Re: [patch] Speed up find_pc_section

[Ulrich Weigand <Ulrich.Weigand@de.ibm.com>]

Paul Pluzhnikov <ppluzhnikov@google.com> wrote on 08/17/2009 11:15:20 PM:

> On Mon, Aug 17, 2009 at 12:54 PM, Ulrich Weigand<uweigand@de.ibm.com> 
wrote:
> 
> > Section can partially overlap with overlays ...
> 
> Yes, the same thing also is happening on Darwin; some discussion here:
>   http://sourceware.org/ml/gdb-patches/2009-08/msg00201.html
> 
> The thing is: when the assumptions are violated, (AFAICT) 
find_pc_section
> couldn't have worked correctly before my patch, except by accident.

Which assumptions?   For overlay sections, find_pc_section would
delegate its work to find_pc_mapped_section anyway:

  s = find_pc_mapped_section (pc);
  if (s)
    return s;

so the rest of find_pc_section does not need to handle overlay sections 
...

It just shouldn't abort simply because overlays are present anywhere.

> What is the simplest way to test overlays?

The only target with overlays I know to test is the SPU.  (Apparently, 
d10v
and m32r also support overlays, but I don't know how to test those.)

I'd be happy to test patches on the SPU ...


Mit freundlichen Gruessen / Best Regards

Ulrich Weigand

-- 
  Dr. Ulrich Weigand | Phone: +49-7031/16-3727
  STSM, GNU compiler and toolchain for Linux on System z and Cell/B.E.
  IBM Deutschland Research & Development GmbH
  Vorsitzender des Aufsichtsrats: Martin Jetter | Geschäftsführung: Erich 
Baier
  Sitz der Gesellschaft: Böblingen | Registergericht: Amtsgericht 
Stuttgart, HRB 243294