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
<konradybcio>
robclark: the venus driver has some special handling for your chromebooks..
<robclark>
ahh, ok.. so presumably someone could copy that for iris for el2 boot?
newyear25 has quit [Read error: Connection reset by peer]
hawer has joined #aarch64-laptops
sally has quit [Quit: sally]
sally has joined #aarch64-laptops
nothorseface has joined #aarch64-laptops
nothorseface has quit []
deathmist has quit [Ping timeout: 480 seconds]
deathmist has joined #aarch64-laptops
nothorseface has joined #aarch64-laptops
tobhe has joined #aarch64-laptops
tobhe_ has quit [Ping timeout: 480 seconds]
<travmurav[m]>
robclark anthony25 windows nowdays always wants to run under hyperv (for stuff like "hyp based security") so it always boots under hyper-v which resides in el2
<travmurav[m]>
Qcom fw nor booting uefi in el2 feels like some stupid pile of ugly hacks to me, like "well we have this uefi that boots in el1 anyway, it's too hard to make it bbr compliant, let's put el2 swith into dtrm and call it a security feature" -_-
<travmurav[m]>
But with the actual question, yes when qhee/gunyah is gone we need to handle fw differently and, also, handle some resources in linux manually
<travmurav[m]>
Like iommu, some clocks and pds
<travmurav[m]>
Afaik 7c1 Venus is the only example in linux that works in el2 due to the special cros codepath
<travmurav[m]>
(Which, obviously, works for woa in el2 too, which I've confirmed)
<travmurav[m]>
There is some stuff for hexagons but I've mentioned before I couldn't get it to boot on 7c1 el2
hexdump01 has joined #aarch64-laptops
<JensGlathe[m]>
Woke up to HTC Rx errors on the X13s - msi-map was disabled.
<steev>
still none here, at almost 10 hours of uptime
alfredo has joined #aarch64-laptops
nothorseface has quit []
nothorseface has joined #aarch64-laptops
alfredo has quit [Ping timeout: 480 seconds]
nothorseface has quit [Remote host closed the connection]
nothorseface has joined #aarch64-laptops
<JensGlathe[m]>
now running with interrupt-names disabled, too
<macc24>
anthony25: there used to be pwrseq stuff that blocked merging of it, and similar patches for other stuff. not sure if it's still around or not
xroumegue has quit [Ping timeout: 480 seconds]
ungeskriptet has joined #aarch64-laptops
<craftyguy>
JensGlathe[m], steev : I have not seen any of those HTC Rx errors in over 1 day since disabling msi-map for pcie4
<craftyguy>
previously it was like, at least every hour or so (and I often had to reload the driver to recover)
nothorseface has quit [Ping timeout: 480 seconds]
nothorseface has joined #aarch64-laptops
nothorseface has quit []
<JensGlathe[m]>
It is now one shared interrupt line, apparently - for the 20 handlers it registered
jglathe_angrybox has quit [Remote host closed the connection]
<JensGlathe[m]>
and again HTC Rx errors, on the X13s
<JensGlathe[m]>
so the mitigation with msi-map fails here
<jhovold>
JensGlathe[m]: we know it's not a fix, but is certainly sounds like it reduces the likelyhood of hitting this at least in setups like craftyguy's
<jhovold>
does it seem to happen as often as before (e.g. a couple of times a week) for you?
<JensGlathe[m]>
Just my observation. X13s had 2 occurrences within 12 hours, the wdk2023s zero
<jhovold>
and with gic its enabled, you see this how often? roughly as frequently?
<jhovold>
craftyguy: did you try disabling the modem altogether? we still don't know if your hang is related to it, but may be worth a try
<jhovold>
craftyguy: just to rule out other changes, do the ath11k errors come back immediately if you add back the msi-map property?
<JensGlathe[m]>
I tried to debug/trace it, armed the Rx interrupt handler with a global mutex (assumed that it might be some re-entrance thing) all nope. If you get HTC Rx size errors, the cb gets called with garbled data.
<JensGlathe[m]>
Fun fact: You can get HTC Rx errors even with no connection established, IF may be up, but no link. Its definitely less often than when the connection is up.
