2025-01-02 10:26:57 <ncopa> clandmeter: btw, I have not set up the CI for the riscv machine we replaced storage for
2025-01-02 10:27:18 <ncopa> so riscv64 CI is on reduced capacity
2025-01-02 10:30:42 <clandmeter> Ok, not sure if i have time today. need to fix some other stuff at the house
2025-01-02 10:31:14 <clandmeter> i also think i know how to fix the bootloader menu not showing up on serial
2025-01-02 10:49:52 <mps> clandmeter: you use grub?
2025-01-02 10:53:57 <ncopa> we want be able to use grub for the pioneer machines
2025-01-02 10:54:39 <ncopa> or let me put it this way: we want be able to build and update kernels for pioneer without needing to rebuild and update the boot loader(s) with erevery kernel update
2025-01-02 10:57:13 <mps> does it use u-boot
2025-01-02 10:57:44 <mps> iirc when I looked at it u-boot is in 'chain' there
2025-01-02 11:09:51 <ncopa> no
2025-01-02 11:10:32 <ncopa> well it may use uboot, but I don't know how, and i dont find much documenation for u-boot on pioneer/sg2042
2025-01-02 11:11:17 <ncopa> they use u-root, which basically is a linux kernel with nvme drivers that will kexec kernel from nvme
2025-01-02 11:11:29 <ncopa> or they use edk2
2025-01-02 11:11:34 <ncopa> which loads grub
2025-01-03 19:35:21 <mps> anyone of you use riscv64 as desktop?
2025-01-03 19:43:07 <ncopa> no
2025-01-03 19:43:23 <ncopa> still too slow compared to the i9 and macbooks
2025-01-03 19:46:41 <mps> I've built telegram-desktop on riscv64 and not sure should it be pushed to aports. it is slow to build
2025-01-04 22:12:49 <mps> anyone knows why linux-edge 6.12.8 is not built in 3.21-stable
2025-01-04 22:13:06 <mps> or is not uploaded
2025-01-06 07:44:08 <ncopa> which arch?
2025-01-06 07:44:49 <ncopa> it is built https://build.alpinelinux.org/buildlogs/build-3-21-aarch64/community/linux-edge/
2025-01-06 07:45:15 <ncopa> it is uploaded https://dl-master.alpinelinux.org/alpine/v3.21/community/aarch64/linux-edge-6.12.8-r0.apk
2025-01-06 07:48:32 <mps> riscv64
2025-01-06 07:48:56 <mps> sorry, forgot to mention
2025-01-06 07:52:56 <ncopa> oh. not good :-/
2025-01-06 08:02:02 <ncopa> seems like it hadnt built anything since 29 Dec
2025-01-06 08:02:11 <ncopa> same with build-3-20-riscv64
2025-01-06 08:02:19 <ncopa> thanks for the ping mps
2025-01-06 08:24:54 <clandmeter> ncopa: there is an issue with that box
2025-01-06 08:25:05 <clandmeter> when it boots it need manual fw rules
2025-01-06 08:25:17 <clandmeter> i think this is what blocks the builds
2025-01-06 08:25:35 <clandmeter> im testing edk2 on pioneer2
2025-01-06 08:32:32 <ncopa> 👍
2025-01-06 08:42:27 <clandmeter> different error now
2025-01-06 08:42:35 <clandmeter> like reported on github
2025-01-06 09:17:37 <mps> I deleted my OTP on an old phone and now cannot login to gitlab.a.o
2025-01-06 09:18:02 <mps> how can I solve problem
2025-01-06 09:58:27 <ncopa> mps: I disabled the two factor auth for you now
2025-01-06 10:01:02 <mps> ncopa: thank you, I've got page to reregister OTP
2025-01-06 10:01:33 <ncopa> you're welcome
2025-01-06 12:00:24 <mps> huh, world is becoming too much complicated. my phone can't scan QR code to enable 2FA :|
2025-01-06 12:04:43 <clandmeter> install a compatible app maybe?
