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
ungeskriptet has quit [Quit: Ping timeout (120 seconds)]
ungeskriptet has joined #aarch64-laptops
<clee>
robclark: oh awesome, thanks! I'll try that. has Qualcomm posted patches for cpufreq support or GPU acceleration yet?
<robclark>
gpu patches are posted and I merged them into msm-next today.. I think dt bits will land in 6.11 as well.. and I saw patches for cpufreq on list
<clee>
excellent.
<robclark>
not sure if this will all land in 6.11, but ofc there are integration branches
<robclark>
I kinda think the long pole will be per-device dts
<robclark>
since there are a lot of devices and we don't just get things automagically because acpi
<robclark>
if you just want to take $random_distro installer and have things work.. it is going to be a few months or more (depending on how bleeding edge your preferred $random_distro is)
<robclark>
if you are comfortable building your own kernel and doing dts hacking, then join the party ;-)
<clee>
well, I just finished building a custom rk3588 kernel to try to get my Rock5B working, but it panics 100% of the time when I rsync a 5+GB file to a local disk on it
<clee>
so hopefully my luck will be better here
<robclark>
I'd say it is defn still "rough".. if you aren't comfy installing a distro without an actual installer (or don't mind using windows+WSL for a bit) then wait
<robclark>
but I expect things will move reasonably fast
derzahl has quit [Ping timeout: 480 seconds]
<clee>
well, I have working netboot for all my _other_ arm64 machines here…
<clee>
surely it can't be that hard to get this laptop to netboot too
<clee>
all right, thanks for the pointers, I'll go play around and try some stuff and report back
<robclark>
if I hit f12 at boot to bring up the boot selection menu, with a usb-eth plugged in, I see a pxe boot option.. I've not tried it
<robclark>
cobbling together my own esp on thumb drive has been good enough for me so far
<robclark>
I might try to clean up the yoga slim dts a bit, but I'm not really an acpi expert (and mostly want to get things to the point where I can start working on a7xx gallium, since until now I haven't had an a7xx device.. turnip/vk + zink is already usable but that is suboptimal when most things on linux are still using gl)
derzahl has joined #aarch64-laptops
Kelsar has quit [Ping timeout: 480 seconds]
calebccff has joined #aarch64-laptops
derzahl has quit [Ping timeout: 480 seconds]
possiblemeatball has quit [Remote host closed the connection]
cachoeira has quit [Remote host closed the connection]
cachoeira has joined #aarch64-laptops
srinik has joined #aarch64-laptops
Kelsar has joined #aarch64-laptops
albsen[m] has joined #aarch64-laptops
<albsen[m]>
I'm now about to connect 3 screens to my x13s! yay! on 6.10.rc3
<albsen[m]>
s/about/able/
<clee>
ooooh, very nice
<clee>
I officially have gotten netboot working on the Yoga Slim 7X
<albsen[m]>
nice
<clee>
now I just need to build a newer kernel image with more hardware-enablement patches
<HdkR>
albsen[m]: MST working now or something?
<clee>
robclark: thanks so much for the tip about F12, I'd been using a SIM removal tool to push the tiny pinhole button on the bottom of the case to get into the BIOS and that suuuuucks
<albsen[m]>
HdkR: multi screen? kernel 6.10.rc3 works for me. for wayland gdm needs v46 or v45.6 (i think). I'm currently booting into text mode and starting sway to get wayland to work.
<albsen[m]>
one screen on DP out (usbc I belive), and one via HDMI from my satchi adapter
<HdkR>
albsen[m]: Yea, displayport multi-stream transport. Otherwise you wouldn't be able to run three screens off two type-c ports :D
<clee>
the built-in screen counts as 1!
<albsen[m]>
I have 3 on now
<HdkR>
"connecting" an internal screen is a bit of a weird statement
<HdkR>
But I guess technically true if you've disconnected it :P
<albsen[m]>
1 internal, 2 external. 2.5k + 1080p
<albsen[m]>
dont have a 4k screen, or 8k
<albsen[m]>
not sure what the max supported is
<HdkR>
2-lanes of displayport from each type-C port, so 4k30 each. Until we can dynamically choose 4-lanes of DP anyway :)
<HdkR>
Would be cool if the display controller on the X13s was good enough to run a tiled 8k display although. I tried it on Windows and it wasn't possible there
<albsen[m]>
my 1080p asus is doing 144hz
<albsen[m]>
HdkR: ok, so 1 internal + 2x 4k 30hz is the current max?
