[eepro100] Problems at 100fd with eepro100

Vernon McPherron vernon@typeset.net
Wed, 24 May 2000 18:14:53 -0600


On Tue, 23 May 2000 14:38:09 Donald Becker wrote:
> On Tue, 23 May 2000, Vernon McPherron wrote:
> 
> > > A duplex mismatch won't prevent ping from working, but it will make 
> > > stream traffic (e.g. file transfers) very slow. 
> > >
> > > A negotiation mismatch might prevent communication, but will only happen
> > > if you limit the advertised media types. 
> > 
> > Hmm...  What about nearly all packets being dropped OR seriously delayed?
> > 
> > > What driver version? The one from scyld.com? 
> > 
> > Yes.  Let me check the version.
> > "eepro100.c:v1.10a 4/15/00 Donald Becker <becker@scyld.com>\n";
> > That's the one.
> 
> That version has interrupt-blocked recovery, so it pass a few packets (with
> horrible delays) even with a broken interrupt line.
> 
> Check the interrupt count in /proc/net/dev.

Since I've got it working looks ok.  Here is what I get on my SMP machine with
version: 
eepro100.c:v1.09j-t 9/29/99 Donald Becker
http://cesdis.gsfc.nasa.gov/linux/drivers/eepro100.html
eepro100.c: $Revision: 1.32 $ 2000/03/30 Modified by Andrey V. Savochkin
<saw@saw.sw.com.sg> and others

in /proc/net/dev looks ok.
Inter-|   Receive                                                |  Transmit
 face |bytes    packets errs drop fifo frame compressed multicast|bytes   
packets errs drop fifo colls carrier compressed
    lo:       0       0    0    0    0     0          0         0        0     
 0    0    0    0     0       0          0
 tunl0:       0       0    0    0    0     0          0         0        0     
 0    0    0    0     0       0          0
  gre0:       0       0    0    0    0     0          0         0        0     
 0    0    0    0     0       0          0
  eth0: 3971750    5254    0    0    0     0          0         0   400055   
5269    0    0    0     0       0          0

or at least I assume that is what I should see.  
Now would I have the same problems with the other driver (the one that was
causing me problems) if I didn't use APIC?  


> > Well I started on 2.3.99-pre6, and then when I had problems, I patched the
> > kernel to 2.3.99-pre8.  It DID compile, but I got the same errors that I
> > was getting on my SMP machine.  Seriously delayed packets, made it appear
> > that nothing was getting across the link.  
> 
> Due to political reasons, my code isn't designed for the frequently changing
> 2.3.99 kernels.  But the problem is likely common between the two drivers.

Yeah.  Are there big differences between the two drivers?