00:18 <bluerise> I know this channel is about laptops, but is anyone working on a device tree for Windows DevKit? :) OpenBSD does work in ACPI mode, but having a device tree would be nice as well.

00:42 <HdkR> kettenis: Someone needs to wire that up for me then :)

01:39 <steev> if i had one, i might bluerise, but, i don't have one, so no

01:39 <steev> t here

04:24 <leezu> hexdump0815: Reverting c44094ee fixes the wifi issue on lazor. Unfortunately there are also some external display regressions on 6.1. I observed "arm-smmu 15000000.iommu: Unhandled context fault: fsr=0x402, iova=0x03ff0e00, fsynr=0x7f0023, cbfrsynra=0x800, cb=5" upon connecting a monitor and dpu errors upon disconnecting. dmesg for display regression at

04:24 <leezu> https://termbin.com/asse I didn't have time yet to look further into the display issues.

04:28 <leezu> The c44094ee revert is upstream at b7d9aae4

14:54 <init> steev: just saw your push of rc8, any recommandation on the defconfigs i should use?

15:50 <steev> i would use the laptop_defconfig iirc

16:21 <init> thx

16:27 <init> well, so many changes. ended up using my rc7 working one.

17:42 <hexdump0815> leezu: thanks a lot for bisecting and the info

18:24 <steev> oh, sounds like you guys need dma-coherent stffs too

19:02 <HdkR> steev: Did you ever test X13s with gpu firmware again?

19:12 <steev> not in the past week, i poked at it a bit ago, but was failing :)

19:13 <steev> needed to focus on the kali release but that's out the door so now i need to put out the post release fires and then can look at things

19:22 <HdkR> cool cool

19:53 <hexdump0815> steev: do you mean dma-coherent too for sc7180? i guess no - ?

20:18 <HdkR> What is the general choice with zswap these days? LZO+zbud? LZ4? What's the differences between zbud, z3fold, and zsmalloc?

21:50 <steev> idk, i use the defaults

21:50 <steev> i think the current default is lz4

22:09 <HdkR> Interesting. I don't currently use it just because it wasn't enabled in the kernel config but I'm planning on tinkering with it

22:09 <HdkR> I've had problems with the Adreno driver faulting from paged out buffers, so getting CPU assets more compressed would be beneficial.

22:47 <hexdump0815> HdkR: i have very good experiences with zstd+zsmalloc with zswap percentage set to 20% and plenty of swap space (1-2x ram) as the compressed memory is counted as swap as well (although not written to disk)

22:50 <hexdump0815> or more correct: even if not written to disk

22:54 <HdkR> Fancy