llvm-project/lld/ELF/SymbolTable.cpp

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

329 lines
12 KiB
C++
Raw Normal View History

//===- SymbolTable.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
//
//===----------------------------------------------------------------------===//
2015-10-14 03:51:57 +08:00
//
// Symbol table is a bag of all known symbols. We put all symbols of
2016-01-06 04:47:37 +08:00
// all input files to the symbol table. The symbol table is basically
2015-10-14 03:51:57 +08:00
// a hash table with the logic to resolve symbol name conflicts using
// the symbol types.
//
//===----------------------------------------------------------------------===//
#include "SymbolTable.h"
#include "Config.h"
#include "LinkerScript.h"
#include "Symbols.h"
#include "SyntheticSections.h"
#include "lld/Common/ErrorHandler.h"
#include "lld/Common/Memory.h"
#include "lld/Common/Strings.h"
#include "llvm/ADT/STLExtras.h"
using namespace llvm;
using namespace llvm::object;
using namespace llvm::ELF;
using namespace lld;
using namespace lld::elf;
SymbolTable *elf::symtab;
void SymbolTable::wrap(Symbol *sym, Symbol *real, Symbol *wrap) {
// Redirect __real_foo to the original foo and foo to the original __wrap_foo.
int &idx1 = symMap[CachedHashStringRef(sym->getName())];
int &idx2 = symMap[CachedHashStringRef(real->getName())];
int &idx3 = symMap[CachedHashStringRef(wrap->getName())];
idx2 = idx1;
idx1 = idx3;
if (real->exportDynamic)
sym->exportDynamic = true;
if (!real->isUsedInRegularObj && sym->isUndefined())
sym->isUsedInRegularObj = false;
// Now renaming is complete, and no one refers to real. We drop real from
// .symtab and .dynsym. If real is undefined, it is important that we don't
// leave it in .dynsym, because otherwise it might lead to an undefined symbol
// error in a subsequent link. If real is defined, we could emit real as an
// alias for sym, but that could degrade the user experience of some tools
// that can print out only one symbol for each location: sym is a preferred
// name than real, but they might print out real instead.
memcpy(real, sym, sizeof(SymbolUnion));
real->isUsedInRegularObj = false;
}
// Find an existing symbol or create a new one.
Symbol *SymbolTable::insert(StringRef name) {
// <name>@@<version> means the symbol is the default version. In that
// case <name>@@<version> will be used to resolve references to <name>.
//
// Since this is a hot path, the following string search code is
// optimized for speed. StringRef::find(char) is much faster than
// StringRef::find(StringRef).
size_t pos = name.find('@');
if (pos != StringRef::npos && pos + 1 < name.size() && name[pos + 1] == '@')
name = name.take_front(pos);
[Coding style change] Rename variables so that they start with a lowercase letter This patch is mechanically generated by clang-llvm-rename tool that I wrote using Clang Refactoring Engine just for creating this patch. You can see the source code of the tool at https://reviews.llvm.org/D64123. There's no manual post-processing; you can generate the same patch by re-running the tool against lld's code base. Here is the main discussion thread to change the LLVM coding style: https://lists.llvm.org/pipermail/llvm-dev/2019-February/130083.html In the discussion thread, I proposed we use lld as a testbed for variable naming scheme change, and this patch does that. I chose to rename variables so that they are in camelCase, just because that is a minimal change to make variables to start with a lowercase letter. Note to downstream patch maintainers: if you are maintaining a downstream lld repo, just rebasing ahead of this commit would cause massive merge conflicts because this patch essentially changes every line in the lld subdirectory. But there's a remedy. clang-llvm-rename tool is a batch tool, so you can rename variables in your downstream repo with the tool. Given that, here is how to rebase your repo to a commit after the mass renaming: 1. rebase to the commit just before the mass variable renaming, 2. apply the tool to your downstream repo to mass-rename variables locally, and 3. rebase again to the head. Most changes made by the tool should be identical for a downstream repo and for the head, so at the step 3, almost all changes should be merged and disappear. I'd expect that there would be some lines that you need to merge by hand, but that shouldn't be too many. Differential Revision: https://reviews.llvm.org/D64121 llvm-svn: 365595
2019-07-10 13:00:37 +08:00
auto p = symMap.insert({CachedHashStringRef(name), (int)symVector.size()});
int &symIndex = p.first->second;
bool isNew = p.second;
