From mboxrd@z Thu Jan 1 00:00:00 1970 Content-Type: multipart/mixed; boundary="===============7383869164302353211==" MIME-Version: 1.0 From: Denis Kenzior Subject: Re: [PATCH v2 08/11] dbus: Fix the kdbus message encoding Date: Tue, 12 Apr 2016 19:14:26 -0500 Message-ID: <570D8F62.1080106@gmail.com> In-Reply-To: List-Id: To: ell@lists.01.org --===============7383869164302353211== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Hi Andrew, > But we'll still need a memcpy, unless we use a separate io vector when > sending the message to the kdbus FD, but I'd have to check if it > doesn't break the splitting rules. > So one thought here would be to add extra space in the beginning of the = body used by gvariant builder. That way we can fill it with the variant = contents when we finalize the builder. We might need to keep an extra = book-keeping variable, something like body_offset, to know where the = body actually begins. It will be always 0 for dbus-1. Perhaps you can think of something better. Regards, -Denis --===============7383869164302353211==--