Re: CONSTANT_P_RTX and PowerPC wrong integers and fpmem alignment

[Michael Meissner <meissner at cygnus dot com>]

On Tue, Jan 12, 1999 at 02:44:13PM -0500, David Edelsohn wrote:
> 	While the code indirectly aligned fpmem offset on AIX, it ensured
> that fpmem offset ALWAYS was misaligned for the SVR4 ABI (which uses a
> slightly different ordering).  New alignment code recently was added to
> ensure fpmem alignment directly which seems to make the indirect alignment
> redundant (and always would add an additional stack word on SVR4).  I now
> have removed the former code.  If this causes any problems (eABI?), we can
> restore the code but need to better document and correctly doument its
> purpose.

Great!  Now hopefully both camps will be happy.

BTW, we have a bunch of PPC patches coming down the pike from some development
work we've been doing (750 support, some new optimizations, etc.).

I also have a rather large patch that I'm in the middle of to add alternatives
for all of the . instructions if we couldn't allocate cr0 for the task that it
would be helpful to have some other eyes stare at (the motivation is to avoid
register spill errors if cr0 is used by asm statements, and to avoid doing
mcrf's if we can, since these serialize the high end machines).

-- 
Michael Meissner, Cygnus Solutions (Massachusetts office)
4th floor, 955 Massachusetts Avenue, Cambridge, MA 02139, USA
meissner@cygnus.com,	617-354-5416 (office),	617-354-7161 (fax)