[ELF2] --strip-all/-s command line implemented

[grimar]

Summary

-s, --strip-all - Strip all symbols
Implementation removes .strtab and .symtab sections from output.

Test from lld\test\elf branch was taken as a base, but was improved in part of checks.

[ruiu]

ELF/Writer.cpp

113–115

This is a bit confusing. Let's do this

std::unique_ptr<SymbolTableSection<ELFT>> SymTab;
if (!Config->StripAll) {
  SymTab.reset(new SymbolTableSection<ELFT>(*Symtab, *Out<ELFT>::StrTab);
  Out<ELFT>::SymTab = SymTab.get();
}

[grimar]

ELF/Writer.cpp

113–115

I thought about something like that, just didn`t know will it keep code consistency or not. I agree, that will be better.

btw, the same can be applied for RelaPlt and GotPlt since we dont need to create them when lazy relocs are off. (not for this patch)

[davide]

Structure looks good, just few comments.

ELF/Config.h

59

This is not sorted.

ELF/Options.td

86

Can't you just say "strip all symbols" as gold and ld do?

test/elf2/strip-all.s

1

Please do not use obj2yaml, but use assembler for tests. This is the standard for the new ELF linker.

[grimar]

ELF/Config.h

59

Will fix.
I forgot that after first two characters there are still other ones, sorry :)

ELF/Options.td

86

I took it from man ld, but will replace if you wish.

test/elf2/strip-all.s

1

Ok. I expected that. Thats explains why I never saw its usings in new ELF branch.

[davide]

ELF/Options.td

86

This is what I see on FreeBSD. I would not deviate from that as it's shorter and captures the essence anyway (which I think is better for --help output)
davide@ganondorf:~ % /usr/local/bin/ld.bfd --help | grep strip-all

-s, --strip-all             Strip all symbols

davide@ganondorf:~ % /usr/local/bin/ld.gold --help | grep strip-all

-s, --strip-all             Strip all symbols

Thanks,

Davide

[grimar]

Review comments addressed.

[ruiu]

LGTM

ELF/Config.h

struct Configuration {
   bool DiscardLocals;
   bool DiscardNone;
   bool EnableNewDtags;
   bool ExportDynamic;
   bool Mips64EL = false;
   bool NoInhibitExec;
   bool NoUndefined;
   bool Shared;
+  bool StripAll;
   bool Static = false;
   bool Verbose;
   bool ZNodelete = false;
   bool ZNow = false;
   ELFKind EKind = ELFNoneKind;
   uint16_t EMachine = llvm::ELF::EM_NONE;
   uint64_t EntryAddr = -1;
 };
 
 extern Configuration *Config;
 
 } // namespace elf2
 } // namespace lld
 
 #endif

ELF/Driver.cpp

void LinkerDriver::createFiles(opt::InputArgList &Args) {
   Config->DiscardAll = Args.hasArg(OPT_discard_all);
   Config->DiscardLocals = Args.hasArg(OPT_discard_locals);
   Config->DiscardNone = Args.hasArg(OPT_discard_none);
   Config->EnableNewDtags = !Args.hasArg(OPT_disable_new_dtags);
   Config->ExportDynamic = Args.hasArg(OPT_export_dynamic);
   Config->NoInhibitExec = Args.hasArg(OPT_noinhibit_exec);
   Config->NoUndefined = Args.hasArg(OPT_no_undefined);
   Config->Shared = Args.hasArg(OPT_shared);
+  Config->StripAll = Args.hasArg(OPT_strip_all);
   Config->Verbose = Args.hasArg(OPT_verbose);
 
   Config->DynamicLinker = getString(Args, OPT_dynamic_linker);
   Config->Entry = getString(Args, OPT_entry);
   Config->Fini = getString(Args, OPT_fini, "_fini");
   Config->Init = getString(Args, OPT_init, "_init");
   Config->OutputFile = getString(Args, OPT_o);
   Config->SoName = getString(Args, OPT_soname);

