bfin-gdb

[Ralf Corsepius]

On Thu, 2008-05-01 at 08:02 -0500, Joel Sherrill wrote:
> Ralf Corsepius wrote:
> > On Thu, 2008-05-01 at 11:40 +0200, Ralf Corsepius wrote:
> >   
> >> Hi,
> >>
> >> Is anybody actively working with bfin targets and using bfin-gdb?
> >>     
> No one on the SC AFAIK but there are users.  One is in the
> class this week and I know of others.
> >> ATM, rtems's bfin-gdb is still based on gdb-6.5 (very obsolete) with its
> >> bfin-specific patches actually being a bunch of hacks, but a real port
> >> and seemingly doesn't have an active upstream nor an active user-base.
> >>
> >> That said, I'd seriously advise those of you working with bfin targets
> >> to try porting RTEMS's bfin-gdb to gdb-6.8 or (better) to have support
> >> for bfin integrated into upstream gdb.
> >>
> >> Otherwise, I fear, we will not be able to discontinue supporting
> >>     
> >
> > Grr, sorry, this should have been ....
> >   ... we will not be able to continue supporting ...
> >
> >   
> >> bfin-gdb rather soon (bfin-binutils and bfin-gcc are part of upstream
> >> FSF binutils rsp. GCC)
> >>     
> >
> >   
> I wonder what the issue is with bfin-gdb.  Do you know who
> wrote the code? 
> 
> It may just require a friendly nudge to get it merged if the
> paperwork is in place. 
The patch contains this copyright notices by Analog Devices, is mostly
derived from original FSF gdb code and appears to be GPLv2'ed.

The patch itself once had been been submitted by Alain Schaeffer.

The problem with this patch are more of a technical nature. It contains
questionable spots (say: hacks) which I would expect upstream FSF-gdb to
rejected this patch.

>From a mere RTEMS perspective, the problem is gdb having changed to an
extend, porting (and hopefully cleaning up) the code for current gdb
would be tedious and would require somebody with actual hardware to  be
able to test it.

Ralf