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
<kuruczgy[m]>
[ 0.156544] CPU: All CPU(s) started at EL2
<kuruczgy[m]>
Success
<kuruczgy[m]>
hogliux: Bluetooth does work for me in EL2
krei-se has quit [Ping timeout: 480 seconds]
<kuruczgy[m]>
Next question: what's needed to get /dev/kvm?
<kuruczgy[m]>
maz: Are there any KVM patches that are not in Johan's x1e branch? (E.g. I heard that you even got nested virt working?)
<robclark>
I think just enable it in kernel config.. unless you need nested virt, then you probably need maz's patch
chrisl has joined #aarch64-laptops
<kuruczgy[m]>
Oh yes you are right, CONFIG_KVM is not even enabled 🤦. I don't see why it is not though, I guess time to debug that first.
krei-se has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
<kuruczgy[m]>
Hm DP altmode does not seem to be working in EL2. Does that also depend on the aDSP?
<farchord>
Hey guys, how's the X1E situation in 6.13? Do we still need jhovold's kernel?
<kuruczgy[m]>
I don't think you "need" it. I think I even booted mainline 6.11 successfully, and I am sure 6.12 mainline works too. But it still has useful fixes (and I do carry quite a few patches on top of Johan's branch as well).
<farchord>
10-4. Is there updates on audio support?
<kuruczgy[m]>
I don't think anyone is working actively on it right now unfortunately. AFAIK the current situation is bad quality audio, without speaker protection. For me bluetooth is just easier to use at that point.
<ppd[m]>
something something I need to not be doing exams to do speaker protection stuff
<ppd[m]>
(sorry kuru)
<robclark>
kuruczgy[m]: yeah, I think DP depends on adsp, which seems to be doing some DP related things that I'd normally see in the EC (I guess adsp is in the low power island and can do EC things without a power cost, allowing for a cheaper EC to reduce BOM cost)
<robclark>
so on x13s at least, that external monitors didn't work when booted in EL2
hexa- has quit [Quit: WeeChat 4.4.3]
hexa- has joined #aarch64-laptops
davidinux is now known as Guest1169
davidinux has joined #aarch64-laptops
Guest1169 has quit [Ping timeout: 480 seconds]
chrisl has joined #aarch64-laptops
weirdtreething has quit [Remote host closed the connection]
chrisl has quit [Ping timeout: 480 seconds]
weirdtreething has joined #aarch64-laptops
weirdtreething has quit [Remote host closed the connection]
weirdtreething has joined #aarch64-laptops
<kuruczgy[m]>
Well, starting QEMU with KVM crashes the machine consistently :/
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
tobhe_ has joined #aarch64-laptops
tobhe has quit [Ping timeout: 480 seconds]
<steev>
progress
<bamse>
robclark: tcpm is implemented in adsp firmware, so you need remoteproc to have working altmode
<macc24>
robclark: so like the intel management engine but more annoying?
