01:46 <steev> alrigh, based on 20221118 - pushed out the audio bits, but i'm not sure what i'm missing config-wise,

01:47 <steev> https://github.com/steev/linux/commit/5d9faf06dd99e462e2a0d9bd305fa948a155f637 rx_macro is probe deferring

01:51 <steev> srinik: ^ (and note that with that defconfig both tx/rx macros do get enabled)

09:00 <jenneron[m]> mothenjoyer69@ I've found that my galaxy book s artifacts only when modem is enabled (maybe I will catch it later with disabled modem?) at least with disabled modem it now seems fine

09:01 <jenneron[m]> definitely a hardware issue because it happens in both Linux and Windows

09:02 <jenneron[m]> I'm still not sure whether it could be caused by efforts of porting Linux

15:18 <gwolf> Cannot contribute much to the channel besides loving the machine ;-)

16:58 <mani_s> steev, johan fixed the nvme timeout issue

16:59 <mani_s> you just need to add "dma-coherent" property to the pcie node in dts

16:59 <mani_s> this is present in downstream but somehow we didn't use it in upstream for any SoCs

17:19 <steev> ahhh

17:19 <steev> well that is exciting

17:37 <steev> mani_s: that one? https://paste.debian.net/hidden/659edcd7

20:07 <steev> hm, most of the other boards put it into the dtsi, so i did that... but i can't reboot a bunch to test yet

21:13 <calebccff> I managed to get some signs of life on the C630 using the same kernel we run on the SDM845 phones, it flashes the kernel logs for about a second with efifb and then goes blank with the kb backlight on

21:13 <calebccff> is there a mechanism to dump logs from this thing or get serial out?

21:24 <jenneron[m]> calebccff: maybe too obvious suggestion, but make sure all panel-related configs are enabled, check CONFIG_DRM_TI_SN65DSI86 and CONFIG_DRM_PANEL_EDP

21:25 <calebccff> no... not too obvious at all...

21:25 <jenneron[m]> for taking logs in such cases I usually connect my phone to USB and use SSH, but that works only when device is booted into rootfs

21:27 <jenneron[m]> also, if you want pmOS on it (i'm not sure why do you want that kernel we use on phones), dtbloader is not necessary, check how https://gitlab.com/postmarketOS/pmaports/-/tree/master/device/testing/device-microsoft-surface-rt works

21:33 <calebccff> cheers, I guess I'll go for something similar here

21:36 <jenneron[m]> this device is armv7, so change `--format="arm-efi"` and rename bootarm.efi > bootaa64.efi

21:37 <jenneron[m]> also, this device package does not install grub modules related to efi gop because it is broken on surface rt

21:37 <jenneron[m]> you can use this instead: grub-mkimage --prefix="/grub" --output="bootaa64.efi" --format="arm64-efi" --compression="xz" all_video disk efi_gop fat gzio iso9660 linux normal part_gpt part_msdos

21:37 <calebccff> I've been using device-tablet-x64uefi as a base and modifying it for aarch64 lol

21:38 <jenneron[m]> surface rt example is a bit more appropriate because it uses dtb, but that one is useful too

21:54 <steev> android on the c630... why do you make santa sad

21:58 <jenneron[m]> android?

21:58 <calebccff> steev: pmOS is based on Alpine Linux

21:58 <steev> oh

21:58 <steev> i always think it's a custom android for some reason

21:58 <calebccff> we scrap all the proprietary android stuff and run mainline + mobile optimised userspace

21:59 <steev> if you feel like, you can always poke at our defconfig https://github.com/steev/linux/blob/linux-6.0.y/arch/arm64/configs/distro_defconfig that's what i use on my c630

22:00 <calebccff> oh nice, thanks for the link!

22:00 <calebccff> I finally got it booted, although the screen is blank for like 15 seconds during the boot process, any idea what's going on there?

22:01 <calebccff> it just flashes up once and then goes black until i guess drm/msm fbdev emulation loads

22:01 <steev> timeout looking for firmware, possibly

22:01 <jenneron[m]> calebccff: you probably want panel modules into initramfs

22:01 <steev> or that

22:01 <calebccff> ah makes sense

22:01 <calebccff> thanks both

22:01 <steev> actually, let me update the config again

22:01 <jenneron[m]> it's possible that some thing like dispcc or disabling unused clocks kills efifb, and you get panel only when booted to rootfs

22:36 <javierm> calebccff: what I use usually is "earlycon efi=debug debug clk_ignore_unused pd_ignore_unused"

22:36 <javierm> toprevent the case that jenneron[m] mentioned

22:38 <calebccff> javierm: thanks, that looks handy. Is {clk,pd}_ignore_unused still necessary on 6.0/6.1?

22:41 <javierm> calebccff: I'm not familiar with that particular device, I was talking in general :)

22:42 <javierm> calebccff: fwiw in the HP X2 Chromebook I still need to boot with deferred_probe_timeout=30 in v6.1

22:43 <javierm> but that's not your problem since you at some point get the msm drm driver probed while I don't

23:00 <steev> yeah they are calebccff

23:00 <steev> clk will be for a very long time :(

23:01 <steev> actually, i don't use pd_ignore_unused here on my c630

23:28 <calebccff> steev: what clock is missing?

23:37 <steev> calebccff: it's not that, afaik, it requires reworking the qcom clk stuff

23:40 <calebccff> steev: oh right, I'm booting without it here and stuff seems to be working? just wondering what to look for