m68k patch reviews

[Bernardo Innocenti]

On Saturday 23 August 2003 01:04, Peter Barada wrote:

> >If your versions of GCC haven't got too much exposure already,
> >we'd better rename your flavour to interrupt_handler. It all
> >depends on how much code depends on it.
>
> I think the fiddes patches had 'attribute(handler)' support in them.

I've just checked and in the patch set dated 20010415 it was already
called attribute((interrupt)).

> Please check the RTEMS patches to see what(if anything) was done there.

I've checked their patches here:

  ftp://ftp.rtems.com/pub/rtems/snapshots/c_tools/source/

There seems to be nothing about an interrupt attribute.


> Also ask on the ColdFire list at wildrice.com before you usuper
> attribute(interrupt) to be Amiga specific.

Well, you know them much better than me. I'll forward this thread to
them...


> Perhaps you could create a m68k-amiga configuration that can usurp
> interrupt(attribute) for Amiga's purpose and leave the rest of the
> m68k universe alone.

Let's do so: I'll repost the patch with the corrections requested by
Richard so I can get it approved. Meanwhile, if I don't get enough
complaints from the ColdFire ml people, I'll rename the attribute to
interrupt_handler. Otherwise, I'll check it in as is and leave to the
Amiga people the (quite trivial) task of making it work differently
on their target.

-- 
  // Bernardo Innocenti - Develer S.r.l., R&D dept.
\X/  http://www.develer.com/

Please don't send Word attachments - http://www.gnu.org/philosophy/no-word-attachments.html