It has been pointed out to me by Andrew and Fernando Pablo Lopez-Lezcano that we do a lot of sleeping in code that probably should reschedule rather than sleep,radeon_do_wait_for_idle being a prime example, this has a DRM_UDELAY(1), this messes up audio latencys,
Now I'm not sure that wholesale replacing these with a conditional schedule is correct, but in most cases it looks like there should be no issues doing this, theses functions are all usually called from a process context,
My main worry is what to do on BSD about this?, should I just define a DRM_SCHED( delay ) and on Linux do a schedule and do a delay on BSD until someone steps up to figure it out?
That's been the traditional approach - filp maps to pid on BSD, or at least it did initially. Doing the same thing with delays should be fine.
What you will notice though is a performance cliff when you start to hit the delay occasionally. Or at least that was the case on early 2.4 when I looked at this last.
Keith
-------------------------------------------------------
This SF.Net email is sponsored by: Oracle 10g
Get certified on the hottest thing ever to hit the market... Oracle 10g. Take an Oracle 10g class now, and we'll give you the exam FREE.
http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click
--
_______________________________________________
Dri-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/dri-devel
