Bug 13688

Summary: xorg-server-1.4.0.90 freezes
Product: xorg Reporter: Bernd Steinhauser <linux>
Component: Driver/fglrxAssignee: atilinuxbugsunassigned
Status: RESOLVED FIXED QA Contact: Xorg Project Team <xorg-team>
Severity: major    
Priority: medium    
Version: unspecified   
Hardware: Other   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
Full log
none
emerge --info output none

Description Bernd Steinhauser 2007-12-16 11:13:23 UTC
Yesterday I upgraded from xorg-server-1.4 to xorg-server-1.4.0.90.
With the first one I didn't expect any freezes, that could be related to the xorg-server for about 2 months (user 1.3 before).
Since yesterday my computer freezed several times.

I found messages like this in the log, they appear very often:
tossed event which came in late
mieqEnequeue: out-of-order valuator event; dropping.

When this does happen, the server doesn't work anymore, the mouse is still movable, but doesn't respond to any keyboard action or mouse clicks.
The opened windows are not updated anymore. 

SSH login is still possible, but killing the X server won't work (even with -9), a reboot is needed.

I'll attach the full log.
Comment 1 Bernd Steinhauser 2007-12-16 11:14:01 UTC
Created attachment 13142 [details]
Full log
Comment 2 Bernd Steinhauser 2007-12-16 11:16:32 UTC
Created attachment 13143 [details]
emerge --info output

emerge --info output.
Maybe this could be helpful.
Comment 3 Daniel Stone 2007-12-16 13:39:05 UTC
This just means that the server is hung.  Unfortunately, debugging this is impossible with a proprietary driver.
Comment 4 Bernd Steinhauser 2007-12-16 13:56:27 UTC
(In reply to comment #3)
> This just means that the server is hung.  Unfortunately, debugging this is
> impossible with a proprietary driver.
> 

Hm, ok.
I've opened a bug in there database.
http://ati.cchtml.com/show_bug.cgi?id=936
Comment 5 lonefox 2007-12-21 13:32:33 UTC
I have a way to reproduce this with open source drivers. Of course I can't be sure that it is the same bug, but symptoms are exactly same.
- Start X using IceWM as window manager (I use IceWM 1.3.1, but I don't believe the version matters).
- Press Control-Alt-Del to bring up the IceWM system dialog.
- Sometimes the server freezes immediately. If not, clicking one of the buttons in the dialog triggers it always.
I have tested this in three different systems: one with RadeonHD driver and two with xf86-video-ati. Same behavior in all three.

BTW, the server actually can be killed but then the screen is stuck in graphics mode, so it looks like nothing happened. Doing "DISPLAY=:0.0 startx" in ssh shell brings it back up, but it freezes again if I try to switch to text console.
Comment 6 James Cloos 2007-12-21 14:06:31 UTC
The icewm crash is probably fixed by commit 50e80c39870adfdc84fdbc00dddf1362117ad443.  See bug 13511 for details of that issue.

I’m not sure whether the OP’s issue is the same.
Comment 7 Bernd Steinhauser 2007-12-21 14:51:56 UTC
(In reply to comment #6)
> The icewm crash is probably fixed by commit
> 50e80c39870adfdc84fdbc00dddf1362117ad443.  See bug 13511 for details of that
> issue.
> 
> I’m not sure whether the OP’s issue is the same.
> 

I also don't think it is the same.
Because here the X server can't be killed.
(Meaning, that I kill it and just nothing happends, the process stays there and doesn't die.)

ATI released a new driver version, I'll try again with that one when I've got time.

I'm actually not sure if this is a xorg-server1-4.0.90+ati-drivers issue or if this is a xorg-server issue?
Comment 8 lonefox 2007-12-21 23:09:51 UTC
That patch fixes the IceWM crash... but replaces it with another bug. Now when I press backspace first time after using the dialog, the server quits without any error message, like if I had pressed control-alt-backspace instead.
Comment 9 Bernd Steinhauser 2008-01-09 01:09:36 UTC
With the newer version of fglrx (7.12 aka 8.443.1) it seems to work. See:
https://bugs.gentoo.org/show_bug.cgi?id=202553

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.