-----BEGIN PGP SIGNED MESSAGE-----
>>>>> "Andi" == Andi Kleen <ak@xxxxxxx> writes:
Andi> On Fri, Feb 08, 2002 at 03:59:37PM -0500, Michael Richardson wrote:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> >>>>> "Andi" == Andi Kleen <ak@xxxxxxx> writes:
Andi> On Thu, Feb 07, 2002 at 05:59:45PM -0500, Michael Richardson wrote:
>> >> We think that /proc/net/rt_cache might have info, but we don't >>
>> see it. Is there another place that we should look?
Andi> /proc/net/rt_cache has a MTU column. You can also get it via ip
Andi> route get <ip address>
Andi> route --cache unfortunately doesn't show it.
>> Do they expire very quickly? Can we extend it? Is there some debugging
>> we can enable to log the MTU changes? Does it get deleted when the TCP
>> connection gets closed?
Andi> Depends on your workload. Normally not.
Is the timeout configurable somewhere?
What kind of work would affect our workload.
Andi> Just add a few printks.
I was considering heading in that direction. Would some printk's and a
patch to enable them via some sysctl be welcome?
Adding printk's helps for debugging this, but it does not help us do
regression tests of non-hacked kernels :-)
We use User-Mode-Linux builds such that we can easily capture console
output and control ethernet I/O. Having to apply additional patches to the
kernel kind of defeats this.
] ON HUMILITY: to err is human. To moo, bovine. | firewalls [
] Michael Richardson, Sandelman Software Works, Ottawa, ON |net architect[
] mcr@xxxxxxxxxxxxxxxxxxxxxx http://www.sandelman.ottawa.on.ca/ |device driver[
] panic("Just another NetBSD/notebook using, kernel hacking, security guy"); [
-----BEGIN PGP SIGNATURE-----
Comment: Finger me for keys
-----END PGP SIGNATURE-----