<HdkR>
I believe so
<albsen[m]>
interesting, I guess I'm lucky that I use 2.5k with 60hz; really dont like working on below 60
<albsen[m]>
this little machine (x13s) is a beast. its silent, can charge from my phone charger (lowest I used 5w to charge); runs all my apps at the same time (32gb ram), has super fast wifi speeds and a replace-able SSD. just need to get the webcam working (its supported but I didn't have time to put the correct pieces together on my debian install)
<albsen[m]>
* replace-able SSD and powers 2 external screens. just
jhovold has joined #aarch64-laptops
<HdkR>
It's almost perfect
srinik has quit [Ping timeout: 480 seconds]
alfredo has joined #aarch64-laptops
alfredo1 has joined #aarch64-laptops
alfredo has quit [Ping timeout: 480 seconds]
alfredo1 is now known as alfredo
srinik has joined #aarch64-laptops
<strongtz[m]>
hmm I think the upstram msm driver lacks DP MST support...
<strongtz[m]>
*upstream
<clee>
oh hey strongtz[m], got any tips for me getting my Yoga Slim 7X running Linux?
<strongtz[m]>
clee: for now you'll have to build your rootfs on your own, and manually install the grub bootloader and rootfs to your nvme
<clee>
already done that, just need a working kernel
<clee>
currently cloning the msm repo
<clee>
(also have netboot working, which is nice, lets me test different kernel versions much faster)
<strongtz[m]>
clee then my repo with the correct kernel config should just work
<clee>
link?
<clee>
(also, what is the "correct" kernel config?)
srinik has quit [Ping timeout: 480 seconds]
matthias_bgg has joined #aarch64-laptops
smpl has joined #aarch64-laptops
ungeskriptet is now known as Guest375
ungeskriptet has joined #aarch64-laptops
Guest375 has quit [Ping timeout: 480 seconds]
srinik has joined #aarch64-laptops
alfredo1 has joined #aarch64-laptops
jhovold has quit [Quit: WeeChat 4.2.1]
alfredo has quit [Ping timeout: 480 seconds]
alfredo1 is now known as alfredo
alfredo has quit [Quit: alfredo]
juergh_ has joined #aarch64-laptops
juergh has quit [Ping timeout: 480 seconds]
jhovold has joined #aarch64-laptops
Kelsar has quit [Remote host closed the connection]
Kelsar has joined #aarch64-laptops
craftyguy has quit [Remote host closed the connection]
craftyguy has joined #aarch64-laptops
craftyguy has quit [Remote host closed the connection]
craftyguy has joined #aarch64-laptops
craftyguy has quit [Remote host closed the connection]
possiblemeatball has quit [Remote host closed the connection]
possiblemeatball has joined #aarch64-laptops
alfredo has joined #aarch64-laptops
possiblemeatball has quit [Remote host closed the connection]
possiblemeatball has joined #aarch64-laptops
possiblemeatball has quit [Remote host closed the connection]
<robclark>
bamse (or anyone in the know about dsp things): What is cdsp_dtb.mbn / adsp_dtb.mbn ? I don't see anything with "dtb" in the name from the win fw
possiblemeatball has joined #aarch64-laptops
<robclark>
abelvesa: ^^^
<bamse>
robclark: those are exactly what it says, devicetree blobs for cdsp and adsp ;)
<bamse>
robclark: that said, i don't know the answer to your question :(
<robclark>
heh.. are they signed or can I just use the crd ones?
<bamse>
they are supposed to be signed
<robclark>
hmm
<bamse>
robclark: does this imply that you got past your boot crash? or just circumvented it?
<bamse>
"just running this random tree i found on the internet" ;)
<robclark>
heheh
<robclark>
well I did look at strongtz's commits to make sure there was no hidden `rm -rf /` :-P
<bamse>
would be nice to have some commit messages, to understand those hacks...but the delta ontop of abel's tree looks quite nice
<bamse>
right, need to do that git log --author "Jia Tan" before you run the code
<robclark>
if strongtz isn't interested in posting the yoga slim dts then I'll clean up the history and add commit msg when I get more things work.. otherwise he can take my fixups
<robclark>
*working
<clee>
hmmm, that x1e80100 defconfig sure does seem to be missing a lot of options when I run `make oldconfig` with the work/sakuramist-x1e80100 branch…
cachoeira has quit [Remote host closed the connection]
cachoeira has joined #aarch64-laptops
<robclark>
clee: I just catted the x1e defconfig to my existing config which had various distro things... and that seemed to work
<robclark>
bamse: so.. how would I know if adsp/cdsp were working?
<JensGlathe[m]>
Shouldn't it be the usual suspects like sound and battery moneitor?
<clee>
jhovold: don't suppose you have any notion of how different the X13s and Yoga Slim 7X are internally?
<clee>
(how much of your X13s work can I copy, basically? :)
<robclark>
not directly answering the question, the diff btwn asus and slim acpi tables isn't all that much
<bluerise>
I think the keyboard or touchpad is on a different I2C bus
<robclark>
yeah
<agl>
clover[m]: I have again problems with the grafik under x13s and EndeavourOS ARM.
