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)
shawnguo has joined #aarch64-laptops
hexdump01 has joined #aarch64-laptops
hexdump0815 has quit [Ping timeout: 480 seconds]
iivanov has joined #aarch64-laptops
alpernebbi has joined #aarch64-laptops
spawacz has joined #aarch64-laptops
alpernebbi has quit [Quit: alpernebbi]
booda has joined #aarch64-laptops
booda has quit []
* rfs613
wonders if he should try the debian image on his c630 (currenly using the original ubuntu one)
<gwolf>
rfs613: By all means!
<gwolf>
I got a C630 last week, and am more than happy with the Debian image provided by aarch64-laptops
<rfs613>
i guess the main difference, other the the os itself, is the use of dtbloader
<steev>
si
<rfs613>
gwolf: congrats ;-) mine is like a year or more old... but I've never had much luck with wifi under linux
<steev>
don't let the domain worry you, it's debian, not kali
<rfs613>
steev: cool, what's different in those images?
<steev>
it uses my 5.12 kernel and not the 5.11 from aarch64-laptops, also has the latest versions of all the userlands
booda has joined #aarch64-laptops
<steev>
you still don't get network during install, but that's because the userland isn't running during d-i
<steev>
i should look into how to do that
<gwolf>
rfs613: And I am using steev's hand-crafted kernel
<rfs613>
that's okay I have a wired ethernet dongle
<gwolf>
The only hardware I am missing support for is HDMI-over-USB-C :-(
<steev>
sorry :(
<steev>
bamse will get to it at some point
<gwolf>
some days ago bamse provided some pointers to start peeking in that general direction, but I've not started doing that yet
<gwolf>
steev: don't be sorry, I'm still more than thankful for your help
<steev>
i'm glad you're liking the device. i do too, it's my daily driver
<rfs613>
if I could get wifi reliable i might be able to use mine more... for now it mostly boots into windows :-(
<steev>
wifi is stable here :)
<rfs613>
and by reliable I mean, not having to boot 10x until it happens to come up
<gwolf>
Mine has not had a glitch so far.
<gwolf>
steev: I didn't expect not to require my previous-main laptop so soon!
<gwolf>
I'm still fiddling bits with sway (wayland works *way* better than xorg), but that's peanuts.
<rfs613>
cool. I'll try your image steev, that should at least narrow down if my hardware is funky, or if it's my cross-compile of your kerne + .config
<steev>
there are a massive number of people in the user list here that you can thank for that gwolf
<steev>
i was starting to list them but then i realized it's basically everyone
<gwolf>
:-]
<rfs613>
gwolf: ah cool, i used sway also, until forced back to i3 due to screen-sharing difficulties...
<gwolf>
rfs613: they went away for me, and I *do* need screen-sharing for my daily work
<rfs613>
gwolf: that's good to know, it was about a year back that I switched back to i3
<gwolf>
and that's also the reason why I switched basically all of my family's Gnome sessions to Xorg a while back
<rfs613>
steev: installing debian over top of ubuntu, should just work?
booda has quit []
<gwolf>
rfs613: no, don't go that route
<steev>
you'd wanna format it
<gwolf>
unless you add a "mkfs" somewhere on that "on top"
<steev>
just rsync your homdir somewhere, then rsync it back
<rfs613>
I assumed that I could tell the debian installer to format the disk
<steev>
oh, then yes
<rfs613>
I recently un-installed linux from an x86 laptop, and it took some effort to really get rid of all traces of it (remove from uefi etc)
<steev>
there is one change i would make as well, after install
<bamse>
rfs613: wifi only comes up on your c630 10% of the boots?
<rfs613>
bamse: something like that.
<bamse>
rfs613: using the debian image or something you brew yourself?
<steev>
old ubuntu install
<rfs613>
i started with the ubuntu 18 image, upgraded it to 20.04, plus a bunch of different kernels based on steev's repo + .config
<bamse>
odd, steev have we have any other reports of that?
