jjsullivan1 has joined #nixos-aarch64
h0m1 has quit [Ping timeout: 246 seconds]
h0m1 has joined #nixos-aarch64
noonien has quit [Quit: Connection closed for inactivity]
<samueldr> aaaaaaaaaaaaaand TIME!
<samueldr> I'm on the desktop of the rootfs image, on a device which I did no work other than use the (soon to be published) autoporter, and adapting the kernel config according to what a similar device does
justanotheruser has joined #nixos-aarch64
rajivr has joined #nixos-aarch64
orivej has quit [Ping timeout: 246 seconds]
<ashkitten> autoporter :o
<ashkitten> so you just give it a stock rom and it automatically adds a new device to the mobile-nixos tree then?
<samueldr> not even give it, it slurps data from the "android dumps" project
<samueldr> so I can produce a skeleton for a device I don't have!
<samueldr> though a goal is to eventually allow giving a ROM to the tool, rather than rely on a (sometimes flaky) third party service
<samueldr> though giving a ROM is a whole can of worms and a half
<samueldr> as there are *multiple* ways a ROM can be packaged
<ashkitten> oh hell yeah
<samueldr> though the work has mostly been done in a way where the hardcoding is superficial
<samueldr> don't know if you saw it on #nixos-chat
<ashkitten> wow sick
<samueldr> with the details drakonis shared it should become much much better at detecting quirks
<samueldr> boooo, motorola downcases their manufacturer name and device names in the config
<samueldr> (I guess that's anyway how they want it written now)
<samueldr> just tried running it on motorola def
<ashkitten> oh you have motorola-def too?
<ashkitten> have you managed to make the second slot bootable
<ashkitten> cuz i havent
<Ke> did anyone here get that helios64 yet
<samueldr> ashkitten: I don't have it
<samueldr> I just wanted to check since it's a "known" device around here
<samueldr> I know more about def than many devices I don't own :)
<samueldr> so it's an additional data point to test
<samueldr> Ke: still not here, though it should be on the seas, probably soon arriving in the country
<samueldr> Ke: anything particular in mind?
<Ke> I want to know, how loud it is
<samueldr> right, then that won't be something I can do without it :)
<Ke> of course if there is something you would not expect from rk3399 board, I would like to know
<samueldr> the board itself shouldn't _require_ active cooling
<Ke> but I have expectation that it would just work
<samueldr> yeah
<Ke> well it has 2 80mm fans still
<samueldr> I assume so too
<samueldr> yeah, that's what I was looking, if it had fans
<samueldr> so I guess it'll depend on that
<samueldr> and the noise of the disks
<samueldr> pwm controlled, so likely it can be quiet
<Ke> I am just thinking, what I should do, if I can't make mcbin work
<Ke> it's maybe not worth spending too much time with, as I believe it's slightly broken
<samueldr> once I have one I'll be able to say if it's good or not, probably also run tests for people with a "test array" of disks
<samueldr> I'll have to play around with setups since I never did anything like that
<Ke> ok actually the m2 sata is shared, so it has too few disk slots
<samueldr> "it" being?
<Ke> helios64
<samueldr> I'm confused to the meaning of what you said
<samueldr> ah
<Ke> if you use the m.2 sata on helios64 you will lose one disk slot
<samueldr> yes
<samueldr> and you needed 5 + m.2 for whatever you're doing?
<Ke> well my disk array currently has 4 disks and 2 ssds
<samueldr> I don't *really* know what I'll be doing with mine :)
<Ke> not that I really need the ssds
<samueldr> I mean, as a setup
<samueldr> it'll most likely replace the old 2012 computer that's serving as "poor man's NAS"
<samueldr> probably also learn a bit about zfs, since it seems to be all the rage with storage
<Ke> ZFS with 4g might not be good
<Ke> of ram I mean
<samueldr> as far as I've heard, it seems it wouldn't be an isssue
<samueldr> issue*
s1ng0c has joined #nixos-aarch64
<samueldr> but proper information is hard to come by with all the parroted info :/
s1ng0c has quit [Quit: WeeChat 2.9]
quinn has quit [Ping timeout: 260 seconds]
quinn has joined #nixos-aarch64
<sphalerite> Ke: I'm using zfs on my NAS with 4G, it's fine. dedup is completely out of the question of course, but 4G is fine if you don't need amazing performance.
