You are not logged in.

#1 2014-11-17 14:41:38

belette
Guest

DRM Intel Errors

Hi !

From a fresh installation, I got multiple errors during booting process.
I activated mmio_debug=1 as suggested into journalctl logs entries.

This is the log (I got 25 times similar logs during log process)

Appreciate any idea smile

belette

WARNING: CPU: 3 PID: 226 at drivers/gpu/drm/i915/intel_uncore.c:528 hsw_unclaimed_reg_debug+0x6d/0x80 [i915]()
Nov 17 17:21:35 archfit kernel: [drm:intel_uncore_check_errors] *ERROR* Unclaimed register before interrupt
Nov 17 17:21:35 archfit kernel: Unclaimed register detected after reading register 0x650c0
Nov 17 17:21:35 archfit kernel: Modules linked in: snd_hda_codec_realtek snd_hda_codec_generic nls_iso8859_1 nls_cp437 vfat fat iTCO_wdt iTCO_vendor_
Nov 17 17:21:35 archfit kernel:  dm_crypt dm_mod sd_mod crc_t10dif crct10dif_pclmul crct10dif_common crc32_pclmul crc32c_intel ghash_clmulni_intel at
Nov 17 17:21:35 archfit kernel: CPU: 3 PID: 226 Comm: systemd-udevd Not tainted 3.17.2-1-ARCH #1
Nov 17 17:21:35 archfit kernel: Hardware name: CompuLab Ltd. Intense-PC2 (IPC2)/Intense-PC2 (IPC2), BIOS IPC2_3.330.2 X64 08/29/2014
Nov 17 17:21:35 archfit kernel:  0000000000000000 000000007c6450cc ffff880406237190 ffffffff815367d0
Nov 17 17:21:35 archfit kernel:  ffff8804062371d8 ffff8804062371c8 ffffffff8107054d ffff8804068e0000
Nov 17 17:21:35 archfit kernel:  00000000000650c0 0000000000000000 00000000000650c0 ffff8804068e0000
Nov 17 17:21:35 archfit kernel: Call Trace:
Nov 17 17:21:35 archfit kernel:  [<ffffffff815367d0>] dump_stack+0x4d/0x6f
Nov 17 17:21:35 archfit kernel:  [<ffffffff8107054d>] warn_slowpath_common+0x7d/0xa0
Nov 17 17:21:35 archfit kernel:  [<ffffffff810705cc>] warn_slowpath_fmt+0x5c/0x80
Nov 17 17:21:35 archfit kernel:  [<ffffffffa04d516d>] hsw_unclaimed_reg_debug+0x6d/0x80 [i915]
Nov 17 17:21:35 archfit kernel:  [<ffffffffa04d81f3>] gen6_read32+0x93/0x150 [i915]
Nov 17 17:21:35 archfit kernel:  [<ffffffffa04e7868>] haswell_write_eld+0x78/0x450 [i915]
Nov 17 17:21:35 archfit kernel:  [<ffffffffa04ed467>] intel_write_eld+0x67/0xa0 [i915]
Nov 17 17:21:35 archfit kernel:  [<ffffffffa0506271>] intel_ddi_pre_enable+0x51/0x1f0 [i915]
Nov 17 17:21:35 archfit kernel:  [<ffffffffa04f0fa4>] haswell_crtc_enable+0x3d4/0xa70 [i915]
Nov 17 17:21:35 archfit kernel:  [<ffffffffa04bb477>] ? i915_gem_object_ggtt_unpin+0x47/0x90 [i915]
