[i855] Xserver freezing on Toshiba Satellite A10 (Intel 82852/855GM)
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
xf86-video-intel |
Fix Released
|
Medium
|
|||
xserver-xorg-video-intel (Ubuntu) |
Fix Released
|
High
|
Unassigned |
Bug Description
Binary package hint: xserver-
After installing Xubuntu Karmic on this laptop, I'm having regular Xserver lockups, as in once per session. The mouse pointer still moves, but clicking does nothing and the keyboard is disabled. I can use the Alt-SysRq-REISUB commands, but the reboot that follows doesn't complete. The only way to clear this is by a complete power-down. The time before the freeze occurs varies but would usually be less than 15 minutes.
I have tried using the standard Xserver-
Hardy has run without issue on this machine, and as far as I can tell, Jaunty also. (I ran it for an hour or more from the live CD without problems.)
The batchbuffer dump archive is attached. I reinstalled the standard Xserver-
PS: This was spun off from bug 447892, since it doesn't seem to be the same freeze.
Jane Atkinson (irihapeti) wrote : | #1 |
Bryce Harrington (bryce) wrote : | #2 |
tags: | added: needs-xorglog |
Changed in xserver-xorg-video-intel (Ubuntu): | |
status: | New → Incomplete |
Jane Atkinson (irihapeti) wrote : | #3 |
I've attached both Xorg.0.conf and Xorg.0.log.old, because I'm not sure which one, if any, is the one needed.
The laptop was rebooted immediately after the freeze (via ssh) and hung while rebooting, and then again after powering off completely. (I retrieved the log files just now via PartedMagic liveCD.) So maybe the necessary info has been overwritten. :(
I can run the whole batchbuffer thing again and include the Xorg.0.log file this time, if it will help.
Jane Atkinson (irihapeti) wrote : | #4 |
Jane Atkinson (irihapeti) wrote : | #5 |
Should have added this in the initial report: This bug is a duplicate of / related to https:/
Also: Xorg.0.log from the time of the freeze was provided as part of the initial batchbuffer dump archive.
summary: |
- Xserver freezing on Toshiba Satellite A10 (Intel 82852/855GM) + [i855] Xserver freezing on Toshiba Satellite A10 (Intel 82852/855GM) |
Jane Atkinson (irihapeti) wrote : | #6 |
I see that this bug is still marked "incomplete".
Is there something else I should be doing to help progress?
tags: |
added: 855gm freeze karmic removed: needs-xorglog |
Geir Ove Myhr (gomyhr) wrote : | #7 |
Irihapeti, in bug 447892 comment 23 it was suggested to use xorg-edgers with the option DebugFlushCaches. Could you use that for your batchbuffer dump, since it will most likely make it easier for the developers to make sense of it.
Also, except for the batchbuffer dump, we prefer to have files attached separately and not in an archive. That makes it much easier to look at them through the web interface. It would be nice if you could copy /var/log/Xorg.0.log at the same time as you are making the batchbuffer dump (when you are logged in via SSH and the computer is frozen).
Changed in xserver-xorg-video-intel (Ubuntu): | |
status: | Incomplete → Confirmed |
description: | updated |
Jane Atkinson (irihapeti) wrote : | #8 |
@Geir Ove Myhr:
Sorry if this is a stupid question, but there appear to be links to two versions of xorg-edgers attached to that message. Which one do I use?
Geir Ove Myhr (gomyhr) wrote : | #9 |
Sorry, I thought the link to xorg-edgers was more visible in that bug report. The right one is https:/
Jane Atkinson (irihapeti) wrote : | #10 |
@Geir Ove Myhr
Ahh, I was thinking I had to download stuff from git and compile it!! A ppa sounds much easier for me to handle.
So, if I understand correctly, I need to install the .deb from the ppa, create an xorg.conf file with:
Section "Device"
Identifier "my-old-i8xx"
Driver "intel"
Option "DebugFlushCaches" "1"
EndSection
recreate the freeze, and post a batchbuffer archive plus the xorg.0.log as a separate text file.
Have I got that right?
Geir Ove Myhr (gomyhr) wrote : Re: [Bug 477256] Re: [i855] Xserver freezing on Toshiba Satellite A10 (Intel 82852/855GM) | #11 |
> Have I got that right?
Almost...
> So, if I understand correctly, I need to install the .deb from the ppa,
The xorg-edgers is made so that the packages depend on each other in
the right way. So while it may work to only install the
xserver-
doesn't (if the xorg-server in the PPA is too new compared to the one
in Karmic, for example). The way it is meant is to add the PPA to
/etc/apt/
apt-get dist-upgrade`. It used to be quite cumbersome to back out
again, but now there is the ppa-purge script in there that does that
easily (`sudo apt-get install ppa-purge` to get it after having
xorg-edgers).
Otherwise, you got it all right :-)
Jane Atkinson (irihapeti) wrote : | #12 |
Thanks. It goes to show that it pays to ask for clarification, even if I risk looking like an idiot :) As you might have guessed, I'm new to bug reporting.
Jane Atkinson (irihapeti) wrote : | #13 |
- batchbuffer dump Edit (285.3 KiB, application/x-tar)
Here are the latest logs, after following the instructions given above. I hope these are more useful.
Jane Atkinson (irihapeti) wrote : | #14 |
Geir Ove Myhr (gomyhr) wrote : | #15 |
Thank you for the new files. At bug 476817 which is similar to this (855GM freeze bug) it is reported that adding acpi=off to the kernel command line avoids the freezes. Could you check if this is the case for you as well? To edit the kernel command line, press Esc on boot the show the Grub boot menu, 'e' to edit the default entry and then you can add acpi=off after "quiet splash".
Jane Atkinson (irihapeti) wrote : | #16 |
That appears to have fixed the freezing. I used the system for over 3 hours, mainly web browsing. Previously, the limit was one hour, maximum, and often a lot less.
