Debugging: General Approaches

GoogleGroups
1

Hi:
  I'm having difficulty getting my usb ethernet adapter working. I
want to turn this process of fixing it into an general exercise I can
teach my students.

Here's the background for my problem:
I have an older image that boots up fine and eth0 works fine (as does
the usb keyboard and mouse).
Here's what 'uname -a' says about it:
Linux beagleboard 2.6.27-omap1 #2 Sun Jan 11 11:23:31 CET 2009 armv7l
unknown unknown GNU/Linux

I have a newer image where eth0 doesn't work (but the keyboard and
mouse do). Same hardware as before. uname -a says:
Linux beagleboard 2.6.28-omap1 #1 Fri Mar 20 01:11:29 EDT 2009 armv7l
GNU/Linux

So the question is: how to I track down this problem? Here's what I
did:

1. Run ifconfig
a. Old Image
# ifconfig
eth0 Link encap:Ethernet HWaddr 00:60:6E:00:03:16
          inet addr:137.112.41.132 Bcast:137.112.41.255 Mask:
255.255.255.0
          inet6 addr: fe80::260:6eff:fe00:316/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
          RX packets:1182 errors:0 dropped:0 overruns:0 frame:0
          TX packets:826 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:78470 (76.6 KiB) TX bytes:68979 (67.3 KiB)

lo Link encap:Local Loopback
          inet addr:127.0.0.1 Mask:255.0.0.0
          inet6 addr: ::1/128 Scope:Host
          UP LOOPBACK RUNNING MTU:16436 Metric:1
          RX packets:12 errors:0 dropped:0 overruns:0 frame:0
          TX packets:12 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:600 (600.0 b) TX bytes:600 (600.0 b)

b New Image
# ifconfig
lo Link encap:Local Loopback
          inet addr:127.0.0.1 Mask:255.0.0.0
          UP LOOPBACK RUNNING MTU:16436 Metric:1
          RX packets:43 errors:0 dropped:0 overruns:0 frame:0
          TX packets:43 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:3184 (3.1 KiB) TX bytes:3184 (3.1 KiB)

usb0 Link encap:Ethernet HWaddr 6A:AC:00:2A:D4:6D
          UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:0 (0.0 b) TX bytes:0 (0.0 b)

Hmmm.... So it looks like eth0 never got set up.

2. Check dmesg.
a. Old Image
eth0: register 'dm9601' at usb-musb_hdrc-1.4, Davicom DM9601 USB
Ethernet, 00:60:6e:00:03:16
usbcore: registered new interface driver dm9601

b. New Image
usb 2-1.4: new full speed USB device using musb_hdrc and address 5
usb 2-1.4: device v0a46 p9601 is not supported
usb 2-1.4: configuration #1 chosen from 1 choice

Hmmm.... I'm guessing the driver didn't get installed, but why?

3. Search Discussions

I found some a discussion that suggests the newer kernel may not work
(http://groups.google.com/group/beagleboard/browse_thread/thread/
b4e1af42a8ad353f/850e76675db42165?lnk=gst&q=eth0#850e76675db42165),
but I don't know where to go from here.

4. Post a question.
And here we are. Where do I go from here?

BIG Question 1: Is this a reasonable way to approach a problem? Are
there other steps I should add? Is there more (or less) information
that should be put in the original post?

BIG Question 2: How do I get eth0 working on the new image?

Thanks...

--Mark

2

Try doing ‘ifconfig usb0 down’. If this works, the I’d suspect a routing tables issue. I don’t know how to configure the routing tables in Linux, but I suspect there is plenty of literature on that.

You may be correct that the device driver is not installed, but I think you may be reading that into the error message. Do an ‘lsusb’ and ‘lsmod’ to check. There are probably some other better ways to see if the driver loaded.

3

Mark,

...well, you didn't say where you got your kernel source tree - that
would help...

The USB bus is populated by devices which may be identified by Vendor
ID and Product ID.

Your 'dmesg' outputs show the device to be a Davicom DM9601.

The first kernel has found this device and loaded a driver.

The second kernel states that the device Vendor: 0a46, Product: 9601
is not supported.

This could be as simple as the device is not included in your .config

A quick look-see in linux-2.6/drivers/net/usb shows a dm9601.c file.

The same copy of linux has an omap3_beagle_defconfig but that file
doesn't include the DM9601 driver by default.

'make <blah, blah, blah> menuconfig' shows out of the box, the kernel
I just downloaded doesn't include the DM9601 driver even as a loadable
module.

Look in
'Device Drivers ---> Network device support ---> USB Network Adapters --->'

Once I select the DM9601 adapter and Exit out, my .config includes this line:
CONFIG_USB_NET_DM9601=y

...it wasn't there before!

- dan

4

Dan:
  I think you are on to the solution. Unfortunately I'm too new at
this to take the next step. Two questions:

1. In the short term, how do I copy the driver to the correct
location on the SD card and configure the present system to use the
driver?

2. In the long term, how do I tell bitbake to include the correct
driver when building a new image?

Thanks again for all the help...

--Mark

5

Mark,

Well, if you are using 'bitbake' you are using Angstrom or some other
OpenEmbedded variant.

I find bitbake to be archaic. It is going to take something like:

# bitbake <your_target_kernel> -c menuconfig
# bitbake <your_target_kernel> -f -c compile
# bitbake <your_target_kernel>
# bitbake <your_usual_recipe>

Angstrom and OE in general are amazing techinology - as long as you
stick with the stock recipes and don't go changing things in
individual packages...

Just my humble opinion...
- dan