ChanServ changed the topic of #aarch64-laptops to: Linux support for AArch64 Laptops (Asus NovaGo TP370QL - HP Envy x2 - Lenovo Mixx 630 - Lenovo Yoga C630)
<steev> can you show dmesg output
<steev> i see i2c-hid and that *should* cover some things
<pierro78> keyboard working fine if I boot without lightdm
<steev> sounds like maybe you don't have libinput installed
<steev> xserver-xorg-input-libinput
<pierro78> and also I can boot with an empty usb slot (which was not booting before)
<pierro78> xserver-xorg-input-libinput : got following error when trying to apt install : libwacom2 : Depends: libwacom-common (= 1.8-2) but 2.4.0-2 is to be installed
<pierro78> OK I am rebooting with lightdm and send you dmesg
<jenneron[m]> pierro78: please upload dmesg to dpaste.com or something similar
<jenneron[m]> yeah
<jenneron[m]> i'm surprised it boots with ACPI when 2 USBs are busy, we can take some useful information from there
<jenneron[m]> i don't have a second otg adapter though
<pierro78> thanks*
<jenneron[m]> your usb ethernet adapter needs firmware installed
<jenneron[m]> it also wants a qcmpss8180_nm.mbn, but i don't have one in my windows installation
<pierro78> oh it doesn t draw power from my charger (my usb cable show the consummed wattage )
<jenneron[m]> i haven't looked into battery and charger so far
<steev> jenneron[m]: are you sure? it should be a qcsubsys_ext_mpss8180.inf_arm64_<commit> folder in the FileRepository
<pierro78> OK 2:15 am ... need some sleep ...
<steev> jenneron[m]: for battery/charger on my flex5g which is also sc8180x i had to use https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/qcom/LENOVO/21BX/battmgr.jsn
<jenneron[m]> pierro78: try to rename /lib/firmware/qcom/sc8180x/sm-w767/qcmpss8180_XEF.mbn to /lib/firmware/qcom/sc8180x/sm-w767/qcmpss8180_nm.mbn
<steev> jenneron[m]: if its actually called XEF on your system, you could always edit that in the dts file to point to that instead of _nm.mbn
<jenneron[m]> steev: should i do that? we can just call this _nm.mbn in firmware package
<jenneron[m]> but anyway renaming now is easier to test
<jenneron[m]> pierro78: probably not rename, but copy it, just in case
<steev> if it's XEF on the actual system, might be easier if/when they release updates
<jenneron[m]> i see
<jenneron[m]> thanks for the tips, but where i can find a battmgr.jsn or a reference for it?
<steev> but it's up to you, since you're writing the dts :)
<steev> i just linked you a lil bit ago
<steev> i just dropped that into my LENOVO/82AK directory, and it picked it up
<jenneron[m]> yeah but this config is lenovo flex 5g specific, isn't it?
<steev> no
<steev> the 21BX is the thinkpad
<jenneron[m]> ok, thanks
<steev> 81JL = c630, 82AK = flex5g, 21BX/21BY = thinkpad x13s
<pierro78> still no touchpad / keyboard ... but now it s drawing power from charger !
<steev> if the keyboard/touchpad aren't working in lightdm, but they do on the console, it means you're missing a driver
<steev> userspace, not kernel
<jenneron[m]> pierro78: please dmesg with copied firmware
<steev> i'd check /var/log/lightdm/* and /var/log/Xorg*
<pierro78> (new dmesg)
<jenneron[m]> hm, i probably need to adjust reserved memory
<pierro78> the strange thing is keyboard and touchpad were working when booting from internal ssd ...
<jenneron[m]> at least it worked in sway for someone who tested kernel for me before
<pierro78> it s like usb is not working anymore when booting from usb
<jenneron[m]> if keyboard works in console, then it's a userspace issue
<pierro78> (or maybe not : keyboard works when no lightdm)
<jenneron[m]> you should check xorg logs
<jenneron[m]> it's probably better to avoid xorg actually
<jenneron[m]> you should have working gpu now, despite that error
<pierro78> had a look at /var/log/lightdm/* and /var/log/Xorg* ... no special error ...
<pierro78> OK I have removed lightdm ... keyboard working ... charger charging ...
<pierro78> there was an issue with gdm3 ... how can I start wayland with gdm3 ?
<pierro78> withOUT gdm3 ?
