svn commit: r253255 - head/sys/netgraph/bluetooth/drivers/ubt

[Craig Rodrigues]

On Sat, Jul 13, 2013 at 11:33 AM, Mikolaj Golub <trociny at freebsd.org> wrote:
>
> Currently I don't have any pluggable device so my testing would not be
> very useful. I tried Marko's patch to test kldload/kldunload change,
> which is not in your patch.


OK, just to double check, based on reading your post:
http://lists.freebsd.org/pipermail/svn-src-all/2013-July/071848.html ,
with Marko's kldload/kldunload change, you encountered kernel panics in PF
and IPFW?
I also encountered kernel panics in PF with that patch.
Maybe we can clean up those modules in a separate pass,
and apply the kldload/kldunload fix some time in the future.


>
>
> Concerning your combined patch, as I said previously I don't like very
> much that dance with is_default_vnet. What is a purpose of it? As for
> me, it only makes the code less readable. I would just set curvnet to
> vnet0, as it is in Marko's patch.
>

Fine by me.  I have re-rolled my combined patch to remove the
is_default_vnet check
and go with Marko's code to set curvnet to vnet0:

(1)  Combine Adrian and Marko's patches, but omit Marko's changes to
kldload/kldunload:

http://people.freebsd.org/~rodrigc/vnet-patch-adrian+marko-p2.txt

(2)  Revert my Bluetooth patch:

http://people.freebsd.org/~rodrigc/patch-revert-GRN-253255.txt

--
Craig