[Ksummit-2012-discuss] [ATTEND] depth of our git tree structure; HID subsystem; kernel bugzilla; stable review
Guenter Roeck
linux at roeck-us.net
Fri Jun 15 22:32:13 UTC 2012
On Fri, Jun 15, 2012 at 11:32:23PM +0200, Jiri Kosina wrote:
> Hi everybody,
>
> I'd like to attend Kernel Summit this year. Selection of the topics I'd
> consider worth discussing:
>
> (1) HID subsystem, which I am happy to maintain, has grown from "feeding
> data mechanically into input layer" in several aspects -- it can be
> seen as a "leader" in multitouch in Linux these days, and we are
> extending its coverage of transport protocols as well
> (not only USB/Bluetooth, but expanding to to I2C, IIO, and userspace
> transport drivers mostly because of Bluetooth-LE).
> Thus, as we are gaining "cross-subsystem" coverage, discussion with
> affected subsystem maintainers would be welcome. If "general KS
> public" is interested, I can of course give a "state of the union"
> short overview presentation.
>
> (2) This might be more question to Linus rather than discussion topic, but
> still ... Time to time I have a gut feeling that our development
> "tree structure" is not as deep as it could/should be, mostly because
> many people are sending pull requests directly to Linus instead of
> going through appropriate "upstream" trees.
> I'd like to know whether it's just me having this feeling, or there
> are other maintainers sharing this. If the latter is the case,
> discussion might be useful.
>
> (3) Kernel bugzilla; do we really want it for general public bug reports?
> For trees I am maintaining, I very much prefer all bug reports and
> patch reviews/discussions to happen via e-mail; consequently, I am
> not very careful when it comes to reports received through
> bugzilla.kernel.org, and tend to drop bugzilla-reported issues on the
> floor silently. I understand that users/bug reporters might be
> confused in which cases to prefer bugzilla over mailinglist (and vice
> versa), so it'd be nice to actually try to improve the situation
> here.
>
> (4) I am a responsible maintainer of kernels for all SUSE enterprise
> products. As such, I am dealing with -stable trees on a regular
> basis. It's sometimes difficult to decide about -stable tree merge
> into our enterprise tree (depending on the development stage of the
> enterprise kernel), due to the amount of changes it flows into
> -stable. I'd like to understand whether all the patches are really
> always necessary, and how can we help with review/filtering.
>
Doing the same for my employer, I have not had much (if any) trouble with that.
There may be some glitches (I usually wait for a week or two before pulling in a
new kernel), but overall I think the stable maintainers are doing an excellent job.
My problems in respect to -stable are:
- stable tree submission process, which doesn't apply to all subsystems.
Would be great to have a single well defined process for all -stable
submissions. That doesn't mean that all patches sent to -stable have to be
handled by the same person, but the "patches for the XXX subsystem are not
handled here" feedback is a bit odd for newcomers.
- stable release selection process. Would be great to have some predictable
process, ie to know ahead of time if a release is going to be a stable release.
That would make long term release planning much easier.
Guenter
> (5) If James is invited so that he can (in the order of several
> magnitudes) increase the number of people in the group who
> actually know how to dress for formal ocasions properly, I'd like to
> pay my debt to him (I have promised to let him taste a really good
> plum brandy, but the brain cell that was responsible for not
> forgetting this died prematurely last year).
>
> Thanks,
>
> --
> Jiri Kosina
> SUSE Labs
> _______________________________________________
> Ksummit-2012-discuss mailing list
> Ksummit-2012-discuss at lists.linux-foundation.org
> https://lists.linux-foundation.org/mailman/listinfo/ksummit-2012-discuss
>
More information about the Ksummit-2012-discuss
mailing list