Linux-6.x kernel

Hardware Librem 5
1

The today apt update updated my L5 to this kernel:

purism@pureos:~$ uname -a
Linux pureos 6.0.0-1-librem5 #1 SMP PREEMPT Wed Oct 5 09:24:53 UTC 2022 aarch64 GNU/Linux

Is there any compressed list what’s new with this?

4 Likes
2

Here’s a compressed list

https://kernelnewbies.org/Linux_6.0

It’s still very long though, so if you by condensed mean only contains info related to Librem 5 I don’t think that’s available. From a quick search though, the only relevant thing I could find was a compression option for the hardware jpeg encoder.

1 Like
3

I found three related ChangeLog-6.0 commits: 43b240d41b5173eaebb825f19c423b4278435d2c, 69ed3dd6be9cf5d75252940b9a927dff4bab7860 and e62563db857f81d75c5726a35bc0180bed6d1540.

4

It’s mostly just a rebase of our downstream patches on top of upstream 6.0 tree. The only meaningful difference when it comes to downstream changes (already present in 5.19, but 5.19 was never pushed out to be the default) is a change to lower SoC voltages a bit in order to reduce power consumption and produce less heat.

There are also two other minor changes since 5.18 (multicolor LED driver enabled in the config and alternative vendor-provided touchscreen driver included in the tree), but these have no effect on regular users at the moment.

7 Likes
5

In case your question is in that direction:
There’s absolutely nothing fancy about 6.0, but Linus was in the mood of switching to a new major version after 20 releases for 5.x.
In fact, it’s expected that 6.1 will have bigger changes than 6.0 (although maybe not very relevant for L5 users).

3 Likes
6

At least those who are having pre-orders on batch Fir can find (just another speculation of mine, quite unofficial one) themselves there: “The new NXP i.MX93 SoC, the follow-up to the popular i.MX6 and i.MX8 embedded SoCs, now using Cortex-A55 cores and the Ethos-U65 NPU.”

P.S. Not that I’m writing here about NXP® i.MX93 SoC (Cortex-A55), please do not understand myself wrongly, I’m writing about follow up traces related to this article: “94 and above for performance and I/O richness, all with the same ML, security and energy efficiency.” Additionally same article is in accordance with this kind/main thread/post: Will Linux phones get decent processors in the future?. Therefore this post is actually related toward NXP i.MX94 SoC (or some other/particular one within here linked, officially introduced NXP® i.MX9x family.

7

Yeah he mentioned that. Basically when the minor versions number starts getting too confusing like say; going over x.20.x he prefers to do major version to help keeping it simple.

1 Like
8

Document changes to the upcoming kernel release 6.0.0pureos2 librem5:

[ Sebastian Krzyszkowiak ]

  • ASoC: wm8962: Add WM8962_ADC_MONOMIX to “3D Coefficients” mask
  • ASoC: wm8962: Don’t disable SYSCLK if it’s enabled via DAPM
  • ASoC: wm8962: Don’t report a microphone if it’s shorted to ground on plug
  • ASoC: wm8962: Don’t mark WM8962_ADDITIONAL_CONTROL_4 as volatile
  • dts: imx8mq-librem5: Don’t set mic-cfg for wm8962
  • ASoC: wm8962: wm8962_mic_work: Allow a one-shot operation without IRQ
  • hack: ASoC: simple-card-utils: Call into wm8962’s mic detection
  • hack: ASoC: wm8962: Trigger mic detection on out_pga_event
  • Revert “drm: etnaviv: Unmap gems on gem_close”
  • d/control: Add breaks relationship for Mesa without BO fixes

[ Lucas Stach ]

  • drm/etnaviv: move idle mapping reaping into separate function
  • drm/etnaviv: reap idle mapping if it doesn’t match the softpin address

Edit: Rebased: To kernel release 6.0.2pureos1 Librem5

4 Likes