Nov 17 17:21:35 archfit kernel:  [<ffffffffa04eb5da>] __intel_set_mode+0x87a/0x16d0 [i915]
Nov 17 17:21:35 archfit kernel:  [<ffffffffa04f3da6>] intel_set_mode+0x16/0x30 [i915]
Nov 17 17:21:35 archfit kernel:  [<ffffffffa04f4f54>] intel_crtc_set_config+0xaf4/0xef0 [i915]
Nov 17 17:21:35 archfit kernel:  [<ffffffff8119f7e1>] ? alloc_pages_current+0xb1/0x160
Nov 17 17:21:35 archfit kernel:  [<ffffffffa039b324>] drm_mode_set_config_internal+0x64/0xe0 [drm]
Nov 17 17:21:35 archfit kernel:  [<ffffffffa03d51a3>] restore_fbdev_mode+0xb3/0xe0 [drm_kms_helper]
Nov 17 17:21:35 archfit kernel:  [<ffffffffa03d51f3>] drm_fb_helper_restore_fbdev_mode_unlocked+0x23/0x40 [drm_kms_helper]
Nov 17 17:21:35 archfit kernel:  [<ffffffffa03d7418>] drm_fb_helper_set_par+0x28/0x70 [drm_kms_helper]
Nov 17 17:21:35 archfit kernel:  [<ffffffffa050008a>] intel_fbdev_set_par+0x1a/0x60 [i915]
Nov 17 17:21:35 archfit kernel:  [<ffffffff8130b3b8>] fbcon_init+0x578/0x600
Nov 17 17:21:35 archfit kernel:  [<ffffffff81383a08>] visual_init+0xb8/0x120
Nov 17 17:21:35 archfit kernel:  [<ffffffff81385f26>] do_bind_con_driver+0x176/0x3f0
Nov 17 17:21:35 archfit kernel:  [<ffffffff81386744>] do_take_over_console+0xb4/0x1b0
Nov 17 17:21:35 archfit kernel:  [<ffffffff8130b4a3>] do_fbcon_takeover+0x63/0xd0
Nov 17 17:21:35 archfit kernel:  [<ffffffff8130fb1d>] fbcon_event_notify+0x6cd/0x7d0
Nov 17 17:21:35 archfit kernel:  [<ffffffff8108f02d>] notifier_call_chain+0x4d/0x70
Nov 17 17:21:35 archfit kernel:  [<ffffffff8108f3f7>] __blocking_notifier_call_chain+0x47/0x60
Nov 17 17:21:35 archfit kernel:  [<ffffffff8108f426>] blocking_notifier_call_chain+0x16/0x20
Nov 17 17:21:35 archfit kernel:  [<ffffffff81315e6b>] fb_notifier_call_chain+0x1b/0x20
Nov 17 17:21:35 archfit kernel:  [<ffffffff8131806e>] register_framebuffer+0x20e/0x360
Nov 17 17:21:35 archfit kernel:  [<ffffffffa03d7247>] drm_fb_helper_initial_config+0x377/0x520 [drm_kms_helper]
Nov 17 17:21:35 archfit kernel:  [<ffffffffa04d8d18>] ? hsw_write32+0xa8/0x160 [i915]
Nov 17 17:21:35 archfit kernel:  [<ffffffffa0500c6f>] intel_fbdev_initial_config+0x1f/0x30 [i915]
Nov 17 17:21:35 archfit kernel:  [<ffffffffa0526c81>] i915_driver_load+0x1051/0x10b0 [i915]
Nov 17 17:21:35 archfit kernel:  [<ffffffff8147059b>] ? netlink_broadcast_filtered+0x12b/0x380
Nov 17 17:21:35 archfit kernel:  [<ffffffff812ac300>] ? kobj_ns_drop+0x50/0x50
Nov 17 17:21:35 archfit kernel:  [<ffffffff812ac638>] ? kobject_uevent_env+0x178/0x540