ELF/Options.td

 def script : Separate<["--"], "script">, HelpText<"Read linker script">;
 
 def shared : Flag<["-"], "shared">,
   HelpText<"Build a shared object">;
 
 def soname : Joined<["-"], "soname=">,
   HelpText<"Set DT_SONAME">;
 
+def strip_all : Flag<["--"], "strip-all">,
+  HelpText<"Omit all symbol information from the output file">;
+
 def sysroot : Joined<["--"], "sysroot=">,
   HelpText<"Set the system root">;
 
 def undefined : Joined<["--"], "undefined=">,
   HelpText<"Force undefined symbol during linking">;
 
 def verbose : Flag<["--"], "verbose">;
 

 def alias_init_init : Joined<["-"], "init=">, Alias<init>;
 def alias_l__library : Joined<["--"], "library=">, Alias<l>;
 def alias_o_output : Joined<["--"], "output=">, Alias<o>;
 def alias_rpath_rpath : Joined<["-"], "rpath=">, Alias<rpath>;
 def alias_shared_Bshareable : Flag<["-"], "Bshareable">, Alias<shared>;
 def alias_soname_h : Separate<["-"], "h">, Alias<soname>;
 def alias_soname_soname : Separate<["-"], "soname">, Alias<soname>;
 def alias_script_T : Separate<["-"], "T">, Alias<script>;
+def alias_strip_all: Flag<["-"], "s">, Alias<strip_all>;
 def alias_undefined_u : Separate<["-"], "u">, Alias<undefined>;
 
 // Options listed below are silently ignored now.
 def O3 : Flag<["-"], "O3">;
 def build_id : Flag<["--"], "build-id">;
 def eh_frame_hdr : Flag<["--"], "eh-frame-hdr">;
 def end_group : Flag<["--"], "end-group">;
 def fatal_warnings : Flag<["--"], "fatal-warnings">;
 def gc_sections : Flag<["--"], "gc-sections">;
 def hash_style : Joined<["--"], "hash-style=">;
 def no_add_needed : Flag<["--"], "no-add-needed">;
 def no_fatal_warnings : Flag<["--"], "no-fatal-warnings">;
 def no_warn_mismatch : Flag<["--"], "no-warn-mismatch">;
 def start_group : Flag<["--"], "start-group">;
-def strip_all : Flag<["--"], "strip-all">;
 def version_script : Separate<["--"], "version-script">;
 def warn_common : Flag<["--"], "warn-common">;
 def warn_shared_textrel : Flag<["--"], "warn-shared-textrel">;

ELF/OutputSections.h

 template <class ELFT>
 class SymbolTableSection final : public OutputSectionBase<ELFT> {
 public:
   typedef typename llvm::object::ELFFile<ELFT>::Elf_Shdr Elf_Shdr;
   typedef typename llvm::object::ELFFile<ELFT>::Elf_Sym Elf_Sym;
   typedef typename llvm::object::ELFFile<ELFT>::Elf_Sym_Range Elf_Sym_Range;
   typedef typename llvm::object::ELFFile<ELFT>::uintX_t uintX_t;
   SymbolTableSection(SymbolTable<ELFT> &Table,
-                     StringTableSection<ELFT> &StrTabSec);
+                     StringTableSection<ELFT> *StrTabSec);
 
   void finalize() override;
   void writeTo(uint8_t *Buf) override;
   void addLocalSymbol(StringRef Name);
   void addSymbol(SymbolBody *Body);
-  StringTableSection<ELFT> &getStrTabSec() const { return StrTabSec; }
   unsigned getNumSymbols() const { return NumVisible + 1; }
 
   ArrayRef<SymbolBody *> getSymbols() const { return Symbols; }
 
 private:
   void writeLocalSymbols(uint8_t *&Buf);
   void writeGlobalSymbols(uint8_t *Buf);
 
   static uint8_t getSymbolBinding(SymbolBody *Body);
 