[Coding style change] Rename variables so that they start with a lowercase letter This patch is mechanically generated by clang-llvm-rename tool that I wrote using Clang Refactoring Engine just for creating this patch. You can see the source code of the tool at https://reviews.llvm.org/D64123. There's no manual post-processing; you can generate the same patch by re-running the tool against lld's code base. Here is the main discussion thread to change the LLVM coding style: https://lists.llvm.org/pipermail/llvm-dev/2019-February/130083.html In the discussion thread, I proposed we use lld as a testbed for variable naming scheme change, and this patch does that. I chose to rename variables so that they are in camelCase, just because that is a minimal change to make variables to start with a lowercase letter. Note to downstream patch maintainers: if you are maintaining a downstream lld repo, just rebasing ahead of this commit would cause massive merge conflicts because this patch essentially changes every line in the lld subdirectory. But there's a remedy. clang-llvm-rename tool is a batch tool, so you can rename variables in your downstream repo with the tool. Given that, here is how to rebase your repo to a commit after the mass renaming: 1. rebase to the commit just before the mass variable renaming, 2. apply the tool to your downstream repo to mass-rename variables locally, and 3. rebase again to the head. Most changes made by the tool should be identical for a downstream repo and for the head, so at the step 3, almost all changes should be merged and disappear. I'd expect that there would be some lines that you need to merge by hand, but that shouldn't be too many. Differential Revision: https://reviews.llvm.org/D64121 llvm-svn: 365595
2019-07-10 13:00:37 +08:00
if (!isNew)
return symVector[symIndex];
[Coding style change] Rename variables so that they start with a lowercase letter This patch is mechanically generated by clang-llvm-rename tool that I wrote using Clang Refactoring Engine just for creating this patch. You can see the source code of the tool at https://reviews.llvm.org/D64123. There's no manual post-processing; you can generate the same patch by re-running the tool against lld's code base. Here is the main discussion thread to change the LLVM coding style: https://lists.llvm.org/pipermail/llvm-dev/2019-February/130083.html In the discussion thread, I proposed we use lld as a testbed for variable naming scheme change, and this patch does that. I chose to rename variables so that they are in camelCase, just because that is a minimal change to make variables to start with a lowercase letter. Note to downstream patch maintainers: if you are maintaining a downstream lld repo, just rebasing ahead of this commit would cause massive merge conflicts because this patch essentially changes every line in the lld subdirectory. But there's a remedy. clang-llvm-rename tool is a batch tool, so you can rename variables in your downstream repo with the tool. Given that, here is how to rebase your repo to a commit after the mass renaming: 1. rebase to the commit just before the mass variable renaming, 2. apply the tool to your downstream repo to mass-rename variables locally, and 3. rebase again to the head. Most changes made by the tool should be identical for a downstream repo and for the head, so at the step 3, almost all changes should be merged and disappear. I'd expect that there would be some lines that you need to merge by hand, but that shouldn't be too many. Differential Revision: https://reviews.llvm.org/D64121 llvm-svn: 365595
2019-07-10 13:00:37 +08:00
Symbol *sym = reinterpret_cast<Symbol *>(make<SymbolUnion>());
symVector.push_back(sym);
[Coding style change] Rename variables so that they start with a lowercase letter This patch is mechanically generated by clang-llvm-rename tool that I wrote using Clang Refactoring Engine just for creating this patch. You can see the source code of the tool at https://reviews.llvm.org/D64123. There's no manual post-processing; you can generate the same patch by re-running the tool against lld's code base. Here is the main discussion thread to change the LLVM coding style: https://lists.llvm.org/pipermail/llvm-dev/2019-February/130083.html In the discussion thread, I proposed we use lld as a testbed for variable naming scheme change, and this patch does that. I chose to rename variables so that they are in camelCase, just because that is a minimal change to make variables to start with a lowercase letter. Note to downstream patch maintainers: if you are maintaining a downstream lld repo, just rebasing ahead of this commit would cause massive merge conflicts because this patch essentially changes every line in the lld subdirectory. But there's a remedy. clang-llvm-rename tool is a batch tool, so you can rename variables in your downstream repo with the tool. Given that, here is how to rebase your repo to a commit after the mass renaming: 1. rebase to the commit just before the mass variable renaming, 2. apply the tool to your downstream repo to mass-rename variables locally, and 3. rebase again to the head. Most changes made by the tool should be identical for a downstream repo and for the head, so at the step 3, almost all changes should be merged and disappear. I'd expect that there would be some lines that you need to merge by hand, but that shouldn't be too many. Differential Revision: https://reviews.llvm.org/D64121 llvm-svn: 365595
2019-07-10 13:00:37 +08:00
// *sym was not initialized by a constructor. Fields that may get referenced
// when it is a placeholder must be initialized here.
