Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 18 Jun 2000 22:44:13 -0600
From:      Warner Losh <imp@village.org>
To:        Dave Preece <dave.preece@kbgroup.co.nz>
Cc:        freebsd-hackers@FreeBSD.ORG
Subject:   Re: Quickie: C++ statically linked into kernel? 
Message-ID:  <200006190444.WAA53529@harmony.village.org>
In-Reply-To: Your message of "Mon, 19 Jun 2000 16:26:45 %2B1200." <67B808B0DD93D211ABEE0000B498356B02BCC0@internet.kbgroup.co.nz> 
References:  <67B808B0DD93D211ABEE0000B498356B02BCC0@internet.kbgroup.co.nz>  

next in thread | previous in thread | raw e-mail | index | archive | help
In message <67B808B0DD93D211ABEE0000B498356B02BCC0@internet.kbgroup.co.nz> Dave Preece writes:
: I'm writing some C++ code that currently uses divert sockets off the
: firewall and for performance reasons moving the code kernel mode is looking
: like a (long term) good idea.
: 
: The question is: Should I bite the bullet and start writing in pure C now
: (and therefore save pain later)? Can I just provide an API using extern "C"
: and link it in (with some calling functions in ipfw.c, of course)? Would an
: lkm be the best way to go?

I've run C++ code in the kernel.  You couldn't easily use: templates,
exceptions, global ctors, and sometimes you had to be careful with
automatic instantiation of things.  I implemented only new and delete
and was able to get sample code to run in the kernel.  I punted at
that point due to the pain in actually knowing if these things were
being used or not.

Warner


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200006190444.WAA53529>