lld: unquote possibly quoted `EXTERN("symbol")` entry in linker script

[luciang]

Summary

gold accepts quoted strings:

gold/yyscript.y

252┊ file_cmd:
253┊           EXTERN '(' extern_name_list ')'

312┊ /* A list of external undefined symbols.  We put the lexer into
313┊    expression mode so that commas separate names; this is what the GNU
314┊    linker does.  */
315┊
316┊ extern_name_list:
317┊             { script_push_lex_into_expression_mode(closure); }
318┊           extern_name_list_body
319┊             { script_pop_lex_mode(closure); }
320┊         ;
321┊
322┊ extern_name_list_body:
323┊           string
324┊             { script_add_extern(closure, $1.value, $1.length); }
325┊         | extern_name_list_body string
326┊             { script_add_extern(closure, $2.value, $2.length); }
327┊         | extern_name_list_body ',' string
328┊             { script_add_extern(closure, $3.value, $3.length); }
329┊         ;
330┊

1123┊ /* A string can be either a STRING or a QUOTED_STRING.  Almost all the
1124┊    time we don't care, and we use this rule.  */
1125┊ string:
1126┊           STRING
1127┊             { $$ = $1; }
1128┊         | QUOTED_STRING
1129┊             { $$ = $1; }
1130┊         ;

binutils requires quoted strings for some kinds of symbols, e.g.:

• it accepts quoted symbols with @ in name:

$ echo 'EXTERN("__libc_start_main@@GLIBC_2.2.5")' > a.script
$ g++ a.script
/usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../../lib64/crt1.o: In function `_start':
(.text+0x20): undefined reference to `main'
collect2: error: ld returned 1 exit status

• but rejects them if unquoted:

$ echo 'EXTERN(__libc_start_main@@GLIBC_2.2.5)' > a.script
$ g++ a.script
a.script: file not recognized: File format not recognized
collect2: error: ld returned 1 exit status

To maintain compatibility with existing linker scripts support quoted strings in lld as well.

[ruiu]

LGTM

[luciang]

Update test/ELF/undefined-opt.s to verify quoted symbol correctly unquoted

[smeenai]

Do you need someone to commit this for you?

[luciang]

Yes, I need help committing this.
Thank you very much! :)

[smeenai]

I'll give @ruiu a bit to commit this, since he reviewed it, otherwise I can take care of that for you.

lld/trunk/ELF/ScriptParser.cpp

void ScriptParser::readEntry() {
   if (Config->Entry.empty())
     Config->Entry = Tok;
   expect(")");
 }
 
 void ScriptParser::readExtern() {
   expect("(");
   while (!errorCount() && !consume(")"))
-    Config->Undefined.push_back(next());
+    Config->Undefined.push_back(unquote(next()));
 }
 
 void ScriptParser::readGroup() {
   bool Orig = InputFile::IsInGroup;
   InputFile::IsInGroup = true;
   readInput();
   InputFile::IsInGroup = Orig;
   if (!Orig)

lld/trunk/test/ELF/linkerscript/linkerscript.s

 # REQUIRES: x86
 # RUN: llvm-mc -filetype=obj -triple=x86_64-unknown-linux %s -o %t
 # RUN: llvm-mc -filetype=obj -triple=x86_64-unknown-linux \
 # RUN:   %p/Inputs/libsearch-st.s -o %t2.o
 
-# RUN: echo "EXTERN( undef undef2 )" > %t.script
+# RUN: echo "EXTERN( undef undef2 \"undef3\" \"undef4@@other\")" > %t.script
 # RUN: ld.lld %t -o %t2 %t.script
 # RUN: llvm-readobj %t2 > /dev/null
 
 # RUN: echo "OUTPUT_FORMAT(elf64-x86-64) /*/*/ GROUP(\"%t\" )" > %t.script
 # RUN: ld.lld -o %t2 %t.script
 # RUN: llvm-readobj %t2 > /dev/null
 
 # RUN: rm -f %t.out

lld/trunk/test/ELF/undefined-opt.s

 # RUN: llvm-readobj --symbols %t3 | FileCheck --check-prefix=TWO-UNDEFINED %s
 # TWO-UNDEFINED: Symbols [
 # TWO-UNDEFINED: Name: abs
 # TWO-UNDEFINED: Name: big
 # TWO-UNDEFINED: Name: bar
 # TWO-UNDEFINED: Name: zed
 # TWO-UNDEFINED: ]
 # Now the same logic but linker script is used to set undefines
-# RUN: echo "EXTERN( bar abs )" > %t.script
+# RUN: echo "EXTERN( bar \"abs\" )" > %t.script
 # RUN: ld.lld -o %t3 %t.o %tar.a %t.script
 # RUN: llvm-readobj --symbols %t3 | FileCheck --check-prefix=TWO-UNDEFINED %s
 
 # Added undefined symbol may be left undefined without error, but
 # shouldn't show up in the dynamic table.
 # RUN: ld.lld -shared -o %t4 %t.o %tar.a -u unknown
 # RUN: llvm-readobj --dyn-symbols %t4 | \
 # RUN:     FileCheck --check-prefix=UNK-UNDEFINED-SO %s