MX 17/18 Repository: The Liquorix Kernel Thread

[Stevo]

Virtual Box modules for host, or in a guest? Vbox 5.2.16 in the test repo should work for the host modules.

The last 4.14 kernel builds do include the Meltdown mitigation, though are older than the 4.15 stock Debian kernel in MX 17.1, so I can't really see the point of that version over a newer Liquorix version. The mitigations have been optimized in more recent kernels, and may even be out for 32-bit in a future kernel--maybe 4.19.

[Stevo]

The new 4.17-10 (based on 4.17.14) release is now in the test repo. This now has a workaround for the MuQSS scheduler bug that was causing to the use of excess CPU cycles, so should improve the battery life somewhat. CPU use at idle should now approach that of the Debian kernel.

[Stevo]

We now have 4.17-13 (based on 4.17.17 and 4.17.18 rc1) in the test repo. Changes over the last few releases include:

linux-liquorix (4.17-13ubuntu1~xenial) xenial; urgency=medium

* merge 4.17.18-rc updates
* merge bfq updates from algodev-github/bfq-mq
* merge intel bugfix
- drm/i915: Increase LSPCON timeout
* update version to 17.3

linux-liquorix (4.17-12) unstable; urgency=medium

* remove modification to cpufreq/ondemand (it changed unrelated thing)
* merge hrtimeout muqss patches, preserve 1000hz timer config
* update version to 17.2

linux-liquorix (4.17-11) unstable; urgency=medium

* merge 4.17.17
* merge 4.17.18-rc
* cpufreq/ondemand: raise minimum up_threshold to 10%
* update version to 17.1

[dr-kart]

4.17.0-17.3-liquorix-amd64 #1 ZEN SMP PREEMPT liquorix 4.17-13~mx17+1 (2018-08-21) x86_64 GNU/Linux

Works well with Nvidia 390.77 video driver. Thank you, Steven.

[asqwerth]

Time to finally change kernels...

I hadn't seen the need to change from the 4.15 liquorix until now, with this new vulnerability.

On another note, 4.16 kernels didn't seem to work well in some distros of mine where there are frequent kernel updates/changes (Korora, and an Arch derivative). With 4.16 kernel, those distros kept freezing/hanging. Once the latest kernel series moved to 4.17 in those distros, it was fine.

[Stevo]

Time to finally change kernels...

I hadn't seen the need to change from the 4.15 liquorix until now, with this new vulnerability.

On another note, 4.16 kernels didn't seem to work well in some distros of mine where there are frequent kernel updates/changes (Korora, and an Arch derivative). With 4.16 kernel, those distros kept freezing/hanging. Once the latest kernel series moved to 4.17 in those distros, it was fine.

I haven't read up on the new L1TF problem, but I got the impression that it required actual hardware access for a desktop system. It was considered much more of a problem for virtualized systems on network servers.

I haven't tested if there's actually any Meltdown kpti mitigation for 32-bit systems yet, either. I'll see what the checker reports in Virtual Box.

[asqwerth]

I would prefer not to change, to be honest. 4.15 liquorix works nicely on both MX15/16 as well as MX17 installations, on both my PC and new laptop.

[stsoh]

I would prefer not to change, to be honest................

imo, always go with the latest stable kernel 4.17.18 (self compiled).
pro - fix faults, improve codes/securities.
con - new fix, come with new error, cycle continue endlessly.

[Stevo]

We now have 4.18-4 in the test repo, based on vanilla kernel release 4.18.5. Upgraders should also use the latest Nvidia driver from backports as necessary, or other dkms packages and firmware from our MX test repo if your hardware requires those.

All 4.17 Liquorix kernel versions have been removed except for the last release.

[cinclus]

https://plus.google.com/communities/117 ... 4346186936
https://drive.google.com/drive/folders/ ... DugCW492cp
There is a kernel patch-set from John Brodie for building a 32-Bit Liquorix-kernel (4.16.16-zen) for my Asus T100TA.
Question:
The 32-Bit Liquorix kernel-configuration file is ".config-i686-zen (baytrail)". Would that be suitable for building a live-kernel for MX-17?
Any comments highly appreciated.