However, two odd things happened. Firstly, that the screen suddenly went blank for no apparent reason. Moving the mouse did not restore it but touching the touchpad did. This happened twice during the session. Secondly, the laptop did not shut down properly. The final message on the screen reads:
login: init: statd main process (2206) killed by KILL signal
init: statd main process ended, respawning
[11176.911297] System halted.
I had to pull out the plug to turn it off completely.
Jane Atkinson (irihapeti) wrote : | #17 |
I found a fix for the shutdown problem here:
http://
It may take a while before I know if that's caused any other problems.
Geir Ove Myhr (gomyhr) wrote : | #18 |
A couple of more things I would like you to test in order to narrow this down (when booting without acpi=off)
- Could you try to disable desktop effects and see if you stil have freezes
- If you start the computer and leave it at the login screen, does it still hang, or do you have to log in first?
- Could you try a recent mainline kernel from http://
Jane Atkinson (irihapeti) wrote : | #19 |
I'd like to clarify a couple of things, if I may.
I think that desktop effects are already turned off. I'm not that familiar with Xubuntu, and I'm just running with the default settings. The compositor is turned off in the Window Manager Tweaks application. Is there something else I need to do as well?
I've never had the computer hang if I leave it and do nothing, even when logged in. I seem to have to be doing something active such as scrolling through documents or changing tabs in Firefox. How long would I need to leave inactive it to be sure?
So far, the new settings have meant I can actually use the computer for several hours. There is still this mysterious sudden blanking of the screen, but it's not a freeze or xserver crash and touching the touchpad or pressing an arrow key reactivates it.
By the way, I had to uninstall the xorg-edgers version of xserver-
Jane Atkinson (irihapeti) wrote : | #20 |
Installed mainline kernel v2.6.32-rc7, dated 13 November. The machine locked up as usual, but with one difference: I could use cntrl-alt-F2 to access tty2. Running TOP showed nothing unusual. I tried restarting gdm but it did no good. On reboot it hung at "loading modem drivers" or similar wording.
Jane Atkinson (irihapeti) wrote : | #21 |
Further tests:
I installed the xorg.conf file from https:/
Without the addition of acpi=off to Grub, the laptop hangs at a blank screen with a cursor once it's got past the initial XFCE mouse screen.
If I add acpi=off to the kernel line, then it boots completely and behaves much the same as the Intel driver. That is, I still get these unexpected screen blackouts at more or less 3/4 hour to 1 hour intervals. Moving the mouse (usb) does not revive the screen but pressing an arrow key or touching the touchpad does. No data is lost, but it's a bit disconcerting.
I booted into Ubuntu 8.04 for several hours this morning, and it's not happening there, so I don't think there's a hardware fault.
Jane Atkinson (irihapeti) wrote : | #22 |
Upgraded to the latest standard Karmic kernel, along with various other updates (but none for xserver-
Ben (benjamin-buffet) wrote : | #23 |
I confirm the problem for this laptop.
I also install xubuntu on my Toshiba Satellite A10 and there are random freeze.
@Irihapeti have you found a temporary solution or must I install xubuntu 9.04 ?
Jane Atkinson (irihapeti) wrote : | #24 |
@Ben
Yes, there is a workaround.
1) Add "acpi=off" (without the quotes characters) to kernel line in Grub. It should end like this: .... acpi=off quiet splash
2)Edit /etc/modules and add a new line: "apm power_off=1"
I'm still getting the screen suddenly go blank from time to time, but I restore it by pressing an arrow key or by touching the touchpad. No data is lost.
Ben (benjamin-buffet) wrote : | #25 |
@Irihapeti thanx ;)
I will try as soon as possible, but I just downgrad to 9.04 as I need to work with the Toshiba
Philip (philip-askthelocal) wrote : | #26 |
I run fedora 9 on a toshiba satellite A10 and have only this morning seen this same problem. I wonder if there is an element of hardware failure here? This is a laptop I've used since 2004.
Jane Atkinson (irihapeti) wrote : | #27 |
I installed linux-image-
The kernels behaved well, as far as I can tell, except that I had to boot using the recovery mode. Otherwise, boot hung at the point where the listing of modem drivers occurs.
As version 9 didn't freeze, I didn't bother with testing version 8.
Jane Atkinson (irihapeti) wrote : | #28 |
I have tried various kernels as per https:/
The last kernel I can use without lockup is 2.6.31.2
I notice that this version, plus rc9 and .1, seem to cause a lot of heavy CPU use which really slows down Firefox in particular. Also, on bootup I get an error message:
kernel: [ 0.076432] pci 0000:00:1d.0: BAR 4: address space collision on of device [0xcfe0-0xcfff]
kernel: [ 0.076529] pci 0000:00:1d.0: BAR 4: can't allocate resource
Geir Ove Myhr (gomyhr) wrote : | #29 |
Thank you for testing.
> The last kernel I can use without lockup is 2.6.31.2
Strange... Olie reported on bug 447892 that the last one working for
him was 2.6.31.1 and that 2.6.31.2 would freeze. It is possible that
the problems are slightly different. Could you check the kernels I
have compiled and put at http://
Olie everything up to 599-git522bb74 works without freeze and the
kernels from 600-git7abf3aa and above would freeze. Could you double
check that 600, 799, and 800 works without freeze for you? It is
entirely possible, but it means we have more work to do on this one,
and I would like to be sure before we start.
I would also really like to see some test of Lucid. It would be too
bad if it turns out that some later changes fixes this in Lucid, and
we go through a lot of trouble for nothing (it is unlikely that this
can be fixed with a Karmic update).
Jane Atkinson (irihapeti) wrote : | #30 |
I think that these are different problems, one reason being that the lockup I experience can sometimes take an hour to appear, whereas bug 447892 seems to cause lockups almost immediately. I can test the custom kernels you mention and see what happens. Presumably, if 800 works, I don't need to test the other two?