<jenneron[m]> which issue?
<pierro78> black screen
<jenneron[m]> with dtb?
<pierro78> with like a small "dash" cursor at the top left
<pierro78> I can try again
<jenneron[m]> i'm not sure where you can find gdm logs
<pierro78> same black screen issue ...
<jenneron[m]> do you want to run gnome or what?
<pierro78> I wanted to run wayland instead of xorg ?
<jenneron[m]> but which DE
<jenneron[m]> if you want to run it without gdm, there will be different commands for various DEs. i think it won't work anyway, but at least it will print logs to stdout
<pierro78> I have gnome
<jenneron[m]> for gnome run `gnome-shell --wayland` in console
<pierro78> that s the debiancdimage default DE
falk689 has quit [Remote host closed the connection]
falk689 has joined #aarch64-laptops
<pierro78> run `gnome-shell --wayland` in console => wow : keyboard and touchpad working !!!
<pierro78> and charger still charging !!!
<pierro78> (I think it was not charging when lightdm/xorg was started)
<jenneron[m]> so, gnome works, right?
pierro78_ has joined #aarch64-laptops
<pierro78_> yes gnome is working :)
<jenneron[m]> does it use gpu? it should print in settings that it uses "FD680" or something like this
<pierro78_> where in settings ??
<pierro78_> I looked for gpu but it didn t find it
<jenneron[m]> in the info about device or system, i don't remember exactly
<jenneron[m]> should be at the bottom
<pierro78_> graphics unknown :(
<jenneron[m]> try to run glmark2-es2-wayland (install it firstly)
<pierro78> wow octane v2 27600 ... was only about 6400 before ...
<pierro78> (in chromium)
<pierro78_> glmark2-es2-wayland seems to run beautifully :)
<jenneron[m]> i'm not sure where i'm supposed to find reserved-memory ranges, it may take some time
<jenneron[m]> ok, i've found some in msinfo32
<steev> pierro78: would you mind posting the lightdm and xorg logs?
pierro78__ has joined #aarch64-laptops
<steev> you can do a number of things for the logs for gnome.. most deal with journalctl; you can do journalctl -f -o cat /usr/bin/gnome-shell
<pierro78__> going to sleep now
<pierro78__> thanks btw !
pierro78__ has quit [Quit: Page closed]
<steev> [ 29.636] (II) No input driver specified, ignoring this device.
<steev> i'm assuming it's the lack of xserver-xorg-input-libinput
<pierro78> running apt dist-upgrade now ... that was a difference I had when this was working ...
hexdump0815 has joined #aarch64-laptops
<pierro78> OK I could install xserver-xorg-input-libinput after dist-upgrade and now touchpad and keyboard are working with xfce ! :)
<pierro78> really time to (hapilly) go to sleep now ;)
hexdump01 has quit [Ping timeout: 480 seconds]
<steev> :)
pierro78 has quit [Remote host closed the connection]
alpernebbi has quit [Ping timeout: 480 seconds]
alpernebbi has joined #aarch64-laptops
hexdump01 has joined #aarch64-laptops
hexdump0815 has quit [Ping timeout: 480 seconds]
<jenneron[m]> does anyone know how to get adsp and mpss memory regions from windows or ACPI?
<travmurav[m]> jenneron: iirc the region size is in the driver .ini file near the blob and the blobs are probably relocatable with some alignment
* travmurav[m] should also fix the reserved memory as his is almost double of the windows's reserved 300mb
<travmurav[m]> actually, blobs have entry address in the elf header so maybe that also can be used
<jenneron[m]> i found some values in registry, e.g. for adsp MemoryAlignment 0x00100000 and MemoryReservation 0x02000000
<travmurav[m]> for mpss/adsp that actually seem to have a sane entry at least
<travmurav[m]> yeah
<jenneron[m]> so, for entry i should look into blob's elf header and reserve 0x2000000 from there
<travmurav[m]> qzed: has a nice reserved memory table in some surface dts actually so maybe can give some hints
<travmurav[m]> but I'd probably do it that way, yeah
<travmurav[m]> actually for modem it's a bit interesting
<travmurav[m]> since there are two
<travmurav[m]> modem and non-modem
<travmurav[m]> not sure if the _nm blob is small compared to non-modem so it'd be possible to save some space
<jenneron[m]> should i rely on blobs' size or MemoryReservation in registry?
