1f7c5fbb1SRui Ueyama //===- LinkerScript.cpp ---------------------------------------------------===// 2f7c5fbb1SRui Ueyama // 3f7c5fbb1SRui Ueyama // The LLVM Linker 4f7c5fbb1SRui Ueyama // 5f7c5fbb1SRui Ueyama // This file is distributed under the University of Illinois Open Source 6f7c5fbb1SRui Ueyama // License. See LICENSE.TXT for details. 7f7c5fbb1SRui Ueyama // 8f7c5fbb1SRui Ueyama //===----------------------------------------------------------------------===// 9f7c5fbb1SRui Ueyama // 10f7c5fbb1SRui Ueyama // This file contains the parser/evaluator of the linker script. 11f7c5fbb1SRui Ueyama // 12f7c5fbb1SRui Ueyama //===----------------------------------------------------------------------===// 13f7c5fbb1SRui Ueyama 14717677afSRui Ueyama #include "LinkerScript.h" 15f7c5fbb1SRui Ueyama #include "Config.h" 16f7c5fbb1SRui Ueyama #include "Driver.h" 171ebc8ed7SRui Ueyama #include "InputSection.h" 189381eb10SRui Ueyama #include "Memory.h" 19652852c5SGeorge Rimar #include "OutputSections.h" 20e77b5bf6SAdhemerval Zanella #include "ScriptParser.h" 2193c9af42SRui Ueyama #include "Strings.h" 22f7c5fbb1SRui Ueyama #include "SymbolTable.h" 2355518e7dSRui Ueyama #include "Symbols.h" 243fb5a6dcSGeorge Rimar #include "SyntheticSections.h" 25467c4d55SEugene Leviant #include "Target.h" 26bbe38602SEugene Leviant #include "Writer.h" 2722886a28SEugene Zelenko #include "llvm/ADT/STLExtras.h" 288c6a5aafSRui Ueyama #include "llvm/ADT/SmallString.h" 2922886a28SEugene Zelenko #include "llvm/ADT/StringRef.h" 30960504b9SRui Ueyama #include "llvm/ADT/StringSwitch.h" 3122886a28SEugene Zelenko #include "llvm/Support/Casting.h" 32652852c5SGeorge Rimar #include "llvm/Support/ELF.h" 3322886a28SEugene Zelenko #include "llvm/Support/Endian.h" 3422886a28SEugene Zelenko #include "llvm/Support/ErrorHandling.h" 35f7c5fbb1SRui Ueyama #include "llvm/Support/FileSystem.h" 3622886a28SEugene Zelenko #include "llvm/Support/MathExtras.h" 37f03f3cc1SRui Ueyama #include "llvm/Support/Path.h" 3822886a28SEugene Zelenko #include <algorithm> 3922886a28SEugene Zelenko #include <cassert> 4022886a28SEugene Zelenko #include <cstddef> 4122886a28SEugene Zelenko #include <cstdint> 4222886a28SEugene Zelenko #include <iterator> 4322886a28SEugene Zelenko #include <limits> 4422886a28SEugene Zelenko #include <memory> 4522886a28SEugene Zelenko #include <string> 4622886a28SEugene Zelenko #include <tuple> 4722886a28SEugene Zelenko #include <vector> 48f7c5fbb1SRui Ueyama 49f7c5fbb1SRui Ueyama using namespace llvm; 50652852c5SGeorge Rimar using namespace llvm::ELF; 511ebc8ed7SRui Ueyama using namespace llvm::object; 52e38cbab5SGeorge Rimar using namespace llvm::support::endian; 53f7c5fbb1SRui Ueyama using namespace lld; 54e0df00b9SRafael Espindola using namespace lld::elf; 55f7c5fbb1SRui Ueyama 56884e786dSGeorge Rimar LinkerScriptBase *elf::ScriptBase; 5707320e40SRui Ueyama ScriptConfiguration *elf::ScriptConfig; 58717677afSRui Ueyama 598f1f3c40SMeador Inge template <class ELFT> static SymbolBody *addRegular(SymbolAssignment *Cmd) { 603dabfc6bSRafael Espindola uint8_t Visibility = Cmd->Hidden ? STV_HIDDEN : STV_DEFAULT; 618f1f3c40SMeador Inge Symbol *Sym = Symtab<ELFT>::X->addUndefined( 628f1f3c40SMeador Inge Cmd->Name, /*IsLocal=*/false, STB_GLOBAL, Visibility, 638f1f3c40SMeador Inge /*Type*/ 0, 648f1f3c40SMeador Inge /*CanOmitFromDynSym*/ false, /*File*/ nullptr); 6520d03194SEugene Leviant 668f1f3c40SMeador Inge replaceBody<DefinedRegular<ELFT>>(Sym, Cmd->Name, /*IsLocal=*/false, 678f1f3c40SMeador Inge Visibility, STT_NOTYPE, 0, 0, nullptr, 688f1f3c40SMeador Inge nullptr); 698f1f3c40SMeador Inge return Sym->body(); 70ceabe80eSEugene Leviant } 71ceabe80eSEugene Leviant 728f1f3c40SMeador Inge template <class ELFT> static SymbolBody *addSynthetic(SymbolAssignment *Cmd) { 738f1f3c40SMeador Inge uint8_t Visibility = Cmd->Hidden ? STV_HIDDEN : STV_DEFAULT; 74afaa9343SEugene Leviant const OutputSectionBase *Sec = 75afaa9343SEugene Leviant ScriptConfig->HasSections ? nullptr : Cmd->Expression.Section(); 768f1f3c40SMeador Inge Symbol *Sym = Symtab<ELFT>::X->addUndefined( 778f1f3c40SMeador Inge Cmd->Name, /*IsLocal=*/false, STB_GLOBAL, Visibility, 788f1f3c40SMeador Inge /*Type*/ 0, 798f1f3c40SMeador Inge /*CanOmitFromDynSym*/ false, /*File*/ nullptr); 80afaa9343SEugene Leviant 818f1f3c40SMeador Inge replaceBody<DefinedSynthetic>(Sym, Cmd->Name, 0, Sec); 828f1f3c40SMeador Inge return Sym->body(); 83ceabe80eSEugene Leviant } 84ceabe80eSEugene Leviant 8522375f24SRui Ueyama static bool isUnderSysroot(StringRef Path) { 8622375f24SRui Ueyama if (Config->Sysroot == "") 8722375f24SRui Ueyama return false; 8822375f24SRui Ueyama for (; !Path.empty(); Path = sys::path::parent_path(Path)) 8922375f24SRui Ueyama if (sys::fs::equivalent(Config->Sysroot, Path)) 9022375f24SRui Ueyama return true; 9122375f24SRui Ueyama return false; 9222375f24SRui Ueyama } 9322375f24SRui Ueyama 94b2b70975SGeorge Rimar // Sets value of a symbol. Two kinds of symbols are processed: synthetic 95b2b70975SGeorge Rimar // symbols, whose value is an offset from beginning of section and regular 96b2b70975SGeorge Rimar // symbols whose value is absolute. 97b2b70975SGeorge Rimar template <class ELFT> 98b2b70975SGeorge Rimar static void assignSymbol(SymbolAssignment *Cmd, typename ELFT::uint Dot = 0) { 99b2b70975SGeorge Rimar if (!Cmd->Sym) 1008f1f3c40SMeador Inge return; 1018f1f3c40SMeador Inge 102b2b70975SGeorge Rimar if (auto *Body = dyn_cast<DefinedSynthetic>(Cmd->Sym)) { 103b2b70975SGeorge Rimar Body->Section = Cmd->Expression.Section(); 104193b158bSRafael Espindola if (Body->Section) 105193b158bSRafael Espindola Body->Value = Cmd->Expression(Dot) - Body->Section->Addr; 106b2b70975SGeorge Rimar return; 107db741e72SEugene Leviant } 108b2b70975SGeorge Rimar 109b2b70975SGeorge Rimar cast<DefinedRegular<ELFT>>(Cmd->Sym)->Value = Cmd->Expression(Dot); 1108f1f3c40SMeador Inge } 1118f1f3c40SMeador Inge 1128f1f3c40SMeador Inge template <class ELFT> static void addSymbol(SymbolAssignment *Cmd) { 1131602421cSRui Ueyama if (Cmd->Name == ".") 1148f1f3c40SMeador Inge return; 1158f1f3c40SMeador Inge 1168f1f3c40SMeador Inge // If a symbol was in PROVIDE(), we need to define it only when 1178f1f3c40SMeador Inge // it is a referenced undefined symbol. 1181602421cSRui Ueyama SymbolBody *B = Symtab<ELFT>::X->find(Cmd->Name); 1198f1f3c40SMeador Inge if (Cmd->Provide && (!B || B->isDefined())) 1208f1f3c40SMeador Inge return; 1218f1f3c40SMeador Inge 1228f1f3c40SMeador Inge // Otherwise, create a new symbol if one does not exist or an 1238f1f3c40SMeador Inge // undefined one does exist. 1248f1f3c40SMeador Inge if (Cmd->Expression.IsAbsolute()) 1258f1f3c40SMeador Inge Cmd->Sym = addRegular<ELFT>(Cmd); 1268f1f3c40SMeador Inge else 1278f1f3c40SMeador Inge Cmd->Sym = addSynthetic<ELFT>(Cmd); 128b2b70975SGeorge Rimar 129b2b70975SGeorge Rimar // If there are sections, then let the value be assigned later in 130b2b70975SGeorge Rimar // `assignAddresses`. 131b2b70975SGeorge Rimar if (!ScriptConfig->HasSections) 1328f1f3c40SMeador Inge assignSymbol<ELFT>(Cmd); 133ceabe80eSEugene Leviant } 134ceabe80eSEugene Leviant 135076fe157SGeorge Rimar bool SymbolAssignment::classof(const BaseCommand *C) { 136076fe157SGeorge Rimar return C->Kind == AssignmentKind; 137076fe157SGeorge Rimar } 138076fe157SGeorge Rimar 139076fe157SGeorge Rimar bool OutputSectionCommand::classof(const BaseCommand *C) { 140076fe157SGeorge Rimar return C->Kind == OutputSectionKind; 141076fe157SGeorge Rimar } 142076fe157SGeorge Rimar 143eea3114fSGeorge Rimar bool InputSectionDescription::classof(const BaseCommand *C) { 144eea3114fSGeorge Rimar return C->Kind == InputSectionKind; 145eea3114fSGeorge Rimar } 146eea3114fSGeorge Rimar 147eefa758eSGeorge Rimar bool AssertCommand::classof(const BaseCommand *C) { 148eefa758eSGeorge Rimar return C->Kind == AssertKind; 149eefa758eSGeorge Rimar } 150eefa758eSGeorge Rimar 151e38cbab5SGeorge Rimar bool BytesDataCommand::classof(const BaseCommand *C) { 152e38cbab5SGeorge Rimar return C->Kind == BytesDataKind; 153e38cbab5SGeorge Rimar } 154e38cbab5SGeorge Rimar 15522886a28SEugene Zelenko template <class ELFT> LinkerScript<ELFT>::LinkerScript() = default; 15622886a28SEugene Zelenko template <class ELFT> LinkerScript<ELFT>::~LinkerScript() = default; 157f34d0e08SRui Ueyama 158e0be2901SRui Ueyama template <class ELFT> static StringRef basename(InputSectionBase<ELFT> *S) { 159e0be2901SRui Ueyama if (S->getFile()) 160e0be2901SRui Ueyama return sys::path::filename(S->getFile()->getName()); 161e0be2901SRui Ueyama return ""; 162e0be2901SRui Ueyama } 163e0be2901SRui Ueyama 16407320e40SRui Ueyama template <class ELFT> 16507320e40SRui Ueyama bool LinkerScript<ELFT>::shouldKeep(InputSectionBase<ELFT> *S) { 166e0be2901SRui Ueyama for (InputSectionDescription *ID : Opt.KeptSections) 167e0be2901SRui Ueyama if (ID->FilePat.match(basename(S))) 168cf43f179SEugene Leviant for (SectionPattern &P : ID->SectionPatterns) 169f91282e1SRui Ueyama if (P.SectionPat.match(S->Name)) 170eea3114fSGeorge Rimar return true; 171eea3114fSGeorge Rimar return false; 172eea3114fSGeorge Rimar } 173eea3114fSGeorge Rimar 174575208caSGeorge Rimar static bool comparePriority(InputSectionData *A, InputSectionData *B) { 175575208caSGeorge Rimar return getPriority(A->Name) < getPriority(B->Name); 176575208caSGeorge Rimar } 177575208caSGeorge Rimar 178c0028d3dSRafael Espindola static bool compareName(InputSectionData *A, InputSectionData *B) { 179042a3f20SRafael Espindola return A->Name < B->Name; 1800702c4e8SGeorge Rimar } 181742c3836SRui Ueyama 182c0028d3dSRafael Espindola static bool compareAlignment(InputSectionData *A, InputSectionData *B) { 183742c3836SRui Ueyama // ">" is not a mistake. Larger alignments are placed before smaller 184742c3836SRui Ueyama // alignments in order to reduce the amount of padding necessary. 185742c3836SRui Ueyama // This is compatible with GNU. 186742c3836SRui Ueyama return A->Alignment > B->Alignment; 187742c3836SRui Ueyama } 188742c3836SRui Ueyama 189c0028d3dSRafael Espindola static std::function<bool(InputSectionData *, InputSectionData *)> 190be394db3SGeorge Rimar getComparator(SortSectionPolicy K) { 191be394db3SGeorge Rimar switch (K) { 192be394db3SGeorge Rimar case SortSectionPolicy::Alignment: 193c0028d3dSRafael Espindola return compareAlignment; 194be394db3SGeorge Rimar case SortSectionPolicy::Name: 195be394db3SGeorge Rimar return compareName; 196be394db3SGeorge Rimar case SortSectionPolicy::Priority: 197be394db3SGeorge Rimar return comparePriority; 198be394db3SGeorge Rimar default: 199be394db3SGeorge Rimar llvm_unreachable("unknown sort policy"); 200be394db3SGeorge Rimar } 201742c3836SRui Ueyama } 2020702c4e8SGeorge Rimar 20348c3f1ceSRui Ueyama template <class ELFT> 204e71a3f8aSRafael Espindola static bool matchConstraints(ArrayRef<InputSectionBase<ELFT> *> Sections, 20506ae6836SGeorge Rimar ConstraintKind Kind) { 2068f66df92SGeorge Rimar if (Kind == ConstraintKind::NoConstraint) 2078f66df92SGeorge Rimar return true; 208e746e52cSRafael Espindola bool IsRW = llvm::any_of(Sections, [=](InputSectionData *Sec2) { 209d3190795SRafael Espindola auto *Sec = static_cast<InputSectionBase<ELFT> *>(Sec2); 2101854a8ebSRafael Espindola return Sec->Flags & SHF_WRITE; 21106ae6836SGeorge Rimar }); 212e746e52cSRafael Espindola return (IsRW && Kind == ConstraintKind::ReadWrite) || 213e746e52cSRafael Espindola (!IsRW && Kind == ConstraintKind::ReadOnly); 21406ae6836SGeorge Rimar } 21506ae6836SGeorge Rimar 21607171f21SGeorge Rimar static void sortSections(InputSectionData **Begin, InputSectionData **End, 217ee924709SRui Ueyama SortSectionPolicy K) { 218ee924709SRui Ueyama if (K != SortSectionPolicy::Default && K != SortSectionPolicy::None) 21907171f21SGeorge Rimar std::stable_sort(Begin, End, getComparator(K)); 220ee924709SRui Ueyama } 221ee924709SRui Ueyama 222d3190795SRafael Espindola // Compute and remember which sections the InputSectionDescription matches. 223be94e1b6SRafael Espindola template <class ELFT> 224e71a3f8aSRafael Espindola void LinkerScript<ELFT>::computeInputSections(InputSectionDescription *I) { 2254dc07becSRui Ueyama // Collects all sections that satisfy constraints of I 2264dc07becSRui Ueyama // and attach them to I. 2274dc07becSRui Ueyama for (SectionPattern &Pat : I->SectionPatterns) { 22807171f21SGeorge Rimar size_t SizeBefore = I->Sections.size(); 2298c6a5aafSRui Ueyama 2308c6a5aafSRui Ueyama for (InputSectionBase<ELFT> *S : Symtab<ELFT>::X->Sections) { 231f94efdddSRui Ueyama if (!S->Live || S->Assigned) 2328c6a5aafSRui Ueyama continue; 2338c6a5aafSRui Ueyama 234e0be2901SRui Ueyama StringRef Filename = basename(S); 235e0be2901SRui Ueyama if (!I->FilePat.match(Filename) || Pat.ExcludedFilePat.match(Filename)) 236e0be2901SRui Ueyama continue; 237e0be2901SRui Ueyama if (!Pat.SectionPat.match(S->Name)) 238e0be2901SRui Ueyama continue; 239d3190795SRafael Espindola I->Sections.push_back(S); 240f94efdddSRui Ueyama S->Assigned = true; 241f94efdddSRui Ueyama } 242d3190795SRafael Espindola 243ee924709SRui Ueyama // Sort sections as instructed by SORT-family commands and --sort-section 244ee924709SRui Ueyama // option. Because SORT-family commands can be nested at most two depth 245ee924709SRui Ueyama // (e.g. SORT_BY_NAME(SORT_BY_ALIGNMENT(.text.*))) and because the command 246ee924709SRui Ueyama // line option is respected even if a SORT command is given, the exact 247ee924709SRui Ueyama // behavior we have here is a bit complicated. Here are the rules. 248ee924709SRui Ueyama // 249ee924709SRui Ueyama // 1. If two SORT commands are given, --sort-section is ignored. 250ee924709SRui Ueyama // 2. If one SORT command is given, and if it is not SORT_NONE, 251ee924709SRui Ueyama // --sort-section is handled as an inner SORT command. 252ee924709SRui Ueyama // 3. If one SORT command is given, and if it is SORT_NONE, don't sort. 253ee924709SRui Ueyama // 4. If no SORT command is given, sort according to --sort-section. 