Subject: Re: A good, cheap, wire-speed PCI-e NIC for NetBSD?
To: None <firstname.lastname@example.org>
From: Jonathan Stone <jonathan@Pescadero.dsg.stanford.edu>
Date: 11/04/2005 17:20:52
In message <20051104215709.GA22405@panix.com>, Thor Lancelot Simon writes:
>On Fri, Nov 04, 2005 at 11:43:01AM -0800, Jonathan Stone wrote:
>> * The S&K SK-9E22, which is yet another sk/skc combo, of which
>> I vaguely heard Thor commenting about compatibiliity issues
>> with various variants (Marvell? I forget).
>> SysKonnect provides binary-only FreeBSD-5 drivers, and a
>> Linux source driver from which TSO support could be
>> reverse-engineered into our sk/skc code.
>The newer Marvell chips are "Yukon 2/2XL". The general register layout
>and operation of the chip are similar to those of the original Yukon
>but they're different enough that a separate driver might make more
>sense. The checksum offload works totally differently, and, as you
>note, there's TSO support.
>The Linux driver supports every Yukon variant that's ever seen the light
>of day and some, AFAICT, that haven't. So it's a mess. Still, it is
>possible to read between the ifdefs and look at the header files and
>see what would be needed to support the newer parts. They certainly look
>nice... it helps that the Linux driver (the Marvell-supplied one, not the
>other one, which I think still doesn't support the Y2 chips) has separate
>source files containing functions for some of the newer parts.
Hmm. thanks, that's definitely food for thought.
I was hoping Jason (or someone else acquainted with stge(4) might
comment on the DGE-560T. ICplus.com.tw has a datasheet available, and
their FreeBSD-4 driver matches the same PCI vendor/device IDs as the
Sundance/Ala board. I was hoping against hope that (modulo TSO) the
DGE-560T might Just Work...