Red Hat Bugzilla – Bug 80423
(savage mx) video server not initialized properly upon cold boot on Toshiba 2805 laptop
Last modified: 2008-01-17 12:49:42 EST
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2) Gecko/20021126
Description of problem:
When I do a cold boot into RedHat Linux (Phoebe) and the X starts, the image on
the screen is garbled: the background is dark and only the mouse pointer is
shown. The mouse pointer leaves persistent trails when moved.
When I boot first in Debian (which uses XFree 4.2.1) and then reboot into RedHat
the greeter appears fine and I can log in into an user account.
The relevant section from /etc/X11/XF86Config:
VendorName "Videocard vendor"
BoardName "S3 Savage/MX"
Modes "1024x768" "800x600" "640x480"
It happens even if I change the color depth to 16.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Do an install, choose graphical login
2. Shut down the computer.
3. Start up the computer.
Actual Results: Screen is garbled.
Expected Results: Screen properly displays greeter.
Please attach the complete X server log file and config file, uncompressed,
using the link below.
Created attachment 88929 [details]
XFree86 config file generated by the installer
Created attachment 88931 [details]
XFree86 startup log
I had started up RedHat both cold and warm (reboot from Debian). The
/var/log/XFree86.0.log files are identical except for the time stamp.
XFree86-Servers does not exist, it is "XFree86" component. Changing component.
Please try adding the "noaccel" option to the device section of your
XFree86 config file. If that makes the problem go away, we will refine
our troubleshooting and narrow it down.
Adding the noaccel option makes the problem go away.
Ok, now remove the noaccel option, and instead, try using the XaaNoxxxxxx
options that are listed on the XF86Config manpage. Try them one at a time
to see if any one option makes the problem go away. If not, try 2 at
a time. If you can narrow down a minimum number of options which resolve
the problem, please update the report and I will let the driver author
know where the problems lay for this hardware in hopes he can fix it for
Huh, you are kidding, right? 18 * 18 combinations to test?
Oh, well, I have tried two combinations and here are the results:
gdm greeter screen has no images, only an edit box (which has whiter
background than the rest of the screen).
GNOME desktop appears fine
Mozilla is fine, scrolling is slow
Evolution is fine
gdm greeter: normal
GNOME desktop: normal
Evolution: all the menus are transparent, displaying what is actually
Mozilla: scrolling is all messed up but the menus are ok
Before I overwrite the 8.0 install on my S3 Savage/MX laptop with Phoebe: I've
got it working with XFree86-188.8.131.52-0.20021018.1, without any options set or
reset in XF86Config.
For gdm to display the login screen properly:
For the gdm screen to be cleared before the ssh-ask-password window is opened,
and for the gnome start-up screen to be displayed properly:
For gnome-terminal contents to be displayed properly, especially when the window
is moved or partially covered/uncovered:
For GNU Emacs menus and xchat gray-on-black contents to display correctly:
For the border of the Back/Next/Cancel buttons in evolution's start-up screen to
However, even with noaccel, the traditional X cursor is displayed after login,
and even in the login screen after a switch to a text VC and back, right in the
center of the screen, atop of everything, and it remains there until X is
terminated, without anything ever being able to get me rid of it. Suggestions?
Use KDE. Seriously, the problem occurs only in Gnome.
To me, this only implies that there's room for improvement in KDE, for it to
take advantage of acceleration in cases that it doesn't :-)
I get the problem with KDE too, the fix does work though.
Disabling gpm also seems to get rid of the X cursor for me...
I'm having a similar problem with an "S3 Savage/MX" adapter during
installation... X seems to start correctly, but only the mouse cursor seems to
be intact... the background stays garbled with the remains of what the BIOS
displayed before linux was booted. As a result graphical install is not possible...
I tried linux lowres and linux resolution=1024x768 with no significant changes
and certainly without success.
I should mention that what I posted above applies to Phoebe - BETA4 8.0.93
I'm also having problems with X after installation - GDM arrears no-so-garbled
at first, but with strange redish borders. When logging in, the screen looks
like a TV with week reception. navigatable, but very very noisy and butt ugly!
i have tried with Option "noaccel" with no change in behaviour
I've pinged the upstream driver maintainer to see if he might have a fix
for this problem on Savage/MX in current CVS XFree86.
Worst case, if we don't find a fix before our final release, I'll change the
default driver for this one chip to "vesa" so that there is working video
out of the box.
Can someone test "vesa" out in the mean time and provide comments back on
wether it works or not?
The vesa driver seems to work just fine here!
Vesa works fine here with IBM T21 (S3 Inc. 86C270-294 Savage/IX-MV (rev 13)) and
the performance is at least tolerable.
Alexandres list of XaaNo... options also makes it work, only window moving in
painful because of the redraw slowness, "shadowfb" helps somewhat in that case.
John Murga says:
> Disabling gpm also seems to get rid of the X cursor for me...
This is no longer necessary in Phoebe2.
The need for disabling a number of acceleration options is still there, though.
I believe this is a duplicate of bug #80346
Could everyone please examine that bug report, and the ones duped against
it, and comment on it. Please test the driver suggested in the report
as well and let me know if it resolves the issues for you all.
In't very likeli that this report and 80346 are regarding the same issue.
However I'm not experiencing quite the same problems, it seems, the only
workaround that has worked for me, has been using the vesa driver.
I'v also not been able to successfully have the updated driver installed for
test and so i can't see if that fixes my problems as well, until the source for
that module can be provided. (I'm getting unresolved symbol error and X aborts
As far as I'm concerned, this bug is a duplicate of bug #80346...
I have installed the suggested driver and it eliminated all the display problems
that I was having.
*** This bug has been marked as a duplicate of 80346 ***
Changed to 'CLOSED' state since 'RESOLVED' has been deprecated.