mirror of
https://github.com/intel/llvm.git
synced 2026-01-13 19:08:21 +08:00
Add the `--{no-}separate-cstring-literal-sections` option to emit
cstring literals into sections defined by their section name. This
allows for changes like https://github.com/swiftlang/swift/pull/84300
and https://github.com/swiftlang/swift/pull/84236 to actually have an
affect. The default behavior has not changed.
The reason this is useful is because strings in different sections might
have different access patterns at runtime. By splitting these strings
into separate sections, we may reduce the number of page faults during
startup. For example, the ObjC runtime accesses all strings in
`__objc_classname` before main.
293 lines
11 KiB
C++
293 lines
11 KiB
C++
//===- MapFile.cpp --------------------------------------------------------===//
|
|
//
|
|
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
|
|
// See https://llvm.org/LICENSE.txt for license information.
|
|
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
//
|
|
// This file implements the -map option, which maps address ranges to their
|
|
// respective contents, plus the input file these contents were originally from.
|
|
// The contents (typically symbols) are listed in address order. Dead-stripped
|
|
// contents are included as well.
|
|
//
|
|
// # Path: test
|
|
// # Arch: x86_84
|
|
// # Object files:
|
|
// [ 0] linker synthesized
|
|
// [ 1] a.o
|
|
// # Sections:
|
|
// # Address Size Segment Section
|
|
// 0x1000005C0 0x0000004C __TEXT __text
|
|
// # Symbols:
|
|
// # Address Size File Name
|
|
// 0x1000005C0 0x00000001 [ 1] _main
|
|
// # Dead Stripped Symbols:
|
|
// # Size File Name
|
|
// <<dead>> 0x00000001 [ 1] _foo
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
#include "MapFile.h"
|
|
#include "ConcatOutputSection.h"
|
|
#include "Config.h"
|
|
#include "InputFiles.h"
|
|
#include "InputSection.h"
|
|
#include "OutputSegment.h"
|
|
#include "Symbols.h"
|
|
#include "SyntheticSections.h"
|
|
#include "Target.h"
|
|
#include "lld/Common/ErrorHandler.h"
|
|
#include "llvm/ADT/DenseMap.h"
|
|
#include "llvm/Support/Parallel.h"
|
|
#include "llvm/Support/TimeProfiler.h"
|
|
|
|
using namespace llvm;
|
|
using namespace llvm::sys;
|
|
using namespace lld;
|
|
using namespace lld::macho;
|
|
|
|
struct CStringInfo {
|
|
uint32_t fileIndex;
|
|
StringRef str;
|
|
};
|
|
|
|
struct MapInfo {
|
|
SmallVector<InputFile *> files;
|
|
SmallVector<Defined *> deadSymbols;
|
|
DenseMap<const OutputSection *,
|
|
SmallVector<std::pair<uint64_t /*addr*/, CStringInfo>>>
|
|
liveCStringsForSection;
|
|
SmallVector<CStringInfo> deadCStrings;
|
|
};
|
|
|
|
static MapInfo gatherMapInfo() {
|
|
MapInfo info;
|
|
for (InputFile *file : inputFiles) {
|
|
bool isReferencedFile = false;
|
|
|
|
if (isa<ObjFile>(file) || isa<BitcodeFile>(file)) {
|
|
uint32_t fileIndex = info.files.size() + 1;
|
|
|
|
// Gather the dead symbols. We don't have to bother with the live ones
|
|
// because we will pick them up as we iterate over the OutputSections
|
|
// later.
|
|
for (Symbol *sym : file->symbols) {
|
|
if (auto *d = dyn_cast_or_null<Defined>(sym))
|
|
// Only emit the prevailing definition of a symbol. Also, don't emit
|
|
// the symbol if it is part of a cstring section (we use the literal
|
|
// value instead, similar to ld64)
|
|
if (d->isec() && d->getFile() == file &&
|
|
!isa<CStringInputSection>(d->isec())) {
|
|
isReferencedFile = true;
|
|
if (!d->isLive())
|
|
info.deadSymbols.push_back(d);
|
|
}
|
|
}
|
|
|
|
// Gather all the cstrings (both live and dead). A CString(Output)Section
|
|
// doesn't provide us a way of figuring out which InputSections its
|
|
// cstring contents came from, so we need to build up that mapping here.
