1228 lines · cpp
1//===-- gold-plugin.cpp - Plugin to gold for Link Time Optimization ------===//2//3// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.4// See https://llvm.org/LICENSE.txt for license information.5// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception6//7//===----------------------------------------------------------------------===//8//9// This is a gold plugin for LLVM. It provides an LLVM implementation of the10// interface described in http://gcc.gnu.org/wiki/whopr/driver .11//12//===----------------------------------------------------------------------===//13 14#include "llvm/ADT/ScopeExit.h"15#include "llvm/ADT/Statistic.h"16#include "llvm/Bitcode/BitcodeReader.h"17#include "llvm/Bitcode/BitcodeWriter.h"18#include "llvm/CodeGen/CommandFlags.h"19#include "llvm/Config/config.h" // plugin-api.h requires HAVE_STDINT_H20#include "llvm/Config/llvm-config.h"21#include "llvm/IR/Constants.h"22#include "llvm/IR/DiagnosticPrinter.h"23#include "llvm/LTO/LTO.h"24#include "llvm/Object/Error.h"25#include "llvm/Remarks/HotnessThresholdParser.h"26#include "llvm/Support/CachePruning.h"27#include "llvm/Support/Caching.h"28#include "llvm/Support/CommandLine.h"29#include "llvm/Support/FileSystem.h"30#include "llvm/Support/ManagedStatic.h"31#include "llvm/Support/MemoryBuffer.h"32#include "llvm/Support/Path.h"33#include "llvm/Support/TargetSelect.h"34#include "llvm/Support/Threading.h"35#include "llvm/Support/TimeProfiler.h"36#include "llvm/Support/raw_ostream.h"37#include "llvm/TargetParser/Host.h"38#include <list>39#include <plugin-api.h>40#include <string>41#include <system_error>42#include <utility>43#include <vector>44 45// FIXME: remove this declaration when we stop maintaining Ubuntu Quantal and46// Precise and Debian Wheezy (binutils 2.23 is required)47#define LDPO_PIE 348 49#define LDPT_GET_SYMBOLS_V3 2850 51// FIXME: Remove when binutils 2.31 (containing gold 1.16) is the minimum52// required version.53#define LDPT_GET_WRAP_SYMBOLS 3254 55using namespace llvm;56using namespace lto;57 58static codegen::RegisterCodeGenFlags CodeGenFlags;59 60// FIXME: Remove when binutils 2.31 (containing gold 1.16) is the minimum61// required version.62typedef enum ld_plugin_status (*ld_plugin_get_wrap_symbols)(63 uint64_t *num_symbols, const char ***wrap_symbol_list);64 65static ld_plugin_status discard_message(int level, const char *format, ...) {66 // Die loudly. Recent versions of Gold pass ld_plugin_message as the first67 // callback in the transfer vector. This should never be called.68 abort();69}70 71static ld_plugin_release_input_file release_input_file = nullptr;72static ld_plugin_get_input_file get_input_file = nullptr;73static ld_plugin_message message = discard_message;74static ld_plugin_get_wrap_symbols get_wrap_symbols = nullptr;75 76namespace {77struct claimed_file {78 void *handle;79 void *leader_handle;80 std::vector<ld_plugin_symbol> syms;81 off_t filesize;82 std::string name;83};84 85/// RAII wrapper to manage opening and releasing of a ld_plugin_input_file.86struct PluginInputFile {87 void *Handle;88 std::unique_ptr<ld_plugin_input_file> File;89 90 PluginInputFile(void *Handle) : Handle(Handle) {91 File = std::make_unique<ld_plugin_input_file>();92 if (get_input_file(Handle, File.get()) != LDPS_OK)93 message(LDPL_FATAL, "Failed to get file information");94 }95 ~PluginInputFile() {96 // File would have been reset to nullptr if we moved this object97 // to a new owner.98 if (File)99 if (release_input_file(Handle) != LDPS_OK)100 message(LDPL_FATAL, "Failed to release file information");101 }102 103 ld_plugin_input_file &file() { return *File; }104 105 PluginInputFile(PluginInputFile &&RHS) = default;106 PluginInputFile &operator=(PluginInputFile &&RHS) = default;107};108 109struct ResolutionInfo {110 bool CanOmitFromDynSym = true;111 bool DefaultVisibility = true;112 bool CanInline = true;113 bool IsUsedInRegularObj = false;114};115 116}117 118static ld_plugin_add_symbols add_symbols = nullptr;119static ld_plugin_get_symbols get_symbols = nullptr;120static ld_plugin_add_input_file add_input_file = nullptr;121static ld_plugin_set_extra_library_path set_extra_library_path = nullptr;122static ld_plugin_get_view get_view = nullptr;123static bool IsExecutable = false;124static bool SplitSections = true;125static std::optional<Reloc::Model> RelocationModel;126static std::string output_name = "";127static std::list<claimed_file> Modules;128static DenseMap<int, void *> FDToLeaderHandle;129static StringMap<ResolutionInfo> ResInfo;130static std::vector<std::string> Cleanup;131 132namespace options {133 enum OutputType {134 OT_NORMAL,135 OT_DISABLE,136 OT_BC_ONLY,137 OT_ASM_ONLY,138 OT_SAVE_TEMPS139 };140 static OutputType TheOutputType = OT_NORMAL;141 static unsigned OptLevel = 2;142 // Currently only affects ThinLTO, where the default is the max cores in the143 // system. See llvm::get_threadpool_strategy() for acceptable values.144 static std::string Parallelism;145 // Default regular LTO codegen parallelism (number of partitions).146 static unsigned ParallelCodeGenParallelismLevel = 1;147#ifdef NDEBUG148 static bool DisableVerify = true;149#else150 static bool DisableVerify = false;151#endif152 static std::string obj_path;153 static std::string extra_library_path;154 static std::string triple;155 static std::string mcpu;156 // Tells plugin to use unified lto157 static bool unifiedlto = false;158 // When the thinlto plugin option is specified, only read the function159 // the information