Speed up --start-lib and --end-lib. --{start,end}-lib give files grouped by the options the archive file semantics. That is, each object file between them acts as if it were in an archive file whose sole member is the file. Therefore, files between --{start,end}-lib are linked to the final output only if they are needed to resolve some undefined symbols. Previously, the feature was implemented this way: 1. We read a symbol table and insert defined symbols to the symbol table as lazy symbols. 2. If an undefind symbol is resolved to a lazy symbol, that lazy symbol instantiate ObjFile class for that symbol, which re-insert all defined symbols to the symbol table. So, if an ObjFile is instantiated, defined symbols are inserted to the symbol table twice. Since inserting long symbol names is not cheap, there's a room to optimize here. This patch optimzies it. Now, LazyObjFile remembers symbol handles and passed them over to a new ObjFile instance, so that the ObjFile doesn't insert the same strings. Here is a quick benchmark to link clang. "Original" is the original lld with unmodified command line options. For "Case 1" and "Case 2", I extracted all files from archive files and replace .a's in a command line with .o's wrapped with --{start,end}-lib. I used the original lld for Case 1" and use this patch for Case 2. Original: 5.892 Case 1: 6.001 (+1.8%) Case 2: 5.701 (-3.2%) So, interestingly, --{start,end}-lib are now faster than the regular linking scheme with archive files. That's perhaps not too surprising, though, because for regular archive files, we look up the symbol table with the same string twice. Differential Revision: https://reviews.llvm.org/D62188 llvm-svn: 361473
2019-05-23 17:53:30 +08:00
sym->setName(name);
sym->symbolKind = Symbol::PlaceholderKind;
sym->versionId = VER_NDX_GLOBAL;
sym->visibility = STV_DEFAULT;
sym->isUsedInRegularObj = false;
sym->exportDynamic = false;
2019-08-13 17:12:52 +08:00
sym->inDynamicList = false;
sym->canInline = true;
sym->referenced = false;
sym->traced = false;
sym->scriptDefined = false;
sym->partition = 1;
return sym;
}
Symbol *SymbolTable::addSymbol(const Symbol &newSym) {
Symbol *sym = insert(newSym.getName());
sym->resolve(newSym);
return sym;
ELF: New symbol table design. This patch implements a new design for the symbol table that stores SymbolBodies within a memory region of the Symbol object. Symbols are mutated by constructing SymbolBodies in place over existing SymbolBodies, rather than by mutating pointers. As mentioned in the initial proposal [1], this memory layout helps reduce the cache miss rate by improving memory locality. Performance numbers: old(s) new(s) Without debug info: chrome 7.178 6.432 (-11.5%) LLVMgold.so 0.505 0.502 (-0.5%) clang 0.954 0.827 (-15.4%) llvm-as 0.052 0.045 (-15.5%) With debug info: scylla 5.695 5.613 (-1.5%) clang 14.396 14.143 (-1.8%) Performance counter results show that the fewer required indirections is indeed the cause of the improved performance. For example, when linking chrome, stalled cycles decreases from 14,556,444,002 to 12,959,238,310, and instructions per cycle increases from 0.78 to 0.83. We are also executing many fewer instructions (15,516,401,933 down to 15,002,434,310), probably because we spend less time allocating SymbolBodies. The new mechanism by which symbols are added to the symbol table is by calling add* functions on the SymbolTable. In this patch, I handle local symbols by storing them inside "unparented" SymbolBodies. This is suboptimal, but if we do want to try to avoid allocating these SymbolBodies, we can probably do that separately. I also removed a few members from the SymbolBody class that were only being used to pass information from the input file to the symbol table. This patch implements the new design for the ELF linker only. I intend to prepare a similar patch for the COFF linker. [1] http://lists.llvm.org/pipermail/llvm-dev/2016-April/098832.html Differential Revision: http://reviews.llvm.org/D19752 llvm-svn: 268178
2016-05-01 12:55:03 +08:00
}
Symbol *SymbolTable::find(StringRef name) {
auto it = symMap.find(CachedHashStringRef(name));
if (it == symMap.end())
return nullptr;
Symbol *sym = symVector[it->second];
if (sym->isPlaceholder())
return nullptr;
return sym;
}
// A version script/dynamic list is only meaningful for a Defined symbol.