|
|
for (const Section *sec : file->sections) {
|
|
for (const Subsection &subsec : sec->subsections) {
|
|
if (auto isec = dyn_cast<CStringInputSection>(subsec.isec)) {
|
|
auto &liveCStrings = info.liveCStringsForSection[isec->parent];
|
|
for (const auto &[i, piece] : llvm::enumerate(isec->pieces)) {
|
|
if (piece.live)
|
|
liveCStrings.push_back({isec->parent->addr + piece.outSecOff,
|
|
{fileIndex, isec->getStringRef(i)}});
|
|
else
|
|
info.deadCStrings.push_back({fileIndex, isec->getStringRef(i)});
|
|
isReferencedFile = true;
|
|
}
|
|
} else {
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
} else if (const auto *dylibFile = dyn_cast<DylibFile>(file)) {
|
|
isReferencedFile = dylibFile->isReferenced();
|
|
}
|
|
|
|
if (isReferencedFile)
|
|
info.files.push_back(file);
|
|
}
|
|
|
|
// cstrings are not stored in sorted order in their OutputSections, so we sort
|
|
// them here.
|
|
for (auto &liveCStrings : info.liveCStringsForSection)
|
|
parallelSort(liveCStrings.second, llvm::less_first());
|
|
return info;
|
|
}
|
|
|
|
// We use this instead of `toString(const InputFile *)` as we don't want to
|
|
// include the dylib install name in our output.
|
|
static void printFileName(raw_fd_ostream &os, const InputFile *f) {
|
|
if (f->archiveName.empty())
|
|
os << f->getName();
|
|
else
|
|
os << f->archiveName << "(" << path::filename(f->getName()) + ")";
|
|
}
|
|
|
|
// For printing the contents of the __stubs and __la_symbol_ptr sections.
|
|
static void printStubsEntries(
|
|
raw_fd_ostream &os,
|
|
const DenseMap<lld::macho::InputFile *, uint32_t> &readerToFileOrdinal,
|
|
const OutputSection *osec, size_t entrySize) {
|
|
for (const Symbol *sym : in.stubs->getEntries())
|
|
os << format("0x%08llX\t0x%08zX\t[%3u] %s\n",
|
|
osec->addr + sym->stubsIndex * entrySize, entrySize,
|
|
readerToFileOrdinal.lookup(sym->getFile()),
|
|
sym->getName().str().data());
|
|
}
|
|
|
|
static void printNonLazyPointerSection(raw_fd_ostream &os,
|
|
NonLazyPointerSectionBase *osec) {
|
|
// ld64 considers stubs to belong to particular files, but considers GOT
|
|
// entries to be linker-synthesized. Not sure why they made that decision, but
|
|
// I think we can follow suit unless there's demand for better symbol-to-file
|
|
// associations.
|
|
for (const Symbol *sym : osec->getEntries())
|
|
os << format("0x%08llX\t0x%08zX\t[ 0] non-lazy-pointer-to-local: %s\n",
|
|
osec->addr + sym->gotIndex * target->wordSize,
|
|
target->wordSize, sym->getName().str().data());
|
|
}
|
|
|
|
static uint64_t getSymSizeForMap(Defined *sym) {
|
|
if (sym->identicalCodeFoldingKind == Symbol::ICFFoldKind::Body)
|
|
return 0;
|
|
return sym->size;
|
|
}
|
|
|
|
void macho::writeMapFile() {
|
|
if (config->mapFile.empty())
|
|
return;
|
|
|
|
TimeTraceScope timeScope("Write map file");
|
|
|
|
// Open a map file for writing.
|
|
std::error_code ec;
|
|
raw_fd_ostream os(config->mapFile, ec, sys::fs::OF_None);
|
|
if (ec) {
|
|
error("cannot open " + config->mapFile + ": " + ec.message());
|
|
return;
|
|
}
|
|
|
|
os << format("# Path: %s\n", config->outputFile.str().c_str());
|
|
os << format("# Arch: %s\n",
|
|
getArchitectureName(config->arch()).str().c_str());
|
|
|
|
MapInfo info = gatherMapInfo();
|
|
|
|
os << "# Object files:\n";
|
|
os << format("[%3u] %s\n", 0, (const char *)"linker synthesized");
|
|
uint32_t fileIndex = 1;
|
|
DenseMap<lld::macho::InputFile *, uint32_t> readerToFileOrdinal;
|
|
for (InputFile *file : info.files) {
|
|
os << format("[%3u] ", fileIndex);
|
|
printFileName(os, file);
|
|
os << "\n";
|
|
readerToFileOrdinal[file] = fileIndex++;
|
|
}
|
|
|
|
os << "# Sections:\n";
|
|
os << "# Address\tSize \tSegment\tSection\n";
|
|
for (OutputSegment *seg : outputSegments)
|
|
for (OutputSection *osec : seg->getSections()) {
|
|
if (osec->isHidden())
|
|
continue;
|
|
|
|
os << format("0x%08llX\t0x%08llX\t%s\t%s\n", osec->addr, osec->getSize(),
|
|
seg->name.str().c_str(), osec->name.str().c_str());
|
|
}
|
|
|
|
// Helper lambda that prints all symbols from one ConcatInputSection.
