[fglrx] crashes on VT switch

[Johannes Gijsbers]

About 33-50% of the time when I close the lid of my Asus A2500L, I get logged
out from X. When I reopen the lid, I'm back at the GDM login screen. As you can
imagine, this can be quite irritating when running download processes or playing
music. I'll be happy to provide any information you might need to debug this.

The /etc/default/acpi-support config file hasn't been changed from the
acpi-support 0.18 default.

[Thom May (thombot)]

What do you have in /var/log/acpid when this happens? Is htere anything
interesting in syslog?

[Johannes Gijsbers (johannes)]

I can not reproduce the bug right now (sometimes it'll work correctly for
stretches of time), but it's bound to pop up again soon, I think. So here are
excerpts from the night I reported the bugs.

I find this in syslog:

Feb 28 01:34:06 surfboy gdm[32269]: gdm_slave_xioerror_handler: Fatal X error -
Restarting :0

acpid from around the same time:

[Mon Feb 28 01:34:05 2005] received event "button/lid LID 00000080 00000028"
[Mon Feb 28 01:34:05 2005] notifying client 971[1000:1000]
[Mon Feb 28 01:34:05 2005] executing action "/etc/acpi/lid.sh"
[Mon Feb 28 01:34:05 2005] BEGIN HANDLER MESSAGES
xscreensaver-command: warning: $DISPLAY is not set: defaulting to ":0.0".
XIO: fatal IO error 104 (Connection reset by peer) on X server ":0.0"

      after 0 requests (0 known processed) with 0 events remaining.

[Mon Feb 28 01:34:05 2005] END HANDLER MESSAGES
[Mon Feb 28 01:34:05 2005] action exited with status 0
[Mon Feb 28 01:34:05 2005] completed event "button/lid LID 00000080 00000028"

2 minutes later in syslog:

Feb 28 01:35:47 surfboy gdm[12623]: gdm_slave_xioerror_handler: Fatal X error -
Restarting :0

and in acpid around that time:

[Mon Feb 28 01:35:51 2005] received event "button/lid LID 00000080 0000002a"
[Mon Feb 28 01:35:51 2005] notifying client 12908[1000:1000]
[Mon Feb 28 01:35:51 2005] client has disconnected
[Mon Feb 28 01:35:51 2005] executing action "/etc/acpi/lid.sh"
[Mon Feb 28 01:35:51 2005] BEGIN HANDLER MESSAGES
xscreensaver-command: warning: $DISPLAY is not set: defaulting to ":0.0".
Xlib: connection to ":0.0" refused by server

Xlib: Invalid MIT-MAGIC-COOKIE-1 key

xscreensaver-command: can't open display :0.0
[Mon Feb 28 01:35:53 2005] END HANDLER MESSAGES
[Mon Feb 28 01:35:53 2005] action exited with status 0
[Mon Feb 28 01:35:53 2005] completed event "button/lid LID 00000080 0000002a"

I'll try to get a more recent log when the bug shows up again. If you need the
full logs from that night, just ask.

[Johannes Gijsbers (johannes)]

Okay, it happened again. This time I immediately grabbed the relevant logs.

/var/log/acpid:

[Sat Mar 5 05:22:32 2005] received event "button/lid LID 00000080 00000001"
[Sat Mar 5 05:22:32 2005] notifying client 7996[1000:1000]
[Sat Mar 5 05:22:32 2005] executing action "/etc/acpi/lid.sh"
[Sat Mar 5 05:22:32 2005] BEGIN HANDLER MESSAGES
xscreensaver-command: throttled.

xscreensaver-command: activating and locking.