<jenneron[m]> is it possible to work unstable when not all required memory is reserved though enough for blob?
<travmurav[m]> 0x2000000 looks like 35 mb that is less than 60mb of my modem-enabled blob tho so maybe it 's correct size for non-modem
<travmurav[m]> jenneron[m]: the remoteproc memory will certainly be protected
<jenneron[m]> rmtfs value from ACPI is going to overlap with xbl from sc8180x.dtsi
<travmurav[m]> jenneron: what ini file I meant is like this btw
<travmurav[m]> /mnt/Windows/System32/DriverStore/FileRepository/qcsubsys_ext_mpss7180.inf_arm64_7a58e200ce416ba6/qcsubsys_ext_mpss7180.inf
<travmurav[m]> ah
<travmurav[m]> it has two memoryreservationn-s
<travmurav[m]> for apq and msm
<jenneron[m]> ok, i've found (at least i think so) mpss and adsp regions, the only ones complained in kmsg
<travmurav[m]> so 0x2000000 is correct from that as well as the MemoryAddress x86000000 that seems the same as the elf
<travmurav[m]> (at least I guess this is similar enough between my 7c and yours 8cx)
<jenneron[m]> travmurav: 0x2000000 is for adsp not mpss
<jenneron[m]> for mpss it's 0xa000000 which is bigger, 160mb
<travmurav[m]> well, in my file it's either 0x2000000 for apq or 0x8400000 for msm so I suppose it's different for 8cx for "big" firmware
<jenneron[m]> i'm wondering if irc users will receive anything when i send a file through matrix
<travmurav[m]> I guess my adsp is a bit bigger too so guess no point comparing numbers as is on different platforms
<jenneron[m]> i would rather send a dtb than a patch
<travmurav[m]> it should send a link
<jenneron[m]> pierro78: if you see any link above, boot with this dtb, test wifi and take dmesg
flowriser has quit [Remote host closed the connection]
iivanov has joined #aarch64-laptops
jhovold has joined #aarch64-laptops
pierro78 has joined #aarch64-laptops
<abelvesa> steev: WRT fw_devlink cyclic dependency fixup going wrong, have a look at this thread: https://lore.kernel.org/all/CAGETcx8F0wP+RA0KpjOJeZfc=DVG-MbM_=SkRHD4UhD2ReL7Kw@mail.gmail.com/
<abelvesa> steev: also, Saravana sent a patchset that fixes it: https://lore.kernel.org/all/CAGETcx_nVXbHzZ3+_aR4SZtSnSBU=Rfp8Qm2jOs7zGZRaH_88A@mail.gmail.com/
<abelvesa> steev: this is the actual fix (from that last thread): https://lore.kernel.org/all/20220810060040.321697-6-saravanak@google.com/
<abelvesa> steev: but gregkh dropped the whole series from his queue, so ...
<pierro78> (ip a doesnt show no wifi interface)
<pierro78> lspci -k doesn t return anything ... lsusb doesn t seem to show wifi card
matthias_bgg has joined #aarch64-laptops
<pierro78> also I installed debian on a microsd in a usb 2 card reader ... tried to boot from internal microsd port and it almost booted (errors about ufs ... I think the FS ids have changed and it doesn t find the FS ?? ...) ... I tried to put the microsd in a faster USB3 card reader but the boot entry I previously wasn't showed by the "ESP boot chooser" ... there was another older entry but it didn't work ... maybe I can add another boot entry
<pierro78> (working fine again from the USB2 card reader)
<pierro78> it s drawing power from the charger but it's only like 7Watts or less when the battery needs charging ...
<pierro78> ... if I just shutdown the galaxy book s then the book s will charge (about 30 watts I don t remember)
shawnguo9 has joined #aarch64-laptops
c00k has quit [Ping timeout: 480 seconds]
Penguinpee has quit [Quit: Leaving]
Penguinpee has joined #aarch64-laptops
<jenneron[m]> pierro78: dmesg?
