Re: Trying to make usb networking work

From: Andrew Zabolotny <zap_at_homelink.ru>
Date: Tue, 24 Feb 2004 19:07:37 +0300

On Tue, 24 Feb 2004 15:00:12 +0100
Michael Opdenacker <zumbi3_at_free.fr> wrote:

> # mount /proc
> # insmod
> /lib/modules/2.6.1-rmk0-hh0/kernel/drivers/usb/gadget/pxa2xx_udc.ko (I
> haven't installed modutils yet in my home made initrd)
> <6>pxa2xx_udc: version 7-Nov-2003
> pxa2xx_udc: version 7-Nov-2003
If you disable "Low-level kernel debugging" in kernel config, you won't
get these funny duplicate messages. This option is good in the
beginning, but as you have a working serial console there's no need to
use it anymore.

> # ifconfig usb0 inet 192.168.0.200 netmask 255.255.255.0 broadcast
> 192.168.0.255
> # route add 192.168.0.201 dev usb0
I think you don't have to add a explicit route since when a interface
comes up, a line is automatically added to the route table.

> Then I make ping tries on both sides. This is unsuccessful, but
> re-running ifconfig is interesting:

> There are more transmitted packets, but just because I couldn't run
> both ifconfig commands at the same time. It seems that all transmitted
> packets are received by the PC host.
To better see it you can use tcpdump or something similar. It would be
also very helpful to run tcpdump on iPAQ (get the ipk file and unpack it
- it's either a .tar.gz file or an .a file (unpack with ar x)).

Some thoughts: it is possible that some GPIO haven't been set up
correctly, and this prevents UDC from receiving packets. A possible
cause could be the H2200_GPIO_USB_PULL_UP_N signal, but if it is really
a inverted signal, it looks like it is initialized correctly. However,
if it is really a non-inverted signal, you can try to change
SET_H2200_GPIO_N(USB_PULL_UP, 0); to SET_H2200_GPIO_N (..., 1) and vice
versa - 1 to 0 in h2200.c, then see if this makes any difference.

--
Greetings,
   Andrew
Received on Tue Feb 24 2004 - 16:07:39 EST

This archive was generated by hypermail 2.2.0 : Mon Jul 25 2005 - 17:19:25 EDT