[i865] pretty unstable

[Edmundo]

I think I have finally recovered acceleration on 865G after losing it on jaunty... however there's a heavy flicker affecting the screen. It's like if the screen is scrolled to the right of the screen and moved back to its normal position over and over in a very short amount of time. I think it's somehow tied to video refreshes (when something changes on the screen) and the CPU usage.

Another problem is that after a while of these heavy flickering there's a condition (go figure what it is) that will make video pseudo-freeze. I get to move the mouse, the screen is positioned normally but there's no more change in the screen. Keyboard gets unresponsive too. I get to reboot with alt-sysrq-x, though.

The box is close to unusable right now. I tried to report this bug on GNU/Linux like 4 times and every time it pseudo-freezed (as I said). I'm reporting it on güindous (and it makes me so happy, you know).

Hope I can help somehow to solve this 865g problems once and for all.

[Timo Aaltonen (tjaalton)]

Render acceleration should be disabled for now, since it doesn't seem like getting fixed anytime soon. There's a fresh patch to disable it in the Debian package.

[Edmundo (eantoranz)]

So what should I do?

[Edmundo (eantoranz)]

As a workaround, you can get a working environment with vesa.

First, blacklist i915. That's done adding a line saying:
blacklist i915 to /etc/modprobe.d/blacklist.conf

By the way, that blacklist worked with the 2.6.31-5 kernel. With 2.6.31-4 it didn't.

Then configure xorg (/etc/X11/xorg.conf) to use the vesa driver (if you don't have a /etc/X11/corg.conf file it will require that you create one by running sudo X --configure with the gdm/kdm/xdm/whateverdm service down and copy it there for xorg to use it). That will be enough to get a working environment instead of a screen with a heavy case of parkinson's desease.... no acceleration though.... but at least you will be able to work.

[Edmundo (eantoranz)]

Hi, guys!

I was testing the 2.6.31-10 kernel (just upgraded the box). The 'parkinson's desease' problem from boot up is gone. After I removed the the i915 from the blacklist, I got a stable screen on boot up.

However, when I removed the xorg.conf (that I use to stick with vesa) I think I got the hardware acceleration back _but_ the box pseudo-hangs when kde is still starting my environment (past the splash screen, while loading the applications). It happened twice after updating so I'm sticking with vesa for now. I was able to reboot with alt-sysrq key combinations (that's why I said it pseudo-crashes).

What can I do so that I gather some information for you regarding these situation?

[Tomasz Czapiewski (xeros)]

I'm not sure if it's the same bug but I've just installed today's (2009-10-05) Kubuntu Karmic cdimage on i865G hardware and since boot from LiveCD til now (installed and upgraded system on hard drive) I'm getting proper display on intel driver but Xorg eats ~100% CPU and even text typing in Konsole lags so much as it looks like working on SSH terminal on heavy loaded server/internet connection - letters are arriving few seconds after typing them. But when I switch to console (CTRL+ALT+F1 the Xorg doesn't eat even 1% CPU and system is very responsive and I've get ~99% CPU idle). After switching back to KDE4.3.1 in Xorg (ALT+F7) there's ~100% CPU eaten by Xorg etc. So I've installed Kubuntu by most of the time switching to console to idle the CPU. Is there any workaround for this - other than switching to Vesa driver?

[Tomasz Czapiewski (xeros)]

Todays (kernel?) upgrade fixed the 100% CPU used by Xorg but there's indeed instability of intel driver (sometimes I can see drm error):
$ dmesg | grep drm
[ 9.007420] [drm] Initialized drm 1.1.0 20060810
[ 9.215105] [drm] fb0: inteldrmfb frame buffer device
[ 9.215118] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0
[ 9.330064] [drm:i915_handle_error] *ERROR* EIR stuck: 0x00000010, masking
[ 9.347234] [drm] DAC-5: set mode 1024x768 c

[Tomasz Czapiewski (xeros)]

$ dmesg | grep error
[ 9.330058] render error detected, EIR: 0x00000010
[ 9.330064] [drm:i915_handle_error] *ERROR* EIR stuck: 0x00000010, masking
[ 9.330075] render error detected, EIR: 0x00000010

[GonzO (gonzo)]

I can confirm random all-but-mouse freezes (system still live, able to ssh in and reboot).