<jenneron[m]> with new dtb
<jenneron[m]> steev: there is a charging issue^
jelly has joined #aarch64-laptops
matthias_bgg has quit [Quit: Leaving]
jelly has quit [Remote host closed the connection]
jelly has joined #aarch64-laptops
go4godvin is now known as Guest801
Guest801 has quit []
go4godvin has joined #aarch64-laptops
go4godvin is now known as Guest802
Guest802 is now known as go4godvin
frytaped has joined #aarch64-laptops
frytaped has quit [Quit: WeeChat 3.5]
frytaped has joined #aarch64-laptops
matthias_bgg has joined #aarch64-laptops
frytaped has quit [Quit: WeeChat 3.5]
frytaped has joined #aarch64-laptops
iivanov has quit [Quit: Leaving...]
frytaped has quit [Quit: WeeChat 3.5]
frytaped has joined #aarch64-laptops
iivanov has joined #aarch64-laptops
frytaped has quit [Quit: WeeChat 3.5]
go4godvin is now known as testing
testing is now known as frytaped
pierro78__ has joined #aarch64-laptops
<clover[m]> <szclsya[m]> "to preserve the battery, since I..." <- Sounds useful, I use a feature like that on my snapdragon phone
pierro78__ has quit [Remote host closed the connection]
<pierro78> jenneron : here is dmesg : https://paste.debian.net/1251340/
<pierro78> efi (or maybe usb c hub management) a little finicky : I wasn t able to boot on the correct esp partition with no ethernet cable in the usb c hub (I was at the office)
matthias_bgg has quit [Quit: Leaving]
<steev> abelvesa: oh interesting, and i see, greg wants fixes only, which does make sense, and apparently reading through the thread, there is some (possible?) other breakage
<abelvesa> steev: I was hoping Saravana would at least suggest applying some patches from the set, since they are not really related between them
<abelvesa> we could dig to see exactly which patch fixes our issue and maybe tell Saravana to send just that one as a fix
<steev> i'll give the one you linked specifically a whirl, and see what happens
<abelvesa> since we are dragging this problem for quite some time now
<steev> part of the problem is that it's not every time
<steev> robclark might be a good candidate as well to testing, since iirc, he's passing some fw_devlink stuff in his commandline
<robclark> iirc I might still have fw_devlink=on .. I've not had to disable it for a while (but also haven't had time to move to 6.0-rc yet)
<robclark> so might be out of the loop if that was broken again recently
<steev> i think it's just that every few boots, not always, sometimes the display doesn't come up properly
robher has joined #aarch64-laptops
<steev> i definitely see it more often on the flex5g (and even occasionally the x13s) than the c630
<steev> so will eb testing all 3
<steev> jhovold: welcome back, hope the vacation was relaxing
<jenneron[m]> pierro78: move file /lib/firmware/qcom/sc8180x/sm-w767/qcmpss8180_mm.mbn back to /lib/firmware/qcom/sc8180x/sm-w767/qcmpss8180_XEF.mbn
<jenneron[m]> reboot, take dmesg again
<abelvesa> robclark: I've been using fw_devlink=permissive lately because of the specific dispcc and dsi-phy cyclic dependency
<steev> abelvesa: to be sure i'm on the right one, it's the DL_FLAG_CYCLE patch?
<abelvesa> steev: I think that's the one that should fix it, yes
<abelvesa> there should not be any device listed in /sys/kernel/debug/devices_deferred after applying that patch
<jenneron[m]> pierro78: make sure sha256sum of this file matches ecefb772cc57531edff255d59bf5639e35f00359b438eb2f4586218110e65663
<steev> abelvesa: fwiw, applying the whole set on the x13s seems to break things, but i was being lazy with b4 to apply them :D
<steev> i'm not sure what exactly, it says it's trying to fixup mdss to mdss, and then spits me out inside the initramfs because it won't mount root
<abelvesa> steev: can you pastebin the log ?
<steev> abelvesa: i'll try to get it again after work, i just gave it a quick test and since it spit me out in the initrd, i rebooted to a different kernel (applying just the DL_FLAG_CYCLE does not show the issue)
<jenneron[m]> travmurav: what does MemoryAlignment mean?
<steev> will also test c630 and flex after work
<travmurav[m]> jenneron: well.. memory alignment? :D so the address divides by that number would be an explanation I think
<jenneron[m]> i see..
<jenneron[m]> also, i don't know what to do with regions which are not defined in registry, e.g. xbl
<travmurav[m]> Iirc uefiplat.cfg had some stuff but unsure of what usefulness
<jenneron[m]> travmurav: is it on xbl partition?