Nov 17 17:21:35 archfit kernel:  [<ffffffff813a7a57>] ? get_device+0x17/0x30
Nov 17 17:21:35 archfit kernel:  [<ffffffff813ad7d5>] ? klist_class_dev_get+0x15/0x20
Nov 17 17:21:35 archfit kernel:  [<ffffffff8152abf8>] ? klist_node_init+0x38/0x60
Nov 17 17:21:35 archfit kernel:  [<ffffffff8152acb6>] ? klist_add_tail+0x46/0x50
Nov 17 17:21:35 archfit kernel:  [<ffffffff813a933b>] ? device_add+0x1db/0x6d0
Nov 17 17:21:35 archfit kernel:  [<ffffffffa0394fcd>] drm_dev_register+0xad/0x100 [drm]
Nov 17 17:21:35 archfit kernel:  [<ffffffffa03978d8>] drm_get_pci_dev+0xd8/0x200 [drm]
Nov 17 17:21:35 archfit kernel:  [<ffffffffa049145b>] i915_pci_probe+0x3b/0x60 [i915]
Nov 17 17:21:35 archfit kernel:  [<ffffffff812e6835>] local_pci_probe+0x45/0xa0
Nov 17 17:21:35 archfit kernel:  [<ffffffff812e6785>] ? pci_match_device+0xe5/0x110
Nov 17 17:21:35 archfit kernel:  [<ffffffff812e6991>] pci_device_probe+0x101/0x150
Nov 17 17:21:35 archfit kernel:  [<ffffffff813ac7f3>] driver_probe_device+0xa3/0x410
Nov 17 17:21:35 archfit kernel:  [<ffffffff813acc2b>] __driver_attach+0x8b/0x90
Nov 17 17:21:35 archfit kernel:  [<ffffffff813acba0>] ? __device_attach+0x40/0x40
Nov 17 17:21:35 archfit kernel:  [<ffffffff813aa613>] bus_for_each_dev+0x73/0xc0
Nov 17 17:21:35 archfit kernel:  [<ffffffff813ac27e>] driver_attach+0x1e/0x20
Nov 17 17:21:35 archfit kernel:  [<ffffffff813abe60>] bus_add_driver+0x180/0x250
Nov 17 17:21:35 archfit kernel:  [<ffffffff813ad4c4>] driver_register+0x64/0xf0
Nov 17 17:21:35 archfit kernel:  [<ffffffff812e5f9b>] __pci_register_driver+0x4b/0x50
Nov 17 17:21:35 archfit kernel:  [<ffffffffa0397b0a>] drm_pci_init+0x10a/0x140 [drm]
Nov 17 17:21:35 archfit kernel:  [<ffffffffa0570000>] ? 0xffffffffa0570000
Nov 17 17:21:35 archfit kernel:  [<ffffffffa0570089>] i915_init+0x89/0x90 [i915]
Nov 17 17:21:35 archfit kernel:  [<ffffffff81002148>] do_one_initcall+0xd8/0x210
Nov 17 17:21:35 archfit kernel:  [<ffffffff811ac176>] ? kfree+0x196/0x1a0
Nov 17 17:21:35 archfit kernel:  [<ffffffff811905a2>] ? __vunmap+0xa2/0x100
Nov 17 17:21:35 archfit kernel:  [<ffffffff810f222c>] load_module+0x1dcc/0x26a0
Nov 17 17:21:35 archfit kernel:  [<ffffffff810ee750>] ? store_uevent+0x70/0x70
Nov 17 17:21:35 archfit kernel:  [<ffffffff810f2bcd>] SyS_init_module+0xcd/0x120
Nov 17 17:21:35 archfit kernel:  [<ffffffff8153c7a9>] system_call_fastpath+0x16/0x1b
Nov 17 17:21:35 archfit kernel: ---[ end trace 445ee4af8ac4aba4 ]---

