03:08 <dgilmore> https://www.irccloud.com/pastebin/BKntUQut/

03:08 <dgilmore> i had to remove and reload qcom_q6v5_pas

03:09 <dgilmore> I assume it would be unecessary if i forced the firmware into the initramfs

03:29 <steev> should be able to poke somewhere in sysfs to restart the remote procs

03:37 <dgilmore> yeah

03:37 <dgilmore> ideally when / is mounted it is retried

05:28 <steev> no, i mean, it's possible, i just don't remember where

05:40 <agl> I'am on holiday for 5 days ...

06:04 <valpackett> <Jasper[m]> "@[Val Packett] @jglathe_sdbox2..." <- that's the silly misconfigured compander, RX_COMP thingy

06:04 <valpackett> <valpackett> "+ https://github.com/jglathe/..."; <- here for headphones, but do delete it for speakers too

06:07 <valpackett> on the latitude/inspiron, RX_COMP on speakers does a very obvious overboost on attacks and makes the volume curve totally bonkers

06:09 <valpackett> but on headphones it uhhhh..... somehow effectively pans to one side all the sounds quieter than some threshold, lolwtf

06:35 <travmurav[m]> last time I tried to cset those knobs it didn't fix anything, decided to try again today after actually making a setup to verify the jank

06:36 <travmurav[m]> but now after I turned it off once, it seems to have fixed the jank but now I can't get the jank back by turning the knobs back on???

06:54 <travmurav[m]> meh can't tell if those settings just don't apply instantly, now it's stuck with jank again

07:10 <Jasper[m]> I just rebooted

07:26 <JensGlathe[m]> Updated the branches accordingly, thanks for the hint Val Packett https://github.com/jglathe/alsa-ucm-conf/tree/jg/master

07:54 <trchen> dgilmore: I think you might have a race between ADSP booting and qrtr-smd.ko loading. Did you build qrtr-smd as module or builtin? I solved my probe issue by making CONFIG_QRTR_SMD=y.

08:00 <Jasper[m]> <valpackett> "that's the silly misconfigured..." <- I applied that change aswell

08:01 <Jasper[m]> On the heaphonejack

08:01 <Jasper[m]> * On the headphonejack

08:01 <Jasper[m]> I'll go over it again tomorrow, but I'm fairly sure I did it properly

10:22 <crpalmer> dgilmore, you don't by any chance have qcom_q6v5_pas blacklisted in /etc/modules.d (or whatever the right folder is)? If the installer was booted with that blacklisted then it gets carried forward to your installed imaged.

11:42 <jdb[m]> @valy_: similar issue with the Surface Pro 9 5G, I never got to get the built-in eDP display working so far

11:48 <angeryboi[m]> os-prober situation with ubuntu 25.10 got slightly better, but still doesn't work 🤔

11:48 <tobhe[m]> apparmor fix is WIP

11:49 <angeryboi[m]> Ah, ok thanks

11:50 <tobhe[m]> https://launchpad.net/ubuntu/+source/apparmor there is a newer version in questing-proposed that might actually fix it

11:56 <xeliteCJF> I update my ucm2, using jglathe/alsa-ucm-conf/tree/jg/master . Bluetooth & Headphones working but still get no sound from my speakers

11:56 <xeliteCJF> qcom-soundwire 6b10000.soundwire: qcom_swrm_irq_handler: SWR CMD error, fifo status 0x4e00c706, flushing fifo

11:56 <xeliteCJF> end up with the soundwire SWR error.

12:18 <dgilmore> trchen: I do not think that is it, adding the firmware to the initramfs fixes it. I need to boot an older kernel and see when it was loaded

12:43 <anthony25> I'll need to look at it closely and paste the error message, but anyone's having issue with ath12k and cryptsetup in initramfs? If the laptop (slim 7x) stays too long in initramfs (for example because I rebooted and went making a coffee), something seems to timeout and I don't have wifi after

12:43 <anthony25> I build ath12k as a module, and include the firmware in initramfs

12:44 <anthony25> 6.17-rc6 linaro's branch

12:47 <crpalmer> dgilmore, oh man, I completely forgot. I have a dracut file in /etc/dracut.conf.d/ that adds the firmware to the initramfs: https://pastebin.com/raw/RE0HbYdC

12:51 <dgilmore> anthony25: I've seen that here on my t14s

12:52 <dgilmore> crpalmer: it was not necessary to do before.

