i915 GPU hang with WARNING: at /build/buildd/linux-3.2.0/drivers/gpu/drm/i915/i915_drv.c:413 __gen6_gt_wait_for_fifo+0x94/0xa0 [i915]()

Attaching output of 'apport-bug libdrm-intel1 --save' ran before I rebooted and ran `apport-bug linux`.

Pasting relevant dmesg crash output from the apport-bug file here:

 [102164.197484] ------------[ cut here ]------------
 [102164.197530] WARNING: at /build/buildd/linux-3.2.0/drivers/gpu/drm/i915/i915_drv.c:413 __gen6_gt_wait_for_fifo+0x94/0xa0 [i915]()
 [102164.197537] Hardware name: 4294CTO
 [102164.197541] Modules linked in: parport_pc ppdev usbhid hid ip6table_filter ip6_tables ebtable_nat ebtables xt_state ipt_REJECT xt_CHECKSUM iptable_mangle xt_tcpudp iptable_filter snd_hda_codec_hdmi snd_hda_codec_conexant joydev ipt_MASQUERADE iptable_nat nf_nat nf_conntrack_ipv4 nf_conntrack nf_defrag_ipv4 ip_tables x_tables bridge kvm_intel stp kvm rfcomm bnep binfmt_misc arc4 snd_hda_intel snd_hda_codec snd_hwdep snd_pcm snd_seq_midi snd_rawmidi psmouse serio_raw thinkpad_acpi nvram snd_seq_midi_event tpm_tis wmi mac_hid wacom snd_seq iwlwifi mac80211 snd_timer snd_seq_device btusb mei(C) bluetooth cfg80211 i915 snd drm_kms_helper drm soundcore snd_page_alloc i2c_algo_bit video coretemp lp parport e1000e sdhci_pci sdhci btrfs zlib_deflate libcrc32c
 [102164.197653] Pid: 19024, comm: Xorg Tainted: G C 3.2.0-20-generic #33-Ubuntu
 [102164.197658] Call Trace:
 [102164.197675] [<ffffffff8106726f>] warn_slowpath_common+0x7f/0xc0
 [102164.197683] [<ffffffff810672ca>] warn_slowpath_null+0x1a/0x20
 [102164.197704] [<ffffffffa01939d4>] __gen6_gt_wait_for_fifo+0x94/0xa0 [i915]
 [102164.197725] [<ffffffffa0194055>] i915_write32+0xe5/0xf0 [i915]
 [102164.197761] [<ffffffffa01d0241>] ring_write_tail+0x21/0x30 [i915]
 [102164.197793] [<ffffffffa01d29c6>] gen6_render_ring_flush+0xd6/0xf0 [i915]
 [102164.197818] [<ffffffffa01a555e>] i915_gem_flush_ring.part.26+0x3e/0xd0 [i915]
 [102164.197841] [<ffffffffa01a5f99>] i915_gem_flush_ring+0x19/0x20 [i915]
 [102164.197866] [<ffffffffa01aaf60>] i915_gem_execbuffer_move_to_gpu+0x1b0/0x200 [i915]
 [102164.197890] [<ffffffffa01ab64b>] i915_gem_do_execbuffer.isra.8+0x69b/0x940 [i915]
 [102164.197902] [<ffffffff8108f02b>] ? __hrtimer_start_range_ns+0x16b/0x3a0
 [102164.197926] [<ffffffffa01abdb3>] i915_gem_execbuffer2+0xa3/0x270 [i915]
 [102164.197950] [<ffffffffa01295d4>] drm_ioctl+0x444/0x510 [drm]
 [102164.197974] [<ffffffffa01abd10>] ? i915_gem_execbuffer+0x420/0x420 [i915]
 [102164.197984] [<ffffffff8101dbf4>] ? restore_user_xstate+0x54/0xa0
 [102164.197993] [<ffffffff81189d2a>] do_vfs_ioctl+0x8a/0x340
 [102164.198001] [<ffffffff8118a071>] sys_ioctl+0x91/0xa0
 [102164.198010] [<ffffffff81664502>] system_call_fastpath+0x16/0x1b
 [102164.198016] ---[ end trace dd6b4ef05c93f21a ]---
 [102164.203168] ------------[ cut here ]------------
 [102164.203188] WARNING: at /build/buildd/linux-3.2.0/drivers/gpu/drm/i915/i915_drv.c:413 __gen6_gt_wait_for_fifo+0x94/0xa0 [i915]()
 [102164.203194] Hardware name: 4294CTO
 [102164.203197] Modules linked in: parport_pc ppdev usbhid hid ip6table_filter ip6_tables ebtable_nat ebtables xt_state ipt_REJECT xt_CHECKSUM iptable_mangle xt_tcpudp iptable_filter snd_hda_codec_hdmi snd_hda_codec_conexant joydev ipt_MASQUERADE iptable_nat nf_nat nf_conntrack_ipv4 nf_conntrack nf_defrag_ipv4 ip_tables x_tables bridge kvm_intel stp kvm rfcomm bnep binfmt_misc arc4 snd_hda_...