<robclark>
idk about _more_ but otherwise I guess so.. hw folks in chase of the almighty BoM do what they do
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
hexdump0815 has joined #aarch64-laptops
hexdump01 has quit [Ping timeout: 480 seconds]
deathmist has quit [Ping timeout: 480 seconds]
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
<JensGlathe[m]>
kuruczgy I'm using lxc/lxd on the EL2 box, only additional kernel config is CONFIG_KVM. I have it always on, let it complain in dmesg when its started in EL1
<JensGlathe[m]>
qemu is also on it, though
<JensGlathe[m]>
oh no its not
<JensGlathe[m]>
qemu-system-arm: QEMU emulator version 9.0.2 (Debian 1:9.0.2+ds-4ubuntu5.1)
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
sally has quit [Quit: sally]
martiert_work has quit [Quit: WeeChat 4.4.3]
alfredo has joined #aarch64-laptops
alfredo1 has joined #aarch64-laptops
alfredo has quit [Ping timeout: 480 seconds]
alfredo1 is now known as alfredo
chrisl has joined #aarch64-laptops
sally has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
jhovold has joined #aarch64-laptops
martiert_work has joined #aarch64-laptops
alfredo has quit [Ping timeout: 480 seconds]
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
hogliux has joined #aarch64-laptops
<hogliux>
kuruczgy[m]: I updated my EL2 kernel and now bluetooth also works for me
<hogliux>
kuruczgy[m]: in EL2
<jhovold>
hogliux: bluetooth is controlled over uart so should indeed work fine also at el2
<jhovold>
it the adsp services you lose (e.g. usb-c orientation detection, usb altmode, battery, and audio)
<hogliux>
jhovold: yes, yes - not sure what I did wrong that bluetooth wasn't working
<jhovold>
hogliux: btw, audio sometimes not probing with the in-kernel pd-mapper is indeed still an issue
<hogliux>
jhovold: also my qemu runs just fine. I've been using linux in EL2 almost daily for a month now
<jhovold>
I'm seeing it on both the crd and t14s, which machine did you have?
<hogliux>
jhovold: I've got slim 7x. So for me I had the same issue with the in kernel pd-mapper
<jhovold>
right, yeah, we need to get that sorted
<hogliux>
jhovold: yeah, so I've got it sorted for me: I have an init bash script in my initrd which loads all the necessary kernel modules
<jhovold>
i saw your comment about that, the audio probe failure happens while the qcom audio stack is probe deferring like mad
<hogliux>
jhovold: as long as I load snd_soc_wsa884x in initrd and then let qcom_pd_mapper load later when udev starts properly when ubuntu is booting
<jhovold>
so if you preload a bunch of modules you may change timing of events so that you no longer see the issue
<hogliux>
jhovold: it seems to work for me
<hogliux>
exactly
davidinux is now known as Guest1211
davidinux has joined #aarch64-laptops
<jhovold>
hogliux: how often did you hit it without your workaround?
<jhovold>
running a boot loop here, it could happen anything from every 5 to 70 boots, but imagine those number will differ quite a bit between setups
hogliux has quit [Remote host closed the connection]
hogliux has joined #aarch64-laptops
<hogliux>
jhovold: never
Guest1211 has quit [Ping timeout: 480 seconds]
<jhovold>
i meant *without* your workaround
<hogliux>
jhovold: oh without my workaround - I don't know: seemed like 50/50
<jhovold>
oh, that often, not good
<jhovold>
one in fifty boots is also way too often, but still
<hogliux>
Hmmm don't quote me on that. I've always had my init script in initrd and added pre-loading snd_soc_wsa884x pretty early on. So I can't really reliably say how often I saw it without the workaround
<hogliux>
jhovold: kuruczgy: Hmmmm - qemu installed via apt also locks up my machine. I've been using lxd with kvm just fine though. I wonder what's different in qemu.
<jhovold>
ok, fair enough, it's still happening way to often and we need to fix it
alfredo has joined #aarch64-laptops
<jhovold>
I haven't tried slbounce myself yet so no idea what's up with that
<jhovold>
bumble[m]: regarding the new x13s bios, i got a report that about missing usb hotplug events after a bios update
<jhovold>
not fully confirmed, but this was with the second most recent bios, hopefully the last one would fix any such issues if it was indeed a fw bug
<jhovold>
but perhaps check with someone running the latest before updating
<maz>
kuruczgy[m]: there are a few patches that workaround some creative behaviours.
<hogliux>
maz: qemu patches?
<maz>
kernel patches.
sally has quit [Quit: sally]
<maz>
I don't touch QEMU unless I want to feel pathologically sick. :)
<hogliux>
so these are kernel patches to get KVM working?
<maz>
this is what I wrote.
<hogliux>
do you links to these patches?
<maz>
no. once I'm satisfied that they do the right thing, I will post them. not before next week anyway.
