ChanServ 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
<RyanR>
FarchordSteveCossette[m], How do you run a Linux operating system on your smartphone?
<robclark>
amstan: I've got an x1e yoga 7x.. which is pretty close to becoming my new daily driver :-)
<FarchordSteveCossette[m]>
RyanR: in most cases right now you can't
<amstan>
robclark: fancy! i need to shop around for one too
pbsds has joined #aarch64-laptops
<amstan>
robclark: what don't you like about it?
* FarchordSteveCossette[m]
feels like he's being trolled.
<amstan>
Farchord (Steve Cossette): right? lol
baryluk3 has quit [Remote host closed the connection]
<robclark>
FarchordSteveCossette[m], RyanR: there is https://postmarketos.org/ but that is only adjacently on-topic here (probably more on topic in #linux-msm and other places)
<FarchordSteveCossette[m]>
robclark: Oh well yeah.... but I'm guessing the question was just a random throw....
<amstan>
... it's how the LLMs learn
<robclark>
amstan: very little.. we don't have external DP working yet, that is the main reason I'm still using the x13s.. but the 7x defn has a nicer screen and I think better kb feel (although dislike lack of non-fn pgup/pgdn keys)
<robclark>
but way faster than x13s.. which is pretty nice for kernel and mesa build time ;-)
<amstan>
robclark: how goes the adreno on those? did you keep being awesome on newer chips than trogdor?
<robclark>
there are a bunch of x1e laptops.. and dts posted for a few of them (crd, ofc.. but also yoga 7x, the vivobook, and lenovo t14s so far).. so fun times for aarch64-laptops
<RyanR>
robclark, Does the touchscreen not work on this Linux distribution?
<robclark>
turnip, the vk driver, already works.. I've got wip branch for gallium (gl/gles) which works pretty well but still debugging a few cts fails
<steev>
amstan: welcome back!
* robclark
touches screen... yup it works
<amstan>
steev: \o
<HdkR>
Oh, the touch screen does work. I'll need to disable that
<HdkR>
There we go, removed it from the DT
<amstan>
is arch kinda dead for this now? or are some people here still keeping it alive for their personal purposes?
<robclark>
afaict linaro efforts are debian, but there seem to be at least a few doing the arch thing.. and some fedora.. and openbsd
<robclark>
so defn doesn't feel like asahi style x1e distro situation
<robclark>
I guess it helps that we are a lot closer to having things in upstream kernel
<amstan>
how's the bios/bootloader on these windows machines?
<robclark>
uefi.. I'm using fedora's grub.. but dt instead of acpi (I just use dtb= kernel cmdline to load dtb from esp, but there are other ways)
<amstan>
but it's fairly straight forward? no annoying stuff to jump over?
<amstan>
i guess there's still that /dev/nvmep0n0 namespace that you have wifi settings in?
<robclark>
disable secureboot, disable bitlocker to copy over some fw files.. but other than dt vs acpi it is surprisingly normal
<robclark>
not afaict
<amstan>
you know what i mean, the wifi calibration data that needs to be baked on every device
<amstan>
i know bamse had some scripts that would make wifi go together with that data
<robclark>
for ath12k you might need to patch board-2.bin for cal data... but I guess proper fw will get posted in time
<robclark>
it is loaded from /lib/firmware/.. just like it is on x86 laptops
<FarchordSteveCossette[m]>
Using an external kb and mouse lol
bluerise_ has joined #aarch64-laptops
bluerise has quit [Ping timeout: 480 seconds]
<colemickens>
if anyone has a spare usb stick and wants to try a nixos image I've put together for the yoga7x, let me know.
hexdump0815 has joined #aarch64-laptops
hexdump01 has quit [Ping timeout: 480 seconds]
iivanov has joined #aarch64-laptops
iivanov has quit [Quit: Leaving...]