from intermediate files and write a combined160 // global index for the ThinLTO backends.161 static bool thinlto = false;162 // If false, all ThinLTO backend compilations through code gen are performed163 // using multiple threads in the gold-plugin, before handing control back to164 // gold. If true, write individual backend index files which reflect165 // the import decisions, and exit afterwards. The assumption is166 // that the build system will launch the backend processes.167 static bool thinlto_index_only = false;168 // If non-empty, holds the name of a file in which to write the list of169 // oject files gold selected for inclusion in the link after symbol170 // resolution (i.e. they had selected symbols). This will only be non-empty171 // in the thinlto_index_only case. It is used to identify files, which may172 // have originally been within archive libraries specified via173 // --start-lib/--end-lib pairs, that should be included in the final174 // native link process (since intervening function importing and inlining175 // may change the symbol resolution detected in the final link and which176 // files to include out of --start-lib/--end-lib libraries as a result).177 static std::string thinlto_linked_objects_file;178 // If true, when generating individual index files for distributed backends,179 // also generate a "${bitcodefile}.imports" file at the same location for each180 // bitcode file, listing the files it imports from in plain text. This is to181 // support distributed build file staging.182 static bool thinlto_emit_imports_files = false;183 // Option to control where files for a distributed backend (the individual184 // index files and optional imports files) are created.185 // If specified, expects a string of the form "oldprefix:newprefix", and186 // instead of generating these files in the same directory path as the187 // corresponding bitcode file, will use a path formed by replacing the188 // bitcode file's path prefix matching oldprefix with newprefix.189 static std::string thinlto_prefix_replace;190 // Option to control the name of modules encoded in the individual index191 // files for a distributed backend. This enables the use of minimized192 // bitcode files for the thin link, assuming the name of the full bitcode193 // file used in the backend differs just in some part of the file suffix.194 // If specified, expects a string of the form "oldsuffix:newsuffix".195 static std::string thinlto_object_suffix_replace;196 // Optional path to a directory for caching ThinLTO objects.197 static std::string cache_dir;198 // Optional pruning policy for ThinLTO caches.199 static std::string cache_policy;200 // Additional options to pass into the code generator.201 // Note: This array will contain all plugin options which are not claimed202 // as plugin exclusive to pass to the code generator.203 static std::vector<const char *> extra;204 // Sample profile file path205 static std::string sample_profile;206 // Debug new pass manager207 static bool debug_pass_manager = false;208 // Directory to store the .dwo files.209 static std::string dwo_dir;210 /// Statistics output filename.211 static std::string stats_file;212 // Asserts that LTO link has whole program visibility213 static bool whole_program_visibility = false;214 215 // Optimization remarks filename, accepted passes and hotness options216 static std::string RemarksFilename;217 static std::string RemarksPasses;218 static bool RemarksWithHotness = false;219 static std::optional<uint64_t> RemarksHotnessThreshold = 0;220 static std::string RemarksFormat;221 222 // Context sensitive PGO options.223 static std::string cs_profile_path;224 static bool cs_pgo_gen = false;225 226 // When true, MergeFunctions pass is used in LTO link pipeline.227 static bool merge_functions = false;228 229 // Time trace options.230 static std::string time_trace_file;231 static unsigned time_trace_granularity = 500;232 233 static void process_plugin_option(const char *opt_)234 {235 if (opt_ == nullptr)236 return;237 llvm::StringRef opt = opt_;238 239 if (opt.consume_front("mcpu=")) {240 mcpu = std::string(opt);241 } else if (opt.consume_front("extra-library-path=")) {242 extra_library_path = std::string(opt);243 } else if (opt.consume_front("mtriple=")) {244 triple = std::string(opt);245 } else if (opt.consume_front("obj-path=")) {246 obj_path = std::string(opt);247 } else if (opt == "emit-llvm") {248 TheOutputType = OT_BC_ONLY;249 } else if (opt == "save-temps") {250 TheOutputType = OT_SAVE_TEMPS;251 } else if (opt == "disable-output") {252 TheOutputType = OT_DISABLE;253 } else if (opt == "emit-asm") {254 TheOutputType = OT_ASM_ONLY;255 } else if (opt == "unifiedlto") {256 unifiedlto = true;257 } else if (opt == "thinlto") {258 thinlto = true;259 } else if (opt == "thinlto-index-only") {260 thinlto_index_only = true;261 } else if (opt.consume_front("thinlto-index-only=")) {262 thinlto_index_only = true;263 thinlto_linked_objects_file = std::string(opt);264 } else if (opt == "thinlto-emit-imports-files") {265 thinlto_emit_imports_files = true;266 } else if (opt.consume_front("thinlto-prefix-replace=")) {267 thinlto_prefix_replace = std::string(opt);268 if (thinlto_prefix_replace.find(';') == std::string::npos)269 message(LDPL_FATAL, "thinlto-prefix-replace expects 'old;new' format");270 } else if (opt.consume_front("thinlto-object-suffix-replace=")) {271 thinlto_object_suffix_replace = std::string(opt);272 if (thinlto_object_suffix_replace.find(';') == std::string::npos)273 