2.4.17[16] USB problem

2.4.17[16] USB problem

[Nils Faerber]

Hi all!
I recently see a strange USB problem which I find in 2.4.17 and 2.4.16 and
maybe even earlier.
What happens is that a special device attached via a USB HUB is not
detected anymore, specifically usb.c claims that it cannot set the new
address.
The strange thing is:
- It worked with that device (Brainboxes Bluetooth USB dongle) with
earlier kernels.
- The device works when directly connected (and not via extra HUB).
- hub.c tells me that it HAS successfully assigned a new address to that
device!
- And above all, only that special device is affected. All other devices I
have to test (USB mouse and USB webcam) work perfectly.

The error messages also show a strange behaviour of the Linux USB system:
Why does hub.c set the address and then usb.c tries the same again?

I am sure not I am not suffering the earlier mentioned "device not
accepting new address" problem due to interrupt routing problems (since
all other devices work and the machine I am using is a UP notebook).
And I am also sure that this device used to work connected to the HUB with
earlier USB releases.

So please if anyone has new insights on this problem please CC me since I
am not subscribed to lkml...
Many thanks in advance!
CU
  nils faerber

-- 
kernel concepts          Tel: +49-271-771091-12
Dreisbachstr. 24         Fax: +49-271-771091-19
D-57250 Netphen          D1 : +49-170-2729106
--

Re: 2.4.17[16] USB problem

[Greg KH]

On Wed, Feb 13, 2002 at 07:16:51PM +0100, Nils Faerber wrote:
> Hi all!
> I recently see a strange USB problem which I find in 2.4.17 and 2.4.16 and
> maybe even earlier.
> What happens is that a special device attached via a USB HUB is not
> detected anymore, specifically usb.c claims that it cannot set the new
> address.
> The strange thing is:
> - It worked with that device (Brainboxes Bluetooth USB dongle) with
> earlier kernels.

This device is flawed.  I have had _lots_ of reports of this problem
with these devices.  Talk to Brainbox and ask if their devices have
passed the USB certification tests (odds are it hasn't :)

thanks,

greg k-h

Re: 2.4.17[16] USB problem

[Martin Diehl]

On Wed, 13 Feb 2002, Nils Faerber wrote:

> What happens is that a special device attached via a USB HUB is not
> detected anymore, specifically usb.c claims that it cannot set the new
> address.
> The strange thing is:
> - It worked with that device (Brainboxes Bluetooth USB dongle) with
> earlier kernels.

Is this a buspowered device and if so, how much power does this drain from
the USB? I'm asking because I've just solved (kind of) a similar issue
with a 400mA device when I introduced the missing
connect-detect-to-port-reset debounce delay.

> The error messages also show a strange behaviour of the Linux USB system:
> Why does hub.c set the address and then usb.c tries the same again?

hub.c selects the address and calls usb_new_device from usb.c, which sets
(applies) the new address.

Martin