Last edited by belette (2014-11-18 18:05:28)

#2 2014-11-17 14:50:03

Trilby
Inspector Parrot
Registered: 2011-11-29
Posts: 30,480
Website

Re: DRM Intel Errors

Welcome to the forums.  Could you please edit your post to use code tags for that log output.

While these are not the same errors I've seen, it seems intel graphic acceleration has been very buggy with recent versions (and/or recent kernels).  In most cases selecting a different acceleration method (e.g. UXA) avoids the issue.  But is this only when X is starting, or do you get the same errors if you boot to a tty?

EDIT: also, is this a crash and non-responsive system, or are you just seeing errors in the logs in an otherwise functional system?


"UNIX is simple and coherent" - Dennis Ritchie; "GNU's Not Unix" - Richard Stallman

Offline

#3 2014-11-18 18:19:27

belette
Guest

Re: DRM Intel Errors

Thanks Trilby for your response (sorry for the code tag-changed now)

This is before X is started just after the initial ramdisk and for sure related to i915 drivers, so I don't know if changing the Acceleration Method will change the problem as booting in tty has the problem as well sad

No crash or non-responsive system but annoying logs and it seems very poor graphical performance (I still need to test it but without any tool and only with my eyes I can confirm..)

#4 2014-11-18 19:31:12

89c51
Member
Registered: 2012-06-05
Posts: 741

Re: DRM Intel Errors

As Trilby said intel sucks balls in the last few kernels. I get a similar error in the lts kernel. 3.17.x boots in a blank screen for one of my systems. Hope 3.18 fixes these.

Offline

#5 2014-11-18 20:02:17

belette
Guest

Re: DRM Intel Errors

Thanks ! Have you tweak the driver / Acceleration method to improve perf?

#6 2014-11-18 20:19:23

Trilby
Inspector Parrot
Registered: 2011-11-29
Posts: 30,480
Website

Re: DRM Intel Errors

I suspect passing the kernel parameter "nomodeset" would avoid the boot-time errors (worth checking at least for diagnostic purposes).  I don't know if this would impact later loading when X starts.

I also had all sorts of intel graphics related issues several weeks ago.  Switching from the default 'sna' to 'uxa' acceleration avoided all the errors, but did lead to reduced performance.  This reduction, however, was only noticable in a couple very graphics-intensive programs I use (bioacoustic analysis with an 'interactive' multi-layered spectrogram).  Luckily for me, 'sna' has worked fine since 3.17.  But following some of the history of these issues it seems to affect each bit of hardware differently so it is really hard to compare notes other than suggesting alternative acceleration methods as a potential fix/work-around.


"UNIX is simple and coherent" - Dennis Ritchie; "GNU's Not Unix" - Richard Stallman

Offline

#7 2014-11-18 22:53:54

89c51
Member
Registered: 2012-06-05
Posts: 741

Re: DRM Intel Errors

belette wrote:

Thanks ! Have you tweak the driver / Acceleration method to improve perf?

The intel netbook i use is an x less setup (testing wayland basically).

Didn't touch a single thing.

Nomodset boots 3.17 for me but it's unusable for my usecase since KMS doesn't work (hence no WL).


@Trilby
off topic but do you use custom software for spectrograms and acoustics?? I'm thinking of testing the acoustics of a room and want to know what is available for linux.

Offline

#8 2014-11-18 23:18:41

Trilby
Inspector Parrot
Registered: 2011-11-29
Posts: 30,480
Website

Re: DRM Intel Errors

The tool I was referring to was a very task-specific creation of mine to be featured in two upcoming publications (any day now ... waiting on a couple coauthors): https://github.com/BehaviorEnterprises/Fex

Most of the tools I'm familiar with are for working with animal communication signals.  But sound is sound.  Audacity is very user-friendly without sacrificing too much in power.  The Seewave package for R is very powerful, but not so user-friendly (and could work without X).  If you'd want any more input feel free to send me an email.


"UNIX is simple and coherent" - Dennis Ritchie; "GNU's Not Unix" - Richard Stallman

Offline

#9 2014-11-18 23:31:38

89c51
Member
Registered: 2012-06-05
Posts: 741

Re: DRM Intel Errors

Thanks for the answer. I was only aware of Spek.

Offline

#10 2014-11-19 15:43:06

belette
Guest

Re: DRM Intel Errors

Thanks Trilby
I just tried to nomodeset and now only one screen is up and running (I have 3 screens connected, 2 in HDMI 1 in DP).
I got this error during the boot process...

snd_hda_intel 0000:00:03:.0:Cannot turn on display power on i915

Board footer

Powered by FluxBB