Reply-To: akpm@linux-foundation.org


(switched to email.  Please respond via emailed reply-to-all, not via the
bugzilla web interface).

(Might be an alsa bug?)

On Thu,  4 Sep 2008 13:43:41 -0700 (PDT)
bugme-daemon@bugzilla.kernel.org wrote:

> http://bugzilla.kernel.org/show_bug.cgi?id=11499
> 
>            Summary: BUG & Oops with USB sound card 0d8c:000c C-Media
>                     Electronics
>            Product: Drivers
>            Version: 2.5
>      KernelVersion: 2.6.27-rc5
>           Platform: All
>         OS/Version: Linux
>               Tree: Mainline
>             Status: NEW
>           Severity: normal
>           Priority: P1
>          Component: USB
>         AssignedTo: greg@kroah.com
>         ReportedBy: kernel6org@pook.es
> 
> 
> Latest working kernel version:?
> Earliest failing kernel version:2.6.27-rc5
> Distribution: Debian
> Hardware Environment:Intel Core 2 Duo E6550
> Software Environment:2.6.27-rc5 i686 32 bit
> Problem Description:using the USB sound card causes an Oops
> 
> Steps to reproduce:
> 
> I just bought a Philips SHM6105/00 VoIP headset that came with a USB sound
> card
> which is identified as a 0d8c:000c C-Media Electronics, Inc. Audio Adapter. I
> see to be able play sound ok but recording and playing at the same time
> causes
> problems. This morning I got a total USB freeze (even the light on my USB
> mouse
> went off) and tonight I got the "BUG". I guess that I have to reboot as I
> have
> lots of hung processes.
> 
> I as using a vanilla kernel.
> 
> [46192.928097] usb 4-1.2: reset full speed USB device using ehci_hcd and
> address 7
> [46866.188431] BUG: unable to handle kernel paging request at 00100100
> [46866.188435] IP: [<f8974611>] :snd_pcm:snd_pcm_attach_substream+0x79/0x2a8
> [46866.188446] *pde = 00000000 
> [46866.188449] Oops: 0000 [#1] SMP 
> [46866.188451] Modules linked in: bnep rfcomm l2cap tun binfmt_misc kvm_intel
> kvm video output ac battery bridge stp ip6t_LOG nf_conntrack_ipv6 ipt_REJECT
> xt_limit nf_conntrack_ipv4 xt_state nf_conntrack xt_comment ts_bm xt_string
> xt_length xt_tcpudp xt_physdev ipt_LOG ip6table_filter ip6_tables
> iptable_filter ip_tables x_tables ipv6 vfat fat loop isofs zlib_inflate
> sha256_generic aes_generic cbc dm_crypt crypto_blkcipher dm_snapshot
> dm_mirror
> dm_log dm_mod coretemp it87 hwmon_vid usb_storage btusb bluetooth
> snd_usb_audio
> snd_usb_lib snd_seq_dummy snd_seq_oss usbhid snd_seq_midi hid
> snd_seq_midi_event snd_seq ff_memless snd_rawmidi snd_seq_device snd_hwdep
> snd_hda_intel snd_mixer_oss snd_pcm ide_cd_mod i2c_i801 cdrom parport_pc
> parport ehci_hcd iTCO_wdt button snd_timer i2c_core uhci_hcd snd intel_agp
> evdev soundcore snd_page_alloc pcspkr
> [46866.188505] 
> [46866.188507] Pid: 28719, comm: arecord Not tainted (2.6.27-rc5 #1)
> [46866.188509] EIP: 0060:[<f8974611>] EFLAGS: 00210212 CPU: 0
> [46866.188518] EIP is at snd_pcm_attach_substream+0x79/0x2a8 [snd_pcm]
> [46866.188520] EAX: 00007030 EBX: f712ed70 ECX: 00100100 EDX: ef1faa40
> [46866.188522] ESI: e556fe78 EDI: 00100100 EBP: ef3b1e00 ESP: e556fe24
> [46866.188524]  DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068
> [46866.188526] Process arecord (pid: 28719, ti=e556e000 task=ef1faa40
> task.ti=e556e000)
> [46866.188528] Stack: ef1ef5c0 00000001 f712ed4c e556fe48 e556fe78 ef3b1f10
> ef1ef5c0 f8976c60 
> [46866.188533]        e556fe48 00000000 fffffff2 ef3b1e00 f8976d5f e556fe78
> 00000001 ef3b1f20 
> [46866.188537]        00000000 ef1faa40 c021ed6b ef3b1f24 ef3b1f24 ef1ef5c0
> 00000000 f88fe960 
> [46866.188542] Call Trace:
> [46866.188545]  [<f8976c60>] snd_pcm_open_substream+0x11/0x6c [snd_pcm]
> [46866.188557]  [<f8976d5f>] snd_pcm_open+0xa4/0x179 [snd_pcm]
> [46866.188566]  [<c021ed6b>] default_wake_function+0x0/0x8
> [46866.188574]  [<f88fa4b8>] snd_open+0xb7/0x10f [snd]
> [46866.188584]  [<c027c8e8>] chrdev_open+0x132/0x149
> [46866.188587]  [<c027c7b6>] chrdev_open+0x0/0x149
> [46866.188590]  [<c0278f78>] __dentry_open+0x10d/0x1fc
> [46866.188593]  [<c02790e4>] nameidata_to_filp+0x1c/0x2c
> [46866.188595]  [<c0283879>] do_filp_open+0x34b/0x6a0
> [46866.188601]  [<f88fa778>] snd_card_file_remove+0xdc/0xe4 [snd]
> [46866.188611]  [<f88fbf4a>] snd_ctl_empty_read_queue+0x13/0x49 [snd]
> [46866.188620]  [<c028c423>] alloc_fd+0xb9/0xc5
> [46866.188623]  [<c0278d99>] do_sys_open+0x3f/0xb2
> [46866.188626]  [<c0278e50>] sys_open+0x1e/0x23
> [46866.188628]  [<c02038a3>] sysenter_do_call+0x12/0x2f
> [46866.188631]  =======================
> [46866.188632] Code: c7 8b 8b 70 01 00 00 81 c3 70 01 00 00 eb 1c 64 8b 15 00
> d0 58 c0 8b 41 0c 3b 82 04 01 00 00 75 08 8b 71 10 83 fe ff 75 0f 89 f9 <8b>
> 39
> 0f 18 07 90 39 d9 75 da 83 ce ff 8b 44 24 08 e8 6f 22 8c 
> [46866.188657] EIP: [<f8974611>] snd_pcm_attach_substream+0x79/0x2a8
> [snd_pcm]
> SS:ESP 0068:e556fe24
> [46866.188670] ---[ end trace 3974142a4c620b4a ]---
> 
> 