// A CommonSymbol will be converted to a Defined in replaceCommonSymbols().
// A lazy symbol may be made Defined if an LTO libcall extracts it.
static bool canBeVersioned(const Symbol &sym) {
return sym.isDefined() || sym.isCommon() || sym.isLazy();
}
// Initialize demangledSyms with a map from demangled symbols to symbol
// objects. Used to handle "extern C++" directive in version scripts.
//
// The map will contain all demangled symbols. That can be very large,
// and in LLD we generally want to avoid do anything for each symbol.
// Then, why are we doing this? Here's why.
//
// Users can use "extern C++ {}" directive to match against demangled
// C++ symbols. For example, you can write a pattern such as
// "llvm::*::foo(int, ?)". Obviously, there's no way to handle this
// other than trying to match a pattern against all demangled symbols.
// So, if "extern C++" feature is used, we need to demangle all known
// symbols.
StringMap<std::vector<Symbol *>> &SymbolTable::getDemangledSyms() {
if (!demangledSyms) {
demangledSyms.emplace();
std::string demangled;
for (Symbol *sym : symVector)
if (canBeVersioned(*sym)) {
StringRef name = sym->getName();
size_t pos = name.find('@');
if (pos == std::string::npos)
demangled = demangleItanium(name);
else if (pos + 1 == name.size() || name[pos + 1] == '@')
demangled = demangleItanium(name.substr(0, pos));
else
demangled =
(demangleItanium(name.substr(0, pos)) + name.substr(pos)).str();
(*demangledSyms)[demangled].push_back(sym);
}
}
return *demangledSyms;
}
std::vector<Symbol *> SymbolTable::findByVersion(SymbolVersion ver) {
if (ver.isExternCpp)
return getDemangledSyms().lookup(ver.name);
if (Symbol *sym = find(ver.name))
if (canBeVersioned(*sym))
return {sym};
return {};
}
std::vector<Symbol *> SymbolTable::findAllByVersion(SymbolVersion ver,
bool includeNonDefault) {
std::vector<Symbol *> res;
SingleStringMatcher m(ver.name);
auto check = [&](StringRef name) {
size_t pos = name.find('@');
if (!includeNonDefault)
return pos == StringRef::npos;
return !(pos + 1 < name.size() && name[pos + 1] == '@');
};
if (ver.isExternCpp) {
for (auto &p : getDemangledSyms())
if (m.match(p.first()))
for (Symbol *sym : p.second)
if (check(sym->getName()))
res.push_back(sym);
return res;
}
for (Symbol *sym : symVector)
if (canBeVersioned(*sym) && check(sym->getName()) &&
m.match(sym->getName()))
res.push_back(sym);
return res;
}
void SymbolTable::handleDynamicList() {
for (SymbolVersion &ver : config->dynamicList) {
std::vector<Symbol *> syms;
if (ver.hasWildcard)
syms = findAllByVersion(ver, /*includeNonDefault=*/true);
else
syms = findByVersion(ver);
2019-08-13 17:12:52 +08:00
for (Symbol *sym : syms)
sym->inDynamicList = true;
}
}
// Set symbol versions to symbols. This function handles patterns containing no
// wildcard characters. Return false if no symbol definition matches ver.
bool SymbolTable::assignExactVersion(SymbolVersion ver, uint16_t versionId,
StringRef versionName,
bool includeNonDefault) {
// Get a list of symbols which we need to assign the version to.
std::vector<Symbol *> syms = findByVersion(ver);
auto getName = [](uint16_t ver) -> std::string {
if (ver == VER_NDX_LOCAL)
return "VER_NDX_LOCAL";
if (ver == VER_NDX_GLOBAL)
return "VER_NDX_GLOBAL";
return ("version '" + config->versionDefinitions[ver].name + "'").str();
};
// Assign the version.
for (Symbol *sym : syms) {
// For a non-local versionId, skip symbols containing version info because
// symbol versions specified by symbol names take precedence over version
// scripts. See parseSymbolVersion().
