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
xroumegue has joined #aarch64-laptops
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
Lucanis_ has joined #aarch64-laptops
<HdkR>
Weird, my Sabrent USB dock seemingly overheated and prevented my Yoga from booting. I don't even know what would cause that
chrisl has joined #aarch64-laptops
<HdkR>
JensGlathe[m]: Finally got around to testing that dt patch. Looks like it breaks the internal display and I don't think external is working
chrisl has quit [Ping timeout: 480 seconds]
landwork has joined #aarch64-laptops
chrisl has joined #aarch64-laptops
<landwork>
question, regarding ubuntu forums (An error occurred: We're sorry, but new users are temporarily limited to 3 replies in the same topic.)
<landwork>
how do I get rid of that, I have been a user for months now.
<landwork>
super annoying.
chrisl has quit [Ping timeout: 480 seconds]
<tobhe>
landwork: I am not sure tbh. will try to find out tomorrow
<landwork>
thanks :)
<landwork>
I also wanted to ask, I have slim7x just updated to latest as of like 20 mins ago. Got a black screen failed boot left it for 10 mins, then repowered and it's fine now.
<landwork>
but I wanted to ask if there was something I can do to get sound working?
<landwork>
i can help with testing too
<landwork>
just need to be nudged in the right direction.
<landwork>
very nice of lenovo adding the official firmware now :)
<landwork>
maybe the higher ups are frustrated that windows isn't driving sales ;)
<landwork>
we all know a solution to that...
<tobhe>
the main problem with sound support is that we aren't sure how safe the current kernel parts are in terms of hardware damage
<landwork>
i'm ok to risk that within reason.
<tobhe>
there needs to be some kind of safeguard to make sure users don't roast their speakers
<landwork>
is it just -> "snd-soc-x1e80100.i_accept_the_danger=1" to override?
<tobhe>
that and installing sof firmare + alsa ucm configs
<tobhe>
audioreach topology firmware I meant
<tobhe>
I don't have a slim7x here so no idea where to find the alsa config
<landwork>
ok, I try go hunting, that's a start :)
<landwork>
thanks tobhe.
<robclark>
HdkR: you need a later dt patch that removes the inadvertent change to edp bl power.. when I dropped that (and enabled the new retimer module in my config) I have internal+external display working on 7x
<HdkR>
I see
<robclark>
HdkR: CONFIG_TYPEC_MUX_PS883X=m and https://gitlab.freedesktop.org/robclark/msm/-/commits/deleteme (which has a bunch of unrelated stuff I'm working on but you can just cherry-pick the top commit which has the fixup needed for edp backlight to work)
landwork has quit [Quit: Leaving.]
<HdkR>
robclark: Some reference to non-existent node `ec_int_n_default` sadly
<robclark>
not sure what your baseline is.. the branch I'm using is on top of jhovold/wip/x1e80100-6.13
<HdkR>
A bit of a merge conflict but it compiled at least
tobhe_ has joined #aarch64-laptops
<robclark>
you might also be interested in the GMU bw and ACD/opp patches in my branch for moar faster fps (the NArmstrong and APOommen ones)
hexdump01 has quit [Ping timeout: 480 seconds]
<HdkR>
more faster fps is always nice
<robclark>
I think we are still leaving some fps on the table, or at least my attempts to getting LLCC WNA isn't giving the gain that akhil thinks it should.. but at least hitting the higher avail opp's is worth an extra 10% or more
<robclark>
and 10% ain't nothing
<HdkR>
I'm fighting for every percentage over here when a game is at like 55 FPS :D
tobhe has quit [Ping timeout: 480 seconds]
<HdkR>
robclark: Sadly same result
<robclark>
you mean, edp not working too?
<HdkR>
yea
<robclark>
hmm, look at the diff, is it touching anything with edp in the name? External display "works" in the sense that I have one monitor, and it seems to work fine, but that is a limited sample set
<HdkR>
Doesn't seem like it
<robclark>
double check you copied the dtb to wherever, and all that sorta thing?
<HdkR>
Yea, little ritual of modifying grub with a new entry, doing a %s/<old version>/<new version> on the three lines and then :!ls -lah <three files> to make sure the targets exist.