[Sat Mar 5 05:22:35 2005] END HANDLER MESSAGES
[Sat Mar 5 05:22:35 2005] action exited with status 0
[Sat Mar 5 05:22:35 2005] completed event "button/lid LID 00000080 00000001"
[Sat Mar 5 05:22:40 2005] received event "button/lid LID 00000080 00000002"
[Sat Mar 5 05:22:40 2005] notifying client 7996[1000:1000]
[Sat Mar 5 05:22:40 2005] client has disconnected
[Sat Mar 5 05:22:40 2005] executing action "/etc/acpi/lid.sh"
[Sat Mar 5 05:22:40 2005] BEGIN HANDLER MESSAGES
xscreensaver-command: warning: $DISPLAY is not set: defaulting to ":0.0".
Xlib: connection to ":0.0" refused by server
Xlib: Invalid MIT-MAGIC-COOKIE-1 key
xscreensaver-command: can't open display :0.0
[Sat Mar 5 05:22:43 2005] END HANDLER MESSAGES
[Sat Mar 5 05:22:43 2005] action exited with status 0
[Sat Mar 5 05:22:43 2005] completed event "button/lid LID 00000080 00000002"
[Sat Mar 5 05:23:13 2005] client connected from 18248[1000:1000]
[Sat Mar 5 05:23:13 2005] 1 client rule loaded

/var/log/syslog:

Mar 5 05:22:36 surfboy udev[17969]: removing device node '/dev/vcsa7'
Mar 5 05:22:36 surfboy udev[17968]: removing device node '/dev/vcs7'
Mar 5 05:22:36 surfboy gdm[6434]: gdm_slave_xioerror_handler: Fatal X error -
Restarting :0
Mar 5 05:22:36 surfboy gconfd (johannes-7571): Received signal 15, shutting
down cleanly
Mar 5 05:22:36 surfboy gconfd (johannes-7571): Exiting
Mar 5 05:22:37 surfboy udev[18009]: removing device node '/dev/vcsa7'
Mar 5 05:22:37 surfboy udev[18002]: removing device node '/dev/vcs7'
Mar 5 05:22:38 surfboy udev[18027]: creating device node '/dev/vcs7'
Mar 5 05:22:38 surfboy udev[18029]: creating device node '/dev/vcsa7'
Mar 5 05:22:50 surfboy gconfd (johannes-18126): starting (version 2.9.91), pid
18126 user 'johannes'
Mar 5 05:22:50 surfboy gconfd (johannes-18126): Resolved address
"xml:readonly:/etc/gconf/gconf.xml.mandatory" to a read-only configuration
source at position 0
Mar 5 05:22:50 surfboy gconfd (johannes-18126): Resolved address
"xml:readwrite:/home/johannes/.gconf" to a writable configuration source at
position 1
Mar 5 05:22:50 surfboy gconfd (johannes-18126): Resolved address
"xml:readonly:/etc/gconf/gconf.xml.defaults" to a read-only configuration source
at position 2
Mar 5 05:22:59 surfboy gconfd (johannes-18126): Resolved address
"xml:readwrite:/home/johannes/.gconf" to a writable configuration source at
position 0

[Thom May (thombot)]

This looks suspiciously like an X problem.

[Daniel Stone (daniels)]

Wonder if it's X or GNOME dying. Johannes, could you please attach
/var/log/Xorg.0.log.old from after a crash?

[Johannes Gijsbers (johannes)]

Hrm. I haven't been able to reproduce this since your message, Daniel. I even
tried closing and opening the lid a dozen times or so, but no cigar. I did
upgrade my system in the meantime (I'm tracking hoary+a couple of other
sources), so one of those upgrades probably fixed the problem. I recommend
closing the bug, but I'll post the X logs if it ever happens again.

[Daniel Stone (daniels)]

Closing as WORKSFORME per Johannes's comment; this is fodder for reopening if it
pops back up, though. Thanks for following this one up.

[L.Lopez (ldotlopez)]

I can reproduce the bug every time I close the lid.

I have a partial fix for workaround this bug.
Erasing "chvt ..." from /etc/acpi/lid.sh solves the problem at least on my box.

My logs dont say anything special diferent from the rest of bug's comments, at
least at first sight.

VGA Info:
 VGA compatible controller: ATI Technologies Inc RV350 [Mobility Radeon 9600 M10]

Running Ubuntu Hoary with kernel 2.6.10-31, xorg (6.8.2-5.1) and fglrx drivers
(6.8.0-8.8.25-0ubuntu8)

[Daniel Stone (daniels)]

i believe this problem is fixed in 8.10.19/8.12.3.

[Daniel Stone (daniels)]

8.12.10 is now in Breezy.