"Transmit timed out" with EtherExpress Pro100B

[Osma Ahvenlampi]

"Robert G. Brown" <rgb@phy.duke.edu> writes:
> Well, VLAN's are great if you can do it with your switch.  Isolate the
> Macs off by themselves where the can chatter away.  Spurious network
> traffic eats CPU in addition to bandwidth.  I think that some of the

My switches support it, but I'm not going to start going through that
hassle for just about any reason, with our people moving from desk to
desk quite often and taking their laptops with them. In any case,
VLANs would be impossible now, because this machine is supposed to 
become the primary file server (AppleShare and SMB). Pretty difficult
to isolate it from the Macs and keep that working..

> problems may come from deeper in the kernel, though, and not just in the
> card.  One reason I think this is that they are just (re)surfacing as a
> new generation of very fast systems like the 2300's appears.  With U2W
> SCSI controllers, fast ethernet controllers, and a heavy task mix you
> are pushing some kernel latencies, possibly past some point of failure.

Keep in mind this machine isn't really under any load at all yet. I
have it sitting by my desk, with most network services installed and
enabled, but it doesn't have the RAID connected yet, and no one but me is
really accessing at all. It went crazy without anyone touching it,
just by being on the network and seeing the multicast traffic between
the Macs and their current file server. It hasn't really been pushed
to any limit yet, and it's even connected to a 10Mbps switch port at
the moment.

It's remained quiet and behaved itself now that I patched the driver
to not use the hardware multicast filtering at all. I'll receive the
rest of the hardware next week and put it to a better stress test
with all the components configured, and that's when I'll see if the
patch makes the network interface too slow, or if it goes crazy again.

> Has anybody tried the 2.0.36pre stuff?  Alan has released a bunch of
> patches that are supposed to address some of the kernel bugs in 2.0.35;
> perhaps one of them "fixes" this...

I suppose I could.. I was going to wait for the official release
though, since I don't particularly think of constant kernel patching
and recompiling as "fun".

> That's a lot of room, if you think about it... especially if the problem
> is load dependent.  Do you have an SMP system?  Do you think that this
> could be an interrupt resolution issue?

I don't think the machine has had a loadavg over 0.1 at any point
during it's life here. It's a single-CPU system, and the only shared
interrupt in it is the one used by aic7xxx (both 7880 and 7890 are on
interrupt 11):

[root@pe2300 /root]# cat /proc/interrupts
 0:    8110722   timer
 1:       1114   keyboard
 2:          0   cascade
 3:     495090 + serial
 8:          1 + rtc
11:      72524   aic7xxx, aic7xxx
13:          1   math error
14:     348484   Intel EtherExpress Pro 10/100 Ethernet

-- 
You know better than to trust a strange computer.
Osma Ahvenlampi <oa@spray.fi>