<maz>
you can use KVM without patches anyway. you will just get some odd behaviours.
sally has joined #aarch64-laptops
<kuruczgy[m]>
maz: oh, I was hoping to apply them, I though early feedback would be useful for you too if I find that anything is wrong with them. Anyway, please post a link as soon as they are available somewhere publicly :)
<bumble[m]>
@_oftc_jhovold:matrix.org thanks
lak has joined #aarch64-laptops
nothorseface has joined #aarch64-laptops
<maz>
kuruczgy[m]: early feedback is not really useful for code that is still in development. once I decide the code is in a decent enough state, I will post it to the relevant mailing list, and at that point I will very much welcome your review.
<maz>
and yes, it takes time (after all, the NV code has been worked on for 8 years).
<vickdu31[m]>
Does it mean 0,8% among all new laptop sold in Q324?
chrisl has quit [Ping timeout: 480 seconds]
svarbanov has joined #aarch64-laptops
chrisl has joined #aarch64-laptops
<JensGlathe[m]>
sounds like it
chrisl has quit [Ping timeout: 480 seconds]
<\[m]>
that's a lot imo? or that's bad?
<\[m]>
so tomorrow they'll for sure cut the prices for those by 20% right? 😄
<JensGlathe[m]>
I guess its a lot, but depends on their projections. No idea what qcom actually expected, how they handled the Dev Kit (although this thing is fast) doesnt bode well
nothorseface has quit [Ping timeout: 480 seconds]
<\[m]>
but you can test in the cloud, no worries ahhaha
<vickdu31[m]>
Well if they would do ARM server SoC it might work out better and linux dev would speed up, we have seen some weird server farms of Mac minis on Asahi...
nothorseface has joined #aarch64-laptops
<HdkR>
Any SoC that sits somewhere betwen X1E and GH200 would be a boon.
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
deathmist has joined #aarch64-laptops
ungeskriptet has quit [Remote host closed the connection]
alfredo has quit [Ping timeout: 480 seconds]
chrisl has joined #aarch64-laptops
hogliux has joined #aarch64-laptops
alfredo has joined #aarch64-laptops
<hogliux>
kuruczgy: can't seem to get qemu working on EL2 - it always just immedietly reboots the laptop. I tried latest qemu master, 9.2.0 and the default version on ubuntu 24.04 - none of them seem to work. I wish I had some kind of error log to see why it failed
hogliux has quit [Remote host closed the connection]
<kuruczgy>
hogliux: I would try strace, see if there is maybe a specific KVM ioctl where it fails
<travmurav[m]>
kuruczgy: switch-if-overlay -> it's not merged because I didn't feel it exactly aligns with my current "philosophy" of slbounce (that is, not being linux specific). also iirc there was some other more technical issue with it. I can probably put it under a compile flag though...
hogliux has joined #aarch64-laptops
<hogliux>
travmurav: I really like the solution with grub and efidriver. This way you can easily select if you want to boot into EL2 or not from the grub menu.
<travmurav[m]>
well this was one of the reason I've implemented that thing for grub :P
<travmurav[m]>
still I also use sd-boot primarily so need to think of that too...
<travmurav[m]>
I guess ideally someone just fixes linux in el2 :P
chrisl has quit [Ping timeout: 480 seconds]
<hogliux>
yes!
alfredo has quit [Ping timeout: 480 seconds]
chrisl has joined #aarch64-laptops
<macc24>
HdkR: doesn't M4 (pro/max) already serve that purpose?
chrisl has quit [Ping timeout: 480 seconds]
craftyguy has quit [Remote host closed the connection]
craftyguy has joined #aarch64-laptops
alfredo has joined #aarch64-laptops
ungeskriptet has joined #aarch64-laptops
hogliux has quit [Quit: Page closed]
nothorseface has quit []
nothorseface has joined #aarch64-laptops
chrisl has joined #aarch64-laptops
<maz>
hogliux: make sure you pass 'id_aa64mmfr0.ecv=1' on the command-line. otherwise, you trip some firmware bug.
