------- Comment #3 from hp at gcc dot gnu dot org 2009-08-17 16:32 -------
(In reply to comment #2)
> Although I think assigning r8 to pseudo 47 could result in wrong reload
> behaviour because r8 is used as a frame pointer even before the RA. It looks
> very strange to me that r8 used as a frame pointer is *not a fixed* register.
There's nothing wrong with that in a port, besides trigging reload bugs, that
is. ...and now that you mention it, I guess it might inconvenience the
register allocator.
> If I wrote the port, I'd rather use a virtual register for the frame pointer
> and eliminate it into hard frame pointer register (r8) or to stack pointer as
> all other ports do.
Not all ports, e.g. the m68k port does the same (A6). IMHO, requiring to
*manually* allocate a virtual frame-pointer in a port only to eliminate it to a
hard register is a wart; it's just complicating the port where middle-end gcc
could do the work.
> This probably would solve the problem.
I'd say it would cover it up, not solve it.
But thanks for looking, I'll have another look myself and try and see if I can
fix the actual reload bug. I'm not ruling out eventually using a virtual frame
pointer, albeit pragmatically a fix, it just doesn't strike me as TRT.
--
hp at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
AssignedTo|unassigned at gcc dot gnu |hp at gcc dot gnu dot org
|dot org |
Status|UNCONFIRMED |ASSIGNED
Ever Confirmed|0 |1
Last reconfirmed|0000-00-00 00:00:00 |2009-08-17 16:32:08
date| |
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41085