   SymbolTable<ELFT> &Table;
-  StringTableSection<ELFT> &StrTabSec;
+  StringTableSection<ELFT> *StrTabSec;
   std::vector<SymbolBody *> Symbols;
   unsigned NumVisible = 0;
   unsigned NumLocals = 0;
 };
 
 template <class ELFT>
 class RelocationSection final : public OutputSectionBase<ELFT> {
   typedef typename llvm::object::ELFFile<ELFT>::Elf_Rel Elf_Rel;

ELF/OutputSections.cpp

if (Config->DiscardNone)
     return true;
 
   // ELF defines dynamic locals as symbols which name starts with ".L".
   return !(Config->DiscardLocals && SymName.startswith(".L"));
 }
 
 template <class ELFT>
 SymbolTableSection<ELFT>::SymbolTableSection(
-    SymbolTable<ELFT> &Table, StringTableSection<ELFT> &StrTabSec)
+    SymbolTable<ELFT> &Table, StringTableSection<ELFT> *StrTabSec)
     : OutputSectionBase<ELFT>(
-          StrTabSec.isDynamic() ? ".dynsym" : ".symtab",
-          StrTabSec.isDynamic() ? llvm::ELF::SHT_DYNSYM : llvm::ELF::SHT_SYMTAB,
-          StrTabSec.isDynamic() ? (uintX_t)llvm::ELF::SHF_ALLOC : 0),
+          (StrTabSec && StrTabSec->isDynamic()) ? ".dynsym" : ".symtab",
+          (StrTabSec && StrTabSec->isDynamic()) ? llvm::ELF::SHT_DYNSYM
+                                                : llvm::ELF::SHT_SYMTAB,
+          (StrTabSec && StrTabSec->isDynamic()) ? (uintX_t)llvm::ELF::SHF_ALLOC
+                                                : 0),
       Table(Table), StrTabSec(StrTabSec) {
   typedef OutputSectionBase<ELFT> Base;
   typename Base::Elf_Shdr &Header = this->Header;
 
   Header.sh_entsize = sizeof(Elf_Sym);
   Header.sh_addralign = ELFT::Is64Bits ? 8 : 4;
 }
 
 template <class ELFT> void SymbolTableSection<ELFT>::finalize() {
   this->Header.sh_size = getNumSymbols() * sizeof(Elf_Sym);
-  this->Header.sh_link = StrTabSec.SectionIndex;
+  this->Header.sh_link = StrTabSec->SectionIndex;
   this->Header.sh_info = NumLocals + 1;
 
-  if (!StrTabSec.isDynamic()) {
+  if (!StrTabSec->isDynamic()) {
     std::stable_sort(Symbols.begin(), Symbols.end(),
                      [](SymbolBody *L, SymbolBody *R) {
                        return getSymbolBinding(L) == STB_LOCAL &&
                               getSymbolBinding(R) != STB_LOCAL;
                      });
     return;
   }
   size_t I = 0;
   for (SymbolBody *Body : Symbols)
     Body->setDynamicSymbolTableIndex(++I);
 }
 
 template <class ELFT>
 void SymbolTableSection<ELFT>::addLocalSymbol(StringRef Name) {
-  StrTabSec.add(Name);
+  StrTabSec->add(Name);
   ++NumVisible;
   ++NumLocals;
 }
 
 template <class ELFT>
 void SymbolTableSection<ELFT>::addSymbol(SymbolBody *Body) {
-  StrTabSec.add(Body->getName());
+  StrTabSec->add(Body->getName());
   Symbols.push_back(Body);
   ++NumVisible;
 }
 
 template <class ELFT> void SymbolTableSection<ELFT>::writeTo(uint8_t *Buf) {
   Buf += sizeof(Elf_Sym);
 
