[ELF] - Implemented --defsym option

[grimar]

Summary

gnu ld description of option is:

--defsym=symbol=expression
Create a global symbol in the output file, containing the absolute address given by expression. You may use this option as many times as necessary to define multiple symbols in the command line. A limited form of arithmetic is supported for the expression in this context: you may give a hexadecimal constant or the name of an existing symbol, or use "+" and "-" to add or subtract hexadecimal constants or symbols. If you need more elaborate expressions, consider using the linker command language from a script. Note: there should be no white space between symbol, the equals sign ("="), and expression.

In compare with D32082, this patch does not support math expressions and absolute symbols. It implemented via code similar to --wrap.
That covers 1 of 3 possible --defsym cases.

[davide]

ELF/SymbolTable.cpp

174–178 ↗

(On Diff #95586)

this bit is untested.

[grimar]

• Addressed review comment.

ELF/SymbolTable.cpp

174–178 ↗

(On Diff #95586)

Fixed.

[grimar]

• Updated testcase to show we can use aliasing.

test/ELF/defsym.s

16 ↗

(On Diff #95591)

Forgot to mention this moment. We have 2 symbols 'foo1' in output here. That is because I don't
change name of alias symbol body in alias().
To change the name in symtab I think I need to make SymbolBody::Name public
and remove getter, that allows to set new name, but also involves changes in many places. Do we want it ?
Aliasing itself works without that I believe.

[ruiu]

ELF/Driver.cpp

884 ↗

(On Diff #95693)

I'd support just 1.

887 ↗

(On Diff #95693)

Make this a function that returns a vector of pair<StringRef, StringRef> and use alias function in link().

[grimar]

• Addressed review comments.

ELF/Driver.cpp

884 ↗

(On Diff #95693)

Done.

887 ↗

(On Diff #95693)

Done.

[ruiu]

Generally looking good, a few nits.

Please remove "#2" from the patch title. It's not relevant for people who will be reading this patch after you submit this, as they don't know and don't care about how many you have tried before.

ELF/Driver.cpp

886 ↗

(On Diff #95926)

Expr is not an expr, but a symbol name, so Expr is not a good name. I'd name them From and To.

888 ↗

(On Diff #95926)

Remove =.

Since we expect a symbol name, the error should be something like error("-defsym: symbol name expected, but got " + To);

ELF/Options.td

28 ↗

(On Diff #95926)

"Define a symbol alias"

ELF/SymbolTable.cpp

176 ↗

(On Diff #95926)

= is not part of the option name but a separator.

Also, we are not handling an expression, so don't use the word "expression".

Error message should be error("-defsym: undefined symbol: " + Name);

[grimar]

• Addressed review comments.

ELF/Driver.cpp

886 ↗

(On Diff #95926)

Done.

888 ↗

(On Diff #95926)

Done.

ELF/Options.td

28 ↗

(On Diff #95926)

Done.

ELF/SymbolTable.cpp

176 ↗

(On Diff #95926)

Done.

[ruiu]

LGTM

lld/trunk/ELF/Driver.cpp

if (S.getAsInteger(0, V)) {
     error("-image-base: number expected, but got " + S);
     return 0;
   }
   if ((V % Config->MaxPageSize) != 0)
     warn("-image-base: address isn't multiple of page size: " + S);
   return V;
 }
 
+// Parses --defsym=alias option.
+static std::vector<std::pair<StringRef, StringRef>>
+getDefsym(opt::InputArgList &Args) {
+  std::vector<std::pair<StringRef, StringRef>> Ret;
+  for (auto *Arg : Args.filtered(OPT_defsym)) {
+    StringRef From;
+    StringRef To;
+    std::tie(From, To) = StringRef(Arg->getValue()).split('=');
+    if (!isValidCIdentifier(To))
+      error("--defsym: symbol name expected, but got " + To);
+    Ret.push_back({From, To});
+  }
+  return Ret;
+}
+
 // Do actual linking. Note that when this function is called,
 // all linker scripts have already been parsed.
 template <class ELFT> void LinkerDriver::link(opt::InputArgList &Args) {
   SymbolTable<ELFT> Symtab;
   elf::Symtab<ELFT>::X = &Symtab;
   Target = createTarget();
 
   Config->MaxPageSize = getMaxPageSize(Args);
template <class ELFT> void LinkerDriver::link(opt::InputArgList &Args) {
   // Some symbols (such as __ehdr_start) are defined lazily only when there
   // are undefined symbols for them, so we add these to trigger that logic.
   for (StringRef Sym : Script->Opt.ReferencedSymbols)
     Symtab.addUndefined(Sym);
 
   for (auto *Arg : Args.filtered(OPT_wrap))
     Symtab.wrap(Arg->getValue());
 
+  // Handle --defsym=sym=alias option.
+  for (std::pair<StringRef, StringRef> &Def : getDefsym(Args))
+    Symtab.alias(Def.first, Def.second);
+
   // Now that we have a complete list of input files.
   // Beyond this point, no new files are added.
   // Aggregate all input sections into one place.
   for (elf::ObjectFile<ELFT> *F : Symtab.getObjectFiles())
     for (InputSectionBase *S : F->getSections())
       if (S && S != &InputSection::Discarded)
         InputSections.push_back(S);
   for (BinaryFile *F : Symtab.getBinaryFiles())

lld/trunk/ELF/Options.td

 
 def build_id: F<"build-id">, HelpText<"Generate build ID note">;
 
 def build_id_eq: J<"build-id=">, HelpText<"Generate build ID note">;
 
 def compress_debug_sections : J<"compress-debug-sections=">,
   HelpText<"Compress DWARF debug sections">;
 
