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> well, i'm not sure what the latest windows restore usb is, honestly. unless you made one ON the c630, because an amd64 one wouldn't work afaik
iivanov has joined #aarch64-laptops
iivanov has quit [Ping timeout: 480 seconds]
iivanov has joined #aarch64-laptops
<steev> shawnguo: for aarch64-laptops/linux 5.13 branch, would you mind pulling in https://patches.linaro.org/patch/460614/
<shawnguo> steev: done
<shawnguo> only compile testing, trust you and bamse on runtime testing ;)
<steev> si, it works here once i had modemmanager installed (or i should say, i get the modem device, i don't actually have a sim card for the c630)
<bamse> steev: apparently i don't have the ipa firmware on this machine, so have to look at the other one...but i think windows has a .elf and you should just copy that as /lib/firmware/ipa_fws.mdt
<bamse> steev: but alex was recently adding support for specifying the firmware name in dt, so it might be that you can use firmware-name to point it to qcom/LENOVO/81JL/ipa_fws.elf
<steev> bamse: yeah my problem was not having userland stuff, without modemmanager, it would give that timeout error
<steev> i still haven't tracked down why audio is broken on one of the c630s, but not the other
<steev> getting errno 5, so no idea why that's happening
alpernebbi has joined #aarch64-laptops
<bamse> steev: you got a timeout from the kernel if you don't have the user space stuff?
<bamse> steev: i upgraded my userspace on my daily driver-c630...so now i don't have audio anymore...so i guess it's time to spend the effort and bump the kernel
<derzahl> steev: you can download a usb creator from lenovo for the c630. Can run on any windows. I ended up finding the img that I ripped from one I created months ago though
<steev> bamse: yeah
<steev> derzahl: ah yeah, that's the one i used
<steev> bamse: on one c630 i just get timeouts on audio, not sure why, on the other though, audio works fine
<steev> userland side, everything is the same, same firmwares, same ucm2 directory
<steev> http://dpaste.com/3PE5L95ZL is the dmesg output of the non-working audio
<steev> https://dpaste.com/3PE5L95ZL#line-442 is where things seem to start going down hill?
<steev> rebooted it into windows and audio works there
<derzahl> does anyoneremember if there was some trick to boot the c630 from usb? tried a few different usb disks with the restore img. usb boot is enabled in the bios. I hit fn-12 to get into the boot menu but it only shows "windows boot manager", no usb option
alpernebbi has quit [Quit: alpernebbi]
<steev> usb has to be smaller than 64gb iirc
<steev> if you're on the latest bios
<steev> i'd have to check the lenovo forums, where i bitched about it, and they said they'd have someone look into it... and then promptly ignored it for the rest of time
<steev> First is that it no longer boots off of USB keys that are larger than 32GB without a little... help... If you put 2 USB keys that are larger than 32GB in (so 1 in each USB port), then the one on the left side (sim card, power led) will show up as "UEFI: USB Hard Drive", whereas the one on the right (headphone jack, power button) will show up as "UEFI: USB USB Hard Drive" and it will boot from the device in the left port.
<steev> If only 1 USB key larger than 32GB is plugged in, it shows up as "UEFI: USB USB Hard Drive" and will not boot from it, even when selecting it with the boot prompt. It simply boots into Windows. Booting from a 128GB or even 256GB USB key was possible without the hoops of connecting multiple devices on the 1.23 firmware, so it would be nice if whatever change was made that makes this not work, could be reverted?
<macc24> does anyone know if cameras are supposed to work on krane?
<robclark> is it a usb cam or mipi? The former should usually "just work"(TM), the latter not so much
<macc24> uh
<macc24> i don't know to be honest
<macc24> in dts it's i2c device that is hooked up to "seninf" <- whatever that is
<macc24> and seninf node points at "camisp"
<robclark> that sounds like mipi
<robclark> you can check lsusb tho to be sure
<macc24> i'm pretty confident there's no camera on usb
<macc24> on duet
<robclark> if it has two cameras, it's possible one is usb and the other mipi
<macc24> i kinda hope it's midi as i think it would have better image quality?
