amd64 new kernel kvm problem (triggers netstat, etc. weirdness)

[Gregory Steuck <greg_(_at_)_y2004_(_dot_)_nest_(_dot_)_cx>]

I ran into the same problem Jared reported in June:
http://www.monkey.org/openbsd/archive/misc/0406/msg01077.html

I happen to remember how it occurred. I booted a snapshot kernel and
then rebooted back into release kernel. dmesg.boot demonstrating the
sequence is at http://greg.nest.cx/dmesg.boot (scroll through wi0
chatter)

kvm_mkdb.testdb now mistakenly thinks that I am still running the
same kernel:

% sudo kvm_mkdb -v
Password:
kvm_mkdb: /var/db/kvm_bsd.db already up to date

Even though:

% uname -a
OpenBSD home.nest.cx 3.6 GENERIC#136 amd64
% strings /var/db/kvm_bsd.db| grep OpenBSD
_nfs_readrpcOpenBSD 3.6-current (GENERIC) #182: Fri Nov 19 02:21:11 MST 2004

Here is gdb output confirming my hypothesis:

Breakpoint 1, testdb (dbname=0x7f7ffffef610 "/var/db/kvm_bsd.db")
    at /usr/local/src/usr.sbin/kvm_mkdb/testdb.c:103
103             ret = memcmp(kversion, dbversion, dbversionlen) == 0;
(gdb) p kversion 
$1 = "OpenBSD 3.6-current (GENERIC) #182: Fri Nov 19 02:21:11 MST 2004\n    deraadt_(_at_)_amd64_(_dot_)_openbsd_(_dot_)_org:/usr/src/sys/arch/amd64/compile/GENERIC\n\000iM9Щr;\203\004$\004H\213|$\020Hc\004$L\215<е\000\000\000\000L\211ЧХ\225%\002\000H\205юH\211бt5L\211ХM\215|\027ПH+D$\020H\211T$\020L\215,"...
(gdb) p dbversion
$2 = "OpenBSD 3.6-current (GENERIC) #182: Fri Nov 19 02:21:11 MST 2004\n", '\0' <repeats 203 times>, "\004\000\000\000hCйJ\000\000\000\000y:<\a\000\000\000\000\006\000\000\000\000\000\000\000Cс\210D\000\000\000\000hОЧЪ\177\177\000\000\030ОЧЪ\177\177\000\0000.\200F\000\000\000\000\030ОЧЪ\177\177\000\000\000\000\000\000\000\000\000\000\001\000\000\000\036\001\000\000\000\000\000\000\000\000\000\000(@йJ", '\0' <repeats 20 times>, "hCйJ\000\000\000\000x\230\211D\000\000\000\000\b\000\000"...

Apparently kversion is not correct given uname output above.

Thanks
Greg