[alsa-devel] v2.6.25.3 BUG (due to usb audio device?)
This locks up KDE due to kmixer somehow trying to stay attached to the audio device after resume from suspend-to-disk. This is the second time of dealing with this (or a similar bug) with the usb audio device:
http://mailman.alsa-project.org/pipermail/alsa-devel/2008-March/006867.html
I got it working by using the head of the git tree recommended in that thread (unfortunately I don't have the hash it was related to).
http://avuton.googlepages.com/config.gz http://avuton.googlepages.com/proc-version http://avuton.googlepages.com/lspci-vvv
[ 1432.710102] BUG: unable to handle kernel paging request at 00100100 [ 1432.710102] IP: [<c039be3d>] snd_ctl_dev_disconnect+0x5d/0x90 [ 1432.710102] *pde = 00000000 [ 1432.710102] Oops: 0000 [#1] PREEMPT [ 1432.710102] Modules linked in: netconsole configfs snd_pcm_oss snd_mixer_oss [ 1432.710102] [ 1432.710102] Pid: 176, comm: khubd Not tainted (2.6.25.3 #1) [ 1432.710102] EIP: 0060:[<c039be3d>] EFLAGS: 00010246 CPU: 0 [ 1432.710102] EIP is at snd_ctl_dev_disconnect+0x5d/0x90 [ 1432.710102] EAX: 00000000 EBX: 00100100 ECX: 00020004 EDX: 0000001d [ 1432.710102] ESI: f7c7dd68 EDI: f7c7dc00 EBP: f7c7dd4c ESP: f7d33e68 [ 1432.710102] DS: 007b ES: 007b FS: 0000 GS: 0000 SS: 0068 [ 1432.710102] Process khubd (pid: 176, ti=f7d32000 task=f7d30000 task.ti=f7d32000) [ 1432.710102] Stack: 00000000 f7d654e0 f7c7dc00 f7c7dd40 00000000 c039e0b6 c01a8630 f7d654e0 [ 1432.710102] c039e35a 00000000 f7c7dc00 c05502c0 f7c7e080 c039a561 f7d33ec4 f7d6d9c0 [ 1432.710102] f7d6c400 f7d6c41c c03be425 f7c7dc00 c04f0755 f7c3dc5c f7d6c400 f7d6c41c [ 1432.710102] Call Trace: [ 1432.710102] [<c039e0b6>] snd_device_disconnect+0x56/0x80 [ 1432.710102] [<c01a8630>] sysfs_ilookup_test+0x0/0x10 [ 1432.710102] [<c039e35a>] snd_device_disconnect_all+0x2a/0x50 [ 1432.710102] [<c039a561>] snd_card_disconnect+0x141/0x1b0 [ 1432.710102] [<c03be425>] usb_audio_disconnect+0x55/0x130 [ 1432.710102] [<c035ba18>] usb_unbind_interface+0x38/0x90 [ 1432.710102] [<c030df81>] __device_release_driver+0x51/0x90 [ 1432.710102] [<c030e3ad>] device_release_driver+0x1d/0x40 [ 1432.710102] [<c030d881>] bus_remove_device+0x51/0x80 [ 1432.710102] [<c030bf3c>] device_del+0xdc/0x150 [ 1432.710102] [<c035902a>] usb_disable_device+0x7a/0xe0 [ 1432.710102] [<c0354be9>] usb_disconnect+0x89/0xe0 [ 1432.710102] [<c0355996>] hub_thread+0x2f6/0xc90 [ 1432.710102] [<c0116e0b>] pick_next_task_fair+0x7b/0xb0 [ 1432.710102] [<c012dc40>] autoremove_wake_function+0x0/0x40 [ 1432.710102] [<c03556a0>] hub_thread+0x0/0xc90 [ 1432.710102] [<c012d912>] kthread+0x42/0x70 [ 1432.710102] [<c012d8d0>] kthread+0x0/0x70 [ 1432.710102] [<c0104b1f>] kernel_thread_helper+0x7/0x18 [ 1432.710102] ======================= [ 1432.710102] Code: 18 b9 01 00 00 00 ba 03 00 00 00 c7 04 24 00 00 00 00 e8 d7 c7 d7 ff 8d 43 20 b9 04 00 02 00 ba 1d 00 00 00 e8 65 9a dd ff 8b 1b <8b> 03 0f 18 00 90 39 f3 75 c9 89 e8 e8 12 5a d9 ff 89 fa 31 c0 [ 1432.710102] EIP: [<c039be3d>] snd_ctl_dev_disconnect+0x5d/0x90 SS:ESP 0068:f7d33e68 [ 1432.710127] ---[ end trace 1c91addd529faaa2 ]---
At Sun, 11 May 2008 09:53:01 -0700, Avuton Olrich wrote:
This locks up KDE due to kmixer somehow trying to stay attached to the audio device after resume from suspend-to-disk. This is the second time of dealing with this (or a similar bug) with the usb audio device:
http://mailman.alsa-project.org/pipermail/alsa-devel/2008-March/006867.html
I got it working by using the head of the git tree recommended in that thread (unfortunately I don't have the hash it was related to).
