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)
e1eph4nt has joined #aarch64-laptops
e1eph4nt has quit [Ping timeout: 480 seconds]
e1eph4nt has joined #aarch64-laptops
e1eph4nt has quit [Ping timeout: 480 seconds]
e1eph4nt has joined #aarch64-laptops
hexdump0815 has joined #aarch64-laptops
hexdump01 has quit [Ping timeout: 480 seconds]
pundir has joined #aarch64-laptops
robclark has joined #aarch64-laptops
arnd has joined #aarch64-laptops
_alice has joined #aarch64-laptops
exit70 has joined #aarch64-laptops
jhovold has joined #aarch64-laptops
e1eph4nt has quit [Ping timeout: 480 seconds]
e1eph4nt has joined #aarch64-laptops
e1eph4nt has quit [Ping timeout: 480 seconds]
e1eph4nt has joined #aarch64-laptops
jhovold has quit [Quit: WeeChat 3.5]
jhovold has joined #aarch64-laptops
e1eph4nt has quit [Ping timeout: 480 seconds]
e1eph4nt has joined #aarch64-laptops
rpirea__ has quit []
rpirea has joined #aarch64-laptops
iivanov has joined #aarch64-laptops
e1eph4nt has quit [Ping timeout: 480 seconds]
e1eph4nt has joined #aarch64-laptops
e1eph4nt has quit [Ping timeout: 480 seconds]
e1eph4nt has joined #aarch64-laptops
e1eph4nt has quit [Ping timeout: 480 seconds]
e1eph4nt has joined #aarch64-laptops
e1eph4nt has quit [Ping timeout: 480 seconds]
e1eph4nt has joined #aarch64-laptops
e1eph4nt has quit [Ping timeout: 480 seconds]
e1eph4nt has joined #aarch64-laptops
hexdump01 has joined #aarch64-laptops
hexdump0815 has quit [Ping timeout: 480 seconds]
<jenneron[m]>
<qzed> "on the sc8180x GFX_INTERRUPT is..." <- GIC_SPI interrupts are +32 in ACPI
e1eph4nt has quit [Ping timeout: 480 seconds]
e1eph4nt has joined #aarch64-laptops
e1eph4nt has quit [Ping timeout: 480 seconds]
e1eph4nt has joined #aarch64-laptops
e1eph4nt has quit [Ping timeout: 480 seconds]
e1eph4nt has joined #aarch64-laptops
hexdump0815 has joined #aarch64-laptops
hexdump01 has quit [Ping timeout: 480 seconds]
e1eph4nt has quit [Ping timeout: 480 seconds]
hexdump01 has joined #aarch64-laptops
hexdump0815 has quit [Ping timeout: 480 seconds]
e1eph4nt has joined #aarch64-laptops
xroumegue has quit [Ping timeout: 480 seconds]
xroumegue has joined #aarch64-laptops
hexdump0815 has joined #aarch64-laptops
hexdump01 has quit [Ping timeout: 480 seconds]
hexdump01 has joined #aarch64-laptops
hexdump0815 has quit [Ping timeout: 480 seconds]
hexdump0815 has joined #aarch64-laptops
hexdump01 has quit [Ping timeout: 480 seconds]
e1eph4nt has quit [Ping timeout: 480 seconds]
e1eph4nt has joined #aarch64-laptops
iivanov has quit [Quit: Leaving...]
e1eph4nt has quit [Ping timeout: 480 seconds]
e1eph4nt has joined #aarch64-laptops
e1eph4nt has quit [Ping timeout: 480 seconds]
e1eph4nt has joined #aarch64-laptops
JoshuaAshton has joined #aarch64-laptops
<JoshuaAshton>
Hello!
<JoshuaAshton>
Anyone here interested in Samsung Galaxy Book S?
e1eph4nt has quit [Ping timeout: 480 seconds]
<steev>
there are a few people
e1eph4nt has joined #aarch64-laptops
<HdkR>
More compute class chips to figure out the quirks of :P
<JoshuaAshton>
HdkR: Hi :frog:
<JoshuaAshton>
Trying to figure out why when foollowing the exact same steps as everyone else, my Book S doesn't boot into gdm/display stuff properly and just hangs. Whereas other people have gotten a working desktop and/or can switch to tty (doesn't work, already hung here)
<JoshuaAshton>
inb4 entirely different sku of laptop somehow :S
<JoshuaAshton>
Wouldn't surprise me...