13:01 <anthony25> dgilmore: ok great, thanks, so it's not just me

13:03 <dgilmore> booting the fedora kernel without CONFIG_PINCTRL_SM8550_LPASS_LPI being unset the adsp is started 13 seconds after poweron, at which point the rootfilesystem is mounted. with it being set to a module the adsp is started 2 seconds after poweron, causing the firmware to not be available. and the need to get it manually restarted after boot. or putting the firmware i the initramfs

13:10 <rfs613> some spammers seem to have found the Aarch64-laptops mailing list...

13:10 <robclark> dgilmore: this is what I do for remoteproc situation.. maybe there is a better way to trigger it based on rootfs being mounted..

13:10 <robclark> https://www.irccloud.com/pastebin/UNZPDueC/

13:35 <dgilmore> robclark: cheers, that is probably a better solution

13:52 <robclark> dgilmore: maybe there is some way to turn that into an rpm.. I guess the script part could check for the existence of the sysfs file so it should be safe to install anywhere?

13:58 <dgilmore> robclark: wouldn't be hard to do. I wonder if we couldn't change the driver to try again. I thought that's what's done with the GPU

14:05 <robclark> for the GPU, we have userspace opening /dev/dri/whatever to trigger trying again.. for remoteproc, I don't think there is a useful trigger to hook onto

14:05 <robclark> (although I wish iris would do similar to the gpu, that would avoid power draw when iris is enabled but fw is missing)

14:40 <maz> mini-x landed. cracked it open, very close to the devkit. all 3 m.2 connectors are there. also, unpopulated x8 slot, jtag and uart.

14:45 <maz> PCB pics: http://www.loen.fr/tmp/mini-x/

14:45 <maz> I'll switch it on once I have the UART populated! :)

14:50 <robclark> nice

15:13 <xeliteCJF> <maz> very nice! I fly home to Hong Kong Saturday, will have a look what is the price tag for it. I like to retire my intel workstation for it.

15:19 <maz> xeliteCJF: it was GBP480 last week on the Lenovo web site, but it's now back to about 600, which is pretty steep. hopefully you can score something at a lower price in HK.

16:01 <xeliteCJF> HK has no VAT. That alone makes a difference

18:49 <zdykstra1> it's very expensive in the US, too. I think ~800 or ~900 when I last checked.

18:51 <JensGlathe[m]> My neighbour spent ~890€ incl Win11 Pro

18:59 <maz> OK, we have a UART! (dinner time)

19:01 <JensGlathe[m]> neededto solder?

19:01 <JensGlathe[m]> * needed to solder?

19:30 <clover[m]> looks like they finally released bluez 5.84 an hour ago. lets hope this fixes the bluetooth issue

19:31 <xeliteCJF> <maz> well done!

19:32 <xeliteCJF> Nothing is better then a serial console to dig into it. That is worth it

19:42 <maz> JensGlathe[m]: yup. the pads are under the CPU heatsink, so a bit tricky to route things. I've botched it for now, but I may come back to it at some point.

19:43 <maz> I've only connected RX so tar, until I can convince myself that the levels are indeed 3.3v.

19:43 <maz> so far*

19:45 <maz> F1 gets you into a PC-like setup, from which you can disable secure boot. it feels like a PC (terrifying!)

19:45 <JensGlathe[m]> Hah

19:52 <maz> serial console is just as broken as on the devkit (unsurprisingly). all the control characters are ignored, but grub manages to output something!

19:53 <JensGlathe[m]> why not use kb and diskplay

19:53 <JensGlathe[m]> s/diskplay/display/

19:55 <maz> because I don't have a display to attach to this machine, and I definitely don't want to stick a display in the rack.

19:56 <maz> ah. in the fine QC fashion, ExitBootServices is immediately followed by a reset.

19:58 <maz> JensGlathe[m]: I appreciate that my usage of these machines is a bit unconventional.

20:16 <valpackett> @_oftc_xeliteCJF:matrix.org: make sure speakers aren't disabled in the bios?

20:19 <valpackett> <travmurav[m]> "but now after I turned it off..." <- oh yeah many knobs kinda just don't apply at runtime correctly and it does seem random sometimes

20:19 <valpackett> that's why it took me some time to discover RX_COMP

22:36 <clover[m]> archlinux/archiso has an experimental branch that supports other architectures now. Happy to report that I could get it booting on my x13s with a minimal patch

22:37 <clover[m]> Still lacking device tree support but huge progress