I found this link and I read it and it gives no reason why the settings fo an LTsSkernel are different. The previous kernels and the following kernels all have
CONFIG_PREEMPT_VOLUNTARY not set
I was asking this question here because we have experts in the forum who are closer to the Arch devs or kernel devs. And they may know more about the topic. But thank you for your contribution.
See here. This was not able to be inplemented in kernel 5.15 but was enabled with 6.1 - the logic being “Enabling CONFIG_PREEMPT_DYNAMIC=y is a clear all around benefit allowing system administrators to choose the best mode for their use case [from the kernel config line].”
6.1 when it was linux did not change that setting:
you are expecting something that simply is not there. The LTS kernel is maintained by a completely different maintainer than the other two official kernels. Therefore, they also follow different philosophies and configs.
Maintainers of the regular kernel: Jan Alexander Steffens David Runge Tobias Powalowski Levente Polyak
the LTS kernel maintainer clearly sees that as a valid reason to change the default for the LTS kernel, while the other maintainers don’t see the necessity (or no one raised that issue for the other kernels).
This is an interesting read. Thank you for the link. Although the comments on this page are not making things clearer.
The maintainer Andreas Radke is saying:
And default CONFIG_PREEMPT=y (low latency desktop) is not desired for LTS kernels.
How come? I would say that this is an opinion based on the assumption that LTS kernels are for servers. And servers dont want CONFIG_PREEMPT=y. Well, that is debatable.
Anyways, I find it very confusing that an official arch kernel is changing the config if it is moving from linux to linux-lts. That does not make sense to me. A server admin can easily change the preemption on the kernel commandline, thanks to CONFIG_PREEMPT_DYNAMIC=y, but all the arch desktop users should not need to do that. In fact I am pretty sure that they do not even know that they have to do that to get the same system responsiveness as always.
If I add to the kernel comandline of the lts kernel preempt=full
How can I tell that the setting is active and not just sillently ignored? I gave it a try but I do not see any hint in the journal that preemption has changed. How can I check that?