|
|
auto printOne = [&](const ConcatInputSection *isec) {
|
|
for (Defined *sym : isec->symbols) {
|
|
if (!(isPrivateLabel(sym->getName()) && getSymSizeForMap(sym) == 0)) {
|
|
os << format("0x%08llX\t0x%08llX\t[%3u] %s\n", sym->getVA(),
|
|
getSymSizeForMap(sym),
|
|
readerToFileOrdinal.lookup(sym->getFile()),
|
|
sym->getName().str().data());
|
|
}
|
|
}
|
|
};
|
|
// Shared function to print one or two arrays of ConcatInputSection in
|
|
// ascending outSecOff order. The second array is optional; if provided, we
|
|
// interleave the printing in sorted order without allocating a merged temp
|
|
// array.
|
|
auto printIsecArrSyms = [&](ArrayRef<ConcatInputSection *> arr1,
|
|
ArrayRef<ConcatInputSection *> arr2 = {}) {
|
|
// Print both arrays in sorted order, interleaving as necessary.
|
|
while (!arr1.empty() || !arr2.empty()) {
|
|
if (!arr1.empty() && (arr2.empty() || arr1.front()->outSecOff <=
|
|
arr2.front()->outSecOff)) {
|
|
printOne(arr1.front());
|
|
arr1 = arr1.drop_front();
|
|
} else if (!arr2.empty()) {
|
|
printOne(arr2.front());
|
|
arr2 = arr2.drop_front();
|
|
}
|
|
}
|
|
};
|
|
|
|
os << "# Symbols:\n";
|
|
os << "# Address\tSize \tFile Name\n";
|
|
for (const OutputSegment *seg : outputSegments) {
|
|
for (const OutputSection *osec : seg->getSections()) {
|
|
if (auto *textOsec = dyn_cast<TextOutputSection>(osec)) {
|
|
printIsecArrSyms(textOsec->inputs, textOsec->getThunks());
|
|
} else if (auto *concatOsec = dyn_cast<ConcatOutputSection>(osec)) {
|
|
printIsecArrSyms(concatOsec->inputs);
|
|
} else if (is_contained(in.cStringSections, osec)) {
|
|
const auto &liveCStrings = info.liveCStringsForSection.lookup(osec);
|
|
uint64_t lastAddr = 0; // strings will never start at address 0, so this
|
|
// is a sentinel value
|
|
for (const auto &[addr, info] : liveCStrings) {
|
|
uint64_t size = 0;
|
|
if (addr != lastAddr)
|
|
size = info.str.size() + 1; // include null terminator
|
|
lastAddr = addr;
|
|
os << format("0x%08llX\t0x%08llX\t[%3u] literal string: ", addr, size,
|
|
info.fileIndex);
|
|
os.write_escaped(info.str) << "\n";
|
|
}
|
|
} else if (osec == (void *)in.unwindInfo) {
|
|
os << format("0x%08llX\t0x%08llX\t[ 0] compact unwind info\n",
|
|
osec->addr, osec->getSize());
|
|
} else if (osec == in.stubs) {
|
|
printStubsEntries(os, readerToFileOrdinal, osec, target->stubSize);
|
|
} else if (osec == in.lazyPointers) {
|
|
printStubsEntries(os, readerToFileOrdinal, osec, target->wordSize);
|
|
} else if (osec == in.stubHelper) {
|
|
// yes, ld64 calls it "helper helper"...
|
|
os << format("0x%08llX\t0x%08llX\t[ 0] helper helper\n", osec->addr,
|
|
osec->getSize());
|
|
} else if (osec == in.got) {
|
|
printNonLazyPointerSection(os, in.got);
|
|
} else if (osec == in.tlvPointers) {
|
|
printNonLazyPointerSection(os, in.tlvPointers);
|
|
} else if (osec == in.objcMethList) {
|
|
printIsecArrSyms(in.objcMethList->getInputs());
|
|
}
|
|
// TODO print other synthetic sections
|
|
}
|
|
}
|
|
|
|
if (config->deadStrip) {
|
|
os << "# Dead Stripped Symbols:\n";
|
|
os << "# \tSize \tFile Name\n";
|
|
for (Defined *sym : info.deadSymbols) {
|
|
assert(!sym->isLive());
|
|
os << format("<<dead>>\t0x%08llX\t[%3u] %s\n", getSymSizeForMap(sym),
|
|
readerToFileOrdinal[sym->getFile()],
|
|
sym->getName().str().data());
|
|
}
|
|
for (CStringInfo &cstrInfo : info.deadCStrings) {
|
|
os << format("<<dead>>\t0x%08zX\t[%3u] literal string: ",
|
|
cstrInfo.str.size() + 1, cstrInfo.fileIndex);
|
|
os.write_escaped(cstrInfo.str) << "\n";
|
|
}
|
|
}
|
|
}
|