virtualbox-ose 4.0.8 fails

Matt sendtomatt at
Fri Jun 24 18:35:40 UTC 2011

On 06/24/11 10:52, Jung-uk Kim wrote:
> On Friday 24 June 2011 01:14 pm, Matt wrote:
>> It fails a couple ways actually, first on an isDVD in a disk system
>> request...commenting out the inq_(something, not in front of
>> machine with recent svn) parts of that code yields virtualbox
>> compiling, but failing during kmod compile due to the recent change
>> (without revision bump) from cpumask_t to cpuset_t.
>> It seems like recent CAM changes and CPU change are going to
>> require some changes to virtualbox in HostHardwareFreeBSD.c and
>> mp-r0drv.c at least. Even though OS revision was not bumped,
>> perhaps Makefile can switch on presence of cpuset userland utility?
>> Luckily I only csup'd a machine I don't really need Vbox on, so I'm
>> holding back all other machines until Vbox maintainers sort out the
>> issue.
> You should be able to build the kmod with this patch.
> Just drop this patch in ports/emulators/virtualbox-ose-kmod/files and
> rebuild.
> Please note the revision wasn't set right for the obvious reason,
> though.  Do we really need revision bump, BTW?  "Current" means "no
> seat belt" anyway. ;-)
> Cheers,
> Jung-uk Kim
Thanks for the patch. I had read a comment somewhere complaining about 
detecting cpuset_t or cpumask_t regarding osrevision. Not really an 
issue, because it can be tested for without a bump.

Who needs seatbelts anyway... :). CURRENT & a cold beer is good enough 
for my home systems. Certainly prevents boredom anyway.

The Virtualbox error (not kmod error) looked like it was using an 
undefined struct to determine drive types, which I assume has been removed.

If you &> the make output into a file and search for isDVD, you'll find 
that particular error, if still present. I just commented out the parts 
of the struct we don't have anymore, and it did compile...definitely 
could be dangerous, I haven't actually launched virtualbox with that 
"fix" could make for subtle or major problems. Your mileage & 
seatbelt may vary :)



More information about the freebsd-current mailing list