robclark changed the topic of #aarch64-laptops to: Linux support for AArch64 Laptops (Chrome OS Trogdor Devices - Asus NovaGo TP370QL - HP Envy x2 - Lenovo Mixx 630 - Lenovo Yoga C630 - Lenovo ThinkPad X13s - and various other snapdragon laptops) - https://oftc.irclog.whitequark.org/aarch64-laptops
<dgilmore> One glaring issue to resolve
todi1 has joined #aarch64-laptops
todi has quit [Ping timeout: 480 seconds]
hexdump0815 has joined #aarch64-laptops
hexdump01 has quit [Ping timeout: 480 seconds]
<Jasper[m]> <dgilmore> "One glaring issue to resolve..." <- What did it end up being?
<travmurav[m]> HdkR: btw I see no rendering bugs with the aarch64/gles version of beatsaber
<travmurav[m]> apparently the version I had was the same as the one you tested lol
jhovold has quit [Ping timeout: 480 seconds]
xyaon has joined #aarch64-laptops
_alice_ is now known as _alice
f_ has quit [Read error: Connection reset by peer]
f_ has joined #aarch64-laptops
Lucanis has quit [Ping timeout: 480 seconds]
Lucanis has joined #aarch64-laptops
Kelsar_ has quit []
Kelsar has joined #aarch64-laptops
baozich has joined #aarch64-laptops
jkm has joined #aarch64-laptops
<robclark> dgilmore: the json files in /lib/firmware are actually read by userspace (pd-mapper I think).. I think there was a patch to make it understand compressed fw, otherwise those files need to be uncompressed
<Jasper[m]> <robclark> "dgilmore: the json files in /lib..." <- The json files are not compressed, the firmware is
<xyaon> besides acpi tables, where else could i find any info on gpio pin bindings? are most parameters that are in the devicetree hardcoded in windows drivers or it just works in a different way there?
<xyaon> i would be thankful for any kind of resource on how this kind of porting works
<robclark> Jasper[m]: ok.. the fedora qcom-firmware package had them compressed, or at least it did when I was setting up my x13s.. I had to manually decompress 'em
<Jasper[m]> robclark: Oh, interesting. I could concatenate them fine
baozich has quit [Ping timeout: 480 seconds]
baozich has joined #aarch64-laptops
<dgilmore> robclark: yeah, the issue is not the compressed firmware, pd-mapper reads them fine. It's that the remoteproc symlinks are not created so the jsn files can not be found
baozich has quit [Ping timeout: 480 seconds]
<dgilmore> robclark: pd-mapper in fedora has the patches applied to read the compressed files
<dgilmore> Jasper[m]: not sure the cause yet, but that the symlinks for remoteproc is the problem
laine has joined #aarch64-laptops
<robclark> ahh, ok
<Jasper[m]> <dgilmore> "https://www.irccloud.com/..."; <- I remember my dmesg having the same lines
<dgilmore> qcom_q6v5_pas is not being loaded
<dgilmore> with qcom_q6v5_pas installed pd-mapper starts fine
<dgilmore> I am going to make sure it is in the initd and see if it gets autoloaded
<Jasper[m]> dgilmore: It was one of the steps to make usb boot work iirc, maybe it got copied over to the installed system?
<dgilmore> Jasper[m]: /etc/modprobe.d/anaconda-denylist.conf:blacklist qcom_q6v5_pas
<Jasper[m]> I'll check it out in a sec
<dgilmore> so anaconda while not doing the bootargs right blacklisted the module
<dgilmore> I removed /etc/modprobe.d/anaconda-denylist.conf and rebuilt the initrd to make sure it was not in there also adn rebooted, pd-mapper started fine on boot
<steev> usb boot still works though?
<Jasper[m]> steev: Is this a remark or a question? :p
<steev> it was a question for dan
<dgilmore> steev: probably not, but I am using an install on NVMe
<dgilmore> the issue with usb booting is that the stroage gets reprobed and /dev/sda becomes /dev/sdb and everything gets unhappy
xyaon has quit [Remote host closed the connection]
<dgilmore> https://paste.centos.org/view/99a6cde1 is dmesg on boot with the latest f39 kernel. I should boot from the rawhide kernel again
<dgilmore> I will try boot the installer without adding the modprobe.blacklist line
<dgilmore> well livecd
<Jasper[m]> Aww, firmware is still not automagically getting my dtb from the ESP
<Jasper[m]> Firmware booting grub is also a bit of a coinflip, it seems to hang on the Lenovo logo from time to time
<Jasper[m]> <dgilmore> "Jasper: /etc/modprobe.d/..." <- Removing this worked for me
<Jasper[m]> I'm on 6.7.0-rc5
<dgilmore> Jasper[m]: did you tell the firmware to use devicetree?
