Re: c4x - TMS320C30-40

[Joel Sherrill <joel dot sherrill at OARcorp dot com>]

"Svein E. Seldal" wrote:
> 
> Joel Sherrill wrote:
> > I verified that the config.sub patch produces the results I think we
> > all want.
> 
> Then I will start the commit process. I cant promise anything on
> reaching binutils-2.14 or the gcc-3.3.1, but I'll certainly try.
> 
> BTW! For rtems the most critical part is the config.sub in the /cvs/src
> repository, not neccessarily the one in gcc, right?

Logically RTEMS is just another package that periodically grabs a
config.sub
and may patch it as needed. 

Honestly, I would rather it be right in binutils and gcc releases since
otherwise we will be stuck shipping patches for them.  binutils has been
patch free for RTEMS for a few releases now.

> > With all the time in config.sub, I think there are extra '*'s in a
> > couple
> > of other places that shouldn't have them.  It looks to me that c54x*,
> > c55x*, and c6x* have the same mistake c4x did.  Could someone double
> > check me on this?
> 
> Yes, I am aware of this. This is the reason for the (ti)c4x having this
> bug. I copied the configuration from the c54x* configuration, assuming
> that it were correct. This is probably the same thing the c55x and the
> c6x has done as well.
> 
> However, I dont think I want to change them, because there have not been
> any protests coming from the communities. Please remember that when you
> change the toplevel config, you will need to change the sub-directory
> config's as well. So I leave it for the maintainers to change their
> config, if this turns out to be a problem.

I assumed this would be the decision since I wouldn't want to tinker
with
something like that either. :)

> Regards,
> Svein

-- 
Joel Sherrill, Ph.D.             Director of Research & Development
joel@OARcorp.com                 On-Line Applications Research
Ask me about RTEMS: a free RTOS  Huntsville AL 35805
Support Available                (256) 722-9985