I understand your point about going through a load of trouble for nothing, and I'd be happy to test Lucid on this laptop. As it happens, I have a spare partition on it which I can use for Lucid without upsetting the Karmic install.
Geir Ove Myhr (gomyhr) wrote : | #31 |
> Presumably, if 800 works, I
> don't need to test the other two?
No, that shouldn't be necessary.
> The last kernel I can use without lockup is 2.6.31.2
So that means that the freezes start at 2.6.31.3 (and we're not
talking -rc3 here...)? I find that even stranger after looking at the
changelog:
http://
The only difference between 2.6.31.2 and 2.6.31.3 is a single patch
that fixes typos in what seems to be an unrelated part of the kernel.
Jane Atkinson (irihapeti) wrote : | #32 |
Having tested the kernels you compiled for Olie, I can confirm that I am getting the freeze on 600, 799 and 800, but not 599.
It may simply be that I didn't test the 2.6.31.2 kernel for long enough - it sometimes takes an hour or so before I get a freeze.
I'm about to install Lucid on the spare partition and see how that behaves.
Geir Ove Myhr (gomyhr) wrote : | #33 |
On Sat, Jan 2, 2010 at 10:10 PM, Irihapeti wrote:
> Having tested the kernels you compiled for Olie, I can confirm that I am
> getting the freeze on 600, 799 and 800, but not 599.
Great! That makes life easier for us :-)
Jane Atkinson (irihapeti) wrote : | #34 |
I am also getting the freeze with Lucid. Unfortunately the machine won't boot with acpi=off added to the kernel line.
Gianfranco (gianfranco-alongi) wrote : | #35 |
I can confirm the random freeze behavior with Ubuntu 9.10 on a Toshiba 10 Satellite, just freshly installed. Started googling the problem and found this bug report.
X freezes (gnome) from time to time and only a full power-reboot fixes the problem. Mouse can be moved but nothing else. Can't access ttys with Ctrl+Alt+F1 , .... , Sound still works (had internet radio on). Happens regardless of programs running, sometimes none, sometimes just Firefox.
greping the logs shows a lot of render errors, but I have no idea wether this is the root of the problem
/var/log/
/var/log/
/var/log/
/var/log/
/var/log/
/var/log/
/var/log/
/var/log/
/var/log/
/var/log/
/var/log/
/var/log/
/var/log/
/var/log/
/var/log/
/var/log/
Will try to modify grub with acpi=off and see wether it fixes the problem.
Gianfranco (gianfranco-alongi) wrote : | #36 |
I have not seen any freezes since adding acpi=off, but I am logged out every time I visit this swedish hifi page,
www.agrenshifi.
This is happening with Firefox 3.5.6 and ubuntu 9.10 on the Toshiba A10 Satellite. After being logged out, Firefox is unable to restore the tabs and the "oops this is embarasing" message is shown. Any thoughts?
I do not know if this is related in any way.
Geir Ove Myhr (gomyhr) wrote : | #37 |
Gianfranco, when you are logged out visiting that web page, what happens is probably that something crashes xorg. You will probably see a backtrace in the end of /var/log/
Geir Ove Myhr (gomyhr) wrote : | #38 |
[This comment goes to freeze bugs on intel 855GM]
At bug 456902 (similar freezes on 845G) there is currently an effort to find where xserver-
Apologies if someone already reported that Jaunty with Karmic kernel also freezes. In that case, testing older -intel and libdrm in Karmic is probably pointless. I'm loosing the overview as there are many similar bug reports on 845G, 855GM and 945G/945GM, and I forget exactly what is already tested for each chipset.
Jane Atkinson (irihapeti) wrote : | #39 |
I added Brian's PPA to my sources.list and installed the current versions of the packages. Unfortunately, I got the lock-up after only a few minutes.
Software versions are:
Kernel: 2.6.31-17-generic
libdrm: 1:2.4.11-2+karmic3
xserver-
No older versions of Brian's packages are currently available, otherwise I'd test them.
Jane Atkinson (irihapeti) wrote : | #40 |
@Geir
I just remembered that I originally tested a "retro" (Jaunty, I think) version of xserver-
I'm also running Lucid on this laptop and for the last few days I've been able to boot with nomodeset and use the vesa driver with no lockups after many hours. As I don't use desktop effects on this machine, I'm quite happy with the workaround.
Geir Ove Myhr (gomyhr) wrote : | #41 |
Irihapeti, there is a report at bug 496187 that libdrm 1:2.4.11-2+karmic3 and xserver-
Jane Atkinson (irihapeti) wrote : | #42 |
I've installed those versions. After about an hour of active use, it's locked up, but a little differently from previous freezes. I have set the clock in the panel to display seconds, so that I can tell easily when there's a graphics freeze, but this time the clock is still counting seconds. However, I can't do anything with the screen other than move the mouse.
During the time it was working, it was extremely heavy on the processor, maxing out at 100% just about every time I scrolled something in Firefox. Playing Freecell caused it to use over 700 MB of swap at one point. So, not really usable, even when working - more like walking through molasses!
Jane Atkinson (irihapeti) wrote : | #43 |
I noticed that there was an xserver-
I don't know if the different desktop environment would make any difference. Someone in another thread mentioned that no lockups occurred under IceWM, which is why I wonder if it might affect things.
Jane Atkinson (irihapeti) wrote : | #44 |
The laptop acquired an extra 512MB of RAM yesterday, and I thought it might be useful to reinstall Xubuntu with all updates and see what happens. Installation method was to install a command-line system first, then install Xubuntu-desktop plus do other updates before rebooting and logging into the Xubuntu desktop.
Unfortunately, the system froze twice. The first time after 1-3/4 hours and the second a lot sooner - probably around 30 minutes, maybe less.