<robclark>
could be a different panel.. do you get fbcon? Can you ssh to it? In _theory_ different panel should be ok, as long as the power-on delay is enough and the panel has a valid edid. But perhaps theory doesn't always match reality?
<JoshuaAshton>
I get fbcon, yes, but right now it's trying to boot into gdm or something then going black with a `_` permenantly and I can't go to TTY
<JoshuaAshton>
I don't have anything I know I can easily chroot with for aarch64 like I do a regular PC hmb
<JoshuaAshton>
Really don't know what I am doing in the arm space yet if I am honest, just trying to figure stuff out with this laptop so far
<robclark>
oh, with ACPI boot you don't get any of the nice shiny things
<JoshuaAshton>
I think someone made a device tree for this laptop, I'll have to take a look at that
<JoshuaAshton>
If I am honest I really dont get the point of device tree, why can't things just use ACPI like Windows + Linux on x86?
<HdkR>
It's probably booting ACPI since that ISO won't have a DT for the device. So it'll need to get compiled, installed, and then set to be used
e1eph4nt has quit [Ping timeout: 480 seconds]
<JoshuaAshton>
I love that /sys/class/power_supply is empty, that isnt terrifying at all... Guess that needs DT too? :S
<robclark>
The major thing blocking ACPI boot w/ gpu/video/cpufreq/etc/etc is that windows on arm things uses PEP thing for providing all clk/bandwidth/power stuff, and linux completely doesn't support such a mechanism.. but there are a lot of other ways where various kernel frameworks we depend on are tied heavily to DT so even with PEP it would be a lot of work to get reasonable ACPI support
<JoshuaAshton>
Apparently there is also GRUB_DEVICETREE hm
<robclark>
yeah, that sounds about right, I think.. or at least the `devicetree /path/to/some.dtb` part sounds right
<JoshuaAshton>
I guess I will try it, seems weird that it wouldn't need to be special-cased like initrd stuff though, so my expectations are currently low
<HdkR>
Does `devicetree <dt>` work as a linux command line option? I thought it needed to be on a new line like initrd?
<JoshuaAshton>
Thats what I was thinking
<robclark>
I mean, it is a grub cmd
<JoshuaAshton>
ah
<robclark>
oh, ic
<JoshuaAshton>
but I am not sure how to specify that in /etc/default/grub config... thats easy to do in grub.cfg
<robclark>
yeah, it needs to be on newline
e1eph4nt has joined #aarch64-laptops
<JoshuaAshton>
Doesnt seem like mkconfig supports it that I can tell... :S
<JoshuaAshton>
I'll derp it in .cfg manually for now... maybe I am missing something obvious
<HdkR>
I modified some file manually to input the dtb I care about. that way each time I make install I don't need to type it again
<steev>
JoshuaAshton: jenneron[m] does the gbs stuff iirc
<steev>
and pierro78 is the guinnea pig using it all
<jenneron[m]>
<JoshuaAshton> "Trying to figure out why when..." <- gdm used to take about 2 minutes for some reason, you can start gnome manually from tty
<JoshuaAshton>
Ah
<JoshuaAshton>
jenneron[m]: I tried your kernel and it just hangs for me, on boot, doesnt even get to fbcon or anything... I am using the device tree posted too fwiw
<JoshuaAshton>
the 5.19 one
<JoshuaAshton>
Maybe I am missing something
<JoshuaAshton>
I still have 5.11 and it boots fine (with the two USB thing)
<jenneron[m]>
use my kernel + dtb from the tree + galaxy-book-s_defconfig
<JoshuaAshton>
Which is apparently just a build of yours
<jenneron[m]>
note that my galaxy book s now suffers from graphical glitches, but i'm not sure whether it's related to linux efforts or not
<jenneron[m]>
but these glitches are permanent (in windows too), i had to order a new laptop
<JoshuaAshton>
Something something bad PLL clock broke display? :P
<JoshuaAshton>
I have had that on a Lumia before :(
<jenneron[m]>
i don't know what broke it, i'm more thinking of improper usage of voltage + scaling on GPU, since it seems to be different on each boot
<jenneron[m]>
but, again, i don't know for sure what broke it
<JoshuaAshton>
yeah no display with your 5.19 kernel here, unless its been updated since that build was posted
<jenneron[m]>
hopefully, we will get DVFS for GPU, etc, we will get better implementations of everything, and, hopefully, possibility to break a device will naturally disappear
<jenneron[m]>
JoshuaAshton: display and gpu should work since i posted it here
<JoshuaAshton>
well i dont even get fbcon, it doesnt end up booting with it :(
<jenneron[m]>
you can clone my branch and build it yourself (git clone https://gitlab.com/jenneron/linux -b galaxy-book-s-5.19.0 --depth 1) and build it yourself
<jenneron[m]>
i can't help with those builds, but pierro78 may help you
<jenneron[m]>
also, make sure your grub.cfg provides device tree, kernel config, this thing built with, doesn't support ACPI
<JoshuaAshton>
yeah it does
<JoshuaAshton>
same thing targeting 5.11 adds a bunch of constant error spam so its definitely there
<jenneron[m]>
can you send your grub.cfg
<JoshuaAshton>
sec
<jenneron[m]>
also, make sure you provide efi=novamap cmdline
<JoshuaAshton>
Then I copied the kernel over and ran `update-initramfs`
<jenneron[m]>
and modules?