25407171f21SGeorge Rimar InputSectionData **Begin = I->Sections.data() + SizeBefore; 25507171f21SGeorge Rimar InputSectionData **End = I->Sections.data() + I->Sections.size(); 25607171f21SGeorge Rimar if (Pat.SortOuter != SortSectionPolicy::None) { 25707171f21SGeorge Rimar if (Pat.SortInner == SortSectionPolicy::Default) 25807171f21SGeorge Rimar sortSections(Begin, End, Config->SortSection); 259ee924709SRui Ueyama else 26007171f21SGeorge Rimar sortSections(Begin, End, Pat.SortInner); 26107171f21SGeorge Rimar sortSections(Begin, End, Pat.SortOuter); 26207171f21SGeorge Rimar } 263ee924709SRui Ueyama } 264be94e1b6SRafael Espindola } 265be94e1b6SRafael Espindola 266be94e1b6SRafael Espindola template <class ELFT> 267be94e1b6SRafael Espindola void LinkerScript<ELFT>::discard(ArrayRef<InputSectionBase<ELFT> *> V) { 268be94e1b6SRafael Espindola for (InputSectionBase<ELFT> *S : V) { 269be94e1b6SRafael Espindola S->Live = false; 270be94e1b6SRafael Espindola reportDiscarded(S); 271be94e1b6SRafael Espindola } 272be94e1b6SRafael Espindola } 273be94e1b6SRafael Espindola 27406ae6836SGeorge Rimar template <class ELFT> 2750b9ce6a4SRui Ueyama std::vector<InputSectionBase<ELFT> *> 27606ae6836SGeorge Rimar LinkerScript<ELFT>::createInputSectionList(OutputSectionCommand &OutCmd) { 2770b9ce6a4SRui Ueyama std::vector<InputSectionBase<ELFT> *> Ret; 278e7f912cdSRui Ueyama 27906ae6836SGeorge Rimar for (const std::unique_ptr<BaseCommand> &Base : OutCmd.Commands) { 2807c3ff2ebSRafael Espindola auto *Cmd = dyn_cast<InputSectionDescription>(Base.get()); 2817c3ff2ebSRafael Espindola if (!Cmd) 2820b9ce6a4SRui Ueyama continue; 283e71a3f8aSRafael Espindola computeInputSections(Cmd); 284d3190795SRafael Espindola for (InputSectionData *S : Cmd->Sections) 285d3190795SRafael Espindola Ret.push_back(static_cast<InputSectionBase<ELFT> *>(S)); 2860b9ce6a4SRui Ueyama } 287e71a3f8aSRafael Espindola 2880b9ce6a4SRui Ueyama return Ret; 2890b9ce6a4SRui Ueyama } 2900b9ce6a4SRui Ueyama 291e5d3ca50SPetr Hosek template <class ELFT> 29220d03194SEugene Leviant void LinkerScript<ELFT>::addSection(OutputSectionFactory<ELFT> &Factory, 29320d03194SEugene Leviant InputSectionBase<ELFT> *Sec, 29420d03194SEugene Leviant StringRef Name) { 295e08e78dfSRafael Espindola OutputSectionBase *OutSec; 29628c1597aSRafael Espindola bool IsNew; 297fe12450eSRafael Espindola std::tie(OutSec, IsNew) = Factory.create(Sec, Name); 29828c1597aSRafael Espindola if (IsNew) 29928c1597aSRafael Espindola OutputSections->push_back(OutSec); 30020d03194SEugene Leviant OutSec->addSection(Sec); 30120d03194SEugene Leviant } 30220d03194SEugene Leviant 30320d03194SEugene Leviant template <class ELFT> 30420d03194SEugene Leviant void LinkerScript<ELFT>::processCommands(OutputSectionFactory<ELFT> &Factory) { 3057c3ff2ebSRafael Espindola for (unsigned I = 0; I < Opt.Commands.size(); ++I) { 3067c3ff2ebSRafael Espindola auto Iter = Opt.Commands.begin() + I; 3077c3ff2ebSRafael Espindola const std::unique_ptr<BaseCommand> &Base1 = *Iter; 3080b1b695aSRui Ueyama 3090b1b695aSRui Ueyama // Handle symbol assignments outside of any output section. 3102ab5f73dSRui Ueyama if (auto *Cmd = dyn_cast<SymbolAssignment>(Base1.get())) { 311b0de56b5SRafael Espindola addSymbol<ELFT>(Cmd); 3122ab5f73dSRui Ueyama continue; 3132ab5f73dSRui Ueyama } 3140b1b695aSRui Ueyama 31520d03194SEugene Leviant if (auto *Cmd = dyn_cast<AssertCommand>(Base1.get())) { 31620d03194SEugene Leviant // If we don't have SECTIONS then output sections have already been 317194470cdSGeorge Rimar // created by Writer<ELFT>. The LinkerScript<ELFT>::assignAddresses 31820d03194SEugene Leviant // will not be called, so ASSERT should be evaluated now. 31920d03194SEugene Leviant if (!Opt.HasSections) 32020d03194SEugene Leviant Cmd->Expression(0); 32120d03194SEugene Leviant continue; 32220d03194SEugene Leviant } 3232ab5f73dSRui Ueyama 324ceabe80eSEugene Leviant if (auto *Cmd = dyn_cast<OutputSectionCommand>(Base1.get())) { 3257bd37870SRafael Espindola std::vector<InputSectionBase<ELFT> *> V = createInputSectionList(*Cmd); 3267bd37870SRafael Espindola 3270b1b695aSRui Ueyama // The output section name `/DISCARD/' is special. 3280b1b695aSRui Ueyama // Any input section assigned to it is discarded. 32948c3f1ceSRui Ueyama if (Cmd->Name == "/DISCARD/") { 3307bd37870SRafael Espindola discard(V); 33148c3f1ceSRui Ueyama continue; 33248c3f1ceSRui Ueyama } 3330b9ce6a4SRui Ueyama 3340b1b695aSRui Ueyama // This is for ONLY_IF_RO and ONLY_IF_RW. An output section directive 3350b1b695aSRui Ueyama // ".foo : ONLY_IF_R[OW] { ... }" is handled only if all member input 3360b1b695aSRui Ueyama // sections satisfy a given constraint. If not, a directive is handled 3370b1b695aSRui Ueyama // as if it wasn't present from the beginning. 3380b1b695aSRui Ueyama // 3390b1b695aSRui Ueyama // Because we'll iterate over Commands many more times, the easiest 3400b1b695aSRui Ueyama // way to "make it as if it wasn't present" is to just remove it. 3417c3ff2ebSRafael Espindola if (!matchConstraints<ELFT>(V, Cmd->Constraint)) { 3427c3ff2ebSRafael Espindola for (InputSectionBase<ELFT> *S : V) 343f94efdddSRui Ueyama S->Assigned = false; 3447c3ff2ebSRafael Espindola Opt.Commands.erase(Iter); 345dfbbbc86SGeorge Rimar --I; 3467c3ff2ebSRafael Espindola continue; 3477c3ff2ebSRafael Espindola } 3487c3ff2ebSRafael Espindola 3490b1b695aSRui Ueyama // A directive may contain symbol definitions like this: 3500b1b695aSRui Ueyama // ".foo : { ...; bar = .; }". Handle them. 3517c3ff2ebSRafael Espindola for (const std::unique_ptr<BaseCommand> &Base : Cmd->Commands) 3527c3ff2ebSRafael Espindola if (auto *OutCmd = dyn_cast<SymbolAssignment>(Base.get())) 3537c3ff2ebSRafael Espindola addSymbol<ELFT>(OutCmd); 3547c3ff2ebSRafael Espindola 3550b1b695aSRui Ueyama // Handle subalign (e.g. ".foo : SUBALIGN(32) { ... }"). If subalign 3560b1b695aSRui Ueyama // is given, input sections are aligned to that value, whether the 3570b1b695aSRui Ueyama // given value is larger or smaller than the original section alignment. 3580b1b695aSRui Ueyama if (Cmd->SubalignExpr) { 3590b1b695aSRui Ueyama uint32_t Subalign = Cmd->SubalignExpr(0); 3600b1b695aSRui Ueyama for (InputSectionBase<ELFT> *S : V) 3610b1b695aSRui Ueyama S->Alignment = Subalign; 36220d03194SEugene Leviant } 3630b1b695aSRui Ueyama 3640b1b695aSRui Ueyama // Add input sections to an output section. 3650b1b695aSRui Ueyama for (InputSectionBase<ELFT> *S : V) 3660b1b695aSRui Ueyama addSection(Factory, S, Cmd->Name); 367eea3114fSGeorge Rimar } 36848c3f1ceSRui Ueyama } 369db24d9c3SGeorge Rimar } 370e63d81bdSEugene Leviant 3710b1b695aSRui Ueyama // Add sections that didn't match any sections command. 37220d03194SEugene Leviant template <class ELFT> 37393c64025SGeorge Rimar void LinkerScript<ELFT>::addOrphanSections( 37493c64025SGeorge Rimar OutputSectionFactory<ELFT> &Factory) { 3758c6a5aafSRui Ueyama for (InputSectionBase<ELFT> *S : Symtab<ELFT>::X->Sections) 3768f9026baSRafael Espindola if (S->Live && !S->OutSec) 37755518e7dSRui Ueyama addSection(Factory, S, getOutputSectionName(S->Name)); 378e63d81bdSEugene Leviant } 379e63d81bdSEugene Leviant 380e08e78dfSRafael Espindola template <class ELFT> static bool isTbss(OutputSectionBase *Sec) { 38104a2e348SRafael Espindola return (Sec->Flags & SHF_TLS) && Sec->Type == SHT_NOBITS; 382a940e539SRafael Espindola } 383a940e539SRafael Espindola 384d3190795SRafael Espindola template <class ELFT> void LinkerScript<ELFT>::output(InputSection<ELFT> *S) { 385d3190795SRafael Espindola if (!AlreadyOutputIS.insert(S).second) 386ceabe80eSEugene Leviant return; 387e08e78dfSRafael Espindola bool IsTbss = isTbss<ELFT>(CurOutSec); 388d3190795SRafael Espindola 389d3190795SRafael Espindola uintX_t Pos = IsTbss ? Dot + ThreadBssOffset : Dot; 390d3190795SRafael Espindola Pos = alignTo(Pos, S->Alignment); 39104a2e348SRafael Espindola S->OutSecOff = Pos - CurOutSec->Addr; 392d3190795SRafael Espindola Pos += S->getSize(); 393d3190795SRafael Espindola 394d3190795SRafael Espindola // Update output section size after adding each section. This is so that 395d3190795SRafael Espindola // SIZEOF works correctly in the case below: 396d3190795SRafael Espindola // .foo { *(.aaa) a = SIZEOF(.foo); *(.bbb) } 39704a2e348SRafael Espindola CurOutSec->Size = Pos - CurOutSec->Addr; 398d3190795SRafael Espindola 399b889744eSMeador Inge // If there is a memory region associated with this input section, then 400b889744eSMeador Inge // place the section in that region and update the region index. 401b889744eSMeador Inge if (CurMemRegion) { 402b889744eSMeador Inge CurMemRegion->Offset += CurOutSec->Size; 403b889744eSMeador Inge uint64_t CurSize = CurMemRegion->Offset - CurMemRegion->Origin; 404b889744eSMeador Inge if (CurSize > CurMemRegion->Length) { 405b889744eSMeador Inge uint64_t OverflowAmt = CurSize - CurMemRegion->Length; 406b889744eSMeador Inge error("section '" + CurOutSec->Name + "' will not fit in region '" + 407b889744eSMeador Inge CurMemRegion->Name + "': overflowed by " + Twine(OverflowAmt) + 408b889744eSMeador Inge " bytes"); 409b889744eSMeador Inge } 410b889744eSMeador Inge } 411b889744eSMeador Inge 4127252ae52SRafael Espindola if (IsTbss) 4137252ae52SRafael Espindola ThreadBssOffset = Pos - Dot; 4147252ae52SRafael Espindola else 415d3190795SRafael Espindola Dot = Pos; 4162de509c3SRui Ueyama } 417ceabe80eSEugene Leviant 418d3190795SRafael Espindola template <class ELFT> void LinkerScript<ELFT>::flush() { 41965499b90SRafael Espindola if (!CurOutSec || !AlreadyOutputOS.insert(CurOutSec).second) 42065499b90SRafael Espindola return; 42165499b90SRafael Espindola if (auto *OutSec = dyn_cast<OutputSection<ELFT>>(CurOutSec)) { 422d3190795SRafael Espindola for (InputSection<ELFT> *I : OutSec->Sections) 423d3190795SRafael Espindola output(I); 42465499b90SRafael Espindola } else { 42504a2e348SRafael Espindola Dot += CurOutSec->Size; 426d3190795SRafael Espindola } 427d3190795SRafael Espindola } 42897403d15SEugene Leviant 429d3190795SRafael Espindola template <class ELFT> 430e08e78dfSRafael Espindola void LinkerScript<ELFT>::switchTo(OutputSectionBase *Sec) { 431d3190795SRafael Espindola if (CurOutSec == Sec) 432d3190795SRafael Espindola return; 433d3190795SRafael Espindola if (AlreadyOutputOS.count(Sec)) 434d3190795SRafael Espindola return; 435d3190795SRafael Espindola 436d3190795SRafael Espindola flush(); 437d3190795SRafael Espindola CurOutSec = Sec; 438d3190795SRafael Espindola 43904a2e348SRafael Espindola Dot = alignTo(Dot, CurOutSec->Addralign); 440e08e78dfSRafael Espindola CurOutSec->Addr = isTbss<ELFT>(CurOutSec) ? Dot + ThreadBssOffset : Dot; 441b71d6f7aSEugene Leviant 442b71d6f7aSEugene Leviant // If neither AT nor AT> is specified for an allocatable section, the linker 443b71d6f7aSEugene Leviant // will set the LMA such that the difference between VMA and LMA for the 444b71d6f7aSEugene Leviant // section is the same as the preceding output section in the same region 445b71d6f7aSEugene Leviant // https://sourceware.org/binutils/docs-2.20/ld/Output-Section-LMA.html 446b71d6f7aSEugene Leviant CurOutSec->setLMAOffset(LMAOffset); 447d3190795SRafael Espindola } 448d3190795SRafael Espindola 449d3190795SRafael Espindola template <class ELFT> void LinkerScript<ELFT>::process(BaseCommand &Base) { 450e38cbab5SGeorge Rimar // This handles the assignments to symbol or to a location counter (.) 451d3190795SRafael Espindola if (auto *AssignCmd = dyn_cast<SymbolAssignment>(&Base)) { 45297403d15SEugene Leviant if (AssignCmd->Name == ".") { 45397403d15SEugene Leviant // Update to location counter means update to section size. 45414460e02SGeorge Rimar uintX_t Val = AssignCmd->Expression(Dot); 45514460e02SGeorge Rimar if (Val < Dot) 45614460e02SGeorge Rimar error("unable to move location counter backward for: " + 45714460e02SGeorge Rimar CurOutSec->Name); 45814460e02SGeorge Rimar Dot = Val; 45904a2e348SRafael Espindola CurOutSec->Size = Dot - CurOutSec->Addr; 460d3190795SRafael Espindola return; 46197403d15SEugene Leviant } 462b2b70975SGeorge Rimar assignSymbol<ELFT>(AssignCmd, Dot); 463d3190795SRafael Espindola return; 46497403d15SEugene Leviant } 465e38cbab5SGeorge Rimar 466e38cbab5SGeorge Rimar // Handle BYTE(), SHORT(), LONG(), or QUAD(). 467e38cbab5SGeorge Rimar if (auto *DataCmd = dyn_cast<BytesDataCommand>(&Base)) { 46804a2e348SRafael Espindola DataCmd->Offset = Dot - CurOutSec->Addr; 469e38cbab5SGeorge Rimar Dot += DataCmd->Size; 47004a2e348SRafael Espindola CurOutSec->Size = Dot - CurOutSec->Addr; 471e38cbab5SGeorge Rimar return; 472e38cbab5SGeorge Rimar } 473e38cbab5SGeorge Rimar 474b2d99d6aSMeador Inge if (auto *AssertCmd = dyn_cast<AssertCommand>(&Base)) { 475b2d99d6aSMeador Inge AssertCmd->Expression(Dot); 476b2d99d6aSMeador Inge return; 477b2d99d6aSMeador Inge } 478b2d99d6aSMeador Inge 479e38cbab5SGeorge Rimar // It handles single input section description command, 480e38cbab5SGeorge Rimar // calculates and assigns the offsets for each section and also 481e38cbab5SGeorge Rimar // updates the output section size. 482d3190795SRafael Espindola auto &ICmd = cast<InputSectionDescription>(Base); 483d3190795SRafael Espindola for (InputSectionData *ID : ICmd.Sections) { 4843fb5a6dcSGeorge Rimar // We tentatively added all synthetic sections at the beginning and removed 4853fb5a6dcSGeorge Rimar // empty ones afterwards (because there is no way to know whether they were 4863fb5a6dcSGeorge Rimar // going be empty or not other than actually running linker scripts.) 4873fb5a6dcSGeorge Rimar // We need to ignore remains of empty sections. 4883fb5a6dcSGeorge Rimar if (auto *Sec = dyn_cast<SyntheticSection<ELFT>>(ID)) 4893fb5a6dcSGeorge Rimar if (Sec->empty()) 4903fb5a6dcSGeorge Rimar continue; 4913fb5a6dcSGeorge Rimar 492d3190795SRafael Espindola auto *IB = static_cast<InputSectionBase<ELFT> *>(ID); 493d3190795SRafael Espindola switchTo(IB->OutSec); 494d3190795SRafael Espindola if (auto *I = dyn_cast<InputSection<ELFT>>(IB)) 495d3190795SRafael Espindola output(I); 49665499b90SRafael Espindola else 49765499b90SRafael Espindola flush(); 498ceabe80eSEugene Leviant } 499ceabe80eSEugene Leviant } 500ceabe80eSEugene Leviant 5018f66df92SGeorge Rimar template <class ELFT> 5022b074553SRafael Espindola static OutputSectionBase * 5032b074553SRafael Espindola findSection(StringRef Name, const std::vector<OutputSectionBase *> &Sections) { 5042b074553SRafael Espindola auto End = Sections.end(); 5052b074553SRafael Espindola auto HasName = [=](OutputSectionBase *Sec) { return Sec->getName() == Name; }; 5062b074553SRafael Espindola auto I = std::find_if(Sections.begin(), End, HasName); 507e08e78dfSRafael Espindola std::vector<OutputSectionBase *> Ret; 5082b074553SRafael Espindola if (I == End) 5092b074553SRafael Espindola return nullptr; 5102b074553SRafael Espindola assert(std::find_if(I + 1, End, HasName) == End); 5112b074553SRafael Espindola return *I; 5128f66df92SGeorge Rimar } 5138f66df92SGeorge Rimar 514b889744eSMeador Inge // This function searches for a memory region to place the given output 515b889744eSMeador Inge // section in. If found, a pointer to the appropriate memory region is 516b889744eSMeador Inge // returned. Otherwise, a nullptr is returned. 