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)
<steev>
bamse: so far in 5.19 testing, the only thing i've noticed (aside from the spam of bam dma driver), is that i2c seems.... moody
jenneron has quit [Read error: Connection reset by peer]
jenneron has joined #aarch64-laptops
alpernebbi has quit [Ping timeout: 480 seconds]
alpernebbi has joined #aarch64-laptops
<bamse>
steev: isn't that the device which you might not have?
<steev>
bamse: maybe? but when that error happens i 100% do not have touchpad
<bamse>
sounds unlikely to be the case then
<steev>
i'm going back to torvalds master with just the patch to remove the print of the bam dma underflow, and the defconfig now to verify things work
<bamse>
steev: and this is on the flex or the yoga?
<steev>
c630
<bamse>
okay, i just booted v5.19-rc1 on mine a couple of times and i2c seems fine
<steev>
i have not attempted 5.19 on the flex yet, i was letting you and qzed sort things out there first lol
<steev>
the fresh compile is just finishing up now
<steev>
the annoying thing is, based on the output, i *should* have sound, but i'm just showing dummy output
<bamse>
what does /sys/kernel/debug/gpio say about gpio85 and gpio86?
<bamse>
hmm, no that's the touchscreen
<bamse>
gpio40 and gpio41 then
<bamse>
and gpio37
<steev>
give me a bit, was attempting to poweroff and ran into some devfreq thing, and eventually got a ufshcd_wl_runtime_suspend failed (why is it suspend when trying to power off?)
<bamse>
so you're saying that we need to do some more polishing before we ship this thing? ;)
<steev>
maybe just a lil bit
<steev>
that one might be systemd's fault, something about failed unmounting /run/credentials/systemd-sysusers.service
<steev>
mine might be a little busted too though? no keyboard in grub
<steev>
alright looking like this boot isn't gonna have touchpad either, so lemme see what the kernel says about those files
<bamse>
that happens from time to time...
<bamse>
on both 845 and sc8180x...sc8280xp seems more polished in that regard
<bamse>
the grub keyboard thing i mean
<steev>
hm, didn't get wifi this time either tho
<steev>
blargh, can't get wifi back without krzk's patch
<steev>
i wish i knew how to poke around the remote procs more
<steev>
https://paste.debian.net/1243275/ that's gpio with not working touchpad, 40 says in low func0 2mA pull down, 41 is in high func1 2ma pull up, and 37 is in high func0 2mA pull up
<steev>
oddly... the audio seems... powered down?
<steev>
am i reading that file correctly?
<bamse>
gpio40 and gpio41 should be the i2c lines...in which case "in" is bad...
<bamse>
well, "in" and "func0"
<bamse>
but i got the same here...so perhaps i'm looking at the wrong gpios
<bamse>
ahh gpio41 and gpio42...and htose are func1
hexdump0815 has joined #aarch64-laptops
<bamse>
that's better then :)
<steev>
oh good, well, not entirely better but at least they look correct, even though it's not working
<bamse>
and you don't happen to have the device at 0x2c on that same bus?
<steev>
the battery?
hexdump01 has quit [Ping timeout: 480 seconds]
<bamse>
no, the alternative source touchpad
<steev>
it might be but i'm not applying the battery patch
<steev>
oh, uh, one moment
<bamse>
looking at the dts, we have 2 different touchpads, one on 0x15 and one on 0x2c...and we just kept both enabled, as the one you don't have will time out quickly
<bamse>
except that it seems "quickly" has changed?
<bamse>
i'm more curioius to know why i get at least 1452 of those warnings...
