Re: [Dri-devel] DRM+ATI-Radeon-Mobility

[Michel <mi...@da...>]

On Thu, 2004-01-15 at 03:10, Wlodzimierz Kozlowski wrote: 
> I observe the following problem concerning, most probably,  radeon.o module 
> compiled from the kernel linux-2.4.23 source (generally later than 2.4.18):
> 
> An effort of launching a graphic application using DRI results in total 
> freezing of the keyboard, mouse, screen so that only switch off can be 
> done (glxinfo shows that DRI , AGP and all the rest are enabled).

Can you post the output of

LIBGL_DEBUG=verbose glxinfo

as well as any relevant parts of the X server log and kernel output?

> However , I have observed that 
> launching the graphic application using DRI with XFree86-4.1.0  under 
> kernel 2.4.18 (Linux kernel 2.4.18 from http://www.kernel.org)
> and subsequent warm restart to 2.4.23 removes this unrequired effect
> as long as the machine is switched on (subsequently, a number of WARM
> restarts can be done and this will not restore the freezing effect). 
>   I have found this has most probably nothing with APM or using 
> the EXT3 and occurs with  XFree86 subversions 4.1.0, 4.2.0 and 
> apparently does not depend on the XFree86 subversion number.

Have you tried 4.3 or newer though?

>   Assignment of RAM intervals appears to be the only difference between 
> log files [...]

What RAM intervals?


> (others report the problem preserves also with Linux kernel 2.6.0,
> e.g. 
> https://sourceforge.net/mailarchive/forum.php?thread_id=3713780&forum_id=6511

That URL doesn't work for me, but maybe that's just sf.net...


All in all, it sounds to me like newer DRMs fail to initialise something
they should, at least with older DRI userland.


-- 
Earthling Michel Dänzer      |     Debian (powerpc), X and DRI developer
Software libre enthusiast    |   http://svcs.affero.net/rm.php?r=daenzer