2025-01-06 12:10:12 <ncopa> maybe clean the camera lens? :)
2025-01-06 12:10:43 <mps> ncopa: :)
2025-01-06 12:11:46 <mps> I'm using FreeOPT+
2025-01-06 15:05:20 <clandmeter> ncopa: seems the edk2 relocation error happens when its loading the grub stub 
2025-01-06 15:05:51 <clandmeter> when renaming it on nvme it does go further, but still does not show me efi shell 
2025-01-06 15:15:53 <clandmeter> i just broke the boot sdcard
2025-01-06 15:16:02 <clandmeter> what a joy
2025-01-06 18:20:55 <mps> tried two otp programs on two phones, no one can recognize QR code on gitlab.a.o when I'm trying to enable 2FA, and looks like gitlab.a.o will be lodked for me in near day or two /o\
2025-01-06 18:21:38 <mps> can someone tell which program works on android
2025-01-06 18:23:05 <ikke> I use authy, but that's from twillio
2025-01-06 18:25:32 <mps> authy cannot be downloaded if user is not registered, iiuc?
2025-01-06 18:26:25 <ikke> iirc, an account is not required
2025-01-06 18:34:03 <mps> ah yes, https://www.authy.com/download/
2025-01-06 18:40:06 <mps> otpclient on alpine says "can't find a valid qrcode" when I feed image to it
2025-01-06 18:53:04 <mps> oh, otpclient and set it manually worked
2025-01-06 18:59:15 <mps> I guess now is all ok
2025-01-06 18:59:49 <mps> maybe 'hope' instead of 'guess' is better term
2025-01-07 08:41:28 <iggy> I just use oathtool
2025-01-07 09:44:20 <mps> iggy: first time hear for it, looks interesting. thanks
2025-01-07 15:10:34 <valerius> I use oathtool on my desktop, you can also import the keys into a program called "Aegis" on Android
2025-01-07 15:10:52 <valerius> an "app", sorry
2025-01-07 15:10:55 <valerius> 
ACTION strokes grey beard

2025-01-07 16:05:52 <mps> could be qrcode read by oathtool?
2025-01-07 16:06:27 <mps> to reformulate, could oathtool read qrcode?
2025-01-07 16:22:23 <iggy> mps: I don't think so, but every 2fa prompt I've ever seen either showed the key or there was a little button/link you could click to show it
2025-01-07 16:46:46 <panekj> mps: FreeOTP+ works on Android
2025-01-07 16:48:21 <mps> panekj: yes, but can't read qrcode on gitlab.a.o
2025-01-07 16:48:44 <panekj> what can I say, except for "works for me"
2025-01-07 16:49:13 <panekj> I had it set up before and just did it again to check if it didn't break and it worked
2025-01-07 16:49:30 <mps> could be problem with lineageOS on my phone
2025-01-07 16:49:40 <panekj> maybe you have time wrong?
2025-01-07 16:50:03 <mps> no, time is ok
2025-01-07 16:50:32 <panekj> does it just not understand the qrcode?
2025-01-07 16:51:14 <mps> though have to add that I don't have much experience with OTP, I use it only when must
2025-01-07 16:52:00 <mps> panekj: yes, it says "can't find a valid qrcode" when trying to read it
2025-01-07 16:52:20 <ikke> mps: what if you try some kind of qr-code scanner?
2025-01-07 16:52:32 <panekj> I would recommend using just the text code to add otp
2025-01-07 16:52:33 <mps> qrcode image is in firefox window
2025-01-07 16:53:23 <mps> same phone quite fine read qrcodes in telegram
2025-01-07 16:53:35 <mps> and one from DigiCert
2025-01-07 16:54:16 <mps> ikke: iiuc this will be useful only to test if qr-code is ok?
2025-01-07 16:55:30 <mps> anyway, what iggy says looks like it is a lot easier to use command line tools than these on android
2025-01-07 19:38:34 <valerius> I switched from lineageOS to GrapheneOS
2025-01-07 19:38:40 <valerius> I think I got Aegis from F-Droid
2025-01-07 19:45:03 <mps> 
ACTION wants to throw away all android devices

2025-01-07 19:45:45 <valerius> GrapheneOS is realistically the best option these days, unless you want to sacrifice basic functionality to use a niche platform
2025-01-07 19:46:03 <valerius> it at least makes me feel like I'm the one in control of the device for the first time ever in mobile devices
2025-01-08 04:10:03 <sms> Hi, does anyone have any experience with alpine on the star64 by pine64?
2025-01-08 04:10:46 <sms> I'm trying to build a wifi driver currently and was wondering if anyone had some advice
2025-01-08 07:18:01 <ncopa> looks like it is StarFive JH7110, so it should boot at least?
2025-01-08 07:18:52 <ncopa> do you know which wifi driver you need?