<agl>
After an Upgrade.
<steev>
check that the firmwares exist
<agl>
steev: I have once more updatet the x13s-firmware, but it does not work
<steev>
in what way? black screen at boot? gdm not starting?
<agl>
steev: The kernel starts and I can per SSH a login. I have at this moment not seen the Console output. gdm does not start and the Xorg proces does also not start.
<agl>
dmesg have the folowin output:[ 11.999361] arm-smmu 3da0000.iommu: deferred probe timeout, ignoring dependency
<agl>
[ 11.999368] arm-smmu 3da0000.iommu: probe with driver arm-smmu failed with error -110
<robclark>
abelvesa: hmm, building kernel on the x1e yoga... it got the fans to come on, but also appears to have locked the thing up
<abelvesa>
robclark: hm..
<abelvesa>
any logs?
<steev>
agl: and lastly, are you sure you're using the correct dtb?
<frozen_cheese[m]>
fwiw, I have a kernel that at least starts to load on the Yoga 7x. Kernel 6.9.6 built from source, handmade EFI boot from USB. But after about a page of boot messages and penguin graphics, the screen goes black and doesn't come back. Might be the dtb as I'm using the CRD image.
<frozen_cheese[m]>
Also, in case anybody didn't know, at least on the Yoga 7x, F2 will open the "BIOS" screen
<robclark>
abelvesa: not sure yet
<robclark>
frozen_cheese[m]: you probably don't want to use crd dtb.. but fwiw I was hitting a reboot on the yoga which I traced to /lib/udev/rules.d/80-drivers.rules
<robclark>
(this was on a kernel that is a bit closer to linux-next)
<frozen_cheese[m]>
yeah, was trying to find a better dtb, but was not successful thusfar
<agl>
steev: The right .dtb comes with the kernel 6.9.5 from clover
<steev>
not sure a better one exists yet frozen_cheese[m] :(
<agl>
steev: I have the sc8280xp-lenovo-x13s...dtb from clover which comes with the linux kernel package 6.9.5
<steev>
can you pastebin `grep 8280 <kernelconfig>` wherever the package installs the kernel config used to? i don't have any idea where it may be
todi_away has quit []
<colemickens>
robclark: seems like you're hacking on the yoga, do you have your own tree with that patch, or are you collabing with strongz on that x1e branch?
<robclark>
abelvesa: ok 2nd time from scratch kernel build succeeded without fireworks... for same config, on x13s it was 33m21s.. on x1 yoga (x1e78100) it was 8m59s
<Jasper[m]>
robclark: Hahaha nice speed bump
<abelvesa>
robclark: good to hear
<robclark>
yeah, no kidding
<JosDehaes[m]>
I have a yoga slim coming in 2 weeks, sounds like by then it will mostly work 😁
<abelvesa>
robclark: so what still doesn't work then?
<Jasper[m]>
Someone I know compiled edk2/mu on their devkit and now on a surface. Went from a 10min build to less than a minute
<abelvesa>
robclark: do you have gpu working?
<abelvesa>
Jasper[m]: less than a minute? from scratch? highly doubt it.
<Jasper[m]>
abelvesa: Pretty sure it was yeah
<Jasper[m]>
devkit being the one with sc8280xp, not the new one
<abelvesa>
with sc8280xp even less possible? what kernel config was the guy using?
<abelvesa>
... with sc8280xp even less possible!
<Jasper[m]>
abelvesa: Nah not kernel, Mu/EDK2
<abelvesa>
oh, I see
<abelvesa>
no idea how to compare then
<JensGlathe[m]>
my kernel (6.10.0-rc5 for pop / ubuntu with the ubuntu-mainline build scripts) builds 33m+ on sc8280xp devkit, and 17m+ cross on a 12 core / 24 thr vm on a R9-3950x. x1e appears to be pretty fast.
<srinik>
robclark: thanks for the review, will sendout a v2 with touchpad, which is working with the suggested changes
* colemickens
can't contain his excitement
* colemickens
has 11 days left til receiving the Yoga 7x :D
<agl>
steev: Sorry here is it raining, and I'am at home now. I have a look to my x13s: The screen is black and I can not switch to any console. But I can make a login per SSH from an another Notebook.
<steev>
okay the config options are there. i've no idea, and i'll leave it to clover whenever he's back around
<agl>
ok
<agl>
Thank you
srinik has quit [Ping timeout: 480 seconds]
<robclark>
yeah, my kernel config is closer to distro config with everything enabled, no doubt a minimal kernel would just be a couple of minutes
<robclark>
abelvesa: yeah, gpu working (with the hack for chip-id)
<robclark>
abelvesa: I don't think wifi is working.. or at least I'm not sure what I'm missing (maybe some firmware?)
