WARNING: at /build/buildd/linux-2.6.28/drivers/gpu/drm/i915/i915_gem.c:2475 i915_gem_idle+0x2ec/0x310 [i915]()

[Martin Olsson]

Intel G45 (and possibly other chipsets) always print the following error into dmesg during shutdown (happens on every single shutdown and it dectected by apport during the next boot):

WARNING: at /build/buildd/linux-2.6.28/drivers/gpu/drm/i915/i915_gem.c:2475 i915_gem_idle+0x2ec/0x310 [i915]()
Modules linked in: vboxnetflt vboxdrv binfmt_misc i915 drm ppdev bnep ipt_MASQUERADE iptable_nat nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_state nf_conntrack ipt_REJECT xt_tcpudp iptable_filter ip_tables x_tables bridge stp kvm_intel kvm input_polldev video output coretemp it87 hwmon_vid lp parport joydev snd_hda_intel snd_usb_audio snd_pcm_oss snd_mixer_oss iTCO_wdt iTCO_vendor_support snd_seq_dummy intel_agp snd_pcm usbhid snd_seq_oss snd_usb_lib snd_seq_midi snd_seq_midi_event snd_rawmidi snd_seq snd_timer snd_seq_device snd_hwdep psmouse snd soundcore snd_page_alloc serio_raw pcspkr pata_it8213 r8169 mii floppy fbcon tileblit font bitblit softcursor [last unloaded: vboxdrv]
Pid: 4192, comm: Xorg Not tainted 2.6.28-11-generic #35-Ubuntu
Call Trace:
 [<ffffffff802509bf>] warn_on_slowpath+0x5f/0x90
 [<ffffffffa0524034>] ? i915_gem_retire_requests+0x114/0x170 [i915]
 [<ffffffffa052618c>] i915_gem_idle+0x2ec/0x310 [i915]
 [<ffffffffa0526203>] i915_gem_leavevt_ioctl+0x23/0x50 [i915]
 [<ffffffffa050180a>] drm_ioctl+0x10a/0x330 [drm]
 [<ffffffff8021bb60>] ? restore_i387_xstate+0x120/0x150
 [<ffffffffa05261e0>] ? i915_gem_leavevt_ioctl+0x0/0x50 [i915]
 [<ffffffff802f640d>] vfs_ioctl+0x7d/0xa0
 [<ffffffff802f6775>] do_vfs_ioctl+0x75/0x230
 [<ffffffff802f69c9>] sys_ioctl+0x99/0xa0
 [<ffffffff8021253a>] system_call_fastpath+0x16/0x1b
---[ end trace 30147e9762b57ebd ]---

I've seen this error+stacktrace in dmesg every now and then but right around when intel ddx driver 2.6.3 hit jaunty repros this become 100% reproducible (i.e. happens every shutdown).

ProblemType: KernelOops
Annotation: Your system might become unstable now and might need to be restarted.
Architecture: amd64
DistroRelease: Ubuntu 9.04
Failure: oops
MachineType: Gigabyte Technology Co., Ltd. EG45M-DS2H
Package: linux-image-2.6.28-11-generic 2.6.28-11.35
ProcCmdLine: root=UUID=0ceb2ba0-df2e-4da2-a28c-3be483363631 ro quiet splash crashkernel=384M-2G:64M@16M,2G-:128M@16M
ProcVersionSignature: Ubuntu 2.6.28-11.35-generic
SourcePackage: linux
Title: WARNING: at /build/buildd/linux-2.6.28/drivers/gpu/drm/i915/i915_gem.c:2475 i915_gem_idle+0x2ec/0x310 [i915]()

[Jaco (jacotb)]

Same for me, on x86. (HP Pavilion dv1668ea; Core Duo; Intel GMA 945)

[Gregory Oschwald (osch0001)]

Same here on a Thinkpad x60s.

[warp (naquadareactor)]

Már akartam kérdeni h akkor hogy is lesz ez, mert pénteken nagyszülők
jönnek át hozzájuk.
Menni tudok persze de ő?

[warp (naquadareactor)]

The last comment was got here by error.
(Accidentally replied to the bug subscr.mail, it does not contain any personal information, but should be deleted)
Anyway, i got the same error, on a Thinkpad T500

[Emiel Kollof (emiel-kollof)]

Happens to me on my EeePC 900

[hardtoneselector (info-franken-reggae)]

hmmm it always happens if I play with firefox a ajax based game and once it happen if I played the same game in seamonkey, the games is travianer.sevengames.de

it happend never If I didn't played this game?!?

[AphoxemaG (xeristian)]

Same on a Wind U100

[Martin Olsson (mnemo)]

Still happens after mesa7.4 upload today.

[Martin Olsson (mnemo)]

I've investigated this some more... I wasn't seeing this error in "dmesg" so I wondered if the error was valid. However, after poking around a bit I understand that the error actually happens every time the machine shuts down. kerneloops and apport both will not report it until the next time the machine boots though and then of course the dmesg is clean.

