ChanServ changed the topic of #linux-msm to:
Danct12 has joined #linux-msm
Daaanct12 is now known as Daanct12
Daaanct12 has joined #linux-msm
Daaanct12 has quit [Remote host closed the connection]
Daaanct12 has joined #linux-msm
Daanct12 has quit [Ping timeout: 480 seconds]
marvin24 has joined #linux-msm
marvin24_ has quit [Ping timeout: 480 seconds]
svarbanov_ has quit [Remote host closed the connection]
svarbanov_ has joined #linux-msm
Danct12 has quit [Quit: WeeChat 3.8]
<DylanVanAssche> srinik: Any chance you could have a look at https://lore.kernel.org/linux-arm-msm/20230327184204.498032-1-me@dylanvanassche.be/T/#t since the merge window of 6.3 is closing fast 🙂 ?
ruisilva has joined #linux-msm
<flamingradian[m]> patches are not normally applied during merge windows from my exlerience
ruisilva has quit [Quit: WeeChat 3.8]
ruisilva has joined #linux-msm
ruisilva has quit [Quit: WeeChat 3.8]
<steev> DylanVanAssche: is there a v3 that doesn't have the debug prints re: caleb's request?
<DylanVanAssche> steev: Not yet as that request just came in, I have to spin up a v3 still.
<steev> oh got ya, i didn't notice the date :)
<steev> i'm always interested in 845 stuff because of my c630
<DylanVanAssche> steev: Ah yes, the SDM850 chipset?
<steev> i love the thinkpad, don't get me wrong, it has the superior keyboard, but even though the c630 is like 5 years old, it's battery life is still godly
<DylanVanAssche> Does it has sensors actually like accelerometer, proximity, etc. ?
<steev> it should - calebccff mentioned some thing the other day about it, and pointed at some repo on gitlab/postmarketos, but i haven't had a chance to look into it fully
<DylanVanAssche> steev: Well if you want to try it out and are lost in all the parts you need, feel free to ask 🙂
<DylanVanAssche> You need patches in several projects ranging from the kernel to libqmi, iio-sensor-proxy, firmware files etc.
<steev> the firmware files are likely in our UFS
<DylanVanAssche> Yes, normally in the vendor and DSP partitions
<calebccff> steev: you already found the "firmware files", the JSON ones at least, those should be all you really need... although calibration might be a bit off heh
<calebccff> oh slpi.mbn is needed too...
<steev> slpi mbn i already have in place, but i don't recall the patchset
<DylanVanAssche> calebccff: There's also some .so libraries that are uploaded to the DSP via FastRPC. Related to CHRE from Google/Android. Not sure if that applies to c630 as well
<steev> we do have some of those somewhere in the c630 stuff
<calebccff> \o/ oh nice
<steev> https://github.com/steev/linux/commits/c630-next-20230330 looks like i'd applied a portion of it to next 20230330
<steev> so i'll grab those, apply mine, and... then start digging into userspace i suppose :)
<steev> calebccff: that doesn't work :(
<steev> or maybe it does
<steev> if they aren't applied to 6.3 yet, i guess they might
<steev> i just know my audio on my c630 with 6.3.1 is still quite poppy
<calebccff> Dylan has been trying to migrate us over to pipewire on the phones, iirc it's almost all working? we don't have the WSA codec though so things might be different
<steev> let me yank those back in too... i dropped pretty much everything from 6.3 to do some other testing and seem to have lost the work :(
<steev> we are using pipewire in kali, so it could very well be
<steev> [ 0.000000] OF: reserved mem: 0x0000000096700000..0x0000000097afffff (20480 KiB) nomap non-reusable slpi@96700000
<steev> [ 4.194607] remoteproc remoteproc2: Booting fw image qcom/sdm850/LENOVO/81JL/qcslpi850.mbn, size 6783124
<steev> it's at least loading the slpi firmware :)
<steev> DylanVanAssche: pointers towards the userland stuff? :)
<calebccff> steev \o/
<calebccff> steev: this should have everything you need https://gitlab.com/postmarketOS/pmaports/-/merge_requests/4050
<steev> everything built here, except libssc, because of an undefined reference to atan2
<DylanVanAssche> steev: ah I may need to add a dep on a math lib
<steev> yeah, it just needs -lm added
<DylanVanAssche> Weird that I don't have it locally
<steev> maybe because i'm on gcc12 and it's a lot stricter?