<maz>
*on the kernel command-line.
chrisl has quit [Ping timeout: 480 seconds]
davidinux has quit [Quit: WeeChat 4.3.1]
nothorseface has quit []
hogliux has joined #aarch64-laptops
<hogliux>
maz: thank you! I can confirm that fixes the issue with qemu
<maz>
hogliux: nothing to do with QEMU, really. everything to do with a broken firmware, and KVM using the advertised features...
<maz>
but glad it works for you.
phire_ has joined #aarch64-laptops
phire is now known as Guest1244
phire_ is now known as phire
Guest1244 has quit [Ping timeout: 480 seconds]
nothorseface has joined #aarch64-laptops
alfredo has quit [Quit: alfredo]
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
<kuruczgy>
maz: Can confirm that this fixes the crashing with qemu. Thank you very much!
<maz>
np
nothorseface has quit []
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
hogliux has quit [Remote host closed the connection]
<robclark>
then capturing 0.8% of the market is really good..
<robclark>
JosDehaes[m]: the interesting figure to see is not % of market but % of market by price segment (since mid cost purwa only just recently started shipping and low cost one hasn't started shipping yet).. idk the breakdown of low/mid/high end laptops or if the figures for total market size include desktops/macs/chromebooks, so as given the figures are a bit useless. If high end windows laptops are 5% of the total market
<Dylanger>
<vickdu31[m]> "Well if they would do ARM server..." <- I think my next "big boy" PC that I use to do builds etc will end up being either a Mac Mini or some sort of aarch64 device hopefully
<robclark>
JosDehaes[m]: so looking at the canalys report we can kinda tease out segments.. they say 20% of market for AI-PCs.. so that makes qc ~4% of the segment. The problem is that 20% figure also includes apple, so it isn't just windows PCs. And it looks like apple is nearly half of what the count as "AI-PCs", so qc is closer to 8% of the segment. Which is actually pretty solid.
Caterpillar has joined #aarch64-laptops
alfredo has joined #aarch64-laptops
<JensGlathe[m]>
and ~1,5M units running?
<robclark>
JensGlathe[m]: IIRC it was ~0.7M .. but for a single quarter (I guess Q3 should be Jul-Sept, so basically just as the first x1e's were trickling out).. I guess double that for the year, especially as they fill out the mid and low end, seems doable.
<JensGlathe[m]>
that's not zero, and tbh not many on the 2nd hand market as of now
<robclark>
yeah, I think qc will do fine in the win laptop market
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
alfredo1 has joined #aarch64-laptops
alfredo has quit [Ping timeout: 480 seconds]
alfredo1 is now known as alfredo
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
alfredo has quit [Quit: alfredo]
deathmist1 has joined #aarch64-laptops
deathmist has quit [Read error: Connection reset by peer]
deathmist1 is now known as deathmist
nothorseface has joined #aarch64-laptops
chrisl has joined #aarch64-laptops
<JosDehaes[m]>
difficult to know what numbers to believe: lies, damn lies and statistics 😁
chrisl has quit [Ping timeout: 480 seconds]
nothorseface has quit []
jhovold has quit [Ping timeout: 480 seconds]
<JensGlathe[m]>
same game as everywhere
<JensGlathe[m]>
offers on ebay are quite a good metric for quality
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
tobhe_ is now known as tobhe
<steev>
What version of qt does fedora have currently dgilmore ?
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
echanude has quit [Ping timeout: 480 seconds]
<HdkR>
macc24: Technically yes, but the bugged PCIe makes it harder to care about :D
<macc24>
are there any non-server arm platforms with completely normal pcie?
<HdkR>
macc24: Jetson Orin pretty much
chrisl has joined #aarch64-laptops
chrisl has quit [Ping timeout: 480 seconds]
Caterpillar has quit [Quit: Konversation terminated!]