message(LDPL_FATAL,274 "thinlto-object-suffix-replace expects 'old;new' format");275 } else if (opt.consume_front("cache-dir=")) {276 cache_dir = std::string(opt);277 } else if (opt.consume_front("cache-policy=")) {278 cache_policy = std::string(opt);279 } else if (opt.size() == 2 && opt[0] == 'O') {280 if (opt[1] < '0' || opt[1] > '3')281 message(LDPL_FATAL, "Optimization level must be between 0 and 3");282 OptLevel = opt[1] - '0';283 } else if (opt.consume_front("jobs=")) {284 Parallelism = std::string(opt);285 if (!get_threadpool_strategy(opt))286 message(LDPL_FATAL, "Invalid parallelism level: %s",287 Parallelism.c_str());288 } else if (opt.consume_front("lto-partitions=")) {289 if (opt.getAsInteger(10, ParallelCodeGenParallelismLevel))290 message(LDPL_FATAL, "Invalid codegen partition level: %s", opt_ + 5);291 } else if (opt == "disable-verify") {292 DisableVerify = true;293 } else if (opt.consume_front("sample-profile=")) {294 sample_profile = std::string(opt);295 } else if (opt == "cs-profile-generate") {296 cs_pgo_gen = true;297 } else if (opt == "merge-functions") {298 merge_functions = true;299 } else if (opt.consume_front("cs-profile-path=")) {300 cs_profile_path = std::string(opt);301 } else if (opt == "new-pass-manager") {302 // We always use the new pass manager.303 } else if (opt == "debug-pass-manager") {304 debug_pass_manager = true;305 } else if (opt == "whole-program-visibility") {306 whole_program_visibility = true;307 } else if (opt.consume_front("dwo_dir=")) {308 dwo_dir = std::string(opt);309 } else if (opt.consume_front("opt-remarks-filename=")) {310 RemarksFilename = std::string(opt);311 } else if (opt.consume_front("opt-remarks-passes=")) {312 RemarksPasses = std::string(opt);313 } else if (opt == "opt-remarks-with-hotness") {314 RemarksWithHotness = true;315 } else if (opt.consume_front("opt-remarks-hotness-threshold=")) {316 auto ResultOrErr = remarks::parseHotnessThresholdOption(opt);317 if (!ResultOrErr)318 message(LDPL_FATAL, "Invalid remarks hotness threshold: %s",319 opt.data());320 else321 RemarksHotnessThreshold = *ResultOrErr;322 } else if (opt.consume_front("opt-remarks-format=")) {323 RemarksFormat = std::string(opt);324 } else if (opt.consume_front("stats-file=")) {325 stats_file = std::string(opt);326 } else if (opt.consume_front("time-trace=")) {327 time_trace_file = std::string(opt);328 } else if (opt.consume_front("time-trace-granularity=")) {329 unsigned Granularity;330 if (opt.getAsInteger(10, Granularity))331 message(LDPL_FATAL, "Invalid time trace granularity: %s", opt.data());332 else333 time_trace_granularity = Granularity;334 } else {335 // Save this option to pass to the code generator.336 // ParseCommandLineOptions() expects argv[0] to be program name. Lazily337 // add that.338 if (extra.empty())339 extra.push_back("LLVMgold");340 341 extra.push_back(opt_);342 }343 }344}345 346static ld_plugin_status claim_file_hook(const ld_plugin_input_file *file,347 int *claimed);348static ld_plugin_status all_symbols_read_hook(void);349static ld_plugin_status cleanup_hook(void);350 351extern "C" ld_plugin_status onload(ld_plugin_tv *tv);352ld_plugin_status onload(ld_plugin_tv *tv) {353 InitializeAllTargetInfos();354 InitializeAllTargets();355 InitializeAllTargetMCs();356 InitializeAllAsmParsers();357 InitializeAllAsmPrinters();358 359 // We're given a pointer to the first transfer vector. We read through them360 // until we find one where tv_tag == LDPT_NULL. The REGISTER_* tagged values361 // contain pointers to functions that we need to call to register our own362 // hooks. The others are addresses of functions we can use to call into gold363 // for services.364 365 bool registeredClaimFile = false;366 bool RegisteredAllSymbolsRead = false;367 368 for (; tv->tv_tag != LDPT_NULL; ++tv) {369 // Cast tv_tag to int to allow values not in "enum ld_plugin_tag", like, for370 // example, LDPT_GET_SYMBOLS_V3 when building against an older plugin-api.h371 // header.372 switch (static_cast<int>(tv->tv_tag)) {373 case LDPT_OUTPUT_NAME:374 output_name = tv->tv_u.tv_string;375 break;376 case LDPT_LINKER_OUTPUT:377 switch (tv->tv_u.tv_val) {378 case LDPO_REL: // .o379 IsExecutable = false;380 SplitSections = false;381 break;382 case LDPO_DYN: // .so383 IsExecutable = false;384 RelocationModel = Reloc::PIC_;385 break;386 case LDPO_PIE: // position independent executable387 IsExecutable = true;388 RelocationModel = Reloc::PIC_;389 break;390 case LDPO_EXEC: // .exe391 IsExecutable = true;392 RelocationModel = Reloc::Static;393 break;394 default:395 message(LDPL_ERROR, "Unknown output file type %d", tv->tv_u.tv_val);396 return LDPS_ERR;397 }398 break;399 case LDPT_OPTION:400 options::process_plugin_option(tv->tv_u.tv_string);401 break;402 case LDPT_REGISTER_CLAIM_FILE_HOOK: {403 ld_plugin_register_claim_file callback;404 callback = tv->tv_u.tv_register_claim_file;405 406 if (callback(claim_file_hook) != LDPS_OK)407 return LDPS_ERR;408 409 registeredClaimFile = true;410 } break;411 case LDPT_REGISTER_ALL_SYMBOLS_READ_HOOK: {412 ld_plugin_register_all_symbols_read callback;413 callback = tv->tv_u.tv_register_all_symbols_read;414 415 if (callback(all_symbols_read_hook) != LDPS_OK)416 return LDPS_ERR;417 418 RegisteredAllSymbolsRead = true;419 } break;420 case LDPT_REGISTER_CLEANUP_HOOK: {421 ld_plugin_register_cleanup callback;422 callback = tv->tv_u.tv_register_cleanup;423 424 if (callback(cleanup_hook) != LDPS_OK)425 return LDPS_ERR;426 } break;427 case LDPT_GET_INPUT_FILE:428 