<JoshuaAshton>
Do I need to do more than that for modules? AFAIU that's all I should need to do right?
<jenneron[m]>
for modules, you may also want to add something like KBUILD_BUILD_VERSION="5.19.0-galaxy-book-s"
<jenneron[m]>
JoshuaAshton: modules are in /lib/modules
<JoshuaAshton>
Derp. Okay
rpirea_ has joined #aarch64-laptops
rpirea has quit [Read error: Connection reset by peer]
<JoshuaAshton>
Yeah no output on this kernel still
<steev>
it's a debian system?
<steev>
instead of copying stuff over randomly, just build it with the bindeb-pkg target, and at the end of the build it'll spit out a deb file you can dpkg -i
<steev>
so like; make ARCH=arm64 CROSS_COMPILE=aarch64-linux-gnu- galaxy-book-s_defconfig; make ARCH=arm64 CROSS_COMPILE=aarch64-linux-gnu- -j$(nproc) bindeb-pkg
<JoshuaAshton>
Cool
<JoshuaAshton>
I will try that
<steev>
that's how i do it with my (non galaxy-book-s devices)
<JoshuaAshton>
wasn't aware bindeb-pkg was a thing. I am not a debian user by trade :b
<JoshuaAshton>
Had to make it specify -d to dpkg-buildpackage as obviously it doesnt think the deps are installed haha
<JoshuaAshton>
What's the correct way to build the dts -> dtb? I tried
<JoshuaAshton>
But it just throws that it has a syntax error
<HdkR>
it should get built when you build the kernel if the option is enabled
<HdkR>
ends up in the same directory as the dts
<JoshuaAshton>
ah dtbs_install, thanks
<JoshuaAshton>
__
<JoshuaAshton>
Yeah exact same result with building kernel + own dtb this way
<JoshuaAshton>
Just black screens forever, no kernel output
<clover[m]>
steev: anything cool in rc5?
<steev>
probably not
<steev>
for the most part, rcX is just bug fixes of rcX-1
<steev>
new goodies might be in 6.1
<steev>
unless someone posts something to the list, or brings it to my attention :D
klardotsh has joined #aarch64-laptops
<steev>
might check out jhovold's irq patches though
e1eph4nt has quit [Ping timeout: 480 seconds]
<JoshuaAshton>
jenneron[m]: Any more ideas? I tried removing quiet and stuff to grab more info but it doesnt seem to even get that far... This is just automatically added to grub from update-grub and me adding devicetree param to grub.cfg so theres no typos or whatever.
<jenneron[m]>
JoshuaAshton: does usb work?
<jenneron[m]>
you can connect an ethernet adapter or a phone as modem
<jenneron[m]>
and ssh into it
<JoshuaAshton>
No, that doesnt work
<JoshuaAshton>
As I said it fails really early
<jenneron[m]>
so not only display?
<JoshuaAshton>
Correct. It doesnt even get to fbcon or anything
<jenneron[m]>
I mean besides display, usb doesn't work too?
<jenneron[m]>
no screen output doesn't always mean that kernel is not booted
<JoshuaAshton>
Nothing works. It boots to a black screen
<jenneron[m]>
also, what messages does EFI stub print?
<JoshuaAshton>
there are no messages, the last thing visible is grun
<JoshuaAshton>
grub
<jenneron[m]>
does it print anything useful?
<JoshuaAshton>
Aside from `Booting blah kernel`, no
<JoshuaAshton>
5.11 works fine
e1eph4nt has joined #aarch64-laptops
<qzed>
do you have efi=novamap set? although I think it should have printed something before then...