517b889744eSMeador Inge template <class ELFT> 518b889744eSMeador Inge MemoryRegion *LinkerScript<ELFT>::findMemoryRegion(OutputSectionCommand *Cmd, 519b889744eSMeador Inge OutputSectionBase *Sec) { 520b889744eSMeador Inge // If a memory region name was specified in the output section command, 521b889744eSMeador Inge // then try to find that region first. 522b889744eSMeador Inge if (!Cmd->MemoryRegionName.empty()) { 523b889744eSMeador Inge auto It = Opt.MemoryRegions.find(Cmd->MemoryRegionName); 524b889744eSMeador Inge if (It != Opt.MemoryRegions.end()) 525b889744eSMeador Inge return &It->second; 526b889744eSMeador Inge error("memory region '" + Cmd->MemoryRegionName + "' not declared"); 527b889744eSMeador Inge return nullptr; 528b889744eSMeador Inge } 529b889744eSMeador Inge 530b889744eSMeador Inge // The memory region name is empty, thus a suitable region must be 531b889744eSMeador Inge // searched for in the region map. If the region map is empty, just 532b889744eSMeador Inge // return. Note that this check doesn't happen at the very beginning 533b889744eSMeador Inge // so that uses of undeclared regions can be caught. 534b889744eSMeador Inge if (!Opt.MemoryRegions.size()) 535b889744eSMeador Inge return nullptr; 536b889744eSMeador Inge 537b889744eSMeador Inge // See if a region can be found by matching section flags. 538b889744eSMeador Inge for (auto &MRI : Opt.MemoryRegions) { 539b889744eSMeador Inge MemoryRegion &MR = MRI.second; 5408a8a953eSRui Ueyama if ((MR.Flags & Sec->Flags) != 0 && (MR.NegFlags & Sec->Flags) == 0) 541b889744eSMeador Inge return &MR; 542b889744eSMeador Inge } 543b889744eSMeador Inge 544b889744eSMeador Inge // Otherwise, no suitable region was found. 545b889744eSMeador Inge if (Sec->Flags & SHF_ALLOC) 546b889744eSMeador Inge error("no memory region specified for section '" + Sec->Name + "'"); 547b889744eSMeador Inge return nullptr; 548b889744eSMeador Inge } 549b889744eSMeador Inge 5500b1b695aSRui Ueyama // This function assigns offsets to input sections and an output section 5510b1b695aSRui Ueyama // for a single sections command (e.g. ".text { *(.text); }"). 552d3190795SRafael Espindola template <class ELFT> 553d3190795SRafael Espindola void LinkerScript<ELFT>::assignOffsets(OutputSectionCommand *Cmd) { 554b71d6f7aSEugene Leviant if (Cmd->LMAExpr) 555b71d6f7aSEugene Leviant LMAOffset = Cmd->LMAExpr(Dot) - Dot; 5562b074553SRafael Espindola OutputSectionBase *Sec = findSection<ELFT>(Cmd->Name, *OutputSections); 5572b074553SRafael Espindola if (!Sec) 558d3190795SRafael Espindola return; 559b889744eSMeador Inge 560165088aaSPetr Hosek // Handle align (e.g. ".foo : ALIGN(16) { ... }"). 561165088aaSPetr Hosek if (Cmd->AlignExpr) 562165088aaSPetr Hosek Sec->updateAlignment(Cmd->AlignExpr(0)); 563165088aaSPetr Hosek 564b889744eSMeador Inge // Try and find an appropriate memory region to assign offsets in. 565b889744eSMeador Inge CurMemRegion = findMemoryRegion(Cmd, Sec); 566b889744eSMeador Inge if (CurMemRegion) 567b889744eSMeador Inge Dot = CurMemRegion->Offset; 568b889744eSMeador Inge switchTo(Sec); 5690b1b695aSRui Ueyama 570d3190795SRafael Espindola // Find the last section output location. We will output orphan sections 571d3190795SRafael Espindola // there so that end symbols point to the correct location. 572d3190795SRafael Espindola auto E = std::find_if(Cmd->Commands.rbegin(), Cmd->Commands.rend(), 573d3190795SRafael Espindola [](const std::unique_ptr<BaseCommand> &Cmd) { 574d3190795SRafael Espindola return !isa<SymbolAssignment>(*Cmd); 575d3190795SRafael Espindola }) 576d3190795SRafael Espindola .base(); 577d3190795SRafael Espindola for (auto I = Cmd->Commands.begin(); I != E; ++I) 578d3190795SRafael Espindola process(**I); 5792506cb4dSEugene Leviant flush(); 580b31dd370SGeorge Rimar std::for_each(E, Cmd->Commands.end(), 581b31dd370SGeorge Rimar [this](std::unique_ptr<BaseCommand> &B) { process(*B.get()); }); 582d3190795SRafael Espindola } 583d3190795SRafael Espindola 58407fe6129SRafael Espindola template <class ELFT> void LinkerScript<ELFT>::removeEmptyCommands() { 5856d38e4dbSRafael Espindola // It is common practice to use very generic linker scripts. So for any 5866d38e4dbSRafael Espindola // given run some of the output sections in the script will be empty. 5876d38e4dbSRafael Espindola // We could create corresponding empty output sections, but that would 5886d38e4dbSRafael Espindola // clutter the output. 5896d38e4dbSRafael Espindola // We instead remove trivially empty sections. The bfd linker seems even 5906d38e4dbSRafael Espindola // more aggressive at removing them. 5916d38e4dbSRafael Espindola auto Pos = std::remove_if( 5926d38e4dbSRafael Espindola Opt.Commands.begin(), Opt.Commands.end(), 5936d38e4dbSRafael Espindola [&](const std::unique_ptr<BaseCommand> &Base) { 5940b1b695aSRui Ueyama if (auto *Cmd = dyn_cast<OutputSectionCommand>(Base.get())) 5952b074553SRafael Espindola return !findSection<ELFT>(Cmd->Name, *OutputSections); 5960b1b695aSRui Ueyama return false; 5976d38e4dbSRafael Espindola }); 5986d38e4dbSRafael Espindola Opt.Commands.erase(Pos, Opt.Commands.end()); 59907fe6129SRafael Espindola } 60007fe6129SRafael Espindola 6016a53737cSRafael Espindola static bool isAllSectionDescription(const OutputSectionCommand &Cmd) { 6026a53737cSRafael Espindola for (const std::unique_ptr<BaseCommand> &I : Cmd.Commands) 6036a53737cSRafael Espindola if (!isa<InputSectionDescription>(*I)) 6046a53737cSRafael Espindola return false; 6056a53737cSRafael Espindola return true; 6066a53737cSRafael Espindola } 6076d38e4dbSRafael Espindola 6086a53737cSRafael Espindola template <class ELFT> void LinkerScript<ELFT>::adjustSectionsBeforeSorting() { 6099546fffbSRafael Espindola // If the output section contains only symbol assignments, create a 6109546fffbSRafael Espindola // corresponding output section. The bfd linker seems to only create them if 6119546fffbSRafael Espindola // '.' is assigned to, but creating these section should not have any bad 6129546fffbSRafael Espindola // consequeces and gives us a section to put the symbol in. 6139546fffbSRafael Espindola uintX_t Flags = SHF_ALLOC; 614f93b8c29SRafael Espindola uint32_t Type = SHT_NOBITS; 6159546fffbSRafael Espindola for (const std::unique_ptr<BaseCommand> &Base : Opt.Commands) { 6169546fffbSRafael Espindola auto *Cmd = dyn_cast<OutputSectionCommand>(Base.get()); 6179546fffbSRafael Espindola if (!Cmd) 6189546fffbSRafael Espindola continue; 6192b074553SRafael Espindola if (OutputSectionBase *Sec = 6202b074553SRafael Espindola findSection<ELFT>(Cmd->Name, *OutputSections)) { 6212b074553SRafael Espindola Flags = Sec->Flags; 6222b074553SRafael Espindola Type = Sec->Type; 6239546fffbSRafael Espindola continue; 6249546fffbSRafael Espindola } 6259546fffbSRafael Espindola 6266a53737cSRafael Espindola if (isAllSectionDescription(*Cmd)) 6276a53737cSRafael Espindola continue; 6286a53737cSRafael Espindola 62995642b95SRui Ueyama auto *OutSec = make<OutputSection<ELFT>>(Cmd->Name, Type, Flags); 6309546fffbSRafael Espindola OutputSections->push_back(OutSec); 6319546fffbSRafael Espindola } 632f7a17448SRafael Espindola } 633f7a17448SRafael Espindola 634f7a17448SRafael Espindola template <class ELFT> void LinkerScript<ELFT>::adjustSectionsAfterSorting() { 635f7a17448SRafael Espindola placeOrphanSections(); 636f7a17448SRafael Espindola 637f7a17448SRafael Espindola // If output section command doesn't specify any segments, 638f7a17448SRafael Espindola // and we haven't previously assigned any section to segment, 639f7a17448SRafael Espindola // then we simply assign section to the very first load segment. 640f7a17448SRafael Espindola // Below is an example of such linker script: 641f7a17448SRafael Espindola // PHDRS { seg PT_LOAD; } 642f7a17448SRafael Espindola // SECTIONS { .aaa : { *(.aaa) } } 643f7a17448SRafael Espindola std::vector<StringRef> DefPhdrs; 644f7a17448SRafael Espindola auto FirstPtLoad = 645f7a17448SRafael Espindola std::find_if(Opt.PhdrsCommands.begin(), Opt.PhdrsCommands.end(), 646f7a17448SRafael Espindola [](const PhdrsCommand &Cmd) { return Cmd.Type == PT_LOAD; }); 647f7a17448SRafael Espindola if (FirstPtLoad != Opt.PhdrsCommands.end()) 648f7a17448SRafael Espindola DefPhdrs.push_back(FirstPtLoad->Name); 649f7a17448SRafael Espindola 650f7a17448SRafael Espindola // Walk the commands and propagate the program headers to commands that don't 651f7a17448SRafael Espindola // explicitly specify them. 652f7a17448SRafael Espindola for (const std::unique_ptr<BaseCommand> &Base : Opt.Commands) { 653f7a17448SRafael Espindola auto *Cmd = dyn_cast<OutputSectionCommand>(Base.get()); 654f7a17448SRafael Espindola if (!Cmd) 655f7a17448SRafael Espindola continue; 656f7a17448SRafael Espindola if (Cmd->Phdrs.empty()) 657f7a17448SRafael Espindola Cmd->Phdrs = DefPhdrs; 658f7a17448SRafael Espindola else 659f7a17448SRafael Espindola DefPhdrs = Cmd->Phdrs; 660f7a17448SRafael Espindola } 6616a53737cSRafael Espindola 6626a53737cSRafael Espindola removeEmptyCommands(); 6639546fffbSRafael Espindola } 6649546fffbSRafael Espindola 66515c57951SRafael Espindola // When placing orphan sections, we want to place them after symbol assignments 66615c57951SRafael Espindola // so that an orphan after 66715c57951SRafael Espindola // begin_foo = .; 66815c57951SRafael Espindola // foo : { *(foo) } 66915c57951SRafael Espindola // end_foo = .; 67015c57951SRafael Espindola // doesn't break the intended meaning of the begin/end symbols. 67115c57951SRafael Espindola // We don't want to go over sections since Writer<ELFT>::sortSections is the 67215c57951SRafael Espindola // one in charge of deciding the order of the sections. 67315c57951SRafael Espindola // We don't want to go over alignments, since doing so in 67415c57951SRafael Espindola // rx_sec : { *(rx_sec) } 67515c57951SRafael Espindola // . = ALIGN(0x1000); 67615c57951SRafael Espindola // /* The RW PT_LOAD starts here*/ 67715c57951SRafael Espindola // rw_sec : { *(rw_sec) } 67815c57951SRafael Espindola // would mean that the RW PT_LOAD would become unaligned. 6795fcc99c2SRafael Espindola static bool shouldSkip(const BaseCommand &Cmd) { 68015c57951SRafael Espindola if (isa<OutputSectionCommand>(Cmd)) 68115c57951SRafael Espindola return false; 68215c57951SRafael Espindola const auto *Assign = dyn_cast<SymbolAssignment>(&Cmd); 68315c57951SRafael Espindola if (!Assign) 68415c57951SRafael Espindola return true; 6855fcc99c2SRafael Espindola return Assign->Name != "."; 68615c57951SRafael Espindola } 68715c57951SRafael Espindola 6886697ec29SRui Ueyama // Orphan sections are sections present in the input files which are 6896697ec29SRui Ueyama // not explicitly placed into the output file by the linker script. 6906697ec29SRui Ueyama // 6916697ec29SRui Ueyama // When the control reaches this function, Opt.Commands contains 6926697ec29SRui Ueyama // output section commands for non-orphan sections only. This function 6936697ec29SRui Ueyama // adds new elements for orphan sections to Opt.Commands so that all 6946697ec29SRui Ueyama // sections are explicitly handled by Opt.Commands. 6956697ec29SRui Ueyama // 6966697ec29SRui Ueyama // Writer<ELFT>::sortSections has already sorted output sections. 6976697ec29SRui Ueyama // What we need to do is to scan OutputSections vector and 6986697ec29SRui Ueyama // Opt.Commands in parallel to find orphan sections. If there is an 6996697ec29SRui Ueyama // output section that doesn't have a corresponding entry in 7006697ec29SRui Ueyama // Opt.Commands, we will insert a new entry to Opt.Commands. 7016697ec29SRui Ueyama // 7026697ec29SRui Ueyama // There is some ambiguity as to where exactly a new entry should be 7036697ec29SRui Ueyama // inserted, because Opt.Commands contains not only output section 7046697ec29SRui Ueyama // commands but other types of commands such as symbol assignment 7056697ec29SRui Ueyama // expressions. There's no correct answer here due to the lack of the 7066697ec29SRui Ueyama // formal specification of the linker script. We use heuristics to 7076697ec29SRui Ueyama // determine whether a new output command should be added before or 7086697ec29SRui Ueyama // after another commands. For the details, look at shouldSkip 7096697ec29SRui Ueyama // function. 71093c64025SGeorge Rimar template <class ELFT> void LinkerScript<ELFT>::placeOrphanSections() { 711aab6d5c5SRafael Espindola // The OutputSections are already in the correct order. 712aab6d5c5SRafael Espindola // This loops creates or moves commands as needed so that they are in the 713aab6d5c5SRafael Espindola // correct order. 714aab6d5c5SRafael Espindola int CmdIndex = 0; 7155fcc99c2SRafael Espindola 7165fcc99c2SRafael Espindola // As a horrible special case, skip the first . assignment if it is before any 7175fcc99c2SRafael Espindola // section. We do this because it is common to set a load address by starting 7185fcc99c2SRafael Espindola // the script with ". = 0xabcd" and the expectation is that every section is 7195fcc99c2SRafael Espindola // after that. 7205fcc99c2SRafael Espindola auto FirstSectionOrDotAssignment = 7215fcc99c2SRafael Espindola std::find_if(Opt.Commands.begin(), Opt.Commands.end(), 7225fcc99c2SRafael Espindola [](const std::unique_ptr<BaseCommand> &Cmd) { 7235fcc99c2SRafael Espindola if (isa<OutputSectionCommand>(*Cmd)) 7245fcc99c2SRafael Espindola return true; 7255fcc99c2SRafael Espindola const auto *Assign = dyn_cast<SymbolAssignment>(Cmd.get()); 7265fcc99c2SRafael Espindola if (!Assign) 7275fcc99c2SRafael Espindola return false; 7285fcc99c2SRafael Espindola return Assign->Name == "."; 7295fcc99c2SRafael Espindola }); 7305fcc99c2SRafael Espindola if (FirstSectionOrDotAssignment != Opt.Commands.end()) { 7315fcc99c2SRafael Espindola CmdIndex = FirstSectionOrDotAssignment - Opt.Commands.begin(); 7325fcc99c2SRafael Espindola if (isa<SymbolAssignment>(**FirstSectionOrDotAssignment)) 7335fcc99c2SRafael Espindola ++CmdIndex; 7345fcc99c2SRafael Espindola } 7355fcc99c2SRafael Espindola 736e08e78dfSRafael Espindola for (OutputSectionBase *Sec : *OutputSections) { 737652852c5SGeorge Rimar StringRef Name = Sec->getName(); 738aab6d5c5SRafael Espindola 739aab6d5c5SRafael Espindola // Find the last spot where we can insert a command and still get the 74015c57951SRafael Espindola // correct result. 