Hi Robbie,

Can you run a test to see if this bug is related to RC6, by disabling it?

To disable RC6, do the following:
Hold down Left-Shift key to enter the Grub menu on boot
Hit 'e' to edit the kernel command line
Append i915.i915_enable_rc6=0 as a kernel boot parameter
For example: linux /boot/vmlinuz-3.2.0-17-generic <...> quiet splash vt.handoff=7 i915.i915_enable_rc6=0
Press Ctrl-x to boot

Thanks in advance.

Thank you for taking the time to file a bug report on this issue.

However, given the number of bugs that the Kernel Team receives during any development cycle it is impossible for us to review them all. Therefore, we occasionally resort to using automated bots to request further testing. This is such a request.

We have noted that there is a newer version of the development kernel than the one you last tested when this issue was found. Please test again with the newer kernel and indicate in the bug if this issue still exists or not.

You can update to the latest development kernel by simply running the following commands in a terminal window:

    sudo apt-get update
    sudo apt-get dist-upgrade

If the bug still exists, change the bug status from Incomplete to Confirmed. If the bug no longer exists, change the bug status from Incomplete to Fix Released.

If you want this bot to quit automatically requesting kernel tests, add a tag named: bot-stop-nagging.

 Thank you for your help, we really do appreciate it.

Thank you for taking the time to file a bug report on this issue.

However, given the number of bugs that the Kernel Team receives during any development cycle it is impossible for us to review them all. Therefore, we occasionally resort to using automated bots to request further testing. This is such a request.

We have noted that there is a newer version of the development kernel than the one you last tested when this issue was found. Please test again with the newer kernel and indicate in the bug if this issue still exists or not.

You can update to the latest development kernel by simply running the following commands in a terminal window:

    sudo apt-get update
    sudo apt-get dist-upgrade

If the bug still exists, change the bug status from Incomplete to Confirmed. If the bug no longer exists, change the bug status from Incomplete to Fix Released.

If you want this bot to quit automatically requesting kernel tests, add a tag named: bot-stop-nagging.

 Thank you for your help, we really do appreciate it.

I'm seeing very similar issues, with a similar stack in dmesg in 3.2.0-24-generic-pae in X (from linux-image-3.2.0.24-generic-pae/3.2.0-24.37)

Manifests itself as X going black for a second or two before redispaying the desktop, unreliable input after it comes back and general flakiness. Occasional hangs.

I'm installing 3.4.0-030400rc4 from http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.4-rc4-precise/ to see if it's any different.

The 3.4.0 rc didn't help. i915.i915_enable_rc6=0 caused an oops with 3.4.x (in NFS, so possibly unrelated) whereas it seems to have helped with graphics stability in 3.2.

@Andrew,

Do you have a way to reproduce this bug?

@jsalisbury, yes - boot with either the 3.2.x or 3.4.x kernels: log in using Unity 3D; start Chrome, Firefox and a Terminal; use for a few minutes.

Going back to a 3.0.x (3.0.0-19-generic-pae) kernel seems to be fine.

@Andrew

Would you be able to test some kernels and assist with a kernel bisect since you can reproduce this issue?

If so, it would be great if you could test some of the upstream kernels[0] to identify when the regression was introduced.

The first kernel of interest would be the first avaible v3.1 kernel, which is available at:
http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.1-rc2-oneiric/

Thanks in advance!

[0] http://kernel.ubuntu.com/~kernel-ppa/mainline/

@Joseph, sure - no problem. Interestingly, although the Oneiric kernel seems fine, apport-gpu-intel.py keeps popping up after booting. It doesn't seem to properly succeed, though, so maybe it's picking up previous indicators somehow (no idea how it works)

I have upgraded from ubuntu 11.10 to 12.04, I am also having same issue with the similar dmesg!
I have tried with
    sudo apt-get update
    sudo apt-get dist-upgrade
During up-gradation to kernel-3.2.0-24.37 gives an error which says

    Error! Bad return status for module build on kernel: 3.2.0-24-generic (x86_64)
    Consult /var/lib/dkms/blcr/0.8.2/build/make.log for more information.

