All of lore.kernel.org
 help / color / mirror / Atom feed
From: Ian Campbell <ian.campbell@citrix.com>
To: Doug Goldstein <cardoe@cardoe.com>,
	Jonathan Creekmore <jonathan.creekmore@gmail.com>,
	xen-devel@lists.xenproject.org
Cc: Keir Fraser <keir@xen.org>,
	Ian Jackson <ian.jackson@eu.citrix.com>,
	Jan Beulich <jbeulich@suse.com>, Tim Deegan <tim@xen.org>
Subject: Re: [PATCH 0/4] Allow schedulers to be selectable through Kconfig
Date: Wed, 6 Jan 2016 14:45:41 +0000	[thread overview]
Message-ID: <1452091541.21055.58.camel@citrix.com> (raw)
In-Reply-To: <567448E9.2030702@cardoe.com>

On Fri, 2015-12-18 at 11:56 -0600, Doug Goldstein wrote:
> In the end I really see the primary people that build Xen on their own
> as project maintainers (XenServer, Qubes, OpenXT) or distro maintainers
> (Ubuntu, Debian, Gentoo, Yocto) or "expert" users. Most people will use
> Xen as it comes packaged for them already and the Kconfig changes give
> the project/distro/experts the flexibility they need to build up Xen
> without maintaining downstream patches. So these won't really be shiny
> knobs for users to twiddle.

In this context I would consider the distro maintainers (at least for
generic, non-Xen specialised distros, e.g. XenServer and QUBES don't fit
here) to be users.

I don't think we want such distros to all be shipping subtly varying
versions of Xen any more than I want individual end users to be doing so.
Users of community (rather than Enterprise/Commercial) distros use many of
our upstream resources, including reporting bugs on xen-users and -devel.

But of course you (quite naturally, I think) expect that they will want to
play with these things, and I think that they will be no less prone to
fiddling with the shiny nobs than a individual user would be.

I don't want to distro users coming to use for support because they tried
to follow http://wiki.xenproject.org/wiki/Credit2_Scheduler_Development#Usi
ng_Credit2 and it didn't work, only to find that their distro has decided
to turn off credit2 in the packages, for whatever random reason the package
maintainer had.

I don't see this as contrary to your stated goals (e.g. ripping out all the
other schedulers), but I consider you to be within the expert camp for
wanting to do so (and having the chops to handle whatever pieces you find
yourselves with). I have no objections at all to allowing experts such as
yourselves to configure things and I applaud you for doing this in an
upstream way (it is the right thing to do).

My concern is that while you rightly consider yourselves expert enough and
are building something for a specific (and AIUI targeted) use case many
normal users tend to think that if they are expert enough to find and flip
the switch then they are expert enough to deal with the consequences, when
they are not and/or they do not have the specific use case which the switch
was added to support i.e. they want common or garden Xen and we want that
to mean the same for everyone.

It's those people (including general purpose distro maintainers) who I
think need to be strongly discouraged from messing with these options
because there will be a strong gravity towards them doing so.

> What provides the defaults that users should use are the defconfig
> files. The reason this doesn't work out so well for Linux is that there
> are really not defaults that work for everyone.

... and sadly by switching Xen to Kconfig this expectation _is_ going to
flow over to us, once someone sees "make menuconfig" mentioned somewhere
they are going to get their Linux head on and start messing with it.

>  But on the flip
> side of the coin Ian Campbell has been helping Debian to upstream
> changes they need [2].

BTW, I'm also a Debian Developer (although not the Xen package maintainer),
so I am doing this with as much downstream as upstream hat on.

Ian.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

  parent reply	other threads:[~2016-01-06 14:47 UTC|newest]

Thread overview: 49+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-12-17 20:59 [PATCH 0/4] Allow schedulers to be selectable through Kconfig Jonathan Creekmore
2015-12-17 20:59 ` [PATCH 1/4] build: Hook the schedulers into Kconfig Jonathan Creekmore
2015-12-18  1:35   ` Dario Faggioli
2015-12-18  2:20     ` Jonathan Creekmore
2015-12-18  9:19       ` Dario Faggioli
2015-12-18 10:52     ` George Dunlap
2015-12-18 11:23       ` Dario Faggioli
2015-12-18  8:57   ` Andrew Cooper
2015-12-18 16:43     ` Jonathan Creekmore
2015-12-17 20:59 ` [PATCH 2/4] build: Alloc space for sched list in the link file Jonathan Creekmore
2015-12-18  9:01   ` Andrew Cooper
2015-12-18 16:40     ` Jonathan Creekmore
2015-12-18 16:48       ` Andrew Cooper
2015-12-18 17:07         ` Jan Beulich
2015-12-18 17:10           ` Andrew Cooper
2015-12-18 17:11           ` Jonathan Creekmore
2015-12-18 17:23             ` Andrew Cooper
2015-12-17 20:59 ` [PATCH 3/4] sched: Register the schedulers into the list Jonathan Creekmore
2015-12-18  1:42   ` Dario Faggioli
2015-12-18  9:03   ` Andrew Cooper
2015-12-17 20:59 ` [PATCH 4/4] sched: Use the auto-generated list of schedulers Jonathan Creekmore
2015-12-18  1:47   ` Dario Faggioli
2015-12-18  9:12   ` Andrew Cooper
2015-12-18 16:44     ` Jonathan Creekmore
2015-12-18 10:50   ` Jan Beulich
2015-12-18 16:00     ` Jonathan Creekmore
2015-12-18 16:43       ` Jan Beulich
2015-12-18 17:24         ` Jonathan Creekmore
2015-12-18 17:30           ` Andrew Cooper
2015-12-18 10:39 ` [PATCH 0/4] Allow schedulers to be selectable through Kconfig Jan Beulich
2015-12-18 10:45 ` Ian Campbell
2015-12-18 10:58   ` Jan Beulich
2015-12-18 11:08   ` Juergen Gross
2015-12-18 11:19     ` Ian Campbell
2015-12-18 11:30     ` Jan Beulich
     [not found]     ` <5673FC6202000078000C122B@suse.com>
2015-12-18 11:41       ` Juergen Gross
2015-12-18 17:56   ` Doug Goldstein
2015-12-18 18:25     ` Andrew Cooper
2016-01-06 14:45     ` Ian Campbell [this message]
2016-01-07 14:01       ` Jonathan Creekmore
2016-01-07 14:37         ` Jan Beulich
2016-01-07 15:00           ` Ian Campbell
2016-01-07 15:18             ` Doug Goldstein
2016-01-07 15:31               ` Ian Campbell
2016-01-07 15:43               ` Jonathan Creekmore
2016-01-07 15:54                 ` Jan Beulich
2016-01-07 15:47               ` Jan Beulich
2016-01-07 15:30             ` Ian Campbell
2016-01-07 15:51               ` Jan Beulich

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1452091541.21055.58.camel@citrix.com \
    --to=ian.campbell@citrix.com \
    --cc=cardoe@cardoe.com \
    --cc=ian.jackson@eu.citrix.com \
    --cc=jbeulich@suse.com \
    --cc=jonathan.creekmore@gmail.com \
    --cc=keir@xen.org \
    --cc=tim@xen.org \
    --cc=xen-devel@lists.xenproject.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.