[Intel-gfx] Problem using an Mesa based App with recent xorg/mesa/xf86-video-intel (loop?)

Jesse Barnes jbarnes at virtuousgeek.org
Tue Mar 2 20:50:05 CET 2010


On Tue, 2 Mar 2010 19:50:45 +0100
Florian Mickler <florian at mickler.org> wrote:

> On Tue, 2 Mar 2010 09:32:57 -0800
> Jamey Sharp <jamey at minilop.net> wrote:
> 
> > On Tue, Mar 2, 2010 at 4:37 AM, Stephan Raue
> > <mailinglists at openelec.tv> wrote:
> > > i have problems running an Application that depends on Mesa. It
> > > seems there is an loop after starting this App and before the GUI
> > > loads. I use
> > 
> > A loop? Or just a hang?
> > 
> > > #1  0xb5e47fb6 in *__GI___poll (fds=0xb5ec7ff4, nfds=1,
> > > timeout=-1) at ../sysdeps/unix/sysv/linux/poll.c:87
> > > #2  0xb5d5d268 in _xcb_conn_wait (c=0x9571ca0, cond=0xbf9dbbe0,
> > > vector=0x0, count=0x0) at xcb_conn.c:306
> > > #3  0xb5d5ee09 in xcb_wait_for_reply (c=0x9571ca0, request=30,
> > > e=0xbf9dbc6c) at xcb_in.c:390
> > 
> > Judging only by the stack trace, it looks like your application is
> > waiting for a reply that never comes. There are a variety of bugs
> > that can cause that. The sequence number (30) looks reasonable for
> > application startup, which rules out one or two things.
> > 
> > For the rest, if the Intel/Mesa folks don't have an answer, it would
> > help if you could get a trace of the X wire protocol using something
> > like Wireshark.
> > 
> > Jamey
> > _______________________________________________
> > xorg-devel mailing list
> > xorg-devel at lists.x.org
> > http://lists.x.org/mailman/listinfo/xorg-devel
> 
> 
> 
> i'm seeing the same with glxgears, see below for a gdb run...
> 
> if i login glxgears runs fine, until i start up some programs (licq,
> psi, twinkle, skype)... then it hangs sometimes and doesnt redraw the
> screen anymore. if i start it after that point it never draws...
> 
> seems to me, like some event goes missing somewhere... ? 
> 
> anyway, i'm on amd64 with intel from git, using the new event-driven
> invalidate stuff...
> 
> what debugging would be helpful here? dmesg doesn't show anything
> (abnormal) with kernel drm.debug=1 ...
> 
> cheers,
> Flo
> 
> p.s.: i verified with strace that it hangs on that poll... 

So the server is hanging when the client tries to get buffers?  Can you
see what it's doing at the time?

-- 
Jesse Barnes, Intel Open Source Technology Center



More information about the Intel-gfx mailing list