ov511 verbose startup.

ov511 verbose startup.

[Dave Jones]

The changes to ov511 in 2.5.3 seem to generate excessive
amounts of blurb on boot up for me..

ov511.c: USB OV511+ camera found
ov511.c: model: Creative Labs WebCam 3
ov511.c: Sensor is an OV7620
ov511.c: Device registered on minor 0
usbdevfs: USBDEVFS_CONTROL failed dev 2 rqt 128 rq 6 len 137 ret -75
usb_control/bulk_msg: timeout
usbdevfs: USBDEVFS_CONTROL failed dev 2 rqt 128 rq 6 len 18 ret -110
usb_control/bulk_msg: timeout
usbdevfs: USBDEVFS_CONTROL failed dev 2 rqt 128 rq 6 len 18 ret -110
...
repeat last two lines another dozen or so times...

-- 
| Dave Jones.        http://www.codemonkey.org.uk
| SuSE Labs

Re: ov511 verbose startup.

[Greg KH]

On Thu, Jan 31, 2002 at 02:34:57AM +0100, Dave Jones wrote:
> The changes to ov511 in 2.5.3 seem to generate excessive
> amounts of blurb on boot up for me..
> 
> ov511.c: USB OV511+ camera found
> ov511.c: model: Creative Labs WebCam 3
> ov511.c: Sensor is an OV7620
> ov511.c: Device registered on minor 0
> usbdevfs: USBDEVFS_CONTROL failed dev 2 rqt 128 rq 6 len 137 ret -75
> usb_control/bulk_msg: timeout
> usbdevfs: USBDEVFS_CONTROL failed dev 2 rqt 128 rq 6 len 18 ret -110
> usb_control/bulk_msg: timeout
> usbdevfs: USBDEVFS_CONTROL failed dev 2 rqt 128 rq 6 len 18 ret -110
> ...
> repeat last two lines another dozen or so times...

What userspace program are you using that is talking to the usb device
through usbfs?  Or is this usbutils trying to determine what driver to
load?

thanks,

greg k-h

Re: ov511 verbose startup.

[Mark McClelland]

Greg KH wrote:

>On Thu, Jan 31, 2002 at 02:34:57AM +0100, Dave Jones wrote:
>
>>The changes to ov511 in 2.5.3 seem to generate excessive
>>amounts of blurb on boot up for me..
>>
>>ov511.c: USB OV511+ camera found
>>ov511.c: model: Creative Labs WebCam 3
>>ov511.c: Sensor is an OV7620
>>ov511.c: Device registered on minor 0
>>usbdevfs: USBDEVFS_CONTROL failed dev 2 rqt 128 rq 6 len 137 ret -75
>>usb_control/bulk_msg: timeout
>>usbdevfs: USBDEVFS_CONTROL failed dev 2 rqt 128 rq 6 len 18 ret -110
>>usb_control/bulk_msg: timeout
>>usbdevfs: USBDEVFS_CONTROL failed dev 2 rqt 128 rq 6 len 18 ret -110
>>...
>>repeat last two lines another dozen or so times...
>>
>
>What userspace program are you using that is talking to the usb device
>through usbfs?  Or is this usbutils trying to determine what driver to
>load?
>
I have been getting reports of -75 (babble?) and -110 errors with both 
control and iso transfers. The problematic kernels seem to be 2.4.17+, 
with uhci HCD. IIRC, nearly all of the reports mentioned a Via chipset. 
There are sporadic reports of corrupted iso packets (blocks of zeros 
inserted randomly) with uhci under 2.4.17 as well.

This is the first case of a usbfs-related ov511 error that I have seen. 
Very strange.

Greg (if you know): usbfs is not allowed to access claimed interfaces, 
correct? (ie. ones that are implicitly claimed because of a successful 
return from probe()). Are interfaces treated as claimed while probe() is 
active, so that user-space "probes" cannot interfere with driver probes()?

I use usb-uhci, uhci, and usb-ohci regularly with ov511, and I have 
never seen any of these problems. There are clearly a number of factors 
involved here.

-- 
Mark McClelland
mmcclell@bigfoot.com

Re: ov511 verbose startup.

[Greg KH]

On Wed, Jan 30, 2002 at 09:31:07PM -0800, Mark McClelland wrote:
> 
> Greg (if you know): usbfs is not allowed to access claimed interfaces, 
> correct? (ie. ones that are implicitly claimed because of a successful 
> return from probe()). Are interfaces treated as claimed while probe() is 
> active, so that user-space "probes" cannot interfere with driver probes()?

Yes, but odds are, it is trying to read the configuration of the device,
and we don't have control pipe locking, yet :)

Dave, does this problem go away on 2.5.3-pre6?
And which host controller driver are you using?

thanks,

greg k-h

Re: ov511 verbose startup.

[Dave Jones]

On Wed, 30 Jan 2002, Greg KH wrote:

> Yes, but odds are, it is trying to read the configuration of the device,
> and we don't have control pipe locking, yet :)
> Dave, does this problem go away on 2.5.3-pre6?

Yes. From what I recall. I'll build a pre6 later to double check.

> And which host controller driver are you using?

CONFIG_USB_UHCI_ALT=y

-- 
| Dave Jones.        http://www.codemonkey.org.uk
| SuSE Labs

Re: ov511 verbose startup.

[Dave Jones]

On Wed, Jan 30, 2002 at 09:31:24PM -0800, Greg KH wrote:
 > Dave, does this problem go away on 2.5.3-pre6?

 Great, heisenbug. Subsequent reboots under any kernel
 haven't managed to reproduce the problem.

-- 
| Dave Jones.        http://www.codemonkey.org.uk
| SuSE Labs

Re: ov511 verbose startup.

[Dave Jones]

On Wed, 30 Jan 2002, Greg KH wrote:

> > usbdevfs: USBDEVFS_CONTROL failed dev 2 rqt 128 rq 6 len 18 ret -110
> > usb_control/bulk_msg: timeout
> > usbdevfs: USBDEVFS_CONTROL failed dev 2 rqt 128 rq 6 len 18 ret -110
> > ...
> > repeat last two lines another dozen or so times...
> What userspace program are you using that is talking to the usb device
> through usbfs?  Or is this usbutils trying to determine what driver to
> load?

More than likely usbutils. There shouldn't be anything else running
on startup that touches usb.

-- 
| Dave Jones.        http://www.codemonkey.org.uk
| SuSE Labs