<JoshuaAshton>
yes
<JoshuaAshton>
Sometimes it doesnt even go to a black screen and hangs on the last grub frame
<steev>
earlycon=efifb ?
<JoshuaAshton>
it's a 50/50 dice roll
<JoshuaAshton>
will try that
<steev>
also console=efifb
<JoshuaAshton>
oh that prints stuff
<JoshuaAshton>
then it goes black
<JoshuaAshton>
:frog:
<steev>
time to break out the vidya recorder
<JoshuaAshton>
yea haha
<qzed>
or maybe the boot_delay kernel parameter
<qzed>
or both
<steev>
or maybe blacklist msm in case it's getting that far and trying to switch over, and video=efifb
<JoshuaAshton>
I set it to 500 sooo should be 500ms but it didnt slow anything down :/
<steev>
that's why i suggest vidya
<jenneron[m]>
it needs CONFIG_BOOT_PRINTK_DELAY enabled
<JoshuaAshton>
ugh
<JoshuaAshton>
maybe it pooped out to dmesg of last boot?
<JoshuaAshton>
doubtful...
e1eph4nt has joined #aarch64-laptops
<jenneron[m]>
<steev> "or maybe blacklist msm in case..." <- I doubt it will help since even usb doesn't work
<jenneron[m]>
also, maybe stupid question, but GPU firmware is not required for drm + panel, right?
<steev>
i'm just going on the assumption that the "black screen" is what happens when the efifb -> msm changeover is supposed to happen
<steev>
you should be able to have no gpu if you have no gpu enabled and amd,imageon in the dtsi, like how sc8280xp does it
<JoshuaAshton>
Looks like it dies shortly after usb init by the log
e1eph4nt has quit [Ping timeout: 480 seconds]
e1eph4nt has joined #aarch64-laptops
<robclark>
for display without gpu, you don't want "amd,imageon"
<robclark>
for gpu without display, you do
<robclark>
so steev has that backwards ;-)
<robclark>
to have display without gpu, just skip having gmu fw ;-)
<bamse>
status = "disabled" on the gpu nodes works too
<bamse>
not sure when that started working, but thanks robclark :)
<JoshuaAshton>
well, what should I try next then, hearing conflicting things
<robclark>
bamse: seems like probably 9ca7ad6c7706edeae331c1632d0c63897418ebad .. so thank jhugo ;-)
<steev>
oh
<steev>
this is why you shouldn't trust my judgement :D
<steev>
i also don't quite follow how i have gpu without display, when i have display without gpu, because we've got llvmpipe
<steev>
disclaimer that i'm not a very smart man
<robclark>
so.. if this is a690 and you haven't recompiled mesa yourself to add a690 to the device table, that would be one reason you could end up w/ llvmpipe
* robclark
loosing track of who is working on which laptop(s)
<steev>
yeah, mesa doesn't have a690, and we don't have gpu either
<steev>
galaxybook is a680 though, and it *should* be possible to get gpu there, i do on my flex 5g at least
falk689 has quit [charon.oftc.net dacia.oftc.net]
<robclark>
yeah, ok, current mesa shouldn't be the issue for a680
<jenneron[m]>
<JoshuaAshton> "Looks like it dies shortly after..." <- anything useful in logs?
<JoshuaAshton>
the last thing printed before it dies in my video is a line about USB
<jenneron[m]>
I really don't like the glitch on last seconds
<jenneron[m]>
it looks like my glitches which now happen even in windows
<JoshuaAshton>
well it shouldnt be trying to do anything whacky with the display, i blacklisted msm and video=efifb :/
<JoshuaAshton>
the only time ive seen displays break is from PLL clock stuff getting involved which shouldnt be happening here
<JoshuaAshton>
jenneron[m]: Do you have any other kernel branches? Maybe something 5.15 or around that time to try and see if it has the same issue?
<JoshuaAshton>
id be kind of weirded out if a display like this was able to be killed/damaged by sw, I've abused the shit out of my Steam Deck's display when bringing up the dynamic refresh rate stuff with all sorts of crazy stuff and its been fine
<jenneron[m]>
weird, but the fact is that my laptop is broken
<JoshuaAshton>
idk what tech the screen on the Galaxy Book S, uses though... on Deck it's a eDP -> MIPI bridge
<JoshuaAshton>
yeah, ends after USB stuff
<JoshuaAshton>
I tried with all USB stuff disconnected too
<jenneron[m]>
it's just eDP here
<JoshuaAshton>
well, eDP internally or does it also go through some bridge... itll just show as eDP otherwise -- I can find out once I have DRM working in any way here :-P
<jenneron[m]>
JoshuaAshton: it's not the problem
<JoshuaAshton>
The USB stuff, yeah...