I've run the laptop for many hours at a time while using lxde/openbox or openbox on its own. I've tried using just the xfce4 package as well and it seems to be fine, though this might need a bit more testing.
Therefore, I'm inclined to conclude that one or more of the Xubuntu-desktop components is/are interacting with xserver-
Geir Ove Myhr (gomyhr) wrote : | #45 |
On Sat, Jan 23, 2010 at 7:56 PM, Irihapeti wrote:
> Therefore, I'm inclined to conclude that one or more of the Xubuntu-
> desktop components is/are interacting with xserver-
> cause the freezes.
Yes, that is probably the case. This is useful information. Often
there are specific system calls that makes it hang, and different
desktop environments make different system calls.
I meant forward your bug report to intel developers back in November
when you got the batchbuffer dump with xorg-egers, but it looked like
there would soon be some progress on some other 855GM freeze bugs
which would give us some test cases and possibly a fix, so I waited
for that. Eventually identification of the kernel patch that triggered
the freezes came out of it, but I haven't seen any progress upstream
otherwise. At this time they would need driver and kernel versions at
least at the Lucid level. Do you have Lucid installed already or could
you install it? Standard Lucid would probably be okay, but it would be
better with xorg-edgers on top of that and then the DebugFushCaches
xorg.conf option for the batchbuffer dump as described on
https:/
to know if it also freezes with mainline kernel v2.6.33-rc5 which is
currently the newest, but I don't think that is needed for the
batchbuffer dump (you may want to avoid copying i915_regs in that
batchbuffer dump, since reading it may cause hangs - I think that one
has been removed in recent 2.6.33 kernels).
If you can get this information, I will report this upstream as soon
as I have the opportunity.
Jane Atkinson (irihapeti) wrote : | #46 |
@Geir
Lucid is already on the laptop, with lxde and openbox as desktop environments. I can make an image of the installation (so that I can put it back later) and then install Xubuntu-desktop and do some testing. I take it that ideally you'd like batchbuffer dumps from both the standard Ubuntu kernel plus the latest mainline kernel. I hope to have something to report in a day or two, other commitments allowing. :)
Geir Ove Myhr (gomyhr) wrote : | #47 |
Yes, I think they would only need one of the batchbuffer dumps, but since I don't know which one would be best, both would be ideal.
Jane Atkinson (irihapeti) wrote : | #48 |
- visual artifacts with xorg-edgers drivers Edit (21.5 KiB, image/png)
I've been testing with Xubuntu,
ubuntu kernel version 2.6.32-11
libdrm2 version 2.4.17+git20100120
xserver-
The first test was after installing xubuntu-desktop metapackage, but leaving all my other stuff installed.
Uptime was 6.5 hours with no freezes. I even turned on Xubuntu's window compositor to see if that would make a difference, and nothing happened.
Not being sure that something I'd previously installed was affecting the results, I did a fresh install of Xubuntu and ran another test.
Same result. Uptime of nearly 4.5 hours, including some use of the window compositor, and no freeze.
I am noticing some visual artifacts: occasional streaking (white background with pixellated colour) and bits being left out of type. I've attached a small screenshot of the latter.
Interestingly, the letter affected varies from instance to instance, but when it happens, the same letter is affected universally throughout the page. Often, a screen refresh or scrolling up and down will fix it.
I think that the xorg-edgers drivers may be heavier on RAM than the standard ones.
Geir Ove Myhr (gomyhr) wrote : | #49 |
This is actually somewhat bad news -- we can't send this upstream if we can't reproduce it with the newest drivers. Could you purge the xorg-edgers again? This is most easily done with the ppa-purge script from the ppa-purge package which is included in xorg-edgers. If that also doesn't freeze, it is good, but if it does, we need to find out what made the difference. Hopefully, only upgrading libdrm2 and xsererver-
As for the font problem, this is not uncommon when running xorg-edgers. Since this is bleeding edge code it typically has more bugs than released code. Here, we are focusing on the freeze bug. You may track xorg-edgers and report bugs like this upstream, but expect a much rougher ride with bugs coming and going at a high rate.
Jane Atkinson (irihapeti) wrote : | #50 |
- dri_debug-20100126.tgz Edit (333.8 KiB, application/x-tar)
@Geir
Standard Xubuntu Lucid froze after 5 hours of uptime. If it's of any relevance, I was playing Freecell when it happened, and I was trying to move a group of cards from the left-hand-most pile. On previous occasions, this action has caused freezes. I was playing Freecell at various times during the test, so am unclear why it took so long for the freeze to occur.
The batchbuffer files are attached. I was unable to get all of them by ssh because trying to get the i915 stuff caused a complete lockup. I had to get :0.log by booting a rescue CD and copying it that way. I hope that it's still usable.
Jane Atkinson (irihapeti) wrote : | #51 |
>so am unclear why it took so long for the freeze to occur.
I just remembered that I'd turned on desktop effects not long previously, to try and cause something to happen, if it was going to. That may be a contributing factor.
Geir Ove Myhr (gomyhr) wrote : | #52 |
> The batchbuffer files are attached. I was unable to get all of them by
> ssh because trying to get the i915 stuff caused a complete lockup. I had
> to get :0.log by booting a rescue CD and copying it that way. I hope
> that it's still usable.
It is a known problem that reading i915_regs can cause a lockup on
some computers. This file has been removed in recent kernels. You can
copy the other i915 files by using
sudo cp -r /sys/kernel/
instead of
sudo cp -r /sys/kernel/
What would make life easier is if you found a way to reproduce the
freeze more quickly. There is a gnome-specific script that was good at
reproducing some freezes on 965GM a while ago. That could be a
starting point, but would require some research into how to automate
things in other WMs.
Otherwise, I'm a little bit surprised the xorg-edgers should make a
difference regarding this freeze. Maybe you could go back to
xorg-edgers and enable desktop effects there as well, then maybe you
can get a freeze and a nice bathbuffer dump.