<HdkR>
Oh wait, I didn't have that mux enabled in the config
<robclark>
oh, yeah, you defn need to enable CONFIG_TYPEC_MUX_PS883X (at least as =m, but maybe =y depending on your setup)
<HdkR>
Okay, enabling the mux fixes edp as well
<HdkR>
and I can see DP-1-2-3
<robclark>
\o/
<HdkR>
But I left my type-C to DP adapter somewhere and my USB hub with HDMI doesn't seem to work with it
<HdkR>
Oh wait, I have one that definitely worked with X13s and it isn't working with X1E
<HdkR>
USB and DP prioritizing USB maybe? Do I need an adapter that is only DP-alt?
<robclark>
idk.. the one monitor I tried is usb-c (with an internal hub for eth/etc)... it just worked
<JensGlathe[m]>
What does lsusb -tv say
<robclark>
g-s was a bit finicky with some scaling factors but that seemd to be g-s issues
mbuhl has quit [Remote host closed the connection]
<JensGlathe[m]>
Sounds to me like the smb2360 lines are missing in the dt. On newer kernels you need to explicitly enable them. Otherwise the eUSB2 repeaters are off and the type-c ports don’t work correctly
<JensGlathe[m]>
It is in my originally posted patch
<HdkR>
`OpenGL ES 3.2 Mesa 24.2.3-1ubuntu1` Apparently during an upgrade to oracular, it blew away my mesa install. So I guess I could update and try again
<HdkR>
I track Mesa pretty closely because of gamer tendencies
chrisl has joined #aarch64-laptops
<HdkR>
I wouldn't be surprised if I'm missing some random drivers that need to be enabled but aren't
chrisl has quit [Ping timeout: 480 seconds]
<JensGlathe[m]>
you could just try the already built packages I posted for 6.14-rc1
<JensGlathe[m]>
its ubuntu_x1edefconfig, which is the same but qualcomm only
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
paddymahoney has joined #aarch64-laptops
<albsen[m]>
Jens Glathe: I still cant get the 6.13 kernels to work. the new 6.12 from the repo installed without issues and works.
<JensGlathe[m]>
Have you tried 6.14-rc1
<albsen[m]>
I will now
<albsen[m]>
just saw it
<smoorgborg[m]>
<HdkR> "I track Mesa pretty closely..." <- What do you play on ARM?
<HdkR>
smoorgborg[m]: I don't actually play games on ARM, I just do a ton of game testing on it. Due to FEX's primary focus being games.
<HdkR>
So dogfooding basically
icecream95 has quit [Ping timeout: 480 seconds]
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
srini_ has joined #aarch64-laptops
srini_ has quit [Ping timeout: 480 seconds]
<anthony25>
I'm looking at the bluetooth support for the yoga slim 7x, and I was looking at what was done for CRD and the t14s
<anthony25>
is the bluetooth/wifi card on the t14s removable? or how is it different?
<anthony25>
but I get weird disconnection when idle, I'm wondering if it's not the issue
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
<JensGlathe[m]>
Using the powerseq variant on the HP X14 and ThinkBook 16, works well
<JensGlathe[m]>
Actually using it everywhere
<anthony25>
do you have an example?
<anthony25>
or you mean taking the wcn6855-pmu block of the hp x14 and adapting it for wcn7850?