<steev>
tables has had god awful issues with his wifi too, but he's also seeing them in windows so his may be a hardware issue
<bamse>
the qualcomm wifi is rather shitty, in windows and android as well
<rfs613>
for me windows works fine, although I would say the range of the wifi is not particularly good. Eg. in some parts of house it loses signal, while other devices work fine at the same place. And this is on 2.4G, with 5G its worse.
<bamse>
rfs613: it's comparable in performance to my old macbook pro...but it does reconnect every now and then
<bamse>
rfs613: and sometimes when i walk through the house it stays on an old ap, so i need to manually toggle wifi off/on to trick it to reconnect to a closer mesh node
<rfs613>
i've only got one ap ;-)
<rfs613>
hous isn't that big ;-)
<steev>
mine is fairly solid but tbh, i have just a small 1br so everything is near the wifi :(
<bamse>
regardless, i haven't seen any problems with getting wifi up and running - not since the early days
<bamse>
rfs613: steev knows this stuff better, but i think it would make sense to ditch your old install
<rfs613>
yep, i've just finished dd'ing the new image, and backing up my old home... so we will see shortly.
<steev>
yeah, wifi was kinda... iffy on older kernels
<bamse>
yeah..."i've finally grown my hair back"-iffy
<bamse>
man it was a struggle to get that up and running
<gwolf>
FWIW, my house is big, and I have only one AP. I not had any low signal issues.
<rfs613>
hrm, do i need to use soemthign other than dd to copy the iso to usb stick? I'm using same stick as the original ubuntu image, on left side port, but bios doesn't seem to want to boot from it.
<steev>
what size is the usb stick?
<steev>
also which one are you using? 11 or unstable?
<rfs613>
64gb... same one i used previously to get ubuntu loaded on the c630 in fact
<steev>
hm, if you've updated your firmware, that isn't gonna work
<steev>
the latest uefi firmware won't boot off anything > 32GB
<rfs613>
ah, i guess windows might have updated that, though I thought i had booted from this stick after that point.
<steev>
i used to do a trick to get it to boot, by using a smaller usb drive, with the same uuid's as a 256gb, and then once it loaded the kernel/initrd i'd yank the smaller one, and it'd boot just fine
<steev>
it required keeping everything in sync though and was a pita
<rfs613>
16gb stick works... it shows 3 choices in addition to window boot manager
<rfs613>
i'm guessing i want partition 1
<rfs613>
part1 seems to be debian, part2 is efi
<rfs613>
grub says version 2.04-20kali1
<rfs613>
:-)
<steev>
oh, maybe it did get poisoned by my build system :(
<steev>
should still work tho and install from debian?
<rfs613>
will find out soon, just in a meeting for a bit...
<steev>
the one other change i'd make to a system, something like mkdir -p /etc/systemd/system.conf.d && echo -e "[Manager]\nDefaultTimeoutStopSec=8s" > /etc/systemd/systemd.conf.d/default-timeout.conf
<steev>
this stops it from taking centuries to kill wpa_supplicant
<macc24>
oh right, i should hack ath10k to give up when it loses device to get lazor to poweroff less violently
<macc24>
amstan: i noticed that usb power doesn't go off in suspend... do you think this could be related to lazor not-really suspending?
<rfs613>
in debian installer for hostname, does one enter the fqdn or just hostname?
<gwolf>
both should work
<gwolf>
if you have a fqdn, better.
<rfs613>
looks like it rejects dots, so no fqdn I guess
<gwolf>
that's quite odd to read... I am not that familiar with the latest d-i, but have used d-i for over 15 years... it always accepted periods.
<BlitzWorks>
ben installer prompts for hostname and 'domain name' speratly witht he domain anme being append to the chosen hostname iirc
<robclark>
macc24: iirc we can't power down the usb hub, because things hanging off usb could be generating wakeup events.. or something roughly along those lines
<macc24>
robclark: ,_, i just want that blue light on power button gone
* robclark
hands macc24 a sharpie
* macc24
draws a mustache on robclark
* macc24
hands back the sharpie
<rfs613>
steev: earlier you said "there is one change i would make as well, after install", was this the timeout for killing wpa_supplicant, or something else?