Jane Atkinson (irihapeti) wrote : | #53 |
I've had a couple more freezes since I left the last message, and they didn't take nearly as long to happen. I still don't know why - neither of them involved having desktop effects enabled, and I hadn't done any updates in the meantime.
I'll put the xorg-edgers stuff back and see what I can achieve.
Jane Atkinson (irihapeti) wrote : | #54 |
- dri_debug-20100127.tgz Edit (24.9 KiB, application/x-tar)
I got the xorg-edgers to freeze in the end by using a modified version of the repro.sh.
The batchbuffer dump is attached.
I did get one error message when running it:
cp: reading `/sys/kernel/
Now I'll go and get the latest mainline kernel and see what happens.
Jane Atkinson (irihapeti) wrote : | #55 |
- Batchbuffer dump - mainline kernel Edit (254.5 KiB, application/x-tar)
Batchbuffer dump from xorg-edgers and mainline kernel (2.6.33-
In freedesktop.org Bugzilla #26326, Geir Ove Myhr (gomyhr) wrote : | #56 |
Originally reported by Irihapeti at:
https:/
[Problem]
This GPU hang only happens every few hours during normal use with xfce with desktop effects enabled. It doesn't seem to happen without desktop effects or with openbox or lxde, but since it doesn't happen very often, it is hard to say for sure. Irihapeti can trigger it more often by using a script to stress the video system.
One of the two attached batchbuffer dumps is with default Lucid kernel (2.6.32) and the other with 2.6.33-rc5. Both are with recent xorg-edgers which give recent git versions of the xorg, driver, mesa, etc.
Irihapeti found that freezes started after the kernel commit
[e517a5e97080bb
(in the 2.6.31 stable tree with intel driver 2.9.0) but since the kernel and driver has progressed quite a bit since then, this freeze may be of a different nature.
[Original bug report from Ubuntu Karmic (2.6.31 kernel and driver 2.9.0)]
Binary package hint: xserver-
After installing Xubuntu Karmic on this laptop, I'm having regular Xserver lockups, as in once per session. The mouse pointer still moves, but clicking does nothing and the keyboard is disabled. I can use the Alt-SysRq-REISUB commands, but the reboot that follows doesn't complete. The only way to clear this is by a complete power-down. The time before the freeze occurs varies but would usually be less than 15 minutes.
I have tried using the standard Xserver-
Hardy has run without issue on this machine, and as far as I can tell, Jaunty also. (I ran it for an hour or more from the live CD without problems.)
The batchbuffer dump archive is attached. I reinstalled the standard Xserver-
PS: This was spun off from Launchpad bug 447892, since it doesn't seem to be the same freeze.
In freedesktop.org Bugzilla #26326, Geir Ove Myhr (gomyhr) wrote : | #57 |
Created an attachment (id=32911)
batchbuffer dump with 2.6.33-rc5
In freedesktop.org Bugzilla #26326, Geir Ove Myhr (gomyhr) wrote : | #58 |
Created an attachment (id=32912)
batchbuffer dump with kernel 2.6.32
In freedesktop.org Bugzilla #26326, Geir Ove Myhr (gomyhr) wrote : | #59 |
Created an attachment (id=32913)
dmesg output
In freedesktop.org Bugzilla #26326, Geir Ove Myhr (gomyhr) wrote : | #60 |
Created an attachment (id=32914)
Xorg.0.log
Geir Ove Myhr (gomyhr) wrote : | #61 |
I have reported the bug upstream at https:/
Changed in xserver-xorg-video-intel (Ubuntu): | |
status: | Confirmed → Triaged |
importance: | Undecided → High |
Changed in xserver-xorg-video-intel: | |
status: | Unknown → Confirmed |
Jane Atkinson (irihapeti) wrote : | #62 |
> Please register at bugs.freedeskto
Done.
Jane Atkinson (irihapeti) wrote : | #63 |
@Geir
I installed your libdrm version as mentioned in Bug 447892. Initially, I just added the PPA to my sources.list and updated. That installed a total of four upgraded packages, and my laptop locked up while I was running the repro.sh.
Then I downgraded the other three packages to the standard Ubuntu versions, just leaving libdrm as your version. I have run the script a number of times and I can't get the laptop to lock up.
I'll keep using it like this for a while and see what happens.
Geir Ove Myhr (gomyhr) wrote : Re: [Bug 477256] Re: [i855] Xserver freezing on Toshiba Satellite A10 (Intel 82852/855GM) | #64 |
What packages were upgraded and which one did you keep? I'm not sure
about how the libdrm source package is split into binary packages, but
I would think that the patch only would affect libdrm-intel1.
Jane Atkinson (irihapeti) wrote : | #65 |
The upgraded packages were libdrm-intel1, libdrm-nouveau1, libdrm-radeon1, and libdrm2. I removed - downgraded, rather - the first three and kept libdrm2. I figure that -nouveau1 and -radeon1 wouldn't be doing anything on an intel system and that the problem was most likely to be with -intel1.
Still no lockups so far.
Jane Atkinson (irihapeti) wrote : | #66 |
After several hours of using the laptop (not all in one session), I finally got a lockup. It was different from the usual ones I experience because the mouse pointer also froze. The only option was a total power-down. (I've removed the battery some time ago to make this easier.)
tags: | added: xubuntu |
In freedesktop.org Bugzilla #26326, Chris Wilson (ickle) wrote : | #67 |
This looks like the CPU/GPU coherency issue with a twist. Both gpu dumps have recorded PGTBL_ERR=0x49, which is unusual. Those seem to be the unaligned fb issue, but there are several other preceding oopses in the dmesg as all. I'd like to see this bug confirmed on a recent kernel with i915_error_state capturing the batch buffer before declaring it to indeed be the same bug.
summary: |
- [i855] Xserver freezing on Toshiba Satellite A10 (Intel 82852/855GM) + [i855] [i855] Xserver freezing on Toshiba Satellite A10 (Intel + 82852/855GM) |
summary: |
- [i855] [i855] Xserver freezing on Toshiba Satellite A10 (Intel - 82852/855GM) + [i855] Xserver freezing on Toshiba Satellite A10 (Intel 82852/855GM) |
Geir Ove Myhr (gomyhr) wrote : | #68 |
Irihapeti, did you see that Chris Wilson requested some more information upstream? You can use the latest kernel from http://
Jane Atkinson (irihapeti) wrote : | #69 |
Geir:
No, I wasn't aware of that. Thanks for letting me know. I probably won't be able to do anything about it until tomorrow, though.