<bamse>
ahh "slimbam"...so it's audio trying to do something awesome
<bamse>
how do i get grub to load the grub.cfg from my new partition? grub-install claims to have completed successfully, but loads the grub.cfg from the old partition
<steev>
that, i'm not sure of; i'd probably drop to the grub shel and set prefix=(hdX,gptX)/boot/grub
<steev>
i'm not sure if that works from the grub shell though, or only from the grub rescue shell
<steev>
but conceivably, it'd be set prefix=blah/boot/grub; insmod normal; normal
<steev>
that's what i have to do on my flex5g because a windows 11 update a month ago (or so) fucked something up and now grub simply refuses to load properly and always drops me into the rescue shell
<bamse>
right, that works fine...but i want it to be permanent
<steev>
and unlike the c630, there's no way to get restore media
<bamse>
wonderful
<steev>
oh
<steev>
you might need to do it via efi menu?
<steev>
in the efi shell
<steev>
now that i think about it... maybe i can do that on the flex5g too
<steev>
i'd have to dig out my debian installer cd
<bamse>
hmm, launching the efi shell and launching grubaa64.efi took me to the right grub config
<steev>
bamse: one other thing i've noticed - if i remotely reboot the c630, with the display having been off/blanked, when it reboots, it forgets what the previous brightness was at and sets it to the default
<bamse>
hmm, interesting
<steev>
oh
<steev>
i rebooted into 5.18.1
<steev>
i'm definitely on 3-0015
<steev>
steev@limitless:~$ ls /sys/bus/i2c/devices/3-00*
<steev>
/sys/bus/i2c/devices/3-0015:
<steev>
driver name power supplier:platform:3400000.pinctrl uevent
<bamse>
perhaps they are waiting for someone to add freedreno support?
<steev>
maybe you can only get it from AT&T currently?
<steev>
i wouldn't know, i absolutely refuse to deal with at&t anymore
<bamse>
hehe, t-mobile did such a bad job that i'm happy with at&t
<steev>
sprint was just bought by tmo :(
<steev>
they're hounding me because the.... galaxybook2 isn't gonna work on their network, i think it is
<bamse>
i had to buy a new phone...because at&t shut down their 3g network
<steev>
welcome to modern times sir
<bamse>
i could do data, but as soon as it tried to make a phone call i was kicked out of the network
<bamse>
you would think that my snapdragon 855-device would be new enough
<bamse>
but apparently i had some european "international" version
<steev>
wow
<bamse>
so it would only do voice over 3g here
<steev>
that doesn't sound too bad, but i don't do many voice calls
<steev>
i should look into what they're trying to tell me at some point, since i think the changeover or whatever happens on the 10th
<steev>
oh, 6/30
<bamse>
worse is that they kicked out my car...and that's just not possible to debug :/
<steev>
and yeah, it's the galaxybook2
<steev>
oh, yeah, that's way worse
<bamse>
ahh, the c630 still works on at&t's new network
<steev>
hm
<steev>
they do offer the galaxy book go 5g as a possible replacement for the galaxybook 2
<bamse>
and all you need to do is give theem money, what a deal!
<steev>
yeah.... "free" upgrade
<steev>
just... pay more
<steev>
otoh, it's a 8cxg2
<bamse>
but for a data-only lte device you should be fine
<steev>
i... don't really need it tbh
<steev>
i was only using the galaxybook2 because that is what microsoft recommended to us for wsl2 testing on arm64
<steev>
why the hell they supported wsl2 and the c630 doesn't, i dunno
<bamse>
sweet, ran geekbench on c630, with v5.19-rc1 + some-battery...still working nicely
<steev>
nice
<steev>
okay so basically the issue seems to be that for some reason, 3-002c responds with *something* when it should instead simply nack?
<bamse>
8% faster single core performance than windows and 9% multicore
<bamse>
steev: so if you disable 3-002c 3-0015 starts working?
<steev>
i mean, it just randomly works, i haven't tried checking if removing the 2c device works full time
<steev>
i guess i can try that tomorrow since it's midnight now
<bamse>
perhaps we're screwing up the state of the i2c controller when the 2c fails
<steev>
well, it's not new
<steev>
i just checked the log with 5.18.1 and it has the same HID message
<bamse>
when did it start breaking?
<steev>
with 5.19
<bamse>
yeah, but it's still working, right?
<bamse>
i mean, 3-002c complains...but 3-0015 worked?