get_input_file = tv->tv_u.tv_get_input_file;429 break;430 case LDPT_RELEASE_INPUT_FILE:431 release_input_file = tv->tv_u.tv_release_input_file;432 break;433 case LDPT_ADD_SYMBOLS:434 add_symbols = tv->tv_u.tv_add_symbols;435 break;436 case LDPT_GET_SYMBOLS_V2:437 // Do not override get_symbols_v3 with get_symbols_v2.438 if (!get_symbols)439 get_symbols = tv->tv_u.tv_get_symbols;440 break;441 case LDPT_GET_SYMBOLS_V3:442 get_symbols = tv->tv_u.tv_get_symbols;443 break;444 case LDPT_ADD_INPUT_FILE:445 add_input_file = tv->tv_u.tv_add_input_file;446 break;447 case LDPT_SET_EXTRA_LIBRARY_PATH:448 set_extra_library_path = tv->tv_u.tv_set_extra_library_path;449 break;450 case LDPT_GET_VIEW:451 get_view = tv->tv_u.tv_get_view;452 break;453 case LDPT_MESSAGE:454 message = tv->tv_u.tv_message;455 break;456 case LDPT_GET_WRAP_SYMBOLS:457 // FIXME: When binutils 2.31 (containing gold 1.16) is the minimum458 // required version, this should be changed to:459 // get_wrap_symbols = tv->tv_u.tv_get_wrap_symbols;460#pragma GCC diagnostic push461#pragma GCC diagnostic ignored "-Wcast-function-type"462 get_wrap_symbols = (ld_plugin_get_wrap_symbols)tv->tv_u.tv_message;463#pragma GCC diagnostic pop464 break;465 default:466 break;467 }468 }469 470 if (!registeredClaimFile) {471 message(LDPL_ERROR, "register_claim_file not passed to LLVMgold.");472 return LDPS_ERR;473 }474 if (!add_symbols) {475 message(LDPL_ERROR, "add_symbols not passed to LLVMgold.");476 return LDPS_ERR;477 }478 479 if (!RegisteredAllSymbolsRead)480 return LDPS_OK;481 482 if (!get_input_file) {483 message(LDPL_ERROR, "get_input_file not passed to LLVMgold.");484 return LDPS_ERR;485 }486 if (!release_input_file) {487 message(LDPL_ERROR, "release_input_file not passed to LLVMgold.");488 return LDPS_ERR;489 }490 491 return LDPS_OK;492}493 494static void diagnosticHandler(const DiagnosticInfo &DI) {495 std::string ErrStorage;496 {497 raw_string_ostream OS(ErrStorage);498 DiagnosticPrinterRawOStream DP(OS);499 DI.print(DP);500 }501 ld_plugin_level Level;502 switch (DI.getSeverity()) {503 case DS_Error:504 Level = LDPL_FATAL;505 break;506 case DS_Warning:507 Level = LDPL_WARNING;508 break;509 case DS_Note:510 case DS_Remark:511 Level = LDPL_INFO;512 break;513 }514 message(Level, "LLVM gold plugin: %s", ErrStorage.c_str());515}516 517static void check(Error E, std::string Msg = "LLVM gold plugin") {518 handleAllErrors(std::move(E), [&](ErrorInfoBase &EIB) -> Error {519 message(LDPL_FATAL, "%s: %s", Msg.c_str(), EIB.message().c_str());520 return Error::success();521 });522}523 524template <typename T> static T check(Expected<T> E) {525 if (E)526 return std::move(*E);527 check(E.takeError());528 return T();529}530 531/// Called by gold to see whether this file is one that our plugin can handle.532/// We'll try to open it and register all the symbols with add_symbol if533/// possible.534static ld_plugin_status claim_file_hook(const ld_plugin_input_file *file,535 int *claimed) {536 MemoryBufferRef BufferRef;537 std::unique_ptr<MemoryBuffer> Buffer;538 if (get_view) {539 const void *view;540 if (get_view(file->handle, &view) != LDPS_OK) {541 message(LDPL_ERROR, "Failed to get a view of %s", file->name);542 return LDPS_ERR;543 }544 BufferRef =545 MemoryBufferRef(StringRef((const char *)view, file->filesize), "");546 } else {547 int64_t offset = 0;548 // Gold has found what might be IR part-way inside of a file, such as549 // an .a archive.550 if (file->offset) {551 offset = file->offset;552 }553 ErrorOr<std::unique_ptr<MemoryBuffer>> BufferOrErr =554 MemoryBuffer::getOpenFileSlice(sys::fs::convertFDToNativeFile(file->fd),555 file->name, file->filesize, offset);556 if (std::error_code EC = BufferOrErr.getError()) {557 message(LDPL_ERROR, EC.message().c_str());558 return LDPS_ERR;559 }560 Buffer = std::move(BufferOrErr.get());561 BufferRef = Buffer->getMemBufferRef();562 }563 564 *claimed = 1;565 566 Expected<std::unique_ptr<InputFile>> ObjOrErr = InputFile::create(BufferRef);567 if (!ObjOrErr) {568 handleAllErrors(ObjOrErr.takeError(), [&](const ErrorInfoBase &EI) {569 std::error_code EC = EI.convertToErrorCode();570 if (EC == object::object_error::invalid_file_type ||571 EC == object::object_error::bitcode_section_not_found)572 *claimed = 0;573 else574 message(LDPL_FATAL,575 "LLVM gold plugin has failed to create LTO module: %s",576 EI.message().c_str());577 });578 579 return *claimed ? LDPS_ERR : LDPS_OK;580 }581 582 std::unique_ptr<InputFile> Obj = std::move(*ObjOrErr);583 584 Modules.emplace_back();585 claimed_file &cf = Modules.back();586 587 cf.handle = file->handle;588 // Keep track of the first handle for each file descriptor, since there are589 // multiple in the case of an archive. This is used later in the case of590 // ThinLTO parallel backends to ensure that each file is only opened and591 // released once.592 auto LeaderHandle =593 FDToLeaderHandle.insert(std::make_pair(file->fd, file->handle)).first;594 cf.leader_handle = LeaderHandle->second;595 // Save the filesize since for parallel ThinLTO backends we can only596 // invoke get_input_file once per archive (only for the leader handle).597 cf.filesize = file->filesize;598 // In the case of an archive library, all but the first member must have a599 // non-zero offset, which we can append to the file name to obtain a600 // unique name.601 cf.name = file->name;602 if (file->offset)603 cf.name += ".llvm." + std::to_string(file->offset) + "." +604 sys::path::filename(Obj->getSourceFileName()).str();605 