Re: This merge window...

[Leon Romanovsky <leon-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>]

[-- Attachment #1: Type: text/plain, Size: 3531 bytes --]

On Sun, Jul 16, 2017 at 08:46:43AM -0400, Doug Ledford wrote:
> On 7/15/2017 11:57 PM, Leon Romanovsky wrote:
> > On Sat, Jul 15, 2017 at 02:12:42PM -0400, Doug Ledford wrote:
> >> On 7/14/2017 10:28 AM, Christopher Lameter wrote:
> >>> On Thu, 13 Jul 2017, Sagi Grimberg wrote:
> >>>
> >>>> Join on the awesomeness.
> >>>
> >>> It would even be more awesome if we had multiple people that can work the
> >>> release process so that we do not have a single person bottleneck.
> >>>
> >>> Doug: Would you please do some more detailed planning and delegate things
> >>> as possible? A couple of developers should be able to review the current
> >>> state of things and help to move things forward if there is a bottleneck
> >>> currently or coming up.
> >>>
> >>
> >> I am, without a doubt, reworking my process.  I will not be syncing up
> >> against Dave Miller's tree any more.  When I take patches, they should
> >> apply cleanly to my tree and work without dependencies.  If there are
> >> dependencies, then people need to send the dependencies through Dave's
> >> tree (assuming that's where they go) in release X, and then send the
> >> dependent code to my tree in release X + 1.
> >
> > There are two parts in this proposal, while second part (dependency) is
> > fair enough and possible to meet, it is unclear to me the first part
> > (apply cleanly).
> >
> > 1. On which branch should we send our topics?
>
> You can always use -rc1 as a safe starting point.  I'm going to try to
> open things up around then anyway.

No problem, it is the same as we do it anyway for our shared code.

>
> > 2. Are you acknowledge that this branch is going to be updated +/- on daily basis?
>
> Not necessarily daily, but certainly every 2-3 days.

2-3 days in some countries is the same as to say daily :)

>
> > 3. How do you see us submitting multiple topics? Sequential submission - like DaveM,
> > and it should be applied close to submissions (if no objections and so on ..).
> > Or parallel submissions - like it is now and the conflicts are unavoidable.
>
> Not the parallel submissions we are using now.  It generates far too
> many conflicts and such.  What I would prefer to see is one submission,
> then two or three days (so the first submission has had some bake time),
> then the next one and it should assume the first is applied and apply
> cleanly on top of 4hat.

I'm totally in for it. It will simplify greatly the whole my process and
will allow to do more cross-tree work.

Can we move to pull-request model? So our whole submission queue will participate
in linux-next and we will move our testing to use it as a testing branch instead of
our artificial queue-next [1]?

It will allow to all of us to have trees in linux-next, because we will
have same SHA1 and Stephen won't have any merge conflicts between trees.

>
> > 4. If 2 and 3 are not going to change, should we wait till late rc and submit all (100+)
> > patches at one or two shots to avoid merge conflicts?
> >
> > I would be glad to get the whole picture of submission process, before
> > we are moving forward.
>
> Is what I wrote above clear enough?

Thanks, it is much clear.
When do you expect to update your branches so we will be able to prepare our shared code?

[1]
https://git.kernel.org/pub/scm/linux/kernel/git/mellanox/linux.git/log/?h=queue-next

>
>
> --
> Doug Ledford <dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
>     GPG Key ID: B826A3330E572FDD
>     Key fingerprint = AE6B 1BDA 122B 23B4 265B  1274 B826 A333 0E57 2FDD
>




[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]