<JensGlathe[m]>
with its-gic enabled I see it more often. The x13s was off for a few days, I was surprised to get that many HTC rx occurrences.
<jhovold>
steev: thanks for checking the modem model, another datapoint if there's some router feature that make some people hit this more easily
<jhovold>
as I mentioned, I've only seen this once in a year, while craftyguy cannot go five minutes without hitting this currently (his setup may come in handy for debugging and testing fixes)
hogliux_ has joined #aarch64-laptops
<_mike>
sometihng about this kernel i guess is not stable, when i try to parallel build it crashes
<anthony25>
oh sorry, didn't see you answer macc24
<hogliux_>
Yes, I'll try to submit a patch for this. But just like you recently, it would be my first one. So need to read up on it.
<hogliux_>
I wasn't sure if macc24's comment was related to my bluetooth patch? Honestly, all I did was take the same changes that were used on the think pad and applied it to the slim 7x
<anthony25>
after that, I just ended up setting up a basic nmutt config to answer in plain text, tweaking thunderbird for that was too much pain
<hogliux_>
OK I'll read through that. I'll try to submit it later today.
<anthony25>
that would be so nice, thanks a lot!
<hogliux_>
Thanks for getting the lid switch working!
<anthony25>
you're welcome, and thanks again for the bluetooth, I use it quite often actually (more than the lid switch :p)
<hogliux_>
Now I only have three things missing from my slim 7x want list: 1) webcam 2) USB4 PCIe pass-through 3) get adsp working in EL2
<anthony25>
I'm a simple man, 1) < 1W in suspend and I'll be happy :D
<hogliux_>
just use hibernation
<hogliux_>
:-)
<hogliux_>
My old Intel laptop had so bad suspend power draw during suspend that I just got used to hibernating anyway
<anthony25>
hibernation doesn't work
<anthony25>
I used to do the same on my previous laptop, suspend-then-hibernate
<hogliux_>
Hmmm hibernation works for me on my slim 7x. But it was fiddly to get working. I need to rmmod the adsp module for example and then reload some of them in the initrd in the right order.
<kalebris_>
hogliux_: do you have a list of modules on what needs to be disabled/re-enabled and the order of reenabling them?
kalebris_ is now known as kalebris
<anthony25>
oh interesting!
hogliux_ has quit [Remote host closed the connection]
hogliux_ has joined #aarch64-laptops
hogliux has joined #aarch64-laptops
hogliux_ has quit [Read error: Connection reset by peer]
<hogliux>
anthony25: kalebris: you need to `rmmod qcom_q6v5_pas`. This is how my initrd init script looks like: https://controlc.com/83bdaa9a
<hogliux>
Of course, you'll also need to have the relevant firmware in the initrd as well
<hogliux>
And then I hibernate with `pm-hibernate`
<anthony25>
I'll try first with the rmmod trick
<anthony25>
right now if I try the hibernate, the laptop just freezes
<hogliux>
One downside though is that hibernation takes really long, much longer than on my 7 year old laptop
<anthony25>
to hibernate or resume?
<hogliux>
You also need to be sure that efivars is working and that your EC works. When I type in `pm-hibernate` I see my laptop LED pulsing like it's gone into suspend. It then takes a good 2-3 minutes to then turn off.
<hogliux>
both
<anthony25>
wow ok
<hogliux>
If you are using a lot of RAM it can take up to 6 minutes or so to hibernate. So be patient. If you hit any key during hibernation process, the laptop wakes up but the screen just turns blue. But even then, eventually it will turn off
<hogliux>
But I use hibernation constantly and don't have any issues with it. Rarely the Wifi crashes after resuming and I need to rmmod/modprobe the wifi ath driver.
<hogliux>
But that's the only issue I have (apart from it taking pretty long)
<anthony25>
hmm, could it be that writes on nvme are heavily throttled after you rmmod qcom_q6v5_pas?
<hogliux>
No I don't think so. Because during testing I always hibernated without removing the module. The computer will still resume then, but it will then crash after 2-3 minutes or so because it constantly tries to reboot the adsp.