<steev>
prior to the rc1 tests, it was rock solid
<steev>
correct
<steev>
i remember way back in the day there were i2c issues where the touchpad would randomly not work, but hadn't seen that in forever
<steev>
oh wait, no, i'm misremembering, that was back when we were doing something with the rising edge
<bamse>
hmm, computer says that we haven't changed the geni-i2c driver since v5.18
<bamse>
anyway...it's tomorrow now...time to get some sleep
<bamse>
i bet this problem is still with us when we wake up :(
<steev>
yeah, i'll test 5.17 and see if the message was there - i see it in 5.18's log
<steev>
sometime tomorrow
<bamse>
in v5.17..v5.18 vkoul changed the interrupt handler...
<bamse>
hmm, no not the interrupt handler side, only the xfer side
<steev>
stop, go to bed
<bamse>
yes! good night
<steev>
night
jhovold has joined #aarch64-laptops
iivanov has joined #aarch64-laptops
matthias_bgg has joined #aarch64-laptops
matthias_bgg_ has joined #aarch64-laptops
matthias_bgg has quit [Ping timeout: 480 seconds]
<robclark>
steev: hmm, neat splat.. probably related to some of my GEM work that went in this merge window.. but not entirely sure how can doing anything to trigger it (ie. new codepaths, other than a few new GEM_WARN_ON()s are things you shouldn't hit unless running freedreno in a vm ;-))
<steev>
robclark: afaict, what triggered it was display off in gnome
<robclark>
hmm
iivanov has quit []
<steev>
I’m not exactly sure since I was afk when it happened, I just happened to notice the kernel panic on the remote connection
jenneron_ has joined #aarch64-laptops
jenneron has quit [Ping timeout: 480 seconds]
matthias_bgg_ has quit [Ping timeout: 480 seconds]
flowriser has joined #aarch64-laptops
flowriser has quit [Ping timeout: 480 seconds]
flowriser has joined #aarch64-laptops
jenneron__ has joined #aarch64-laptops
Erisa4 has joined #aarch64-laptops
djakov_ has joined #aarch64-laptops
jenneron_ has quit [reticulum.oftc.net helix.oftc.net]
travmurav[m] has quit [reticulum.oftc.net helix.oftc.net]
djakov has quit [reticulum.oftc.net helix.oftc.net]
ajhalaney[m] has quit [reticulum.oftc.net helix.oftc.net]
qzed has quit [reticulum.oftc.net helix.oftc.net]
Erisa has quit [reticulum.oftc.net helix.oftc.net]
travmurav[m] has joined #aarch64-laptops
qzed has joined #aarch64-laptops
ajhalaney[m] has joined #aarch64-laptops
jhovold has quit [Ping timeout: 480 seconds]
<steev>
bamse: yep, with 2c removed from the dtb, i reliably get touchpad, have tried about 15-20 reboots now without issue
<bamse>
steev: okay, that certainly seems to indicate that 0015 inherits the error state from 002c (in the i2c-geni driver)
<steev>
okay so i need to poke around in there
<bamse>
yeah...
<bamse>
as i said yesterday, vkoul has spent some time in this code not too long ago, so might be worth having a chat with him as well
<bamse>
but i think the key thing is to understand why the error on 2c affects the next read on 15...
<bamse>
i.e. is it just some state variables, do we not clear the error interrupt bits, or do we return a failure and then receive the error interrupt as we're attempting the next xfer
<qzed>
bamse: any breakage I should expect when rebasing onto 5.19?
<bamse>
qzed: no, it looks pretty good
<qzed>
neat, gonna do that now then
<bamse>
qzed: and i intend to get the sc8180x, primus and flex dts on the list this week...just trying to wrap up some sc8280xp work first
<bamse>
qzed: hoping to see that enable you to post some prox patches as well
<qzed>
awesome! yeah, I think I can add some basic stuff
<robclark>
only breakage I know of in -rc1 is lazor specific
<bamse>
i need to double check, but i think the only pieces left in my tree are dp-related
<qzed>
altough I might need to clean up some things...