[Bug dynamic-link/14538] Complicated x86-64 elf_machine_dynamic/elf_machine_load_address

["hjl.tools at gmail dot com" <sourceware-bugzilla@sourceware.org>]

http://sourceware.org/bugzilla/show_bug.cgi?id=14538

--- Comment #1 from H.J. Lu <hjl.tools at gmail dot com> 2012-09-01 05:11:21 UTC ---
The new version:

/* Return the link-time address of _DYNAMIC.  Conveniently, this is the
   first element of the GOT.  This must be inlined in a function which
   uses global data.  */
static inline ElfW(Addr) __attribute__ ((unused))
elf_machine_dynamic (void)
{
  /* This produces a PC32 reloc that resolves to zero at link time, so in
     fact just loads from the GOT register directly.  By doing it without
     an asm we can let the compiler choose any register.  */
  extern const ElfW(Addr) _GLOBAL_OFFSET_TABLE_[] attribute_hidden;
  return _GLOBAL_OFFSET_TABLE_[0];
}


/* Return the run-time load address of the shared object.  */
static inline ElfW(Addr) __attribute__ ((unused))
elf_machine_load_address (void)
{
  /* Compute the difference between the runtime address of _DYNAMIC as seen
     by a PC32 reference, and the link-time address found in the special
     unrelocated first GOT entry.  */
  extern ElfW(Dyn) bygotoff[] asm ("_DYNAMIC") attribute_hidden;
  return (ElfW(Addr)) &bygotoff - elf_machine_dynamic ();
}

It generates:

        movq    _DYNAMIC(%rip), %rax 
        movq    %rdx, start_time(%rip)
        leaq    _DYNAMIC(%rip), %rdx 
        movq    %rdx, %r13 
        subq    _GLOBAL_OFFSET_TABLE_(%rip), %r13 
        testq   %rax, %rax 
        movq    %rdx, 2472+_rtld_local(%rip)
        movq    %r13, 2456+_rtld_local(%rip)
        je      .L993

instead of

#APP
# 75 "../sysdeps/x86_64/dl-machine.h" 1
        lea _dl_start(%rip), %r13 
        sub 1f(%rip), %r13 
        .section        .data.rel.ro
1:      .quad _dl_start
        .previous

# 0 "" 2
#NO_APP
        movq    %rdx, start_time(%rip)
        movq    %r13, %rdx 
        addq    _DYNAMIC@GOTPCREL(%rip), %rdx 
        movq    %r13, 2456+_rtld_local(%rip)
        movq    (%rdx), %rax 
        movq    %rdx, 2472+_rtld_local(%rip)
        testq   %rax, %rax
        je      .L994

The new one avoids one load and reduces one GOT entry.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.