Re: post-reload crossjump tweak

[Jan Hubicka <jh at suse dot cz>]

> > On Thu, Apr 17, 2003 at 06:41:19PM +0200, Jan Hubicka wrote:
> > > That is wrong because R1 is clobbered at the moment.  What to do here?
> > > Only solution I can think of is to add instruction doing both call and
> > > restore at once and splitting it post-reload.
> > 
> > This is *exactly* why ia64 and pa64 have moved to post-reload
> > splitting of call instructions.
> 
> Hmm, to fully obsolette it we would probably need something like
> backward copy propagation or my register coalescing patch in order to
> elliminate the copis.  I think it is not worthwhile so I will simply
> kill the hunk as it is dead now and having dead and ugly code in the
> tree is not good idea.

Something got wrong with my mailer.  This mail is reffering to the
dropping of hunk dealing with the REG_EQUIV notes in crossjumping.
> 
> Honza
> > 
> > 
> > r~