<steev>
yeah that timeout i mentioned yeah
<steev>
also i wish we had better wifi firmware, it doesn't support rawmode :(
<rfs613>
i'll settle for 'it connects most of the time' ;-)
<steev>
well, aside from trying to use rawmode=1 cryptomode=1 (which fails because the firmware doesn't support rawmode), i haven't had issues with wifi in forever
<rfs613>
i'm very curious to see how mine behaves once this install finishes...
<steev>
just be sure to follow the readme to install the dtbloader
<rfs613>
yup, and i'm wondering if my old ubuntu might make an apparance there again... we will see soon
<steev>
i don't believe it'll get rid of the ubuntu stuff in the efi partition,no, you'd have to do that manually
<rfs613>
hrm, got to install grub, but it failed immediately with "An installation step failed"
<rfs613>
maybe because grub is already installed?
<steev>
check the log
<rfs613>
Configuring 'grub-installer' failed with error code 1
<steev>
it should install to /efi/debian not /efi/ubuntu so it shouldn't matter
<rfs613>
oh, and error mounting /target/sys/firmware/efivars
<steev>
i wonder
<rfs613>
indeed /target/sys is empty
<steev>
it may be that the iso is poisoned from kali :(
<rfs613>
this was the debian-11 one by the way.
<rfs613>
I can do bcdedit from windows side (did that on original ubuntu install)
<steev>
check that /efi has a debian directory and it has the grub thingie in it?
<rfs613>
i have /target/boot/efi/EFI/{debian,ubuntu}
<steev>
the efivars bit is fine though
<rfs613>
DtbLoader.efi is present in debian, while grubaa64.efi is in ubuntu
<steev>
the grub in ubuntu is the ubuntu one, not the debian one
<steev>
you can try manually running
<steev>
ah crud, i can't remember the incantation
<rfs613>
there's another error too, in grub-postinst line 16, db_input not found, db_go not found
<steev>
i have no idea what those are
<steev>
check grub-install's options, one of them should be something along the lines of --no-removable
<steev>
it'd be something along the lines of... --target=aarch64-efi --efi-directory=/boot/efi --bootloader-id=Debian --no-removable
<rfs613>
i seem to have grub-installer that seems to be a busybox built-in... and not grub-install
<steev>
you probably want to chroot into the installation in /target
<rfs613>
ah yes
<steev>
chroot /target /bin/bash
<rfs613>
that's better
<steev>
i just can't remember the option(s) needed to succeed
<steev>
oh, it's arm64-efi
<steev>
and it's probably --no-nvram --no-extra-removable ?
<steev>
so, something along the lines of grub-install --target=arm64-efi --efi-directory=/boot/efi --bootloader-id=Debian --no-extra-removable --no-nvram
<steev>
so, something along the lines of grub-install --target=arm64-efi --efi-directory=/boot/efi --bootloader-id=Debian --no-extra-removable --no-nvram /dev/sda
<rfs613>
i've got --force-extra-removable or --removable, but dont' seem to like --no-extra-removable
<steev>
ah, maybe it's just --no-nvram then
<rfs613>
it prints a whole pile of "unknown device /dev/sdaN" with N cycling throuh all 5 partitions on my disk. Then "instalation finished. No error reported"
<steev>
check if Debian shows a grubaarch64.efi now?
<rfs613>
yup, it's there now
<steev>
so, it might work now
<steev>
i'd go back to the installer and see if there are any other steps past installing the bootloader
<steev>
also leave the chroot
<rfs613>
there's 'continue wihtout bootloader' and 'finish the installation', then 'change debconf priority'
* rfs613
picks the first one
<rfs613>
tells me i will have to manually set root=/dev/sda5 on kernel commandline. Ok
<steev>
shouldn't need to, but maybe you will :)(
<rfs613>
ok, in Shell.efi now
* steev
crosses his fingers
<rfs613>
ooh the colours ;-)
<rfs613>
hmm, backspace doesn't work in shell.efi
<rfs613>
hmm, instead of changing the entry used for windows boot manager, can we add a new entry?