cole-h has quit [Ping timeout: 240 seconds]
<julm> I'm also using ZFS with only 4G of RAM and no-dedup, so far the arc_summary report looks fine to me, though I'm not running really intensive tasks
aforemny has joined #nixos-aarch64
V has quit [Quit: V]
V has joined #nixos-aarch64
s1ng0c has joined #nixos-aarch64
<s1ng0c> quit
s1ng0c has quit [Client Quit]
s1ng0c has joined #nixos-aarch64
justanotheruser has quit [Ping timeout: 272 seconds]
s1ng0c has quit [Remote host closed the connection]
user2 has joined #nixos-aarch64
user2 is now known as s1ng0c
s1ng0c has quit [Client Quit]
s1ng0c has joined #nixos-aarch64
justanotheruser has joined #nixos-aarch64
alp has joined #nixos-aarch64
justanotheruser has quit [Ping timeout: 240 seconds]
justanotheruser has joined #nixos-aarch64
tilpner has joined #nixos-aarch64
lafa has joined #nixos-aarch64
orivej has joined #nixos-aarch64
orivej has quit [Ping timeout: 260 seconds]
zupo has joined #nixos-aarch64
s1ng0c has quit [Ping timeout: 256 seconds]
orivej has joined #nixos-aarch64
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
s1ng0c has joined #nixos-aarch64
knerten has joined #nixos-aarch64
alp has quit [Ping timeout: 246 seconds]
alp has joined #nixos-aarch64
h0m1 has quit [Quit: WeeChat 2.9]
h0m1 has joined #nixos-aarch64
alp has quit [Ping timeout: 246 seconds]
Darkmatter66 has quit [Read error: Connection reset by peer]
Darkmatter66 has joined #nixos-aarch64
alp has joined #nixos-aarch64
kahiru has joined #nixos-aarch64
aforemny has quit [Ping timeout: 260 seconds]
cole-h has joined #nixos-aarch64
s1ng0c has quit [Quit: WeeChat 2.9]
zupo has joined #nixos-aarch64
<Ke> hmm. mcbin u-boot looks better in terms of working deterministically
<Ke> now I only need to figure out, how does one boot nixos
<Ke> maybe I could copy-paste my homework https://github.com/bgamari/mcbin.nix/blob/master/mcbin.nix
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zupo has joined #nixos-aarch64
alp has quit [Ping timeout: 272 seconds]
<samueldr> Ke: usually with u-boot and nixos it's "generic distro boot"
<samueldr> and from there, we usually use the "extlinux" support
<samueldr> so as long as it's mainline u-boot, it should be good, if it's not, check that it's newer than (IIRC) mid 2017
<Ke> this is mainline/master from very recent time
<Ke> is there a way to have nixos build generate that extlinux.conf?
<samueldr> yes
<samueldr> boot.loader.generic-extlinux-compatible.enable = true;
<samueldr> you'll also need to set grub.enable to false
alp has joined #nixos-aarch64
<Ke> thanks
<samueldr> it's the default expectation for our U-Boot using ARM ecosystem
<Ke> do I need to specify dtb, I believe u-boot should manage that also?
<samueldr> shouldn't need to specify
<samueldr> if it's mainline u-boot, it has some methods to "know" which FDT to load from the directory
<samueldr> usually the name is the right one, but usually only for mainline-based kernels
<Ke> well it's built for mcbin and even u-boot uses a dtb to my knowledge
<Ke> I believe u-boot should have a dtb
<samueldr> yes, it ends up being built-in
<samueldr> but the way the linux kernel operates they really prefer you use the FDT they decide you should use :(
<samueldr> but leaves the loading to the firmware
<samueldr> which is quite an issue with NixOS due to generations
<Ke> I know this from experience, but mcbin dtb has been very stable for quite some time
<samueldr> still, u-boot with extlinux support _will_ load an FDT and fail if it can't, but as long as mainline linux is used you shouldn't see this as an issue
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
ky0ko has quit [Quit: killed]
ky0ko has joined #nixos-aarch64
rajivr has quit [Quit: Connection closed for inactivity]
aforemny has joined #nixos-aarch64
<Ke> setenv ramdisk_addr_r $initrd_addr
<Ke> setenv kernel_addr_r $kernel_addr
<Ke> sysboot -p mmc 1:1 fat $fileaddr /extlinux/extlinux.conf
<Ke> that's what mcbin needed to boot
<Ke> with master u-boot
<samueldr> configs/mvebu_mcbin-88f8040_defconfig ?