iivanov has joined #aarch64-laptops
iivanov has quit []
iivanov has joined #aarch64-laptops
iivanov has quit [Remote host closed the connection]
iivanov has joined #aarch64-laptops
iivanov has quit [Remote host closed the connection]
iivanov has joined #aarch64-laptops
iivanov has quit []
iivanov has joined #aarch64-laptops
<strongtz[m]>
Farchord (Steve Cossette): ath12k log indicates there's a high possibility that the firmware is wrong
<sera[m]>
do you get a message about SCMI notifications - core enabled? seems like now the mhi interface is broken, but it worked last time.. when you set up the new drive, are you sure all the modules are being loaded? try to see if there are logs from mhi (or just send the whole dmesg if you want)
<sera[m]>
i think the arth12k driver doesn't load the firmware until after it prints the message with the chip_id and etc
srini_ has joined #aarch64-laptops
jhovold has joined #aarch64-laptops
srini_ has quit [Ping timeout: 480 seconds]
pstef has quit [Ping timeout: 480 seconds]
pstef has joined #aarch64-laptops
paddymahoney has joined #aarch64-laptops
ellyq_ has joined #aarch64-laptops
srini_ has joined #aarch64-laptops
ellyq_ has quit [Remote host closed the connection]
ellyq_ has joined #aarch64-laptops
pstef has quit [Ping timeout: 480 seconds]
pstef has joined #aarch64-laptops
<FarchordSteveCossette[m]>
<sera[m]> "do you get a message about..." <- Before we do that, lemme rebuild the kernel once again with the config you guys recommended
<FarchordSteveCossette[m]>
Right now this is using stock kernel provided by Fedora
<sera[m]>
while you wait for that let me write up where to get the adsp/cdsp which is the fw you were missing too
<sera[m]>
put these in /lib/firmware/qcom/x1e80100/ASUSTeK/vivobook-s15/ to fix the pd-mapper errors: qcadsp8380.mbn and adsp_dtbs.elf from the most recent folder in FileRepository with name qcsubsys_ext_adsp8380_ , and qccdsp8380.mbn and cdsp_dtbs.elf from the most recent qcubsys_ext_cdsp8380_
<sera[m]>
* put these in /lib/firmware/qcom/x1e80100/ASUSTeK/vivobook-s15/ to fix the pd-mapper errors: qcadsp8380.mbn and adsp_dtbs.elf from the most recent folder in FileRepository with name qcsubsys_ext_adsp8380_*, and qccdsp8380.mbn and cdsp_dtbs.elf from the most recent qcubsys_ext_cdsp8380_*
<FarchordSteveCossette[m]>
I basically nuked most of fedora'S kernel configs, and kept only the config you guys recommended to avoid any issues. I guess we'll see...
ellyq_ has quit []
<FarchordSteveCossette[m]>
Ehhhh failed… <.<
<FarchordSteveCossette[m]>
Ill check later
<sera[m]>
looks like the script wants the config to be called kernel-6.11.0-something
<sera[m]>
can probably just rename in your tree and try again
srini__ has joined #aarch64-laptops
srini_ has quit [Ping timeout: 480 seconds]
<FarchordSteveCossette[m]>
Nah the script renames it
<FarchordSteveCossette[m]>
It's expecting something else. I guess I gotta reenable the base fedora configs and apply the x1e config on top
* FarchordSteveCossette[m]
shrugs.
<FarchordSteveCossette[m]>
But I think I did it wrong last time.... meh we'll see I guess
smpl has joined #aarch64-laptops
<FarchordSteveCossette[m]>
Hmmm I wonder if some of the modules in the config don't exist in the base 6.11 kernel
<sera[m]>
edit the spec to manually set Arch=arm64 at line 2067
<sera[m]>
looks like it expects the config to be a certain format too with the arch at a certain place
<FarchordSteveCossette[m]>
<sera[m]> "edit the spec to manually set..." <- No if you look, it renames all configs to add the kernel version. That works fine. That spec has everything for all arches, but if you look at the copr build, I'm only building the aarch64, so that's already all it's doing. The non-aarch64 stuff just gets ignored
<sera[m]>
the last failure was because the spec tried to set Arch to blank
<sera[m]>
I tried to build with the spec provided but I ran into that exact issue and setting it manually mostly worked but for some reason I OOM when linking using it
<FarchordSteveCossette[m]>
yeah I tried his last 6.10.0-next kernel, and it failed to boot
<FarchordSteveCossette[m]>
couldn't mount boot device
<sera[m]>
I can build with his config off fedora's tree then yeah
<sera[m]>
yeah the file name needed to changed back, right? name it kernel-6.11.0-aarch64-x1e80100.config and additionally add the # arm64 line to the top
<sera[m]>
It should install fine with the normal rpm package installation
<FarchordSteveCossette[m]>
Oh I'm an idiot
<FarchordSteveCossette[m]>
I uploaded -1, needed to upload -2
<FarchordSteveCossette[m]>
The file name was ok, but I did add # arm64 at the top
<FarchordSteveCossette[m]>
Found unset config items in arm64 aarch64-x1e80100, please set them to an appropriate value
<FarchordSteveCossette[m]>
<.<
craftyguy has quit [Remote host closed the connection]
craftyguy has joined #aarch64-laptops
<sera[m]>
Probably need to go add those lines in as it wants.. I'm build another rpm kernel with the fedora config as a base this time which might work better for you
<FarchordSteveCossette[m]>
Thanks! That would be awesome.