741aab6d5c5SRafael Espindola auto CmdIter = Opt.Commands.begin() + CmdIndex; 742aab6d5c5SRafael Espindola auto E = Opt.Commands.end(); 7435fcc99c2SRafael Espindola while (CmdIter != E && shouldSkip(**CmdIter)) { 744aab6d5c5SRafael Espindola ++CmdIter; 745aab6d5c5SRafael Espindola ++CmdIndex; 746aab6d5c5SRafael Espindola } 747aab6d5c5SRafael Espindola 748aab6d5c5SRafael Espindola auto Pos = 749aab6d5c5SRafael Espindola std::find_if(CmdIter, E, [&](const std::unique_ptr<BaseCommand> &Base) { 750aab6d5c5SRafael Espindola auto *Cmd = dyn_cast<OutputSectionCommand>(Base.get()); 751aab6d5c5SRafael Espindola return Cmd && Cmd->Name == Name; 752aab6d5c5SRafael Espindola }); 753aab6d5c5SRafael Espindola if (Pos == E) { 754aab6d5c5SRafael Espindola Opt.Commands.insert(CmdIter, 755aab6d5c5SRafael Espindola llvm::make_unique<OutputSectionCommand>(Name)); 756aab6d5c5SRafael Espindola ++CmdIndex; 75715c57951SRafael Espindola continue; 75815c57951SRafael Espindola } 75915c57951SRafael Espindola 76015c57951SRafael Espindola // Continue from where we found it. 76115c57951SRafael Espindola CmdIndex = (Pos - Opt.Commands.begin()) + 1; 762652852c5SGeorge Rimar } 763337f903cSRafael Espindola } 764337f903cSRafael Espindola 765337f903cSRafael Espindola template <class ELFT> 76617cb7c0aSRafael Espindola void LinkerScript<ELFT>::assignAddresses(std::vector<PhdrEntry> &Phdrs) { 7677c18c28cSRui Ueyama // Assign addresses as instructed by linker script SECTIONS sub-commands. 768be607334SRafael Espindola Dot = 0; 769652852c5SGeorge Rimar 77006f4743aSRafael Espindola // A symbol can be assigned before any section is mentioned in the linker 77106f4743aSRafael Espindola // script. In an DSO, the symbol values are addresses, so the only important 77206f4743aSRafael Espindola // section values are: 77306f4743aSRafael Espindola // * SHN_UNDEF 77406f4743aSRafael Espindola // * SHN_ABS 77506f4743aSRafael Espindola // * Any value meaning a regular section. 77606f4743aSRafael Espindola // To handle that, create a dummy aether section that fills the void before 77706f4743aSRafael Espindola // the linker scripts switches to another section. It has an index of one 77806f4743aSRafael Espindola // which will map to whatever the first actual section is. 77906f4743aSRafael Espindola auto *Aether = make<OutputSectionBase>("", 0, SHF_ALLOC); 78006f4743aSRafael Espindola Aether->SectionIndex = 1; 78106f4743aSRafael Espindola switchTo(Aether); 78206f4743aSRafael Espindola 783076fe157SGeorge Rimar for (const std::unique_ptr<BaseCommand> &Base : Opt.Commands) { 784076fe157SGeorge Rimar if (auto *Cmd = dyn_cast<SymbolAssignment>(Base.get())) { 7858d083e6aSRui Ueyama if (Cmd->Name == ".") { 7868d083e6aSRui Ueyama Dot = Cmd->Expression(Dot); 7878d083e6aSRui Ueyama } else if (Cmd->Sym) { 788b2b70975SGeorge Rimar assignSymbol<ELFT>(Cmd, Dot); 7898d083e6aSRui Ueyama } 79005ef4cffSRui Ueyama continue; 791652852c5SGeorge Rimar } 792652852c5SGeorge Rimar 793eefa758eSGeorge Rimar if (auto *Cmd = dyn_cast<AssertCommand>(Base.get())) { 794eefa758eSGeorge Rimar Cmd->Expression(Dot); 795eefa758eSGeorge Rimar continue; 796eefa758eSGeorge Rimar } 797eefa758eSGeorge Rimar 798076fe157SGeorge Rimar auto *Cmd = cast<OutputSectionCommand>(Base.get()); 79958e5c4dcSGeorge Rimar if (Cmd->AddrExpr) 80058e5c4dcSGeorge Rimar Dot = Cmd->AddrExpr(Dot); 801d3190795SRafael Espindola assignOffsets(Cmd); 802a14b13d8SGeorge Rimar } 803467c4d55SEugene Leviant 804aab6d5c5SRafael Espindola uintX_t MinVA = std::numeric_limits<uintX_t>::max(); 805193b158bSRafael Espindola for (OutputSectionBase *Sec : *OutputSections) 80604a2e348SRafael Espindola if (Sec->Flags & SHF_ALLOC) 807e08e78dfSRafael Espindola MinVA = std::min<uint64_t>(MinVA, Sec->Addr); 808aab6d5c5SRafael Espindola 8098c495e20SRafael Espindola allocateHeaders<ELFT>(Phdrs, *OutputSections, MinVA); 810fb8978fcSDima Stepanov } 811652852c5SGeorge Rimar 812464daadcSRui Ueyama // Creates program headers as instructed by PHDRS linker script command. 81317cb7c0aSRafael Espindola template <class ELFT> std::vector<PhdrEntry> LinkerScript<ELFT>::createPhdrs() { 81417cb7c0aSRafael Espindola std::vector<PhdrEntry> Ret; 815bbe38602SEugene Leviant 816464daadcSRui Ueyama // Process PHDRS and FILEHDR keywords because they are not 817464daadcSRui Ueyama // real output sections and cannot be added in the following loop. 818bbe38602SEugene Leviant for (const PhdrsCommand &Cmd : Opt.PhdrsCommands) { 819edebbdf1SRui Ueyama Ret.emplace_back(Cmd.Type, Cmd.Flags == UINT_MAX ? PF_R : Cmd.Flags); 82017cb7c0aSRafael Espindola PhdrEntry &Phdr = Ret.back(); 821bbe38602SEugene Leviant 822bbe38602SEugene Leviant if (Cmd.HasFilehdr) 823adca245fSRui Ueyama Phdr.add(Out<ELFT>::ElfHeader); 824bbe38602SEugene Leviant if (Cmd.HasPhdrs) 825adca245fSRui Ueyama Phdr.add(Out<ELFT>::ProgramHeaders); 82656b21c86SEugene Leviant 82756b21c86SEugene Leviant if (Cmd.LMAExpr) { 82817cb7c0aSRafael Espindola Phdr.p_paddr = Cmd.LMAExpr(0); 82956b21c86SEugene Leviant Phdr.HasLMA = true; 83056b21c86SEugene Leviant } 831bbe38602SEugene Leviant } 832bbe38602SEugene Leviant 833464daadcSRui Ueyama // Add output sections to program headers. 834e08e78dfSRafael Espindola for (OutputSectionBase *Sec : *OutputSections) { 83504a2e348SRafael Espindola if (!(Sec->Flags & SHF_ALLOC)) 836bbe38602SEugene Leviant break; 837bbe38602SEugene Leviant 838bbe38602SEugene Leviant // Assign headers specified by linker script 839f7a17448SRafael Espindola for (size_t Id : getPhdrIndices(Sec->getName())) { 840edebbdf1SRui Ueyama Ret[Id].add(Sec); 841865bf863SEugene Leviant if (Opt.PhdrsCommands[Id].Flags == UINT_MAX) 84217cb7c0aSRafael Espindola Ret[Id].p_flags |= Sec->getPhdrFlags(); 843bbe38602SEugene Leviant } 844bbe38602SEugene Leviant } 845edebbdf1SRui Ueyama return Ret; 846bbe38602SEugene Leviant } 847bbe38602SEugene Leviant 848f9bc3bd2SEugene Leviant template <class ELFT> bool LinkerScript<ELFT>::ignoreInterpSection() { 849f9bc3bd2SEugene Leviant // Ignore .interp section in case we have PHDRS specification 850f9bc3bd2SEugene Leviant // and PT_INTERP isn't listed. 851f9bc3bd2SEugene Leviant return !Opt.PhdrsCommands.empty() && 852f9bc3bd2SEugene Leviant llvm::find_if(Opt.PhdrsCommands, [](const PhdrsCommand &Cmd) { 853f9bc3bd2SEugene Leviant return Cmd.Type == PT_INTERP; 854f9bc3bd2SEugene Leviant }) == Opt.PhdrsCommands.end(); 855f9bc3bd2SEugene Leviant } 856f9bc3bd2SEugene Leviant 85793c64025SGeorge Rimar template <class ELFT> uint32_t LinkerScript<ELFT>::getFiller(StringRef Name) { 858f6c3ccefSGeorge Rimar for (const std::unique_ptr<BaseCommand> &Base : Opt.Commands) 859f6c3ccefSGeorge Rimar if (auto *Cmd = dyn_cast<OutputSectionCommand>(Base.get())) 860f6c3ccefSGeorge Rimar if (Cmd->Name == Name) 861f6c3ccefSGeorge Rimar return Cmd->Filler; 86216068aebSRui Ueyama return 0; 863e2ee72b5SGeorge Rimar } 864e2ee72b5SGeorge Rimar 865e38cbab5SGeorge Rimar template <class ELFT> 866e38cbab5SGeorge Rimar static void writeInt(uint8_t *Buf, uint64_t Data, uint64_t Size) { 867e38cbab5SGeorge Rimar const endianness E = ELFT::TargetEndianness; 868e38cbab5SGeorge Rimar 869e38cbab5SGeorge Rimar switch (Size) { 870e38cbab5SGeorge Rimar case 1: 871e38cbab5SGeorge Rimar *Buf = (uint8_t)Data; 872e38cbab5SGeorge Rimar break; 873e38cbab5SGeorge Rimar case 2: 874e38cbab5SGeorge Rimar write16<E>(Buf, Data); 875e38cbab5SGeorge Rimar break; 876e38cbab5SGeorge Rimar case 4: 877e38cbab5SGeorge Rimar write32<E>(Buf, Data); 878e38cbab5SGeorge Rimar break; 879e38cbab5SGeorge Rimar case 8: 880e38cbab5SGeorge Rimar write64<E>(Buf, Data); 881e38cbab5SGeorge Rimar break; 882e38cbab5SGeorge Rimar default: 883e38cbab5SGeorge Rimar llvm_unreachable("unsupported Size argument"); 884e38cbab5SGeorge Rimar } 885e38cbab5SGeorge Rimar } 886e38cbab5SGeorge Rimar 887e38cbab5SGeorge Rimar template <class ELFT> 888e38cbab5SGeorge Rimar void LinkerScript<ELFT>::writeDataBytes(StringRef Name, uint8_t *Buf) { 889e38cbab5SGeorge Rimar int I = getSectionIndex(Name); 890e38cbab5SGeorge Rimar if (I == INT_MAX) 891e38cbab5SGeorge Rimar return; 892e38cbab5SGeorge Rimar 8936e68c5e5SRui Ueyama auto *Cmd = dyn_cast<OutputSectionCommand>(Opt.Commands[I].get()); 8946e68c5e5SRui Ueyama for (const std::unique_ptr<BaseCommand> &Base : Cmd->Commands) 8956e68c5e5SRui Ueyama if (auto *Data = dyn_cast<BytesDataCommand>(Base.get())) 89695c7d8d2SMeador Inge writeInt<ELFT>(Buf + Data->Offset, Data->Expression(0), Data->Size); 897e38cbab5SGeorge Rimar } 898e38cbab5SGeorge Rimar 899b71d6f7aSEugene Leviant template <class ELFT> bool LinkerScript<ELFT>::hasLMA(StringRef Name) { 9008ceadb38SGeorge Rimar for (const std::unique_ptr<BaseCommand> &Base : Opt.Commands) 9018ceadb38SGeorge Rimar if (auto *Cmd = dyn_cast<OutputSectionCommand>(Base.get())) 902b71d6f7aSEugene Leviant if (Cmd->LMAExpr && Cmd->Name == Name) 903b71d6f7aSEugene Leviant return true; 904b71d6f7aSEugene Leviant return false; 9058ceadb38SGeorge Rimar } 9068ceadb38SGeorge Rimar 907c3e2a4b0SRui Ueyama // Returns the index of the given section name in linker script 908c3e2a4b0SRui Ueyama // SECTIONS commands. Sections are laid out as the same order as they 909c3e2a4b0SRui Ueyama // were in the script. If a given name did not appear in the script, 910c3e2a4b0SRui Ueyama // it returns INT_MAX, so that it will be laid out at end of file. 911076fe157SGeorge Rimar template <class ELFT> int LinkerScript<ELFT>::getSectionIndex(StringRef Name) { 9126e68c5e5SRui Ueyama for (int I = 0, E = Opt.Commands.size(); I != E; ++I) 9136e68c5e5SRui Ueyama if (auto *Cmd = dyn_cast<OutputSectionCommand>(Opt.Commands[I].get())) 914076fe157SGeorge Rimar if (Cmd->Name == Name) 915f510fa6bSRui Ueyama return I; 916f510fa6bSRui Ueyama return INT_MAX; 91771b26e94SGeorge Rimar } 91871b26e94SGeorge Rimar 919bbe38602SEugene Leviant template <class ELFT> bool LinkerScript<ELFT>::hasPhdrsCommands() { 920bbe38602SEugene Leviant return !Opt.PhdrsCommands.empty(); 921bbe38602SEugene Leviant } 922bbe38602SEugene Leviant 9239e69450eSGeorge Rimar template <class ELFT> 924ed30ce7aSEugene Leviant const OutputSectionBase *LinkerScript<ELFT>::getOutputSection(const Twine &Loc, 925ed30ce7aSEugene Leviant StringRef Name) { 926afaa9343SEugene Leviant static OutputSectionBase FakeSec("", 0, 0); 92796659df0SGeorge Rimar 928e08e78dfSRafael Espindola for (OutputSectionBase *Sec : *OutputSections) 929b71d6f7aSEugene Leviant if (Sec->getName() == Name) 930afaa9343SEugene Leviant return Sec; 931ed30ce7aSEugene Leviant 932ed30ce7aSEugene Leviant error(Loc + ": undefined section " + Name); 933afaa9343SEugene Leviant return &FakeSec; 93436fac7f0SEugene Leviant } 93536fac7f0SEugene Leviant 936edf75e79SRui Ueyama // This function is essentially the same as getOutputSection(Name)->Size, 937edf75e79SRui Ueyama // but it won't print out an error message if a given section is not found. 938edf75e79SRui Ueyama // 939edf75e79SRui Ueyama // Linker script does not create an output section if its content is empty. 940edf75e79SRui Ueyama // We want to allow SIZEOF(.foo) where .foo is a section which happened to 941edf75e79SRui Ueyama // be empty. That is why this function is different from getOutputSection(). 942edf75e79SRui Ueyama template <class ELFT> 943edf75e79SRui Ueyama uint64_t LinkerScript<ELFT>::getOutputSectionSize(StringRef Name) { 944edf75e79SRui Ueyama for (OutputSectionBase *Sec : *OutputSections) 945edf75e79SRui Ueyama if (Sec->getName() == Name) 946edf75e79SRui Ueyama return Sec->Size; 947edf75e79SRui Ueyama return 0; 948edf75e79SRui Ueyama } 949edf75e79SRui Ueyama 950884e786dSGeorge Rimar template <class ELFT> uint64_t LinkerScript<ELFT>::getHeaderSize() { 9510d4b6d5cSRafael Espindola return elf::getHeaderSize<ELFT>(); 952e32a3598SGeorge Rimar } 953e32a3598SGeorge Rimar 954f6aeed36SEugene Leviant template <class ELFT> 955f6aeed36SEugene Leviant uint64_t LinkerScript<ELFT>::getSymbolValue(const Twine &Loc, StringRef S) { 956884e786dSGeorge Rimar if (SymbolBody *B = Symtab<ELFT>::X->find(S)) 957884e786dSGeorge Rimar return B->getVA<ELFT>(); 958f6aeed36SEugene Leviant error(Loc + ": symbol not found: " + S); 959884e786dSGeorge Rimar return 0; 960884e786dSGeorge Rimar } 961884e786dSGeorge Rimar 962f34f45fdSGeorge Rimar template <class ELFT> bool LinkerScript<ELFT>::isDefined(StringRef S) { 963f34f45fdSGeorge Rimar return Symtab<ELFT>::X->find(S) != nullptr; 964f34f45fdSGeorge Rimar } 965f34f45fdSGeorge Rimar 9662f831dcaSRafael Espindola template <class ELFT> bool LinkerScript<ELFT>::isAbsolute(StringRef S) { 9672f831dcaSRafael Espindola SymbolBody *Sym = Symtab<ELFT>::X->find(S); 9682f831dcaSRafael Espindola auto *DR = dyn_cast_or_null<DefinedRegular<ELFT>>(Sym); 9692f831dcaSRafael Espindola return DR && !DR->Section; 9702f831dcaSRafael Espindola } 9712f831dcaSRafael Espindola 972afaa9343SEugene Leviant // Gets section symbol belongs to. Symbol "." doesn't belong to any 973afaa9343SEugene Leviant // specific section but isn't absolute at the same time, so we try 974afaa9343SEugene Leviant // to find suitable section for it as well. 975afaa9343SEugene Leviant template <class ELFT> 976afaa9343SEugene Leviant const OutputSectionBase *LinkerScript<ELFT>::getSymbolSection(StringRef S) { 97706f4743aSRafael Espindola if (SymbolBody *Sym = Symtab<ELFT>::X->find(S)) 97860aed443SGeorge Rimar return SymbolTableSection<ELFT>::getOutputSection(Sym); 97906f4743aSRafael Espindola return CurOutSec; 980afaa9343SEugene Leviant } 981afaa9343SEugene Leviant 982bbe38602SEugene Leviant // Returns indices of ELF headers containing specific section, identified 983bbe38602SEugene Leviant // by Name. Each index is a zero based number of ELF header listed within 984bbe38602SEugene Leviant // PHDRS {} script block. 