05:33 <strongtz[m]> robclark: Hi, I wonder if the new X1 chip and the rest a7xx series gpu will be supported by freedreno(gallium) in the future?

05:39 <HdkR> Just need someone to backport all the a7xx patches to Freedreno from Turnip :)

05:40 <jglathe_x13s> altacus can you boot from USB on these? Windows Dev Kit 2023 is 8cx3, Surface Pro 9 5G (same M/B) is making progress. It's a question of the dtb mostly as it seems

07:41 <konradybcio> jenneron: regarding these swr errors, can you point me to your tree?

08:10 <krzk> konradybcio: jenneron[m]: that's not x13s but flex-5g which does not have sound.

08:10 <krzk> what's more, that's sc8180x where the sound was not brought up

08:10 <konradybcio> I know

09:01 <jenneron[m]> konradybcio: https://gitlab.com/sc8180x-mainline/linux

09:02 <jenneron[m]> flex 5g dts and sm8150 ASoC driver

09:30 <konradybcio> jenneron: looks like you're not setting the soundwire pin states

09:30 <konradybcio> (no pinctrl-0/names under the swr master)

09:35 <jenneron[m]> ok

09:35 <jenneron[m]> I will take a look

09:39 <konradybcio> same for the speaker pins, though this is probably not fatal

10:10 <jenneron[m]> konradybcio: i'm not sure what pin you want me to configure

10:10 <jenneron[m]> do you mean soundwire interrupt?

10:11 <jenneron[m]> this pin is from wcd9340, not from tlmm

10:14 <jenneron[m]> https://gitlab.com/sc8180x-mainline/linux/-/blob/master/arch/arm64/boot/dts/qcom/sc8180x-wcd9340.dtsi?ref_type=heads#L60

10:48 <konradybcio> is there no swr rx/tx pins on wcd?

11:25 <jenneron[m]> konradybcio: i don't see them

11:25 <jenneron[m]> i don't see hem on sdm845 either https://gitlab.com/sc8180x-mainline/linux/-/blob/master/arch/arm64/boot/dts/qcom/sdm845-wcd9340.dtsi?ref_type=heads

12:42 <robclark> strongtz[m], HdkR: yeah, it'll get a7xx at some point.. like whenever I get my hands on a7xx device (which will probably be the x1)

12:47 <HdkR> Just need more DEs that are full Vulkan :P

12:50 <strongtz[m]> yeah, currently I have to use zink+turnip, which has serious performance issues and graphical artifacts

12:50 <strongtz[m]> and plasma 6 is very unstable with such combination

13:12 <konradybcio> zink+tu is mostly fine, but it also sometimes has issues with seemingly very simple renders

13:13 <konradybcio> e.g. glmark2 has graphical artifacts

13:17 <steev> jenneron[m]: they may also be missing from sdm845 (and this could be why audio doesn't work there anymore either)

14:03 <krzk> jenneron[m]: open existing code for upstreamed board, like x13s and sc8280xp and look there. Whatever you missed, needs to be fixed.

14:04 <krzk> jenneron[m]: another way is to look at audio bringup patches for recent socs, like for sm8650 or x1e80100 and compare as well. You need on SoC sides GPR/APR, soundwire, macro codecs and LPASS LPI. Easy to find all this from commits.

14:05 <krzk> jenneron[m]: then on the board you need sound node and soundwire devices, plus any pinctrl which is not set by DTSI.

14:08 <krzk> although I see that you use wse wcd934x with soundwire interface itself... oh, that's some old kernel 6.7?

14:10 <konradybcio> krzk 8180 seems to be based on an older design, before all the macros

14:10 <konradybcio> wcd934x/936x

14:10 <krzk> konradybcio: yeah, I see now...

14:10 <krzk> so it even predates RB5, more like RB3

14:10 <konradybcio> y

14:10 <bamse> 8180 is based on 8150

14:11 <krzk> so speakers are on soundwire bus coming from wcd934x, not from the SoC

14:11 <travmurav[m]> konradybcio: re thermal-zones not probing: https://github.com/TravMurav/linux/commit/39f857f937dc58e2da280d2a2cd605a09ab88ca1 still don't understand how it works for you, especially for things without cooling-devices

14:12 <konradybcio> power allocator is not the default governor

14:12 <konradybcio> i need to manually select it through sysfs

14:12 <travmurav[m]> duh

14:12 <travmurav[m]> well that explains it lmao

14:14 <travmurav[m]> as in, it is in my config xD

14:14 <konradybcio> welp

14:16 <krzk> konradybcio: jenneron[m]: are you sure of your GPIOs for speakers? first, you put two different ones. second, you made them active low, so double check that it is actually supported.

14:18 <jenneron[m]> <krzk> "jenneron: open existing code for..." <- x13s is different, its soundwire is on SoC, not in codec afaik

14:18 <jenneron[m]> but i don't remember for sure

14:18 <krzk> also double check the address (thus also compatible)

14:18 <jenneron[m]> thus i used c630 as reference

14:18 <krzk> yeah, that's right

14:18 <jenneron[m]> <krzk> "jenneron: then on the board..." <- you probably missed that audio overall works, it is just broken soundwire & speakers

14:19 <jenneron[m]> <krzk> "although I see that you use..." <- yep i will need to rebase it

14:19 <krzk> oh my, this matrix quoting does not make it readable.

14:19 <jenneron[m]> ah

14:20 <krzk> double check if the speakers go out of reset - so the GPIOs and their low/high.

14:20 <krzk> also maybe you will have something in grep . /sys/bus/soundwire/devices/*/status

20:09 <alx__> anyone ever used the neural rpc proc on the x13s before?

22:14 <altacus> jglathe_x13s: I don't have these devices. I was curious about them and wondered if anyone knew anything about them. Would they share the same firmware?

22:20 <jglathe_x13s> altacus: Sort of, partially. the *8280.mbn files are signed in a way that they only load on the designated hardware (my observation) so you need to obtain it from Windows. The fun part is, it boots good enough (most of the time) without them. So you can iterate and debug then

22:22 <jglathe_x13s> If they allow boot from USB and with secure boot off, you could at least try. But they are a bet, still.

22:38 <jglathe_x13s> sc8280xp is quite a treatment performance wise, tho. even Windows runs quite nice on it.

23:04 <altacus> jglathe_x13s: Thank you for the info. I think I will hold off. I have enough devices with only partial linux support as is. The developments for the x13s are very encouraging, it looks like with a bit more patience, most, if not all of the functionality will be there, so I'm very excited!

23:05 <altacus> I'm also extremely curious about the X elite laptops, which according to news article should be released with Linux support baked in.