<Ke> yes
<samueldr> did it somehow load a u-boot environment?
<samueldr> because the defaults AFAICT (now looking) should just work
<samueldr> note that the partition has to be marked as bootable
<Ke> ah
<Ke> thanks
<samueldr> for the default bootcmd to work
<Ke> can it be gpt partition?
<samueldr> yes
<samueldr> I don't remember how to mark it bootable the right way
<samueldr> there is a way though
<Ke> well the big issue at first is that I do not get all interfaces up
<samueldr> I don't think bgamari ends up relying on the default distro boot concept
<Ke> well you see the config there
<samueldr> yeah, that's why I'm saying that :)
<samueldr> though weird: there's absolutely nothing about "initializing [eth2] with u-boot"
<Ke> anyway at least fdisk has LegacyBIOSBootable in expert commands
<Ke> for gpt
<samueldr> good to know, I literally don't know how to change the flag
<samueldr> if it's an ESP, u-boot will honor it with its EFI boot
<samueldr> (without having to mark it as bootable)
<Ke> hmm better not, as the efi boot crashes, I believe
<Ke> maybe I did it wrong
<samueldr> yeah, just stating things
<Ke> right now I am fine with doing things manually, I can easily work out to automate this later
t184256 has left #nixos-aarch64 ["Error from remote client"]
t184256 has joined #nixos-aarch64
justanotheruser has quit [Ping timeout: 260 seconds]
alp has quit [Ping timeout: 272 seconds]
<hexa-> hm, since I migrated my rpi4 to 20.09 I get usb resets … which is a bit meh considering my rootfs is on an SSD connected via UAS
<samueldr> hexa-: the most likely change would be the kernel version for linux_rpi4 in 20.03 vs. 20.09
<samueldr> since it's highly likely it never got udated in 20.03 post-fork
<hexa-> yep
<hexa-> can't wait to have a mainline kernel on that
<hexa-> 4.19.75 ←→ 4.19.118
<samueldr> so you could try changing the kernel version to that older one to confirm it's that or not
<samueldr> assuming the usb resets are consistent
<hexa-> true
<samueldr> if it's not that it could be... anything...
<samueldr> though if it's that, you might want to try and upgrade the 20.09 package to the most up to date 4.19 version of the rpi fork
<hexa-> yep
<samueldr> I'm still hoping that some among you have a spare aarch64 android-based device to test the autoport with
orivej has quit [Ping timeout: 256 seconds]
<hexa-> i don't
<samueldr> I definitely wasn't singling anyone out
<hexa-> :)
<samueldr> but someone with an open PR for a port would be a good candidate
<Ke> I have motog 3rd
alp has joined #nixos-aarch64
justanotheruser has joined #nixos-aarch64
aforemny has quit [Ping timeout: 256 seconds]
<samueldr> if you feel like spending some time on a project with results that are not guaranteed, we can sync up so I can stay updated with your progresses and take notes about possible problems
<samueldr> with the latest changes it detects the last few important annoying bits :)
<samueldr> looking for fun, osprey (osprey_ud2 in android dumps) is an armv7l kernel on an aarch64-capable device
<samueldr> so it's a bit like addison
<samueldr> from a quick glance (really superficial) no obvious aarch64 kernel made available
alp has quit [Ping timeout: 272 seconds]
justan0theruser has joined #nixos-aarch64
justanotheruser has quit [Ping timeout: 240 seconds]
<lopsided98> What's going on with ofborg failing to eval anything on aarch64?
orivej has joined #nixos-aarch64