[PATCH] D30419: [ELF] - Define __bss_start symbol.

[George Rimar via llvm-commits]

>>> +  // __bss_start should point to the first .bss.* output section.
>>> +  for (OutputSection *OS : OutputSections) {
>>> +    bool IsBSS = (In<ELFT>::Common && In<ELFT>::Common->OutSec == OS) ||
>>> +                 OS == Out::Bss || OS == Out::BssRelRo;
>>
>>You can just check if the type is NOBITS, no?
>
>I do not think I can. That would catch .tbss first.
> __bss_start is not documented, so I am refering to BFD imlementation,
>it default script places __bss_start before all .bss and .tbss is listed explicitly before:
>
>  .tbss           : { *(.tbss .tbss.* .gnu.linkonce.tb.*) *(.tcommon) }
>.....
>  . = .;
>  __bss_start = .;
>  .bss            :
>  {
>   *(.dynbss)
>   *(.bss .bss.* .gnu.linkonce.b.*)
>   *(COMMON)
>   /* Align here to ensure that the .bss section occupies space up to
>      _end.  Align after .bss to ensure correct alignment even if the
>      .bss section disappears because there are no input sections.
>      FIXME: Why do we need it? When there is no .bss section, we don't
>      pad the .data section.  */
>   . = ALIGN(. != 0 ? 64 / 8 : 1);
>  }
>
>.So to have the same behavior, I think I should do what I do.

So is it ok for commit with original check ?
>>> +    bool IsBSS = (In<ELFT>::Common && In<ELFT>::Common->OutSec == OS) ||
>>> +                 OS == Out::Bss || OS == Out::BssRelRo;

George.