Re: 3c59x driver 0.25 falls off network

Ove Ewerlid (ewerlid@syscon.uu.se)
Thu, 18 Jul 1996 16:49:53 +0200


Unless you've seen such reports yourself ...
I've had one report that the disabling of the "optimization"
causes problems. This problems seems to be a tricky bastard.

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Subject:=20
Re: 3c595 problems, 2.0.6, possible fix
Date:=20
17 Jul 1996 11:53:21 +0200
From:=20
ig25@fg70.rz.uni-karlsruhe.de (Thomas Koenig)
Reply-To:=20
Thomas.Koenig@ciw.uni-karlsruhe.de
Organization:=20
Universit=E4t Karlsruhe (TH), Germany=20
To:=20
submit-linux-dev-net@ratatosk.yggdrasil.com
Newsgroups:=20
linux.dev.net
References:=20
1

In linux.dev.net, Ove Ewerlid <ewerlid@syscon.uu.se> wrote:

>#if 0
> if (inw(ioaddr + TxFree) > skb->len) /* We actually have free r=
oom. */
> dev->tbusy =3D 0; /* Fake out the check=
below. */
> else
>#endif
>
>and things work much better!
>Your mileage may vary ...

I applied this to 2.0.7, and it didn't work too well:

eth0: 3Com 3c590 Vortex 10Mbps at 0x6000, 00:a0:24:57:ee:81, IRQ 10
Internal config register is 120010, transceivers 0xe138.
8K byte-wide RAM 1:1 Rx:Tx split, 10Mbs AUI interface.
3c59x.c:v0.25 5/17/96 becker@cesdis.gsfc.nasa.gov
[...]

eth0: vortex_open() InternalConfig 00120010.
eth0: vortex_open() irq 10 media status 0800.

and about four minutes later:

eth0: Infinite loop in interrupt, status e011. Disabling functions
(78ee).