Status of itohy-usb1 branch and/or device support in the axe driver?

[Erik Bertelsen <bertelsen.erik%gmail.com@localhost>]

Hi all !

Can anyone enlighten me on the status of the itohy-usb1 branch?

According to BRANCHES, this branch contains a "USB stack overhaul,
mostly DMA related", but
it also contains support for additional devices.

It appears that the original developer's work was discontinued in May
2008 except for a set of
commits on 12 and 13 Nov 2009 by Masao Uebayashi with commit messages
talking about reducing
diffs with HEAD.

My real interest is not in the branch as such, but is in the
additional devices supported by the axe driver, cf.


------------------------------------------------------
Module Name:    src
Committed By:   itohy
Date:           Mon Jun 25 09:21:28 UTC 2007

Modified Files:
       src/sys/dev/usb [itohy-usb1]: if_axe.c

Log Message:
Add support of AX88178 and AX88772 (based on OpenBSD).
(beware new Tx/Rx code, totally untested)

New devices from FreeBSD/OpenBSD:
[..]
- D-Link DUB-E100 rev B1
[..]

To generate a diff of this commit:
cvs rdiff -r1.17.10.5 -r1.17.10.6 src/sys/dev/usb/if_axe.c
------------------------------------------------------

The D-link device mentioned above is just one example needed by
another poster, I have both a Belkin F5D5055
and an Apple USB LAN adapter that seem to be (or could be) covered by
the enhanced axe driver.

(What I'm aiming at is to add a second LAN interface to a Mac Mini and
use it as the gateway between
my home network and the rest of the world. As I also need to run DHCP
and a local nameserver on the
machine implicates that I need/prefer to run NetBSD rather than Mac OS X on it).

I have tried to pull out the relevant changes on the itohy-usb1 branch
to support the enhanced axe driver and
applied it to NetBSD -current, but as lots of other changes have been
made to the USB stack and although
the device configures and the MAC address is printed on the console, I
cannot get any traffic to pass through
the device (and quickly gets a panic).

From my point of view, a good way to proceed is if a developer knowing
the USB code pulls the enhanced axe
driver into -current sources (with or without the rest of the branch).
Then I would be a happy tester.

If it helps, I can of course provide the diff of my manual merging of
the original update, but as I have done this
without really understanding the code that I patched, this may have
limited usefullness.

kind regards
Erik