Subject: Re: possible regression with Xorg radeon drivers?
To: Joerg Sonnenberger <firstname.lastname@example.org>
From: George Michaelson <email@example.com>
Date: 09/06/2006 23:17:37
On Tue, 05 Sep 2006 09:23:47 +0200
Joerg Sonnenberger <firstname.lastname@example.org> wrote:
> On Tue, Sep 05, 2006 at 02:22:36PM +1000, George Michaelson wrote:
> > mplayer is finding it very hard to 'keep up' with any video, any
> > format. trivial mouse/keyboard interactions cause it huge stutter.
> > Everything else is fine (in fact, better than fine: the radeon
> > multi-head stuff is just lovely. I really like Xorg)
> Can you also try to increase the SysV-IPC shmsize? That's another
> possible source of issues. Can you also what the CPU usage of Xorg and
without having done the IPC shmsize increase, I can say that there is
not a stunningly obvious cpu usage spike for either Xorg or mplayer.
I'm not sure how I can best prove this. top does not seem appropriate.
the impact is felt when you do keyboard work on mplayer. the display
goes bi-stable over 2 images (eg for OSD display) and then stabilizes,
then when you keypress again it lags for 5seconds, goes bi-stable, and
then stabilizes again.
so, another theory is that this is actually about wscons/wskbd and
Xorg, not about the XV path. Or, the effects of kbd/mouse interrupt