This is what my /var/log/kern.log looks like:

Apr 2 17:04:17 kingfish kernel: [ 27.992308] r8169: eth0: link up
Apr 2 17:04:24 kingfish kernel: [ 35.309640] 2:2:1: cannot get freq at ep 0x84
Apr 2 17:04:27 kingfish kernel: [ 38.477507] eth0: no IPv6 routers present
Apr 2 18:20:22 kingfish kernel: [ 4593.765194] [drm:i915_gem_object_bind_to_gtt] *ERROR* GTT full, but LRU list empty
Apr 2 18:20:22 kingfish kernel: [ 4593.765199] [drm:i915_gem_object_pin] *ERROR* Failure to bind: -12<3>[drm:i915_gem_evict_something] *ERROR* inactive empty 1 request empty 1 fl$
Apr 2 21:30:07 kingfish kernel: [15978.632809] process `skype.real' is using obsolete setsockopt SO_BSDCOMPAT
Apr 3 09:04:00 kingfish kernel: [57611.436564] hda-intel: IRQ timing workaround is activated for card #0. Suggest a bigger bdl_pos_adj.
Apr 3 09:04:14 kingfish kernel: [57625.563513] 2:2:1: cannot get freq at ep 0x84
Apr 3 12:50:42 kingfish kernel: [71213.459685] npviewer.bin[8144]: segfault at ff9bea2c ip 00000000ff9bea2c sp 00000000ffed2a8c error 14
Apr 3 14:30:52 kingfish kernel: [77223.579987] [drm:i915_gem_object_bind_to_gtt] *ERROR* GTT full, but LRU list empty
Apr 3 14:30:52 kingfish kernel: [77223.579992] [drm:i915_gem_object_pin] *ERROR* Failure to bind: -12<3>[drm:i915_gem_evict_something] *ERROR* inactive empty 1 request empty 1 fl$
Apr 3 14:31:17 kingfish kernel: [77248.591892] ------------[ cut here ]------------
Apr 3 14:31:17 kingfish kernel: [77248.591895] WARNING: at /build/buildd/linux-2.6.28/drivers/gpu/drm/i915/i915_gem.c:2475 i915_gem_idle+0x2ec/0x310 [i915]()
Apr 3 14:31:17 kingfish kernel: [77248.591897] Modules linked in: i915 drm binfmt_misc ppdev bnep vboxnetflt vboxdrv ipt_MASQUERADE iptable_nat nf_nat nf_conntrack_ipv4 nf_defrag$
Apr 3 14:31:17 kingfish kernel: [77248.591943] Pid: 3414, comm: Xorg Not tainted 2.6.28-11-generic #38-Ubuntu
Apr 3 14:31:17 kingfish kernel: [77248.591945] Call Trace:
Apr 3 14:31:17 kingfish kernel: [77248.591952] [<ffffffff802509bf>] warn_on_slowpath+0x5f/0x90
Apr 3 14:31:17 kingfish kernel: [77248.591958] [<ffffffffa0523fe4>] ? i915_gem_retire_requests+0x114/0x170 [i915]
Apr 3 14:31:17 kingfish kernel: [77248.591963] [<ffffffffa052611c>] i915_gem_idle+0x2ec/0x310 [i915]
Apr 3 14:31:17 kingfish kernel: [77248.591968] [<ffffffffa0526193>] i915_gem_leavevt_ioctl+0x23/0x50 [i915]
Apr 3 14:31:17 kingfish kernel: [77248.591979] [<ffffffffa05017fa>] drm_ioctl+0x10a/0x330 [drm]
Apr 3 14:31:17 kingfish kernel: [77248.591983] [<ffffffff8021bb70>] ? restore_i387_xstate+0x120/0x150
Apr 3 14:31:17 kingfish kernel: [77248.591988] [<ffffffffa0526170>] ? i915_gem_leavevt_ioctl+0x0/0x50 [i915]
Apr 3 14:31:17 kingfish kernel: [77248.591992] [<ffffffff802f631d>] vfs_ioctl+0x7d/0xa0
Apr 3 14:31:17 kingfish kernel: [77248.591995] [<ffffffff802f6685>] do_...

[Michael Gratton (mjog)]

Are any of you using the xorg-edgers PPA? I think this started happening around the time I switched to that, and might be related to VT switching.

[Zack Evans (zevans23)]

More consistent for me on recent kernels too. I don't get an "oops" though, just a total console freeze. I can still log in remotely and get shut down cleanly but only if I kill the X server process itself.

No relation to VT switching - there used to be problems but those bugs all solved for me some time ago and I think unrelated.

I am using the xorg-edgers PPA but I see the lockups if I switch back to Jaunty versions.