<macc24> because usb is limited and not made for cameras
<macc24> s/midi/mipi/
<robclark> mipi *can* be better than usb, but typically requires giant huge piles of isp driver (at least to get decent image quality)..
<macc24> and there is nothing mtk isp related in /the/ mainline tree
<robclark> right, that is typically the problem w/ mipi cameras.. I vaguely remember hearing that someone somewhere is working on a new driver subsystem that can handle complex ISPs better than v4l2, so maybe when that eventually happens you start seeing some of the isp driver stuff showing up upstream
<macc24> damn i kinda wish mediatek engineers used irc
<macc24> robclark: would that be compatible with... v4l2 things or would it require brand new programs for everything?
<robclark> no idea really.. but I suspect that, at least to get good image quality, it will require some userspace parts
<macc24> well due to all the linux phones showing up there are some programs for cameras
<robclark> in android phones, there are large userspace components.. there are probably as many engineers working on camera driver as gpu driver
<macc24> android phones are chasing literally state of the art quality
<steev> just connect the camera over ethernet, ez
<robclark> true, but without ISP magic, the image quality would be actually quite terrible
<macc24> steev: built-in security system xD
<derzahl> crap. well i was using 16GB usb drives. theyre cheap POS's. but they didnt show up at allin the boot menu. currently writing a 64GB samsung but i guess that wont work:(
<robclark> basically usb cameras have the complex isp bits in the camera themselves so all the hard parts are transparent to the OS
<derzahl> ill have to look around and see what else i have
<steev> derzahl: fwiw, i use https://www.amazon.com/gp/product/B01EZ0X23W/ - i've got like 8 of them at this point (i like them because of the dual usb component letting me use them with all my computers)
<macc24> robclark: oh fuck
<steev> robclark: fwiw... fw_devlink=off reduced it... but i've run into [drm:dpu_encoder_frame_done_timeout:2161] [dpu error]enc31 frame done timeout again
<steev> seems to happen when i'm using alacritty
<steev> setup here is xfce on xorg
<robclark> steev: maybe you can throw in a dump_stack() where it hits the timeout so I can (hopefully) see the call path?
<steev> would i put that as part of the DPU_ERROR_ENC or before it, or after it?
<steev> eh, i'll put it before, what's the worst that could happen
<robclark> just at the end of `dpu_encoder_frame_done_timeout()` should probably work from the looks of it
<steev> ah okay, i was gonna put it before the message, but that works too
<derzahl> ok i found i 32gb sandisk drive. i think thats the one i used with the c630 before
<derzahl> gonna try that
<steev> it would be nice if i could get the stupid audio drivers to stop spamming the shit out of dmesg
<steev> hm, maybe if i blacklist snd_soc_sdm845 it will block it
<steev> now to hopefully trigger it
<steev> robclark: hasn't happened yet, but i did get this thingie
<steev> [ 1357.465370] [drm:dpu_core_irq_callback_handler [msm]] *ERROR* no registered cb, idx:27 enable_count:0
<steev> that might be ignorable if dmitry's patch to get rid of the irq stuff gets in?
<derzahl> lol. of course. now the usb creator utility cant even write to the sandisk. lets see. got some other cheapo 16GB sticks. and a 32GB sdcard with usb adapter i can try
iivanov has quit [Remote host closed the connection]
<derzahl> wow, this crappy old sdcard adapter thats cracked and needs to be held together with tape and oldass 32gb sdcard is the only thing that boots on the c630
<derzahl> another odd thing i was getting with your 5.12 kernel steev. After including the i2c_hid and i2c_hid_of modules in the initrd, I was able to see the luks pw prompt and type the password. But when booting with encrypted ZFS root, I could only type the password with 5.11rc7. 5.12 has no keyboard functionality at the ZFS pw prompt. wonder if im missing another module or something. But that wouldnt make sense, because 5.12
<derzahl> works for a luks pw prompt