FreeBSD on Eeepc 1000h

[Pyun YongHyeon]

On Sun, Nov 02, 2008 at 09:33:21AM -0800, Jeremy Chadwick wrote:
 > On Sun, Nov 02, 2008 at 01:56:27PM +0100, Sven Aluoor wrote:
 > > On Sun, 2 Nov 2008 08:42:53 +0100
 > > Matthias Apitz <guru at unixarea.de> wrote:
 > > > El d?a Sunday, November 02, 2008 a las 02:48:13AM +0300, Boris
 > > > Samorodov escribi?:
 > > > > Seems that you may be interested at:
 > > > > http://docs.freebsd.org/cgi/getmsg.cgi?fetch=343551+0+current/freebsd-current
 > > 
 > > Hi Boris
 > > 
 > > This link reference to an empty document
 > > 
 > > > Known issues and FreeBSD work for the eeePC series are described here:
 > > > http://wiki.freebsd.org/AsusEee
 > > > 
 > > > HIH
 > > 
 > > Hi Matthias
 > > 
 > > Thanks for the wikipage. I quote from there:
 > > 
 > > Atheros L1 FastEthernet
 > > This one is second generation of L1 controller which is supposed to be
 > > heavily modified by Atheros after acquiring Attansic. This controller
 > > uses different bits/meaning in register layouts and seems to use
 > > different descriptor formats. This is not supported by the age(4)
 > > driver. 
 > > 
 > > In clear text: at the moment there is no driver available? Is this
 > > correct?

That's not true anymore. There *IS* a driver, ale(4). See below.

 > 
 > This might be applicable:
 > 
 > I recently send Yong-Hyeon PYUN an Asus P5Q SE motherboard, solely to
 > develop a driver for the Attansic L1 chip that's on the motherboard.
 > After a few weeks he returned to me a driver called ate(4), which
 > supports the Atheros AR8121/AR8113/AR8114 NICs.  The description I
 > got, as well as the driver:
 > 
 > "Ok, here is a working driver, ate(4), for Atheros AR8121/AR8113
 > and AR8114. It passed minimal functional requirement of network
 > driver of FreeBSD but it still needs more test and cleanups.
 > I'm not sure you have hardwares to test ate(4) but if you have
 > it give it a try. Note it seems that the hardware, at least AR8121,
 > has checksum offload bug so I disabled Tx side checksum offload.
 > TSO seems to work though.
 > 
 > The Tx performance is 920Mbps or higher but Rx performance
 > is 850~870 Mbps which is lower than Tx. It also requires a lot of
 > CPU cycles to push the hardware to limit but I think it comes from
 > hardware limitation not from ate(4) itself."
 > 
 > There is absolutely no guarantee that this is the same exact NIC used on
 > the Asus Eee, but there's a chance.
 > 

One of FreeBSD developer also confirmed that ale(4) works on his
Eeepc 1000H. :-)

 > This driver isn't publicly available yet, and I do not feel comfortable
 > sending it to people here without Yong-Hyeon's permission.  I've CC'd
 > him so he can put it up on freebsd.org somewhere and let folks test it,
 > assuming it will/does work with the Asus Eee.
 > 

I've sent CFT to CURRENT ML and waiting for feedbacks from users.
http://marc.info/?t=122533988800003&r=1&w=2
(For unknown reason archiving service of CURRENT ML does'nt seem to
work. :-( )
I'm not subscribed to questions@ so please CC to me.

-- 
Regards,
Pyun YongHyeon