At Sat, 11 Jul 2015 15:48:09 +0200, Richard Brown wrote:

Dear openSUSE Kernel Maintainers,

I assume most of you have been following the mailinglist regarding the discussions about openSUSE Regular Releases, but to very briefly summarise:

- openSUSE 13.2 looks likely to be the last of the 'Factory-snapshot-based' Regular Releases (There is still a possibility volunteers will step up, but the previous team involved in those releases are not continuing them) - openSUSE will be producing a new Regular Release called 'Leap' - 'Leap' will consist of a combination of both openSUSE community packages and packages based on SLE sources - The first version of Leap will be 'Leap 42.1' which we're hoping to release at SUSEcon in November - 'Leap' releases will be aligned with SLE releases, with a 'minor release' around the time of each SLE Service Pack, and a 'major release' before a new SLE Version. - Maintenance of the 'Leap' releases are likely to be similarly aligned - I expect maintenance of each minor release to overlap only for 6 months like with SLE. Maintenance of the last-Leap-minor-release before a Major release is likely to be extended until the first-Leap-minor-release of the new version (eg. Assuming 42.3 is the version that comes out before 43.0, 42.3 should receive maintenance updates until 43.1 is out) - As a general rule of thumb, the expectation is that the SLE sources will provide the 'base' of the operating system, with community packages built ontop

And this is where I come to you. Feedback from both our existing contributing community and random people on the internet has been overall pretty supportive of the idea, however there's one recurring sticking point

The SLE Kernel, it's age, and potential lack of hardware support

While me and others have fought hard explaining that SUSE do a lot of work adding hardware enablement in the SLE Kernel, and the version number is not a good reflection of it's capabilities, this hasn't sent those concerns away

I've also spoken to Takashi, who gave me some pretty good solid examples of specific scenarios where the SLE Kernel is missing stuff's likely to be very important to openSUSE users (Hardware support for Laptops newer than 2 years for example)

So together we've come up with the suggestion that Leap 42.1 could include the Linux 4.1 Kernel

I understand it's expected 4.1 will be an LTS Kernel, which is perfect for our desire for Leap to be seen as the more-stable alternative to Tumbleweed.

I do not think we should set too much in stone right now, especially as much of Leap is still at the 'putting together' stage, but I do know that SUSE are considering a Kernel upgrade for SLE 12 SP2, so my current line of thinking would be that having Kernel 4.1 in Leap 42.1 might end up being a 'stop-gap' before Leap takes whatever new kernel ends up in SLE 12 SP2.

Even if SUSE do not change their kernel in SLE 12 SP2, if 4.1 is going to be an LTS release, it might make sense for us to keep using it even for Leap 42.2, which is another reason choosing Kernel 4.1 now seems like a good idea to me.

I don't want to declare policy on you guys, but I think a guideline of "Favouring LTS Kernels and using the SLE kernel if it makes sense" is probably a good starting point for a policy for "which Kernel should Leap have?"

So I guess I can boil down this long post to a few questions

1. Do you like the idea of Kernel 4.1 in Leap 42.1? 2. If yes, are you/is one of you willing to maintain Kernel 4.1 as the kernel for Leap 42.1? 3. Do you like the idea of favouring LTS Kernels and/or SLE Kernels for future Leap releases?

As a person who proposed such an idea, I'm willing to take the role of this git branch maintainer (openSUSE-Leap-4.1 or whatever) once when we agree with this direction. And I already started some conversation with SUSE kernel teams about the group maintenance of 4.1 kernel, so this shouldn't be a big burden. thanks, Takashi -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org

At Sat, 11 Jul 2015 12:28:25 -0700, grantksupport@operamail.com wrote:

On Sat, Jul 11, 2015, at 12:07 PM, Takashi Iwai wrote:

...

As a person who proposed such an idea, I'm willing to take the role of this git branch maintainer (openSUSE-Leap-4.1 or whatever) once when we agree with this direction. And I already started some conversation with SUSE kernel teams about the group maintenance of 4.1 kernel, so this shouldn't be a big burden.

Would there continue to be the equivalent of a Kernel:Stable repo that could be used on/with a Leap-whatever install?

Similar to what's been available to Opensuse 13.2?

You should be still able to install Kernel:stable kernels as is. Actually the kernel has very little dependency on distro versions, and it's kept compatible. Kernel:stable package works even on SLE11 in most cases for now. Takashi -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org

On 14 July 2015 at 15:19, Jeff Mahoney wrote:

...

1. Do you like the idea of Kernel 4.1 in Leap 42.1?

Given the timeline, it'll be the best match.

Great

...

2. If yes, are you/is one of you willing to maintain Kernel 4.1 as the kernel for Leap 42.1?

It depends on what we mean by support. Security and crash/corruption/etc updates? Sure. We already do that with openSUSE point releases, which will be eclipsed by this. If you mean something closer to what we offer to SLES where we have expected SLAs, etc, then that's going to be more difficult without resources to dedicate to that task.