2025-01-08 09:05:10 <sms> yea it boots, i built an image with the script on this site: https://arvanta.net/alpine/alpine-on-visionfive/
2025-01-08 09:06:36 <sms> i found a driver on github, problem is i get the following error while compiling it: https://github.com/morrownr/rtl8852bu-20240418/issues/5
2025-01-08 09:07:05 <sms> perhaps i'm missing a library or something else necessary to build it
2025-01-08 09:10:59 <mps> sms: on which distro you build it
2025-01-08 09:13:12 <mps> I thought that wifi driver for star64 is in mainline kernel
2025-01-08 09:33:31 <sms> mps i built it on alpine, on the same board i installed it to
2025-01-08 09:34:05 <sms> is it? i tried adding linux-firmware to the image but it didn't recognize the card
2025-01-08 09:36:40 <mps> sms: I'm not sure it is in mainline, I just thought it is but I don't know for sure
2025-01-08 09:39:14 <sms> i dont think so, given the github project i linked is still a wip
2025-01-08 09:40:22 <mps> could you try to boot with linux-lts
2025-01-08 09:41:04 <mps> or linux-edge
2025-01-08 10:10:42 <mps> here is my Makefile for build out-of-tree modules https://tpaste.us/VYpv
2025-01-08 10:11:04 <mps> ofc, you can (and maybe should) replace clang with gcc
2025-01-08 10:12:44 <mps> this example is when I built my mpschar module
2025-01-08 13:12:41 <ncopa> it is probably lts.riscv64.config:CONFIG_TRIM_UNUSED_KSYMS=y
2025-01-08 13:13:36 <ncopa> yeah almost sure that is the issue
2025-01-08 13:14:24 <ncopa> it will fail on edge kernel as well
2025-01-08 13:15:01 <ncopa> $ grep UNUSED_KSYM * | tpaste
2025-01-08 13:15:01 <ncopa> https://tpaste.us/mepg
2025-01-08 13:30:43 <mps> nice, now have to fix this also on aarch64 and x86_64 in linux-edge
2025-01-08 13:31:01 <mps> and armv7
2025-01-08 13:38:46 <mps> u-boot 2025.01 pushed to builders
2025-01-08 13:39:47 <mps> it was one of easiest release to upgrade
2025-01-08 23:40:27 <sms> mps it seems linux-lts and linux-edge don't have the DTB's needed to boot on the star64, so i can't boot with just those packages
2025-01-08 23:41:39 <sms> unless i should manually include the DTB's alongside the linux-lts kernel, but that requires building the kernel myself correct?
2025-01-08 23:42:26 <sms> ncopa nice, that could be it. how do i turn that option off?
2025-01-09 06:52:26 <sms> mps nvm it does
2025-01-09 07:28:18 <mps> sms: iiuc linux-lts and linux-edge boots on star64 at the end?
2025-01-09 07:29:40 <ncopa> sms: im on it
2025-01-09 07:29:41 <mps> sms: to turn of ksyms clean from kernel and modules you have disable these options in kernel config and rebuild kernel and install it
2025-01-09 07:31:23 <mps> maybe time linux-{lts,edge}-dbg :)
2025-01-09 07:31:32 <mps> time for*
2025-01-09 07:32:42 <ncopa> star64 is JH7110 so I expect 6.12 kernel boots there
2025-01-09 07:33:21 <mps> yes, because this I told sms to try boot with lts and edge
2025-01-09 07:34:20 <mps> only don't know how mainline u-boot works on this SBC
2025-01-09 07:35:50 <sms> mps i was confused for a second on how to boot it, i think i got it now, i'll let u know if it works
2025-01-09 07:37:57 <mps> sms: if it can't find dtb in fdtdir you can set 'fdt $path to required dtb' in extlinux.conf
2025-01-09 07:39:51 <sms> i still would need to flash u-boot.itb and u-boot-spl.bin.normal.out from the u-boot-starfive package unto the sd card correct?
