答复: 答复: [COMMIT] romcc 0.63 and some fixes...

[YhLu]

I see. The code is getting larger.

Also the reboot in OS can not work.

Regards

YH

-----邮件原件-----
发件人: [e-mail] [mailto:[e-mail]] 
发送时间: 2004年6月7日 13:29
收件人: YhLu; LinuxBIOS
主题: Re: 答复: [COMMIT] romcc 0.63 and some fixes...

[e-mail] (Eric W. Biederman) writes:
> YhLu  writes:
> > Just sent out the core.
>
> Thanks.  I have done the basic tracking.  And it is clearly a code
> bug.

With the bug fix in place I am seeing what looks like another bug.

For the moment you can avoid this by removing the printo_debug
from generate_row.  That still runs out of registers but the
compiler does not crash.

I would dearly love to see a version of this code that works
when you do broad changes, like you did.  At the moment the code does
work well enough that if you change a couple of functions at a time,
you get some benefit.

Eric

答复: 答复: [COMMIT] romcc 0.63 and some fixes...

YhLu  writes:
> I see. The code is getting larger.

Not quite.  There is less generated code, but it requires
extra registers to hold the return addresses so the register
pressure is higher.

Some of that extra register pressure may be artificial and
go away when this bug is fixed.  Some of it may result from
poor choices of the register allocator, and may be remedied
by enhancing the register allocator.  There is hope of
remedies as the kind of code presented to the register
allocator is very different when code is not inlined.
> Also the reboot in OS can not work.

Hmm.  At least with reboot=hard that should work on the Opterons.
I have not tested that lately though.

Eric