<Jasper[m]> dgilmore: What do you mean? The Linux toggle is on if that's what you mean
<dgilmore> yes
<Jasper[m]> I have left the dtb in $ESP/dtb/qcom/sc8280xp-lenovo-thinkpad-x13s.dtb
<dgilmore> that is incorrect. needs to be $ESP/sc8280xp-lenovo-thinkpad-x13s.dtb
<Jasper[m]> Okay I inferred that it had to be somewhere on the esp, I'll symlink it from the other path
<steev> how do you symlink on vfat?
<Jasper[m]> Does reboot work properly for you? @dgilmore
<Jasper[m]> steev: idk, thought it worked last time I tried, I might be wrong
<dgilmore> Vfat doesn't support symlinks
<dgilmore> Reboot works
<HdkR> travmurav[m]: I guess it was the particular version of turnip that I was running
<Jasper[m]> dgilmore: Well whatever Fedora creates does work with symlinks
<Jasper[m]> The "dtb" folder is one I think, it points to the "dtb-kernelversion" folder
<dgilmore> Jasper[m]: that is not on the ESP
<dgilmore> it is in /boot
<Jasper[m]> Oh wait, you're right. I'm very much used to the old (read: windows) way of doing things
<Jasper[m]> Sorry for that
<dgilmore> cp /boot/dtb-6.7.0-0.rc5.20231215git3f7168591ebf.45.fc40.aarch64/qcom/sc8280xp-lenovo-thinkpad-x13s.dtb /boot/efi/
<dgilmore> that will put it in the right place
<Jasper[m]> dgilmore: How do you recommend keeping that updated?
<dgilmore> occasionally copying the latest over
<dgilmore> it should not change that much
<Jasper[m]> dgilmore: :/
<Jasper[m]> Oh well
<dgilmore> I had not updated the dtb on the ESP since putting it in place until yesterday
<Jasper[m]> <dgilmore> "Reboot works" <- I didn't notice grub had no timeout. The reboot problem/hang on boot I had was due to it failing to boot initially
<Jasper[m]> And then it does show the menu
<Jasper[m]> dgilmore: Guess I'll have to just check for changes upstream from time to time haha
<dgilmore> maybe not ideal but dtbs should probably be removed from the kernel now as they really are stable
<dgilmore> and ideally the firmware actually provides it
<dgilmore> Jasper[m]: maybe when there is a major update copy it, i.e when moving from 6.7 to 6.8
<Jasper[m]> Yeah probably
<Jasper[m]> @dgilmore do you also have dmesg saying that it failed to load some gpu related .mbns? Or do you think that's pre-decryption?
<Jasper[m]> Also a thing with pmic-glink and panel_edp doing an oops
<dgilmore> it does get loaded, looks like it looks in a few locations
<steev> the gpu related stuff is normal, as the gpu firmware is rarely in the initramfs
<dgilmore> that is the full dmesg of 6.7.0-rc5
<Jasper[m]> steev: Ah yeah, thought so.
<Jasper[m]> You can see the panel-edp thing in there
<steev> do we now have a third panel?
<steev> bamse: ^^
<Jasper[m]> I don't think it did that before
<steev> i thought we had timings for both panels in the kernel already
<Jasper[m]> Also sound is still very much improperly configured, but I guess I need to wait for a new alsa-ucm-conf version for that to be addressed
<dgilmore> Jasper[m]: I have a slightly newer UEFI firmware
<Jasper[m]> dgilmore: You reckon that'd help? Don't think much changed there.
<bamse> steev: that's indeed a missing panel definition in the panel-edp driver
<steev> Jasper[m]: can you give me the edid output for your panel?
<steev> bamse: i thought we had the ivo and the boe in there
<Jasper[m]> steev: Got a command for me to fetch that? Hahaha
<dgilmore> does not seem like there is a relevant fix
<bamse> steev: we dpm
<bamse> steev: we don't have a BOE 0x964...