   // All symbols with STB_LOCAL binding precede the weak and global symbols.
   // .dynsym only contains global symbols.
-  if (!Config->DiscardAll && !StrTabSec.isDynamic())
+  if (!Config->DiscardAll && !StrTabSec->isDynamic())
     writeLocalSymbols(Buf);
 
   writeGlobalSymbols(Buf);
 }
 
 template <class ELFT>
 void SymbolTableSection<ELFT>::writeLocalSymbols(uint8_t *&Buf) {
   // Iterate over all input object files to copy their local symbols
   // to the output symbol table pointed by Buf.
   for (const std::unique_ptr<ObjectFile<ELFT>> &File : Table.getObjectFiles()) {
     Elf_Sym_Range Syms = File->getLocalSymbols();
     for (const Elf_Sym &Sym : Syms) {
       ErrorOr<StringRef> SymNameOrErr = Sym.getName(File->getStringTable());
       error(SymNameOrErr);
       StringRef SymName = *SymNameOrErr;
       if (!shouldKeepInSymtab<ELFT>(*File, SymName, Sym))
         continue;
 
       auto *ESym = reinterpret_cast<Elf_Sym *>(Buf);
       Buf += sizeof(*ESym);
-      ESym->st_name = StrTabSec.getFileOff(SymName);
+      ESym->st_name = StrTabSec->getFileOff(SymName);
       ESym->st_size = Sym.st_size;
       ESym->setBindingAndType(Sym.getBinding(), Sym.getType());
       uintX_t VA = 0;
       if (Sym.st_shndx == SHN_ABS) {
         ESym->st_shndx = SHN_ABS;
         VA = Sym.st_value;
       } else {
         const InputSectionBase<ELFT> *Section = File->getSection(Sym);

 template <class ELFT>
 void SymbolTableSection<ELFT>::writeGlobalSymbols(uint8_t *Buf) {
   // Write the internal symbol table contents to the output symbol table
   // pointed by Buf.
   auto *ESym = reinterpret_cast<Elf_Sym *>(Buf);
   for (SymbolBody *Body : Symbols) {
     StringRef Name = Body->getName();
 
-    ESym->st_name = StrTabSec.getFileOff(Name);
+    ESym->st_name = StrTabSec->getFileOff(Name);
 
     const OutputSectionBase<ELFT> *OutSec = nullptr;
     const InputSectionBase<ELFT> *Section = nullptr;
 
     switch (Body->kind()) {
     case SymbolBody::DefinedSyntheticKind:
       OutSec = &cast<DefinedSynthetic<ELFT>>(Body)->Section;
       break;

ELF/Writer.cpp

 template <class ELFT> void lld::elf2::writeResult(SymbolTable<ELFT> *Symtab) {
   // Initialize output sections that are handled by Writer specially.
   // Don't reorder because the order of initialization matters.
   InterpSection<ELFT> Interp;
   Out<ELFT>::Interp = &Interp;
   StringTableSection<ELFT> ShStrTab(".shstrtab", false);
   Out<ELFT>::ShStrTab = &ShStrTab;
   StringTableSection<ELFT> StrTab(".strtab", false);
+  if (!Config->StripAll)
     Out<ELFT>::StrTab = &StrTab;
   StringTableSection<ELFT> DynStrTab(".dynstr", true);
   Out<ELFT>::DynStrTab = &DynStrTab;
   OutputSection<ELFT> Bss(".bss", SHT_NOBITS, SHF_ALLOC | SHF_WRITE);
   Out<ELFT>::Bss = &Bss;
   GotSection<ELFT> Got;
   Out<ELFT>::Got = &Got;
   GotPltSection<ELFT> GotPlt;
   if (Target->supportsLazyRelocations())
     Out<ELFT>::GotPlt = &GotPlt;
   PltSection<ELFT> Plt;
   Out<ELFT>::Plt = &Plt;
-  SymbolTableSection<ELFT> SymTab(*Symtab, *Out<ELFT>::StrTab);
+  SymbolTableSection<ELFT> SymTab(*Symtab, Out<ELFT>::StrTab);
+  if (!Config->StripAll)
     Out<ELFT>::SymTab = &SymTab;
-  SymbolTableSection<ELFT> DynSymTab(*Symtab, *Out<ELFT>::DynStrTab);
+  SymbolTableSection<ELFT> DynSymTab(*Symtab, Out<ELFT>::DynStrTab);
   Out<ELFT>::DynSymTab = &DynSymTab;
   HashTableSection<ELFT> HashTab;
   Out<ELFT>::HashTab = &HashTab;
   bool IsRela = Symtab->shouldUseRela();
   RelocationSection<ELFT> RelaDyn(IsRela ? ".rela.dyn" : ".rel.dyn", IsRela);
   Out<ELFT>::RelaDyn = &RelaDyn;
   RelocationSection<ELFT> RelaPlt(IsRela ? ".rela.plt" : ".rel.plt", IsRela);
   if (Target->supportsLazyRelocations())