<FarchordSteveCossette[m]>
I'm trying to simply remove all the lines in the config that are unset (*="") and see if that works
<sera[m]>
What it probably needs you to do is just look at the lines it points out and set them to a matching value
<FarchordSteveCossette[m]>
Oh I didn't look down enough
<FarchordSteveCossette[m]>
I think I only really need the kernel right? No need for devel or headers
craftyguy has quit [Remote host closed the connection]
<sera[m]>
Think so
<sera[m]>
idk what the packaging put in each
craftyguy has joined #aarch64-laptops
craftyguy has quit [Remote host closed the connection]
craftyguy has joined #aarch64-laptops
* FarchordSteveCossette[m]
tries to put those in and hopes he doesn't hose his install again XD
<FarchordSteveCossette[m]>
Oh right, keyboard and mouse dont work... so I'll need to do that from my orangepi, as doing this with an external keyboard is a paaaain XD
<FarchordSteveCossette[m]>
I'll do that later
srini_ has joined #aarch64-laptops
srini__ has quit [Ping timeout: 480 seconds]
<FarchordSteveCossette[m]>
<sera[m]> "https://wormhole.app/AlQmP#..." <- So, I extracted the rpm (as rpm -i refused to install it alongside the current kernel, I simply extracted the RPM and moved files into place) and then I ran depmod and kernel-install to create the images. Then, I edited the loader file, and...
<sera[m]>
I think it didn't built right with everything built in I guess, the kernel was pretty small
<FarchordSteveCossette[m]>
yeah I'll check this afternoon, I got something coming up, thanks for your hepl!
<FarchordSteveCossette[m]>
*help
<travmurav[m]>
Farchord (Steve Cossette): `earlycon=efifb keep_bootcon` doesn't help either in that case?
<travmurav[m]>
as in, to at least know where it choked
<FarchordSteveCossette[m]>
travmurav[m]: Well, the kernel was like 30mb, so I'm guessing something went wrong in the build
<FarchordSteveCossette[m]>
that was oddly small
<FarchordSteveCossette[m]>
This one is 100mb, makes more sense
<travmurav[m]>
30 sounds reasonable for a kernel tailored to a device tbh
<sera[m]>
not too sure, I just merged rob's config on top of the fedora config for both, but the small one I ran make localyesconfig to bake everything in
<FarchordSteveCossette[m]>
FarchordSteveCossette[m]: Even with that in, nothing. Hard freeze at the same place.
<FarchordSteveCossette[m]>
Sorry, I meant with those kernel args
<travmurav[m]>
ugh sounds very sad then since I'd expect the kernel to write at least a little bit into the efi framebuffer when booting :(
<FarchordSteveCossette[m]>
Alright, here goes nothing
<travmurav[m]>
earlycon=efifb needs pretty much nothing but a running cpu to work xD
<FarchordSteveCossette[m]>
Nope, still stuck at... well, nothing
<FarchordSteveCossette[m]>
lemme try with travmurav 's kernel args
<sera[m]>
with the new kernel, and there's stuff on the boot partition? (the dtb, kernel image, and initramfs)
<travmurav[m]>
<FarchordSteveCossette[m]> "ima_93201a1.jpeg" <- ugh and there is no usual efistub output like here?
<FarchordSteveCossette[m]>
nop
<FarchordSteveCossette[m]>
last line is EFI: Exiting boot services.
