Re: Xorg woes and partial tips

[tlaronde%polynum.com@localhost]

Hello,

On Mon, Sep 20, 2010 at 01:46:48PM +0300, markucz%gmail.com@localhost wrote:
> > Exiting from X/radeon, the VT is fading. If you relaunch (xinit(1)),
> > every time the VT fade more untill you can see strictly nothing.
> > [...]
> > Furthermore, from time to time, the input on the display freezes (no
> > keyboard, no pointer). Only solution : remote connection, say ssh(1),
> > and you have to kill the programs including X. But this is not enough,
> > you must switch to root to remove /tmp/.X* or you will not be able to
> > relaunch X.
> I had the same problems with 3.1.1, using the then-current pkgsrc. There was
> no way to remedy radeon (RV200 if I recall correctly) blooming. It might be a
> long-standing bug in Xorg or, IMO unlikely, a kernel issue. Eventually I
> uninstalled Xorg because I feared I would let the magic smoke out. Never
> tried the vesa trick. Is DRI enabled in your config?

Yes but it is in fact enable by default. The dependencies between DRI,
DRI2, DRM, XAA | EXA etc. etc. are daunting. And to have "some
acceleration" is part of the aim of the game.

If I remember correctly, I switched to Xorg because some third parties
applications did not compile anymore with an "old" version of X. All in
all, even for my own map/CAD software programmed to be light-weight,
VESA is no problem. The problem is primarily with a web browser.

Since firefox is a nightmare to compile (when it succeeds and one of the
terabytes dependencies doesn't fail), I use Opera but I don't know if
the Linux emulation can now cause problems (indirectly with kernel
stuff used by hal and evdev). But my problems happen even
on sessions without Opera running.

One further note: it seems that (this may be linux evdev combining
keyboard and mouse) there is an implicit link between the keyboard and
the mouse. I have  a little less problems if, once the X session is
launched via xinit(1), I immediatly num-lock the keyboard, hence
activate it? When the cursor is grabbed or stored, I come back to
Ctl-NumLock to move the cursor with the num pad; in this case, the
link is explicitely set by myself... because I have no choice.
But it seems that, without setting the keyboard mouse driving,
there is an implicit one that I don't have a hand on.

On the "funny" side, this quote from the xorg.conf(5):

.SH "VIDEOADAPTOR SECTION"
Nobody wants to say how this works.
Maybe nobody knows ...

Perhaps the comment could be added to all the other sections. Sigh...

Best regards,
-- 
        Thierry Laronde <tlaronde +AT+ polynum +dot+ com>
                      http://www.kergis.com/
Key fingerprint = 0FF7 E906 FBAF FE95 FD89  250D 52B1 AE95 6006 F40C