985bbe38602SEugene Leviant template <class ELFT> 986edebbdf1SRui Ueyama std::vector<size_t> LinkerScript<ELFT>::getPhdrIndices(StringRef SectionName) { 987076fe157SGeorge Rimar for (const std::unique_ptr<BaseCommand> &Base : Opt.Commands) { 988076fe157SGeorge Rimar auto *Cmd = dyn_cast<OutputSectionCommand>(Base.get()); 989edebbdf1SRui Ueyama if (!Cmd || Cmd->Name != SectionName) 99031d842f5SGeorge Rimar continue; 99131d842f5SGeorge Rimar 99229c5a2a9SRui Ueyama std::vector<size_t> Ret; 99329c5a2a9SRui Ueyama for (StringRef PhdrName : Cmd->Phdrs) 9942a942c4bSEugene Leviant Ret.push_back(getPhdrIndex(Cmd->Location, PhdrName)); 99529c5a2a9SRui Ueyama return Ret; 996bbe38602SEugene Leviant } 99731d842f5SGeorge Rimar return {}; 99831d842f5SGeorge Rimar } 999bbe38602SEugene Leviant 100029c5a2a9SRui Ueyama template <class ELFT> 10012a942c4bSEugene Leviant size_t LinkerScript<ELFT>::getPhdrIndex(const Twine &Loc, StringRef PhdrName) { 100229c5a2a9SRui Ueyama size_t I = 0; 100329c5a2a9SRui Ueyama for (PhdrsCommand &Cmd : Opt.PhdrsCommands) { 100429c5a2a9SRui Ueyama if (Cmd.Name == PhdrName) 100529c5a2a9SRui Ueyama return I; 100629c5a2a9SRui Ueyama ++I; 100729c5a2a9SRui Ueyama } 10082a942c4bSEugene Leviant error(Loc + ": section header '" + PhdrName + "' is not listed in PHDRS"); 100929c5a2a9SRui Ueyama return 0; 101029c5a2a9SRui Ueyama } 101129c5a2a9SRui Ueyama 101203ff0166SEugene Leviant class elf::ScriptParser final : public ScriptParserBase { 1013c3794e58SGeorge Rimar typedef void (ScriptParser::*Handler)(); 1014c3794e58SGeorge Rimar 1015f7c5fbb1SRui Ueyama public: 101622375f24SRui Ueyama ScriptParser(MemoryBufferRef MB) 101722375f24SRui Ueyama : ScriptParserBase(MB), 101822375f24SRui Ueyama IsUnderSysroot(isUnderSysroot(MB.getBufferIdentifier())) {} 1019f23b2320SGeorge Rimar 102020b6598cSGeorge Rimar void readLinkerScript(); 102120b6598cSGeorge Rimar void readVersionScript(); 1022d0ebd84cSRafael Espindola void readDynamicList(); 1023f7c5fbb1SRui Ueyama 1024f7c5fbb1SRui Ueyama private: 102552a1509eSRui Ueyama void addFile(StringRef Path); 102652a1509eSRui Ueyama 1027f7c5fbb1SRui Ueyama void readAsNeeded(); 102890c5099eSDenis Protivensky void readEntry(); 102983f406cfSGeorge Rimar void readExtern(); 1030f7c5fbb1SRui Ueyama void readGroup(); 103131aa1f83SRui Ueyama void readInclude(); 1032b889744eSMeador Inge void readMemory(); 1033ee59282bSRui Ueyama void readOutput(); 10349159ce93SDavide Italiano void readOutputArch(); 1035f7c5fbb1SRui Ueyama void readOutputFormat(); 1036bbe38602SEugene Leviant void readPhdrs(); 103768a39a65SDavide Italiano void readSearchDir(); 10388e3b38abSDenis Protivensky void readSections(); 103995769b4aSRui Ueyama void readVersion(); 104095769b4aSRui Ueyama void readVersionScriptCommand(); 10418e3b38abSDenis Protivensky 1042113cdec9SRui Ueyama SymbolAssignment *readAssignment(StringRef Name); 1043e38cbab5SGeorge Rimar BytesDataCommand *readBytesDataCommand(StringRef Tok); 104416068aebSRui Ueyama uint32_t readFill(); 104510416564SRui Ueyama OutputSectionCommand *readOutputSectionDescription(StringRef OutSec); 104616068aebSRui Ueyama uint32_t readOutputSectionFiller(StringRef Tok); 1047bbe38602SEugene Leviant std::vector<StringRef> readOutputSectionPhdrs(); 1048a2496cbeSGeorge Rimar InputSectionDescription *readInputSectionDescription(StringRef Tok); 1049db688454SEugene Leviant StringMatcher readFilePatterns(); 105007171f21SGeorge Rimar std::vector<SectionPattern> readInputSectionsList(); 1051a2496cbeSGeorge Rimar InputSectionDescription *readInputSectionRules(StringRef FilePattern); 1052bbe38602SEugene Leviant unsigned readPhdrType(); 1053be394db3SGeorge Rimar SortSectionPolicy readSortKind(); 1054a35e39caSPetr Hosek SymbolAssignment *readProvideHidden(bool Provide, bool Hidden); 1055c96da110SRafael Espindola SymbolAssignment *readProvideOrAssignment(StringRef Tok); 105603fc010eSGeorge Rimar void readSort(); 1057eefa758eSGeorge Rimar Expr readAssert(); 1058708019c4SRui Ueyama 105924e626ccSRui Ueyama uint64_t readMemoryAssignment(StringRef, StringRef, StringRef); 106024e626ccSRui Ueyama std::pair<uint32_t, uint32_t> readMemoryAttributes(); 106124e626ccSRui Ueyama 1062708019c4SRui Ueyama Expr readExpr(); 1063708019c4SRui Ueyama Expr readExpr1(Expr Lhs, int MinPrec); 1064b71d6f7aSEugene Leviant StringRef readParenLiteral(); 1065708019c4SRui Ueyama Expr readPrimary(); 1066708019c4SRui Ueyama Expr readTernary(Expr Cond); 10676ad7dfccSRui Ueyama Expr readParenExpr(); 1068f7c5fbb1SRui Ueyama 106920b6598cSGeorge Rimar // For parsing version script. 107012450b20SRui Ueyama std::vector<SymbolVersion> readVersionExtern(); 107112450b20SRui Ueyama void readAnonymousDeclaration(); 107295769b4aSRui Ueyama void readVersionDeclaration(StringRef VerStr); 107312450b20SRui Ueyama std::vector<SymbolVersion> readSymbols(); 1074e999ddb8SRafael Espindola void readLocals(); 107520b6598cSGeorge Rimar 107607320e40SRui Ueyama ScriptConfiguration &Opt = *ScriptConfig; 107716b0cc9eSSimon Atanasyan bool IsUnderSysroot; 1078f7c5fbb1SRui Ueyama }; 1079f7c5fbb1SRui Ueyama 1080d0ebd84cSRafael Espindola void ScriptParser::readDynamicList() { 1081d0ebd84cSRafael Espindola expect("{"); 1082d0ebd84cSRafael Espindola readAnonymousDeclaration(); 1083d0ebd84cSRafael Espindola if (!atEOF()) 1084d0ebd84cSRafael Espindola setError("EOF expected, but got " + next()); 1085d0ebd84cSRafael Espindola } 1086d0ebd84cSRafael Espindola 108720b6598cSGeorge Rimar void ScriptParser::readVersionScript() { 108895769b4aSRui Ueyama readVersionScriptCommand(); 108920b6598cSGeorge Rimar if (!atEOF()) 109095769b4aSRui Ueyama setError("EOF expected, but got " + next()); 109195769b4aSRui Ueyama } 109295769b4aSRui Ueyama 109395769b4aSRui Ueyama void ScriptParser::readVersionScriptCommand() { 109483043f23SRui Ueyama if (consume("{")) { 109512450b20SRui Ueyama readAnonymousDeclaration(); 109620b6598cSGeorge Rimar return; 109720b6598cSGeorge Rimar } 109820b6598cSGeorge Rimar 109995769b4aSRui Ueyama while (!atEOF() && !Error && peek() != "}") { 110020b6598cSGeorge Rimar StringRef VerStr = next(); 110120b6598cSGeorge Rimar if (VerStr == "{") { 110295769b4aSRui Ueyama setError("anonymous version definition is used in " 110395769b4aSRui Ueyama "combination with other version definitions"); 110420b6598cSGeorge Rimar return; 110520b6598cSGeorge Rimar } 110620b6598cSGeorge Rimar expect("{"); 110795769b4aSRui Ueyama readVersionDeclaration(VerStr); 110820b6598cSGeorge Rimar } 110920b6598cSGeorge Rimar } 111020b6598cSGeorge Rimar 111195769b4aSRui Ueyama void ScriptParser::readVersion() { 111295769b4aSRui Ueyama expect("{"); 111395769b4aSRui Ueyama readVersionScriptCommand(); 111495769b4aSRui Ueyama expect("}"); 111595769b4aSRui Ueyama } 111695769b4aSRui Ueyama 111720b6598cSGeorge Rimar void ScriptParser::readLinkerScript() { 1118f7c5fbb1SRui Ueyama while (!atEOF()) { 1119f7c5fbb1SRui Ueyama StringRef Tok = next(); 1120a27eeccaSRui Ueyama if (Tok == ";") 1121a27eeccaSRui Ueyama continue; 1122a27eeccaSRui Ueyama 112320d03194SEugene Leviant if (Tok == "ASSERT") { 112420d03194SEugene Leviant Opt.Commands.emplace_back(new AssertCommand(readAssert())); 112520d03194SEugene Leviant } else if (Tok == "ENTRY") { 1126a27eeccaSRui Ueyama readEntry(); 1127a27eeccaSRui Ueyama } else if (Tok == "EXTERN") { 1128a27eeccaSRui Ueyama readExtern(); 1129a27eeccaSRui Ueyama } else if (Tok == "GROUP" || Tok == "INPUT") { 1130a27eeccaSRui Ueyama readGroup(); 1131a27eeccaSRui Ueyama } else if (Tok == "INCLUDE") { 1132a27eeccaSRui Ueyama readInclude(); 1133b889744eSMeador Inge } else if (Tok == "MEMORY") { 1134b889744eSMeador Inge readMemory(); 1135a27eeccaSRui Ueyama } else if (Tok == "OUTPUT") { 1136a27eeccaSRui Ueyama readOutput(); 1137a27eeccaSRui Ueyama } else if (Tok == "OUTPUT_ARCH") { 1138a27eeccaSRui Ueyama readOutputArch(); 1139a27eeccaSRui Ueyama } else if (Tok == "OUTPUT_FORMAT") { 1140a27eeccaSRui Ueyama readOutputFormat(); 1141a27eeccaSRui Ueyama } else if (Tok == "PHDRS") { 1142a27eeccaSRui Ueyama readPhdrs(); 1143a27eeccaSRui Ueyama } else if (Tok == "SEARCH_DIR") { 1144a27eeccaSRui Ueyama readSearchDir(); 1145a27eeccaSRui Ueyama } else if (Tok == "SECTIONS") { 1146a27eeccaSRui Ueyama readSections(); 1147a27eeccaSRui Ueyama } else if (Tok == "VERSION") { 1148a27eeccaSRui Ueyama readVersion(); 1149c96da110SRafael Espindola } else if (SymbolAssignment *Cmd = readProvideOrAssignment(Tok)) { 11500df80befSPetr Hosek Opt.Commands.emplace_back(Cmd); 1151e5d3ca50SPetr Hosek } else { 11525761042dSGeorge Rimar setError("unknown directive: " + Tok); 1153f7c5fbb1SRui Ueyama } 1154f7c5fbb1SRui Ueyama } 1155e5d3ca50SPetr Hosek } 1156f7c5fbb1SRui Ueyama 1157717677afSRui Ueyama void ScriptParser::addFile(StringRef S) { 115816b0cc9eSSimon Atanasyan if (IsUnderSysroot && S.startswith("/")) { 11595af1687fSJustin Bogner SmallString<128> PathData; 11605af1687fSJustin Bogner StringRef Path = (Config->Sysroot + S).toStringRef(PathData); 116116b0cc9eSSimon Atanasyan if (sys::fs::exists(Path)) { 11625af1687fSJustin Bogner Driver->addFile(Saver.save(Path)); 116316b0cc9eSSimon Atanasyan return; 116416b0cc9eSSimon Atanasyan } 116516b0cc9eSSimon Atanasyan } 116616b0cc9eSSimon Atanasyan 1167f03f3cc1SRui Ueyama if (sys::path::is_absolute(S)) { 116852a1509eSRui Ueyama Driver->addFile(S); 116952a1509eSRui Ueyama } else if (S.startswith("=")) { 117052a1509eSRui Ueyama if (Config->Sysroot.empty()) 117152a1509eSRui Ueyama Driver->addFile(S.substr(1)); 117252a1509eSRui Ueyama else 117352a1509eSRui Ueyama Driver->addFile(Saver.save(Config->Sysroot + "/" + S.substr(1))); 117452a1509eSRui Ueyama } else if (S.startswith("-l")) { 117521eecb4fSRui Ueyama Driver->addLibrary(S.substr(2)); 1176a1b8fc3bSSimon Atanasyan } else if (sys::fs::exists(S)) { 1177a1b8fc3bSSimon Atanasyan Driver->addFile(S); 117852a1509eSRui Ueyama } else { 1179061f9286SRui Ueyama if (Optional<std::string> Path = findFromSearchPaths(S)) 1180061f9286SRui Ueyama Driver->addFile(Saver.save(*Path)); 1181025d59b1SRui Ueyama else 1182061f9286SRui Ueyama setError("unable to find " + S); 118352a1509eSRui Ueyama } 118452a1509eSRui Ueyama } 118552a1509eSRui Ueyama 1186717677afSRui Ueyama void ScriptParser::readAsNeeded() { 1187f7c5fbb1SRui Ueyama expect("("); 118835da9b6eSRui Ueyama bool Orig = Config->AsNeeded; 118935da9b6eSRui Ueyama Config->AsNeeded = true; 119083043f23SRui Ueyama while (!Error && !consume(")")) 1191cd574a5eSGeorge Rimar addFile(unquote(next())); 119235da9b6eSRui Ueyama Config->AsNeeded = Orig; 1193f7c5fbb1SRui Ueyama } 1194f7c5fbb1SRui Ueyama 1195717677afSRui Ueyama void ScriptParser::readEntry() { 119690c5099eSDenis Protivensky // -e <symbol> takes predecence over ENTRY(<symbol>). 119790c5099eSDenis Protivensky expect("("); 119890c5099eSDenis Protivensky StringRef Tok = next(); 119990c5099eSDenis Protivensky if (Config->Entry.empty()) 120090c5099eSDenis Protivensky Config->Entry = Tok; 120190c5099eSDenis Protivensky expect(")"); 120290c5099eSDenis Protivensky } 120390c5099eSDenis Protivensky 1204717677afSRui Ueyama void ScriptParser::readExtern() { 120583f406cfSGeorge Rimar expect("("); 120683043f23SRui Ueyama while (!Error && !consume(")")) 1207a2acc931SRui Ueyama Config->Undefined.push_back(next()); 120883f406cfSGeorge Rimar } 120983f406cfSGeorge Rimar 1210717677afSRui Ueyama void ScriptParser::readGroup() { 1211f7c5fbb1SRui Ueyama expect("("); 121283043f23SRui Ueyama while (!Error && !consume(")")) { 1213f7c5fbb1SRui Ueyama StringRef Tok = next(); 1214a2acc931SRui Ueyama if (Tok == "AS_NEEDED") 1215f7c5fbb1SRui Ueyama readAsNeeded(); 1216a2acc931SRui Ueyama else 1217cd574a5eSGeorge Rimar addFile(unquote(Tok)); 1218f7c5fbb1SRui Ueyama } 1219f7c5fbb1SRui Ueyama } 1220f7c5fbb1SRui Ueyama 1221717677afSRui Ueyama void ScriptParser::readInclude() { 1222d4500653SGeorge Rimar StringRef Tok = unquote(next()); 1223ec1c75e0SRui Ueyama 1224d4500653SGeorge Rimar // https://sourceware.org/binutils/docs/ld/File-Commands.html: 1225d4500653SGeorge Rimar // The file will be searched for in the current directory, and in any 1226d4500653SGeorge Rimar // directory specified with the -L option. 1227ec1c75e0SRui Ueyama if (sys::fs::exists(Tok)) { 1228ec1c75e0SRui Ueyama if (Optional<MemoryBufferRef> MB = readFile(Tok)) 1229ec1c75e0SRui Ueyama tokenize(*MB); 1230025d59b1SRui Ueyama return; 1231025d59b1SRui Ueyama } 1232ec1c75e0SRui Ueyama if (Optional<std::string> Path = findFromSearchPaths(Tok)) { 1233ec1c75e0SRui Ueyama if (Optional<MemoryBufferRef> MB = readFile(*Path)) 1234ec1c75e0SRui Ueyama tokenize(*MB); 1235ec1c75e0SRui Ueyama return; 1236ec1c75e0SRui Ueyama } 1237ec1c75e0SRui Ueyama setError("cannot open " + Tok); 123831aa1f83SRui Ueyama } 123931aa1f83SRui Ueyama 1240717677afSRui Ueyama void ScriptParser::readOutput() { 1241ee59282bSRui Ueyama // -o <file> takes predecence over OUTPUT(<file>). 1242ee59282bSRui Ueyama expect("("); 1243ee59282bSRui Ueyama StringRef Tok = next(); 1244ee59282bSRui Ueyama if (Config->OutputFile.empty()) 1245cd574a5eSGeorge Rimar Config->OutputFile = unquote(Tok); 1246ee59282bSRui Ueyama expect(")"); 1247ee59282bSRui Ueyama } 1248ee59282bSRui Ueyama 1249717677afSRui Ueyama void ScriptParser::readOutputArch() { 1250*4e01c3e8SGeorge Rimar // OUTPUT_ARCH is ignored for now. 12519159ce93SDavide Italiano expect("("); 1252*4e01c3e8SGeorge Rimar while (!Error && !consume(")")) 12535424e7c7SJustin Bogner skip(); 12549159ce93SDavide Italiano } 12559159ce93SDavide Italiano 1256717677afSRui Ueyama void ScriptParser::readOutputFormat() { 1257f7c5fbb1SRui Ueyama // Error checking only for now. 1258f7c5fbb1SRui Ueyama expect("("); 12595424e7c7SJustin Bogner skip(); 12606836c618SDavide Italiano StringRef Tok = next(); 12616836c618SDavide Italiano if (Tok == ")") 12626836c618SDavide Italiano return; 1263025d59b1SRui Ueyama if (Tok != ",") { 12645761042dSGeorge Rimar setError("unexpected token: " + Tok); 1265025d59b1SRui Ueyama return; 1266025d59b1SRui Ueyama } 12675424e7c7SJustin Bogner skip(); 12686836c618SDavide Italiano expect(","); 12695424e7c7SJustin Bogner skip(); 1270f7c5fbb1SRui Ueyama expect(")"); 1271f7c5fbb1SRui Ueyama } 1272f7c5fbb1SRui Ueyama 1273bbe38602SEugene Leviant void ScriptParser::readPhdrs() { 1274bbe38602SEugene Leviant expect("{"); 127583043f23SRui Ueyama while (!Error && !consume("}")) { 1276bbe38602SEugene Leviant StringRef Tok = next(); 127756b21c86SEugene Leviant Opt.PhdrsCommands.push_back( 127856b21c86SEugene Leviant {Tok, PT_NULL, false, false, UINT_MAX, nullptr}); 1279bbe38602SEugene Leviant PhdrsCommand &PhdrCmd = Opt.PhdrsCommands.back(); 1280bbe38602SEugene Leviant 1281bbe38602SEugene Leviant PhdrCmd.Type = readPhdrType(); 1282bbe38602SEugene Leviant do { 1283bbe38602SEugene Leviant Tok = next(); 1284bbe38602SEugene Leviant if (Tok == ";") 1285bbe38602SEugene Leviant break; 1286bbe38602SEugene Leviant if (Tok == "FILEHDR") 1287bbe38602SEugene Leviant PhdrCmd.HasFilehdr = true; 1288bbe38602SEugene Leviant else if (Tok == "PHDRS") 1289bbe38602SEugene Leviant PhdrCmd.HasPhdrs = true; 129056b21c86SEugene Leviant else if (Tok == "AT") 129156b21c86SEugene Leviant PhdrCmd.LMAExpr = readParenExpr(); 1292865bf863SEugene Leviant else if (Tok == "FLAGS") { 1293865bf863SEugene Leviant expect("("); 1294eb685cd7SRafael Espindola // Passing 0 for the value of dot is a bit of a hack. It means that 1295eb685cd7SRafael Espindola // we accept expressions like ".