<SOLVED> Desktop crash on Intel HD 3000 (Sandy Bridge)

[xenopeek]

A few times per week my desktop shortly hangs (not responsive to mouse or keyboard; mouse pointer also doesn't move), then the screen goes black, and then the desktop returns but now all graphical operations are slow. Maximizing or minimizing a window for example, you can see the window or desktop be drawn painfully slow If I continue to use it, lots of artifacts will appear like garbling fonts.

Sometimes this happens after having used it for a long time, usually shortly after logging in. Logging out or resetting X (ctrl+alt+backspace) doesn't help, I have to reboot to clear the problem. Looking at my syslog I find the following just after the problem occurs (logging is the same each time, this is from the first time I captured it):

Code: Select all

Sep  4 08:29:49 maya kernel: [  188.331322] [drm:i915_hangcheck_elapsed] *ERROR* Hangcheck timer elapsed... GPU hung
Sep  4 08:29:49 maya kernel: [  188.333634] [drm:i915_reset] *ERROR* GPU hanging too fast, declaring wedged!
Sep  4 08:29:49 maya kernel: [  188.333638] [drm:i915_reset] *ERROR* Failed to reset chip.

Just before that final piece of logging, it is reporting tens of times the following sequence:

Code: Select all

Sep  4 08:29:41 maya kernel: [  180.541984] ------------[ cut here ]------------
Sep  4 08:29:41 maya kernel: [  180.541990] WARNING: at /build/buildd/linux-3.2.0/drivers/gpu/drm/i915/i915_drv.c:417 __gen6_gt_wait_for_fifo+0x94/0xa0 [i915]()
Sep  4 08:29:41 maya kernel: [  180.541991] Hardware name: To Be Filled By O.E.M.
Sep  4 08:29:41 maya kernel: [  180.541992] Modules linked in: pci_stub vboxpci(O) vboxnetadp(O) vboxnetflt(O) vboxdrv(O) parport_pc ppdev binfmt_misc ip6t_LOG xt_hl ip6t_rt snd_hda_codec_realtek nf_conntrack_ipv6 nf_defrag_ipv6 ipt_REJECT ipt_LOG xt_limit xt_tcpudp xt_addrtype xt_state ip6table_filter snd_hda_intel snd_hda_codec snd_hwdep ip6_tables nf_conntrack_netbios_ns nf_conntrack_broadcast nf_nat_ftp nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_conntrack_ftp nf_conntrack iptable_filter psmouse snd_pcm ip_tables x_tables snd_seq_midi snd_rawmidi serio_raw snd_seq_midi_event snd_seq snd_timer snd_seq_device snd mac_hid soundcore snd_page_alloc mei(C) w83627ehf hwmon_vid coretemp lp parport i915 drm_kms_helper drm i2c_algo_bit usbhid hid r8169 video
Sep  4 08:29:41 maya kernel: [  180.542013] Pid: 1371, comm: Xorg Tainted: G        WC O 3.2.0-29-generic #46-Ubuntu
Sep  4 08:29:41 maya kernel: [  180.542014] Call Trace:
Sep  4 08:29:41 maya kernel: [  180.542017]  [<ffffffff81066cff>] warn_slowpath_common+0x7f/0xc0
Sep  4 08:29:41 maya kernel: [  180.542018]  [<ffffffff81066d5a>] warn_slowpath_null+0x1a/0x20
Sep  4 08:29:41 maya kernel: [  180.542023]  [<ffffffffa00b4a04>] __gen6_gt_wait_for_fifo+0x94/0xa0 [i915]
Sep  4 08:29:41 maya kernel: [  180.542028]  [<ffffffffa00b50a1>] i915_write32+0xe1/0xf0 [i915]
Sep  4 08:29:41 maya kernel: [  180.542036]  [<ffffffffa00f1b52>] gen6_ring_put_irq+0xa2/0xc0 [i915]
Sep  4 08:29:41 maya kernel: [  180.542042]  [<ffffffffa00f1b88>] blt_ring_put_irq+0x18/0x20 [i915]
Sep  4 08:29:41 maya kernel: [  180.542048]  [<ffffffffa00c445c>] i915_gem_ring_throttle.isra.15+0x27c/0x290 [i915]
Sep  4 08:29:41 maya kernel: [  180.542050]  [<ffffffff8108aa50>] ? add_wait_queue+0x60/0x60
Sep  4 08:29:41 maya kernel: [  180.542055]  [<ffffffffa00c9c42>] i915_gem_throttle_ioctl+0x12/0x20 [i915]
Sep  4 08:29:41 maya kernel: [  180.542060]  [<ffffffffa00585d4>] drm_ioctl+0x444/0x510 [drm]
Sep  4 08:29:41 maya kernel: [  180.542065]  [<ffffffffa00c9c30>] ? i915_gem_busy_ioctl+0x160/0x160 [i915]
Sep  4 08:29:41 maya kernel: [  180.542067]  [<ffffffff8101db0b>] ? check_for_xstate+0x3b/0xc0
Sep  4 08:29:41 maya kernel: [  180.542069]  [<ffffffff8101dbe4>] ? restore_user_xstate+0x54/0xa0
Sep  4 08:29:41 maya kernel: [  180.542071]  [<ffffffff81189c5a>] do_vfs_ioctl+0x8a/0x340
Sep  4 08:29:41 maya kernel: [  180.542073]  [<ffffffff81189fa1>] sys_ioctl+0x91/0xa0
Sep  4 08:29:41 maya kernel: [  180.542074]  [<ffffffff81661ec2>] system_call_fastpath+0x16/0x1b
Sep  4 08:29:41 maya kernel: [  180.542075] ---[ end trace 9f8d86f58ee9860e ]---

