This is the mail archive of the
mailing list for the GCC project.
Re: stormy16: limit SI reload regs
> X-Original-To: firstname.lastname@example.org
> Date: Wed, 30 Apr 2003 21:12:29 -0400
> From: DJ Delorie <email@example.com>
> Cc: firstname.lastname@example.org
> X-OriginalArrivalTime: 01 May 2003 01:12:05.0562 (UTC) FILETIME=[ADF9A1A0:01C30F7E]
> > This was in response to a different change to the xstormy16 backend,
> > not to a change in reload.
> Ok, I'm confused.
Maybe it'd be better if I explained the basic objection:
* You have found a bug in reload.
* You are proposing patches that try to work around the reload bug in
* I would prefer that you fix reload.
> So if I do fix reload to honor HARD_REGNO_NREGS when considering
> candidates, what assurances do I have that nowhere else in gcc will
> choose that register?
You fix those places too, in the event that you find any.
> And if it's safe to just fix reload, then
> what's wrong with just telling reload to use a different class?
Because that leaves reload broken for the next port to stumble over.
- Geoffrey Keating <email@example.com>