Vmware 3.2.1 under FreeBSD 5.1-RELEASE

[Sean Welch]

I'm having an odd issue I've not seen mentioned anywhere else and was
wondering if anyone else is seeing it?

I managed to get the program to run yesterday for the first time (Note:
not a good idea to try to compile and run vmware when you've been 
updating your source but left your binaries alone!!) and everything
seemed to be cool (with the exception of networking -- doesn't seem
to see /usr/local/etc/vmware/config for some reason).  I started it up
to make sure it ran and didn't have any issues.

The next time I actually started the virtual machine and was annoyed
by the above mentioned network issues (works fine in vmware2), but
otherwise all worked as expected.  Until I tried to shut my REAL machine
down.  I got a lot of panics and ended up "Not Very Happy" (TM).

I don't know how to get dumps and I don't know if it is necessary -- I'll
be glad to try anything with explicit directions.

Here's what I've been able to glean:  The trouble seems to stem from
the rtc kernel module.  If I don't start the virtual machine everything 
unloads successfully when I shut down (vmnet, if_tap, vmmon_up, rtc)
but if I *do* start up the virtual machine at some point vmmon_up 
refuses to unload (device busy) and attempting to unload rtc panics the
machine in a major way.

At boot I see this:

WARNING: driver "rtc" used unreserved major device number 202
WARNING: driver "vmmon" used unreserved major device number 200
/dev/vmmon: Module vmmon: registered with major=200 minor=0 tag=$Name: build-570+ $
/dev/vmmon: Module vmmon: initialized

At shutdown I don't have it logged anywhere so here I what I've been
able to scribble down as the messages start flying by:

WARNING: Driver mistake: destroy_dev 202/0

Then something about "Panic" and a message saying "don't do that!"
and  a series of messages alternating between starting with CBB0 and
CBB1 with identical messages following that containing something about
VCARD (I assume it is corrupted memory pages??)

I had some file system corruption before I realized that if I commented
out the explicit effort to unload rtc in /usr/local/etc/rc.d/rtc.sh the kernel
would manage to sync the disks before the panics started (presumably
because it then tries to force unload rtc as part of the final shutdown).  I
don't get problems on the next boot anymore with this setup, but I can't
say it makes me feel warm and fuzzy to see that stuff on shutdown (runs
for 15 seconds or more).

I know this isn't terribly precise in detail but maybe someone can help
diagnose this thing?

Here is uname -a:

FreeBSD NitroPhys.welchsmnet.net 5.1-RELEASE FreeBSD 5.1-RELEASE #0: Thu Jun 5 02:55:42 GMT 2003      root at wv1u.btc.adaptec.com:/usr/obj/usr/src/sys/GENERIC i386

Here is a list of relevant installed ports/packages:

vmware3-3.2.1-2237_1
rtc-2202.03.05.2_2
linux_base-7.1_5

X is the version that came with 5.1-RELEASE. Module aio is loaded and
linprocfs is enabled.  DRI is enabled on a 64MB Radeon Mobility 7500.
ACPI is enabled with a replaced aml.  The machine is a 1GHz Dell Inspiron
8000 with 320 MB of RAM (yeah, I upgraded the graphics card ;-)

                                                                              Sean