05:25 <steev> ooh, sdm845 is getting love

06:10 <steev> aha, i found it, just gotta figure out what the correct way to fix it up is

06:28 <steev> i had tried to pull in bamse's v5 of the stuff, and you had modified one of the previous versions

06:28 <steev> you being jhovold :)

06:29 <steev> https://patchwork.kernel.org/project/linux-arm-msm/patch/20221207220012.16529-11-quic_bjorande@quicinc.com/ this one

07:33 <jhovold> steev: yeah, the rebase on -next required some manual touch ups. I'll post my next branch this week if you want to use that as a base.

09:47 <jhovold> steev: I just pushed a branch based on next-20221213 which you may want to rebase on. I've included qzed UEFI variable work as it seems we will be using this in some form after all.

09:48 <jhovold> https://github.com/jhovold/linux/tree/wip/sc8280xp-next-20221213

15:43 <steev> jhovold: sounds good, will look around

18:13 <steev> jhovold: just a quick test with your config on that branch and i get stuck somewhere in boot, didn't look closer as i need to focus on work things now

19:12 <mjeanson> Hi, is anyone working on the Windows Dev Kit 2023? It uses the same SoC as the x13s, I've been able to start a linux kernel in ACPI mode and with a minimal dtb but both result in a reboot, without a serial port it's hard to debug

19:12 <mjeanson> There might be a UART hidden somewhere on the motherboard

19:13 <steev> likely the iommu is kicking in, and since it doesn't know about the dev board, boom

19:21 <mjeanson> There is a patch adding an entry for it in qcom-smmu at least in ACPI mode, but that might indeed be insufficient

19:21 <mjeanson> https://github.com/linux-surface/kernel/commit/a38ad65d4b42477078d34ea236686007de040743

19:27 <steev> are you using that kernel?

19:28 <mjeanson> I tried this one and your repo with the previously mentionned patch on top

19:29 <steev> maybe a video of it booting and watch in slomo after to see?

19:30 <steev> that is what i've been doing when it comes to the thinkpad

19:30 <mjeanson> Yeah I think that's what I'll have to resort to

19:42 <steev> i suppose i could grab that patch, i'm still on the fence about a devkit

19:48 <mjeanson> It's a nice device, reasonably priced for the specs but it lacks the debugging features of a real devkit, like a UART

19:57 <Leandro[m]> <mjeanson> "It's a nice device, reasonably..." <- Welp time to hunt it down on the production board

19:57 <Leandro[m]> Probably some testpoints or unpopulated connect

20:00 <mjeanson> There is many unused connectors on the board, it's probably a surface pro 9 mainboard without the peripherials

20:00 <mjeanson> The connectors are tiny, I'm not equipped to probe them efficiently

20:01 <Leandro[m]> Wait which device are you guys talking about again?

20:01 <mjeanson> This one https://www.microsoft.com/en-us/d/windows-dev-kit-2023

20:02 <Leandro[m]> Ah

20:03 <Leandro[m]> No schematics leaked yet 😔

20:40 <hexdump0815> mjeanson: here someone else already tried to get it booting, but looks like they gave up for now: https://github.com/linux-surface/surface-pro-x/issues/43

21:29 <mjeanson> yeah I saw this, he's the author of the smmu patch

21:30 <mjeanson> do you have a devkit too?

21:44 <hexdump0815> mjeanson: no - was just interested to see if it is possible to get it working ... still have a samsung galaxy book go around to get working with linux first :)

22:45 <bluerise> steev: The workaround on qcom smmu to check the last domain or so to check if it gets rewritten doesn't work on the x13s/windev kit... at least not on OpenBSD

22:45 <bluerise> I have OpenBSD running on the windev kit, mjeanson, and my issues were mostly related to smmu as well

22:45 <steev> bluerise: oh nice

22:46 <steev> bluerise: when are you gonna hook me up with openbsd for the thinkpad?

22:46 <bluerise> The thing is: the ACPI tables are broken and unusable

22:46 <bluerise> so for x13s we're using device trees already

22:47 <bluerise> but we can still boot with ACPI... for the bare minimum

22:47 <bluerise> steev: 'hook you up'? :)

22:47 <steev> can i download 7.2 and install it on my thinkpad?

22:47 <bluerise> it's not as nice as Linux though. no GPU driver, no wifi,... sigh.

22:47 <bluerise> I'm using a USB WiFi dongle

22:47 <bluerise> steev: I'd propose snapshot, not 7.2 release

22:47 <steev> hey, if its any consolation, we don't have GPU on linux either :P

22:48 <bluerise> no remoteprocs :P yet

22:48 <bluerise> I have a patchset for the UEFI variables over TZ app for the real time clock

22:48 <steev> no remoteproc needed for wifi

22:48 <bluerise> I know, but for the battery management

22:48 <steev> oh... yeah that... thats a bit rough

22:48 <bluerise> yup yup

22:49 <bluerise> if you wanna give it a shot, you can install using a current snapshot http://cdn.openbsd.org/pub/OpenBSD/snapshots/arm64/miniroot72.img

22:49 <bluerise> try it on a usb stick, or the nvme if you don't care about what you have on it

22:51 <steev> i could definitely give it a whirl on a usb stick over the christmas holidays

22:51 <steev> nvme has my windows and linux installs ;)

22:51 <bluerise> hehe

22:51 <steev> i'm hoping santa will hook me up with a tricked out thinkpad and i can actually do destroying things on it

22:52 <bluerise> I replaced Windows/Linux with OpenBSD, as I have no use for the other two on that really... and I can do BIOS updates without Windows :D

22:52 <bluerise> On the windev, I still have the internal NVMe with Windows... only way to get BIOS updates

22:55 <steev> unfortunately... as a linux distro dev... i kinda need linux ;)

22:55 <bluerise> :D

22:56 <mjeanson> So ACPI on the windev is a dead end?

22:57 <bluerise> It appears so. From what I have been told, the windows drivers ship ACPI table overlays because the ACPI tables themselves are sparse (and even incorrect), and each driver ships an ACPI table parser for its overlay.

22:58 <bluerise> I don't see that situation improving anytime soon.

22:58 <qzed> Until someone (maybe you? :) ) volunteers for implementing ACPI PEP support, these things need a DT...

22:58 <bluerise> It's good enough for PCIe and USB (like the ethernet, NVMe, WiFi probably), but apart from that...

22:59 <bluerise> The only question is: who can write a dts for it?

23:01 <steev> so you'd need to RE the windows driver for its acpi overlay?