 template <class ELFT> void Writer<ELFT>::copyLocalSymbols() {
   for (const std::unique_ptr<ObjectFile<ELFT>> &F : Symtab.getObjectFiles()) {
     for (const Elf_Sym &Sym : F->getLocalSymbols()) {
       ErrorOr<StringRef> SymNameOrErr = Sym.getName(F->getStringTable());
       error(SymNameOrErr);
       StringRef SymName = *SymNameOrErr;
       if (!shouldKeepInSymtab<ELFT>(*F, SymName, Sym))
         continue;
+      if (Out<ELFT>::SymTab)
         Out<ELFT>::SymTab->addLocalSymbol(SymName);
     }
   }
 }
 
 // PPC64 has a number of special SHT_PROGBITS+SHF_ALLOC+SHF_WRITE sections that
 // we would like to make sure appear is a specific order to maximize their
 // coverage by a single signed 16-bit offset from the TOC base pointer.
 // Conversely, the special .tocbss section should be first among all SHT_NOBITS
for (auto &P : Symtab.getSymbols()) {
     if (auto *U = dyn_cast<Undefined<ELFT>>(Body))
       if (!U->isWeak() && !U->canKeepUndefined())
         reportUndefined<ELFT>(Symtab, *Body);
 
     if (auto *C = dyn_cast<DefinedCommon<ELFT>>(Body))
       CommonSymbols.push_back(C);
     if (!includeInSymtab<ELFT>(*Body))
       continue;
+    if (Out<ELFT>::SymTab)
       Out<ELFT>::SymTab->addSymbol(Body);
 
     if (isOutputDynamic() && includeInDynamicSymtab(*Body))
       Out<ELFT>::DynSymTab->addSymbol(Body);
   }
   addCommonSymbols(CommonSymbols);
 
   // This order is not the same as the final output order
   // because we sort the sections using their attributes below.
+  if (Out<ELFT>::SymTab)
     OutputSections.push_back(Out<ELFT>::SymTab);
   OutputSections.push_back(Out<ELFT>::ShStrTab);
+  if (Out<ELFT>::StrTab)
     OutputSections.push_back(Out<ELFT>::StrTab);
   if (isOutputDynamic()) {
     OutputSections.push_back(Out<ELFT>::DynSymTab);
     OutputSections.push_back(Out<ELFT>::HashTab);
     OutputSections.push_back(Out<ELFT>::Dynamic);
     OutputSections.push_back(Out<ELFT>::DynStrTab);
     if (Out<ELFT>::RelaDyn->hasRelocs())
       OutputSections.push_back(Out<ELFT>::RelaDyn);
     if (Out<ELFT>::RelaPlt && Out<ELFT>::RelaPlt->hasRelocs())