On 04/09/08 23:06, Andrew Morton wrote:
> (Might be an alsa bug?)

I just tried my other USB sound card (a TerraTec) with twinkle (a VoIP 
client) and it crashed my USB sub-system as well. The light on my mouse 
went out and everything USB went dead (my keyboard is USB so I had to 
use the power button).

My other USB sound card is a 0ccd:0077 TerraTec Electronic GmbH (Aureon 
Dual USB).

As with the C-Media 0d8c:000c device, this sound adapter works fine when 
I use it to play music. It is recording (or playing and recording 
together) that causes problems.

I can boot linux 2.6.26 tomorrow and test if you like.

I'm using my USB sound adapter with my VoIP client because I cannot get 
the new btusb driver to work with my bluetooth headset. Would you like a 
bug report about that as well? (I have already sent emails to 
linux-bluetooth@vger.kernel.org 
http://marc.info/?l=linux-bluetooth&m=122038469902850&w=2). My bluetooth 
headset works with a patched 2.6.23 kernel.

thanks!
Stuart

At Thu, 4 Sep 2008 14:06:02 -0700,
Andrew Morton wrote:
> 
> 
> (switched to email.  Please respond via emailed reply-to-all, not via the
> bugzilla web interface).
> 
> (Might be an alsa bug?)

There is no significant difference regarding USB-audio since 2.6.26.
If it's a regression of usb-audio, it must be in the earlier version,
I guess.

One thing you can try is to turn on CONFIG_SND_DEBUG=y if not done.
This enables some NULL checks.


Takashi

Created attachment 17627 [details]
the config of my 2.6.27-rc5 kernel

I'll recompile to add CONFIG_SND_DEBUG=y this weekend.

Any news

On 23/09/08 23:15, bugme-daemon@bugzilla.kernel.org wrote:
> Any news

The oops only happened once but I stopped using the input part of the 
sound card and am now only using it as an output device (the input sound 
quality is not as good as my Terratec Aureon Dual USB).

Because the oops never happened again I never did recompile with 
CONFIG_SND_DEBUG=y.

I am planning to compile 2.6.27-rc7 soon as I would really like a 
solution to http://bugzilla.kernel.org/show_bug.cgi?id=11514 (bluetooth 
headset works with 2.6.26, only hisses with 2.6.27-rc5) I will add 
CONFIG_SND_DEBUG=y when I do.

Stuart