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
unground has quit [Ping timeout: 480 seconds]
Ablu is now known as Guest4101
Ablu has joined #aarch64-laptops
Guest4101 has quit [Ping timeout: 480 seconds]
<clover[m]>
<steev> "clover: okay i'll bring it..." <- Haha twisted your arm did I?
<steev>
actually, i assumed it was causing my usb issues, but they were caused by the iommu skip stuff that's supposed to increase throughput
<Dylanger>
Prob not the right place to ask but, does anyone know of an OpenWRT room? I have a Ubquiti U6-Pro that needs liberating
rz has quit [Remote host closed the connection]
rz has joined #aarch64-laptops
unground has joined #aarch64-laptops
exeat has quit []
todi has joined #aarch64-laptops
jhovold has quit [Quit: WeeChat 4.0.4]
jhovold has joined #aarch64-laptops
<HdkR>
_[m]1: What errors from AUR for FEX?
<HdkR>
When I update the package I usually compile test it so maybe this last month it broke?
<Jasper[m]>
I'm contemplating replacing my Arch install with Ubuntu
<Jasper[m]>
Because the crashes I have are EXTREMELY annoying
<Jasper[m]>
Doesn't help that I also think Ubuntu is extremely annoying hahahaha
todi has quit [Ping timeout: 480 seconds]
<_[m]1>
<HdkR> "When I update the package I..." <- so I could get the git one and put it month back dunno
<_[m]1>
do all of you use gnome btw? can't get any other DE to function, sway is unresponsive, kde plasma stuck in loading
<_[m]1>
in the end I built fex manually which worked (shrug
<_[m]1>
* in the end I built fex manually which worked :shrug: now just struggling to get the proper chroot configured, the only x86 is an outdated linux mint install I had laying around
<konradybcio>
Jasper: i updated concept to 23.10 and got 2 or 3 random crashes too..
<_[m]1>
it's really badly documented imo
<konradybcio>
same with my own next-based kernel
<konradybcio>
the old ubuntu 6.2 was flawless on that front, i think
<HdkR>
_[m]1: Interesting. If you pull the FEX-2310 tag then it should be the same one that is in the AUR
<_[m]1>
Jasper which crashes do you have?
<emily[m]>
_[m]1: had kde running mostly fine out of the box with that old debian image linaro(?) released a while back at least
<_[m]1>
HdkR oh yeah didn't pay attention to the branch lol
<HdkR>
_[m]1: We do a release every month, which is usually better for the user since we do some testing to ensure the world isn't broken :D
hexdump01 has quit []
<Jasper[m]>
_[m]1: On boot, the LUKS thing
hexdump0815 has joined #aarch64-laptops
<Jasper[m]>
I just hope I did something wrong on Arch that gets fixed with Ubuntu or any other distro
todi has joined #aarch64-laptops
<_[m]1>
<Jasper[m]> "On boot, the LUKS thing" <- if it was only you, I'd agree but both of us?
<_[m]1>
once the kernel stabilises you don't restart that often - btw I bought a second harddrive for installing arch, I have windows/ubuntu dual boot on the original and hmyeah, I was sick of it too couple of times but always go back because ubuntu is annoying in small things
<_[m]1>
though the FDE just works and no reboots, I maybe should go and look at the diffs in config files dunno - but maybe it's pebcac since clover couldn't reproduce
<_[m]1>
maybe it's not fde related but something else? because linux starts booting - last time didn't work to make journalctl persistent though
<_[m]1>
with yay, non git one, the git one had some error about a missing commit it was looking for in one of the dependency repos
<HdkR>
Oh yea, I don't maintain the -git one. I only found out about it like a month ago
<HdkR>
So that looks like a clang and lld version mismatch
<_[m]1>
but they're installed in cache of yay right, it doesn't use system ones?
<HdkR>
I don't know what yay is
<_[m]1>
package manager to install from AUR
<_[m]1>
* to install from AUR
<HdkR>
`/usr/bin/clang++` seems to imply system one, but maybe it uses some chroot thing
<_[m]1>
maybe it dóes also use system build toosls
<_[m]1>
s/toosls/tools/
<HdkR>
You can confirm by checking your clang and lld versions, this new ABI was only added in 17. `clang --version` and `ld.lld --version`
<_[m]1>
yes both are 16
<HdkR>
So they're doing *something*
<clover[m]>
Yeah I maybe got the reboot thing once or twice. I do get crashes every now and then where everything locks up and I have to reboot, about twice a week
<_[m]1>
so it ís FDE related huh?