<robclark>
backspace actually kinda works.. but the display is out of sync.. if you can type without looking at the screen it works ;-)
<rfs613>
oh, actually, i see that my windows boot manager is actually pointing at ubuntu's grubaa64.efi
<rfs613>
robclark: ah cool thanks
<rfs613>
the DevPath for my windows boot manager starts with a HD(1,GPT,uuid-type-string,0x100,0x10400)/\EFI\ubuntu\grubaa64.efi
<rfs613>
does it need that HD(..) thing?
<rfs613>
wiki page seems to show just using "EFI\debian\DtbLoader.efi"
<steev>
i only went with what's in the wiki
<rfs613>
okay, i tried that, the resulting DevPath after dumping is even longer... now starts with VenHw(...)/HD(...)/\EFI\debian\DtbLoader.efi
<rfs613>
seemingly wiht two different uuids now
<steev>
i didn't do anything but what was listed in the wiki, didn't even check if things looked correct
<rfs613>
ok, it rebooted to grub (different version than old ubuntu)
<steev>
that part is good!
<rfs613>
so i guess i have to manually load kernel/initrd/dtb now to get it to boot
<steev>
shouldn't need to do the dtb, but the kernel and initrd yeah, probably need to run update-grub after you get booted
<rfs613>
it's at grub command prompt, to be clear... not a boot menu
<steev>
okay, so you should be able to do something like...
<steev>
linux /boot/vmlinuz-5.12.19-custom root=/dev/sda5 ro pd_ignore_unused clk_ingore_unused
<steev>
initrd /boot/initrd.img-5.12.19-custom
<steev>
but fix my typo
<steev>
ingore->ignore
<steev>
fwiw, grub does support the devicetree command, just that it's not patched into debian's like ubuntu has it, so you CAN pass devicetree /usr/lib/linux-image-5.12.19-custom/qcom/sdm850-lenovo-yoga-c630.dtb if you wanna be doubly sure
booda has quit []
<steev>
you're killing me here, rfs613, did it work?
<rfs613>
i was figuring out the grub stuff again and hadn't noticed you typed it ;-)
<steev>
heh
<rfs613>
i needed (hd5,gpt5)/boot/... fwiw
<steev>
i just grabbed what is in my grub.cfg minus the uuid
<steev>
ah, yeah, you shouldn't need it once you update-grub
<steev>
but once it's booted should be golden
<rfs613>
got kernel messages...
<rfs613>
got X cursor
<steev>
also, make sure you have bitlocker off if you have it on so you can copy the firmware across
<rfs613>
yep, its off... did the firmware copy last time for ubuntu, but will do a fresh version now
<rfs613>
desktop boot... now for update-grub
<steev>
i'll look into the grub thing too... i may have stupidly done an apt source not thinking about the fact that i was building it on a kali box
<steev>
but, fwiw, we only patch in that we're kali, nothing extra on top
<rfs613>
happen to have the grub chainloader syntax for windows handy?
<steev>
it's in the readme
<rfs613>
ah, it is, hadn't gotten that far ;-)
<steev>
supposedly there's a newer way to do it as well, instead of chainloader, you just do "ntldr"
<steev>
but i haven't tested that yet
<rfs613>
hehe, the command about install grub being slow sure is true!
<rfs613>
apparently it's there for a reason, there's even a comment
<steev>
i used to also have a shell.efi entry for... reasons
<steev>
but i lost that somewhere along the way
<rfs613>
ok, grub now has a menu! and i made a typo in the windows entry ;-)
<steev>
i did too, which is pretty funny since i sent the correct thing as a pull request
<rfs613>
did you forget the ")" after "}" ?
<rfs613>
ok, firmware is copied over, rebooting...
<rfs613>
have sound, keyboard okay, but touchpad doesn't seem to work now
<steev>
possibly
<rfs613>
(worked before copying over firmware)
<steev>
hm
<steev>
no, sometimes i2c doesn't work right
<steev>
can you check if it gave some i2c reset error in dmesg?