<travmurav[m]> jenneron: I think it was inside it, yes
<jenneron[m]> pierro78: please make a dump of xbl partition in linux
<travmurav[m]> Could be unpacked with some tool for UEFI unpacking
<abelvesa> steev: sure, no pressure, thanks
<clover[m]> steev: any idea where 3-finger gestures are on GNOME on X13s
iivanov has quit [Quit: Leaving...]
<steev> no idea, i didn't know you could do them in linux
<clover[m]> this should automatically work with a wayland session
<steev> i'll check after work
<steev> oh, i don't think you can customize those
<steev> leezu: looks like you're right, there is something more going on - 3.38, just the patch to add the usb connector to the enum works fine, but on 42, i'm still crashing to gdm screen
<steev> it works after that crash though
<steev> and i can't tell on the flex because i think somehow the usb mp stuff is mucking things about, it's just refusing to enumerate, so i'll need to back my changes to add that out :(
<jenneron[m]> pierro78: i had asked to dump xbl, but there is no need anymore
hexdump01 has quit [Quit: WeeChat 1.9.1]
hexdump0815 has joined #aarch64-laptops
<jenneron[m]> travmurav: i extracted uefiplat.cfg from uefi capsule, but it doesn't seem to be very useful https://dpaste.com/BKU9FL6TC, however there is a uefipil.cfg which is more interesting and it uses different sizes for regions for some reason https://dpaste.com/793ZN5WKS, but still no information about xbl and some other regions
<jenneron[m]> actually xbl region in uefiplat.cfg may have some another name
<leezu> steev: After the crash, for me Gnome switches to X11 for both gdm3 and the session. Are you sure it works in a wayland session for you after the crash?
<jenneron[m]> i see some regions matching sc8180x.dtsi, but no xbl which is going to overlap with mpss from this table
<steev> leezu: let me check again
<steev> leezu: right you are, it's in X, and wayland isn't even available to choose from
<jenneron[m]> also, modem/mpss region entry is different in uefipil.cfg
<steev> leezu: looks like what i'm seeing is https://paste.debian.net/1251380/
<steev> leezu: actually https://paste.debian.net/1251381/ - forgot i didn't have the dbg syms installed
<steev> so looks like the same thing you reported?
<jenneron[m]> pierro78: try this^
<leezu> steev: Yes, it's the same issue as https://gitlab.gnome.org/GNOME/mutter/-/issues/2398 It isn't clear to me yet how to fix this issue. Specifically, plane_assignment=null as get_first_plane_assignment returns null. But is there always supposed to be a plane assignment? Or should meta_kms_crtc_predict_state simply check if there is a null assignment and return if there's none
<leezu> (instead of crashing)?
<leezu> (Essentially I didn't have a chance to study the code in more detail)
<steev> i honestly couldn't say, i'd assume the latter
<steev> maybe robclark would know
<robclark> I've not looked at that code before, so no idea ;-)
<steev> leezu: fwiw, mutter 3.38.6 (with my patch) does not show the issue
<steev> otoh, x13s doesn't have an actual gpu so... maybe thats why?
<robclark> abelvesa: re: fw_devlink issues, this came to my attention recently: https://lore.kernel.org/all/20220819221616.2107893-4-saravanak@google.com/
<robclark> looks like a revert is needed (but hasn't landed yet)
<robclark> (not been following it too closely.. just passing it along)
<clover[m]> steev: using `libinput debug-events` it's picking up 3 finger gestures
<clover[m]> gnome just isnt doing anything with them
<steev> clover[m]: which version of gnome shell? it works for me on sdm850 with 42.4
<steev> i haven't checked 3.38
<clover[m]> what is sdm850?
<clover[m]> DE: GNOME 42.4
<qzed> x11 or wayland? gestures only properly work on wayland
<clover[m]> yeah, im on wayland
<clover[m]> > $ evemu-describe
<clover[m]> Available devices:
<clover[m]> Select the device event number [0-0]:
<clover[m]> evemu-describe showing 0 devices not sure what that means
<clover[m]> echo $XDG_SESSION_TYPE
<clover[m]> is showing x11 oops
<clover[m]> that makes sense now
<clover[m]> ok things are working but its backwards. swiping down does what swiping up should do
<clover[m]> ahh its tied to natural scrolling
<clover[m]> ok i have no more questions lol