And the last part of the make.log file shows
      configure: error: --with-linux argument '3.2.0-24-generic' is neither a kernel version string nor a full path
      make[1]: *** [/var/lib/dkms/blcr/0.8.2/build/config-stamp] Error 1
      make: *** [_module_/var/lib/dkms/blcr/0.8.2/build] Error 2
      make: Leaving directory `/usr/src/linux-headers-3.2.0-24-generic'

@Andrew

Were you able to test the first avaible v3.1 kernel, mentioned in comment #13?

@Rajib

Could you also test that kernel?

I've been doing a bisection and it _seems_ 3.1.x is fine, but 3.2.x onwards isn't. I've not yet tried all of them to narrow down further where the regression occurred.

Can you test the following kernels and report if the bug exists:
v3.1.10: http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.1.10-precise/

and

v3.2-rc2: http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.2-rc2-oneiric/

v3.2-rc1 is also available, but only for amd64.

@Robbie

Are you still seeing this issue?

v3.1.10-precise: fine

v3.2-rc2-oneiric: also fine, but there's no PAE version of that available so it's not an exactly equivalent test.

However, uname output from versions which *are* affected:
  * 3.2.0-23-generic-pae #36-Ubuntu
  * 3.2.0-24-generic-pae #38-Ubuntu

@Andrew,

Can you test some additional v3.2 release candidates to try and see which one introduce this regression?

Possibly try v3.2-rc3 next: http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.2-rc3-oneiric/
then v3.2-rc4: http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.2-rc4-oneiric/
...

3.2-rc3 doesn't have any precompiled packages, just the headers: http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.2-rc3-oneiric/

Installing 3.2-rc4-oneiric now. Previous testing showed 3.2-rc4-precise to be OK (IIRC)

Using 3.2-rc4-oneiric gave a different problem. Unlike the immediate and frequent screen blanks/X reinitialisations, X hung. dmesg showed a GPU hang. Attached the stack in gpu-hang.txt.

@Andrew,

Could you also test 3.2-rc5 to see if it has this bug?

Will do (3.2-rc5-precise, right? Is there any difference between the -oneiric & -precise versions?)

It's also worth noting I'm using the -pae versions where available.

The only diff is config file options. It's probably best to use the -precise version.

With 3.2-rc5-precise I get a full machine lockup (doesn't respond to pings, mouse doesn't move, can't switch to console) after some period of usage (first time was after ~1h, second time after ~5m)

Can you also test 3.2-rc6 to see if the bug is there, since rc5 locks up.

Within a few minutes of using 3.2-rc6, the screen flickered black and back exactly as it does when the problem occurs; it was then immediately followed by the full machine hang I saw with 3.2-rc5.

So, IMHO the problems possibly started with the GPU hang in 3.2-rc4-oneiric, turned into a full machine hang in 3.2-rc5-precise and morphed into the "X/GPU gets reinitialised (toavoid a hang?), but then the whole machine hangs anyway" issue of 3.2-rc6.

Due to all these other bugs, it might be a good test to see if this bug is already fixed in the mainline kernel[0]. If it is, it may be easier to perform a reverse bisect to find the fix. If you have a chance, can you test the latest mainline kernel:

[0] http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.5-rc5-quantal/

3.5 seems as bad as 3.4 - lots of weird graphical effects and glitches. Attached the full dmesg output: lots of problems listed.

Any ideas on further things I could try?

The v3.6-rc3 kernel is now available. Can you test that kernel to see if this bug is fixed in mainline?

The v3.6-rc3 kernel can be downloaded from:
 http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.6-rc3-quantal/

You will need to install both the linux-image and linux-image-extra .deb packages.