|1". 1296eb685cd7SRafael Espindola PhdrCmd.Flags = readExpr()(0); 1297865bf863SEugene Leviant expect(")"); 1298865bf863SEugene Leviant } else 1299bbe38602SEugene Leviant setError("unexpected header attribute: " + Tok); 1300bbe38602SEugene Leviant } while (!Error); 1301bbe38602SEugene Leviant } 1302bbe38602SEugene Leviant } 1303bbe38602SEugene Leviant 1304717677afSRui Ueyama void ScriptParser::readSearchDir() { 130568a39a65SDavide Italiano expect("("); 130686c5fb82SRui Ueyama StringRef Tok = next(); 13076c7ad13fSRui Ueyama if (!Config->Nostdlib) 1308cd574a5eSGeorge Rimar Config->SearchPaths.push_back(unquote(Tok)); 130968a39a65SDavide Italiano expect(")"); 131068a39a65SDavide Italiano } 131168a39a65SDavide Italiano 1312717677afSRui Ueyama void ScriptParser::readSections() { 1313e05336ffSEugene Leviant Opt.HasSections = true; 131418a30962SGeorge Rimar // -no-rosegment is used to avoid placing read only non-executable sections in 131518a30962SGeorge Rimar // their own segment. We do the same if SECTIONS command is present in linker 131618a30962SGeorge Rimar // script. See comment for computeFlags(). 131718a30962SGeorge Rimar Config->SingleRoRx = true; 131818a30962SGeorge Rimar 13198e3b38abSDenis Protivensky expect("{"); 132083043f23SRui Ueyama while (!Error && !consume("}")) { 1321113cdec9SRui Ueyama StringRef Tok = next(); 1322c96da110SRafael Espindola BaseCommand *Cmd = readProvideOrAssignment(Tok); 1323ceabe80eSEugene Leviant if (!Cmd) { 1324ceabe80eSEugene Leviant if (Tok == "ASSERT") 1325eefa758eSGeorge Rimar Cmd = new AssertCommand(readAssert()); 1326ceabe80eSEugene Leviant else 132710416564SRui Ueyama Cmd = readOutputSectionDescription(Tok); 13288e3b38abSDenis Protivensky } 132910416564SRui Ueyama Opt.Commands.emplace_back(Cmd); 1330652852c5SGeorge Rimar } 1331708019c4SRui Ueyama } 13328e3b38abSDenis Protivensky 1333708019c4SRui Ueyama static int precedence(StringRef Op) { 1334708019c4SRui Ueyama return StringSwitch<int>(Op) 13350120e3f2SRui Ueyama .Cases("*", "/", 5) 13360120e3f2SRui Ueyama .Cases("+", "-", 4) 13370120e3f2SRui Ueyama .Cases("<<", ">>", 3) 13389c4ac5f2SRui Ueyama .Cases("<", "<=", ">", ">=", "==", "!=", 2) 13390120e3f2SRui Ueyama .Cases("&", "|", 1) 1340708019c4SRui Ueyama .Default(-1); 1341708019c4SRui Ueyama } 1342708019c4SRui Ueyama 1343db688454SEugene Leviant StringMatcher ScriptParser::readFilePatterns() { 134410416564SRui Ueyama std::vector<StringRef> V; 134583043f23SRui Ueyama while (!Error && !consume(")")) 134610416564SRui Ueyama V.push_back(next()); 1347f91282e1SRui Ueyama return StringMatcher(V); 13480702c4e8SGeorge Rimar } 13490702c4e8SGeorge Rimar 1350be394db3SGeorge Rimar SortSectionPolicy ScriptParser::readSortKind() { 135183043f23SRui Ueyama if (consume("SORT") || consume("SORT_BY_NAME")) 1352be394db3SGeorge Rimar return SortSectionPolicy::Name; 135383043f23SRui Ueyama if (consume("SORT_BY_ALIGNMENT")) 1354be394db3SGeorge Rimar return SortSectionPolicy::Alignment; 135583043f23SRui Ueyama if (consume("SORT_BY_INIT_PRIORITY")) 1356be394db3SGeorge Rimar return SortSectionPolicy::Priority; 135783043f23SRui Ueyama if (consume("SORT_NONE")) 1358be394db3SGeorge Rimar return SortSectionPolicy::None; 1359b2a0abdfSRui Ueyama return SortSectionPolicy::Default; 1360be394db3SGeorge Rimar } 1361be394db3SGeorge Rimar 1362395281cfSGeorge Rimar // Method reads a list of sequence of excluded files and section globs given in 1363395281cfSGeorge Rimar // a following form: ((EXCLUDE_FILE(file_pattern+))? section_pattern+)+ 1364395281cfSGeorge Rimar // Example: *(.foo.1 EXCLUDE_FILE (*a.o) .foo.2 EXCLUDE_FILE (*b.o) .foo.3) 1365af03be19SGeorge Rimar // The semantics of that is next: 1366af03be19SGeorge Rimar // * Include .foo.1 from every file. 1367af03be19SGeorge Rimar // * Include .foo.2 from every file but a.o 1368af03be19SGeorge Rimar // * Include .foo.3 from every file but b.o 136907171f21SGeorge Rimar std::vector<SectionPattern> ScriptParser::readInputSectionsList() { 137007171f21SGeorge Rimar std::vector<SectionPattern> Ret; 1371601e9898SGeorge Rimar while (!Error && peek() != ")") { 1372f91282e1SRui Ueyama StringMatcher ExcludeFilePat; 137383043f23SRui Ueyama if (consume("EXCLUDE_FILE")) { 1374395281cfSGeorge Rimar expect("("); 1375f91282e1SRui Ueyama ExcludeFilePat = readFilePatterns(); 1376395281cfSGeorge Rimar } 1377395281cfSGeorge Rimar 1378601e9898SGeorge Rimar std::vector<StringRef> V; 1379601e9898SGeorge Rimar while (!Error && peek() != ")" && peek() != "EXCLUDE_FILE") 1380395281cfSGeorge Rimar V.push_back(next()); 1381601e9898SGeorge Rimar 1382601e9898SGeorge Rimar if (!V.empty()) 1383f91282e1SRui Ueyama Ret.push_back({std::move(ExcludeFilePat), StringMatcher(V)}); 1384601e9898SGeorge Rimar else 1385601e9898SGeorge Rimar setError("section pattern is expected"); 1386395281cfSGeorge Rimar } 138707171f21SGeorge Rimar return Ret; 1388395281cfSGeorge Rimar } 1389395281cfSGeorge Rimar 1390f8f6f1e7SRui Ueyama // Reads contents of "SECTIONS" directive. That directive contains a 1391f8f6f1e7SRui Ueyama // list of glob patterns for input sections. The grammar is as follows. 1392f8f6f1e7SRui Ueyama // 1393f8f6f1e7SRui Ueyama // <patterns> ::= <section-list> 1394f8f6f1e7SRui Ueyama // | <sort> "(" <section-list> ")" 1395f8f6f1e7SRui Ueyama // | <sort> "(" <sort> "(" <section-list> ")" ")" 1396f8f6f1e7SRui Ueyama // 1397f8f6f1e7SRui Ueyama // <sort> ::= "SORT" | "SORT_BY_NAME" | "SORT_BY_ALIGNMENT" 1398f8f6f1e7SRui Ueyama // | "SORT_BY_INIT_PRIORITY" | "SORT_NONE" 1399f8f6f1e7SRui Ueyama // 1400f8f6f1e7SRui Ueyama // <section-list> is parsed by readInputSectionsList(). 1401a2496cbeSGeorge Rimar InputSectionDescription * 1402a2496cbeSGeorge Rimar ScriptParser::readInputSectionRules(StringRef FilePattern) { 1403c91930a1SGeorge Rimar auto *Cmd = new InputSectionDescription(FilePattern); 14040ed42b0cSDavide Italiano expect("("); 1405f373dd76SRui Ueyama while (!Error && !consume(")")) { 140607171f21SGeorge Rimar SortSectionPolicy Outer = readSortKind(); 140707171f21SGeorge Rimar SortSectionPolicy Inner = SortSectionPolicy::Default; 140807171f21SGeorge Rimar std::vector<SectionPattern> V; 140907171f21SGeorge Rimar if (Outer != SortSectionPolicy::Default) { 14100702c4e8SGeorge Rimar expect("("); 141107171f21SGeorge Rimar Inner = readSortKind(); 141207171f21SGeorge Rimar if (Inner != SortSectionPolicy::Default) { 1413350ece4eSGeorge Rimar expect("("); 141407171f21SGeorge Rimar V = readInputSectionsList(); 14150702c4e8SGeorge Rimar expect(")"); 1416350ece4eSGeorge Rimar } else { 141707171f21SGeorge Rimar V = readInputSectionsList(); 1418350ece4eSGeorge Rimar } 1419350ece4eSGeorge Rimar expect(")"); 142007171f21SGeorge Rimar } else { 142107171f21SGeorge Rimar V = readInputSectionsList(); 14220659800eSGeorge Rimar } 14230702c4e8SGeorge Rimar 142407171f21SGeorge Rimar for (SectionPattern &Pat : V) { 142507171f21SGeorge Rimar Pat.SortInner = Inner; 142607171f21SGeorge Rimar Pat.SortOuter = Outer; 142707171f21SGeorge Rimar } 142807171f21SGeorge Rimar 142907171f21SGeorge Rimar std::move(V.begin(), V.end(), std::back_inserter(Cmd->SectionPatterns)); 143007171f21SGeorge Rimar } 143110416564SRui Ueyama return Cmd; 14320659800eSGeorge Rimar } 14330659800eSGeorge Rimar 1434a2496cbeSGeorge Rimar InputSectionDescription * 1435a2496cbeSGeorge Rimar ScriptParser::readInputSectionDescription(StringRef Tok) { 14360659800eSGeorge Rimar // Input section wildcard can be surrounded by KEEP. 14370659800eSGeorge Rimar // https://sourceware.org/binutils/docs/ld/Input-Section-Keep.html#Input-Section-Keep 1438a2496cbeSGeorge Rimar if (Tok == "KEEP") { 1439e7282797SDavide Italiano expect("("); 1440a2496cbeSGeorge Rimar StringRef FilePattern = next(); 1441a2496cbeSGeorge Rimar InputSectionDescription *Cmd = readInputSectionRules(FilePattern); 14420ed42b0cSDavide Italiano expect(")"); 1443cf43f179SEugene Leviant Opt.KeptSections.push_back(Cmd); 144410416564SRui Ueyama return Cmd; 144510416564SRui Ueyama } 1446a2496cbeSGeorge Rimar return readInputSectionRules(Tok); 14470659800eSGeorge Rimar } 14480659800eSGeorge Rimar 144903fc010eSGeorge Rimar void ScriptParser::readSort() { 145003fc010eSGeorge Rimar expect("("); 145103fc010eSGeorge Rimar expect("CONSTRUCTORS"); 145203fc010eSGeorge Rimar expect(")"); 145303fc010eSGeorge Rimar } 145403fc010eSGeorge Rimar 1455eefa758eSGeorge Rimar Expr ScriptParser::readAssert() { 1456eefa758eSGeorge Rimar expect("("); 1457eefa758eSGeorge Rimar Expr E = readExpr(); 1458eefa758eSGeorge Rimar expect(","); 1459cd574a5eSGeorge Rimar StringRef Msg = unquote(next()); 1460eefa758eSGeorge Rimar expect(")"); 1461eefa758eSGeorge Rimar return [=](uint64_t Dot) { 1462eefa758eSGeorge Rimar uint64_t V = E(Dot); 1463eefa758eSGeorge Rimar if (!V) 1464eefa758eSGeorge Rimar error(Msg); 1465eefa758eSGeorge Rimar return V; 1466eefa758eSGeorge Rimar }; 1467eefa758eSGeorge Rimar } 1468eefa758eSGeorge Rimar 146925150e8bSRui Ueyama // Reads a FILL(expr) command. We handle the FILL command as an 147025150e8bSRui Ueyama // alias for =fillexp section attribute, which is different from 147125150e8bSRui Ueyama // what GNU linkers do. 147225150e8bSRui Ueyama // https://sourceware.org/binutils/docs/ld/Output-Section-Data.html 147316068aebSRui Ueyama uint32_t ScriptParser::readFill() { 1474ff1f29e0SGeorge Rimar expect("("); 147516068aebSRui Ueyama uint32_t V = readOutputSectionFiller(next()); 1476ff1f29e0SGeorge Rimar expect(")"); 1477ff1f29e0SGeorge Rimar expect(";"); 1478ff1f29e0SGeorge Rimar return V; 1479ff1f29e0SGeorge Rimar } 1480ff1f29e0SGeorge Rimar 148110416564SRui Ueyama OutputSectionCommand * 148210416564SRui Ueyama ScriptParser::readOutputSectionDescription(StringRef OutSec) { 1483076fe157SGeorge Rimar OutputSectionCommand *Cmd = new OutputSectionCommand(OutSec); 14842a942c4bSEugene Leviant Cmd->Location = getCurrentLocation(); 148558e5c4dcSGeorge Rimar 148658e5c4dcSGeorge Rimar // Read an address expression. 148758e5c4dcSGeorge Rimar // https://sourceware.org/binutils/docs/ld/Output-Section-Address.html#Output-Section-Address 148858e5c4dcSGeorge Rimar if (peek() != ":") 148958e5c4dcSGeorge Rimar Cmd->AddrExpr = readExpr(); 149058e5c4dcSGeorge Rimar 14918e3b38abSDenis Protivensky expect(":"); 1492246f681eSDavide Italiano 149383043f23SRui Ueyama if (consume("AT")) 1494b71d6f7aSEugene Leviant Cmd->LMAExpr = readParenExpr(); 149583043f23SRui Ueyama if (consume("ALIGN")) 14966ad7dfccSRui Ueyama Cmd->AlignExpr = readParenExpr(); 149783043f23SRui Ueyama if (consume("SUBALIGN")) 1498db24d9c3SGeorge Rimar Cmd->SubalignExpr = readParenExpr(); 1499630c6179SGeorge Rimar 1500246f681eSDavide Italiano // Parse constraints. 150183043f23SRui Ueyama if (consume("ONLY_IF_RO")) 1502efc4066bSRui Ueyama Cmd->Constraint = ConstraintKind::ReadOnly; 150383043f23SRui Ueyama if (consume("ONLY_IF_RW")) 1504efc4066bSRui Ueyama Cmd->Constraint = ConstraintKind::ReadWrite; 15058e3b38abSDenis Protivensky expect("{"); 15068ec77e64SRui Ueyama 150783043f23SRui Ueyama while (!Error && !consume("}")) { 1508ceabe80eSEugene Leviant StringRef Tok = next(); 15092fe07923SGeorge Rimar if (Tok == ";") { 151069750755SGeorge Rimar // Empty commands are allowed. Do nothing here. 15112fe07923SGeorge Rimar } else if (SymbolAssignment *Assignment = readProvideOrAssignment(Tok)) { 1512ceabe80eSEugene Leviant Cmd->Commands.emplace_back(Assignment); 1513b2d99d6aSMeador Inge } else if (BytesDataCommand *Data = readBytesDataCommand(Tok)) { 1514e38cbab5SGeorge Rimar Cmd->Commands.emplace_back(Data); 1515b2d99d6aSMeador Inge } else if (Tok == "ASSERT") { 1516b2d99d6aSMeador Inge Cmd->Commands.emplace_back(new AssertCommand(readAssert())); 1517b2d99d6aSMeador Inge expect(";"); 15188e2eca22SGeorge Rimar } else if (Tok == "CONSTRUCTORS") { 15198e2eca22SGeorge Rimar // CONSTRUCTORS is a keyword to make the linker recognize C++ ctors/dtors 15208e2eca22SGeorge Rimar // by name. This is for very old file formats such as ECOFF/XCOFF. 15218e2eca22SGeorge Rimar // For ELF, we should ignore. 1522b2d99d6aSMeador Inge } else if (Tok == "FILL") { 1523ff1f29e0SGeorge Rimar Cmd->Filler = readFill(); 1524b2d99d6aSMeador Inge } else if (Tok == "SORT") { 152503fc010eSGeorge Rimar readSort(); 1526b2d99d6aSMeador Inge } else if (peek() == "(") { 1527a2496cbeSGeorge Rimar Cmd->Commands.emplace_back(readInputSectionDescription(Tok)); 1528b2d99d6aSMeador Inge } else { 1529ceabe80eSEugene Leviant setError("unknown command " + Tok); 15308e3b38abSDenis Protivensky } 1531b2d99d6aSMeador Inge } 1532b889744eSMeador Inge 1533b889744eSMeador Inge if (consume(">")) 1534b889744eSMeador Inge Cmd->MemoryRegionName = next(); 1535b889744eSMeador Inge 1536076fe157SGeorge Rimar Cmd->Phdrs = readOutputSectionPhdrs(); 15374ebc5620SGeorge Rimar 153883043f23SRui Ueyama if (consume("=")) 15394ebc5620SGeorge Rimar Cmd->Filler = readOutputSectionFiller(next()); 15404ebc5620SGeorge Rimar else if (peek().startswith("=")) 1541ff1f29e0SGeorge Rimar Cmd->Filler = readOutputSectionFiller(next().drop_front()); 15424ebc5620SGeorge Rimar 15437185a1acSGeorge Rimar // Consume optional comma following output section command. 15447185a1acSGeorge Rimar consume(","); 15457185a1acSGeorge Rimar 154610416564SRui Ueyama return Cmd; 1547f71caa2bSRui Ueyama } 15488ec77e64SRui Ueyama 15492c8f1f04SRui Ueyama // Read "=<number>" where <number> is an octal/decimal/hexadecimal number. 