On Wed, Mar 12, 2014 at 12:58 PM, tuchalia <[email protected]> wrote:
> Hi all,
>
> I'm really interested in this possibility of porting the Capsicum framework

That's awesome !

> to OpenBSD. Should l try to port also the Casper daemon to OpenBSD,  or
> only work in the kernel implementation?

Capsicum is a huge project, and realistically, it's impossible to to
port it completely
and get a production grade version by working only 8-12 weeks. The
project clearly exceeds
the mandate of a GSoC, in my opinion.

You might start by planning the kernel implementation, similar to what
Joris did for DragonflyBSD.
(http://leaf.dragonflybsd.org/mailarchive/kernel/2013-04/msg00025.html).

With a kernel implementation, we can do quite a lot of things, like
what happened for OpenSSH 6.5.
OpenSSH 6.5 has capsicum support on FreeBSD, and it doesn't need
casper to sandbox the pre-auth code.

matthew had already started working on some bits.
(https://lists.cam.ac.uk/pipermail/cl-capsicum-discuss/2011-July/msg00002.html)

Your best bet is to speak to damien miller and work out a sensible
plan. If you're really interested
motivated in doing a good & complete work, you should be ready to
spend time after gsoc getting feedback
on the diffs, and improve them.

Personally, I would like to see a working kernel implementation, and
getting at least sshd in base
capsicumised :-) I can help you with the last part, as I'm familiar
with the code. Please talk to damien & matthew
as they might have a different roadmap or radically different ideas in mind.


>
> I've used Capsicum during the last summer, but I only worked with the
> syscall API, that is, no Casper (something that can be fixed quickly).

Very good !

However, I would advise caution here, as there has been at least one
bug that has been reported
with the changes happening in FreeBSD.

(http://unbound.net/pipermail/unbound-users/2014-February/003169.html).
So you have to be prepared
to fix any potential fallout if capsicum port is to be merged into OpenBSD :-)


>
> I know what I should I do with the Capsicum side (in the matter of getting
> some info to have a strong proposal) but I'm not sure about where to look
> when it comes to the OpenBSD side. Should I take a look at the process data
> structure, and how all this is implemented in the kernel? Should I take a
> look somewhere else?

I would advise you to look into Joris's posts to DragonflyBSD lists,
and see what difficulties
he faced when he ported capsicum to dflybsd. Also, please take into
account that OpenBSD kernel
is different from FreeBSD kernel, and that will probably involve a lot
of rewriting.

You are lucky as OpenBSD has a rock solid -current tree. Ideally, you
would have a dedicated physical
box such as a 2nd laptop or a development machine to work on that IMHO.

Read the OpenBSD FAQ for running -current.
http://www.openbsd.org/faq/current.html

>
> Also,  do we have any IRC channel to discuss al this?

I'll send it to you in a personal mail.

>
> Many thanks,

Good luck with your proposal !


-- 
This message is strictly personal and the opinions expressed do not
represent those of my employers, either past or present.

Reply via email to