<anthony25>
ok
<hogliux>
Maybe it's normal. I have the 32GB variant of the slim 7x. Maybe it just takes that long to write it out to disk?!?
<anthony25>
it's pretty weird for me that the time to hibernate seems that long and linear depending on how much ram you're using
alfredo has quit [Ping timeout: 480 seconds]
<anthony25>
hogliux: on an HDD maybe :D
<hogliux>
;-)
hogliux has quit [Quit: Leaving]
akaWolf has joined #aarch64-laptops
kalebris has quit [Quit: WeeChat 4.4.4]
kalebris has joined #aarch64-laptops
svarbanov__ has joined #aarch64-laptops
svarbanov_ has quit [Read error: No route to host]
svarbanov_ has joined #aarch64-laptops
svarbanov__ has quit [Ping timeout: 480 seconds]
nothorseface has joined #aarch64-laptops
alfredo has joined #aarch64-laptops
alfredo1 has joined #aarch64-laptops
alfredo has quit [Ping timeout: 480 seconds]
alfredo1 is now known as alfredo
alfredo1 has joined #aarch64-laptops
hogliux has joined #aarch64-laptops
alfredo1 is now known as alfredo
alfredo has quit [Ping timeout: 480 seconds]
cyrinux has quit []
cyrinux has joined #aarch64-laptops
<JensGlathe[m]>
My X13s seems to be the anti-example: msi-map is on, no HTC Rx errors yet
<JensGlathe[m]>
interrupts are distribute over several cores, but fixed assignment as it seems
hogliux has quit [Quit: Leaving]
hogliux has joined #aarch64-laptops
nothorseface has quit [Remote host closed the connection]
nothorseface has joined #aarch64-laptops
chrisl has joined #aarch64-laptops
hogliux has quit [Quit: Leaving]
flokli has quit [Ping timeout: 480 seconds]
<steev>
jhovold: unfortunately, i can't tell a lot, at&t doesn't give much access to the router, but i do know that it's both 5GHz and 2.4GHz (not 6GHz)
<steev>
JensGlathe[m]: until i switched to at&t i never saw the error, or would see it like johan, once a year
flokli has joined #aarch64-laptops
<JensGlathe[m]>
I had it quite regularly before, every 3 days or so
<steev>
that's just using the debian 6.12.8 kernel
<steev>
i'd be curious if ema ever sees it
<craftyguy>
jhovold: how should I "disable the modem altogether"?
<craftyguy>
It's currently detected by modem manager but I'm not using it for any connections (mm thinks it's "off") but it shows up with "mmcli -L"
<craftyguy>
I'll add back the msi-map prop and see if the errors come back, will report back on it later
<jhovold>
craftyguy: you can change the 'status' property of pcie3a (and pcie3a_phy) to "disabled" in the x13s dts
<steev>
fwiw, 21 hours of uptime and no messages here (but about to reboot for new systemd stuff)
<craftyguy>
IIRC there's a quick way to make changes to the dts and conpile it to dtb without building the whole kernel, where can I learn more about that? 😁
<JensGlathe[m]>
re pmic_glink: These errors "can't establish connectionto <device>" are sort of constant now. There were only a few versions where this didn't happpen. Any idea why that is?
<craftyguy>
would this do anything for these errors on the x13s? "pmic-glink: Failed to create device link (0x180) with supplier usb1-sbu-mux"
<JensGlathe[m]>
no, I guess.
<JensGlathe[m]>
But you sometimes see the splat <i2c device>: incomplete transfer
<steev>
no, those are devlink related
<steev>
the failed to create device link lines i mean, not the i2c splat :)
<craftyguy>
jhovold: I disabled the modem as you suggested, entered s2idle (afaik) and it won't resume again. There's no visible difference between when I suspend and after tapping the enter key or pwr button to wake it. Display remains blank/off, pwr led stays lit, I can enable/disable the keyboard backlight (so something is still alive?), and magic sysrq does nothing (it's enabled/working normally)
<craftyguy>
opened/closed the lid a few times too thinking that might wake it, but nope
<travmurav[m]>
kb backlight is probably handled by ec without linux interaction so doesn't indicate much I'd guess
<craftyguy>
yeah, I am just trying to get it to do _something_ heh. But it seems unable to wake or stuck somewhere in the resume process?
