Date: Wed, 28 May 2014 22:45:39 -0700 From: Adrian Chadd <adrian@freebsd.org> To: Marcel Moolenaar <marcel@xcllnt.net> Cc: Anuranjan Shukla <anshukla@juniper.net>, Gleb Smirnoff <glebius@freebsd.org>, "freebsd-arch@FreeBSD.org Arch" <freebsd-arch@freebsd.org> Subject: Re: Roadmap for ifnet(9) for FreeBSD 11 Message-ID: <CAJ-VmonPP3Sbc%2B1GuUxgjQhwTEaHPHMO72V6JVuq2CMpiMutbA@mail.gmail.com> In-Reply-To: <FA49AADC-A853-49F0-B80C-F0FBA1219177@xcllnt.net> References: <FA49AADC-A853-49F0-B80C-F0FBA1219177@xcllnt.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi! On 28 May 2014 09:34, Marcel Moolenaar <marcel@xcllnt.net> wrote: > All, > Since Juniper already has a working patch that achieves the > objective of making ifnet an opaque type by replacing ifnet > pointer dereferences with function calls in an almost mechanical > fashion, we propose to merge that first. > The patch doesn't break old-style access to struct ifnet, which > means that unconverted and converted drivers coexist. This gives > us time to convert drivers. This also gives Juniper an important > rendezvous point between their and our repos. Hi! Converting all the ifnet dereferencing and converting it to accessor functions/macros seems like a no-brainer pass. You can do this without necessarily converting them to void *. That can be a separate pass. i wonder about the immediate benefits of making it fully opaque versus "mostly opaque-ready for you." Would we get less formalish checking coverage with static analysis tools and the compiler? -a
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAJ-VmonPP3Sbc%2B1GuUxgjQhwTEaHPHMO72V6JVuq2CMpiMutbA>