15502c8f1f04SRui Ueyama // https://sourceware.org/binutils/docs/ld/Output-Section-Fill.html 15512c8f1f04SRui Ueyama // 15522c8f1f04SRui Ueyama // ld.gold is not fully compatible with ld.bfd. ld.bfd handles 15532c8f1f04SRui Ueyama // hexstrings as blobs of arbitrary sizes, while ld.gold handles them 15542c8f1f04SRui Ueyama // as 32-bit big-endian values. We will do the same as ld.gold does 15552c8f1f04SRui Ueyama // because it's simpler than what ld.bfd does. 155616068aebSRui Ueyama uint32_t ScriptParser::readOutputSectionFiller(StringRef Tok) { 1557965827d6SRui Ueyama uint32_t V; 155816068aebSRui Ueyama if (!Tok.getAsInteger(0, V)) 155916068aebSRui Ueyama return V; 1560965827d6SRui Ueyama setError("invalid filler expression: " + Tok); 156116068aebSRui Ueyama return 0; 15628e3b38abSDenis Protivensky } 15638e3b38abSDenis Protivensky 1564a35e39caSPetr Hosek SymbolAssignment *ScriptParser::readProvideHidden(bool Provide, bool Hidden) { 1565a31c91b1SEugene Leviant expect("("); 1566174e0a16SRui Ueyama SymbolAssignment *Cmd = readAssignment(next()); 1567a35e39caSPetr Hosek Cmd->Provide = Provide; 1568174e0a16SRui Ueyama Cmd->Hidden = Hidden; 1569a31c91b1SEugene Leviant expect(")"); 1570a31c91b1SEugene Leviant expect(";"); 157110416564SRui Ueyama return Cmd; 1572eda81a1bSEugene Leviant } 1573eda81a1bSEugene Leviant 1574c96da110SRafael Espindola SymbolAssignment *ScriptParser::readProvideOrAssignment(StringRef Tok) { 1575ceabe80eSEugene Leviant SymbolAssignment *Cmd = nullptr; 1576ceabe80eSEugene Leviant if (peek() == "=" || peek() == "+=") { 1577ceabe80eSEugene Leviant Cmd = readAssignment(Tok); 1578ceabe80eSEugene Leviant expect(";"); 1579ceabe80eSEugene Leviant } else if (Tok == "PROVIDE") { 1580a35e39caSPetr Hosek Cmd = readProvideHidden(true, false); 1581a35e39caSPetr Hosek } else if (Tok == "HIDDEN") { 1582a35e39caSPetr Hosek Cmd = readProvideHidden(false, true); 1583ceabe80eSEugene Leviant } else if (Tok == "PROVIDE_HIDDEN") { 1584a35e39caSPetr Hosek Cmd = readProvideHidden(true, true); 1585ceabe80eSEugene Leviant } 1586ceabe80eSEugene Leviant return Cmd; 1587ceabe80eSEugene Leviant } 1588ceabe80eSEugene Leviant 1589f6aeed36SEugene Leviant static uint64_t getSymbolValue(const Twine &Loc, StringRef S, uint64_t Dot) { 159030835ea4SGeorge Rimar if (S == ".") 159130835ea4SGeorge Rimar return Dot; 1592f6aeed36SEugene Leviant return ScriptBase->getSymbolValue(Loc, S); 1593e32a3598SGeorge Rimar } 1594e32a3598SGeorge Rimar 15952f831dcaSRafael Espindola static bool isAbsolute(StringRef S) { 15962f831dcaSRafael Espindola if (S == ".") 15972f831dcaSRafael Espindola return false; 15982f831dcaSRafael Espindola return ScriptBase->isAbsolute(S); 15992f831dcaSRafael Espindola } 16002f831dcaSRafael Espindola 160130835ea4SGeorge Rimar SymbolAssignment *ScriptParser::readAssignment(StringRef Name) { 160230835ea4SGeorge Rimar StringRef Op = next(); 1603db741e72SEugene Leviant Expr E; 160430835ea4SGeorge Rimar assert(Op == "=" || Op == "+="); 160583043f23SRui Ueyama if (consume("ABSOLUTE")) { 16067c1381a0SRui Ueyama // The RHS may be something like "ABSOLUTE(.) & 0xff". 16077c1381a0SRui Ueyama // Call readExpr1 to read the whole expression. 16087c1381a0SRui Ueyama E = readExpr1(readParenExpr(), 0); 1609009d1742SRui Ueyama E.IsAbsolute = [] { return true; }; 1610db741e72SEugene Leviant } else { 1611db741e72SEugene Leviant E = readExpr(); 1612db741e72SEugene Leviant } 1613f6aeed36SEugene Leviant if (Op == "+=") { 1614f6aeed36SEugene Leviant std::string Loc = getCurrentLocation(); 1615f6aeed36SEugene Leviant E = [=](uint64_t Dot) { 1616f6aeed36SEugene Leviant return getSymbolValue(Loc, Name, Dot) + E(Dot); 1617f6aeed36SEugene Leviant }; 1618f6aeed36SEugene Leviant } 1619f661393aSRafael Espindola return new SymbolAssignment(Name, E); 162030835ea4SGeorge Rimar } 162130835ea4SGeorge Rimar 162230835ea4SGeorge Rimar // This is an operator-precedence parser to parse a linker 162330835ea4SGeorge Rimar // script expression. 162430835ea4SGeorge Rimar Expr ScriptParser::readExpr() { return readExpr1(readPrimary(), 0); } 162530835ea4SGeorge Rimar 162636c1cd23SRui Ueyama static Expr combine(StringRef Op, Expr L, Expr R) { 1627cc4d3e57SGeorge Rimar auto IsAbs = [=] { return L.IsAbsolute() && R.IsAbsolute(); }; 1628cc4d3e57SGeorge Rimar auto GetOutSec = [=] { 1629cc4d3e57SGeorge Rimar const OutputSectionBase *S = L.Section(); 1630cc4d3e57SGeorge Rimar return S ? S : R.Section(); 1631cc4d3e57SGeorge Rimar }; 1632cc4d3e57SGeorge Rimar 163336c1cd23SRui Ueyama if (Op == "*") 163436c1cd23SRui Ueyama return [=](uint64_t Dot) { return L(Dot) * R(Dot); }; 163536c1cd23SRui Ueyama if (Op == "/") { 163636c1cd23SRui Ueyama return [=](uint64_t Dot) -> uint64_t { 163736c1cd23SRui Ueyama uint64_t RHS = R(Dot); 163836c1cd23SRui Ueyama if (RHS == 0) { 163936c1cd23SRui Ueyama error("division by zero"); 164036c1cd23SRui Ueyama return 0; 164136c1cd23SRui Ueyama } 164236c1cd23SRui Ueyama return L(Dot) / RHS; 164336c1cd23SRui Ueyama }; 164436c1cd23SRui Ueyama } 164536c1cd23SRui Ueyama if (Op == "+") 1646cc4d3e57SGeorge Rimar return {[=](uint64_t Dot) { return L(Dot) + R(Dot); }, IsAbs, GetOutSec}; 164736c1cd23SRui Ueyama if (Op == "-") 1648cc4d3e57SGeorge Rimar return {[=](uint64_t Dot) { return L(Dot) - R(Dot); }, IsAbs, GetOutSec}; 1649c8ccd1f1SGeorge Rimar if (Op == "<<") 1650c8ccd1f1SGeorge Rimar return [=](uint64_t Dot) { return L(Dot) << R(Dot); }; 1651c8ccd1f1SGeorge Rimar if (Op == ">>") 1652c8ccd1f1SGeorge Rimar return [=](uint64_t Dot) { return L(Dot) >> R(Dot); }; 165336c1cd23SRui Ueyama if (Op == "<") 165436c1cd23SRui Ueyama return [=](uint64_t Dot) { return L(Dot) < R(Dot); }; 165536c1cd23SRui Ueyama if (Op == ">") 165636c1cd23SRui Ueyama return [=](uint64_t Dot) { return L(Dot) > R(Dot); }; 165736c1cd23SRui Ueyama if (Op == ">=") 165836c1cd23SRui Ueyama return [=](uint64_t Dot) { return L(Dot) >= R(Dot); }; 165936c1cd23SRui Ueyama if (Op == "<=") 166036c1cd23SRui Ueyama return [=](uint64_t Dot) { return L(Dot) <= R(Dot); }; 166136c1cd23SRui Ueyama if (Op == "==") 166236c1cd23SRui Ueyama return [=](uint64_t Dot) { return L(Dot) == R(Dot); }; 166336c1cd23SRui Ueyama if (Op == "!=") 166436c1cd23SRui Ueyama return [=](uint64_t Dot) { return L(Dot) != R(Dot); }; 166536c1cd23SRui Ueyama if (Op == "&") 166636c1cd23SRui Ueyama return [=](uint64_t Dot) { return L(Dot) & R(Dot); }; 1667cc3dd629SRafael Espindola if (Op == "|") 1668cc3dd629SRafael Espindola return [=](uint64_t Dot) { return L(Dot) | R(Dot); }; 166936c1cd23SRui Ueyama llvm_unreachable("invalid operator"); 167036c1cd23SRui Ueyama } 167136c1cd23SRui Ueyama 1672708019c4SRui Ueyama // This is a part of the operator-precedence parser. This function 1673708019c4SRui Ueyama // assumes that the remaining token stream starts with an operator. 1674708019c4SRui Ueyama Expr ScriptParser::readExpr1(Expr Lhs, int MinPrec) { 1675708019c4SRui Ueyama while (!atEOF() && !Error) { 1676708019c4SRui Ueyama // Read an operator and an expression. 167746247b85SRui Ueyama if (consume("?")) 1678708019c4SRui Ueyama return readTernary(Lhs); 167946247b85SRui Ueyama StringRef Op1 = peek(); 1680708019c4SRui Ueyama if (precedence(Op1) < MinPrec) 1681a31c91b1SEugene Leviant break; 16825424e7c7SJustin Bogner skip(); 1683708019c4SRui Ueyama Expr Rhs = readPrimary(); 1684708019c4SRui Ueyama 1685708019c4SRui Ueyama // Evaluate the remaining part of the expression first if the 1686708019c4SRui Ueyama // next operator has greater precedence than the previous one. 1687708019c4SRui Ueyama // For example, if we have read "+" and "3", and if the next 1688708019c4SRui Ueyama // operator is "*", then we'll evaluate 3 * ... part first. 1689708019c4SRui Ueyama while (!atEOF()) { 1690708019c4SRui Ueyama StringRef Op2 = peek(); 1691708019c4SRui Ueyama if (precedence(Op2) <= precedence(Op1)) 1692eda81a1bSEugene Leviant break; 1693708019c4SRui Ueyama Rhs = readExpr1(Rhs, precedence(Op2)); 1694eda81a1bSEugene Leviant } 1695708019c4SRui Ueyama 1696708019c4SRui Ueyama Lhs = combine(Op1, Lhs, Rhs); 1697708019c4SRui Ueyama } 1698708019c4SRui Ueyama return Lhs; 1699708019c4SRui Ueyama } 1700708019c4SRui Ueyama 1701708019c4SRui Ueyama uint64_t static getConstant(StringRef S) { 1702e2cc07bcSMichael J. Spencer if (S == "COMMONPAGESIZE") 1703708019c4SRui Ueyama return Target->PageSize; 1704e2cc07bcSMichael J. Spencer if (S == "MAXPAGESIZE") 1705997f8838SPetr Hosek return Config->MaxPageSize; 1706708019c4SRui Ueyama error("unknown constant: " + S); 1707708019c4SRui Ueyama return 0; 1708708019c4SRui Ueyama } 1709708019c4SRui Ueyama 1710626e0b08SRui Ueyama // Parses Tok as an integer. Returns true if successful. 1711626e0b08SRui Ueyama // It recognizes hexadecimal (prefixed with "0x" or suffixed with "H") 1712626e0b08SRui Ueyama // and decimal numbers. Decimal numbers may have "K" (kilo) or 1713626e0b08SRui Ueyama // "M" (mega) prefixes. 17149f2f7ad9SGeorge Rimar static bool readInteger(StringRef Tok, uint64_t &Result) { 171546247b85SRui Ueyama // Negative number 1716eaeafb2bSSimon Atanasyan if (Tok.startswith("-")) { 1717eaeafb2bSSimon Atanasyan if (!readInteger(Tok.substr(1), Result)) 1718eaeafb2bSSimon Atanasyan return false; 1719eaeafb2bSSimon Atanasyan Result = -Result; 1720eaeafb2bSSimon Atanasyan return true; 1721eaeafb2bSSimon Atanasyan } 172246247b85SRui Ueyama 172346247b85SRui Ueyama // Hexadecimal 17249f2f7ad9SGeorge Rimar if (Tok.startswith_lower("0x")) 17259f2f7ad9SGeorge Rimar return !Tok.substr(2).getAsInteger(16, Result); 17269f2f7ad9SGeorge Rimar if (Tok.endswith_lower("H")) 17279f2f7ad9SGeorge Rimar return !Tok.drop_back().getAsInteger(16, Result); 17289f2f7ad9SGeorge Rimar 172946247b85SRui Ueyama // Decimal 17309f2f7ad9SGeorge Rimar int Suffix = 1; 17319f2f7ad9SGeorge Rimar if (Tok.endswith_lower("K")) { 17329f2f7ad9SGeorge Rimar Suffix = 1024; 17339f2f7ad9SGeorge Rimar Tok = Tok.drop_back(); 17349f2f7ad9SGeorge Rimar } else if (Tok.endswith_lower("M")) { 17359f2f7ad9SGeorge Rimar Suffix = 1024 * 1024; 17369f2f7ad9SGeorge Rimar Tok = Tok.drop_back(); 17379f2f7ad9SGeorge Rimar } 17389f2f7ad9SGeorge Rimar if (Tok.getAsInteger(10, Result)) 17399f2f7ad9SGeorge Rimar return false; 17409f2f7ad9SGeorge Rimar Result *= Suffix; 17419f2f7ad9SGeorge Rimar return true; 17429f2f7ad9SGeorge Rimar } 17439f2f7ad9SGeorge Rimar 1744e38cbab5SGeorge Rimar BytesDataCommand *ScriptParser::readBytesDataCommand(StringRef Tok) { 1745e38cbab5SGeorge Rimar int Size = StringSwitch<unsigned>(Tok) 1746e38cbab5SGeorge Rimar .Case("BYTE", 1) 1747e38cbab5SGeorge Rimar .Case("SHORT", 2) 1748e38cbab5SGeorge Rimar .Case("LONG", 4) 1749e38cbab5SGeorge Rimar .Case("QUAD", 8) 1750e38cbab5SGeorge Rimar .Default(-1); 1751e38cbab5SGeorge Rimar if (Size == -1) 1752e38cbab5SGeorge Rimar return nullptr; 1753e38cbab5SGeorge Rimar 175495c7d8d2SMeador Inge return new BytesDataCommand(readParenExpr(), Size); 1755e38cbab5SGeorge Rimar } 1756e38cbab5SGeorge Rimar 1757b71d6f7aSEugene Leviant StringRef ScriptParser::readParenLiteral() { 1758b71d6f7aSEugene Leviant expect("("); 1759b71d6f7aSEugene Leviant StringRef Tok = next(); 1760b71d6f7aSEugene Leviant expect(")"); 1761b71d6f7aSEugene Leviant return Tok; 1762b71d6f7aSEugene Leviant } 1763b71d6f7aSEugene Leviant 1764708019c4SRui Ueyama Expr ScriptParser::readPrimary() { 17656ad7dfccSRui Ueyama if (peek() == "(") 17666ad7dfccSRui Ueyama return readParenExpr(); 1767708019c4SRui Ueyama 17686ad7dfccSRui Ueyama StringRef Tok = next(); 1769b5f1c3ecSRui Ueyama std::string Location = getCurrentLocation(); 1770708019c4SRui Ueyama 1771eaeafb2bSSimon Atanasyan if (Tok == "~") { 1772eaeafb2bSSimon Atanasyan Expr E = readPrimary(); 1773eaeafb2bSSimon Atanasyan return [=](uint64_t Dot) { return ~E(Dot); }; 1774eaeafb2bSSimon Atanasyan } 1775eaeafb2bSSimon Atanasyan if (Tok == "-") { 1776eaeafb2bSSimon Atanasyan Expr E = readPrimary(); 1777eaeafb2bSSimon Atanasyan return [=](uint64_t Dot) { return -E(Dot); }; 1778eaeafb2bSSimon Atanasyan } 1779eaeafb2bSSimon Atanasyan 1780708019c4SRui Ueyama // Built-in functions are parsed here. 1781708019c4SRui Ueyama // https://sourceware.org/binutils/docs/ld/Builtin-Functions.html. 178296659df0SGeorge Rimar if (Tok == "ADDR") { 1783b71d6f7aSEugene Leviant StringRef Name = readParenLiteral(); 1784ed30ce7aSEugene Leviant return {[=](uint64_t Dot) { 1785ed30ce7aSEugene Leviant return ScriptBase->getOutputSection(Location, Name)->Addr; 1786ed30ce7aSEugene Leviant }, 1787009d1742SRui Ueyama [=] { return false; }, 1788ed30ce7aSEugene Leviant [=] { return ScriptBase->getOutputSection(Location, Name); }}; 178996659df0SGeorge Rimar } 1790b71d6f7aSEugene Leviant if (Tok == "LOADADDR") { 1791b71d6f7aSEugene Leviant StringRef Name = readParenLiteral(); 1792afaa9343SEugene Leviant return [=](uint64_t Dot) { 1793ed30ce7aSEugene Leviant return ScriptBase->getOutputSection(Location, Name)->getLMA(); 1794afaa9343SEugene Leviant }; 1795b71d6f7aSEugene Leviant } 1796eefa758eSGeorge Rimar if (Tok == "ASSERT") 1797eefa758eSGeorge Rimar return readAssert(); 1798708019c4SRui Ueyama if (Tok == "ALIGN") { 17995d804dc8SRui Ueyama expect("("); 18005d804dc8SRui Ueyama Expr E = readExpr(); 18015d804dc8SRui Ueyama if (consume(",")) { 18025d804dc8SRui Ueyama Expr E2 = readExpr(); 18035d804dc8SRui Ueyama expect(")"); 18045d804dc8SRui Ueyama return [=](uint64_t Dot) { return alignTo(E(Dot), E2(Dot)); }; 18055d804dc8SRui Ueyama } 18065d804dc8SRui Ueyama expect(")"); 1807708019c4SRui Ueyama return [=](uint64_t Dot) { return alignTo(Dot, E(Dot)); }; 1808708019c4SRui Ueyama } 1809708019c4SRui Ueyama if (Tok == "CONSTANT") { 1810b71d6f7aSEugene Leviant StringRef Name = readParenLiteral(); 1811b0de56b5SRafael Espindola return [=](uint64_t Dot) { return getConstant(Name); }; 1812708019c4SRui Ueyama } 1813f34f45fdSGeorge Rimar if (Tok == "DEFINED") { 18140ee25a69SRui Ueyama StringRef Name = readParenLiteral(); 18150ee25a69SRui Ueyama return [=](uint64_t Dot) { return ScriptBase->isDefined(Name) ? 1 : 0; }; 1816f34f45fdSGeorge Rimar } 181754c145ceSRafael Espindola if (Tok == "SEGMENT_START") { 181854c145ceSRafael Espindola expect("("); 18195424e7c7SJustin Bogner skip(); 182054c145ceSRafael Espindola expect(","); 18218c658bf8SGeorge Rimar Expr E = readExpr(); 182254c145ceSRafael Espindola expect(")"); 18238c658bf8SGeorge Rimar return [=](uint64_t Dot) { return E(Dot); }; 182454c145ceSRafael Espindola } 1825708019c4SRui Ueyama if (Tok == "DATA_SEGMENT_ALIGN") { 1826708019c4SRui Ueyama expect("("); 1827708019c4SRui Ueyama Expr E = readExpr(); 1828708019c4SRui Ueyama expect(","); 1829708019c4SRui Ueyama readExpr(); 1830708019c4SRui Ueyama expect(")"); 1831f7791bb9SRui Ueyama return [=](uint64_t Dot) { return alignTo(Dot, E(Dot)); }; 1832708019c4SRui Ueyama } 1833708019c4SRui Ueyama if (Tok == "DATA_SEGMENT_END") { 1834708019c4SRui Ueyama expect("("); 1835708019c4SRui Ueyama expect("."); 1836708019c4SRui Ueyama expect(")"); 1837708019c4SRui Ueyama return [](uint64_t Dot) { return Dot; }; 1838708019c4SRui Ueyama } 1839276b4e64SGeorge Rimar // GNU linkers implements more complicated logic to handle 1840276b4e64SGeorge Rimar // DATA_SEGMENT_RELRO_END. We instead ignore the arguments and just align to 1841276b4e64SGeorge Rimar // the next page boundary for simplicity. 