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)
<derzahl> on the c630, is hid_multitouch the driver for the keyboard too? or just the touchpad/touchscreen?
<derzahl> how would you guys go about troubleshooting a 3 minute delay for bringing up my wifi after every boot? im guessing it has something to do with the qrtr/rmtfs services, but I dont see anything telling in the logs between boot and when the wifi device finally becomes available
<steev> firmware loading timeout most likely
<derzahl> what would be the cause?
<derzahl> how to fix?
<derzahl> would it be related to those custom services? Cant really find any documentation on them
<derzahl> i looked through the source and found that rmtfs supports -v, but still dont see any logging
<steev> i can't recall exactly, but you can lower the firmware timeout
<steev> bamse knows it best, and i have a terrible memory on the best of days, even worse when i'm on pain kilers
<bamse> derzahl: looking at my sysfs, the keyboard seems to be hid-generic and the touch etc is hid-multitouch
<bamse> derzahl: the reason for the 3 minute delay is as steev say because the ath10k driver probes for a number of firmware files and each time there's a 60 second timeout for userspace to help out as the kernel doesn't find the file
<bamse> derzahl: i typically hack around it by echo 1 > /sys/class/firmware/timeout
<bamse> but this is certainly something that we need to fix, because it's just utterly broken
<bamse> derzahl: typically you have you systemd start rmtfs, which will cause the modem to come up within seconds, then pd-mapper and tqftpserv will set up the wifi firmware in a few additional seconds and then ath10k will request_firmware() and that just sits there for minutes before returning -ENOENT back to the driver
<bamse> derzahl: and interesting things happens if you attempt to suspend your machine before those minutes has passed
<derzahl> bamse: hm, so how is the debian image configured so that wifi comes up immediatly? just has the firmware timeout set to 1?
<derzahl> sweet. that works.
<derzahl> thanks
<derzahl> i added it via tmpfiles.d
iivanov has joined #aarch64-laptops
<derzahl> so i disabled grub from trying to switch to the display driver and im now able to see the password prompt for my luks encrypted root, but I still cant type the password
<derzahl> its like the keyboard driver isnt loaded
<derzahl> i added i2c_hid to the initrd which i think is the kb driver
iivanov has left #aarch64-laptops [#aarch64-laptops]
iivanov has joined #aarch64-laptops
alpernebbi has joined #aarch64-laptops
<steev> derzahl: with 5.12 it should be i2c_hid_of
<steev> the name changed in 5.12
<steev> [c630 ~] > lsmod | grep i2c
<steev> i2c_hid_of 16384 0
<steev> i2c_qcom_geni 24576 0
<bamse> derzahl: the idea is that userspace will detect entries in /sys/class/firmware and help load firmware, or tell the kernel that it doesn't have the file
<bamse> derzahl: the idea of CONFIG_FW_LOADER_USER_HELPER_FALLBACK that is
<bamse> derzahl: perhaps debian has a fw-loading thing running in userspace?
<derzahl> steev: tried including i2c_hid_of too. still cant type at the pw prompt. think theres another one needed? whats i2c_qcom_geni for?
<derzahl> bamse: isnt that what the rmtfs, qrtr, pd-mapper and tqftpsrv daemons are? im not really clear on why they are required but I thought it was to load firmware
<derzahl> either way, its working great now with your timeout hack. not sure if theres any downsides to setting it to 1sec
<bamse> derzahl: those will get the ath10k firmware loaded and going, but then the ath10k driver loads some calibration data in runtime
<bamse> derzahl: essentially /lib/firmware/qcom/** is loaded by the prio, but then ath10k loads stuff from /lib/firmware/ath10k and those are the files that we're hitting the timeouts on
<derzahl> steev: ok including i2c_qcom_geni makes it work. thanks!
