Freeze with KMS enabled on Radeon RV350

[Yves Glodt]

When I boot lucid with KMS enabled on my Acer Travelmate 8000 (which has an Ati Mobility 9700 (RV350)), X get's corrupted and does not respond to input anymore on KDM launch.

When I switch to a VT I see the following messages:

Mar 12 21:06:13 sparrow kernel: [ 120.041087] [drm:radeon_cs_ioctl] *ERROR* Faild to schedule IB !
Mar 12 21:06:13 sparrow kernel: [ 120.053581] [drm:radeon_ib_schedule] *ERROR* radeon: couldn't schedule IB(7).
Mar 12 21:06:13 sparrow kernel: [ 120.053586] [drm:radeon_cs_ioctl] *ERROR* Faild to schedule IB !
Mar 12 21:06:13 sparrow kernel: [ 120.053926] [drm:radeon_ib_schedule] *ERROR* radeon: couldn't schedule IB(8).

(messages copied from https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-ati/+bug/507148/comments/61 but mine are the same)

Somehow the nomodeset switch stopped working with the -18 kernel, so I am forced to use the -17 as of now which works fine, provided I keep KMS disabled. The -20 kernel does still not work with KMS btw.

[Lollerke (pumba88)]

Use the "radeon.modeset=0" command instead of "nomodeset". It works with the -20 kernel.

[Yves Glodt (yglodt)]

Looks like the new -21 kernel solved the problem for me!

I could boot with KMS and it runs stable for ~ 1.5 hours. Will see if it remains stable throughout the day

[Yves Glodt (yglodt)]

Sorry, today I had freezes again, immediately at kdm, or shortly after white KDE started up. Back to radeon.modeset=0 for now.

This is with all packages updated today.

[Yves Glodt (yglodt)]

My hardware is:

ATI Technologies Inc RV350 [Mobility Radeon 9600 M10]

[Ancoron Luziferis (ancoron)]

Looks exactly like my problem (although I didn't ever get this directly after X start): https://bugs.launchpad.net/ubuntu/+source/linux/+bug/561554

[Jeremy Foshee (jeremyfoshee)]

Hi Yves,

Please be sure to confirm this issue exists with the latest development release of Ubuntu. ISO CD images are available from http://cdimage.ubuntu.com/releases/ . If the issue remains, please run the following command from a Terminal (Applications->Accessories->Terminal). It will automatically gather and attach updated debug information to this report.

apport-collect -p linux 561789

Also, if you could test the latest upstream kernel available that would be great. It will allow additional upstream developers to examine the issue. Refer to https://wiki.ubuntu.com/KernelMainlineBuilds . Once you've tested the upstream kernel, please remove the 'needs-upstream-testing' tag. This can be done by clicking on the yellow pencil icon next to the tag located at the bottom of the bug description and deleting the 'needs-upstream-testing' text. Please let us know your results.

Thanks in advance.

    [This is an automated message. Apologies if it has reached you inappropriately; please just reply to this message indicating so.]

[Yves Glodt (yglodt)]

Hi Jeremy,

I still have this issue, with the -21 and also the -22 kernel.

I executed apport-collect, but fail to see whether it posted the information correctly.

[Jeremy Foshee (jeremyfoshee)]

Yves,
    It doesn't look as if it has. Would you mind trying again?

Thanks!

~JFo

[Leonidas Spyropoulos (artafinde)]

Hey, I got the same problem.
Even with the LiveCD
I had to xforcevesa on LiveCD to make the installation.
Also after the install I made an xorg.conf and load the vesa driver instead because the radeon one keeps freezing.
If I remove the xorg file then the laptop after some time deadlocks (keyboard and clicks, I can move the mouse around though)

Adding the option radeon.modset=0 is not helpful

I attach the lspci, /var/log/dmesg /var/log/messages and Xorg.0.log files

[Leonidas Spyropoulos (artafinde)]

If I modify /etc/modptobe.d/radeon.conf and make option radeon.modset=1 then I can change VT (tty1, tty2 ..) but I have no input on X, neither clicks or keyboard)
All my VT are full of
[drm:radeon_ib_schedule] *ERROR* radeon: couldn't schedule IB(1-15)
[drm:radeon_cs_ioctl] *ERROR* Feailed to schedule IB !