<JoshuaAshton>
Just not sure where to go from here cause I can't get any more details. Right pita
<jenneron[m]>
reading previous dmesg logs, it happened before, i suppose it's because i don't have usb <-> dp switch in dts
<jenneron[m]>
JoshuaAshton: try to dump ACPI tables from windows and decode edid form there, make sure your timings are same
<jenneron[m]>
it's unlikely, but possible that you have another display
<JoshuaAshton>
I already nuked Windows :(
<jenneron[m]>
we also can try to switch to using generic compatible which should read timings
<JoshuaAshton>
oh did you bake an edid in the dts?
<JoshuaAshton>
or something?
<jenneron[m]>
i use compatible with panel which has same timings
<JoshuaAshton>
ah
<jenneron[m]>
obviously reading edid is better, but i had some issues with that
<jenneron[m]>
iirc it worked, but there was some traceback
<JoshuaAshton>
this shouldnt be use with efifb and nomodeset though...
<JoshuaAshton>
and im still getting blackscreen forcing efifb and nomodeset
<jenneron[m]>
i think efifb disabled in that kconfig
<jenneron[m]>
JoshuaAshton: well it shouldn't be a problem for working panel
<JoshuaAshton>
right but efifb is really useful anyway!
<JoshuaAshton>
`# CONFIG_FB_EFI is not set`
<jenneron[m]>
well... i worked on panel relying on someone else's testing, so i disabled EFI FB to get response from them like "display works" and i could be sure that it's my panel implementation, not EFI FB
<jenneron[m]>
JoshuaAshton: i don't think so, it will die like earlycon or will be replaced by qcom drm if panel works
<jenneron[m]>
i suggest you to change compatible to "edp-panel" and rebuild dtb, it doesn't require rebuilding kernel image or modules
<JoshuaAshton>
yeah im doin now
<jenneron[m]>
and of course test it without blocking with cmdline
<JoshuaAshton>
No change
<jenneron[m]>
bad
<jenneron[m]>
are you sure when you connect a phone to second usb, it doesn't suggest to transfer data or anything like this?
<JoshuaAshton>
hm it does
<jenneron[m]>
nice
<JoshuaAshton>
i was testing by looking at ethernet lights
<JoshuaAshton>
on my dock
<JoshuaAshton>
but yeah phone works
<jenneron[m]>
boot into 5.11, install openssh-server
<JoshuaAshton>
if this was a chromebook i would be able to grab serial from it :(
<JoshuaAshton>
I have openssh installed
e1eph4nt has quit [Ping timeout: 480 seconds]
<JoshuaAshton>
and ssh all set up
<jenneron[m]>
JoshuaAshton: this kcondig may lack configs for your Ethernet adapter
<jenneron[m]>
but you can connect with phone
<JoshuaAshton>
jenneron[m]: That would suck because it's just the Steam Deck Dock that works on the default 5.11 kernel
<JoshuaAshton>
jenneron[m]: How? Never done that before.
<jenneron[m]>
for Ethernet adapter enable CONFIG_USB_NET*
<jenneron[m]>
for phone, enable it as usb modem, run ip a, find ip, run nmap -sn x.y.z.0/24 and you will get an ip
<JoshuaAshton>
lol nmap is broke in termux
<JoshuaAshton>
I will just do a new build with all the USB_NET stuff enabled...
<jenneron[m]>
<JoshuaAshton> "if this was a chromebook i would..." <- unfortunately I couldn't find schematics for this
<jenneron[m]>
there were ones for surface pro x and flex 5g, but not for galaxy book s
<jenneron[m]>
you can try to find it though
falk689 has quit [Remote host closed the connection]
<JoshuaAshton>
I will just do a new build with all the USB_NET stuff enabled...
<JoshuaAshton>
idk why that sent twice oops
<JoshuaAshton>
I did that, still no good rip
<jenneron[m]>
you can boot to 5.11 and check dmesg to find which driver does it use
<JoshuaAshton>
I enabled all of them :b
<JoshuaAshton>
ok this boot USB didnt work at all
falk689 has joined #aarch64-laptops
<JoshuaAshton>
Ethernet lights on the adapter... No route to host :S
<jenneron[m]>
you can boot into 5.11 and set up some logging
<jenneron[m]>
like to boot into 5.19, then reboot into 5.11 and read previous logs