Red Hat Bugzilla – Bug 67723
X will start only once per boot on i810
Last modified: 2007-04-18 12:43:47 EDT
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.1a) Gecko/20020610
Description of problem:
After installing Red Hat Linux 7.3 from scratch on a i686 on a MB with the i810
video controller, XFree86 would start just fine. After logging in and then
logging out, X would no longer start.
Further investigation revealed that VT#7 appears to be still busy by the time X
attempts to restart. Even if I switch to init level 3 and leave it like that
for a while, modules i810 and agpgart are still busy and cannot be unloaded,
which seems to indicate something is still holding to them.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1.Install Red Hat Linux 7.3 with the following X settings:
xconfig --card "Intel 810" --videoram 16384 --hsync 30-66 --vsync 50-130
--resolution 1152x864 --depth 16 --startxonboot
2.Install all available errata:
4.X starts up correctly
5.init 3; sleep 60; init 5
Actual Results: X won't start any longer, and a reconfiguration will be
suggested, until the next reboot.
Expected Results: It used to work just fine on Red Hat Linux 7.2.
I'm attaching the XFree86 logs of the successful and the unsuccessful sessions.
Created attachment 63204 [details]
XFree86 log of a working session
Created attachment 63205 [details]
XFree86 log of the following X failure to start
Created attachment 63284 [details]
kernel oops that happens when a users attempts to logout
I am having the same problem. When users logout X fails to restart it self. It
looks like the driver is causing a kernel oops. I have attached a copy of the
oops from the syslog.
FWIW, I've switched to the following configuration to work around the problem:
# Screen 0 "Screen0" 0 0
Screen 0 "Screen1" 0 0
# no known options
VendorName "VESA driver (generic)"
BoardName "VESA driver (generic)"
Modes "1152x864" "1024x768" "800x600" "640x480"
The only problem that I have found with this fix is that i'm limited to a
resolution of 800x600
FWIW, I get the same kernel oops, I just had missed it because I was looking for
something in /var/log/messages at about the time X restarted. D'oh.
Anyway, this was the hint I needed to figure the problem probably had to do with
dri. In fact, commenting out the `Load "dri"' fixed it for now.
Arjan, is this the problem which was just fixed in i8x0 DRM?
If so, please let me know, as I think I can close this then, and dupe
many other bug reports against it.
yup the kernel oops is the bug fixed...
*** This bug has been marked as a duplicate of 67978 ***