<clover[m]>
locks up as in it freezes
<clover[m]>
i can move mouse but i cant run programs and all running programs are unresponsive
<clover[m]>
and i have to hard reboot to fix it
<clover[m]>
i dont think i have FDE issues like you guys
<_[m]1>
lol I have to build the build tools?
<Jasper[m]>
@clover @steev can you check if efi pstore (if it even works on this hw) can be enabled?
<Jasper[m]>
I feel like that could help out
<_[m]1>
clover[m]: so what's the reboot thing then? or you don't do FDE on your system partition?
<Jasper[m]>
Jasper[m]: Hopefully it dumps the logs into that
<agl7>
clover[m]: How can I turn on all messages at boot time on EndeavourOS so that I can see the file system check that I do at every boot. This is done in the "dark".
<clover[m]>
id o
<clover[m]>
* i do
<_[m]1>
anybody wants to share the x86 chroot they made for fex-emu?
<_[m]1>
or some pointers how you build, like you use schroot or debootstrap
<_[m]1>
> Solid state drive users should be aware that, by default, TRIM commands are not enabled by the device-mapper, i.e. block-devices are mounted without the discard option unless you override the default.
<HdkR>
Snapdragon 8 Gen1 and newer and Snapdragon 7 Gen1 has it
<_[m]1>
it takes a long time maybe to build and test the hardware around them?
<HdkR>
X13s is just running a relatively old SoC at this point
<HdkR>
Pi5 probably chose an older design to keep the price lower
<_[m]1>
yeah this I read too in rockpro room, why not more pcie lanes e.g., price related
<_[m]1>
I'd pay more though 🤷
<HdkR>
Rockchip, Allwinner, and Amlogic tend to take a while to release new SoCs and are usually more concerned about price as well
<_[m]1>
https://mntre.com/reform.html anyone has experience with this one? I met the creator(s) at cccamp and it felt like a very nice device - ofc not super practicle but many good design choices imo (e.g. interchangeable CPU module)
<steev>
lots of people say they would pay more... most people actually don't
<HdkR>
It's an i.MX8 module, so if you want super old and or slow it might be fine for you :)
<_[m]1>
* CPU module, replaceable batteries)
<_[m]1>
I'd buy one just for supporting it and fun - not much real use for it
<_[m]1>
I'm pretty happy all my personal devices have ARM cpus now I think of it 😄
<_[m]1>
except the gaming desktop mm
<_[m]1>
but why do people buy iPhones though? 1k for a phone?!
<steev>
that's up to each person, for some it's a status symbol, for some it's because android annoys them so much that if they had an android phone no one would ever be able to talk to them on it because it would always be in a state of getting a new fix implemented
<_[m]1>
Jasper you use btrfs?
<_[m]1>
I would think (rather) superior hardware and user 😝
<_[m]1>
😝
systwi has quit [Ping timeout: 480 seconds]
<_[m]1>
duuude, I removed taht comment and added allow_discard and it booted immediately 😲
systwi has joined #aarch64-laptops
<altacus>
steev: for 6.5.8, should we still use iommu.passthrough=0 ?
<steev>
if you want to
<steev>
i don't because i see issues with my usb->nvme adapter
<altacus>
steev: what are the benefits of setting it?
<steev>
more usb throughput
<altacus>
Thank you!
<Jasper[m]>
<_[m]1> "Jasper you use btrfs?" <- nope, good 'ol ext4
<Jasper[m]>
<Jasper[m]> "I'm contemplating replacing my..." <- ok nvm
<Jasper[m]>
Seems like something went horribly wrong
exeat has joined #aarch64-laptops
<Jasper[m]>
Yeah okay my usb dock seems to reset during boot, I have a passive adapter that does work it seems.
ungeskriptet has quit [Quit: Ping timeout (120 seconds)]
<_[m]1>
was that the cause
<_[m]1>
?
