Yeah this is a small problem with our dmesg filtering, since we match for (drm|intel|i915), and snd-hda-intel also has that :( Currently affected machines are byt-nuc (a baytrail). I'll attach full dmesg shortly. [ 261.127300] snd_hda_intel 0000:00:1b.0: azx_get_response timeout, switching to polling mode: last cmd=0x201f0500 [ 262.131814] snd_hda_intel 0000:00:1b.0: No response from codec, disabling MSI: last cmd=0x201f0500 [ 263.143330] snd_hda_intel 0000:00:1b.0: azx_get_response timeout, switching to single_cmd mode: last cmd=0x201f0500 [ 263.146980] snd_hda_codec_hdmi hdaudioC0D2: HDMI: pin nid 0 not registered [ 263.147730] ------------[ cut here ]------------ [ 263.147746] WARNING: CPU: 0 PID: 315 at sound/hda/hdac_i915.c:91 snd_hdac_display_power+0xf1/0x110 [snd_hda_core]() [ 263.147750] Modules linked in: snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_codec_generic i915 intel_powerclamp snd_hda_intel(-) coretemp snd_hda_codec crct10dif_pclmul crc32_pclmul snd_hwdep snd_hda_core lpc_ich snd_pcm i2c_hid i2c_designware_platform i2c_designware_core r8169 mii sdhci_acpi sdhci mmc_core [ 263.147790] CPU: 0 PID: 315 Comm: kworker/0:3 Tainted: G U W 4.4.0-rc4-gfxbench+ #1 [ 263.147794] Hardware name: \xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff \xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff\xffffffff/DN2820FYK, BIOS FYBYT10H.86A.0038.2014.0717.1455 07/17/2014 [ 263.147803] Workqueue: pm pm_runtime_work [ 263.147807] ffffffffa00c7391 ffff880074b7bbe8 ffffffff813dfe3c 0000000000000000 [ 263.147815] ffff880074b7bc20 ffffffff810746a1 0000000000000000 ffff8800747e67b0 [ 263.147822] ffff880072bbc188 0000000000000008 ffff880036a22290 ffff880074b7bc30 [ 263.147829] Call Trace: [ 263.147837] [<ffffffff813dfe3c>] dump_stack+0x4e/0x82 [ 263.147843] [<ffffffff810746a1>] warn_slowpath_common+0x81/0xc0 [ 263.147847] [<ffffffff81074795>] warn_slowpath_null+0x15/0x20 [ 263.147855] [<ffffffffa00c6671>] snd_hdac_display_power+0xf1/0x110 [snd_hda_core] [ 263.147862] [<ffffffffa012939d>] azx_intel_link_power+0xd/0x10 [snd_hda_intel] [ 263.147873] [<ffffffffa00f930a>] azx_link_power+0x1a/0x30 [snd_hda_codec] [ 263.147880] [<ffffffffa00c2179>] snd_hdac_link_power+0x29/0x40 [snd_hda_core] [ 263.147889] [<ffffffffa00f42a6>] hda_codec_runtime_suspend+0x76/0xa0 [snd_hda_codec] [ 263.147897] [<ffffffffa00f4230>] ? hda_codec_runtime_resume+0x50/0x50 [snd_hda_codec] [ 263.147902] [<ffffffff8152442d>] __rpm_callback+0x2d/0x70 [ 263.147909] [<ffffffffa00f4230>] ? hda_codec_runtime_resume+0x50/0x50 [snd_hda_codec] [ 263.147914] [<ffffffff8152448f>] rpm_callback+0x1f/0x80 [ 263.147922] [<ffffffffa00f4230>] ? hda_codec_runtime_resume+0x50/0x50 [snd_hda_codec] [ 263.147926] [<ffffffff81524978>] rpm_suspend+0x148/0x780 [ 263.147931] [<ffffffff81526526>] pm_runtime_work+0x76/0xc0 [ 263.147936] [<ffffffff8108f3d5>] process_one_work+0x1e5/0x620 [ 263.147940] [<ffffffff8108f339>] ? process_one_work+0x149/0x620 [ 263.147945] [<ffffffff8108f859>] worker_thread+0x49/0x450 [ 263.147949] [<ffffffff8108f810>] ? process_one_work+0x620/0x620 [ 263.147953] [<ffffffff8108f810>] ? process_one_work+0x620/0x620 [ 263.147959] [<ffffffff81095e9a>] kthread+0xea/0x100 [ 263.147964] [<ffffffff8179a6e7>] ? _raw_spin_unlock_irq+0x27/0x50 [ 263.147970] [<ffffffff81095db0>] ? kthread_create_on_node+0x1f0/0x1f0 [ 263.147974] [<ffffffff8179b37f>] ret_from_fork+0x3f/0x70 [ 263.147979] [<ffffffff81095db0>] ? kthread_create_on_node+0x1f0/0x1f0 [ 263.147982] ---[ end trace dfabf752a38ef4ee ]---
Created attachment 120518 [details] full dmesg
Takashi, this started blowing up in our CI very recently, so might be a new patch in your queue (we pull your trees into -nightly). Tracking against DRM/Intel since it affects our CI, so please don't reassign when you want to move this, but please just link. Thanks, Daniel
Could you point the top commit ID of the git branch showing the problem? I'd like to know whether all recent fixes in HD-audio are already included. The message indicates that the access to the codec at suspend failed. GPU domain got powered off before the audio part wrongly?
Bug scrub: Focus, Can you have a look to this and check Mendong
I checked the long-term stats of our CI system for the affected byt-nuc machine and this backtrace seems to have disappeared a while ago. We update sound/for-next+for-linus every time we rebuild -nightly, so as long as its in there we'll pull it in automatically.
FWIW, I have just started seeing this same error on a NUC DN2820 (which is bay trail) after upgrading it from Debian Jessie to Ubuntu Xenial. I have two of these NUCs, and only one showed the problem. It doesn't happen on *every* boot but is fairly common. Unfortunately it is not writing the dmesg to disk, but I do see hda_codec_runtime_suspend and rpm_callback in the backtrace. Both machines have the latest BIOS (0053) On a hunch, I went into the BIOS settings and reset them to defaults. After three further shutdown / restart cycles I have not yet seen the problem recur.
(In reply to Brian Candler from comment #6) > FWIW, I have just started seeing this same error on a NUC DN2820 (which is > bay trail) after upgrading it from Debian Jessie to Ubuntu Xenial. > > I have two of these NUCs, and only one showed the problem. It doesn't happen > on *every* boot but is fairly common. Unfortunately it is not writing the > dmesg to disk, but I do see hda_codec_runtime_suspend and rpm_callback in > the backtrace. Then it's likely a different issue. Please make sure that this happens with the latest vanilla kernel, and open another bug if it's confirmed.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.