Still failing in 3.6-rc3. Every 5 seconds in X there's a freeze (except for the mouse pointer), and then re-rendering tainted areas of the screen is erratic.

dmesg says:

   56.148632] ------------[ cut here ]------------
[ 56.148660] WARNING: at /home/apw/COD/linux/drivers/gpu/drm/i915/intel_pm.c:4009 gen6_gt_check_fifodbg+0x3e/0x60 [i915]()
[ 56.148662] Hardware name: To Be Filled By O.E.M.
[ 56.148663] MMIO read or write has been dropped 2
[ 56.148664] Modules linked in: hidp ipt_MASQUERADE iptable_nat nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_state nf_conntrack iptable_filter ip_tables x_tables ipx p8022 psnap llc p8023 dm_crypt snd_hda_codec_hdmi snd_hda_codec_realtek bnep rfcomm parport_pc ppdev binfmt_misc nfsd nfs_acl nfsv4 auth_rpcgss nfs fscache lockd sunrpc joydev hid_generic btusb bluetooth coretemp kvm_intel kvm snd_hda_intel snd_hda_codec snd_hwdep psmouse microcode snd_pcm snd_seq_midi snd_rawmidi snd_seq_midi_event snd_seq snd_timer snd_seq_device serio_raw snd lpc_ich soundcore snd_page_alloc mac_hid mei lp parport uvesafb hid_apple aesni_intel ablk_helper cryptd aes_i586 usbhid hid i915 r8169 drm_kms_helper drm i2c_algo_bit video
[ 56.148692] Pid: 1392, comm: Xorg Tainted: G W 3.6.0-030600rc3-generic #201208221735
[ 56.148693] Call Trace:
[ 56.148699] [<c1049512>] warn_slowpath_common+0x72/0xa0
[ 56.148710] [<f859ba4e>] ? gen6_gt_check_fifodbg+0x3e/0x60 [i915]
[ 56.148719] [<f859ba4e>] ? gen6_gt_check_fifodbg+0x3e/0x60 [i915]
[ 56.148722] [<c10495e3>] warn_slowpath_fmt+0x33/0x40
[ 56.148731] [<f859ba4e>] gen6_gt_check_fifodbg+0x3e/0x60 [i915]
[ 56.148741] [<f859bac1>] __gen6_gt_force_wake_put+0x21/0x30 [i915]
[ 56.148747] [<f855dd40>] i915_read32+0x110/0x130 [i915]
[ 56.148757] [<f85a035c>] intel_ring_get_active_head+0x2c/0x30 [i915]
[ 56.148766] [<f85a0397>] gen6_ring_get_seqno+0x37/0x50 [i915]
[ 56.148773] [<f856eb6d>] i915_gem_retire_requests_ring+0x2d/0x250 [i915]
[ 56.148777] [<c15e0e40>] ? mutex_lock_interruptible+0x10/0x40
[ 56.148784] [<f856f7f6>] i915_gem_object_flush_active+0x66/0x70 [i915]
[ 56.148792] [<f8571fe3>] i915_gem_busy_ioctl+0x53/0xa0 [i915]
[ 56.148795] [<c12d75c2>] ? _copy_from_user+0x42/0x60
[ 56.148802] [<f8488e88>] drm_ioctl+0x408/0x470 [drm]
[ 56.148812] [<f8571f90>] ? i915_gem_unpin_ioctl+0xf0/0xf0 [i915]
[ 56.148815] [<c12d2a97>] ? timerqueue_del+0x27/0x70
[ 56.148817] [<c106e390>] ? __remove_hrtimer+0x40/0xa0
[ 56.148819] [<c106eba2>] ? hrtimer_try_to_cancel+0x42/0xb0
[ 56.148821] [<c105039b>] ? do_setitimer+0x13b/0x230
[ 56.148823] [<c12d74cf>] ? __copy_from_user_ll+0x1f/0x40
[ 56.148829] [<f8488a80>] ? drm_copy_field+0x80/0x80 [drm]
[ 56.148832] [<c1162d4f>] do_vfs_ioctl+0x7f/0x2f0
[ 56.148834] [<c12d73df>] ? __copy_to_user_ll+0x1f/0x40
[ 56.148836] [<c116302f>] sys_ioctl+0x6f/0x80
[ 56.148838] [<c15e9c9f>] sysenter_do_call+0x12/0x28
[ 56.148840] ---[ end trace 7ee6e6596296fef6 ]---
[ 62.163513] [drm:i915_hangcheck_hung] *ERROR* Hangcheck timer elapsed... GPU hung
[ 62.163516] [drm] capturing error event; look for more information in /debug/dri/0/i915_error_state

This happened to me last night on Precise, 3.2.0-31-generic amd64. This isn't consistently reproducible, but I recall the same symptoms happening a few times over the last few months (I don't have logs from those hangs, sorry).

