USB/BlueTooth oops in 2.6.5

USB/BlueTooth oops in 2.6.5

[Daniel Pittman]

When I try to turn on the BlueTooth interface in my laptop, it turns on
a USB device.  Doing that with 2.6.5 generates the following error.

  Daniel

usb 3-1: new full speed USB device using address 2
Unable to handle kernel NULL pointer dereference at virtual address 00000004
 printing eip:
c02c43b2
*pde = 00000000
Oops: 0000 [#1]
CPU:    0
EIP:    0060:[<c02c43b2>]    Not tainted
EFLAGS: 00010296   (2.6.5-enki) 
EIP is at usb_disable_interface+0x14/0x46
eax: d24ca780   ebx: 00000000   ecx: 00000282   edx: dff6f100
esi: 00000002   edi: 00000000   ebp: d4dedc00   esp: dfdafd48
ds: 007b   es: 007b   ss: 0068
Process khubd (pid: 5, threadinfo=dfdae000 task=dff0e080)
Stack: 00000002 0000000b 00000001 00000002 ce3dcdb0 d4dedc00 c02c4608 d4dedc00 
       d24ca780 0000000b 00000001 00000002 00000001 00000000 00000000 00001388 
       00000000 d24ca780 00000000 d59f9880 d59f9938 d24cac80 c02f283d d4dedc00 
Call Trace:
 [<c02c4608>] usb_set_interface+0xb7/0x173
 [<c02f283d>] hci_usb_probe+0x22f/0x480
 [<c015a392>] alloc_inode+0x146/0x14b
 [<c01715f5>] sysfs_new_inode+0x5d/0xa2
 [<c02bf03b>] usb_probe_interface+0x61/0x6e
 [<c024b86d>] bus_match+0x3f/0x6a
 [<c024b8d9>] device_attach+0x41/0x91
 [<c024ba98>] bus_add_device+0x5b/0x9f
 [<c024aad4>] device_add+0xa1/0x120
 [<c02c4996>] usb_set_configuration+0x1d4/0x25f
 [<c02bfdc2>] usb_new_device+0x250/0x3c3
 [<c02c14f1>] hub_port_connect_change+0x177/0x274
 [<c02c1895>] hub_events+0x2a7/0x2fa
 [<c02c1915>] hub_thread+0x2d/0xe4
 [<c0116159>] default_wake_function+0x0/0x12
 [<c02c18e8>] hub_thread+0x0/0xe4
 [<c0104d1d>] kernel_thread_helper+0x5/0xb

Code: 80 7b 04 00 74 24 31 f6 8b 43 0c 83 c7 01 0f b6 44 30 02 83 

-- 
Glass is, we now know, a 'slow liquid;' and we're slow dust.
        -- Albert Goldbarth

Re: USB/BlueTooth oops in 2.6.5

[Janne Pikkarainen]

On Wed, 2004-04-07 at 09:33, Daniel Pittman wrote:
> When I try to turn on the BlueTooth interface in my laptop, it turns on
> a USB device.  Doing that with 2.6.5 generates the following error.

Enabling "SCO (voice) support" under Device Drivers --> Networking
support --> Bluetooth support --> Bluetooth device drivers --> HCI USB
driver causes the same error for me, after disabling SCO everything
works ok. The SCO really is the root of all evil, you know... ;-)

And this is with my desktop box: AMD Athlon XP 2400+ and Asus A7V-133
motherboard, with some USB bluetooth dongle made by SCM.

I hope this little piece of information helps you.

> usb 3-1: new full speed USB device using address 2
> Unable to handle kernel NULL pointer dereference at virtual address 00000004
>  printing eip:
> c02c43b2
> *pde = 00000000
> Oops: 0000 [#1]
> CPU:    0
> EIP:    0060:[<c02c43b2>]    Not tainted
> EFLAGS: 00010296   (2.6.5-enki) 
> EIP is at usb_disable_interface+0x14/0x46
> eax: d24ca780   ebx: 00000000   ecx: 00000282   edx: dff6f100
> esi: 00000002   edi: 00000000   ebp: d4dedc00   esp: dfdafd48
> ds: 007b   es: 007b   ss: 0068
> Process khubd (pid: 5, threadinfo=dfdae000 task=dff0e080)
> Stack: 00000002 0000000b 00000001 00000002 ce3dcdb0 d4dedc00 c02c4608 d4dedc00 
>        d24ca780 0000000b 00000001 00000002 00000001 00000000 00000000 00001388 
>        00000000 d24ca780 00000000 d59f9880 d59f9938 d24cac80 c02f283d d4dedc00 
> Call Trace:
>  [<c02c4608>] usb_set_interface+0xb7/0x173
>  [<c02f283d>] hci_usb_probe+0x22f/0x480
>  [<c015a392>] alloc_inode+0x146/0x14b
>  [<c01715f5>] sysfs_new_inode+0x5d/0xa2
>  [<c02bf03b>] usb_probe_interface+0x61/0x6e
>  [<c024b86d>] bus_match+0x3f/0x6a
>  [<c024b8d9>] device_attach+0x41/0x91
>  [<c024ba98>] bus_add_device+0x5b/0x9f
>  [<c024aad4>] device_add+0xa1/0x120
>  [<c02c4996>] usb_set_configuration+0x1d4/0x25f
>  [<c02bfdc2>] usb_new_device+0x250/0x3c3
>  [<c02c14f1>] hub_port_connect_change+0x177/0x274
>  [<c02c1895>] hub_events+0x2a7/0x2fa
>  [<c02c1915>] hub_thread+0x2d/0xe4
>  [<c0116159>] default_wake_function+0x0/0x12
>  [<c02c18e8>] hub_thread+0x0/0xe4
>  [<c0104d1d>] kernel_thread_helper+0x5/0xb
> 
> Code: 80 7b 04 00 74 24 31 f6 8b 43 0c 83 c7 01 0f b6 44 30 02 83 