<travmurav[m]>
I'd guess either hung or somehow not receiving interrupts to wake somehow, yeah, but hard to tell
<travmurav[m]>
well the hole on the bottom is there for a reason :^)
<craftyguy>
Does it do anything besides reset(?) the ec?
<travmurav[m]>
it's supposed to cut the battery out fully
<travmurav[m]>
i.e. pressing the button == disconnecting the battery
<craftyguy>
Push/release or push/hold?
<travmurav[m]>
while you hold the button, the battery is not connected
<travmurav[m]>
it's (supposed to be) a dumb inverted battery switch
<craftyguy>
Gotcha. Much faster than holding the pwr button
<travmurav[m]>
(should also unplug the charger if you have it on I guess)
<craftyguy>
Right
<travmurav[m]>
it's actually a very clever idea since it's fully hardware, I like those :D
<craftyguy>
I have no more ideas for debugging this resume issue 😒
nothorseface has quit [Read error: Connection reset by peer]
nothorseface has joined #aarch64-laptops
nothorseface has quit [Read error: Connection reset by peer]
nothorseface has joined #aarch64-laptops
nothorseface has quit [Read error: Connection reset by peer]
nothorseface has joined #aarch64-laptops
alfredo1 has joined #aarch64-laptops
nothorseface has quit [Read error: Connection reset by peer]
nothorseface has joined #aarch64-laptops
alfredo1 has quit []
nothorseface has quit [Read error: Connection reset by peer]
nothorseface has joined #aarch64-laptops
SpieringsAE has joined #aarch64-laptops
<bumble[m]>
just lurking here, but have always had intermittent problems "resuming" x13s with postmarketOS. Sometimes a few weeks go by without any problem and other times it fails to resume every time the lid is closed
SpieringsAE has quit [Remote host closed the connection]
nothorseface has quit [Read error: Connection reset by peer]
nothorseface has joined #aarch64-laptops
srinik has quit [Ping timeout: 480 seconds]
<ema>
steev: hi! I haven't been using the x13s much lately, but I'll take a look next time
<craftyguy>
jhovold: ok after re-adding the msi-map thing... 500 HTC Rx errors in 2 hours of uptime
nothorseface has quit [Ping timeout: 480 seconds]
deathmist1 has joined #aarch64-laptops
deathmist has quit [Ping timeout: 480 seconds]
nothorseface has joined #aarch64-laptops
nothorseface has quit [Read error: Connection reset by peer]
nothorseface has joined #aarch64-laptops
<steev>
ema: thanks :) also, if you have spare time, can you glance over https://salsa.debian.org/kernel-team/linux/-/merge_requests/1313 bryanodonoghue says we should enable UDMABUF as well, and I'm not sure if I should make that arm64 specific, and that's why i have it set as draft still
<steev>
there's also one there for the elan driver which also makes the touchscreen work
nothorseface has quit []
nothorseface has joined #aarch64-laptops
deathmist has joined #aarch64-laptops
deathmist1 has quit [Ping timeout: 480 seconds]
<JensGlathe[m]>
zero HTC rx errors with msi-map, hmm
<bumble[m]>
just now x13s screen will not turn on at all... turning off and on again, screen never never comes on :( though probably un-related to the larger x13s discussion maybe the hardware here is failing
alfredo has joined #aarch64-laptops
<steev>
that would really suck if its the case - not even to the lenovo logo?
<bumble[m]>
its back now
<bumble[m]>
it came back on
chrisl has quit [Ping timeout: 480 seconds]
chrisl has joined #aarch64-laptops
alfredo has quit [Quit: alfredo]
nothorseface has quit [Remote host closed the connection]
nothorseface has joined #aarch64-laptops
first_username has joined #aarch64-laptops
nothorseface has quit []
first_username has left #aarch64-laptops [#aarch64-laptops]
kirill_ has joined #aarch64-laptops
kirill_ is now known as Guest5552
Guest5552 has quit []
nothorseface has joined #aarch64-laptops
nothorseface has quit [Remote host closed the connection]