<steev>
ath12k firmware should be needed, maybe this is another case of the windows board id not showing up, and make you could force it like we used to the x13s?
<robclark>
hmm, idk much about wifi.. IIRC I had to install some userspace stuff on the x13s, not sure if that is still needed
<steev>
you'd need pd-mapper, probably
<steev>
wait, no
<robclark>
right, but I think the in-kernel one is part of the branch I'm using
<steev>
oh, then you shouldn't need pd-mapper :D just the jsn files, and the firmware file(s)
<steev>
robclark: yeah should be upstream already, but the files are named the same on windows iirc
<robclark>
hmm, g-s crashed in in libmpg123.so.0 .. that is kinda terrifying that the compositor has codecs linked against it
alfredo has quit [Ping timeout: 480 seconds]
<clee>
hmmm. I must be building something wrong, because I just keep getting kernel panics
<clee>
I did grab the x1e80100-yoga.dtb from arch/arm64/boot/dts/qcom and pass it in and got further along, but still panic after the 12 penguins show up
shoragan has quit [Quit: quit]
<robclark>
clee: maybe you can catch the panic on camera?
<steev>
what command line arguments are you using
<robclark>
steev: ok, getting further with ath12k built as module:
<steev>
but i'd go anonymix007[m] route if it works, it's easier
<robclark>
hmm, but kernel patch I can carry until I don't need it.. hacked fw gets blown away whenever the linux-firmware stuff is update ;-)
<steev>
true
<anonymix007[m]>
steev: Well, it works. But ath12k itself is sub par imho. The manufacturer even told me to use qcacld3
<steev>
to do what
<clee>
okay, neat, now it just hangs waiting for the network device… I must not have built the right USB ethernet driver into the initrd
<anonymix007[m]>
@steev Sent bugzilla link earlier. In short, wifi dies when uploading files
<anonymix007[m]>
* steev: Sent bugzilla link earlier. In short, wifi dies when uploading files
<steev>
i meant, what do you mean by use qcacld3
<steev>
ugh
<steev>
i'm going to my sister's place to watch their dog for the night and need to take like 5 laptops with me :(
<anonymix007[m]>
steev: I meant in place of ath12k
<steev>
uhh is that a driver? they mean replace the actual hardware?
<anonymix007[m]>
Yes, it is the driver
<steev>
the downstream kernels are not exactly a bastion of quality coding
<anonymix007[m]>
Neither is upstream. Qualcomm does not recommend using ath12k. At least this is what the manufacturer said
<steev>
they also don't recommend linux :P
<steev>
i take their recommendations with a HUGE grain of salt
<HdkR>
I'll take the salt block please
<anonymix007[m]>
I mean... qcacld3 driver is used in Android and works without that issue.
<steev>
you will never convince me that "it works in android" means something is good
<Jasper[m]>
anonymix007[m]: ... he says when the average Snapdragon android phone cannot handle trying to connect to an AP using the wrong password 4 times in a row
<Jasper[m]>
after which it crashes
<anonymix007[m]>
Jasper[m]: Define average
<Jasper[m]>
anonymix007[m]: I often run into this bug (though not because a wrong password) in my sm8250 device
<anonymix007[m]>
Jasper[m]: Just tried on sm8550 (same wcn785x, which is why I even referred to Android here) and it didn't crash
<Jasper[m]>
Ok, just saying it isn't flawless
<robclark>
android drivers tend to.. at least sometimes.. work well with android.. kinda beaten into shape.. but fall over when they encounter different userspace that doesn't do things exactly like android
<anonymix007[m]>
Jasper[m]: It obviously isn't. Whatever the qcacld3 version the manufacturer sent me doesn't even compile with a modern enough kernel
<clee>
right, I don't need NFS boot, I can PXE boot to a local disk too. but, uh, that doesn't seem to work either… using the same UUID that the NVMe disk's grub2 has, looks like my kernel built from strongtz[m]'s branch can't find the NVMe disk
<clee>
booting with `/arm64/vmlinuz ro root=UUID=0cfa01bd-36a2-44c0-bd3e-3ca432e3d07d coherent_pool=2M earlycon=efifb clk_ignore_unused pd_ignore_unused arm64.nopauth efi=novamap systemd.unit=multi-user.target` now
<clee>
well, I don't know why it worked the second time, but I've got a shell, and `lsusb` returns nothing, which explains why the network device is missing
possiblemeatball has quit [Quit: Quit]
<robclark>
I've had problems with using uuid but seems fine if I use root=/dev/nvme0n1p3 .. idk what is going on there
<robclark>
also have to manually modprobe evdev and the r1852 (or whatever it is) module for USB eth (also not sure why that isn't happening automatically)