<DanaG>
I had split my GRUB_CMDLINE_LINUX onto multiple lines by setting GRUB_CMDLINE_LINUX="${GRUB_CMDLINE_LINUX} more args ..." on several lines, and each time I did a major-version update, the updater script would convert every one of those lines into a copy of the single string with all lines merged.
<DanaG>
As far as arm64 machines go, I got that Honeycomb to act as a desktop, but then ran into GPU troubles. Old AMD: works, but has glitches (due to bad ordering of instructions in memcpy), and garbled HDMI audio, and no 4k120 output. New AMD: PSP firmware fails to load, perhaps due to unaligned access or something. Nouveau: slow. NVIDIA: falls off the bus. Intel Xe driver: haven't tried it, but no encode/decode support.
<DanaG>
It wouldn't surprise me if an Asahi machine turns out to be far better than anything I've tried with the LX2160A.
<Jasper[m]>
<_[m]1> "was that the cause..." <- > <@nino:matrix.libre.brussels> was that the cause
<Jasper[m]>
> ?
<Jasper[m]>
Yeah, it was complaining about not being able to find some stuff
<Jasper[m]>
It's installed now, I gagged at GNOME
<Jasper[m]>
There was no mouse acceleration option for the trackpad and it was HORRIBLE
<Jasper[m]>
I'm installing kubuntu instead
<steev>
gnome-tweaks has it hidden away, i think
<Jasper[m]>
Hmmm 3.5mm jack output is very much broken on ubuntu
<_[m]1>
<Jasper[m]> "I'm installing kubuntu instead" <- it does wayland? dunno I can see some workflow with gnome but minimally more what you can't with others - perhaps it grows on you?
<Jasper[m]>
_[m]1: Oh yeah, for a while
<_[m]1>
<steev> "gnome-tweaks has it hidden away,..." <- it's generally under accessability now or no only key repeat
<Jasper[m]>
GNOME hasn't grown on me, it feels very clunky to me
<Jasper[m]>
I've used it at work for a bit
<_[m]1>
i was checking hyprland, flashy stuff but you can't log into it seems
<_[m]1>
```
<_[m]1>
$ sudo ls /var/lib/binfmts/FEX*
<_[m]1>
ls: cannot access '/var/lib/binfmts/FEX*': No such file or directory
<_[m]1>
do I need that?
<HdkR>
Why is it in `/var/lib/binfmts/` ?
<HdkR>
If it is registered in the kernel you'll see `/proc/sys/fs/binfmt_misc/FEX*`
<HdkR>
binfmt_misc files end up in `/usr/share/binfmts/` here
<HdkR>
Looks like they did some manual stuff just so they could install some random emulator
<steev>
yell at clover[m]
<_[m]1>
HdkR: hehe I know but I thought this was x13s specific
<HdkR>
Nah, that's just regular stuff
<_[m]1>
oh my 😅
<_[m]1>
I just want to install steam / dota2 lol
<_[m]1>
I guess that won't be straight forward still
<_[m]1>
this fex-emu looks super complicated what 😲 I sometimes wonder if developing this kind of stuff is like slow and painful or oyu just get good and it's like fun
<HdkR>
A lot of these complications are coming from misunderstandings and bad documentation sadly
<_[m]1>
who needs docs when you have irc logs 😉
<_[m]1>
I haven't improved any doc so I can't complain really
<_[m]1>
and in the end glad to learn new shit which the sidequests inevitably make sure of 🙂
<Jasper[m]>
<_[m]1> "I just want to install steam..." <- I don't think dota2 will work right? @HdkR
<Jasper[m]>
Since it's Source 2
<_[m]1>
it is
<_[m]1>
> Dota 2 is now powered by the Source 2 engine
<HdkR>
The `steam` application that gets installed is actually just a bash script, so it needs to be wrapped in the `FEXBash` program so it runs in an x86-like environment
<_[m]1>
yeah I did that first but didn't launch and still doesn't, probably my fex-emu build is incomplete?
<_[m]1>
maybe I'll try to get clang17 installed
<HdkR>
Steam didn't launch at all under FEXBash? Usually FEX complains if something isn't setup early
<HdkR>
Just an early exit or something?
<_[m]1>
no real errors no
<HdkR>
hm, FEX_OUTPUTLOG=stderr and FEX_SILENTLOG=0 and then rerun under FEXBash