Arch Linux

diederick76

Member

Registered: 2010-02-14

Posts: 169

Crash at bluetooth disconnect

Hi,

Whenever I:

*suspend to disk,
*stop the bluetooth service, or
*pull the bluetooth dongle from its USB port

my notebook crashed with this:

May  3 11:53:58 localhost -- MARK --
May  3 11:58:17 localhost kernel: [ 1755.998609] usb 3-1.2: USB disconnect, device number 4
May  3 11:58:17 localhost kernel: [ 1756.022728] PGD 0 
May  3 11:58:17 localhost kernel: [ 1756.022785] Oops: 0000 [#1] PREEMPT SMP 
May  3 11:58:17 localhost kernel: [ 1756.022910] CPU 3 
May  3 11:58:17 localhost kernel: [ 1756.022960] Modules linked in: hidp rfcomm bnep coretemp snd_hda_codec_hdmi snd_hda_codec_realtek joydev snd_hda_intel snd_hda_codec snd_hwdep snd_pcm snd_page_allo
c arc4 uvcvideo snd_timer videobuf2_vmalloc snd videobuf2_memops btusb atl1c videobuf2_core videodev iwlwifi i915 i2c_algo_bit serio_raw bluetooth drm_kms_helper drm v4l2_compat_ioctl32 mac80211 intel_
agp iTCO_wdt iTCO_vendor_support soundcore cfg80211 psmouse intel_gtt media asus_nb_wmi mxm_wmi asus_wmi i2c_i801 pci_hotplug i2c_core pcspkr sparse_keymap rfkill wmi thermal evdev button ac battery vi
deo cpufreq_userspace cpufreq_powersave cpufreq_conservative acpi_cpufreq mperf processor ext4 crc16 jbd2 mbcache usbhid hid sd_mod ahci libahci libata scsi_mod xhci_hcd ehci_hcd usbcore usb_common
May  3 11:58:17 localhost kernel: [ 1756.024485] 
May  3 11:58:17 localhost kernel: [ 1756.024505] Pid: 66, comm: khubd Not tainted 3.3.4-1-ARCH #1 ASUSTeK Computer Inc. U36SD/U36SD
May  3 11:58:17 localhost kernel: [ 1756.024562] RIP: 0010:[<ffffffffa03401dd>]  [<ffffffffa03401dd>] hci_conn_hash_flush+0x7d/0xc0 [bluetooth]
May  3 11:58:17 localhost kernel: [ 1756.024612] RSP: 0018:ffff8801475d5bc0  EFLAGS: 00010282
May  3 11:58:17 localhost kernel: [ 1756.024638] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 00000001820001cd
May  3 11:58:17 localhost kernel: [ 1756.024668] RDX: 00000001820001ce RSI: ffff8801475d5fd8 RDI: ffffffffa01cf780
May  3 11:58:17 localhost kernel: [ 1756.024697] RBP: ffff8801475d5be0 R08: 00000000820001cd R09: ffff8801475d5fd8
May  3 11:58:17 localhost kernel: [ 1756.024728] R10: 0000000000000000 R11: ffffffff8123e3a9 R12: ffff880147e8c658
May  3 11:58:17 localhost kernel: [ 1756.024759] R13: ffff880147e8c5e0 R14: ffff880146c91800 R15: 0000000000000000
May  3 11:58:17 localhost kernel: [ 1756.024789] FS:  0000000000000000(0000) GS:ffff88014f460000(0000) knlGS:0000000000000000
May  3 11:58:17 localhost kernel: [ 1756.024828] CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
May  3 11:58:17 localhost kernel: [ 1756.024852] CR2: 0000000000000021 CR3: 0000000001805000 CR4: 00000000000406e0
May  3 11:58:17 localhost kernel: [ 1756.024884] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
May  3 11:58:17 localhost kernel: [ 1756.024917] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
May  3 11:58:17 localhost kernel: [ 1756.024950] Process khubd (pid: 66, threadinfo ffff8801475d4000, task ffff880147a43800)
May  3 11:58:17 localhost kernel: [ 1756.024983] Stack:
May  3 11:58:17 localhost kernel: [ 1756.024993]  ffff8801475d5be0 0000000000000000 ffff880147e8c000 ffff880147e8c010
May  3 11:58:17 localhost kernel: [ 1756.025032]  ffff8801475d5c10 ffffffffa033bc41 ffff880147e8c000 0000000000000000
May  3 11:58:17 localhost kernel: [ 1756.025100]  ffff880148eef800 ffff880146c91800 ffff8801475d5c30 ffffffffa033e253
May  3 11:58:17 localhost kernel: [ 1756.025158] Call Trace:
May  3 11:58:17 localhost kernel: [ 1756.025191]  [<ffffffffa033bc41>] hci_dev_do_close+0xa1/0x2c0 [bluetooth]
May  3 11:58:17 localhost kernel: [ 1756.025246]  [<ffffffffa033e253>] hci_unregister_dev+0x63/0x180 [bluetooth]
May  3 11:58:17 localhost kernel: [ 1756.025306]  [<ffffffffa01cdcc0>] btusb_disconnect+0x60/0xd0 [btusb]
May  3 11:58:17 localhost kernel: [ 1756.025388]  [<ffffffffa00111e0>] usb_unbind_interface+0x50/0x180 [usbcore]
May  3 11:58:17 localhost kernel: [ 1756.025447]  [<ffffffff8130abac>] __device_release_driver+0x7c/0xe0
May  3 11:58:17 localhost kernel: [ 1756.025489]  [<ffffffff8130ac3c>] device_release_driver+0x2c/0x40
May  3 11:58:17 localhost kernel: [ 1756.025541]  [<ffffffff8130a6d1>] bus_remove_device+0xe1/0x120
May  3 11:58:17 localhost kernel: [ 1756.025588]  [<ffffffff8130822a>] device_del+0x12a/0x1b0
May  3 11:58:17 localhost kernel: [ 1756.025633]  [<ffffffffa000ef8f>] usb_disable_device+0xaf/0x1f0 [usbcore]
May  3 11:58:17 localhost kernel: [ 1756.025688]  [<ffffffffa0007427>] usb_disconnect+0x87/0x120 [usbcore]
May  3 11:58:17 localhost kernel: [ 1756.025742]  [<ffffffffa0008ddb>] hub_thread+0x54b/0x12a0 [usbcore]
May  3 11:58:17 localhost kernel: [ 1756.025798]  [<ffffffff81072570>] ? abort_exclusive_wait+0xb0/0xb0
May  3 11:58:17 localhost kernel: [ 1756.025856]  [<ffffffffa0008890>] ? usb_remote_wakeup+0x40/0x40 [usbcore]
May  3 11:58:17 localhost kernel: [ 1756.025917]  [<ffffffff81071bf3>] kthread+0x93/0xa0
May  3 11:58:17 localhost kernel: [ 1756.025959]  [<ffffffff81461224>] kernel_thread_helper+0x4/0x10
May  3 11:58:17 localhost kernel: [ 1756.026002]  [<ffffffff81071b60>] ? kthread_freezable_should_stop+0x70/0x70
May  3 11:58:17 localhost kernel: [ 1756.026056]  [<ffffffff81461220>] ? gs_change+0x13/0x13
May  3 11:58:17 localhost kernel: [ 1756.026096] Code: 68 04 00 00 48 85 c0 74 0a be 16 00 00 00 48 89 df ff d0 48 89 df e8 63 fd ff ff 48 8b 03 48 89 45 e8 48 8b 5d e8 49 39 dc 74 33 <0f> b6 53 21 66 c7 43 1e 09 00 80 fa 01 74 14 73 a2 be 16 00 00 
May  3 11:58:17 localhost kernel: [ 1756.033899]  RSP <ffff8801475d5bc0>
May  3 11:58:17 localhost kernel: [ 1756.035999] CR2: 0000000000000021
May  3 11:58:17 localhost kernel: [ 1756.166122] ---[ end trace 27ebb663691f9227 ]---