<travmurav[m]>
ah, so the same
<FarchordSteveCossette[m]>
Yup
<travmurav[m]>
well then it's sad I guess since the kernel for sure contains the efistub code but I guess it dies very early :/
<travmurav[m]>
I guess one possibility the efifb is not enabled in it somehow, that would be annoying
<FarchordSteveCossette[m]>
I even went the very careful way: I 100% deleted the old kernel before
<FarchordSteveCossette[m]>
and confirmed that everything relating to the kernel was gone from /boot
<FarchordSteveCossette[m]>
then I rm -rf /lib/modules/ kernel version folder
<sera[m]>
it's weird, if the root was wrong it should print it can't find the devicetree/kernel
<FarchordSteveCossette[m]>
Well the devicetree is in /boot
<FarchordSteveCossette[m]>
and it finds it
<sera[m]>
I don't know if there's some gotchas related to the binrpm-pkg, but it seems like robclark's build before didn't have those issues
<FarchordSteveCossette[m]>
so that ain't it
<travmurav[m]>
nah it most likely jumped to the kernel proper and died somewhere there if you see the "Exiting BS" line
<FarchordSteveCossette[m]>
sera[m]: Right. IT had some, but it went further XD
<sera[m]>
I can try again without a fedora merge too
<FarchordSteveCossette[m]>
The problem rob's kernel had is, for some reason it didn't find my drive
<FarchordSteveCossette[m]>
When it came to mounting them for userspace
<sera[m]>
yeah that might be related to differences in what options fedora needs in the kernel config
<FarchordSteveCossette[m]>
Lemme ask someone in the fedora kernel team
<FarchordSteveCossette[m]>
Actually
<FarchordSteveCossette[m]>
Not sure that would work either, the config provided by rob has a bunch of settings that overwrites fedora's
<FarchordSteveCossette[m]>
so they might just be wholly incompatible
<sera[m]>
ye let me try a build with fedora config on the new tree
<FarchordSteveCossette[m]>
Do you guys have an iso or an image of what you have that I can try out of the box?
<FarchordSteveCossette[m]>
I wanna see if it just boots in Linux proper with (most) things working
<sera[m]>
Yeah if it doesn't need to be fedora I can get you an arch image in a sec
<sera[m]>
you'll just need to add on your firmware
<sera[m]>
Farchord (Steve Cossette): actually can you just zip up the firmware files and send them over
<FarchordSteveCossette[m]>
Well I'd like it to be a fedora image
<FarchordSteveCossette[m]>
but if it can't, I just want to use another image, get things working, and then apply that to fedora
<FarchordSteveCossette[m]>
if that makes sense
<FarchordSteveCossette[m]>
Right now I'm working against myself: I'm working in an environment different than all of you
<FarchordSteveCossette[m]>
<sera[m]> "Farchord (Steve Cossette..." <- Yup one sec
<FarchordSteveCossette[m]>
Actually just got an inpromptu meeting.... it'll have to wait a bit
indy has quit [Ping timeout: 480 seconds]
<FarchordSteveCossette[m]>
<sera[m]> "put these in /lib/firmware/qcom..." <- IS it possible there'S a mistype? I found the qcsubsys folder, but not the qcubsys one
<sera[m]>
Can you upload it somewhere else? Can't seem to be able to dl it from the matrix server. Grab the GPU zap shader too, should be a file called qcdxkmsuc8380.mbn under the qcdx8380.inf_arm64 folder. And yeah, the cdsp folder should be qcsubsys_ext_cdsp
<sera[m]>
Ah it finally downloaded. You're missing the cdsp_dtbs.elf file too
<FarchordSteveCossette[m]>
These files are unavailable
<FarchordSteveCossette[m]>
The sender left before finishing the upload. You will not be able to download the files. Ask the sender to re-upload the files to Wormhole.