paddymahoney has quit []
<JensGlathe[m]>
The other way round, but yeah
<JensGlathe[m]>
And also sc8280xp-blackrock, x1p42100-lenovo-thinkbook-16
<JensGlathe[m]>
Oh and yoga 7x 😁
<anthony25>
haha, but I'm doing it for the yoga 7x :D
<anthony25>
I would like to send the patch already included in Ubuntu, but compared to the other laptops, it doesn't have all the regulators
<anthony25>
so I don't really know if it's on purpose of missing
<anthony25>
*or
chrisl has joined #aarch64-laptops
<albsen[m]>
Jens Glathe: tried 6.14-rc1 doesn't work as well. :/
<albsen[m]>
I'll start now trying the older ones, 6.13-3 and so on, lets see
chrisl has quit [Ping timeout: 480 seconds]
<albsen[m]>
Jens Glathe: 6.13-3 is booting 🎉
<albsen[m]>
I'll step up now to 6.14-4 and so on to see when it stops working
<albsen[m]>
meant 6.13
<JensGlathe[m]>
This has the fix from @jhovold
<JensGlathe[m]>
I removed the fix ~6.13-6
<albsen[m]>
so, likely 6.13-4 and 5 should boot as well
<albsen[m]>
will try
patrickm has quit [Ping timeout: 480 seconds]
<JensGlathe[m]>
I will put it in again, need to do a build later anyway
<JensGlathe[m]>
6.14-rc1-1
patrickm has joined #aarch64-laptops
bluerise has joined #aarch64-laptops
<JensGlathe[m]>
anthony25: if there is one missing, it was unintentional.
<albsen[m]>
6.13-4 works as well
jhovold has joined #aarch64-laptops
<JensGlathe[m]>
edp-wwan is for pcie5, not needed
chrisl has joined #aarch64-laptops
srini_ has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
patrickm has quit [Ping timeout: 480 seconds]
patrickm has joined #aarch64-laptops
SpieringsAE has joined #aarch64-laptops
SpieringsAE has quit []
<M31415jd[m]>
Hey, does anybody have any idea on how to approach the trackpad not working issue on the Dell Latitude (https://bugs.launchpad.net/ubuntu-concept/+bug/2084870/comments/6) and Dell Inspiron (my guess is that they are closely related, and a single patch should work for both). The x1e80100-dell-xps13-9345.dtb works reasonably well. I tried to get the hid-descr-addr and reg values from the ACPI table from the device manager in
<M31415jd[m]>
windows. The values I got (0x2c and 0x2 respectively) however don't work.
<M31415jd[m]>
Would love to help with my limited linux and hardware knowledge
SpieringsAE has joined #aarch64-laptops
SpieringsAE has quit []
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
SpieringsAE has joined #aarch64-laptops
<SpieringsAE>
grrrrr this hdmi port on my asus is eluding me
<SpieringsAE>
wait just got an idea
<SpieringsAE>
hahaaa pin 126 is the hotplug detect
<JensGlathe[m]>
Oh
<JensGlathe[m]>
31415jd: this would be 0x2c, 0x20 usually
<JensGlathe[m]>
It was on the thinkbook and also sort of a guess
SpieringsAE has quit [Remote host closed the connection]
SpieringsAE has joined #aarch64-laptops
<SpieringsAE>
hmm enabling mdss_dp2 made the internal display unhappy
<SpieringsAE>
but I feel like there is something else wrong, given all of the display outputs not working anymore
jhovold has quit [Quit: WeeChat 4.4.3]
<Jasper[m]>
@d0pefish nice
<robclark>
d0pefish: cool.. if you can, feel free to reply to the patch with Tested-by tag (idk if you are subscribed to the list or have a setup where you can import an mbox to reply (mutt and I think thunderbird can do this, not sure about others)
<d0pefish>
I would certainly like to do this - I'll need to get a "real" email client set up :) I have just been reading Patchwork and lore.kernel.org through my browser so far
<robclark>
either patchwork or lore should let you download an mbox.. I guess you could even use git-send-email to reply, if you have that setup
<robclark>
anyways we'll try to pull it in for msm-fixes for v6.14
patrickm has quit [Remote host closed the connection]
<anthony25>
by empty I mean everything stays at 0x0
<anthony25>
let me test by building your kernel, that is just with the dtb built from your repo
<SpieringsAE>
it is litterally the only GpioInt definition that has a number there
<JensGlathe[m]>
Hmm I have all type-c ports working on the DevKit with the same definition
<anthony25>
I don't know if the screen only charges the laptop if it detects a display-port device, but only this one port charges when connected to my screen
<anthony25>
(no issue with my charger, though)
<JensGlathe[m]>
but 4k@60 is spicy with 2 lanes
<JensGlathe[m]>
do you have another screen or a dongle with less res?
<anthony25>
shouldn't the laptop be able to support up to 4K 120Hz?
<anthony25>
I didn't test on windows
<JensGlathe[m]>
we have only 2 lanes in Linux yet
<anthony25>
ok
<JensGlathe[m]>
so quite limited
<anthony25>
the screen lists a bunch of resolution, so it should try 1080p 60Hz no?
<anthony25>
I can try on another screen
<anthony25>
but it would be a USB-C to HDMI cable
<JensGlathe[m]>
thats okay
<JensGlathe[m]>
they should work too
chrisl has joined #aarch64-laptops
<anthony25>
ok so I tried on my 1440p 60Hz
<anthony25>
(I'm too fancy, I don't have less :<)
<anthony25>
you're right, it's better
<anthony25>
the 2 ports on the left work
<anthony25>
but the training fails on the right port
<anthony25>
hmm wait, my other 1440p screen has a "console mode" that makes it simulate 4K 30Hz by default
<anthony25>
same thing, the 2 left ports work, not the port on the right
<JensGlathe[m]>
Ok I also have 1080p displays (27" wow)
chrisl has quit [Ping timeout: 480 seconds]
<JensGlathe[m]>
great for testing, otherwise a bit low res
<SpieringsAE>
jeah that is very low dpi damn
<anthony25>
ha wait, I can maybe try a 1080p screen
<SpieringsAE>
I already feel rough on my old 1080p 15.6" laptop
<SpieringsAE>
that thinkbook is also 1080p right?
<JensGlathe[m]>
this link training issue with the repeaters improves steadily, now that many people complain about it
<JensGlathe[m]>
1900x1200 16"
<anthony25>
yeah the link training error happens on the 1080p 60hz screen too
<anthony25>
but the 2 other ports work
<JensGlathe[m]>
thsat's something
<anthony25>
why would the 2 left ports act differently on 4K 60Hz?
<JensGlathe[m]>
different wiring
<JensGlathe[m]>
wire lngths
<SpieringsAE>
retimer should be close to the port though
<SpieringsAE>
and make it more even
<JensGlathe[m]>
As I said, link training is an issue that is hopefully worked on
<anthony25>
I don't have the link training error at all on the 4K 60Hz display
<anthony25>
it just prints `link_ready = false`
<JensGlathe[m]>
Not sure why this
<anthony25>
anyway, I don't think it's an issue with what you added, it works on lower resolution, maybe it's on the edge of being stable by pushing this much
<JensGlathe[m]>
what is your distro, mesa
<anthony25>
tumbleweed, mesa 24.3.4
<JensGlathe[m]>
sounds quite new
<anthony25>
I mean I can try on Ubuntu, but I don't think it's a userland problem
<JensGlathe[m]>
Ubuntu has some crazy patches in there
<SpieringsAE>
tumbleweed is rolling so yeah, its newest mesa
SpieringsAE has quit [Remote host closed the connection]
<anthony25>
I built your kernel, I have the same results
<JensGlathe[m]>
hmm was to be expected
<anthony25>
I mean it looks good, beside the link training problem!
<JensGlathe[m]>
I guess this is where we're at for now
<JensGlathe[m]>
and 4k@60 on 2 lanes is quite something
<JensGlathe[m]>
I get 4k@30
chrisl has joined #aarch64-laptops
<anthony25>
and about the 120Hz, the qualcomm GPU doesn't support DSC?
<anthony25>
or it's not a port limitation and just the GPU cannot display more than 120Hz?
<JensGlathe[m]>
I don't know
<JensGlathe[m]>
with HBR3 and 4 lanes you would have enough bandwidth for 4k@120
<konradybcio>
the GPU has nothing to do with that
<konradybcio>
the fps is probably (lumag, robclark?) only limited by the pixel clock which is effectively A * res * refreshrate
<anthony25>
JensGlathe[m]: if it makes you feel better… my 4K screen only works on 1 port on Windows too
<JensGlathe[m]>
oh interesting
<JensGlathe[m]>
I'm fine anyway
<JensGlathe[m]>
what is it with the othe rresolutions
<JensGlathe[m]>
is only one port HBR3 per chance?
<anthony25>
it doesn't detect the screen at all
<JensGlathe[m]>
144p, 1080p?