<steev>
the touchpad at least on these, afaik, doesn't use a firmware file
<rfs613>
i2c_hid_of: probe of 3-002c failed with error -61
<steev>
yeah
<steev>
just reboot, it should come back up
<rfs613>
ok, and just noticed touchscreen works
<steev>
si
<rfs613>
so i can see "wifi not connected", let's see...
<steev>
for sound, you're gonna wanna install alsa-ucm-conf, i don't think it is installed by the installer
<rfs613>
i'm hearing sound, in fact
<steev>
oh good
<rfs613>
so network connected to my 2.4G wifi on first try
<rfs613>
getting "failed to reset device" for the same i2c_hid on next boot
<steev>
reboot again, it should do it
<rfs613>
but wifi came up again
<steev>
seems the wifi not being stable moved to the i2c :P
<rfs613>
indeed
<rfs613>
i could probably almost live with that ;-)
<steev>
it would be nice to know what windows does
<rfs613>
3rd boot, have mouse and wifi
<rfs613>
i mean touchpad
<rfs613>
well, this is already better
<steev>
\o/
<steev>
i'll look into the grub thing
<rfs613>
is there a specific reason for setting apt to 'sid' rather than staying on eg. bullseye?
<steev>
uhhhh
<steev>
that's probably just me overlooking it
<steev>
it was based on sid, i tried to point it at bullseye since it was finally released
<steev>
and bullseye->kali is easier than sid->kali just because sid has some newer stuff
<rfs613>
volume control seems to be binary: off or loudish
* gwolf
peeks here now and then...
<gwolf>
FWIW, I'd still have left it to sid, as it's easier for me to do Debian work on a Sid machine -- but yes, I also believe it'd be much more natural if the installer delivered a Bullseye system
<steev>
yeah that's been mentioned rfs, sadly, i haven't seen srini in a bit but i believe bamse mentioned it to him
<steev>
gwolf: well that's why there's also the unstable iso :P
<rfs613>
steev: no big deal for me... keyboard, screen, wifi... and we're good to go
<steev>
:D
<steev>
i'm glad to hear you're back up and running already
<rfs613>
i see that suspend after 20 minutes is enable... last I heard though there was no power managment. Is this still correct?
<steev>
wish i knew what was going on with tables' machine but it really does sound like a hardware issue, except he said that it works, until it tries to update to the latest firmware in windows update
<steev>
that is not still correct
<steev>
HOWEVER
<steev>
remember that thing about the touchpad?
<steev>
it sometimes likes to go ahead and check out when coming back from suspend
<rfs613>
sounds like on any power-up event then
<steev>
no
<steev>
it's not every time
<steev>
and can't tell what is going on there really
<rfs613>
yeah, so something not sequenced, or a hardware issue that needs papering over
<steev>
feel free to debug it if/when you run into it. bamse gave me a command "libinput debug-gui" or something like that
<steev>
but i haven't actually suspended mine in a while because i've been having it busy doing stuff
<rfs613>
ok, will see, for now it's just fun that I can actually use this (and change caps-lock to control finally!)
<steev>
ayyy two of us
<steev>
bamse: did you go through and clear up a bunch of patches today? i swore i saw some about irq cleanups and... now i don't see them in patchwerk
hexdump01 has left #aarch64-laptops [#aarch64-laptops]
<bamse>
steev: i did rebase my branch late friday and posted the pcie controller patches today...not sure if there was anything related to irqs though
<steev>
ah no, i meant in patchwork itself, earlier i saw a bunch of stuff posted today, but now all i'm seeing is shawn's
<steev>
it's not a big deal, as just curiours
<bamse>
steev: i did mark a bunch of "not applicable" patches as such a couple hours back
<steev>
ahh
<steev>
that probably did it
booda has joined #aarch64-laptops
<rfs613>
time make LOCALVERSION="-custom" -j8 bindeb-pkg ==> real 51m53.561s
<rfs613>
about 10x slower than my x86, but not terrible i guess
FizzBuzz has joined #aarch64-laptops
<rfs613>
... and it booted w/ touchpad and wifi working. Cool!