# CHECK-NO-DEDUP-NOT: _named_cfstring
# CHECK: (__TEXT,__text) section
# CHECK-NEXT: _foo1:
# CHECK-NEXT: _foo2: