Re: The problem with stabs and sign extension

["H . J . Lu" <hjl at lucon dot org>]

On Wed, Aug 08, 2001 at 02:12:07PM -0700, Daniel Jacobowitz wrote:
> The problem seems, to my inexperienced eye, to be in stabsread.c:
> 
> #define INTERNALIZE_SYMBOL(intern, extern, abfd)                        \
>   {                                                                     \
>     (intern).n_type = bfd_h_get_8 (abfd, (extern)->e_type);             \
>     (intern).n_strx = bfd_h_get_32 (abfd, (extern)->e_strx);            \
>     (intern).n_desc = bfd_h_get_16 (abfd, (extern)->e_desc);            \
>     (intern).n_value = bfd_h_get_32 (abfd, (extern)->e_value);          \
>   }
> 
> n_value is a CORE_ADDR.  bfd_h_get_32 returns a bfd_vma, without doing sign
> extension.  For MIPS, we want sign extension to have happened here.  Right? 
> It does if we're reading mdebug in (because ECOFF_SIGNED_32 is defined in
> BFD).
> 
> On the other hand, I'm sure other targets don't want sign extension here. 
> How should we handle this?

You can do

#define INTERNALIZE_SYMBOL(intern, extern, abfd)                        \
  {                                                                     \
    (intern).n_type = bfd_h_get_8 (abfd, (extern)->e_type);             \
    (intern).n_strx = bfd_h_get_32 (abfd, (extern)->e_strx);            \
    (intern).n_desc = bfd_h_get_16 (abfd, (extern)->e_desc);            \
    if (bfd_get_sign_extend_vma (abfd))                                 \
      (intern).n_value = bfd_h_get_signed_32 (abfd, (extern)->e_value); \
    else                                                                \
      (intern).n_value = bfd_h_get_32 (abfd, (extern)->e_value);        \
  }

But I don't think it is the main problem. From what I can see,
dbxread.c has no clue whatsoever about the sign extension. I
don't know how hard to fix it.


H.J.