That's still not quite right :-) What Randy meant is that the sign-off
must come /after/ the patch description:
From: Random J Hacker <random@hacker.com>
/*
* Above line not really required if it's your own patch that you're
* sending. But required when you're "relaying" someone else's patch,
* or if there are multiple people in the sign-off chain below,
* to differentiate as to who was the author of the patch.
*/
[PATCH] <subsystem>: <one-line summary>
/*
* Above line not really required if Subject: of email already such.
* But may be required if you're replying on some other LKML thread,
* whose subject may have nothing to do with this patch.
*/
<a few line(s) of patch description,
maybe a couple of paragraphs even.>
Signed-off-by: Random J Hacker <random@hacker.com>
Signed-off-by: Bob Z Maintainer <bob@maintainer.com>
/*
* It is generally preferable to order the sign-off's above so that
* they reflect the patch journey -- so the original author preferably
* comes first, then those who relayed it or picked it up from the
* author (often the maintainer), and so on, till the person who is
* finally relaying it currently (so from the above it seems this is
* being sent out by Bob). If this is the first time a patch is being
* sent, there would naturally be just a single sign-off line.
*/
Then three hyphens, followed by diffstat and patch ... so the below
looks good :-)
[ All this is not too important, admittedly, but causes least amount of
processing time to be wasted on the recipient's end, and also does not
confuse scripts that may be used to extract patches (and git commit
command-line arguments) from mails automatically. ]
-
