On Wed, 13 Mar 1996 18:00:26 -0800 
 buhrow@cats.ucsc.edu (Brian Buhrow) wrote:

 > rumpleteazer# netstat -i
 > Name  Mtu   Network     Address              Ipkts Ierrs    Opkts Oerrs  Coll
 > ep0   1500  <Link>      0.a0.24.5a.c9.9b    313642 12171   119250     0    10
 > ep0   1500  128.114.129 rumpleteazer.UCSC   313642 12171   119250     0    10
 > lo0   32768 <Link>                             366     0      366     0     0
 > lo0   32768 your-net    LOCALHOST              366     0      366     0     0
 > sl0*  296   <Link>                               0     0        0     0     0
 > sl1*  296   <Link>                               0     0        0     0     0
 > ppp0* 1500  <Link>                               0     0        0     0     0
 > ppp1* 1500  <Link>                               0     0        0     0     0
 > tun0* 1500  <Link>                               0     0        0     0     0
 > tun1* 1500  <Link>                               0     0        0     0     0
 > rumpleteazer# exit
 > Script done on Wed Mar 13 17:53:44 1996
 > 
 > 	Can anyone explain why the input error rate is so high?
 > We do have this machine on a very busy subnet, but this error rate seems
 > like it would have a severe performance impact
 > on the ethernet performance of this machine.  Yet, the driver logs no
 > messages from the kernel.

It could be a number of things - incoming packets aren't serviced quick 
enough, maybe ... could be another hose on the wire spewing garbled 
packets...

Take a look at epread() ... it checks the status register.  If debugging 
is enabled on the interface, then error strings are printed out, 
otherwise not.  Note that printfs can be expensive, and sometimes only 
make the problem worse...

Another thing that might increase the error count is if epget() can't 
allocate an mbuf ... that appears to be it, really.

 > Is this error rate all right?  Is this just over-sensativity from the
 > driver?  Is it polling the card, coming up empty and incrementing the input
 > error count?

It's definitely not "polling the card and coming up empty".  Ethernet 
drivers pretty much have to be interrupt driven :-)

--------------------------------------------------------------------------
Jason R. Thorpe                                       thorpej@nas.nasa.gov
NASA Ames Research Center                               Home: 408.866.1912
NAS: M/S 258-6                                          Work: 415.604.0935
Moffett Field, CA 94035                                Pager: 415.428.6939