606 for (auto &Sym : Obj->symbols()) {607 cf.syms.push_back(ld_plugin_symbol());608 ld_plugin_symbol &sym = cf.syms.back();609 sym.version = nullptr;610 StringRef Name = Sym.getName();611 sym.name = strdup(Name.str().c_str());612 613 ResolutionInfo &Res = ResInfo[Name];614 615 Res.CanOmitFromDynSym &= Sym.canBeOmittedFromSymbolTable();616 617 sym.visibility = LDPV_DEFAULT;618 GlobalValue::VisibilityTypes Vis = Sym.getVisibility();619 if (Vis != GlobalValue::DefaultVisibility)620 Res.DefaultVisibility = false;621 switch (Vis) {622 case GlobalValue::DefaultVisibility:623 break;624 case GlobalValue::HiddenVisibility:625 sym.visibility = LDPV_HIDDEN;626 break;627 case GlobalValue::ProtectedVisibility:628 sym.visibility = LDPV_PROTECTED;629 break;630 }631 632 if (Sym.isUndefined()) {633 sym.def = LDPK_UNDEF;634 if (Sym.isWeak())635 sym.def = LDPK_WEAKUNDEF;636 } else if (Sym.isCommon())637 sym.def = LDPK_COMMON;638 else if (Sym.isWeak())639 sym.def = LDPK_WEAKDEF;640 else641 sym.def = LDPK_DEF;642 643 sym.size = 0;644 sym.comdat_key = nullptr;645 int CI = Sym.getComdatIndex();646 if (CI != -1) {647 // Not setting comdat_key for nodeduplicate ensuress we don't deduplicate.648 std::pair<StringRef, Comdat::SelectionKind> C = Obj->getComdatTable()[CI];649 if (C.second != Comdat::NoDeduplicate)650 sym.comdat_key = strdup(C.first.str().c_str());651 }652 653 sym.resolution = LDPR_UNKNOWN;654 }655 656 if (!cf.syms.empty()) {657 if (add_symbols(cf.handle, cf.syms.size(), cf.syms.data()) != LDPS_OK) {658 message(LDPL_ERROR, "Unable to add symbols!");659 return LDPS_ERR;660 }661 }662 663 // Handle any --wrap options passed to gold, which are than passed664 // along to the plugin.665 if (get_wrap_symbols) {666 const char **wrap_symbols;667 uint64_t count = 0;668 if (get_wrap_symbols(&count, &wrap_symbols) != LDPS_OK) {669 message(LDPL_ERROR, "Unable to get wrap symbols!");670 return LDPS_ERR;671 }672 for (uint64_t i = 0; i < count; i++) {673 StringRef Name = wrap_symbols[i];674 ResolutionInfo &Res = ResInfo[Name];675 ResolutionInfo &WrapRes = ResInfo["__wrap_" + Name.str()];676 ResolutionInfo &RealRes = ResInfo["__real_" + Name.str()];677 // Tell LTO not to inline symbols that will be overwritten.678 Res.CanInline = false;679 RealRes.CanInline = false;680 // Tell LTO not to eliminate symbols that will be used after renaming.681 Res.IsUsedInRegularObj = true;682 WrapRes.IsUsedInRegularObj = true;683 }684 }685 686 return LDPS_OK;687}688 689static void freeSymName(ld_plugin_symbol &Sym) {690 free(Sym.name);691 free(Sym.comdat_key);692 Sym.name = nullptr;693 Sym.comdat_key = nullptr;694}695 696/// Helper to get a file's symbols and a view into it via gold callbacks.697static const void *getSymbolsAndView(claimed_file &F) {698 ld_plugin_status status = get_symbols(F.handle, F.syms.size(), F.syms.data());699 if (status == LDPS_NO_SYMS)700 return nullptr;701 702 if (status != LDPS_OK)703 message(LDPL_FATAL, "Failed to get symbol information");704 705 const void *View;706 if (get_view(F.handle, &View) != LDPS_OK)707 message(LDPL_FATAL, "Failed to get a view of file");708 709 return View;710}711 712/// Parse the thinlto-object-suffix-replace option into the \p OldSuffix and713/// \p NewSuffix strings, if it was specified.714static void getThinLTOOldAndNewSuffix(std::string &OldSuffix,715 std::string &NewSuffix) {716 assert(options::thinlto_object_suffix_replace.empty() ||717 options::thinlto_object_suffix_replace.find(';') != StringRef::npos);718 StringRef SuffixReplace = options::thinlto_object_suffix_replace;719 auto Split = SuffixReplace.split(';');720 OldSuffix = std::string(Split.first);721 NewSuffix = std::string(Split.second);722}723 724/// Given the original \p Path to an output file, replace any filename725/// suffix matching \p OldSuffix with \p NewSuffix.726static std::string getThinLTOObjectFileName(StringRef Path, StringRef OldSuffix,727 StringRef NewSuffix) {728 if (Path.consume_back(OldSuffix))729 return (Path + NewSuffix).str();730 return std::string(Path);731}732 733// Returns true if S is valid as a C language identifier.734static bool isValidCIdentifier(StringRef S) {735 return !S.empty() && (isAlpha(S[0]) || S[0] == '_') &&736 llvm::all_of(llvm::drop_begin(S),737 [](char C) { return C == '_' || isAlnum(C); });738}739 740static bool isUndefined(ld_plugin_symbol &Sym) {741 return Sym.def == LDPK_UNDEF || Sym.def == LDPK_WEAKUNDEF;742}743 744static void addModule(LTO &Lto, claimed_file &F, const void *View,745 StringRef Filename) {746 MemoryBufferRef BufferRef(StringRef((const char *)View, F.filesize),747 Filename);748 Expected<std::unique_ptr<InputFile>> ObjOrErr = InputFile::create(BufferRef);749 750 if (!ObjOrErr)751 message(LDPL_FATAL, "Could not read bitcode from file : %s",752 toString(ObjOrErr.takeError()).c_str());753 754 unsigned SymNum = 0;755 std::unique_ptr<InputFile> Input = std::move(ObjOrErr.get());756 auto InputFileSyms = Input->symbols();757 assert(InputFileSyms.size() == F.syms.size());758 std::vector<SymbolResolution> Resols(F.syms.size());759 for (ld_plugin_symbol &Sym : F.syms) {760 const InputFile::Symbol &InpSym = InputFileSyms[SymNum];761 SymbolResolution &R = Resols[SymNum++];762 763 ld_plugin_symbol_resolution Resolution =764 (ld_plugin_symbol_resolution)Sym.resolution;765 766 ResolutionInfo &Res = ResInfo[Sym.name];767 768 switch (Resolution) {769 case LDPR_UNKNOWN:770 llvm_unreachable("Unexpected resolution");771 772 case LDPR_RESOLVED_IR:773 case LDPR_RESOLVED_EXEC:774 case