Do I do the testing with the standard xserver-
Geir Ove Myhr (gomyhr) wrote : Re: [Bug 477256] Re: [i855] Xserver freezing on Toshiba Satellite A10 (Intel 82852/855GM) | #70 |
> Do I do the testing with the standard xserver-
> need to install the xorg-edgers version for the test?
I don't think it will make a difference, but if you'd like to err on
the safe side, then use xorg-edgers. Most of the code involved is in
the kernel anyway, and these newer kernels have some debugging
capabilities that the previous ones didn't have.
In freedesktop.org Bugzilla #26326, Geir Ove Myhr (gomyhr) wrote : | #71 |
Created an attachment (id=33988)
i915_error_state from intel-drm-next kernel
This one has PGTBL_ER: 0x00000000, so it may be different from the rest. Also, IPEHR doesn't match the instruction before the ACTHD address, so I guess this is another GPU/CPU incoherency hang.
In freedesktop.org Bugzilla #26326, Geir Ove Myhr (gomyhr) wrote : | #72 |
Created an attachment (id=33989)
dmesg output from when i915_error_state was captured
In freedesktop.org Bugzilla #26326, Geir Ove Myhr (gomyhr) wrote : | #73 |
Created an attachment (id=33990)
Xorg.0.log from when i915_error_state was captured
Jane Atkinson (irihapeti) wrote : | #74 |
- i915_error_state Edit (756.9 KiB, text/plain)
I got the testing done sooner than expected. The logs are attached.
Jane Atkinson (irihapeti) wrote : | #75 |
Jane Atkinson (irihapeti) wrote : | #76 |
Jane Atkinson (irihapeti) wrote : | #77 |
The logs were copied from the laptop via ssh, as in the past, while it was still hung.
I've been having other problems with X and gdm on the laptop, but I think these are to do with recent updates and probably not just intel things.
Geir Ove Myhr (gomyhr) wrote : | #78 |
Thank you. I have uploaded them to the upstream report. The dump seemed a bit different this time from before. Before you had a page table error, which you could see as PGTBL_ER: 0x00000049 in the dump output and it would also be visible in the output of dmesg. Now it says PGTBL_ER: 0x00000000 in the i915_error_state, which means no page table error is detected. It could be that there is more than one bug here, or that some problem is fixed in the newer kernel. Could you try a couple of times and see if you get any page table errors (it's right at the start of i915_error_state). If you do, please upload the files directly to the upstream bug report.
Jane Atkinson (irihapeti) wrote : | #79 |
I've tried a couple more times, and both times show PGTBL_ER: 0x00000000
Geir Ove Myhr (gomyhr) wrote : | #80 |
Okay, thank you. Let's see if Chris can make some sense of the dump upstream, then.
Geir Ove Myhr (gomyhr) wrote : | #81 |
Irihapeti, one more thing... Do you have the possibility to test with standard Lucid (i.e no xorg-edgers or mainline kernel) if you still get freezes with "nomodeset". From Karmic, the experience is that it doesn't help, but it would be nice to know if this has changed in Lucid so we know if it may be used as a workaround.
Jane Atkinson (irihapeti) wrote : | #82 |
Sorry, it totally refuses to boot with "nomodeset". In recovery mode, I get as far as the recovery menu, choose "continue", get the message "rc.sysinit....." and then it stops.
"nomodeset" with the vesa driver used to be a workaround for me, but I haven't used it for a while. Should I test that?
Jane Atkinson (irihapeti) wrote : | #83 |
Update:
Just discovered something that might make a difference to the bug reports.
The laptop won't currently run without an xorg.conf.failsafe file.
Section "Device"
Identifier "Configured Video Device"
Driver "fbdev"
EndSection
Section "Monitor"
Identifier "Configured Monitor"
EndSection
Section "Screen"
Identifier "Default Screen"
Monitor "Configured Monitor"
Device "Configured Video Device"
EndSection
I might try a vesa version and see if it will work.
Geir Ove Myhr (gomyhr) wrote : Re: [Bug 477256] Re: [i855] Xserver freezing on Toshiba Satellite A10 (Intel 82852/855GM) | #84 |
On Fri, Mar 12, 2010 at 4:22 PM, Irihapeti <email address hidden> wrote:
> Sorry, it totally refuses to boot with "nomodeset". In recovery mode, I
> get as far as the recovery menu, choose "continue", get the message
> "rc.sysinit....." and then it stops.
Okay. It has been disabled for a while but we re-enabled it in
xserver-
or -1ubuntu13, then never mind.
> "nomodeset" with the vesa driver used to be a workaround for me, but I
> haven't used it for a while. Should I test that?
No, that shouldn't be necessary.
In freedesktop.org Bugzilla #26326, Geir Ove Myhr (gomyhr) wrote : | #85 |
(In reply to comment #5)
> This looks like the CPU/GPU coherency issue with a twist. Both gpu dumps have
> recorded PGTBL_ERR=0x49, which is unusual.
We have several automatic bug reports in ubuntu with PGTBL_ER=0x49, so it shouldn't be too unusual.
https:/
https:/
https:/
https:/
The first three are on i845 with 2.6.32 drm (which is what we have most bug reports from) and the last one is i855 on 2.6.31. I haven't seen it in on 2.6.33 drm or drm-intel-next yet.
