On 25 May 2016, at 01:24, Greg Freemyer email@example.com wrote:
Nov 2016 - Release Leap 42.2 with 4.4 kernel
Spring 2017 - Release the Leap 42.2 "advanced kernel" (ie. the 2017 LTS kernel).
Fall 2017 - Release Leap 42.3 with the 2017 LTS kernel (no longer designated advanced).
Fall 2017 - Drop all support for the 4.4 kernel when Leap 42.3 is released
Spring 2018 - Release the 42.3 "advanced kernel" (ie. the 2018 LTS kernel)
Spring 2018 - Drop 2017 LTS kernel support (when Leap 42.2 support is dropped).
The end result is that over the summer 2 kernels are supported: the current LTS kernel and the previous year's LTS kernel.
But during the winter only one kernel is supported.
From the kernel team's perspective they may even be able to leverage SLE kernel support at all times because they will have to support SLE kernels longer than the 12 months I'm proposing. That assumes the SLE will be supporting each of the annual LTS kernels.
Greg, please consider the user impact of what you are proposing.
Two kernels in the same distro? Users having to switch to one during the lifecycle of the distro? That complexity and that forced changing brings with it risk - new kernels break things for some people. Some people don't want to take those risks. Those are the people we make Leap for.
Leap is never going to support the latest and greatest hardware. It's not meant to. It's meant to be a reliable, dependable, workhorse of a distribution that people can put their faith into.
A more moderate pace of change,like the one we currently have, is the best way to accomplish that.
For people like you who buy USB 3.1 Gen 2 Ludicrous speed cards so early in the protocols existence the Kernel doesn't support it yet, we have Tumbleweed.
You cannot have it both ways. If you want to leave life on the edge, and your hardware purchases suggest you do, then Tumbleweed is the best platform for you. You cannot crave stability - you're buying hardware so new that no software support is available. 4.6 is just the first version to add support for USB 3.1 Gen 2. There will be bugs. Things will not be stable for USB 3.1 Gen 2 for some kernel versions yet.
And before you argue '3.1 Gen 2 is going to take over the world in 12 months, I'd like to point out that USB 3 came out in 2008 and took years before becoming ubiquitous. 3.1 is only just now appearing on new mainstream machines and even then that is Gen 1, fully supported stuff I can find ONE motherboard, a high end MSI gaming board with a Gen 2 card in it
So even if they get it right first time and Kernel 4.6 has no USB 3.1 Gen 2 bugs, it's still a whole bunch of change for a very narrow benefit.
We're not building Leap to cover every edge case. USB 3.1 Gen 2 is an edge case now, it will be in November, and I do not feel it's support is a compelling argument for playing loose and risk with the entire concept and purpose of what we're doing with Leap.
Now, I am not saying that the SLE kernel will make sense for Leap 42.3. Just like we did for 42.1, we are in the position to make the right choice for us. We can talk about that next year, after 42.2's release, when we have very real facts to base our discussions on, such as the state of the upstream kernels at the time and our experience of them within Tumbleweed. But I am certain that choosing 4.4 sets us in a very good foundation for the future. Being an LTS kernel means we can be certain it will be well supported for the full lifespan of 42.2 and, if we choose to not do a kernel upgrade in 42.3, then we know that we still be on a very well supported, stable, LTS kernel.
And I guess that's my point, right now, based on the information we have and the points raised to date, I am utterly convinced that the SLE 12 SP2 Kernel (4.4 with whatever backports are already there) is the correct choice for Leap 42.2
I'm open to debate on that point but I feel the arguments need to be stronger than the USB 3.1 Gen 2 case.
I'm don't think it's too productive to worry too far into the future such as Leap 42.3 and beyond - that requires speculation or crystal balls, which aren't very useful, or time machines, which we don't even have support for in Tumbleweed yet ;)