2025-01-09 07:40:20 <sms> i made a label for lts kernel on extlinux.conf
2025-01-09 07:42:30 <mps> we didn't enabled u-boot for star64
2025-01-09 07:42:47 <mps> we is me
2025-01-09 07:42:55 <mps> but no one else afaik
2025-01-09 07:44:04 <mps> probably you don't have to install new u-boot on board, just set path to dtb in 'fdt' directive in extlinux.conf
2025-01-09 07:44:41 <mps> u-boot-starfive is tested only on visionfive V2
2025-01-09 07:47:12 <mps> I see this 'board: support Star64 board' in u-boot git log. So probably we can enable it
2025-01-09 07:47:32 <mps> sms: could you post me official url for this SBC
2025-01-09 07:48:30 <mps> I will be AFK for some time, and later today I will look if we can add this SBC to our u-boot
2025-01-09 07:52:01 <mps> sms: nvm, I found url
2025-01-09 07:55:34 <sms> sorry for not responding. nice
2025-01-09 07:56:25 <sms> mps i've been using u-boot-starfive to boot on the star64, and it works
2025-01-09 07:57:01 <sms> i thought it made sense since its the same cpu as well. the board's spi comes with those dtbs as well i'm pretty sure
2025-01-09 07:57:12 <mps> looking u-boot source it could be that this SBC could work with visionfive V2 u-boot
2025-01-09 07:57:46 <mps> sms: ah nice, you just confirmed my guess
2025-01-09 07:58:16 <mps> so we can add in docs this
2025-01-09 07:58:27 <mps> BBL
2025-01-09 07:58:38 <sms> also tested linux-lts, it boots
2025-01-09 07:59:17 <sms> that grep UNUSED_KSYMS that ncopa did, where do i run that to see if the option is enabled?
2025-01-09 08:27:48 <sms> linux-edge also boots mps
2025-01-09 08:36:29 <mps> sms: it is in alpine aports tree
2025-01-09 08:37:52 <mps> sms: nice and thanks for info. maybe you can write some short notes/guide how to boot alpine on this SBC.
2025-01-09 08:38:39 <mps> would be nice to add this notes to https://arvanta.net/alpine/alpine-on-visionfive/
2025-01-09 08:54:35 <sms> sure
2025-01-09 11:39:33 <mps> sms: maybe you could build akms aport for rtl8852bu device
2025-01-09 12:44:56 <mps> ok, I've found reason why my BPI F3 sometimes hungs on boot from nvme with kernel 6.13-rc1. for unknown reason (at least for me) it doesn't initialize nvme successfully or it doesn't load nvme modules
2025-01-09 12:45:32 <mps> building kernel with nvne modules as =y it never hungs
2025-01-09 12:46:07 <mps> could be flaky nvme device or driver is not yet stable
2025-01-09 18:58:35 <mps> I narrowed problem with boot hungs. It is nlplug-findfs problem
2025-01-09 18:59:07 <mps> here is what appears on console https://tpaste.us/Eb4D
2025-01-09 18:59:25 <mps> but nvme driver is in-kernel
2025-01-09 19:02:16 <mps> ncopa: ^ do you have any idea
2025-01-09 20:53:32 <sms> mps might you hv a guide on hand for building a port? i'm a little new xd
2025-01-09 20:55:18 <mps> sms: this is old but still could be usable https://wiki.alpinelinux.org/wiki/Developer_Documentation
2025-01-09 20:55:52 <mps> i.e. to setup alpine development environment and use it
2025-01-09 20:56:32 <ncopa> looks like kernel goes boom after nlplug-findfs fails to write
2025-01-09 20:56:43 <ncopa> its definitively a kernel problem
2025-01-09 20:56:49 <ncopa> nlplug-finds loads kernel modules
2025-01-09 20:57:02 <ncopa> and writes output (logs etc)
2025-01-09 20:58:50 <mps> sms: if you want to skip and go straight to work (like me) then https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package
2025-01-09 20:59:18 <sms> thanks mps!
2025-01-09 20:59:52 <mps> ncopa: btw, linux-spacemit doesn't have this problem
2025-01-09 21:02:36 <mps> what makes me 'angry' is that I can't reproduce problem in controlled way. sometimes it hangs sometimes doesn't
2025-01-09 21:03:09 <mps> anyway, will continue to find what is culprit
2025-01-09 22:18:44 <sms> wierd, driver now compiles and installs fine, but i still can't see the wlan interface
2025-01-09 22:20:22 <sms> am i missing something?
2025-01-09 22:23:13 <mps> sms: modprobe?
