Re: hme goes down with SENTFRAME,RXTERR

[Pierre Pronchery <khorben%defora.org@localhost>]

                Hey,

Miles Nordin wrote:
>>>>>> "pp" == Pierre Pronchery <khorben%defora.org@localhost> writes:
> 
>     pp>               Hello, I run occasionally into this error:
>     pp> hme0: status=200100<SENTFRAME,RXTERR> It happens with moderate
>     pp> traffic: playing music over NFS was apparently enough to
>     pp> trigger it.
> 
> sound like this:
> 
>  http://mail-index.netbsd.org/port-sparc64/2005/07/19/0008.html
> 
> follow the thread---it happened on hme also.

Ok, I believe I read most of the thread now. Here's what I have to add
so far:
- I don't use any tunneling protocol;
- I don't use altq or any firewall rule;
- I am on a vlan, and using it exclusively on this machine;
- some non-tagged packets are hitting the card from time to time;
- I also have IPv6 configured and working.
I'll try what Martin suggested next time I observe such a stall.

It is probably not related and not harming, but I do have noticed
something strange regarding network statistics with vlans:
- incoming packets are properly counted on the underlying interface and
the vlan interface,
- whereas outgoing packets are only counted on the vlan interface.

FWIW, I observe this behavior with the wmnd applet and in my own code
too (using the SIOCGIFDATA ioctl). But maybe this should go to tech-net@
instead?

HTH,
-- 
khorben