<steev> Jasper[m]: `edid-decode < /sys/class/drm/card0-eDP-1/edid` should work, if you have the edid-decode binary from.. well in debian it's edid-decode
<Jasper[m]> Same for fedora I think, already had a different package installed hah
<Jasper[m]> One sec
<dgilmore> Jasper[m]: looks like https://github.com/alsa-project/alsa-ucm-conf/commit/ddf329098b4427851465709d091f8d8770c7c130 from upstream is not in a released versions of alsa-ucm
<Jasper[m]> (cc: @bamse) recommend saving that^ locally since the paste disappears after a day
<Jasper[m]> dgilmore: Yep! Don't know how they schedule their releases
<steev> dgilmore: that is correct, once 1.2.11 is out, it should be
<dgilmore> steev: 1.2.10 is the latest tagged version
<steev> dgilmore: correct
<steev> i have no idea what their release cadence is, unfortunately, but that commit is indeed needed
<dgilmore> I might just add the patch
<Jasper[m]> dgilmore: I guess copying over the folder would work? It's mostly just config files after all
<dgilmore> Jasper[m]: you could make the two changes manually
<Jasper[m]> Tried that, speaker volume didn't really change, earbuds are now not eardeafening but you have to put it on 100% to hear something
<Jasper[m]> A sidegrade
<Jasper[m]> dgilmore: It's 4 or 5 commits I think
<Jasper[m]> firefox also still doesn't like pipewire hahahaha
<dgilmore> probably do want all 5 commits
<Jasper[m]> I just cloned the repo and copied over everything
<Jasper[m]> It changed things, not entirely ideal, but yeah
<Jasper[m]> <steev> "bamse: i thought we had the..." <- I checked psref real quick, there's three panel options. 2 300nits panels (one of which touch) and 1 400nits panel. I don't have touch so I guess it might be the 300 nits panel?
hexdump0815 has quit [Quit: WeeChat 3.8]
hexdump0815 has joined #aarch64-laptops
<HdkR> How soon until we get a Snapdragon laptop with HDR support? :P
<steev> Jasper[m]: https://paste.debian.net/hidden/e473bfd7/ you can try this to get rid of the splat, but i'm not sure the timings are fully correct, just a guess
<Jasper[m]> <steev> "Jasper: https://paste.debian.net..." <- Ouch, gonna have to do the work to package that for fedora
<Jasper[m]> Ehhh, I'll save it somewhere, maybe I'll come back to that
jkm_ has joined #aarch64-laptops
jkm has quit [Ping timeout: 480 seconds]
<steev> ? you should be able to just drop a patch in; or i can
<steev> i don't know what fedora's git is based on though
<steev> but since i don't have that panel, need you to test it :P
<steev> conceivably... the only change you should see is the splat go away
<dgilmore> Jasper[m]: instead of adding any modues to the initrd you can just add rd.driver.blacklist=msm to delay setting up video and it all works.
jkm_ is now known as jkm
<Jasper[m]> <steev> "i don't know what fedora's git..." <- I don't have experience with that either hahahaha, probably hosting a copr repo is the way to do it
<steev> you shouldn't need an entire repo just to add a patch though
<Jasper[m]> steev: You'd need to rebuild the kernel though, no?
<steev> sure, but doesn't fedora let you build packages outside the repos?
<steev> this is just a one off test, not something to continue to carry forward
<Jasper[m]> steev: Haven't looked into how that works, but probably
<steev> or just compile a custom kernel to test it
<Jasper[m]> Yeah...
<Jasper[m]> Okay I'll look into it
<dgilmore> Jasper[m]: https://fedoraproject.org/wiki/Thinkpad_X13s if you could update for anything not clear or needing extra details would be great
<steev> it is nice to see more sc8180x work coming in... i should see how my flex5g is doing, i haven't booted it in a long time :(
<dgilmore> I should spend some of my holiday checking on the yoga c630
<Jasper[m]> <dgilmore> "Jasper: https://fedoraproject...." <- I did not have to regenerate dracut after removing the module blacklist. Audio is quiet on speakers, but loud on headphones (both are quiet with an updated alsa-conf-ucm). Might be worth noting that you need to do the modprobe.blacklist argument in addition to adding the dtb to the esp of the usb drive.
<Jasper[m]> I don't have an account yet so I'll do it tomorrow if you haven't already