<anthony25>
ha you mean my other screens?
chrisl has quit [Ping timeout: 480 seconds]
<JensGlathe[m]>
yea
<anthony25>
that's interesting
<anthony25>
on my 1440p screen that defaults to 4K 30Hz, the 2 left ports work (up to 1440p 90Hz with this cable)
<anthony25>
the right port doesn't work
<anthony25>
let's try on 1080p
<anthony25>
so that's might mean that the link training problem might also happen on windows with this firmware and bios?
<JensGlathe[m]>
I would assume that Lenovo as a member of the VESA club has the resources and will to get the maximum out of their designs
<anthony25>
same thing on 1080p
vanner has joined #aarch64-laptops
vanner has quit [Quit: Leaving]
<anthony25>
I tried another usb-c dock (I don't have a direct cable usb C to HDMI, in fact, so it's an HDMI cable to a USB-C dock), same thing
<anthony25>
and my laptop has the latest bios, afaik
<JensGlathe[m]>
Which reminds me, I wanted to try out that DisplayLink thing
<JensGlathe[m]>
I have a DisplayLink dock
<JensGlathe[m]>
And apparrently there are drivers for it
<anthony25>
<JensGlathe[m]> with HBR3 and 4 lanes you would have enough bandwidth for 4k@120 << yeah, on windows, that's what I get
vanner- has joined #aarch64-laptops
<anthony25>
but my monitor is 4K 240Hz with DSC, so I was wondering where was the limitation
<JensGlathe[m]>
The box I'm sitting on (Windows Dev Kit 2023) has a MiniDP out and it gets 4k@60 onto my Asus VA32??? display, which should be capable of 4k@120
<JensGlathe[m]>
but that being stable and working from DPMS suspend is a quality of life improvement
<anthony25>
so for me, beside using 2 lanes instead of 4, what you did is on par with windows
<anthony25>
I'll keep using your patch, I'll keep you in touch if I encounter any bug with the wifi or bluetooth
<anthony25>
though, to show the stability of windows drivers, just doing this testing broke them and the internal screen renders at the equivalent of a 30Hz :D
<JensGlathe[m]>
quite stable, then
<JensGlathe[m]>
beware of flash-kernel
<JensGlathe[m]>
I had the strangest of experiences
chrisl has joined #aarch64-laptops
<JensGlathe[m]>
I built an image for ASUS Vivobook S 15 X1P-42-100, but I accidentally left /etc/flash-kernel/machine at "Lenovo ThinkBook 16 Gen 7 QOY". Since I don't have the asus, I tried a test boot on the Thinkbook
<JensGlathe[m]>
It booted, rebooted, and started normally with this image, declaring the machine to a ThinkBook (which it is). It added the right dtb from /lib/firmware/<kernel>, using the same link name that was aleady there.
<JensGlathe[m]>
Thinkbook has i2c2 for keyboard and toucpad, so it would be very strange if this worked with the Asus dtb (i2c0 and i2c5?)
<JensGlathe[m]>
Anyway, spooky behaviour
chrisl has quit [Ping timeout: 480 seconds]
<JensGlathe[m]>
So it is enough to set /etc/flash-kernel/machine-id if the name is in the db and an appropriate dtb can be found
<JensGlathe[m]>
albsen: it has the efi boot patch from @jhovold again
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
<wiley[m]>
clover: was checking on the firmware versions on my x13s and noticed I'm still running x13s-firmware because linux-x13s depends on it, but x13s-firmware is from 2024-06, so it doesn't have the updated Bluetooth firmware unless you've shoved it in.
<wiley[m]>
In the bigger picture, though, is there a reason for x13s-firmware to exist at this point or would it be better to just make linux-x13s depend on linux-firmware?
<wiley[m]>
Hrmmm, okay, you're already stuffing bluetooth firmware in there
<wiley[m]>
it looks like it's only needed vs linux-firmware for qcvss8280.mbn at this point?
<wiley[m]>
hrmmm, and a690_gmu.bin as well
<wiley[m]>
(in the big picture, my wifi is still slow and I'm seeing if the updated ath11k firmware makes any difference)