[Openchrome-users] [openChrome] #252: OpenChrome & Screen savers
OpenChrome Trac
trac
Sat Sep 6 02:13:16 PDT 2008
#252: OpenChrome & Screen savers
------------------------------------------+---------------------------------
Reporter: Danix803 | Owner: somebody
Type: defect | Status: new
Priority: critical | Component: xdriver
Version: 0.2.903 | Resolution:
Keywords: DRM crash screensaver AGP DMA |
------------------------------------------+---------------------------------
Changes (by gringinho):
* keywords: => DRM crash screensaver AGP DMA
* priority: major => critical
Comment:
I have a similar problem and have been experiencing this with various
openchrome releases (0.2.900 , 0.2.901 , 0.2.902 , 0.2.903) and different
kernels (2.6.22 - 2.6.26). I use Debian Etch 4.0 with some backports and
KDE 3.5.7, Mesa 6.5.1.
I first thought there were some other problems, but recently looked into
it more. With the 0.2.902 release and kernel 2.6.25 as well as 0.2.903 +
2.6.26 , I enabled the EnableAGPDMA option. That made it possible to see
some more immediate crashes, when going into the KDE control centre and
looking at the ScreenSaver section. With the miniature screensaver running
in the control centre, the screen started having artifacts indicating some
memory overwrite, and it locked up - hard and similar to how it locks up
after having the screensaver running for some time. I am using the OpenGL
screensaver "Flurry". When I flipped through different OpenGL screensavers
in the KDE control centre, the screen artifacts and lock-up happened very
quickly. I was able to go to a virtual console immediately when artifacts
appeared, and looked at logs etc. When I went back to X, it locked hard.
When it locks up, other processes - audio, network etc. are running
normally.
I looked into the /var/log/kern.log and found the following:
Sep 6 04:37:19 lobo kernel: [ 62.006558] [drm] Initialized drm 1.1.0
20060810
Sep 6 04:37:19 lobo kernel: [ 62.042998] ACPI: PCI Interrupt
0000:01:00.0[A] -> GSI 16 (level, low) -> IRQ 16
Sep 6 04:37:19 lobo kernel: [ 62.042998] [drm] Initialized via 2.11.1
20070202 on minor 0
Sep 6 04:37:19 lobo kernel: [ 62.091211] agpgart: Found an AGP 3.5
compliant device at 0000:00:00.0.
Sep 6 04:37:19 lobo kernel: [ 62.091211] agpgart: Putting AGP V3 device
at 0000:00:00.0 into 8x mode
Sep 6 04:37:19 lobo kernel: [ 62.091211] agpgart: Putting AGP V3 device
at 0000:01:00.0 into 8x mode
Sep 6 04:39:01 lobo kernel: [ 177.731472] [drm:via_hook_segment] *ERROR*
Paused at incorrect address. 0xd1fca800, 0xd1fc9800 0x00000000
Sep 6 04:39:01 lobo kernel: [ 177.731545] [drm:via_hook_segment] *ERROR*
Paused at incorrect address. 0xd1fcb800, 0xd1fc9800 0x00000000
Sep 6 04:39:01 lobo kernel: [ 177.731589] [drm:via_hook_segment] *ERROR*
Paused at incorrect address. 0xd1fcc800, 0xd1fc9800 0x00000000
snip...
Sep 6 04:39:42 lobo kernel: [ 218.932883] [drm:via_hook_segment] *ERROR*
Paused at incorrect address. 0xd1eb0200, 0xd1fc9800 0x00000000
Sep 6 04:39:42 lobo kernel: [ 218.932912] [drm:via_hook_segment] *ERROR*
Paused at incorrect address. 0xd1eb0400, 0xd1fc9800 0x00000000
Sep 6 04:39:42 lobo kernel: [ 218.933006] [drm:via_hook_segment] *ERROR*
Paused at incorrect address. 0xd1eb0600, 0xd1fc9800 0x00000000
Sep 6 04:39:43 lobo kernel: [ 219.805870] [drm:via_hook_segment] *ERROR*
Paused at incorrect address. 0xd1eb0800, 0xd1fc9800 0x00000000
The lockup after inactivity has happened with 0.2.900 - 0.2.903 and
with/without EnableAGPDMA - it was only easier to provoke with
EnableAGPDMA and also showed visual artifacts. Before enabling AGPDMA, the
monitor would be in power-saving mode or without sync. When it locks up
hard, the X server becomes unresponsive to everything - all key input is
frozen, even Ctrl-Alt-Backspace etc.
--
Ticket URL: <http://www.openchrome.org/trac/ticket/252#comment:1>
openChrome <http://www.openchrome.org/>
The openChrome project
More information about the Openchrome-users
mailing list