Re: USB/BlueTooth oops in 2.6.5

[Greg KH]

On Wed, Apr 07, 2004 at 04:33:58PM +1000, Daniel Pittman wrote:
> When I try to turn on the BlueTooth interface in my laptop, it turns on
> a USB device.  Doing that with 2.6.5 generates the following error.

It's being worked on:

	http://bugme.osdl.org/show_bug.cgi?id=2423

thanks,

greg k-h

Re: USB/BlueTooth oops in 2.6.5

[Steven Walter]

I get a very similar oops backtrace, but from a different cause.  
Whenever I plug in my Zaurus for the /second/ time (i.e., plug it in, 
let usbnet find it, unplug it, then plug it in again), I get the 
attached oops backtrace.  This did not occur with 2.6.3; unsure about 2.6.4.

Just like in the bluetooth case, usb_disable_interface is causing the 
oops, and is being called by usb_set_interface.  I tried the patch in 
the comments of your linked bug, however that did not fix the problem.

usb0: unregister usbnet usb-0000:00:11.2-1.1, Sharp Zaurus SL-5x00
usb 1-1.1: new full speed USB device using address 7
Unable to handle kernel NULL pointer dereference at virtual address 00000004
 printing eip:
e18c4ef4
*pde = 00000000
Oops: 0000 [#1]
PREEMPT
CPU:    0
EIP:    0060:[<e18c4ef4>]    Not tainted
EFLAGS: 00010286   (2.6.5)
EIP is at usb_disable_interface+0x14/0x50 [usbcore]
eax: de363cc0   ebx: 00000000   ecx: decb6000   edx: dffff0c0
esi: 00000001   edi: 00000000   ebp: de2b5000   esp: decb7d60
ds: 007b   es: 007b   ss: 0068
Process khubd (pid: 5367, threadinfo=decb6000 task=deb0cd00)
Stack: 00000001 0000000b 00000001 00000001 de73f418 de2b5000 e18c5177 
de2b5000
       de363cc0 0000000b 00000001 00000001 00000001 00000000 00000000 
00001388
       00000000 de363cc0 de73f418 de73f418 de73f2c0 00000001 e19890c1 
de2b5000
Call Trace:
 [<e18c5177>] usb_set_interface+0xb7/0x180 [usbcore]
 [<e19890c1>] get_endpoints+0xb1/0x120 [usbnet]
 [<e198925f>] generic_cdc_bind+0xcf/0x220 [usbnet]
 [<e198ac57>] usbnet_probe+0x3c7/0x400 [usbnet]
 [<c0169d22>] dput+0x22/0x210
 [<e18bf081>] usb_probe_interface+0x61/0x80 [usbcore]
 [<c01f9acf>] bus_match+0x3f/0x70
 [<c01f9b41>] device_attach+0x41/0xa0
 [<c01f9d2b>] bus_add_device+0x5b/0xa0
 [<c01f8bd1>] device_add+0xa1/0x130
 [<e18c5516>] usb_set_configuration+0x1d6/0x270 [usbcore]
 [<e18c00e9>] usb_new_device+0x249/0x3c0 [usbcore]
 [<e18c1a18>] hub_port_connect_change+0x178/0x280 [usbcore]
 [<e18c1dea>] hub_events+0x2ca/0x340 [usbcore]
 [<e18c1e8d>] hub_thread+0x2d/0xf0 [usbcore]
 [<c010725e>] ret_from_fork+0x6/0x14
 [<c011a180>] default_wake_function+0x0/0x20
 [<e18c1e60>] hub_thread+0x0/0xf0 [usbcore]
 [<c0105291>] kernel_thread_helper+0x5/0x14
 
Code: 80 7b 04 00 74 26 31 f6 8d 74 26 00 8b 43 0c 47 0f b6 44 30

-- 
--Steven
"A is A."
	-Ayn Rand
GnuPG Fingerprint: 889A 5BED F01D 61BC 930F  A915 DB55 2585 0010 A205

Re: USB/BlueTooth oops in 2.6.5

[Greg KH]

On Fri, Apr 09, 2004 at 01:55:25PM -0400, Steven Walter wrote:
> I get a very similar oops backtrace, but from a different cause.  
> Whenever I plug in my Zaurus for the /second/ time (i.e., plug it in, 
> let usbnet find it, unplug it, then plug it in again), I get the 
> attached oops backtrace.  This did not occur with 2.6.3; unsure about 2.6.4.

Does it occur in the latest -mm kernel?

thanks,

greg k-h