The numbers in the parenthesis differ each time and cycle from 1 to 15, my guess is the IRQ requests

[Didier Misson (dmlinux)]

Seem same problem on Packard Bell EasyNote R7725.
chipset : ATI RV 350 Mobility Radeon 9600 M10

Ubuntu 10.04 just install. No problem with installation.
Problem occurs after reboot.
logon near impossible. Screen display is corrupt.

In console we have a lot of message :

[drm: radeon_ib_schedule] *ERROR* radeon: couldn't schedule IB (8)
[drm: radeon_cs_ioctl] *ERROR* Failed to schedule IB !

same 2 msgs repeat...

I think it was ok before I enable Compiz 3D effect
These 3D effects were ok until we reboot ...

ps : difficult to provide documentation on this PC due to this problem.

[Didier Misson (dmlinux)]

try to uninstall (aptitude remove) compiz and compiz-core
but it doesn't solve the problem
Always error messages and display corrupted.

[Leonidas Spyropoulos (artafinde)]

It's the same for me too, it's difficult to provide documentation from ubuntu installation.
After each freeze I log into the debian installation and mount the partition just to copy the logs.
Try a working livecd for accessing the logs ;)

[Didier Misson (dmlinux)]

I succeed to boot in editing Grub2 menu to have the rescue mode by default.

- We also have an other bug on the Ubuntu 10.04 Live CD : mouse cursor was moving but NOT possible to click !
problem occurs only with the first line "Test Ubuntu 10.04 without installing"
After, I reboot on the CD and choose direcly the 2 line "Install Ubuntu" and no problem with the mouse...

---------

Ok, I send you the logs about this Radeon 9600 M10

[Didier Misson (dmlinux)]

Thanks ;-)

[Yves Glodt (yglodt)]

Jeremy, I am sorry but apport-collect does not want to post, it gives an "401 Unauthorized". How can I reset or re-specify my login data?

yves@yves-laptop:~$ apport-collect -p linux 561789
Traceback (most recent call last):
  File "/usr/share/apport/apport-kde", line 461, in <module>
    UserInterface = MainUserInterface()
  File "/usr/share/apport/apport-kde", line 175, in __init__
    sys.exit(self.run_argv())
  File "/usr/lib/python2.6/dist-packages/apport/ui.py", line 555, in run_argv
    return self.run_update_report()
  File "/usr/lib/python2.6/dist-packages/apport/ui.py", line 489, in run_update_report
    attachment_comment='apport information')
  File "/usr/lib/python2.6/dist-packages/apport/crashdb_impl/launchpad.py", line 346, in update
    bug.lp_save()
  File "/usr/lib/pymodules/python2.6/lazr/restfulclient/resource.py", line 638, in lp_save
    URI(self.self_link), representation, headers)
  File "/usr/lib/pymodules/python2.6/lazr/restfulclient/_browser.py", line 329, in patch
    'PATCH', extra_headers=extra_headers)
  File "/usr/lib/pymodules/python2.6/lazr/restfulclient/_browser.py", line 269, in _request
    raise HTTPError(response, content)
lazr.restfulclient.errors.HTTPError: HTTP Error 401: Unauthorized
Response headers:
---
content-length: 51
content-type: text/plain
date: Mon, 03 May 2010 07:27:25 GMT
server: zope.server.http (HTTP)
status: 401
via: 1.1 wildcard.edge.launchpad.net
x-lazr-oopsid: OOPS-1584EB889
x-powered-by: Zope (www.zope.org), Python (www.python.org)
---
Response body:
---
(<Bug at 0x2aaab17acf10>, 'tags', 'launchpad.Edit')
---

yves@yves-laptop:~$

[Minimec (minimec)]

I was working with Leonidas on that problem on the #ubuntu freenode channel. I am really surprised to see that bug right now, as I own the same card in a Pentium M Centrino Environment like Leonidas and probably some others here.

I am not affected by this bug, using the same card!

Nevertheless that bug reminds me of a comment I made on bug 413249 (https://bugs.launchpad.net/xserver-xorg-driver-ati/+bug/413259/comments/30). You could try to play around with AGP/KMS settings in /etc/modprobe.d/, as described in the mentioned link.

 I repeat that I am not affected by the bug and that I have better GPU performance than ever with that ATI 9600 mobility.

I added my lspci for comparison.