<sera[m]>
ignore the message, try to dl it anyway
<sera[m]>
seems to work
<FarchordSteveCossette[m]>
ODd
<FarchordSteveCossette[m]>
You know, downloading all those images, I sure am glad I got 1.5gbit internet XD
<FarchordSteveCossette[m]>
lmc something, I heard usb-a dont work well on this laptop in linux, ill try with a usb-c hub thingy
<sera[m]>
yeah, try the usb-c port.. hopefully this works, it's the same image that works perfectly for me, just with the vivobook dt
<sera[m]>
it's possible the nvme also has some issues but I don't think @strongtz (who submitted the vivobook dt) had issues with it
<FarchordSteveCossette[m]>
Yup that's MUCH better
<FarchordSteveCossette[m]>
it was usb-c
<FarchordSteveCossette[m]>
*usba
<Painkiller995[m]>
Hey,
<Painkiller995[m]>
I'm new here. I got the Asus Vivobook S 15 and I wanna help out with an aarch64 laptop. I took it apart and put in a 2 TB SSD. I'm ready to start testing stuff out, but I don't know where to begin. Are you guys working on getting Linux to run on it? Any idea where I can find more info about your progress?
<FarchordSteveCossette[m]>
What'S the password .sera ?
<sera[m]>
alarm
<FarchordSteveCossette[m]>
Okay so
<FarchordSteveCossette[m]>
Got battery (nice)
<FarchordSteveCossette[m]>
No wifi
<FarchordSteveCossette[m]>
No sound
<sera[m]>
I think sound is still wip, can you send the logs again to check for wifi?
<FarchordSteveCossette[m]>
120hz, super nice!
<sera[m]>
Check the gpu works too, though I think it should considering that image forces the gpu driver and would probably crash on desktop if it wasn't working
<FarchordSteveCossette[m]>
GPU seems to be working
<FarchordSteveCossette[m]>
Painkiller995: sorry, I've been working on that specific problem since last friday XD
<sera[m]>
ah okay just use your wifi blobs I had you get from windows bwefore
<FarchordSteveCossette[m]>
I'll send some infos after dinner
<sera[m]>
I left the lenovo ones since you didn't include the files there but I think the wifi will work once you put in your files to /lib/firmware/ath12k (delete board-2.bin also)
<sera[m]>
log properly says the id is bus=pci,vendor=17cb,device=1107,subsystem-vendor=105b,subsystem-device=e0ef,qmi-chip-id=2,qmi-board-id=255 from ath12k/WCN7850/hw2.0/board-2.bin if you want to just add your board to board-2 also using the earlier instructions
<Painkiller995[m]>
Farchord (Steve Cossette):
<Painkiller995[m]>
Great, so I'll play with these images in the meantime while you guys update us with more info. 😅
<FarchordSteveCossette[m]>
> Great, so I'll play with these images in the meantime while you guys update us with more info. 😅
<FarchordSteveCossette[m]>
So, tl;dr: if you see the last wormhole link sera posted you should be able to boot from usb. Dont use usb-a, only -c. Type a ports dont work on our laptop in linix yet
<FarchordSteveCossette[m]>
Wifi needs files copied from windows
<FarchordSteveCossette[m]>
Sound is a wip
<FarchordSteveCossette[m]>
Painkiller995: do you know how to boot from an external drive?
<FarchordSteveCossette[m]>
I would not recommend replacing what'S on the nvme drive
<Painkiller995[m]>
Farchord (Steve Cossette):
<Painkiller995[m]>
Yeah, I know. I've already swapped the NVMe with a new, empty one and reinstalled Windows using WinPE. I'll test the Linux image using an external one.
<FarchordSteveCossette[m]>
> Yeah, I know. I've already swapped the NVMe with a new, empty one and reinstalled Windows using WinPE. I'll test the Linux image using an external one.
<FarchordSteveCossette[m]>
ah ok. My bad forgot you mentioned that
<steev>
i always forget about how terrible the matrix to irc bridge is until someone uses matrix features
<Painkiller995[m]>
Farchord (Steve Cossette): The download won't start.
<abby>
steev: heh
<FarchordSteveCossette[m]>
Painkiller995[m]: huh.... it did for me
<abby>
the bridge has been giving farchord a name so long that it covers up my timestamps fully
<FarchordSteveCossette[m]>
I can make it longer if you want
<abby>
only 6 chars longer :)
<Painkiller995[m]>
Farchord (Steve Cossette):It is working! But this wormhole thing is messing with my head.
<FarchordSteveCossette[m]>
Painkiller995[m]: Wait until you play with trying to make Linux work on your laptop for multiple days in a row!