+def defsym: J<"defsym=">, HelpText<"Define a symbol alias">;
+
 def L: JoinedOrSeparate<["-"], "L">, MetaVarName<"<dir>">,
   HelpText<"Add a directory to the library search path">;
 
 def O: Joined<["-"], "O">, HelpText<"Optimize output file size">;
 
 def Tbss: S<"Tbss">, HelpText<"Same as --section-start with .bss as the sectionname">;
 
 def Tdata: S<"Tdata">, HelpText<"Same as --section-start with .data as the sectionname">;

lld/trunk/ELF/SymbolTable.h

public:
   void scanShlibUndefined();
   void scanVersionScript();
 
   SymbolBody *find(StringRef Name);
   SymbolBody *findInCurrentDSO(StringRef Name);
 
   void trace(StringRef Name);
   void wrap(StringRef Name);
+  void alias(StringRef Alias, StringRef Name);
 
 private:
   std::vector<SymbolBody *> findByVersion(SymbolVersion Ver);
   std::vector<SymbolBody *> findAllByVersion(SymbolVersion Ver);
 
   llvm::StringMap<std::vector<SymbolBody *>> &getDemangledSyms();
   void handleAnonymousVersion();
   void assignExactVersion(SymbolVersion Ver, uint16_t VersionId,

lld/trunk/ELF/SymbolTable.cpp

template <class ELFT> void SymbolTable<ELFT>::wrap(StringRef Name) {
 
   // We rename symbols by replacing the old symbol's SymbolBody with the new
   // symbol's SymbolBody. This causes all SymbolBody pointers referring to the
   // old symbol to instead refer to the new symbol.
   memcpy(Real->Body.buffer, Sym->Body.buffer, sizeof(Sym->Body));
   memcpy(Sym->Body.buffer, Wrap->Body.buffer, sizeof(Wrap->Body));
 }
 
+// Creates alias for symbol. Used to implement --defsym=ALIAS=SYM.
+template <class ELFT>
+void SymbolTable<ELFT>::alias(StringRef Alias, StringRef Name) {
+  SymbolBody *B = find(Name);
+  if (!B) {
+    error("-defsym: undefined symbol: " + Name);
+    return;
+  }
+  Symbol *Sym = B->symbol();
+  Symbol *AliasSym = addUndefined(Alias);
+  memcpy(AliasSym->Body.buffer, Sym->Body.buffer, sizeof(AliasSym->Body));
+}
+
 static uint8_t getMinVisibility(uint8_t VA, uint8_t VB) {
   if (VA == STV_DEFAULT)
     return VB;
   if (VB == STV_DEFAULT)
     return VA;
   return std::min(VA, VB);
 }
 

lld/trunk/test/ELF/defsym.s

+# REQUIRES: x86
+# RUN: llvm-mc -filetype=obj -triple=x86_64-pc-linux %s -o %t.o
+# RUN: ld.lld -o %t %t.o --defsym=foo2=foo1
+# RUN: llvm-readobj -t -s %t | FileCheck %s
+# RUN: llvm-objdump -d -print-imm-hex %t | FileCheck %s --check-prefix=USE
+
+## In compare with GNU linkers, symbol defined with --defsym does
+## not get aliased name in symbol table:
+# CHECK:      Symbol {
+# CHECK:        Name: foo1
+# CHECK-NEXT:   Value: 0x123
+# CHECK-NEXT:   Size:
+# CHECK-NEXT:   Binding: Global
+# CHECK-NEXT:   Type:
+# CHECK-NEXT:   Other:
+# CHECK-NEXT:   Section: Absolute
+# CHECK-NEXT: }
+# CHECK-NEXT: Symbol {
+# CHECK-NEXT:   Name: foo1
+# CHECK-NEXT:   Value: 0x123
+# CHECK-NEXT:   Size:
+# CHECK-NEXT:   Binding: Global
+# CHECK-NEXT:   Type:
+# CHECK-NEXT:   Other:
+# CHECK-NEXT:   Section: Absolute
+# CHECK-NEXT: }
+
+## Check we can use foo2 and it that it is an alias for foo1.
+# USE:       Disassembly of section .text:
+# USE-NEXT:  _start:
+# USE-NEXT:    movl $0x123, %edx
+
+# RUN: not ld.lld -o %t %t.o --defsym=foo2=1 2>&1 | FileCheck %s -check-prefix=ERR1
+# ERR1: error: --defsym: symbol name expected, but got 1
+
+# RUN: not ld.lld -o %t %t.o --defsym=foo2=und 2>&1 | FileCheck %s -check-prefix=ERR2
+# ERR2: error: -defsym: undefined symbol: und
+
+.globl foo1
+ foo1 = 0x123
+
+.global _start
+_start:
+  movl $foo2, %edx