1842276b4e64SGeorge Rimar if (Tok == "DATA_SEGMENT_RELRO_END") { 1843276b4e64SGeorge Rimar expect("("); 184497bdc722SRafael Espindola readExpr(); 1845276b4e64SGeorge Rimar expect(","); 1846276b4e64SGeorge Rimar readExpr(); 1847276b4e64SGeorge Rimar expect(")"); 1848276b4e64SGeorge Rimar return [](uint64_t Dot) { return alignTo(Dot, Target->PageSize); }; 1849276b4e64SGeorge Rimar } 18509e69450eSGeorge Rimar if (Tok == "SIZEOF") { 1851b71d6f7aSEugene Leviant StringRef Name = readParenLiteral(); 1852edf75e79SRui Ueyama return [=](uint64_t Dot) { return ScriptBase->getOutputSectionSize(Name); }; 18539e69450eSGeorge Rimar } 185436fac7f0SEugene Leviant if (Tok == "ALIGNOF") { 1855b71d6f7aSEugene Leviant StringRef Name = readParenLiteral(); 1856afaa9343SEugene Leviant return [=](uint64_t Dot) { 1857ed30ce7aSEugene Leviant return ScriptBase->getOutputSection(Location, Name)->Addralign; 1858afaa9343SEugene Leviant }; 185936fac7f0SEugene Leviant } 1860e32a3598SGeorge Rimar if (Tok == "SIZEOF_HEADERS") 1861b0de56b5SRafael Espindola return [=](uint64_t Dot) { return ScriptBase->getHeaderSize(); }; 1862708019c4SRui Ueyama 18639f2f7ad9SGeorge Rimar // Tok is a literal number. 18649f2f7ad9SGeorge Rimar uint64_t V; 18659f2f7ad9SGeorge Rimar if (readInteger(Tok, V)) 1866b0de56b5SRafael Espindola return [=](uint64_t Dot) { return V; }; 18679f2f7ad9SGeorge Rimar 18689f2f7ad9SGeorge Rimar // Tok is a symbol name. 186930835ea4SGeorge Rimar if (Tok != "." && !isValidCIdentifier(Tok)) 1870708019c4SRui Ueyama setError("malformed number: " + Tok); 1871f6aeed36SEugene Leviant return {[=](uint64_t Dot) { return getSymbolValue(Location, Tok, Dot); }, 1872009d1742SRui Ueyama [=] { return isAbsolute(Tok); }, 1873009d1742SRui Ueyama [=] { return ScriptBase->getSymbolSection(Tok); }}; 1874a9c5a528SGeorge Rimar } 1875708019c4SRui Ueyama 1876708019c4SRui Ueyama Expr ScriptParser::readTernary(Expr Cond) { 1877708019c4SRui Ueyama Expr L = readExpr(); 1878708019c4SRui Ueyama expect(":"); 1879708019c4SRui Ueyama Expr R = readExpr(); 1880708019c4SRui Ueyama return [=](uint64_t Dot) { return Cond(Dot) ? L(Dot) : R(Dot); }; 1881708019c4SRui Ueyama } 1882708019c4SRui Ueyama 18836ad7dfccSRui Ueyama Expr ScriptParser::readParenExpr() { 18846ad7dfccSRui Ueyama expect("("); 18856ad7dfccSRui Ueyama Expr E = readExpr(); 18866ad7dfccSRui Ueyama expect(")"); 18876ad7dfccSRui Ueyama return E; 18886ad7dfccSRui Ueyama } 18896ad7dfccSRui Ueyama 1890bbe38602SEugene Leviant std::vector<StringRef> ScriptParser::readOutputSectionPhdrs() { 1891bbe38602SEugene Leviant std::vector<StringRef> Phdrs; 1892bbe38602SEugene Leviant while (!Error && peek().startswith(":")) { 1893bbe38602SEugene Leviant StringRef Tok = next(); 1894da841c16SGeorge Rimar Phdrs.push_back((Tok.size() == 1) ? next() : Tok.substr(1)); 1895bbe38602SEugene Leviant } 1896bbe38602SEugene Leviant return Phdrs; 1897bbe38602SEugene Leviant } 1898bbe38602SEugene Leviant 189995dd718cSGeorge Rimar // Read a program header type name. The next token must be a 190095dd718cSGeorge Rimar // name of a program header type or a constant (e.g. "0x3"). 1901bbe38602SEugene Leviant unsigned ScriptParser::readPhdrType() { 1902bbe38602SEugene Leviant StringRef Tok = next(); 190395dd718cSGeorge Rimar uint64_t Val; 190495dd718cSGeorge Rimar if (readInteger(Tok, Val)) 190595dd718cSGeorge Rimar return Val; 190695dd718cSGeorge Rimar 1907b0f6c590SRui Ueyama unsigned Ret = StringSwitch<unsigned>(Tok) 1908b0f6c590SRui Ueyama .Case("PT_NULL", PT_NULL) 1909b0f6c590SRui Ueyama .Case("PT_LOAD", PT_LOAD) 1910b0f6c590SRui Ueyama .Case("PT_DYNAMIC", PT_DYNAMIC) 1911b0f6c590SRui Ueyama .Case("PT_INTERP", PT_INTERP) 1912b0f6c590SRui Ueyama .Case("PT_NOTE", PT_NOTE) 1913b0f6c590SRui Ueyama .Case("PT_SHLIB", PT_SHLIB) 1914b0f6c590SRui Ueyama .Case("PT_PHDR", PT_PHDR) 1915b0f6c590SRui Ueyama .Case("PT_TLS", PT_TLS) 1916b0f6c590SRui Ueyama .Case("PT_GNU_EH_FRAME", PT_GNU_EH_FRAME) 1917b0f6c590SRui Ueyama .Case("PT_GNU_STACK", PT_GNU_STACK) 1918b0f6c590SRui Ueyama .Case("PT_GNU_RELRO", PT_GNU_RELRO) 1919270173f2SGeorge Rimar .Case("PT_OPENBSD_RANDOMIZE", PT_OPENBSD_RANDOMIZE) 1920cc6e567cSGeorge Rimar .Case("PT_OPENBSD_WXNEEDED", PT_OPENBSD_WXNEEDED) 1921a2a32c2cSGeorge Rimar .Case("PT_OPENBSD_BOOTDATA", PT_OPENBSD_BOOTDATA) 1922b0f6c590SRui Ueyama .Default(-1); 1923bbe38602SEugene Leviant 1924b0f6c590SRui Ueyama if (Ret == (unsigned)-1) { 1925b0f6c590SRui Ueyama setError("invalid program header type: " + Tok); 1926b0f6c590SRui Ueyama return PT_NULL; 1927b0f6c590SRui Ueyama } 1928b0f6c590SRui Ueyama return Ret; 1929bbe38602SEugene Leviant } 1930bbe38602SEugene Leviant 193112450b20SRui Ueyama // Reads a list of symbols, e.g. "{ global: foo; bar; local: *; };". 193212450b20SRui Ueyama void ScriptParser::readAnonymousDeclaration() { 193312450b20SRui Ueyama // Read global symbols first. "global:" is default, so if there's 193412450b20SRui Ueyama // no label, we assume global symbols. 19354524268cSRafael Espindola if (peek() != "local") { 19364524268cSRafael Espindola if (consume("global")) 19374524268cSRafael Espindola expect(":"); 193812450b20SRui Ueyama Config->VersionScriptGlobals = readSymbols(); 19394524268cSRafael Espindola } 1940e999ddb8SRafael Espindola readLocals(); 194112450b20SRui Ueyama expect("}"); 194212450b20SRui Ueyama expect(";"); 194312450b20SRui Ueyama } 194412450b20SRui Ueyama 1945e999ddb8SRafael Espindola void ScriptParser::readLocals() { 19464524268cSRafael Espindola if (!consume("local")) 1947e999ddb8SRafael Espindola return; 19484524268cSRafael Espindola expect(":"); 1949e999ddb8SRafael Espindola std::vector<SymbolVersion> Locals = readSymbols(); 1950e999ddb8SRafael Espindola for (SymbolVersion V : Locals) { 1951e999ddb8SRafael Espindola if (V.Name == "*") { 1952e999ddb8SRafael Espindola Config->DefaultSymbolVersion = VER_NDX_LOCAL; 1953e999ddb8SRafael Espindola continue; 1954e999ddb8SRafael Espindola } 1955e999ddb8SRafael Espindola Config->VersionScriptLocals.push_back(V); 1956e999ddb8SRafael Espindola } 1957e999ddb8SRafael Espindola } 1958e999ddb8SRafael Espindola 195912450b20SRui Ueyama // Reads a list of symbols, e.g. "VerStr { global: foo; bar; local: *; };". 196095769b4aSRui Ueyama void ScriptParser::readVersionDeclaration(StringRef VerStr) { 196120b6598cSGeorge Rimar // Identifiers start at 2 because 0 and 1 are reserved 196220b6598cSGeorge Rimar // for VER_NDX_LOCAL and VER_NDX_GLOBAL constants. 1963da805c48SRui Ueyama uint16_t VersionId = Config->VersionDefinitions.size() + 2; 196420b6598cSGeorge Rimar Config->VersionDefinitions.push_back({VerStr, VersionId}); 196520b6598cSGeorge Rimar 196612450b20SRui Ueyama // Read global symbols. 19674524268cSRafael Espindola if (peek() != "local") { 19684524268cSRafael Espindola if (consume("global")) 19694524268cSRafael Espindola expect(":"); 197012450b20SRui Ueyama Config->VersionDefinitions.back().Globals = readSymbols(); 19714524268cSRafael Espindola } 1972e999ddb8SRafael Espindola readLocals(); 197320b6598cSGeorge Rimar expect("}"); 197420b6598cSGeorge Rimar 197512450b20SRui Ueyama // Each version may have a parent version. For example, "Ver2" 197612450b20SRui Ueyama // defined as "Ver2 { global: foo; local: *; } Ver1;" has "Ver1" 197712450b20SRui Ueyama // as a parent. This version hierarchy is, probably against your 197812450b20SRui Ueyama // instinct, purely for hint; the runtime doesn't care about it 197912450b20SRui Ueyama // at all. In LLD, we simply ignore it. 198012450b20SRui Ueyama if (peek() != ";") 19815424e7c7SJustin Bogner skip(); 198220b6598cSGeorge Rimar expect(";"); 198320b6598cSGeorge Rimar } 198420b6598cSGeorge Rimar 198512450b20SRui Ueyama // Reads a list of symbols for a versions cript. 198612450b20SRui Ueyama std::vector<SymbolVersion> ScriptParser::readSymbols() { 198712450b20SRui Ueyama std::vector<SymbolVersion> Ret; 1988e0fc2421SGeorge Rimar for (;;) { 19891ef90d2fSRafael Espindola if (consume("extern")) { 199012450b20SRui Ueyama for (SymbolVersion V : readVersionExtern()) 199112450b20SRui Ueyama Ret.push_back(V); 19921ef90d2fSRafael Espindola continue; 19931ef90d2fSRafael Espindola } 1994e0fc2421SGeorge Rimar 1995f3965c02SDmitry Mikulin if (peek() == "}" || (peek() == "local" && peek(1) == ":") || Error) 199612450b20SRui Ueyama break; 19970ee25a69SRui Ueyama StringRef Tok = next(); 199812450b20SRui Ueyama Ret.push_back({unquote(Tok), false, hasWildcard(Tok)}); 1999e0fc2421SGeorge Rimar expect(";"); 2000e0fc2421SGeorge Rimar } 200112450b20SRui Ueyama return Ret; 2002e0fc2421SGeorge Rimar } 2003e0fc2421SGeorge Rimar 200412450b20SRui Ueyama // Reads an "extern C++" directive, e.g., 200512450b20SRui Ueyama // "extern "C++" { ns::*; "f(int, double)"; };" 200612450b20SRui Ueyama std::vector<SymbolVersion> ScriptParser::readVersionExtern() { 20077e71415cSRafael Espindola StringRef Tok = next(); 20087e71415cSRafael Espindola bool IsCXX = Tok == "\"C++\""; 20097e71415cSRafael Espindola if (!IsCXX && Tok != "\"C\"") 2010d0ebd84cSRafael Espindola setError("Unknown language"); 201120b6598cSGeorge Rimar expect("{"); 201220b6598cSGeorge Rimar 201312450b20SRui Ueyama std::vector<SymbolVersion> Ret; 20140ee25a69SRui Ueyama while (!Error && peek() != "}") { 20150ee25a69SRui Ueyama StringRef Tok = next(); 20160ee25a69SRui Ueyama bool HasWildcard = !Tok.startswith("\"") && hasWildcard(Tok); 20177e71415cSRafael Espindola Ret.push_back({unquote(Tok), IsCXX, HasWildcard}); 201820b6598cSGeorge Rimar expect(";"); 201920b6598cSGeorge Rimar } 202020b6598cSGeorge Rimar 202120b6598cSGeorge Rimar expect("}"); 202220b6598cSGeorge Rimar expect(";"); 202312450b20SRui Ueyama return Ret; 202420b6598cSGeorge Rimar } 202520b6598cSGeorge Rimar 202624e626ccSRui Ueyama uint64_t ScriptParser::readMemoryAssignment( 202724e626ccSRui Ueyama StringRef S1, StringRef S2, StringRef S3) { 202824e626ccSRui Ueyama if (!(consume(S1) || consume(S2) || consume(S3))) { 202924e626ccSRui Ueyama setError("expected one of: " + S1 + ", " + S2 + ", or " + S3); 203024e626ccSRui Ueyama return 0; 203124e626ccSRui Ueyama } 203224e626ccSRui Ueyama expect("="); 203324e626ccSRui Ueyama 203424e626ccSRui Ueyama // TODO: Fully support constant expressions. 203524e626ccSRui Ueyama uint64_t Val; 203624e626ccSRui Ueyama if (!readInteger(next(), Val)) 203724e626ccSRui Ueyama setError("nonconstant expression for "+ S1); 203824e626ccSRui Ueyama return Val; 203924e626ccSRui Ueyama } 204024e626ccSRui Ueyama 204124e626ccSRui Ueyama // Parse the MEMORY command as specified in: 204224e626ccSRui Ueyama // https://sourceware.org/binutils/docs/ld/MEMORY.html 204324e626ccSRui Ueyama // 204424e626ccSRui Ueyama // MEMORY { name [(attr)] : ORIGIN = origin, LENGTH = len ... } 2045b889744eSMeador Inge void ScriptParser::readMemory() { 2046b889744eSMeador Inge expect("{"); 2047b889744eSMeador Inge while (!Error && !consume("}")) { 2048b889744eSMeador Inge StringRef Name = next(); 204924e626ccSRui Ueyama 2050b889744eSMeador Inge uint32_t Flags = 0; 20518a8a953eSRui Ueyama uint32_t NegFlags = 0; 2052b889744eSMeador Inge if (consume("(")) { 20538a8a953eSRui Ueyama std::tie(Flags, NegFlags) = readMemoryAttributes(); 2054b889744eSMeador Inge expect(")"); 2055b889744eSMeador Inge } 2056b889744eSMeador Inge expect(":"); 2057b889744eSMeador Inge 205824e626ccSRui Ueyama uint64_t Origin = readMemoryAssignment("ORIGIN", "org", "o"); 2059b889744eSMeador Inge expect(","); 206024e626ccSRui Ueyama uint64_t Length = readMemoryAssignment("LENGTH", "len", "l"); 2061b889744eSMeador Inge 2062b889744eSMeador Inge // Add the memory region to the region map (if it doesn't already exist). 2063b889744eSMeador Inge auto It = Opt.MemoryRegions.find(Name); 2064b889744eSMeador Inge if (It != Opt.MemoryRegions.end()) 2065b889744eSMeador Inge setError("region '" + Name + "' already defined"); 2066b889744eSMeador Inge else 20678a8a953eSRui Ueyama Opt.MemoryRegions[Name] = {Name, Origin, Length, Origin, Flags, NegFlags}; 2068b889744eSMeador Inge } 2069b889744eSMeador Inge } 2070b889744eSMeador Inge 2071b889744eSMeador Inge // This function parses the attributes used to match against section 2072b889744eSMeador Inge // flags when placing output sections in a memory region. These flags 2073b889744eSMeador Inge // are only used when an explicit memory region name is not used. 2074b889744eSMeador Inge std::pair<uint32_t, uint32_t> ScriptParser::readMemoryAttributes() { 2075b889744eSMeador Inge uint32_t Flags = 0; 20768a8a953eSRui Ueyama uint32_t NegFlags = 0; 2077b889744eSMeador Inge bool Invert = false; 2078481ac996SRui Ueyama 2079481ac996SRui Ueyama for (char C : next().lower()) { 2080b889744eSMeador Inge uint32_t Flag = 0; 2081b889744eSMeador Inge if (C == '!') 2082b889744eSMeador Inge Invert = !Invert; 2083481ac996SRui Ueyama else if (C == 'w') 2084b889744eSMeador Inge Flag = SHF_WRITE; 2085481ac996SRui Ueyama else if (C == 'x') 2086b889744eSMeador Inge Flag = SHF_EXECINSTR; 2087481ac996SRui Ueyama else if (C == 'a') 2088b889744eSMeador Inge Flag = SHF_ALLOC; 2089481ac996SRui Ueyama else if (C != 'r') 2090b889744eSMeador Inge setError("invalid memory region attribute"); 2091481ac996SRui Ueyama 2092b889744eSMeador Inge if (Invert) 20938a8a953eSRui Ueyama NegFlags |= Flag; 2094b889744eSMeador Inge else 2095b889744eSMeador Inge Flags |= Flag; 2096b889744eSMeador Inge } 20978a8a953eSRui Ueyama return {Flags, NegFlags}; 2098b889744eSMeador Inge } 2099b889744eSMeador Inge 210007320e40SRui Ueyama void elf::readLinkerScript(MemoryBufferRef MB) { 210122375f24SRui Ueyama ScriptParser(MB).readLinkerScript(); 210220b6598cSGeorge Rimar } 210320b6598cSGeorge Rimar 210420b6598cSGeorge Rimar void elf::readVersionScript(MemoryBufferRef MB) { 210522375f24SRui Ueyama ScriptParser(MB).readVersionScript(); 2106f7c5fbb1SRui Ueyama } 21071ebc8ed7SRui Ueyama 2108d0ebd84cSRafael Espindola void elf::readDynamicList(MemoryBufferRef MB) { 2109d0ebd84cSRafael Espindola ScriptParser(MB).readDynamicList(); 2110d0ebd84cSRafael Espindola } 2111d0ebd84cSRafael Espindola 211207320e40SRui Ueyama template class elf::LinkerScript<ELF32LE>; 211307320e40SRui Ueyama template class elf::LinkerScript<ELF32BE>; 211407320e40SRui Ueyama template class elf::LinkerScript<ELF64LE>; 211507320e40SRui Ueyama template class elf::LinkerScript<ELF64BE>; 2116