LDPR_PREEMPTED_IR:775 case LDPR_PREEMPTED_REG:776 case LDPR_UNDEF:777 break;778 779 case LDPR_RESOLVED_DYN:780 R.ExportDynamic = true;781 break;782 783 case LDPR_PREVAILING_DEF_IRONLY:784 R.Prevailing = !isUndefined(Sym);785 break;786 787 case LDPR_PREVAILING_DEF:788 R.Prevailing = !isUndefined(Sym);789 R.VisibleToRegularObj = true;790 break;791 792 case LDPR_PREVAILING_DEF_IRONLY_EXP:793 R.Prevailing = !isUndefined(Sym);794 // Identify symbols exported dynamically, and that therefore could be795 // referenced by a shared library not visible to the linker.796 R.ExportDynamic = true;797 if (!Res.CanOmitFromDynSym)798 R.VisibleToRegularObj = true;799 break;800 }801 802 // If the symbol has a C identifier section name, we need to mark803 // it as visible to a regular object so that LTO will keep it around804 // to ensure the linker generates special __start_<secname> and805 // __stop_<secname> symbols which may be used elsewhere.806 if (isValidCIdentifier(InpSym.getSectionName()))807 R.VisibleToRegularObj = true;808 809 if (Resolution != LDPR_RESOLVED_DYN && Resolution != LDPR_UNDEF &&810 (IsExecutable || !Res.DefaultVisibility))811 R.FinalDefinitionInLinkageUnit = true;812 813 if (!Res.CanInline)814 R.LinkerRedefined = true;815 816 if (Res.IsUsedInRegularObj)817 R.VisibleToRegularObj = true;818 819 freeSymName(Sym);820 }821 822 check(Lto.add(std::move(Input), Resols),823 std::string("Failed to link module ") + F.name);824}825 826static void recordFile(const std::string &Filename, bool TempOutFile) {827 if (add_input_file(Filename.c_str()) != LDPS_OK)828 message(LDPL_FATAL,829 "Unable to add .o file to the link. File left behind in: %s",830 Filename.c_str());831 if (TempOutFile)832 Cleanup.push_back(Filename);833}834 835/// Return the desired output filename given a base input name, a flag836/// indicating whether a temp file should be generated, and an optional task id.837/// The new filename generated is returned in \p NewFilename.838static int getOutputFileName(StringRef InFilename, bool TempOutFile,839 SmallString<128> &NewFilename, int TaskID) {840 int FD = -1;841 if (TempOutFile) {842 std::error_code EC =843 sys::fs::createTemporaryFile("lto-llvm", "o", FD, NewFilename);844 if (EC)845 message(LDPL_FATAL, "Could not create temporary file: %s",846 EC.message().c_str());847 } else {848 NewFilename = InFilename;849 if (TaskID > 0)850 NewFilename += utostr(TaskID);851 std::error_code EC =852 sys::fs::openFileForWrite(NewFilename, FD, sys::fs::CD_CreateAlways);853 if (EC)854 message(LDPL_FATAL, "Could not open file %s: %s", NewFilename.c_str(),855 EC.message().c_str());856 }857 return FD;858}859 860/// Parse the thinlto_prefix_replace option into the \p OldPrefix and861/// \p NewPrefix strings, if it was specified.862static void getThinLTOOldAndNewPrefix(std::string &OldPrefix,863 std::string &NewPrefix) {864 StringRef PrefixReplace = options::thinlto_prefix_replace;865 assert(PrefixReplace.empty() || PrefixReplace.find(';') != StringRef::npos);866 auto Split = PrefixReplace.split(';');867 OldPrefix = std::string(Split.first);868 NewPrefix = std::string(Split.second);869}870 871/// Creates instance of LTO.872/// OnIndexWrite is callback to let caller know when LTO writes index files.873/// LinkedObjectsFile is an output stream to write the list of object files for874/// the final ThinLTO linking. Can be nullptr.875static std::unique_ptr<LTO> createLTO(IndexWriteCallback OnIndexWrite,876 raw_fd_ostream *LinkedObjectsFile) {877 Config Conf;878 ThinBackend Backend;879 880 Conf.CPU = options::mcpu;881 Conf.Options = codegen::InitTargetOptionsFromCodeGenFlags(Triple());882 883 // Disable the new X86 relax relocations since gold might not support them.884 // FIXME: Check the gold version or add a new option to enable them.885 Conf.Options.MCOptions.X86RelaxRelocations = false;886 887 // Toggle function/data sections.888 if (!codegen::getExplicitFunctionSections())889 Conf.Options.FunctionSections = SplitSections;890 if (!codegen::getExplicitDataSections())891 Conf.Options.DataSections = SplitSections;892 893 Conf.MAttrs = codegen::getMAttrs();894 Conf.RelocModel = RelocationModel;895 Conf.CodeModel = codegen::getExplicitCodeModel();896 std::optional<CodeGenOptLevel> CGOptLevelOrNone =897 CodeGenOpt::getLevel(options::OptLevel);898 assert(CGOptLevelOrNone && "Invalid optimization level");899 Conf.CGOptLevel = *CGOptLevelOrNone;900 Conf.DisableVerify = options::DisableVerify;901 Conf.OptLevel = options::OptLevel;902 Conf.PTO.LoopVectorization = options::OptLevel > 1;903 Conf.PTO.SLPVectorization = options::OptLevel > 1;904 Conf.PTO.MergeFunctions = options::merge_functions;905 Conf.PTO.UnifiedLTO = options::unifiedlto;906 Conf.AlwaysEmitRegularLTOObj = !options::obj_path.empty();907 908 if (options::thinlto_index_only) {909 std::string OldPrefix, NewPrefix;910 getThinLTOOldAndNewPrefix(OldPrefix, NewPrefix);911 Backend = createWriteIndexesThinBackend(912 llvm::hardware_concurrency(options::Parallelism), OldPrefix, NewPrefix,913 // TODO: Add support for optional native object path in914 // thinlto_prefix_replace option to match lld.915 /*NativeObjectPrefix=*/"", options::thinlto_emit_imports_files,916 LinkedObjectsFile, OnIndexWrite);917 } else {918 Backend = createInProcessThinBackend(919 llvm::heavyweight_hardware_concurrency(options::Parallelism));920 }921 922 Conf.OverrideTriple = options::triple;923 Conf.DefaultTriple = sys::getDefaultTargetTriple();924 925 Conf.DiagHandler = diagnosticHandler;926 927 switch (options::TheOutputType) {928 case options::OT_NORMAL:929 break;930 931 case options::OT_DISABLE:932 Conf.PreOptModuleHook = [](size_t Task, const Module &M) { return false; };933 break;934 935 case options::OT_BC_ONLY:936 Conf.PostInternalizeModuleHook = [](size_t Task, const Module &M) {937 std::error_code EC;938 SmallString<128> TaskFilename;939 getOutputFileName(output_name, /* TempOutFile */ false, TaskFilename,940 Task);941 raw_fd_ostream OS(TaskFilename, EC, sys::fs::OpenFlags::OF_None);942 if (EC)943 message(LDPL_FATAL, "Failed to write the output file.");944 WriteBitcodeToFile(M, OS, /* ShouldPreserveUseListOrder */ false);945 return false;946 };947 break;948 949 case options::OT_SAVE_TEMPS:950 check(Conf.addSaveTemps(output_name + ".",951 /* UseInputModulePath */ true));952 break;953 case options::OT_ASM_ONLY:954 Conf.CGFileType = CodeGenFileType::AssemblyFile;955 Conf.Options.MCOptions.AsmVerbose = true;956 break;957 }958 959 if (!options::sample_profile.empty())960 Conf.SampleProfile = options::sample_profile;961 962 if (!options::cs_profile_path.empty())963 Conf.CSIRProfile = options::cs_profile_path;964 Conf.RunCSIRInstr = options::cs_pgo_gen;965 966 Conf.DwoDir = options::dwo_dir;967 968 // Set up optimization remarks handling.969 Conf.RemarksFilename = options::RemarksFilename;970 Conf.RemarksPasses = options::RemarksPasses;971 Conf.RemarksWithHotness = options::RemarksWithHotness;972 Conf.RemarksHotnessThreshold = options::RemarksHotnessThreshold;973 Conf.RemarksFormat = options::RemarksFormat;974 975 // Debug new pass manager if requested976 Conf.DebugPassManager = options::debug_pass_manager;977 978 Conf.HasWholeProgramVisibility = options::whole_program_visibility;979 980 Conf.StatsFile = options::stats_file;981 982 Conf.TimeTraceEnabled = !options::time_trace_file.empty();983 Conf.TimeTraceGranularity = options::time_trace_granularity;984 985 LTO::LTOKind ltoKind = LTO::LTOK_Default;986 if (options::unifiedlto)987 ltoKind =988 options::thinlto ? LTO::LTOK_UnifiedThin : LTO::LTOK_UnifiedRegular;989 return std::make_unique<LTO>(std::move(Conf), Backend,990 options::ParallelCodeGenParallelismLevel,991 ltoKind);992}993 994// Write empty files that may be expected by a distributed build995// system when invoked with thinlto_index_only. This is invoked when996// the linker has decided not to include the given module in the997// final link. Frequently the distributed build system will want to998// confirm that all expected outputs are created based on all of the999// modules provided to the linker.1000// If SkipModule is true then .thinlto.bc should contain just1001// SkipModuleByDistributedBackend flag which requests distributed backend1002// to skip the compilation of the corresponding module and produce an empty1003// object file.1004static void writeEmptyDistributedBuildOutputs(const std::string &ModulePath,1005 const std::string &OldPrefix,1006 const std::string &NewPrefix,1007 bool SkipModule) {1008 std::string NewModulePath =1009 getThinLTOOutputFile(ModulePath, OldPrefix, NewPrefix);1010 std::error_code EC;1011 {1012 raw_fd_ostream OS(NewModulePath + ".thinlto.bc", EC,1013 sys::fs::OpenFlags::OF_None);1014 if (EC)1015 message(LDPL_FATAL, "Failed to write '%s': %s",1016 (NewModulePath + ".thinlto.bc").c_str(), EC.message().c_str());1017 1018 if (SkipModule) {1019 ModuleSummaryIndex Index(/*HaveGVs*/ false);1020 Index.setSkipModuleByDistributedBackend();1021 writeIndexToFile(Index, OS, nullptr);1022 }1023 }1024 if (options::thinlto_emit_imports_files) {1025 raw_fd_ostream OS(NewModulePath + ".imports", EC,1026 sys::fs::OpenFlags::OF_None);1027 if (EC)1028 message(LDPL_FATAL, "Failed to write '%s': %s",1029 (NewModulePath + ".imports").c_str(), EC.message().c_str());1030 }1031}1032 1033// Creates and returns output stream with a list of object files for final1034// linking of distributed ThinLTO.1035static std::unique_ptr<raw_fd_ostream> CreateLinkedObjectsFile() {1036 if (options::thinlto_linked_objects_file.empty())1037 return nullptr;1038 assert(options::thinlto_index_only);1039 std::error_code EC;1040 auto LinkedObjectsFile = std::make_unique<raw_fd_ostream>(1041 options::thinlto_linked_objects_file, EC, sys::fs::OpenFlags::OF_None);1042 if (EC)1043 message(LDPL_FATAL, "Failed to create '%s': %s",1044 options::thinlto_linked_objects_file.c_str(), EC.message().c_str());1045 return LinkedObjectsFile;1046}1047 1048/// Runs LTO and return a list of pairs <FileName, IsTemporary>.1049static std::vector<std::pair<SmallString<128>, bool>> runLTO() {1050 // Map to own RAII objects that manage the file opening and releasing1051 // interfaces with gold. This is needed only for ThinLTO mode, since1052 // unlike regular LTO, where addModule will result in the opened file1053 // being merged into a new combined module, we need to keep these files open1054 // through Lto->run().1055 DenseMap<void *, std::unique_ptr<PluginInputFile>> HandleToInputFile;1056 1057 // Owns string objects and tells if index file was already created.1058 StringMap<bool> ObjectToIndexFileState;1059 1060 std::unique_ptr<raw_fd_ostream> LinkedObjects = CreateLinkedObjectsFile();1061 std::unique_ptr<LTO> Lto = createLTO(1062 [&ObjectToIndexFileState](const std::string &Identifier) {1063 ObjectToIndexFileState[Identifier] = true;1064 },1065 LinkedObjects.get());1066 1067 std::string OldPrefix, NewPrefix;1068 if (options::thinlto_index_only)1069 getThinLTOOldAndNewPrefix(OldPrefix, NewPrefix);1070 1071 std::string OldSuffix, NewSuffix;1072 