Crazy question regarding refresh and X
Mike Klinke
lsomike at futzin.com
Sun Sep 12 03:09:10 UTC 2004
On Saturday 11 September 2004 21:46, Jim Cornette wrote:
> Mike Klinke wrote:
> > On Saturday 11 September 2004 19:17, Jim Cornette wrote:
> >>How would I know if DRI is active or not! Below is the output
> >>from lsmod. Is there another program that reveals more info
> >>related to the X server related problems?
> >
> > I use the "glxinfo" utility where the first three lines are:
> >
> >
> > 21:01:32 # glxinfo
> > name of display: :0.0
> > display: :0 screen: 0
> > direct rendering: Yes
> >
> >
> > for example.
> >
> > Regards, Mike Klinke
>
> It appears that drm is answered no for both 16 and 24 depth
> settings. I set the upgraded FC3T2 version to 16 before launching
> and it gave me the exact symptom, a refresh problem like the
> fresh installation displays at the 16 depth.
> I added to the below report. I am not sure if this can now be
> determined to be an error with using a depth too low for the
> computer to work correctly.
>
> What are some thoughts regarding this setting that is not set
> high enough? Should this be a bug marked against
> system-config-display or xorg-x11?
>
> Jim
>
> https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=131934
It's my understanding that the i810 driver cannot run Direct
Rendering mode with bit depths greater than 16. My laptop seems to
confirm this as when I change from the "Millions of Colors" to the
"Thousands of Colors" choices offered by the gnome display
properties the Direct Rendering Mode is always enabled with the 16
bit selection (Thousands). In additon, the operation of X is MUCH
more responsive when using 16 bit. In either case, 16 or 24 bit
operation, (which corresponds to DRI=yes, DRI=no respectively via
gnome's automatic configuration) I don't see the same symptoms that
you are seeing with the bad refresh of the screen. I have yet to
see my box operate in a 16 bit/DRI=no mode and perhaps if I can get
it into that state it'll reflect what you're seeing.
Regards, Mike Klinke
More information about the fedora-test-list
mailing list