Bugzilla – Bug 146462
fbdev: Error about FBIOBLANK argument
Last modified: 2008-04-21 06:54:27 UTC
When an X server is spawned during installation (when starting the installer or when testing a display configuration), there are always errors about "invalid attribute FBIOBLANK" which are perhaps not there in the recent version of X.
Please attach the xorg log (/var/log/Xorg.0.0.log) and `hwinfo --gfx'.
I´m talking about the X server spawned by the installer for which there is no logfile. Also it happens independant of the graphics card. I think the X server is spawned with this argument but it doesn´t exist anymore in recent versions of X.
Stefan, please comment about that. Does the installation X-Server really create no logfile?
AFAIK these errors/warnings can be ignored and happened with earlier installations as well? What exactly does not work with your instllation?
I didn´t say something doesn´t work *g. I just report all errors/warning I see, because IMHO it´s good development practise to eliminate them all. Some day (with a new version of x) it won´t work anymore. Even if X does not quit with this it´s clearly an error, and why do you give those arguments to it?
Good point. Could you help in investigating this by using sax2 -r -m 0=fbdev and attaching /var/log/SaX2.log? Backup your /etc/X11/xorg.conf before.
Created attachment 65857 [details] sax.log
I tried it and attached the logfile "sax.log" (there was no sax2.log). I didn´t exit X before and testing the configuration in sax2 resulted in crazy screen corruption (3 screens at the same time *g) until I restarted. But I think this is not a problem here.
Thanks. Ok, it's only this line [...] (EE) FBDEV(0): FBIOBLANK: Invalid argument
There are some other warnings which should be fixed as well IMHO (tested with beta3). There are some warnings about two non-existing paths in the list of font paths (which don´t exist neither during installation nor when the system is installed) and one warning from the keymap compiler about <RALT> having two symbols (which should be one).
These can be ignored as well.
See comment #5. Do you always only fix bugs which lead to an unusable or crashing system? Perhaps a low priority would be the best choice for that bug.
Sure, the bugreport is already on low priority. ;-)
VESAfb has no blanking implemented. This the ioctl returns EINVAL which causes this (harmless) error message. Maybe it should be changed into an info.
Created attachment 72495 [details] change error to info message
Thanks. Egbert, do you agree to commit this upstream, so we can close this one as FIXED?
I'm not sure if I will commit this or another message that will tell the user that screen blanking isn't supported in the kernel.
> or another message that will tell the user that screen blanking isn't > supported in the kernel. I like this idea!
Assigning to Egbert.
I'll update our patch accordingly.
done. xorg-x11-server.changes: ------------------------------------------------------------------------ Sun Apr 20 05:17:28 CEST 2008 - sndirsch@suse.de - fbdevhw.diff * screen blanking not supported by vesafb of Linux kernel (bnc #146462)
Created attachment 209204 [details] fbdevhw.diff