another Xorg problem: VESA(0): Cannot read int vect

[risto3]

The update patch is now committed upstream. Is there any reason not to cherry-pick an respin?

[curaga]

Yeah, the patch can now be applied to xorg-server. Juanito, do you have time?

http://cgit.freedesktop.org/xorg/xserver/patch/?id=0a78b599b34cc8b5fe6fe82f90e90234e8ab7a56

[Juanito]

I should be able to get around to it in a couple of days time

[Juanito]

patched xorg-server posted

[curaga]

Confirming that vesa is fixed.

[risto3]

I can confirm that I can get Xorg up, but I seem to notice a considerable performance regression compared to the previous version. 

The easiest way to experience it is with a couple of windows up ('top' is a good one) and moving a window around with the mouse... 

there is an extreme lag (sort of like ants scrawling to keep up:-) and noticeable Xorg cpu utilisation.

also, it is unclear where XkbLayout and XkbModel should be put as now there is a
/usr/local/share/X11/xorg.conf.d/10-evdev.conf already.

manually doing 'setxkbmap -layout fr -model pc105'  works fine and my logitech setup works as did previously... including the € and numpad keys (yeah)

[risto3]

also, it is unclear where XkbLayout and XkbModel should be put as now there is a
/usr/local/share/X11/xorg.conf.d/10-evdev.conf already.

copying the above file to /etc/X11/xorg.conf.d/  adding the options mentioned to it and adding the file to /opt/.filetool.lst as well seemed to get over this part.   Perhaps something for the wiki?

[risto3]

I can confirm that I can get Xorg up, but I seem to notice a considerable performance regression compared to the previous version. 

The easiest way to experience it is with a couple of windows up ('top' is a good one) and moving a window around with the mouse... 

there is an extreme lag (sort of like ants scrawling to keep up:-) and noticeable Xorg cpu utilisation.

actually quite psychedelic in VNC ... moving a terminal window e.g. laterally gives impression that
the side of the window buts, and partially disappears, into a transparent marshmallow... until it catches up and takes full form again.  >.,.<

[curaga]

Yeah, I can reproduce: Xorg + vesa maxes one core, while the default Xvesa gets 20-30%. But I'm not sure we can do anything about a regression upstream, as reverting would hurt Intel users as well as the Radeon user who requested this update in the first place.

[risto3]

OK, I filed https://bugs.freedesktop.org/show_bug.cgi?id=89747
Might be useful if everybody who experiences this chirps in.