Bug 20477

Summary: [Mobility M6 LY] System freeze with DefaultDepth 16
Product: xorg Reporter: Bryce Harrington <bryce>
Component: Driver/RadeonAssignee: xf86-video-ati maintainers <xorg-driver-ati>
Status: RESOLVED INVALID QA Contact: Xorg Project Team <xorg-team>
Severity: normal    
Priority: high    
Version: 7.3 (2007.09)   
Hardware: Other   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
Xorg.0.log with 6.11.0 none

Description Bryce Harrington 2009-03-05 01:04:42 UTC 
Created attachment 23554 [details] [review]
Xorg.0.log with 6.11.0

Forwarding this (ancient) bug from a Ubuntu reporter:
https://bugs.edge.launchpad.net/xserver-xorg-driver-ati/+bug/36596

[Problem]
By default, the DefaultDepth is set to 24, however on this hardware this disables DRI. But when setting DefaultDepth to 16 it causes a system freeze.

Possibly there is some relation with bug 16365, which has similar symptoms on the same hardware.

[lspci]
01:00.0 VGA compatible controller [0300]: ATI Technologies Inc Radeon Mobility M6 LY [1002:4c59]
 Subsystem: Compaq Computer Corporation Device [0e11:b11b]
(II) PCI: 00:00:0: chip 8086,3575 card ffff,ffff rev 02 class 06,00,00 hdr 00
(II) PCI: 01:00:0: chip 1002,4c59 card 0e11,b11b rev 00 class 03,00,00 hdr 00
(--) PCI:*(1:0:0) ATI Technologies Inc Radeon Mobility M6 LY rev 0, Mem @ 0xd8000000/27, 0xd0100000/16, I/O @ 0x2000/8

[Original Report]
This seems to be a general problem (it also occurred when i used bug-buddy) but i can reproduce it easily using pan, among other methods.

The steps to reproduce the problem are:
a) run pan
b) update a newsgroup
c) quickly open task manager, stop the update task (using the stop icon)
d) click on the "trash" icon
e) system is frozen,
        e.g. mouse not moving, even ctrl-alt-f1 does not work. I also have a web server (apache2) running, and it is not accessible after the freeze.

Another way to reproduce the problem is:
a) run glxgears
b) move the window around.

Workarounds:
a) After changing from ubuntulooks to clearlooks theme, it occurs much less frequently.

b) After changing from xorg "ati" to "vesa" driver, it has not occurred yet (been using a few hours). (used together with "ubuntulooks" theme )

c) option "RenderAccel 0" works

d) Disabling modules "dri" and "glx" works (at least as of June 2007)

I first reported it against:
- latest dapper as of 20060325,
- pan 0.14.2.91.
- kernel 2.6.15-19-686 #1 SMP PREEMPT Mon Mar 20 17:33:20 UTC 2006 i686 GNU/Linux . (I have also tried on kernel 2.6.15-19-386).

Also reported as still occurring on:
* Gutsy Alpha 4 (2007-08-21) - still freezes at DefaultDepth 16
* Hardy Alpha 6 (2008-03-10) - still freezes at DefaultDepth 16
* -ati 6.9.0 - still freezes at DefaultDepth 16
* -ati 6.11.0 - still freezes at DefaultDepth 16

My problems with "ati" does not seem 3D related, I have tried disabling "dri" and "glx" modules and it still crashes.

My system is a PIII Mobile 1GHz, 384 MB RAM, and Mobility Radeon M6 LY graphics.

I am still able to get a system freeze at defaultDepth 16. At depth 16 DRI is on (at depth 24 there is not enough memory for DRI).

Having said that, I myself do NOT have any requirements for defaultDepth 16 or DRI (just wanted to try some of the desktop effects).
I would not object if defaultDepth 16 is not supported for an 8Megabyte M6.

In trying older drivers, the problem seems to have started between 6.5.7.3 and 6.5.8.
Comment 1 Alex Deucher 2009-03-05 07:48:41 UTC 
Sounds like an AGP problem.  removing the load dri and glx lines from your config doesn't work on newer xservers because they are loaded automatically.  you have to explicitly disable it via:
Option "DRI" "FALSE"

Try changing the AGP mode:
Option "AGPMode" "x"
where x = 1 or 2 or 4

If that doesn't work, try forcing it to pci mode:
Option "BusType" "PCI"
Comment 2 Bryce Harrington 2009-03-05 14:11:27 UTC 
I had him try out the AGPModes already but it didn't seem to work:

  "No luck, it still crashes with the new drivers. I have tried using AGPMode 1,2,4 (mode 8 is unavailable on this machine).  I tested with module version 6.9.0"

Forcing PCI mode hasn't been tried yet, sounds like a good thing to try next.  Ed?
Comment 3 Bryce Harrington 2009-06-01 12:59:05 UTC 
Ed says the PCI option works

'Sweet. The "bustype" "PCI" option works.'

Can this be quirked?
Comment 4 Alex Deucher 2009-06-01 14:21:02 UTC 
This looks like it may be a dup of bug 18871.  Would be useful to see if this is a drm problem.
Comment 5 Christopher M. Penalver 2016-02-24 05:00:06 UTC 
Closing as considered fixed downstream.

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.