[Didier Misson (dmlinux)]

very strange... You have exactly the same detection for the ATI 9600
Mobility M10 in lspci...
but you don't have any problem !

ps : I don't have easyly access to this laptop.
It's not mine...

I bypass the problem in forcing the 2nd GRUB2 entry "rescue mode" (no
problem with graphic)

Thanks,

;-)

2010/5/3 Martin Tschäppät <email address hidden>

> I was working with Leonidas on that problem on the #ubuntu freenode
> channel. I am really surprised to see that bug right now, as I own the
> same card in a Pentium M Centrino Environment like Leonidas and probably
> some others here.
>
> I am not affected by this bug, using the same card!
>
> Nevertheless that bug reminds me of a comment I made on bug 413249
> (https://bugs.launchpad.net/xserver-xorg-driver-
> ati/+bug/413259/comments/30). You could try to play around with AGP/KMS
> settings in /etc/modprobe.d/, as described in the mentioned link.
>
> I repeat that I am not affected by the bug and that I have better GPU
> performance than ever with that ATI 9600 mobility.
>
> I added my lspci for comparison.
>
>
> ** Attachment added: "lspci.txt"
> http://launchpadlibrarian.net/47522163/lspci.txt
>
> --
> Freeze with KMS enabled on Radeon RV350
> https://bugs.launchpad.net/bugs/561789
> You received this bug notification because you are a direct subscriber
> of the bug.
>

--
Didier

[Minimec (minimec)]

> I bypass the problem in forcing the 2nd GRUB2 entry "rescue mode" (no
> problem with graphic)

Hmmm... That would mean, that it could also be a plymouth/xorg problem...

Even then it is surprising, that I am not affected...

2010/5/3 Didier Misson <email address hidden>:
> very strange... You have exactly the same detection for the ATI 9600
> Mobility M10 in lspci...
> but you don't have any problem !
>
> ps : I don't have easyly access to this laptop.
> It's not mine...
>
> I bypass the problem in forcing the 2nd GRUB2 entry "rescue mode" (no
> problem with graphic)
>
> Thanks,
>
> ;-)
>
>
> 2010/5/3 Martin Tschäppät <email address hidden>
>
>> I was working with Leonidas on that problem on the #ubuntu freenode
>> channel. I am really surprised to see that bug right now, as I own the
>> same card in a Pentium M Centrino Environment like Leonidas and probably
>> some others here.
>>
>> I am not affected by this bug, using the same card!
>>
>> Nevertheless that bug reminds me of a comment I made on bug 413249
>> (https://bugs.launchpad.net/xserver-xorg-driver-
>> ati/+bug/413259/comments/30). You could try to play around with AGP/KMS
>> settings in /etc/modprobe.d/, as described in the mentioned link.
>>
>>  I repeat that I am not affected by the bug and that I have better GPU
>> performance than ever with that ATI 9600 mobility.
>>
>> I added my lspci for comparison.
>>
>>
>> ** Attachment added: "lspci.txt"
>>   http://launchpadlibrarian.net/47522163/lspci.txt
>>
>> --
>> Freeze with KMS enabled on Radeon RV350
>> https://bugs.launchpad.net/bugs/561789
>> You received this bug notification because you are a direct subscriber
>> of the bug.
>>
>
>
> --
> Didier
>
> --
> Freeze with KMS enabled on Radeon RV350
> https://bugs.launchpad.net/bugs/561789
> You received this bug notification because you are a direct subscriber
> of the bug.
>

[brk (reachbach)]

I too am experiencing this problem on an ATI Radeon 9600 with 10.04.

[Minimec (minimec)]

I suggest, that you all try the workaround, Didier proposed.

Here's how to do that:

1. Press the power button of your computer
2. After the BIOS screen quickly press and hold the left shift key <-- http://en.wikipedia.org/wiki/Shift_key
3. Choose the recovery kernel (normally 2nd entry in the grub list
4. The recovery boot process will then start and after a while you will have a menu (blue background)
5. Choose the first entry 'resume' to resume the boot process --> boot process will then finish and you see a console login prompt
6. Login with your username and password
7. type sudo start gdm && exit <-- That command will 'start gdm' and log you out off the console, you are in ('exit')
8. You should now have a graphical interface.

