Kernel 2.4.18 and VIA USB

Kernel 2.4.18 and VIA USB

[Will Newton]

I have seen several posts in the archives that mention the following error on 
VIA chipset (usually K7V Athlon boards):

hub.c: USB new device connect on bus1/2, assigned device number 7
usb.c: USB device not accepting new address=7 (error=-110)
hub.c: Cannot enable port 2 of hub 1, disabling port.
hub.c: Maybe the USB cable is bad?

If anyone could shed any light on this or recommend a test to find out more I 
would be very grateful. It may or may not be related to my other USB problem.

I am seeing this error on attempting a bulk transfer to a USB camera:

usbdevfs: USBDEVFS_BULK failed dev 8 ep 0x81 len 4096 ret -110

The thing that strikes me about this is that the endpoint in question is 
specified thus:

			Endpoint Address: 81
			Direction: in
			Attribute: 2
			Type: Bulk
			Max Packet Size: 64
			Interval:   0ms

Max packet size of 64 would seem to indicate that packets of too large a size 
are being sent to the endpoint. Is this a correct analysis?

Any help with either of these problems would be much appreciated. Please CC 
any replies. Thanks.

Re: Kernel 2.4.18 and VIA USB

[Greg KH]

On Sun, May 19, 2002 at 04:56:19PM +0100, Will Newton wrote:
> 
> I have seen several posts in the archives that mention the following error on 
> VIA chipset (usually K7V Athlon boards):
> 
> hub.c: USB new device connect on bus1/2, assigned device number 7
> usb.c: USB device not accepting new address=7 (error=-110)
> hub.c: Cannot enable port 2 of hub 1, disabling port.
> hub.c: Maybe the USB cable is bad?

Which USB host controller driver are you using?
And are you sure your cable isn't bad?  :)

> If anyone could shed any light on this or recommend a test to find out more I 
> would be very grateful. It may or may not be related to my other USB problem.
> 
> I am seeing this error on attempting a bulk transfer to a USB camera:
> 
> usbdevfs: USBDEVFS_BULK failed dev 8 ep 0x81 len 4096 ret -110

Are you using usbdevfs/libusb to send data to this camera (like gphoto
does?)

> The thing that strikes me about this is that the endpoint in question is 
> specified thus:
> 
> 			Endpoint Address: 81
> 			Direction: in
> 			Attribute: 2
> 			Type: Bulk
> 			Max Packet Size: 64
> 			Interval:   0ms
> 
> Max packet size of 64 would seem to indicate that packets of too large a size 
> are being sent to the endpoint. Is this a correct analysis?

No, it's ok to send larger packet sizes to endpoints, the USB host
controller driver divides up the packet into smaller pieces for the
device, based on the endpoint size.

> Any help with either of these problems would be much appreciated. Please CC 
> any replies. Thanks.

Can you check out the 2.4.19-pre kernel tree right now?  Lots of good
USB host controller cleanups have happened in there that might help you
out.

thanks,

greg k-h

Re: Kernel 2.4.18 and VIA USB

[Greg KH]

On Mon, May 20, 2002 at 09:34:42PM +0100, Will Newton wrote:
> On Monday 20 May 2002 6:53 pm, you wrote:
> 
> > > hub.c: USB new device connect on bus1/2, assigned device number 7
> > > usb.c: USB device not accepting new address=7 (error=-110)
> > > hub.c: Cannot enable port 2 of hub 1, disabling port.
> > > hub.c: Maybe the USB cable is bad?
> >
> > Which USB host controller driver are you using?
> 
> usb-uhci.c: $Revision: 1.275 $ time 11:55:11 Apr 14 2002
> usb-uhci.c: High bandwidth mode enabled
> usb-uhci.c: v1.275:USB Universal Host Controller Interface driver

When trying out 2.4.19-pre, also try the uhci.c driver.

> > And are you sure your cable isn't bad?  :)
> 
> No, and it isn't feasible for me to test this right now, as it's a custom (I 
> think) cable and I have only one cable and one USB capable machine.

Ah, custom cable, not good :(

> Also a google search for "Maybe the USB cable is bad?" seems to turn up only 
> VIA users.

Yes, there are a lot of odd VIA problems, but then again, there are a
lot of USB cables out there in the wild that are not compliant with the
USB spec :)


greg k-h

Re: Kernel 2.4.18 and VIA USB

[Kristofer T. Karas]

Greg KH wrote:

> On Sun, May 19, 2002 at 04:56:19PM +0100, Will Newton wrote:
> > hub.c: Cannot enable port 2 of hub 1, disabling port.
> > hub.c: Maybe the USB cable is bad?
>
> Which USB host controller driver are you using?
> And are you sure your cable isn't bad?  :)

Just a long shot, but ....

I get this message when I plug my Microsoft Internet keyboard (which has both PS/2
and USB interfaces) into my VIA-based MB, with both cables connected; if the
legacy PS/2 cable is in use, it causes the USB interface internal to the keyboard
to claim itself dead.  A feature, not a bug.

Kris