This is the mail archive of the
gdb@sourceware.org
mailing list for the GDB project.
Re: help required : floating point for i386
- From: paawan oza <paawan1982 at yahoo dot com>
- To: Hui Zhu <teawater at gmail dot com>
- Cc: gdb ml <gdb at sourceware dot org>
- Date: Sun, 17 May 2009 10:44:21 -0700 (PDT)
- Subject: Re: help required : floating point for i386
Hi,
I think I found the solution.
register_buffer (const struct regcache *regcache, int regnum)
{
return regcache->registers + regcache->descr->register_offset[regnum];
}
where regcache->registers array has different order of storing FPU registers. (not in order st0,st1,st2.....I suppose.)
so we need to pass correct register number.
I am working on that, once it is done...we should be able to record floating point registers : )
Regards,
Oza.
--- On Sun, 5/17/09, paawan oza <paawan1982@yahoo.com> wrote:
> From: paawan oza <paawan1982@yahoo.com>
> Subject: help required : floating point for i386
> To: "Hui Zhu" <teawater@gmail.com>
> Cc: "gdb ml" <gdb@sourceware.org>
> Date: Sunday, May 17, 2009, 2:22 PM
> Hi,
>
> I am working on i386-tdep.c for floating point.
>
> I try to do ;
> record_arch_list_add_reg (ir.regcache,I386_ST0_REGNUM);
> does not deterministically record st0 register
> (sometims it records st1 or sometimes st7)
> why is it so ?
>
> we collect fpx registers in following functions. do we need
> to do something about following function, so it returns
> correct register !!
> i387_supply_fxsave (regcache, -1, fpxregsetp);
>
> please help.
>
> Regards,
> Oza.
>
>
> ? ? ?
>