I see the lockups under a variety of kernels including
- the most recent Ubuntu 2.6.28-11.40
- Various 2.6.29 builds including the Ubuntu RC1 version
- My build of 2.6.30RC1
Useful debug info depends on which kernel it happened with - in my build I have some of the watchdogs switched on which may help? For instance:

Apr 8 16:31:03 vademecum kernel: [ 0.000000] Linux version 2.6.30-rc1ze (root@vademecum) (gcc version 4.3.3 (Ubuntu 4.3.3-5ubuntu4) ) #1 SMP Wed Apr 8 12:16:38 BST 2009
Apr 8 17:20:20 vademecum kernel: [ 3000.768092] INFO: task events/0:9 blocked for more than 120 seconds.
Apr 8 17:20:20 vademecum kernel: [ 3000.768108] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
Apr 8 17:20:20 vademecum kernel: [ 3000.768122] events/0 D 0000029d 0 9 2
Apr 8 17:20:20 vademecum kernel: [ 3000.768141] f7875f40 00000046 3c688907 0000029d f7b22fc0 00000000 00000000 c0568d24
Apr 8 17:20:20 vademecum kernel: [ 3000.768170] c056c7b0 f7836b70 f7836de8 c17097b0 00000000 0009d202 0901f1fd 0000029d
Apr 8 17:20:20 vademecum kernel: [ 3000.768198] c056c7b0 ee0fc700 f7836b70 c1705d24 f7836de8 5eefa7b0 00000278 00000000
Apr 8 17:20:20 vademecum kernel: [ 3000.768225] Call Trace:
Apr 8 17:20:20 vademecum kernel: [ 3000.768256] [<c03ad7cc>] __mutex_lock_slowpath+0xc0/0x124
Apr 8 17:20:20 vademecum kernel: [ 3000.768276] [<c0132f39>] ? add_timer+0x11/0x18
Apr 8 17:20:20 vademecum kernel: [ 3000.768292] [<c03ad65a>] mutex_lock+0x1a/0x28
Apr 8 17:20:20 vademecum kernel: [ 3000.768337] [<f8d01d66>] i915_gem_retire_work_handler+0x22/0x58 [i915]
Apr 8 17:20:20 vademecum kernel: [ 3000.768357] [<c013933f>] worker_thread+0x13f/0x1f0
Apr 8 17:20:20 vademecum kernel: [ 3000.768396] [<f8d01d44>] ? i915_gem_retire_work_handler+0x0/0x58 [i915]
Apr 8 17:20:20 vademecum kernel: [ 3000.768415] [<c013c910>] ? autoremove_wake_function+0x0/0x34
Apr 8 17:20:20 vademecum kernel: [ 3000.768433] [<c0139200>] ? worker_thread+0x0/0x1f0
Apr 8 17:20:20 vademecum kernel: [ 3000.768449] [<c013c5e3>] kthread+0x37/0x6c
Apr 8 17:20:20 vademecum kernel: [ 3000.768464] [<c013c5ac>] ? kthread+0x0/0x6c
Apr 8 17:20:20 vademecum kernel: [ 3000.768481] [<c0103bcf>] kernel_thread_helper+0x7/0x10
Apr 8 17:27:25 vademecum kernel: [ 3425.968132] [drm:i915_gem_idle] *ERROR* hardware wedged

I don't think I get the "wedged" message until I have killed the X server from a remote shell.

Just to demonstrate this happens on a known kernel (and so can report it here as a bug...)
Apr 9 11:44:49 vademecum kernel: [ 0.000000] Linux version 2.6.28-11-generic (buildd@palmer) (gcc version 4.3.3 (Ubuntu 4.3.3-5ubuntu4) ) #40-Ubuntu SMP Fri Apr 3 17:39:51 UTC 2009 (Ubuntu 2.6.28-11.40-generic)
Apr 9 12:24:07 vade...

[Leann Ogasawara (leannogasawara)]

Hi martin,

The following upstream LKML thread seems related to the issue you are seeing:

http://lkml.org/lkml/2009/1/26/323

[Leann Ogasawara (leannogasawara)]

Hi martin,

Just curious if you'd be willing to give the 2.6.30-rc2 Mainline Kernel Build a test - https://wiki.ubuntu.com/KernelMainlineBuilds . I believe it has some i915 related patches. Would be great to know if this helps with your issue or not. Please let us know. Thanks.

[Martin Olsson (mnemo)]

The LKML thread ( http://lkml.org/lkml/2009/1/26/323 ) describes a completely different issue. I've seen both on my intel machine but this bug report only deals with the warn_on_slowpath / i915_gem_leavevt_ioctl dmesg error printed during shutdown.

PS. I also tried 2.6.30-rc2 and then I could not repro the bug, which is nice.

[Martin Olsson (mnemo)]

I also see this issue on karmic with xorg-edgers as of May 16th 2009.

[Leann Ogasawara (leannogasawara)]

Does this issue remain with the latest 2.6.31-4 Karmic kernel?

[Ismail Gjevori (isgjevori)]

i have a fresh install of karmic alpha 3 and it does not happen again.

[Martin Olsson (mnemo)]

Tried again now and it seems it's fixed in latest-karmic-as-of-today.