00:02.0 VGA compatible controller: Intel Corporation 82865G Integrated Graphics Controller (rev 02)

Dell Dimension 3000, 2g Ram.

Vesa allows the box to work.

[Tomasz Czapiewski (xeros)]

After todays upgrade of kernel and xserver-xorg-video-intel (downloaded from "en" mirror) computer works without crash/hang etc. for ~4 hours now.

[Edmundo (eantoranz)]

Well, I just updated. Looks like I got a normal (accelerated....
Yippie!) environment. I removed the xorg.conf file and i915 is not
blacklisted and it's working. I'll let you know later if it remains
like this without crashing.

[Edmundo (eantoranz)]

It doesn't happen all the time but it has already happened twice that
when the box is booting I get to a black screen with nothing at all on
it. No kdm, no mouse pointer, no keyboard input (not even the leds),
no nothing. I'm able to shutdown gracefully if I press the power
button or if I use the alt+sysrq keys.

I checked the old xorg log (broken start) with the new one (working) I
saw that the log got cut on the broken one here:
(II) AIGLX: enabled GLX_MESA_copy_sub_buffer
(II) AIGLX: enabled GLX_SGI_swap_control and GLX_MESA_swap_control
(II) AIGLX: GLX_EXT_texture_from_pixmap backed by buffer objects
(II) AIGLX: Loaded and initialized /usr/lib/dri/i915_dri.so
(II) GLX: Initialized DRI2 GL provider for screen 0
(II) intel(0): Setting screen physical size to 270 x 200

On the normal one, it says the same plus:
(II) config/hal: Adding input device AT Translated Set 2 keyboard
(II) LoadModule: "evdev"
(II) Loading /usr/lib/xorg/modules/input//evdev_drv.so
(II) Module evdev: vendor="X.Org Foundation"
        compiled for 1.6.3, module version = 2.2.5
        Module class: X.Org XInput Driver
        ABI class: X.Org XInput driver, version 4.0

Etc etc... is it something to place on this same bug or it should be
placed on a new bug?

[Edmundo (eantoranz)]

Well, I think we are not through with the problems with 865. More often than not when the box is booting, X won't start and I won't get access to the VTs when that happens. I can shut it down gracefully by pressing power button or by using alt-sysRq keys. From another box through SSH I get to see this on dmesg:

[ 85.663792] [drm:i915_wait_request] *ERROR* something (likely vbetool) disabled interrupts, re-enabling
[ 85.663820] render error detected, EIR: 0x00000010
[ 85.663827] [drm:i915_handle_error] *ERROR* EIR stuck: 0x00000010, masking
[ 85.663838] render error detected, EIR: 0x00000010

Just in case, here's the backtrace of X:

$ sudo gdb proogram 2252
GNU gdb (GDB) 7.0-ubuntu
Copyright (C) 2009 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law. Type "show copying"
and "show warranty" for details.
This GDB was configured as "i486-linux-gnu".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
proogram: No existe el fichero ó directorio.
Adjuntando a process 2252
Leyendo símbolos desde /usr/bin/Xorg...(no debugging symbols found)...hecho.
Leyendo símbolos desde /usr/lib/libpciaccess.so.0...(no debugging symbols found)...hecho.
Símbolos cargados para /usr/lib/libpciaccess.so.0
Leyendo símbolos desde /lib/tls/i686/cmov/libdl.so.2...(no debugging symbols found)...hecho.
Símbolos cargados para /lib/tls/i686/cmov/libdl.so.2
Leyendo símbolos desde /usr/lib/libXfont.so.1...(no debugging symbols found)...hecho.
Símbolos cargados para /usr/lib/libXfont.so.1
Leyendo símbolos desde /usr/lib/libXau.so.6...(no debugging symbols found)...hecho.
Símbolos cargados para /usr/lib/libXau.so.6
Leyendo símbolos desde /usr/lib/libfontenc.so.1...(no debugging symbols found)...hecho.
Símbolos cargados para /usr/lib/libfontenc.so.1
Leyendo símbolos desde /usr/lib/libpixman-1.so.0...(no debugging symbols found)...hecho.
Símbolos cargados para /usr/lib/libpixman-1.so.0
Leyendo símbolos desde /usr/lib/libhal.so.1...(no debugging symbols found)...hecho.
Símbolos cargados para /usr/lib/libhal.so.1
Leyendo símbolos desde /lib/libdbus-1.so.3...(no debugging symbols found)...hecho.
Símbolos cargados para /lib/libdbus-1.so.3
Leyendo símbolos desde /lib/tls/i686/cmov/libpthread.so.0...(no debugging symbols found)...hecho.
[Thread debugging using libthread_db enabled] ...