Jane Atkinson (irihapeti) wrote : | #86 |
Live CD image of 20100419.1 needs "i915.modeset=0 xforcevesa" added to the boot command line, or it just locks up, presumably when trying to start the x server.
tags: | added: iso-testing |
C (ubuntu-caranfil) wrote : | #87 |
Problem present on Dell Latitude X300 in Lucid RC (20100419.1).
The system boots with xforcevesa.
HOWEVER the problem was NOT present in Lucid Beta1 !!!
Jane Atkinson (irihapeti) wrote : | #88 |
The need to use "i915.modeset=0 xforcevesa" still exists in 10.04 final. I hope that there's a note somewhere for people who are likely to encounter this problem.
Jane Atkinson (irihapeti) wrote : | #89 |
Also, I think that C is correct: this is a problem of fairly recent origin.
Antonio J. de Oliveira (ajoliveira) wrote : | #90 |
Hello
It is by far better on Lucid release 32-bit. The only way of booting a toshiba a10 is by putting it in the recovery mode. Neither the display is detected or the wireless adapter work, but that is the lucant-agere bug, described elsewhere. The computer worked fine on Karmic, and is unusable now.
Best regards
Jane Atkinson (irihapeti) wrote : | #91 |
@Antonio
I wouldn't go so far as to say that a Satellite A10 is unusable, but you will need a /etc/X11/xorg.conf file, using the vesa driver. You can use the configuration from message #74 but with "vesa" instead of "fbdev" as the driver.
Antonio J. de Oliveira (ajoliveira) wrote : | #92 |
ok, agreed, of course i can boot and use the stuff, and if you look at my web page, maybe you find some extra tips to get it going, but, the question is, will the regular user be able to do it?
All the best
Jane Atkinson (irihapeti) wrote : | #93 |
@Antonio
I agree with your comment about the regular user. Even finding the info can be a bit tricky. For example, I already knew about needing to use "i915.modeset=0 xforcevesa" in the boot line, but I couldn't find it from the front page of the Ubuntu website, even though I knew that there was a page of known issues somewhere. How is an anxious new user going to cope?
@Anyone
I've encountered a few users who are limited to a resolution of 640x480 or so when they use the vesa driver, and yet their machines are older ones that will do 1024x768. My A10 defaults to 1024x768. What can we do to get a higher resolution for them? Adding modelines doesn't seem to do it.
Chris Halse Rogers (raof) wrote : | #94 |
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-
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?
Changed in xserver-xorg-video-intel (Ubuntu): | |
assignee: | nobody → Chris Halse Rogers (raof) |
C (ubuntu-caranfil) wrote : | #95 |
Will the xserver-
Stenten (stenten) wrote : | #96 |
C,
That's the hope. Please install the new driver and remove your KMS-enabling workaround (so you boot into UMS) and report back. I've added specific instructions for how to do both at [1]. Let me know if you have any questions.
In freedesktop.org Bugzilla #26326, Chris Wilson (ickle) wrote : | #97 |
Another look:
0x077ac270: 0x7d8e0001: 3DSTATE_BUFFER_INFO
0x077ac274: 0x03800c40: color, tiling = fenced, pitch=3136
0x077ac278: 0x07af6000: address
Old tiling bug.
C (ubuntu-caranfil) wrote : | #98 |
- Xorg.0.log Edit (22.1 KiB, application/octet-stream)
I did install the new version but it seems that it does not fix the problems when needing i915.modeset=1
The installation was a little strange but in the end it seems that I have:
xserver-
2:2.12.
but the output from Xorg -version looks like
X.Org X Server 1.7.6
Release Date: 2010-03-17
X Protocol Version 11, Revision 0
Build Operating System: Linux 2.6.24-27-server i686 Ubuntu
Current Operating System: Linux X300 2.6.32-24-generic #39-Ubuntu SMP Wed Jul 28 06:07:29 UTC 2010 i686
Kernel command line: BOOT_IMAGE=
Build Date: 21 July 2010 12:47:34PM
xorg-server 2:1.7.6-2ubuntu7.3 (For technical support please see http://
Current version of pixman: 0.16.4
I have attached Xorg.0.log with more details ...
Jane Atkinson (irihapeti) wrote : | #99 |
I was finding that the packages from raof PPA weren't helping.
I'm currently using the glasen PPA packages successfully.
Changed in xserver-xorg-video-intel: | |
importance: | Unknown → Medium |
status: | Confirmed → Fix Released |
Changed in xserver-xorg-video-intel (Ubuntu): | |
assignee: | Chris Halse Rogers (raof) → nobody |
Changed in xserver-xorg-video-intel (Ubuntu): | |
status: | Triaged → Fix Released |
Chris Halse Rogers (raof) wrote : | #100 |
I'm marking this as fix released - upstream has identified that this was caused by an old libdrm bug.
You will still, of course, hit the GTT incoherency crashes (bug #541511), which are a separate bug.
Changed in xserver-xorg-video-intel: | |
importance: | Medium → Unknown |
Changed in xserver-xorg-video-intel: | |
importance: | Unknown → Medium |
Jeffrey Rolland (jrolland) wrote : | #101 |
This bug still affects me. I'm using an Acer Aspire <http://
Please help.
Chad Miller (cmiller) wrote : | #102 |
First time it's happened to me, after three years with this hardware. Oneiric up-to-date.
linux-image 3.0-2-generic
xserver-
xserver-
Jun 28 16:02:07 localhost kernel: [11050.384059] [drm:i915_
Jun 28 16:02:07 localhost kernel: [11050.384074] [drm] capturing error event; look for more information in /debug/
Jun 28 16:02:07 localhost kernel: [11050.391508] [drm:i915_
X.Org X Server 1.10.2
Release Date: 2011-05-28
X Protocol Version 11, Revision 0
Build Operating System: Linux 2.6.24-27-server i686 Ubuntu
Current Operating System: Linux hypatia 3.0-2-generic #3-Ubuntu SMP Fri Jun 24 19:09:43 UTC 2011 i686
Kernel command line: root=UUID=
Build Date: 16 June 2011 02:43:32AM
xorg-server 2:1.10.2-1ubuntu1 (For technical support please see http://
Current version of pixman: 0.21.8
Before reporting problems, check http://
to make sure that you have the latest version.