<steev> so now i just gotta sort out the firmware... or just grab the axolotl and see if it does stuff even if it's slightly off
<konradybcio> steev if you feel like experimenting, you may do maxcpus=4 on the thinkpad, A78Cs should eat much less power than the X1Cs
<lumag> Stupid question, is there any out-of-tree fork for 8974? I'm trying to bring up the apq8074 dragonboard.
<DylanVanAssche> steev: ah that could be... Will need to add it anyway as a dependency
<lumag> aka_[m], thx
<DylanVanAssche> steev: If you run hexagonrpcd it will complain when the DSP asks for a file you haven't provided
<leezu> abhinav, lumag: 6.3 breaks external display support (USB-C DP) on lazor / sc7180 for me. According to reports on #aarch64-laptops, it still works on aspire1, which is also sc7180 based. The key difference is that HPD detection for type-c is done by the EC event and special driver on aspire1 but mdss dp_hot_plug_det on lazor. Are you aware of this issue / do you have an idea what
<leezu> may be going on?
<DylanVanAssche> I would use the device specific firmware as they are specific for the hardware
<steev> DylanVanAssche: yeah, its more for... seeing what i should even have :)
<steev> i need to reboot it into windows and reinstall wsl to grab all the files again, i seem to have lost them at some point
<lumag> leezu, my first and semi-random suggestion would be to try reverting e17af1c9d861dc177e5b56009bd4f71ace688d97
<steev> still sad they never gave us wsl2 on the c630
<lumag> steev, why do you need wsl? I think all the DSP files are accessible from win partition
<leezu> lumag: Thank you. Let me try that
<lumag> In the worst case you might have to install ntfs-3g plugin for compressed files
<steev> last time i checked, ntfs-3g doesn't work with latest windows 10/11
<lumag> steev, I was able to get it with the external plugin
<steev> which external plugin? because iirc, this also happened with flex 5g
<lumag> I think it was this one
<lumag> I don't have c630 at hand :-(
<steev> will give that a spin later :)
<steev> ohh, i thought you had
<lumag> Not at this moment
<lumag> I will be able to check it later today
<steev> i still need to go back and see if i can figure out why display off in gnome = reboot
<lumag> And I had few issues with 6.3
<steev> my remoteprocs crashed the other day, but i have no idea when or why
<lumag> Mine doesn't boot with 6.3.
<lumag> aka_[m], it seems, BAM_DMA is broken on my 8974. Am I unlucky or is it a generic issue?
<aka_[m]> lumag: broken how exactly
<konradybcio> minecrell: ^
<aka_[m]> on 8976 requires flag for remote-control
<aka_[m]> otherwise reboot system
<lumag> aka_[m], the board stops booting early in the process
<lumag> aka_[m], doen't seem to help
<aka_[m]> i don't even have 8974 device, z3ntu you had fp2 right
<z3ntu> I have a bunch of 8974 devices, yeah
<z3ntu> the 6.3 branch on https://github.com/msm8974-mainline/linux works on all devices we can test
<steev> konradybcio: oh, that's a good point... but what if i want it to just be X1Cs?
<konradybcio> bad idea :P
<konradybcio> you boot on the A78C and hyp may be running on there so they will probably not be 100% off
<konradybcio> good morning Marijn
marvin24_ has joined #linux-msm
<steev> lumag: fwiw, this https://github.com/steev/linux/tree/c630-v6.3 is the tree i'm booting on my c630, with http://sprunge.us/W9fWtJ as my config
<lumag> steev, I will take a look later
<lumag> thank you
<lumag> z3ntu, I had to disable BAM_DMA config option
marvin24 has quit [Ping timeout: 480 seconds]
<lumag> z3ntu, this is my log with the bam-dma: https://pastebin.ubuntu.com/p/5cBHZm7fXX/
<lumag> with this option disabled, it boots fine
<z3ntu> hm weird, haven't seen that before
<z3ntu> also 5.9.0 is a bit bit old ^^
<lumag> z3ntu, same issue with 6.3, just didn't get the backtraces
<z3ntu> hm okay
<z3ntu> but seems maybe dragonboard is a bit special with dma? see https://github.com/msm8974-mainline/linux/commit/338d51889851cb2ac900d2ec75c8185afcde8caf
<z3ntu> (that commit was reverted again so the node is back in)
<lumag> z3ntu, thanks for the pointer, I will take a look at the downstream if there are any additional hooks there
<lumag> ok, probably I did something wrong the first time
<lumag> qcom,controlled-remotely fixes the issue
<lumag> z3ntu, thanks!