[Jukka Alasalmi (jualasal)]

I have exactly the same symptoms as in the comment #12, except that I cannot even shut down my computer by using the power key (except with a long press, which does a HW power off). I haven't yet tried the SysRq method - I'll try the next time the freezing occurs.

I get the issue sometimes when booting up, and sometimes when the computer has been left idle for a while (I don't know if that is just a coincidence). During the boot-up, I sometimes do get a non-blinking text cursor "_" in the top-left corner when the computer freezes.

I'm having a bit different HW, though: i830M. It also has the flicker issue (though maybe not identical with this bug report), which I've filed in the bug #477264.

Here are a couple of patches that might help in the issue, but unfortunately I haven't so far had the time to apply them and compile the kernel in order to test if they help:

http://lists.freedesktop.org/archives/intel-gfx/2009-September/004122.html
http://lists.freedesktop.org/archives/intel-gfx/2009-September/004128.html

[tarverator (tarverator)]

Although I have not had any freezes, I get the same errors at boot with 2.6.31-15-generic:

[ 1.639827] [drm] Initialized drm 1.1.0 20060810
[ 1.816193] [drm] fb0: inteldrmfb frame buffer device
[ 1.816267] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0
[ 1.905713] [drm:i915_handle_error] *ERROR* EIR stuck: 0x00000010, masking
[ 1.912046] [drm] DAC-5: set mode 1280x1024 17

I had no acceleration at all after I upgraded from jaunty to karmic, but I did get video to sort of work without any change to xorg.conf by simply adding the kernel parameter xforcevesa. Video playback still pauses and stutters, however.

[bersyl91 (choppy-free)]

I get the same error on 9.10 server when the system boots.

Currently, it does not affect running, I suppose it's because I'm only using text mode but I need to be sure before putting in production.

[bersyl91 (choppy-free)]

Oh, important note: it's on an i915.

[Chris Halse Rogers (raof)]

Ok! Ladies and Gentlemen! Chris Wilson of upstream fame has done some work to re-integrate a legacy driver for the Intel cards that have been hard done by in the GEM transition.

The xserver-xorg-video-intel packages in https://edge.launchpad.net/~raof/+archive/aubergine have a GEM-less legacy driver re-integrated which is activated when KMS is disabled - which it is by default on your card for Lucid.

It would be useful if you could remove any work-arounds you've used to get your system more stable and test the drivers from this PPA. They should hopefully end up slightly more stable than the drivers in Ubuntu 9.04 (Jaunty), which were the last set of drivers to not use the GEM memory manager, and so significantly more stable than the drivers in Ubuntu 9.10 and Ubuntu 10.04 (Karmic and Maverick).

Could you please test the drivers from this PPA and report your experiences with them?

[Edmundo (eantoranz)]

I'm not using the box at the time (and if I did, it'd be on lucid) so
I think I won't be able to help on this problem at the time.

[Chris Halse Rogers (raof)]

I'm unassigning myself from this bug; I've got the needed feedback for Maverick.

I'll leave this bug open; there's still a reasonable chance we can get a proper fix, and apparently some i8xx documentation has just been released.

[Bryce Harrington (bryce)]

Since there's not really been any notable testing work done on this bug report in over a year I'm closing out the bug.

Unfortunately, upstream is limiting support for i865 chips and it's definitely true that there are stability issues. I think it is probably best if issues with i865 be reported upstream at bugs.freedesktop.org.