[vortex] [CFT] 3c59x.c for 2.2.18-pre

[Andrew Morton]

Koos van den Hout wrote:
> 
> Quoting Andrew Morton who wrote on Sat, Oct 21, 2000 at 07:19:07PM +1100:
> 
> > Does the interface function reliably when the duplex is not forced?
> 
> It works fine, half-duplex. (but I want full-duplex, that's why I got the
> switch.. I know there are issues with recognition of fullduplex on 100
> mbit)

Oh. Your earlier report showed that the NIC was in fact running
half-duplex.  See the vortex-diag output.

> > > 3c59x.c 19Oct00 Donald Becker and others
> > > http://www.scyld.com/network/vortex.html
> > > eth0: 3Com 3c595 Vortex 100baseTx at 0xd000,  00:a0:24:0b:2e:41, IRQ 5
> >                           ^^^^^^^^^
> >
> > hmm.. There seems to be a disagreement between the drivers and
> > `vortex-diag' about the human-readable name of this device.  Could you
> > please confirm that this is in fact a 10/100 NIC?
> 
> Doublechecked, it is. The transfer speed with ftp also suggests it:
> 
> 32670440 bytes received in 9.86 seconds, 3.16 MB/s.
> (report from ncftp in bytes/sec)

OK, that's 100bT, but pretty poor performance.  I find netperf from
netperf.org to be a good, repeatable FTP benchmark.

I'd suggest that you do some testing with netperf - I think you'll find
that full duplex doesn't gain you much at all.

> > You could try this patch, but it's a stab in the dark...
> 
> It looks (this is very subjective!) like I get *less* errors. I still see
> a few errors:
> 
> 3c59x.c 19Oct00 Donald Becker and others
> http://www.scyld.com/network/vortex.html
> eth0: 3Com 3c595 Vortex 100baseTx at 0xd000,  00:a0:24:0b:2e:41, IRQ 5
>   64K word-wide RAM 3:1 Rx:Tx split, autoselect/10baseT interface.
> eth0: Transmit error, Tx status register 90.
> ...
> 
> This is over a *lot* more transfers then the previous report I made!

That could have been dumb luck :(

Did you say that the interface stops working after the Transmit error?  
Or does it successfully reset and restart?  (It should).