<steev> DylanVanAssche: in https://gitlab.com/sdm845-mainline/firmware-shift-sdm845/-/blob/main/etc/qcom/sns_reg.conf - 2-7 were in my qcsensorsconfigcls850.inf_arm64_a4738dda00d434ba on the c630, but 8-10, i have no idea where they are
<steev> should i just copy 2-7 into /etc/qcom and put that as the place where it can find them? since we don't seem to show up in sys
<steev> Invalid FastRPC node: (null)
<steev> right now, all i'm getting though is steev@limitless:~/firmware$ hexagonrpcd
<steev> flamingradian[m]: this isn't an android system though
jhovold has quit [Ping timeout: 480 seconds]
<flamingradian[m]> steev: oh, in that case, the config that you linked might work
<steev> we're just straight linux... and i'm just pulling the firmware file(s) and seeing if it'll even remotely work on this system
<steev> but, dylan said that hexagonrpcd should spit out the files its expecting, so i think i'm missing something since i get the invalid fastrpc node
<flamingradian[m]> probably just command line stuff: -s -f /dev/fastrpc-slpi
<steev> ah, i don't have a fastrpc-slpi node
<steev> only have adsp, cdsp, cdsp-secure, and sdsp
<steev> i also don't have the libchre stuff
<flamingradian[m]> oh, sorry, it's -s -f /dev/fastrpc-sdsp
<lumag> z3ntu, hmm, otg is also semi-broken, only gadget works
<lumag> do you know if that works for other 8974 platforms?
<steev> flamingradian[m]: https://paste.debian.net/1279543/ it's just sitting there after those lines
<flamingradian[m]> steev: right, it needs /usr/lib/qcom/adsp and /usr/lib/qcom/adsp/avs to exist
<steev> none of those files exist in the windows firmware
<flamingradian[m]> just the dir needs to exist
<z3ntu> lumag: last I tried, maybe a year ago it worked on fp2
<z3ntu> lumag: but didn't try much honestly since upower or some other user space component shuts down the system very quickly after unplugging charger since there's no fuel gauge driver yet for 8974 (BMS driver)
<steev> flamingradian[m]: https://paste.debian.net/1279544/ - tried throwing the files into there
<steev> i guess i need to actually dig around my ufs partitions and see if i can find those files
<flamingradian[m]> actually, chre doesn't matter for most sensors and I removed it in a feature branch
<flamingradian[m]> (or rathet, separated it)
<steev> oh, i found some of them
<Marijn[m]> <z3ntu> "lumag: but didn't try much..." <- There are some patches on upower to disable sleep/hibernate actions on critical battery (useful on my laptop where drawing much power turns its status into "unknown"), but none were accepted yet
<lumag> ok, and now a hangcheck for a330
<aka_[m]> Weird, I had no fg added to my msm8976 device and it works just fine, phosh only shows red battery indicator
<steev> okay, the system isn't angry at me at all, i ended up finding the acdb files, though none of them are prefixed with QRD like in the android firmware
<aka_[m]> Do mainline support adcb?
<aka_[m]> Aren't these some calibration files for audio?
<flamingradian[m]> maybe in the future it will
<steev> i'm just making sure all the stuff i have is in place where it is in the firmware repo i'm looking at, is all
leezu has quit [Quit: WeeChat 3.0]
leezu has joined #linux-msm
marvin24 has joined #linux-msm
marvin24_ has quit [Ping timeout: 480 seconds]