$ uname -a
Linux bob 3.3.4-1-ARCH #1 SMP PREEMPT Sat Apr 28 00:21:22 CEST 2012 x86_64 Intel(R) Core(TM) i5-2410M CPU @ 2.30GHz GenuineIntel GNU/Linux

This is consistently since kernel 3.3.0. I have downgraded to kernel 3.2.9 and the problem went away. Is this a problem with the bluetooth driver in the newer kernels?

I have read this thread and used those hints to look through the logs, but my kernel seems untainted and I'm not using the tp_smapi or mei modules. I have also googled for "khub crash" but that didn't show anything useful to me.

Does anybody know what I could do about this, apart from not upgrading my kernel anymore?

Thanks for any help!

Last edited by diederick76 (2012-05-04 09:07:21)

swiftgeek

Member

Registered: 2011-04-19

Posts: 31

Re: Crash at bluetooth disconnect

I can only confirm that
http://pastebin.com/a16aNDJZ

But for me it happens occasionally and sometimes EVEN WITH DISABLED BT! ( rfkilled even days before crash! )

Also there is a small bug in nvidia driver which acts quite similar, but this is a different bug

Last edited by swiftgeek (2012-05-08 20:35:23)

diederick76

Member

Registered: 2010-02-14

Posts: 169

Re: Crash at bluetooth disconnect

I have found out that the problem doesn't occur when the BT device hasn't yet made any connection. Do you think a kernel bug report in order?

Last edited by diederick76 (2012-05-08 20:48:04)

tljm

Member

Registered: 2008-12-20

Posts: 10

Re: Crash at bluetooth disconnect

I can only confirm that this occur on my HP dv6500 as well. I've noticed it does happen on normal and lts kernel as well. Under lts kernel I can disconnect BT device using command:

hciconfig hci0 down

Then I can suspend successfully. However, this is only a kind of workaround, it does not solve the problem.

Last edited by tljm (2012-05-12 21:19:55)

diederick76

Member

Registered: 2010-02-14

Posts: 169

Re: Crash at bluetooth disconnect

I can only confirm that this occur on my HP dv6500 as well. I've noticed it does happen on normal and lts kernel as well. Under lts kernel I can disconnect BT device using command:

hciconfig hci0 down

Then I can suspend successfully. However, this is only a kind of workaround, it does not solve the problem.

That doesn't work on my machine. Anything that I do that stops the bluetooth connection, except for turning the machine off, will give the above results.

tljm

Member

Registered: 2008-12-20

Posts: 10

Re: Crash at bluetooth disconnect

This is a good question. What happens when you turn the machine off? I have experience that even with hciconfig trick turning off my computer take quite long time if any BT connection was active. Few weeks ago (I do not turn my machine off very often) I have discovered that during turning off  one process does not answer to SIGTERM correctly. Guess which one. Yes, bluetoothd.
Something is apparently wrong and unfortunately I have not enough experience to debug this. Does anyone have any idea?

Last edited by tljm (2012-05-13 18:01:47)