If that workaround is working for you, then imho this bug is not related to the opensource ATI driver, but maybe related to plymouth involved in the boot process.

[moriddin (moriddin)]

Same problems here with an acer travelmate 8000 with an ati mobility radeon 9600 m10 (rv350)

The login screen appears normal and I can move around with the mouse cursor here.
After the login i can see the netbook launcher load most of the time and then a few seconds after that the notebook is completely stuck.
I have to turn the computer off by holding the power button down for a hard power off.

I've tried all the things that are mentioned in this thread:
radeon.modeset : no change at all in behavior
kernel nomodeset : X is completely black with only a pink square following the cursor.
normal text boot (through recovery) and start gdm : x loads and the computer freezes like before.

The only work around that worked for me was to just use the vesa driver (slow but stable).

Some extra info:
Last week I tried to upgrade from 9.04 to 9.10 and then 10.04 however this problem appeared in 9.10 so I tried a clean lucid install after it. Too bad this didn't work out either.

Since the computer is completely unresponsive (also over network) I have no idea how to get a helpful errorlog. Also the system doesn't mention a bad shutdown.

[Leonidas Spyropoulos (artafinde)]

I finally got it to work using radeon driver.
This is my radeon.conf file.

options radeon agpmode=-1
options radeon modeset=1

[brk (reachbach)]

I saw exactly the same symptoms as described by another reporter: The screen is completely blank with only a pink square in the place of the mouse pointer.

[moriddin (moriddin)]

Leonidas thanks,

For the moment those settings were exactly what was needed.

Everything seems to work as intended (even compiz). Stil only running for about 10 min now but I feel confident since this is the first time I can use hardware acceleration after login.

[Leonidas Spyropoulos (artafinde)]

Mine is working fine. It should be something about my specific hardware
Acer Travelmate 3201 XMi and with IRQ requests
My guess is that working on AGP can't specify an IRQ for the radeon driver (which is needed on kms support (?) )
Again this is a guess from the knowledge i have and the errors 1-15 which match the IRQ on motherboards available.
So working as pci by adding the "options radeon agpmode=-1" option provide the nessesary IRQ for the driver.

@moriddin:
As for the issue about the 10min I cannot help
I have an uptime of days now :)
Although I have to add I am not running many 3D accelaration software. Only some effect from compiz and video mostly because I have a hardware error with artifacts as well (I say it's hardware because I find it in vesa driver also and in windows). So I don't push my luck ;)
Try get the logs from a livecd after laptop stuck.

PS: My suspend is not working but I don't think it will work ever, so I cannot test if driver is loaded correctly after suspend. When I press the suspend the pc looks like is closing down only to prompt me for password as like is it just opened from suspend. I might check it sometime laters, not my priority now.

[Didier Misson (dmlinux)]

Leonidas,

I cannot test now. I will have the laptop next Friday.
I will try this and report if ok or not.

Thanks for your help.

[davidyu (yuyich)]

Amazing, it works great in my thinkpad T60.
I add only one line:
options radeon agpmode=-1
in /etc/modprobe.d/radeon.conf

 This solve my suspend problem.
Good job.

[Jeremy Foshee (jeremyfoshee)]

This bug report was marked as Incomplete and has not had any updated comments for quite some time. As a result this bug is being closed. Please reopen if this is still an issue in the current Ubuntu development release http://cdimage.ubuntu.com/daily-live/current/ . Also, please be sure to provide any requested information that may have been missing. To reopen the bug, click on the current status under the Status column and change the status back to "New". Thanks.

[This is an automated message. Apologies if it has reached you inappropriately; please just reply to this message indicating so.]

[madbiologist (me-again)]

This is finally fixed in the upstream 3.10 kernel by commit e49f3959a96dc279860af7e86e6dbcfda50580a5 and in the 3.9.6 kernel by commit e335ccae2ae4caef26a1e81711c44aec1cc83bd2

[madbiologist (me-again)]

The Ubuntu 13.10 "Saucy Salamander" kernel is now based on the upstream 3.9.6 kernel.

[madbiologist (me-again)]

The abovementioned patch is now available in the upstream 3.2.47 kernel as commit 62f5e8156da4e5373a10c7c1d15a4327283f64f2

The progress of having this kernel SRU'd to Ubuntu 12.04 "Precise Pangolin" can be followed at https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1193029