[ 11050.394] (EE) intel(0): Detected a hung GPU, disabling acceleration.
[ 11050.394] (EE) intel(0): When reporting this, please include i915_error_state from debugfs and the full dmesg.
[ 11083.687] [mi] EQ overflowing. The server is probably stuck in an infinite loop.
[ 11083.687]
Backtrace:
[ 11083.696] 0: /usr/bin/X (xorg_backtrace
[ 11083.696] 1: /usr/bin/X (mieqEnqueue+0x1d1) [0x80eb3d1]
[ 11083.696] 2: /usr/bin/X (xf86PostMotion
[ 11083.696] 3: /usr/lib/
[ 11083.696] 4: /usr/lib/
[ 11083.696] 5: /usr/lib/
[ 11083.696] 6: /usr/bin/X (0x8048000+0x6ba11) [0x80b3a11]
if it happens again, I'll connect from another machine and get i915_error_state state.
Proskurin Kirill (oloremo) wrote : | #103 |
Still have this bug
Ubuntu 11.04 - 2.6.38-10-generic
All up to date.
00:02.0 VGA compatible controller [0300]: Intel Corporation Core Processor Integrated Graphics Controller [8086:0042] (rev 12) (prog-if 00 [VGA controller])
xserver-
intel-gpu-
libdrm-
libdrm-
libdrm-
libdrm-
libdrm22.
It is appers only then I switching from one virtual desctop to other by Ctrl+Atl+Arrow. Screen freze and nothing helps.
I got:
[drm:i915_
[drm:i915_
Yuriy Voziy (yuretsz) wrote : | #104 |
Got this bug in 11.04
2.6.39-
[28019.516461] [drm:i915_
[28019.516472] [drm] capturing error event; look for more information in /debug/
[28019.520538] [drm:i915_
[28021.643266] [drm:i915_
[28021.643296] [drm:i915_
[28021.643447] [drm:i915_reset] *ERROR* GPU hanging too fast, declaring wedged!
[28021.643452] [drm:i915_reset] *ERROR* Failed to reset chip.
[28078.148431] [drm:drm_
[28078.148442] [drm:drm_
[28083.949848] [drm:drm_
[28083.949855] [drm:drm_
Thierry (thierry38) wrote : | #105 |
I'm also having this bug, running Ubuntu 10.04, with latest kernel available (2.6.38-10 pae).
===
=== Error message
===
[43327.520095] [drm:i915_
[43327.524194] [drm:i915_
[43329.324014] [drm:i915_
[43329.948017] [drm:i915_reset] *ERROR* Failed to reset chip.
===
=== System OS
===
Linux Raistlin4 2.6.38-
===
=== Video Hardware
===
*-pci
product: Mobile 4 Series Chipset Memory Controller Hub
vendor: Intel Corporation
physical id: 100
bus info: pci@0000:00:00.0
version: 07
width: 32 bits
clock: 33MHz
*-display:0
bus info: pci@0000:00:02.0
width: 64 bits
clock: 33MHz
*-display:1 UNCLAIMED
bus info: pci@0000:00:02.1
width: 64 bits
clock: 33MHz
===
=== Xorg main packages
===
xserver-xorg 1:7.5+5ubuntu1
xserver-
libdrm-dev 2.4.18-1ubuntu3
libdrm-intel1 2.4.18-1ubuntu3
libdrm-nouveau1 2.4.18-1ubuntu3
libdrm-radeon1 2.4.18-1ubuntu3
libdrm2 2.4.18-1ubuntu3
===
=== Xorg -version
===
X.Org X Server 1.7.6
Release Date: 2010-03-17
X Protocol Version 11, Revision 0
Build Operating System: Linux 2.6.24-27-server i686 Ubuntu
Current Operating System: Linux Raistlin4 2.6.38-
Kernel command line: BOOT_IMAGE=
Build Date: 08 March 2011 08:22:50AM
xorg-server 2:1.7.6-2ubuntu7.6 (For technical support please see http://
Current version of pixman: 0.16.4
Any...
Turgut Genc (turgutgenc) wrote : | #106 |
2011.09.21 and still having the same driver issue. I'd gladly supply traces and/or error documents.
Lenovo - Thinkpad X201
Ubuntu 11.04
2.6.38-
Everything's up-to-date.
GDM/KDM/
11083.980409] [drm:i915_
[11083.983236] [drm:i915_
[11086.913018] [drm:i915_
[11086.913033] [drm:i915_
[11086.913219] [drm:i915_reset] *ERROR* GPU hanging too fast, declaring wedged!
[11086.913222] [drm:i915_reset] *ERROR* Failed to reset chip.
I'm sick of it, heed me, save us !
oldmankit (oldmankit) wrote : | #107 |
I have been having this problem since I bought a new laptop. Switching to unity-2d has removed all crashes, however, I would like to go back to 3d if I can.
oldmankit (oldmankit) wrote : | #108 |
I believe I may have found a fix for my computer (closing the lid doesn't crash any more, and I haven't had any X crashes since doing this):
`sudo add-apt-repository ppa:ubuntu-
sudo apt-get update
sudo apt-get upgrade`
Hi irihapeti,
Thanks for including the attached files. Could you also include your /var/log/Xorg.0.log (or Xorg.0.log.old) from after reproducing the issue?
[This is an automated message. Apologies if it has reached you inappropriately; please just reply to this message indicating so.]