Oct 10 02:10:54 uk kernel: [875851.869738] WARNING: at /build/buildd/linux-3.2.0/drivers/gpu/drm/i915/i915_drv.c:417 __gen6_gt_wait_for_fifo+0x94/0xa0 [i915]()
Oct 10 02:10:54 uk kernel: [875851.869744] Hardware name: 4286CTO
Oct 10 02:10:54 uk kernel: [875851.869747] Modules linked in: usblp usb_storage ipt_REDIRECT xt_hl btrfs zlib_deflate libcrc32c ufs qnx4 hfsplus hfs minix ntfs msdos jfs xfs reiserfs nls_iso8859_1 nls_cp437 vfat fat mmc_block pci_stub vboxpci(O) vboxnetadp(O) vboxnetflt(O) vboxdrv(O) ip6table_filter ip6_tables ebtable_nat ebtables xt_state ipt_REJECT xt_CHECKSUM ipt_MASQUERADE iptable_nat nf_nat iptable_mangle nf_conntrack_ipv4 nf_conntrack nf_defrag_ipv4 xt_tcpudp iptable_filter ip_tables x_tables bridge stp nfsd nfs parport_pc bnep ppdev rfcomm binfmt_misc lockd fscache auth_rpcgss nfs_acl sunrpc ext2 snd_hda_codec_hdmi snd_hda_codec_conexant snd_hda_intel snd_hda_codec snd_hwdep btusb arc4 bluetooth snd_pcm i915 drm_kms_helper drm snd_seq_midi iwlwifi i2c_algo_bit snd_rawmidi snd_seq_midi_event tpm_tis snd_seq snd_timer joydev thinkpad_acpi lp uvcvideo nvram wmi snd_seq_device video videodev v4l2_compat_ioctl32 mac_hid snd mei(C) psmouse parport serio_raw soundcore mac80211 snd_page_alloc cfg80211 sdhci_p
Oct 10 02:10:54 uk kernel: ci sdhci e1000e
Oct 10 02:10:54 uk kernel: [875851.869885] Pid: 1449, comm: Xorg Tainted: G WC O 3.2.0-31-generic #50-Ubuntu
Oct 10 02:10:54 uk kernel: [875851.869890] Call Trace:
Oct 10 02:10:54 uk kernel: [875851.869901] [<ffffffff81066d7f>] warn_slowpath_common+0x7f/0xc0
Oct 10 02:10:54 uk kernel: [875851.869909] [<ffffffff81066dda>] warn_slowpath_null+0x1a/0x20
Oct 10 02:10:54 uk kernel: [875851.869930] [<ffffffffa02a6a04>] __gen6_gt_wait_for_fifo+0x94/0xa0 [i915]
Oct 10 02:10:54 uk kernel: [875851.869952] [<ffffffffa02a70a1>] i915_write32+0xe1/0xf0 [i915]
Oct 10 02:10:54 uk kernel: [875851.869984] [<ffffffffa02e42f2>] gen6_ring_get_irq+0xb2/0xd0 [i915]
Oct 10 02:10:54 uk kernel: [875851.870014] [<ffffffffa02e4328>] blt_ring_get_irq+0x18/0x20 [i915]
Oct 10 02:10:54 uk kernel: [875851.870039] [<ffffffffa02b8f13>] i915_wait_request+0x183/0x560 [i915]
Oct 10 02:10:54 uk kernel: [875851.870068] [<ffffffffa02e39a1>] ? ring_write_tail+0x21/0x30 [i915]
Oct 10 02:10:54 uk kernel: [875851.870095] [<ffffffffa02e6166>] ? gen6_render_ring_flush+0xd6/0xf0 [i915]
Oct 10 02:10:54 uk kernel: [875851.870121] [<ffffffffa02b9322>] i915_gem_object_wait_rendering+0x32/0x40 [i915]
Oct 10 02:10:54 uk kernel: [875851.870147] [<ffffffffa02be12d>] i915_gem_execbuffer_sync_rings+0xdd/0x160 [i915]
Oct 10 02:10:54 uk kernel: [875851.870173] [<ffffffffa02be31e>] i915_gem_execbuffer_move_to_gpu+0x16e/0x200 [i915]
Oct 10 02:10:54 uk kernel: [875851.870200] [<ffffffffa02bea18>] i915_gem_do_execbuffer.isra.7+0x668/0x9a0 [i915]
Oct 10 02:10:54 uk kernel: [875851.870210] [<ffffffff8108ec1b>] ? __hrtimer_start_range_ns+0x16b/0x3a0
Oct 10 02:10:54 uk kernel: [875851.870235] [<f...

I've now switched to a new NVidia card. Unfortunately that's not without its share of bugs either (e.g. bug #1057000 and bug #993604).

[Expired for linux (Ubuntu) because there has been no activity for 60 days.]