I am wondering about the current status of a driver for the NS83815 ethernet
chip.
>From searching Google, I know some sort of driver exists. In July, Adam J.
Richter ([email protected]) posted a 2.2.16 driver he obtained from Dave
Gotwisner at Wyse Technologies. And Tim Hockin mentioned that he was using
an NSC driver, but had made some minor modifications.
The only source I've seen is the one Mr. Richter posted.
(http://web.gnu.walfield.org/mail-archive/linux-kernel/2000-July/4234.html).
How well does this driver work? From Mr. Richter's email I gather that Alan
Cox gave some feedback and suggested improvements. This makes me worried
about using the "unimproved" version of the driver.
If anyone has improved code for the 2.2.x series I would greatly appreciate
it.
2.2.17 and 18 didn't include the driver. I also gather that Mr. Richter is
(or was) concentrating on a port to the 2.4 series, how is that coming
along?
For what it's worth, the chip seems to have detailed documentation at
http://www.national.com/pf/DP/DP83815.html.
Thanks for any help.
Torrey Hoffman
[email protected]
Torrey Hoffman wrote:
>
> I am wondering about the current status of a driver for the NS83815 ethernet
> chip.
>
> >From searching Google, I know some sort of driver exists. In July, Adam J.
> Richter ([email protected]) posted a 2.2.16 driver he obtained from Dave
> Gotwisner at Wyse Technologies. And Tim Hockin mentioned that he was using
> an NSC driver, but had made some minor modifications.
Use drivers/net/natsemi.c... it's in 2.4.x-test.
--
Jeff Garzik |
Building 1024 | These are not the J's you're lookin' for.
MandrakeSoft | It's an old Jedi mind trick.
> >From searching Google, I know some sort of driver exists. In July, Adam J.
> Richter ([email protected]) posted a 2.2.16 driver he obtained from Dave
> Gotwisner at Wyse Technologies. And Tim Hockin mentioned that he was using
> an NSC driver, but had made some minor modifications.
We're still using a heavily hacked version of the NSC driver on 2.2.x.
When we do 2.4.x, I'll examine the other driver more closely. I can send
you my hacks on the NSC version, if you need.
>From searching Google, I know some sort of driver exists. In July, Adam J.
>Richter ([email protected]) posted a 2.2.16 driver he obtained from Dave
>Gotwisner at Wyse Technologies. And Tim Hockin mentioned that he was using
>an NSC driver, but had made some minor modifications.
This one has a very murky past. Adam & I have both received the same
driver from within NatSemi, which turned out to be a rehash of the
driver originally written by Donald Becker (with his name and the GPL
license removed...). It should be noted that this wasn't "written" by
NatSemi themselves, but by a company called TeamF1.
Since then Donald's original driver (natsemi.c) has been folded into the
2.4 kernel, and I can happily report works well with 2.2 as well. It's
certainly a lot better than the dp83815.c program, which I found had
the tendency to hang the machine completely in high traffic (well, if
you count ftping from a host on the same 10Mb half duplex network as
high load ;).
john
Please, don't even consider using the NSC dp83815 driver.
I too received this NSC (TeamF1) driver, hacked on it, fed the modifications
to NSC and this TeamF1.
Some time October I received the latest and greatest version from NSC
and found no trace of my changes, but some "Shaun Savage" hacked on it
and ported the buggy 2.2 version to 2.4.
He found the tbusy flag gone so implemented this:
if (chip_out_of_descriptors) {
drop_packet();
stats->tx_dropped++;
}
netif_stop_queue(), what's that ???
I must admit that I did not look any further at this driver ...
Please try the driver in 2.4.