I purposefully didn't use the word 'support', which I see as a more loaded term casting images of L3 support, SLA's, blah blah. When I say "maintenance" I mean security, crash/corruption/etc updates, yes. So yes, I believe the same level of maintenance for the Leap kernel as was provided to openSUSE point releases in the past is perfectly sufficient.

...

3. Do you like the idea of favouring LTS Kernels and/or SLE Kernels for future Leap releases?

It's certainly easier from a maintenance perspective to use SLES kernels but I don't think it really hits the sweet spot for the target user for Leap.

The good news is that, outside of the Xen patches, our kernels tend to be fairly close to upstream. The SLES kernels are almost entirely driver backports. So if we're more aggressive about which kernel we use for Leap, we shouldn't need those. I spent some time yesterday and will spend some time today going through what we have in SLE12 and what would need to come forward. For things like Xen, the work is already done thanks to the effort by Jan Beulich. It's also the last kernel that we get stack unwinding relatively inexpensively (effort-wise), so it's a good pick from that perspective as well.

Sounds good, thanks for your perspective Jeff :) -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org

Am Dienstag, 14. Juli 2015, 17:28:27 schrieb Richard Brown:

On 14 July 2015 at 15:19, Jeff Mahoney wrote:

...

...

1. Do you like the idea of Kernel 4.1 in Leap 42.1?

Given the timeline, it'll be the best match.

Great

...

...

2. If yes, are you/is one of you willing to maintain Kernel 4.1 as the kernel for Leap 42.1?

It depends on what we mean by support. Security and crash/corruption/etc updates? Sure. We already do that with openSUSE point releases, which will be eclipsed by this. If you mean something closer to what we offer to SLES where we have expected SLAs, etc, then that's going to be more difficult without resources to dedicate to that task.

I purposefully didn't use the word 'support', which I see as a more loaded term casting images of L3 support, SLA's, blah blah.

When I say "maintenance" I mean security, crash/corruption/etc updates, yes.

So yes, I believe the same level of maintenance for the Leap kernel as was provided to openSUSE point releases in the past is perfectly sufficient.

...

...

3. Do you like the idea of favouring LTS Kernels and/or SLE Kernels for future Leap releases?

It's certainly easier from a maintenance perspective to use SLES kernels but I don't think it really hits the sweet spot for the target user for Leap.

The good news is that, outside of the Xen patches, our kernels tend to be fairly close to upstream. The SLES kernels are almost entirely driver backports. So if we're more aggressive about which kernel we use for Leap, we shouldn't need those. I spent some time yesterday and will spend some time today going through what we have in SLE12 and what would need to come forward. For things like Xen, the work is already done thanks to the effort by Jan Beulich. It's also the last kernel that we get stack unwinding relatively inexpensively (effort-wise), so it's a good pick from that perspective as well.

Sounds good, thanks for your perspective Jeff :)

I hope I'm not too pushy here for a community-project. But isn't it about time to come to a conclusion? Regards! Rainer Fiebig -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org

isn't it opensuse leap 42 supposed o be based on suse 12? so why changing the kernel? suse going also to based their OS on 4.1 kernel ? i understand if their will be newer gnome version or other packages but to say that opensuse leap will be based on suse 12 and change the core/base package that called kernel to 4.1 that i don't understand. can someone explain to me? On 07/17/2015 04:12 PM, Richard Brown wrote:

On 17 July 2015 at 12:20, Jay wrote:

...

...

Sounds good, thanks for your perspective Jeff :)

I hope I'm not too pushy here for a community-project.

But isn't it about time to come to a conclusion?

Your enthusiasm is infectious ;)

I think it's safe to say that the conclusion is almost certainly "we will use 4.1 as the Kernel version"

However, there's a few details about how that need discussing, mostly about the 'How', and it's probably to be better safe than sorry and only announce a firm conclusion once we've got that 'How' figured out.

Luckily, a bunch of our kernel developers just happen to be visiting my office in Nuremberg next week, so we'll be having a conversation about that then and you can expect seeing more movement after that ;)

-- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org

On Saturday 18 July 2015 11.08:51 Stefan Seyfried wrote:

Am 18.07.2015 um 08:55 schrieb nirsuse:

...

isn't it opensuse leap 42 supposed o be based on suse 12? so why changing the kernel?

To be able to install and use it on hardware newer than 2 years?

And fix the noise about having a kernel (even if not true) that is older than 13.2 But then it raise the inverse noise ... -- Bruno Friedmann Ioda-Net Sàrl www.ioda-net.ch openSUSE Member & Board, fsfe fellowship GPG KEY : D5C9B751C4653227 irc: tigerfoot -- To unsubscribe, e-mail: opensuse-kernel+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-kernel+owner@opensuse.org