My inxi -SGx output:

Code: Select all

System:    Host: maya Kernel: 3.2.0-30-generic x86_64 (64 bit, gcc: 4.6.3) Desktop: N/A Distro: Linux Mint 13 Maya
Graphics:  Card: Intel 2nd Generation Core Processor Family Integrated Graphics Controller bus-ID: 00:02.0
           X.Org: 1.11.3 drivers: intel (unloaded: vesa,fbdev) Resolution: 1280x1024@60.0hz
           GLX Renderer: Mesa DRI Intel Sandybridge Desktop GLX Version: 3.0 Mesa 8.0.2 Direct Rendering: Yes

I'm using Linux Mint 13 MATE, without Compiz. I had the same issues on Linux Mint 13 Xfce. My CPU is the Intel Core i5-2500K, which is running stable at 29°C (this is not an overheating problem ).

What I've done so far (waiting for the problem to reoccur between each step), which may have lessened the frequency of the problem somewhat but didn't fix it:

• dist-upgrade, to upgrade to the latest kernel and xorg drivers as available in the default repositories;
• added ubuntu-x-swat PPA for newer stable Intel drivers (https://launchpad.net/~ubuntu-x-swat/+archive/x-updates);
• added i915.semaphores=1 to my boot parameters (it was set to -1 before that).

I've been looking at recommendations, but not sure what my next step should be. Any help would be appreciated.

[xenopeek]

I've upgraded to kernel 3.4, so let's see how that goes. I followed the instructions from here: https://wiki.ubuntu.com/Kernel/MainlineBuilds

I first tried the 3.6-rc7 but that didn't work as it failed to detect monitor correctly (dumped into 640x480 mode) and neither keyboard nor mouse was working...

[xenopeek]

No success with 3.4 kernel; it crashed within 15 minutes, and ended with a hard lock-up. Not even magic sysrq worked I'm trying the 3.4 kernel again without the i915.semaphores=1 tweak.

The logging from the crash (abruptly terminated due to it being a hard lock-up):

Code: Select all

Sep 27 17:58:54 maya kernel: [  795.016318] ------------[ cut here ]------------
Sep 27 17:58:54 maya kernel: [  795.016343] WARNING: at /home/apw/COD/linux/drivers/gpu/drm/i915/i915_drv.c:440 __gen6_gt_wait_for_fifo+0x9e/0xb0 [i915]()
Sep 27 17:58:54 maya kernel: [  795.016346] Hardware name: To Be Filled By O.E.M.
Sep 27 17:58:54 maya kernel: [  795.016348] Modules linked in: pci_stub vboxpci(O) vboxnetadp(O) vboxnetflt(O) vboxdrv(O) snd_hda_codec_realtek xt_hl ip6t_rt parport_pc nf_conntrack_ipv6 nf_defrag_ipv6 ppdev ghash_clmulni_intel aesni_intel cryptd xt_limit aes_x86_64 snd_hda_intel xt_tcpudp xt_addrtype snd_hda_codec snd_hwdep snd_pcm xt_state ip6table_filter snd_seq_midi snd_rawmidi ip6_tables nf_conntrack_netbios_ns nf_conntrack_broadcast snd_seq_midi_event Sep 27 18:00:10 maya kernel: imklog 5.8.6, log source = /proc/kmsg started.

[mank_in]

hello Vincent , maybe your problem is related with power saving features on the intel IGP
I reads in : http://arstechnica.com/civis/viewtopic. ... &t=1174039.
It's like the solution is add line in GRUB menu :

Code: Select all

GRUB_CMDLINE_LINUX_DEFAULT="i915.i915_enable_rc6=0 quiet splash"

I didn't have sandy bridge CPU , but i hope this works

[xenopeek]

Thanks, I'm going to try this next. With the 3.4 kernel (without the semaphores tweak) I've not had a crash yet, though I did have a couple of times where the screen froze for a few seconds. The following from the syslog at that time:

Code: Select all

Sep 28 23:02:04 maya kernel: [  968.286119] ------------[ cut here ]------------
Sep 28 23:02:04 maya kernel: [  968.286129] WARNING: at /home/apw/COD/linux/drivers/gpu/drm/i915/i915_drv.c:398 gen6_gt_check_fifodbg+0x41/0x60 [i915]()
Sep 28 23:02:04 maya kernel: [  968.286132] Hardware name: To Be Filled By O.E.M.
Sep 28 23:02:04 maya kernel: [  968.286134] MMIO read or write has been dropped 3
Sep 28 23:02:04 maya kernel: [  968.286136] Modules linked in: pci_stub vboxpci(O) vboxnetadp(O) vboxnetflt(O) vboxdrv(O) xt_hl ip6t_rt nf_conntrack_ipv6 nf_defrag_ipv6 bnep rfcomm xt_limit snd_hda_codec_realtek bluetooth xt_tcpudp parport_pc ppdev xt_addrtype xt_state ip6table_filter ghash_clmulni_intel ip6_tables nf_conntrack_netbios_ns nf_conntrack_broadcast nf_nat_ftp nf_nat nf_conntrack_ipv4 snd_hda_intel nf_defrag_ipv4 snd_hda_codec nf_conntrack_ftp aesni_intel nf_conntrack cryptd snd_hwdep aes_x86_64 snd_pcm iptable_filter ip_tables x_tables microcode snd_seq_midi snd_rawmidi snd_seq_midi_event psmouse snd_seq serio_raw snd_timer snd_seq_device snd binfmt_misc soundcore snd_page_alloc mei(C) mac_hid w83627ehf hwmon_vid coretemp lp parport i915 usbhid hid r8169 drm_kms_helper drm i2c_algo_bit video
Sep 28 23:02:04 maya kernel: [  968.286188] Pid: 58, comm: kworker/u:4 Tainted: G        WC O 3.4.0-030400-generic #201205210521
Sep 28 23:02:04 maya kernel: [  968.286191] Call Trace:
Sep 28 23:02:04 maya kernel: [  968.286197]  [<ffffffff81052cff>] warn_slowpath_common+0x7f/0xc0
Sep 28 23:02:04 maya kernel: [  968.286201]  [<ffffffff81052df6>] warn_slowpath_fmt+0x46/0x50
Sep 28 23:02:04 maya kernel: [  968.286207]  [<ffffffff8132899a>] ? delay_tsc+0x4a/0x80
Sep 28 23:02:04 maya kernel: [  968.286218]  [<ffffffffa0078491>] gen6_gt_check_fifodbg+0x41/0x60 [i915]
Sep 28 23:02:04 maya kernel: [  968.286229]  [<ffffffffa00784ea>] __gen6_gt_force_wake_put+0x1a/0x20 [i915]
Sep 28 23:02:04 maya kernel: [  968.286241]  [<ffffffffa0078e5b>] i915_read32+0xbb/0x120 [i915]
Sep 28 23:02:04 maya kernel: [  968.286257]  [<ffffffffa00b7df0>] intel_ring_get_active_head+0x30/0x40 [i915]
Sep 28 23:02:04 maya kernel: [  968.286272]  [<ffffffffa00b7e45>] gen6_ring_get_seqno+0x45/0x50 [i915]
Sep 28 23:02:04 maya kernel: [  968.286285]  [<ffffffffa008bc15>] i915_gem_retire_requests_ring+0x35/0x2b0 [i915]
Sep 28 23:02:04 maya kernel: [  968.286298]  [<ffffffffa008e00e>] i915_gem_retire_requests+0x8e/0xa0 [i915]
Sep 28 23:02:04 maya kernel: [  968.286311]  [<ffffffffa008f936>] i915_gem_retire_work_handler+0x66/0x1a0 [i915]
Sep 28 23:02:04 maya kernel: [  968.286324]  [<ffffffffa008f8d0>] ? i915_gem_fault+0x2c0/0x2c0 [i915]
Sep 28 23:02:04 maya kernel: [  968.286329]  [<ffffffff8106ed2b>] process_one_work+0x12b/0x470
Sep 28 23:02:04 maya kernel: [  968.286333]  [<ffffffff81071896>] worker_thread+0x176/0x420
Sep 28 23:02:04 maya kernel: [  968.286337]  [<ffffffff81071720>] ? manage_workers+0x120/0x120
Sep 28 23:02:04 maya kernel: [  968.286341]  [<ffffffff810763ee>] kthread+0x9e/0xb0
Sep 28 23:02:04 maya kernel: [  968.286344]  [<ffffffff81676164>] kernel_thread_helper+0x4/0x10
Sep 28 23:02:04 maya kernel: [  968.286349]  [<ffffffff81076350>] ? kthread_freezable_should_stop+0x70/0x70
Sep 28 23:02:04 maya kernel: [  968.286353]  [<ffffffff81676160>] ? gs_change+0x13/0x13
Sep 28 23:02:04 maya kernel: [  968.286355] ---[ end trace 846cfca0931e2a9f ]---

[xenopeek]

Thanks mank_in, this seems to be the solution. After more searching, I found also this bug report which in comment #4 shares the same solution: https://bugs.freedesktop.org/show_bug.cgi?id=50619. Now I happen to also have an ASRock H67M motherboard, just like other posters on that bug. Might be a coincidence, or not

[xenopeek]

Marking this topic solved No more crashes, or warnings giving in the syslog after adding "i915.i915_enable_rc6=0" to the boot parameters. I'm also on kernel 3.4 and using the ubuntu-x-swat PPA, so three changes in all.