2025-01-09 22:23:23 <mps> sorry for stuppid answer
2025-01-09 22:24:48 <sms> perhaps, the install script that the driver maintainer made doesn't have a modprobe command in it
2025-01-09 22:24:53 <sms> it does have depmod
2025-01-09 22:25:59 <mps> manually `modprobe /path/to/module`
2025-01-09 22:26:20 <sms> ok
2025-01-09 22:26:37 <sms> also i'm not using dkms for this atm, if it matters
2025-01-09 22:26:51 <mps> I understand
2025-01-09 22:27:42 <mps> you can manually copy module to /lib/modules/$KVER/kernel/...' where you want
2025-01-09 22:27:52 <mps> and then run depmod
2025-01-09 22:30:39 <mps> last option is `insmod $module_filename' in dir where is module
2025-01-10 04:34:10 <sms> so, i ran modprobe and the dmesg told me that it loaded, despite "module verification failed: signature and/or required key missing"
2025-01-10 04:34:26 <sms> however, no wlan0, nor any new interface
2025-01-10 04:34:35 <sms> after running ip a
2025-01-10 04:42:12 <sms> lsmod says its loaded too
2025-01-10 07:56:16 <mps> sms: with which kernel you built module?
2025-01-10 07:56:54 <mps> CONFIG_MODULE_SIG should be disabled in this kernel
2025-01-10 07:58:01 <mps> also module shouldn't be striped after build
2025-01-10 07:58:20 <clandmeter> mps:  did you try debug_init?
2025-01-10 07:58:47 <mps> clandmeter: question is to me?
2025-01-10 07:59:10 <clandmeter> i think so
2025-01-10 07:59:26 <mps> aha, what is debug_init
2025-01-10 07:59:37 <clandmeter> cmdline option
2025-01-10 07:59:53 <clandmeter> gives debug info about nlplug-findfs
2025-01-10 07:59:56 <mps> it is added to alpine initramfs?
2025-01-10 08:00:25 <clandmeter> add it to your kernel cmdline 
2025-01-10 08:00:29 <mps> tbh didn't knew it exists
2025-01-10 08:00:36 <clandmeter> RTFS ;-)
2025-01-10 08:01:28 <clandmeter> https://git.alpinelinux.org/mkinitfs/tree/initramfs-init.in#n487
2025-01-10 08:01:35 <mps> will look later when I boot board
2025-01-10 08:02:17 <mps> hah, every day I learn something surprising :)
2025-01-10 08:02:35 <mps> clandmeter: thanks for hint
2025-01-10 08:02:50 <clandmeter> yw
2025-01-10 08:03:01 <clandmeter> maybe it helps debugging the issue
2025-01-10 08:03:27 <mps> anyway, when booted board is quite stable even with kernel 6.13-rc1
2025-01-10 11:36:04 <mps> clandmeter: with debug_init it hangs rarely, about once in 15 reboots
2025-01-10 13:13:10 <ncopa> mps: i believe it is a race that happens when nlplug-findfs loads the modules
2025-01-10 13:14:18 <mps> ncopa: nvme drivers are built in kernel, and also used FS drivers
2025-01-10 13:15:31 <mps> now I build https://github.com/jmontleon/linux-bianbu/commits/linux-6.6.y/ kernel and will test with it
2025-01-10 14:07:06 <sms> i did it for 6.12.9-0-lts mps
2025-01-10 14:11:58 <sms> also for akms it looks like i need a user called akms, didn't find that in the manpages
2025-01-10 14:14:37 <mps> you have to disable CONFIG_TRIM_UNUSED_KSYMS in kernel and rebuild it
2025-01-10 14:16:52 <mps> for akms I'm not sure. you can look at some examples in aports
2025-01-10 14:34:28 <ncopa> f0ae82f0e1337a252070e267c374eb3e7a25c3e2
2025-01-10 14:34:46 <ncopa> thats included in the 6.12.9 kernel
2025-01-10 14:43:45 <sms> yea i was gonna say. btw thanks ncopa!
2025-01-10 14:45:20 <sms> also another funny thing, with akms if i make an AKMBUILD like the example, the build() will only do the MODPOST on Module.symvers, but if i leave build() with just make -j4 it does what it needs to xd
2025-01-10 15:00:44 <sms> also got a working AKMBUILD
2025-01-10 15:01:05 <sms> same problem though
2025-01-10 15:51:14 <mps> sms: you can look example here https://git.alpinelinux.org/aports/tree/testing/rtl8821ce-src/APKBUILD
2025-01-10 18:13:23 <sms> thanks mps