if (!includeNonDefault && versionId != VER_NDX_LOCAL &&
sym->getName().contains('@'))
continue;
// If the version has not been assigned, verdefIndex is -1. Use an arbitrary
// number (0) to indicate the version has been assigned.
if (sym->verdefIndex == UINT32_C(-1)) {
sym->verdefIndex = 0;
sym->versionId = versionId;
}
if (sym->versionId == versionId)
continue;
warn("attempt to reassign symbol '" + ver.name + "' of " +
getName(sym->versionId) + " to " + getName(versionId));
}
return !syms.empty();
}
void SymbolTable::assignWildcardVersion(SymbolVersion ver, uint16_t versionId,
bool includeNonDefault) {
// Exact matching takes precedence over fuzzy matching,
// so we set a version to a symbol only if no version has been assigned
// to the symbol. This behavior is compatible with GNU.
for (Symbol *sym : findAllByVersion(ver, includeNonDefault))
if (sym->verdefIndex == UINT32_C(-1)) {
sym->verdefIndex = 0;
sym->versionId = versionId;
}
}
// This function processes version scripts by updating the versionId
2016-09-03 06:15:08 +08:00
// member of symbols.
// If there's only one anonymous version definition in a version
// script file, the script does not actually define any symbol version,
// but just specifies symbols visibilities.
void SymbolTable::scanVersionScript() {
SmallString<128> buf;
2016-09-03 06:15:08 +08:00
// First, we assign versions to exact matching symbols,
// i.e. version definitions not containing any glob meta-characters.
std::vector<Symbol *> syms;
for (VersionDefinition &v : config->versionDefinitions) {
auto assignExact = [&](SymbolVersion pat, uint16_t id, StringRef ver) {
bool found =
assignExactVersion(pat, id, ver, /*includeNonDefault=*/false);
buf.clear();
found |= assignExactVersion({(pat.name + "@" + v.name).toStringRef(buf),
pat.isExternCpp, /*hasWildCard=*/false},
id, ver, /*includeNonDefault=*/true);
if (!found && !config->undefinedVersion)
errorOrWarn("version script assignment of '" + ver + "' to symbol '" +
pat.name + "' failed: symbol not defined");
};
for (SymbolVersion &pat : v.nonLocalPatterns)
if (!pat.hasWildcard)
assignExact(pat, v.id, v.name);
for (SymbolVersion pat : v.localPatterns)
if (!pat.hasWildcard)
assignExact(pat, VER_NDX_LOCAL, "local");
}
// Next, assign versions to wildcards that are not "*". Note that because the
// last match takes precedence over previous matches, we iterate over the
// definitions in the reverse order.
auto assignWildcard = [&](SymbolVersion pat, uint16_t id, StringRef ver) {
assignWildcardVersion(pat, id, /*includeNonDefault=*/false);
buf.clear();
assignWildcardVersion({(pat.name + "@" + ver).toStringRef(buf),
pat.isExternCpp, /*hasWildCard=*/true},
id,
/*includeNonDefault=*/true);
};
for (VersionDefinition &v : llvm::reverse(config->versionDefinitions)) {
for (SymbolVersion &pat : v.nonLocalPatterns)
if (pat.hasWildcard && pat.name != "*")
assignWildcard(pat, v.id, v.name);
for (SymbolVersion &pat : v.localPatterns)
if (pat.hasWildcard && pat.name != "*")
assignWildcard(pat, VER_NDX_LOCAL, v.name);
}
// Then, assign versions to "*". In GNU linkers they have lower priority than
// other wildcards.
for (VersionDefinition &v : config->versionDefinitions) {
for (SymbolVersion &pat : v.nonLocalPatterns)
if (pat.hasWildcard && pat.name == "*")
assignWildcard(pat, v.id, v.name);
for (SymbolVersion &pat : v.localPatterns)
if (pat.hasWildcard && pat.name == "*")
assignWildcard(pat, VER_NDX_LOCAL, v.name);
}
// Symbol themselves might know their versions because symbols
// can contain versions in the form of <name>@<version>.
// Let them parse and update their names to exclude version suffix.
for (Symbol *sym : symVector)
sym->parseSymbolVersion();
// isPreemptible is false at this point. To correctly compute the binding of a
// Defined (which is used by includeInDynsym()), we need to know if it is
// VER_NDX_LOCAL or not. Compute symbol versions before handling
// --dynamic-list.
handleDynamicList();
}