Hi,

I have recently experienced a "vnode: table is full - increase 
kern.maxvnodes or NVNODE", "locking against myself" panic with a 
dual-opteron machine and 2.0_STABLE amd64 kernel. Similar problem was 
reported here: 
http://mail-index.netbsd.org/current-users/2004/05/28/0004.html and it 
seems to be resolved with kern/29670 (I don't know, but it seems to me 
to be the same problem).

It happened during 'cvs update' on a null-mounted fs (/usr) :
# mount
/dev/raid0a on / type ffs (noatime, local)
/dev/raid4b on /stor1 type ffs (noatime, local)
/stor1/var on /var type null (local)
/stor1/usr on /usr type null (local)
mfs:633 on /tmp type mfs (synchronous, noexec, nosuid, nodev, noatime, 
local)
kernfs on /kern type kernfs (local)
#

(The directories /var and /usr are null-mounted because build.sh script 
doesn't play well with symlinks.)

I wonder wether the suggested patch (kern/29670) will be pulled up to 
2.x or what state this is in.

Thanks,

r.