linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Hauke Mehrtens <hauke@hauke-m.de>
To: Johannes Berg <johannes@sipsolutions.net>
Cc: "Luis R. Rodriguez" <mcgrof@do-not-panic.com>,
	"backports@vger.kernel.org" <backports@vger.kernel.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: Re: Dropping non supported kernels for kernel backports ?
Date: Wed, 08 May 2013 10:32:37 +0200	[thread overview]
Message-ID: <518A0DA5.4060001@hauke-m.de> (raw)
In-Reply-To: <1367996104.8279.5.camel@jlt4.sipsolutions.net>

On 05/08/2013 08:55 AM, Johannes Berg wrote:
> On Tue, 2013-05-07 at 16:53 -0700, Luis R. Rodriguez wrote:
>> Today the backports project provides support to backport down to
>> 2.6.24 for some subsystems. While this is good for users in practice
>> for development and maintenance this is quite a bit of overhead. Apart
>> from older kernels there are also gaps in between stable releases that
>> are not supported. For example 3.8 and 3.4 are supported but anything
>> in between is not (3.5, 3.6, 3.7), but we still do support them on the
>> backports project. At times this may mean a stable fix may get
>> propagated onto a the linux-3.4.y branch but obviously not the the
>> linux-3.5.y branch. If backporting expressing this becomes a bit
>> complex and we have dealt with it. In short, its a pain.
>>
>> I'd like to see what folks thought if we went ahead and *only*
>> supported kernels listed on kernel.org as supported. This would help
>> with the Linux kernel maintainer effort by also persuading users to
>> upgrade to stable releases as well as educating around this.
>>
>> If v3.10 backported releases are too soon to do this I propose we
>> seriously consider it for v3.11 releases. Any thoughts?
> 
> I'm sure I'd still have to support other kernels random customers might
> be on (customers, they don't always do the smart thing ;-) ). I could do
> that myself of course, but then where would I stick the support patches?
> 
> Maybe we could do a compromise and drop support for no longer supported
> kernels, but only a few years later? If we put the cutoff at 3 years now
> we'd drop everything before 2.6.34, for example.
> 
> johannes

I it is not often that something compat depends on gets backported into
some longterm kernel. The majority ~90% of backports added to compat are
not backported to any longterm kernel. If we want to support some kernel
version I do not think it is a hard task to support the versions in between.

Instead of supporting everything back to 2.6.34 I would like to got to
2.6.32 because this was used in Debian squeeze, RedHat 6 and some Ubuntu
long term version.

Dropping support for kernel versions not older than a year but not
supported any more is a bad idea, because the big distributions often
pick a non longterm kernel as their distribution kernel. Than we would
not even support the recent Ubuntu versions for a long time. Ubuntu
12.10 uses kernel 3.5 and Ubuntu 13.04 uses 3.8, both are no longterm
kernel versions.

Hauke

      reply	other threads:[~2013-05-08  8:32 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-05-07 23:53 Dropping non supported kernels for kernel backports ? Luis R. Rodriguez
2013-05-08  6:55 ` Johannes Berg
2013-05-08  8:32   ` Hauke Mehrtens [this message]

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=518A0DA5.4060001@hauke-m.de \
    --to=hauke@hauke-m.de \
    --cc=backports@vger.kernel.org \
    --cc=johannes@sipsolutions.net \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mcgrof@do-not-panic.com \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).