getThinLTOOldAndNewSuffix(OldSuffix, NewSuffix);1073 1074 for (claimed_file &F : Modules) {1075 if (options::thinlto) {1076 auto [It, Inserted] = HandleToInputFile.try_emplace(F.leader_handle);1077 if (Inserted)1078 It->second = std::make_unique<PluginInputFile>(F.handle);1079 }1080 // In case we are thin linking with a minimized bitcode file, ensure1081 // the module paths encoded in the index reflect where the backends1082 // will locate the full bitcode files for compiling/importing.1083 std::string Identifier =1084 getThinLTOObjectFileName(F.name, OldSuffix, NewSuffix);1085 auto ObjFilename = ObjectToIndexFileState.insert({Identifier, false});1086 assert(ObjFilename.second);1087 if (const void *View = getSymbolsAndView(F))1088 addModule(*Lto, F, View, ObjFilename.first->first());1089 else if (options::thinlto_index_only) {1090 ObjFilename.first->second = true;1091 writeEmptyDistributedBuildOutputs(Identifier, OldPrefix, NewPrefix,1092 /* SkipModule */ true);1093 }1094 }1095 1096 SmallString<128> Filename;1097 // Note that getOutputFileName will append a unique ID for each task1098 if (!options::obj_path.empty())1099 Filename = options::obj_path;1100 else if (options::TheOutputType == options::OT_SAVE_TEMPS)1101 Filename = output_name + ".lto.o";1102 else if (options::TheOutputType == options::OT_ASM_ONLY)1103 Filename = output_name;1104 bool SaveTemps = !Filename.empty();1105 1106 size_t MaxTasks = Lto->getMaxTasks();1107 std::vector<std::pair<SmallString<128>, bool>> Files(MaxTasks);1108 1109 auto AddStream =1110 [&](size_t Task,1111 const Twine &ModuleName) -> std::unique_ptr<CachedFileStream> {1112 Files[Task].second = !SaveTemps;1113 int FD = getOutputFileName(Filename, /* TempOutFile */ !SaveTemps,1114 Files[Task].first, Task);1115 return std::make_unique<CachedFileStream>(1116 std::make_unique<llvm::raw_fd_ostream>(FD, true));1117 };1118 1119 auto AddBuffer = [&](size_t Task, const Twine &ModuleName,1120 std::unique_ptr<MemoryBuffer> MB) {1121 auto Stream = AddStream(Task, ModuleName);1122 *Stream->OS << MB->getBuffer();1123 check(Stream->commit(), "Failed to commit cache");1124 };1125 1126 FileCache Cache;1127 if (!options::cache_dir.empty())1128 Cache = check(localCache("ThinLTO", "Thin", options::cache_dir, AddBuffer));1129 1130 check(Lto->run(AddStream, Cache));1131 1132 // Write empty output files that may be expected by the distributed build1133 // system.1134 if (options::thinlto_index_only)1135 for (auto &Identifier : ObjectToIndexFileState)1136 if (!Identifier.getValue())1137 writeEmptyDistributedBuildOutputs(std::string(Identifier.getKey()),1138 OldPrefix, NewPrefix,1139 /* SkipModule */ false);1140 1141 return Files;1142}1143 1144/// gold informs us that all symbols have been read. At this point, we use1145/// get_symbols to see if any of our definitions have been overridden by a1146/// native object file. Then, perform optimization and codegen.1147static ld_plugin_status allSymbolsReadHook() {1148 if (Modules.empty())1149 return LDPS_OK;1150 1151 if (unsigned NumOpts = options::extra.size())1152 cl::ParseCommandLineOptions(NumOpts, &options::extra[0]);1153 1154 // Initialize time trace profiler1155 if (!options::time_trace_file.empty())1156 llvm::timeTraceProfilerInitialize(options::time_trace_granularity,1157 options::extra.size() ? options::extra[0]1158 : "LLVMgold");1159 auto FinalizeTimeTrace = llvm::make_scope_exit([&]() {1160 if (!llvm::timeTraceProfilerEnabled())1161 return;1162 assert(!options::time_trace_file.empty());1163 check(llvm::timeTraceProfilerWrite(options::time_trace_file, output_name));1164 llvm::timeTraceProfilerCleanup();1165 });1166 1167 std::vector<std::pair<SmallString<128>, bool>> Files = runLTO();1168 1169 if (options::TheOutputType == options::OT_DISABLE ||1170 options::TheOutputType == options::OT_BC_ONLY ||1171 options::TheOutputType == options::OT_ASM_ONLY)1172 return LDPS_OK;1173 1174 if (options::thinlto_index_only) {1175 llvm_shutdown();1176 cleanup_hook();1177 exit(0);1178 }1179 1180 for (const auto &F : Files)1181 if (!F.first.empty())1182 recordFile(std::string(F.first.str()), F.second);1183 1184 if (!options::extra_library_path.empty() &&1185 set_extra_library_path(options::extra_library_path.c_str()) != LDPS_OK)1186 message(LDPL_FATAL, "Unable to set the extra library path.");1187 1188 return LDPS_OK;1189}1190 1191static ld_plugin_status all_symbols_read_hook(void) {1192 ld_plugin_status Ret = allSymbolsReadHook();1193 llvm_shutdown();1194 1195 if (options::TheOutputType == options::OT_BC_ONLY ||1196 options::TheOutputType == options::OT_ASM_ONLY ||1197 options::TheOutputType == options::OT_DISABLE) {1198 if (options::TheOutputType == options::OT_DISABLE) {1199 // Remove the output file here since ld.bfd creates the output file1200 // early.1201 std::error_code EC = sys::fs::remove(output_name);1202 if (EC)1203 message(LDPL_ERROR, "Failed to delete '%s': %s", output_name.c_str(),1204 EC.message().c_str());1205 }1206 exit(0);1207 }1208 1209 return Ret;1210}1211 1212static ld_plugin_status cleanup_hook(void) {1213 for (std::string &Name : Cleanup) {1214 std::error_code EC = sys::fs::remove(Name);1215 if (EC)1216 message(LDPL_ERROR, "Failed to delete '%s': %s", Name.c_str(),1217 EC.message().c_str());1218 }1219 1220 // Prune cache1221 if (!options::cache_dir.empty()) {1222 CachePruningPolicy policy = check(parseCachePruningPolicy(options::cache_policy));1223 pruneCache(options::cache_dir, policy);1224 }1225 1226 return LDPS_OK;1227}1228