If the recent git works, it should be the series of my patches:
73d38b13ffb105ab633bd91969c8d218b2de38d4 [ALSA] Fix the race of card instance unregistration f18638dcf0c481eca2430206ebcdc7295aec8623 [ALSA] Clean up snd_card_free*() 9eb70e68f38bbc5996a2193e7b7dc0b5487a08cb [ALSA] usb-audio - Fix race in reconnection
But the patches are intrusive, and I'm not keen on applying them to stable trees...
thanks,
Takashi
On Tue, May 13, 2008 at 4:25 AM, Takashi Iwai tiwai@suse.de wrote:
At Sun, 11 May 2008 09:53:01 -0700, Avuton Olrich wrote:
This locks up KDE due to kmixer somehow trying to stay attached to the audio device after resume from suspend-to-disk. This is the second time of dealing with this (or a similar bug) with the usb audio device:
http://mailman.alsa-project.org/pipermail/alsa-devel/2008-March/006867.html
I got it working by using the head of the git tree recommended in that thread (unfortunately I don't have the hash it was related to).
If the recent git works, it should be the series of my patches:
73d38b13ffb105ab633bd91969c8d218b2de38d4 [ALSA] Fix the race of card instance unregistration f18638dcf0c481eca2430206ebcdc7295aec8623 [ALSA] Clean up snd_card_free*() 9eb70e68f38bbc5996a2193e7b7dc0b5487a08cb [ALSA] usb-audio - Fix race in reconnection
But the patches are intrusive, and I'm not keen on applying them to stable trees...
OK, finally had the time to try the newer master and here's what I get now;
I suspend and resume fine, but the sound card no longer works again after resume and the only thing I see in the dmesg is:
[ 125.873947] usb 3-1: reset full speed USB device using ohci_hcd and address 2 [ 126.017265] snd-usb-audio 3-1:1.0: no reset_resume for driver snd-usb-audio? [ 126.017265] snd-usb-audio 3-1:1.1: no reset_resume for driver snd-usb-audio? [ 126.017265] snd-usb-audio 3-1:1.2: no reset_resume for driver snd-usb-audio?
On Sat, May 17, 2008 at 7:50 AM, Avuton Olrich avuton@gmail.com wrote:
On Tue, May 13, 2008 at 4:25 AM, Takashi Iwai tiwai@suse.de wrote:
At Sun, 11 May 2008 09:53:01 -0700, Avuton Olrich wrote:
This locks up KDE due to kmixer somehow trying to stay attached to the audio device after resume from suspend-to-disk. This is the second time of dealing with this (or a similar bug) with the usb audio device:
http://mailman.alsa-project.org/pipermail/alsa-devel/2008-March/006867.html
I got it working by using the head of the git tree recommended in that thread (unfortunately I don't have the hash it was related to).
If the recent git works, it should be the series of my patches:
73d38b13ffb105ab633bd91969c8d218b2de38d4 [ALSA] Fix the race of card instance unregistration f18638dcf0c481eca2430206ebcdc7295aec8623 [ALSA] Clean up snd_card_free*() 9eb70e68f38bbc5996a2193e7b7dc0b5487a08cb [ALSA] usb-audio - Fix race in reconnection
But the patches are intrusive, and I'm not keen on applying them to stable trees...
OK, finally had the time to try the newer master and here's what I get now;
I suspend and resume fine, but the sound card no longer works again after resume and the only thing I see in the dmesg is:
[ 125.873947] usb 3-1: reset full speed USB device using ohci_hcd and address 2 [ 126.017265] snd-usb-audio 3-1:1.0: no reset_resume for driver snd-usb-audio? [ 126.017265] snd-usb-audio 3-1:1.1: no reset_resume for driver snd-usb-audio? [ 126.017265] snd-usb-audio 3-1:1.2: no reset_resume for driver snd-usb-audio?
