wm0 failed after mpt "recovered from command timeout"

[Edgar Fuß <ef%math.uni-bonn.de@localhost>]

I had an 8.1/amd64 server losing network connectivity after mpt reported
"recovered from command timeout"

There was also a find (operating on /) hanging in tstile and df / hanging
the same way.

According to syslog:

Jan 28 15:54:01 server /netbsd: sd1(mpt0:0:1:0): command timeout
Jan 28 15:54:01 server /netbsd: mpt0: recovered from command timeout
Jan 28 15:54:01 server /netbsd: nfs server XXX: not responding
Jan 28 15:54:01 server nslcd[12312]: [7dbb2e] XXX ldap_result() timed out
[...]
Jan 28 15:55:31 server nslcd[12312]: [d55673] XXX failed to bind to LDAP server XXX: Can't contact LDAP server: Connection timed out
[...]
Jan 28 15:57:52 server /netbsd: wm0: device timeout (lost interrupt)
[...]
Jan 28 16:03:52 server /netbsd: wm0: device timeout (lost interrupt)
[...]
Jan 28 16:06:14 server nslcd [...]
Jan 28 15:57:52 server /netbsd: wm0: device timeout (lost interrupt)

Note the strange date in the last line quoted. This is NOT a copy-paste typo,
but appears as is in a (non-corrupt) messages.0.gz.

Any hints?