test/elf2/strip-all.s

+# Tests the --strip-all (-s) flag.
+# We expect that .symtab and .strtab are stripped from the final executable.
+#
+# The following code was used to generate the object.
+# $ clang -c blah.c -o blah
+#
+# void
+# callMeMaybe(int *v)
+# {
+#   *v += 1;
+# }
+#
+# int
+# main(void)
+# {
+#
+#   int blah = 42;
+#   callMeMaybe(&blah);
+# }
+
+#RUN: yaml2obj -format=elf %s -o=%t.o
+#RUN: ld.lld2 %t.o -e main -o %t1
+#RUN: llvm-objdump -section-headers %t1 | FileCheck %s -check-prefix BEFORE
+#BEFORE:       7 .symtab 00000048
+#BEFORE-NEXT:  8 .shstrtab 0000004f
+#BEFORE-NEXT:  9 .strtab 00000012
+
+#RUN: ld.lld2 %t.o -e main --strip-all -o %t1
+#RUN: llvm-objdump -section-headers %t1 | FileCheck %s -check-prefix AFTER
+#AFTER-NOT: .symtab
+#AFTER: 7 .shstrtab 0000003f
+#AFTER-NOT: .strtab
+
+# Test alias -s
+#RUN: ld.lld2 %t.o -e main -s -o %t1
+#RUN: llvm-objdump -section-headers %t1 | FileCheck %s -check-prefix AFTER
+
+---
+FileHeader:
+  Class:           ELFCLASS64
+  Data:            ELFDATA2LSB
+  OSABI:           ELFOSABI_FREEBSD
+  Type:            ET_REL
+  Machine:         EM_X86_64
+Sections:
+  - Name:            .text
+    Type:            SHT_PROGBITS
+    Flags:           [ SHF_ALLOC, SHF_EXECINSTR ]
+    AddressAlign:    0x0000000000000010
+    Content:         554889E548897DF8488B7DF88B07050100000089075DC3660F1F840000000000554889E54883EC10488D7DFCC745FC2A000000E8C8FFFFFFB8000000004883C4105DC3
+  - Name:            .data
+    Type:            SHT_PROGBITS
+    Flags:           [ SHF_WRITE, SHF_ALLOC ]
+    AddressAlign:    0x0000000000000004
+    Content:         ''
+  - Name:            .bss
+    Type:            SHT_NOBITS
+    Flags:           [ SHF_WRITE, SHF_ALLOC ]
+    AddressAlign:    0x0000000000000004
+  - Name:            .comment
+    Type:            SHT_PROGBITS
+    Flags:           [ SHF_MERGE, SHF_STRINGS ]
+    AddressAlign:    0x0000000000000001
+    Content:         004672656542534420636C616E672076657273696F6E20332E342E312028746167732F52454C454153455F33342F646F74312D66696E616C203230383033322920323031343035313200
+  - Name:            .note.GNU-stack
+    Type:            SHT_PROGBITS
+    AddressAlign:    0x0000000000000001
+    Content:         ''
+  - Name:            .eh_frame
+    Type:            SHT_PROGBITS
+    Flags:           [ SHF_ALLOC ]
+    AddressAlign:    0x0000000000000008
+    Content:         1400000000000000017A5200017810011B0C070890010000180000001C000000000000001700000000410E108602430D060000001800000038000000000000002300000000410E108602430D06000000
+  - Name:            .rela.eh_frame
+    Type:            SHT_RELA
+    Link:            .symtab
+    AddressAlign:    0x0000000000000008
+    Info:            .eh_frame
+    Relocations:
+      - Offset:          0x0000000000000020
+        Symbol:          .text
+        Type:            R_X86_64_PC32
+      - Offset:          0x000000000000003C
+        Symbol:          .text
+        Type:            R_X86_64_PC32
+        Addend:          32
+Symbols:
+  Local:
+    - Name:            .text
+      Type:            STT_SECTION
+      Section:         .text
+    - Name:            .data
+      Type:            STT_SECTION
+      Section:         .data
+    - Name:            .bss
+      Type:            STT_SECTION
+      Section:         .bss
+    - Name:            .comment
+      Type:            STT_SECTION
+      Section:         .comment
+    - Name:            .note.GNU-stack
+      Type:            STT_SECTION
+      Section:         .note.GNU-stack
+    - Name:            .eh_frame
+      Type:            STT_SECTION
+      Section:         .eh_frame
+  Global:
+    - Name:            callMeMaybe
+      Type:            STT_FUNC
+      Section:         .text
+      Size:            0x0000000000000017
+    - Name:            main
+      Type:            STT_FUNC
+      Section:         .text
+      Value:           0x0000000000000020
+      Size:            0x0000000000000023
+...