> [ On Thu, December 17, 1998 at 14:47:40 (-0700), Tim Rightnour wrote: ]
> > Subject: Re: le0: overflow
> >
> > That would explain his lossage with NFS, however my lossage is with X, and 
I
> > thought X used tcp.  If X is udp as well, then I suppose that answers the
> > question completely.
> 
> Hmm....  X11 does normally use TCP (listening on port 6000).
> 
> Are you seeing the problem only with remote X11 clients, or only with
> local ones, or both?

i only use remote X11. no local clients cause the machine is simply too slow
for running x11 apps local.

> I'm sitting here in front of a diskless Sparc-1 with 16MB RAM, and I've
> never seen an overflow error from the le driver with 1.3.1 or 1.3.2, not
> even when the transceiver falls off (is knocked off by the cats).

thats very strange .. perhaps the internal le is broken?? - this machine is
quite old and was used heavily. (btw: cats rule ..)

> I've even gone back to running my NFS mounts over UDP so that the system
> doesn't instantly crash when an NFS server reboots too.

NFS over TCP doesnt make any difference. same errors. .. it just seems to be
that the machine gets a bit slower in response (which is logical).

> Mind you I've only got a bwtwo frame buffer, so this CPU doesn't have to
> do much work, even for X11.  I should probably even run rc5des on it!  ;-)

hehe.. yeah. I bet .. i have on my SLC 15000keys/sec which isnt that bad for
the age of the box. .. in this SS1 is a bwthree. Perhaps I should try to boot
up Xmono and look if it gets better.
 
> It would appear that the local interface "short circuit" routes to the
> loopback interface actually work in 1.3.x -- I always use a fully
> qualified domain name as my DISPLAY, and yet there's lots of traffic on
> my loopback interface (this machine is basically only an X Terminal):
> 
> 20:15 [5] $ uptime
>  8:19PM  up 20 days, 19:32, 2 users, load averages: 0.71, 0.46, 0.34
> 20:16 [6] $ netstat -in
> Name  Mtu   Network       Address              Ipkts Ierrs    Opkts Oerrs  Co
ll
> le0   1500  <Link>        08:00:20:08:34:2f 13232340    13 13164008   315 318
13
> le0   1500  204.92.254    204.92.254.3      13232340    13 13164008   315 318
13
> le0   1500  204.29.161.16 204.29.161.162    13241876    13 13174012   315 318
30
> lo0   32976 <Link>                           1630328     0  1630328     0    
 0
> lo0   32976 127           127.0.0.1          1630328     0  1630328     0    
 0
> [[....]]
> 
> I can't believe netstat still truncates data from columns, even with -n!!!
> 
> So many things still to fix....

gives on my machine:

Name  Mtu   Network       Address              Ipkts Ierrs    Opkts Oerrs  Coll
le0   1500  <Link>        08:00:20:07:b0:d1    65558    18    44053     1   325
le0   1500  194.152.178   194.152.178.34       65558    18    44053     1   325
lo0   32976 <Link>                                70     0       70     0     0
lo0   32976 127           127.0.0.1               70     0       70     0     
0

I tried the "comment-out-this-annoying-kernel-msg"-method which worked quite 
well but it sucks.

later,
chrisjm
-- 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%% PGP-Key-ID: 0x975B2F19 - CJM17(-RIPE) - http://www.enemy.org %%%
%%%% Your mouse has moved. Windows NT must be restarted for the %%%%
%%%% change to take effect. Reboot now?  [ OK ]                 %%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%