Index | Thread | Search

From:
Landry Breuil <landry@openbsd.org>
Subject:
Re: two ASIX devices: DLINK DUB-1312 & TPLINK UE-306
To:
tech@openbsd.org
Date:
Thu, 4 Jan 2024 08:25:00 +0100

Download raw body.

Thread
Le Thu, Jan 04, 2024 at 10:09:57AM +0800, Kevin Lo a écrit :
> On Tue, Dec 19, 2023 at 05:32:48PM +0800, Kevin Lo wrote:
> > 
> > On Sat, Dec 16, 2023 at 10:34:46AM +0100, Landry Breuil wrote:
> > > hi
> > 
> > Hi Landry,
> > 
> > > wanting to replace an axen that constantly spouts 'axen0: checksum err
> > > (pkt#1)' i've bought two 'cheap' usb-eth devices, supposedly also having
> > > AX88179:
> > > 
> > > -DLINK DUB-1312 is supported by axen(4) according to the manpage, but
> > > plugging the device it attaches as cdce, since
> > > https://marc.info/?l=openbsd-tech&m=168233597211047&w=2:
> > > 
> > > cdce0 at uhub0 port 13 "ASIX AX88179A" rev 3.20/2.00 addr 2
> > > cdce0: address 64:29:43:6d:04:94
> > > 
> > > on the device, the P/N is EUB1312...C1G, i guess a new revision.
> > > the complete lsusb output for both devices is attached.
> > > 
> > > tcpbench isnt so bad:
> > > 150610824 bytes sent over 12.985 seconds
> > > bandwidth min/avg/max/std-dev = 92.498/92.804/93.182/0.161 Mbps
> > > 
> > > according to kevlo,
> > > https://github.com/freebsd/freebsd-src/commit/70fbcd451b68b7f6038d8a602cd8d5e1bb890f1d
> > 
> > This is a quick fix for AX88179A. The diff below skips dummy packet headers,
> > which fixes 'axen0: invalid buffer(pkt#1), continue'.
> > 
> > While here, update man page to reflect ax88179a support.
> > 
> > Tested with:
> > axen0 at uhub0 port 15 configuration 1 interface 0 "ASIX AX88179A" rev 3.20/2.00 addr 4
> > axen0: AX88179A, address f8:e4:3b:xx:xx:xx
> > ukphy0 at axen0 phy 3: Generic IEEE 802.3u media interface, rev. 1: OUI 0x00070b, model 0x0006
> > 
> > axen1 at uhub0 port 16 configuration 1 interface 0 "ASIX Elec. Corp. AX88179" rev 3.00/1.00 addr 5
> > axen1: AX88179, address 00:0e:c6:xx:xx:xx
> > rgephy1 at axen1 phy 3: RTL8169S/8110S/8211 PHY, rev. 5
> > 
> > ok?
> 
> Anyone willing to ok?

not a kernel driver hacker, but this makes the device i have work with
axen instead of cdce so definitely an improvement - ok