On 11/09/2016 08:32 AM, Lars Ellenberg wrote:
On Tue, Nov 08, 2016 at 09:52:04AM -0700, Jens Axboe wrote:
This should go into 4.9,
and into all stable branches since and including v4.0,
which is the first to contain the exposing change.
It is correct for all stable branches older than that as well
(which contain the DRBD driver; which is 2.6.33 and up).
It requires a small "conflict" resolution for v4.4 and earlier, with v4.5
we dropped the comment block immediately preceding the kernel_sendmsg().
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Reported-by: Christoph Lechleitner <[email protected]>
Tested-by: Christoph Lechleitner <[email protected]>
Signed-off-by: Richard Weinberger <[email protected]>
Signed-off-by: Lars Ellenberg <[email protected]>
Changing my patch is perfectly fine, but please clearly state it.
I.e. by adding something like that before your S-o-b.
[Lars: Massaged patch to match my personal taste...]
Lars, are you sending a new one? If you do, add the stable tag as well.
So my "change" against his original patch was
- rv = kernel_sendmsg(sock, &msg, &iov, 1, size - sent);
+ rv = kernel_sendmsg(sock, &msg, &iov, 1, iov.iov_len);
to make it "more obviously correct" from looking just at the one line
without even having to read the context. And a more verbose commit message.
I'm fine with you making that change, I do that too sometimes for
patches. But Richard is absolutely right in that you need to make a note
of that. It's no longer the patch he signed off on, so it needs to
reflect that.
Should I sent two patches, one that applies to 4.5 and later,
and one that applies to 2.6.33 ... 4.4, or are you or stable
willing to resolve the trivial "missing comment block" conflict yourself?
Since it's a trivial one liner, let's just do one for the current series
and the stable folks should be able to do that one. If not, they will
let us know.
--
Jens Axboe