Forgot to mention the hash for this one is linus tree: f26a3988917913b3d11b2bd741601a2c64ab9204
At Sat, 17 May 2008 07:54:04 -0700, Avuton Olrich wrote:
On Sat, May 17, 2008 at 7:50 AM, Avuton Olrich avuton@gmail.com wrote:
On Tue, May 13, 2008 at 4:25 AM, Takashi Iwai tiwai@suse.de wrote:
At Sun, 11 May 2008 09:53:01 -0700, Avuton Olrich wrote:
This locks up KDE due to kmixer somehow trying to stay attached to the audio device after resume from suspend-to-disk. This is the second time of dealing with this (or a similar bug) with the usb audio device:
http://mailman.alsa-project.org/pipermail/alsa-devel/2008-March/006867.html
I got it working by using the head of the git tree recommended in that thread (unfortunately I don't have the hash it was related to).
If the recent git works, it should be the series of my patches:
73d38b13ffb105ab633bd91969c8d218b2de38d4 [ALSA] Fix the race of card instance unregistration f18638dcf0c481eca2430206ebcdc7295aec8623 [ALSA] Clean up snd_card_free*() 9eb70e68f38bbc5996a2193e7b7dc0b5487a08cb [ALSA] usb-audio - Fix race in reconnection
But the patches are intrusive, and I'm not keen on applying them to stable trees...
OK, finally had the time to try the newer master and here's what I get now;
I suspend and resume fine, but the sound card no longer works again after resume and the only thing I see in the dmesg is:
[ 125.873947] usb 3-1: reset full speed USB device using ohci_hcd and address 2 [ 126.017265] snd-usb-audio 3-1:1.0: no reset_resume for driver snd-usb-audio? [ 126.017265] snd-usb-audio 3-1:1.1: no reset_resume for driver snd-usb-audio? [ 126.017265] snd-usb-audio 3-1:1.2: no reset_resume for driver snd-usb-audio?
Forgot to mention the hash for this one is linus tree: f26a3988917913b3d11b2bd741601a2c64ab9204
It's: Merge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc
Are you using ppc? Otherwise I suspect it's due to the addition of suspend/resume callbacks in usbaudio.c.
Takashi
On Sun, May 18, 2008 at 12:41 AM, Takashi Iwai tiwai@suse.de wrote:
At Sat, 17 May 2008 07:54:04 -0700, Avuton Olrich wrote:
On Sat, May 17, 2008 at 7:50 AM, Avuton Olrich avuton@gmail.com wrote:
On Tue, May 13, 2008 at 4:25 AM, Takashi Iwai tiwai@suse.de wrote:
At Sun, 11 May 2008 09:53:01 -0700, Avuton Olrich wrote:
This locks up KDE due to kmixer somehow trying to stay attached to the audio device after resume from suspend-to-disk. This is the second time of dealing with this (or a similar bug) with the usb audio device:
http://mailman.alsa-project.org/pipermail/alsa-devel/2008-March/006867.html
I got it working by using the head of the git tree recommended in that thread (unfortunately I don't have the hash it was related to).
If the recent git works, it should be the series of my patches:
73d38b13ffb105ab633bd91969c8d218b2de38d4 [ALSA] Fix the race of card instance unregistration f18638dcf0c481eca2430206ebcdc7295aec8623 [ALSA] Clean up snd_card_free*() 9eb70e68f38bbc5996a2193e7b7dc0b5487a08cb [ALSA] usb-audio - Fix race in reconnection
But the patches are intrusive, and I'm not keen on applying them to stable trees...
OK, finally had the time to try the newer master and here's what I get now;
I suspend and resume fine, but the sound card no longer works again after resume and the only thing I see in the dmesg is:
[ 125.873947] usb 3-1: reset full speed USB device using ohci_hcd and address 2 [ 126.017265] snd-usb-audio 3-1:1.0: no reset_resume for driver snd-usb-audio? [ 126.017265] snd-usb-audio 3-1:1.1: no reset_resume for driver snd-usb-audio? [ 126.017265] snd-usb-audio 3-1:1.2: no reset_resume for driver snd-usb-audio?
Forgot to mention the hash for this one is linus tree: f26a3988917913b3d11b2bd741601a2c64ab9204
It's: Merge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc
Are you using ppc? Otherwise I suspect it's due to the addition of suspend/resume callbacks in usbaudio.c.
Not using ppc, plain jane x86.
participants (2)
-
Avuton Olrich -
Takashi Iwai