<derzahl> bamse: ah ok, got it. well this is great. im actually making progress on this c630 now
<derzahl> should be able to use an encrypted zfs root now
<derzahl> steev: do we still need any of the extra kernel options on 5.12? "efi=novamap pd_ignore_unused clk_ignore_unused" ?
macc24 has quit [Quit: WeeChat 3.2]
macc24 has joined #aarch64-laptops
macc24 has quit [Quit: WeeChat 3.2]
macc24 has joined #aarch64-laptops
<steev> i don't think novamap is required with the debian setup from shawn
<steev> but don't quote me on it\
<steev> i'd leave clk ignore unused enabled still, but i believe pd can go
<derzahl> yeh, i think the notes for the latest debian image said that none of the options are required anymore
<derzahl> not sure if thats because of the kernel or something else they did
<derzahl> so...uh oh. i f'd something up
<derzahl> can someone tell me what device /dev/disk/by-partlabel/modemstl points to on their c630?
<derzahl> and what the partition layout of that root device is
<derzahl> i think i accidentally rewrote the MBR on that device and now my wifi doesnt load
<derzahl> yeh i did
<steev> lrwxrwxrwx 1 root root 10 Apr 12 13:21 modemst1 -> ../../sdf2
<steev> lrwxrwxrwx 1 root root 10 Apr 12 13:21 modemst2 -> ../../sdf3
<derzahl> can you show me "fdisk -l /dev/sdf" ?
<derzahl> also "lsblk -o NAME,LABEL /dev/sdf"
<derzahl> thanks
<derzahl> oh weird, its not showing the labels
<derzahl> i need to know all the labels on the partition on /dev/sdf so i can put them back. and pray it works
<derzahl> awesome, thanks man!
<derzahl> hopefully the UUID didnt change. or if it did, it doesnt matter
<derzahl> fingers crossed
<derzahl> any idea what the deal with all these separate block devices are?
<steev> stuff, and things
<derzahl> hm. wonder how i set partition type to 'unknown'
<derzahl> guess it doesnt matter
<derzahl> as soon as the labels were put back my wifi came right up!
<derzahl> wheew
<derzahl> thanks a ton man. youre a lifesaver
* derzahl should probably dd all those weird little devices to files
<derzahl> 10GB of weird little devices compressed down to ~10MB of image dumps. definitely worth the peace of mind
<derzahl> hey steev, if you were to one day stop making new kernels, is there a doc on how to do it?
<derzahl> are you constantly pulling latest versions of various patches or are you just merging new kernel source into your tree?
<steev> the point i stop making kernels is the point where everything is in the kernel and it doesn't need to be customized
<steev> both
<steev> sometimes i don't pull the latest version of a patch because it's based on -next or something in -next
<derzahl> is there a list of all the patches you apply?
<derzahl> any other tweaks besides applying the patches?
<steev> everything above <X kernel release> (which is usually done via git pull --rebase)
<steev> i might tweak the config, but i try to commit it, so that i don't have to remember to, each time (see above where i mentioned my memory sucks)
<steev> i usually try to apply everything that's still applicable on linus' tree, then yell at bamse that his stuff is broke again
<steev> i also have a bunch of remotes, and i use pwcli to pull in the patches from the mailing list
<steev> disclaimer: i will avoid using alsa-devel's patchwork as much as i possibly can, because despite their web interface showing only a few hundred patches, the cli interface lists something like 60k and it takes like 20-30 minutes to load
<steev> i usually poke through the linux-msm mailing list, and at this point some people will CC me on patches as well because they know i'll test them
<derzahl> ah
<derzahl> well it sounds pretty involved. appreciate your work. I guess I will just keep my fingers crossed that you dont get hit by a bus:)
alpernebbi has quit [Quit: alpernebbi]
<steev> it's not that involved
<steev> can be time consuming is all
mattst88 has joined #aarch64-laptops
<derzahl> so something i did made grub change device names. now hd1, is hd5 according to grub. wtf
<derzahl> any idea what could have caused that? not sure how grub decides disk order
<bamse> derzahl: and you didn't connect/disconnect your usb drive inbetween?
<derzahl> nope, no external drives have been near it
<derzahl> i did accidentally run 'zpool create' on the device the stores the wifi firmware...or something
<derzahl> modemst1 and modemst2 partition