h0m1 has quit [Ping timeout: 265 seconds]
h0m1 has joined #nixos-aarch64
hexa- has quit [Quit: WeeChat 2.7.1]
hexa- has joined #nixos-aarch64
hexa- has quit [Client Quit]
hexa- has joined #nixos-aarch64
<samueldr> danielrf[m]: a bit off-topic for the issues; my end-goal at some point is hopefully to use robotnix to build for a custom rom for a non-google device, maybe even LineageOS if possible
<samueldr> it's why I'm looking at this
<danielrf[m]> very cool. I think it could definitely be extended to do that
<danielrf[m]> feel free to ask any questions you might have along the way
<samueldr> it may be surprising, but my knowledge of the actual android toolchain bits is severly lacking, so I may at some point ask questions
<samueldr> sure :)
<danielrf[m]> yeah, i've spent waay too much time fighting with the android build system
<samueldr> that tooling that looks like the biggest rube goldberg machine scared/revolted me so much that I never dug into it
<danielrf[m]> seriously
<samueldr> the one time I looked seriously into it, to build a GSI for my new phone, it just didn't work as documented
<danielrf[m]> a lot of documentation is outdated. but the GSI stuff is pretty new so it's unfortunately that didn't work
<danielrf[m]> *unfortunate
<samueldr> that was documentation on that GSI project aimed at making fresh GSIs of AOSP from phhusson
<samueldr> with their specially built docker images!
<samueldr> well, the next section, but still, it didn't work out of the box
<samueldr> this only helped cement the gut feeling I had from that
<danielrf[m]> Yeah, there's a few projects using docker or scripts to make AOSP builds. I think nix is a good alternative with all the benefits of reproducibility and purity we get for free
<samueldr> I'm sure nix is a good alternative for that
<danielrf[m]> lots of these scripts do random network access to download stuff
<danielrf[m]> In fact, the new Android 10 build system uses "nsjail" to disallow network access while building
<danielrf[m]> so hopefully things will get better there
<samueldr> hopefully I won't have to deal with that ever again in the future :)
<samueldr> (by not even using android :))
<danielrf[m]> haha, I'd love to as well! hopefully mobile-nixos can be a daily driver at some point :)
<danielrf[m]> it's been on my TODO list to look at if we could use AVB in mobile-nixos btw
<danielrf[m]> I think google's actually done a really good job at boot security with their recent devices
<danielrf[m]> and they even let you use your own keys!
<samueldr> I don't know if other devices allow usage of one's own keys
<samueldr> information is scarce
<samueldr> we may want to rely on AVB up to boot.img, and from boot.img rely on something more generic that also works on other distros/x86_64-linux for verified boot setups
<samueldr> if it even makes sense
<samueldr> I assume AVB with custom keys verifies boot.img first, and then at that point it's a kernel implementation of system implementation of AVB, or something
<danielrf[m]> https://calyxos.org/ and https://grapheneos.org/ are both projects that build ROMs which let you use AVB.. I think calyxos at least is looking into other devices as well
<danielrf[m]> yep, I agree.
<samueldr> yeah, I learned the most about AVB on grapheneOS' docs
<danielrf[m]> the verified boot relies on having a read-only "system.img"--which isn't really something we do in mobile-nixos
<danielrf[m]> so verifying boot.img and using something else for the rest of the system seems the right approach for us
<danielrf[m]> sorry, I should be more precise--they use dm-verity to verify that read-only partitions haven't been modified
<samueldr> nice, calyxos has a neat nugget about that
<samueldr> >> fastboot getvar avb_user_settable_key_set
<samueldr> I guess that if this is available in fastboot, that it must mean we can set our own keys on that device
<samueldr> at a glance only my pixel 2 seems to support that
lovesegfault has quit [Ping timeout: 260 seconds]
lovesegfault has joined #nixos-aarch64
lovesegfault has quit [Client Quit]
ryantrinkle has quit [Ping timeout: 256 seconds]
FRidh has joined #nixos-aarch64
<ashkitten> samueldr: so that means we could set our own keys and relock the bootloader, in theory?
zupo has joined #nixos-aarch64
Darkmatter66_ has joined #nixos-aarch64
Darkmatter66 has quit [Ping timeout: 240 seconds]
<bennofs[m]> IIRC the main problem is that if anything goes wrong or you loose the private key, your device will be a brick
FRidh2 has joined #nixos-aarch64
FRidh has quit [Ping timeout: 265 seconds]
orivej has quit [Ping timeout: 256 seconds]
<DigitalKiwi> well don't do that
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zupo has joined #nixos-aarch64
zupo has quit [Client Quit]
zupo has joined #nixos-aarch64
Darkmatter66 has joined #nixos-aarch64
FRidh has joined #nixos-aarch64
Darkmatter66_ has quit [Ping timeout: 265 seconds]
FRidh2 has quit [Ping timeout: 260 seconds]
ryantrinkle has joined #nixos-aarch64
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
ryantrinkle has quit [Ping timeout: 256 seconds]
greizgh has quit [Quit: greizgh]
greizgh has joined #nixos-aarch64
zupo has joined #nixos-aarch64
ryantrinkle has joined #nixos-aarch64
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Thra11 has joined #nixos-aarch64
<Valodim> further data point: tested one of the pis from yesterday again, same setup same everything with the projector. still works as expected.
FRidh2 has joined #nixos-aarch64
FRidh has quit [Ping timeout: 265 seconds]
zupo has joined #nixos-aarch64
<danielrf[m]> Benno Fünfstück: Re-locking the bootloader on pixel devices is generally safe--as long as you do not disable fastboot as well
<danielrf[m]> as long as you have access to fastboot, you can unlock (even if you've lost your keys)
<danielrf[m]> the only thing is that unlocking / relocking performs a factory reset--so you'd lose your data
vika_nezrimaya has joined #nixos-aarch64
t184256 has left #nixos-aarch64 [#nixos-aarch64]
t184256 has joined #nixos-aarch64
orivej has joined #nixos-aarch64
pbb has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.]
pbb has joined #nixos-aarch64
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
asbachb has joined #nixos-aarch64
<asbachb> Just to be sure: When config this: https://nixos.wiki/wiki/Distributed_build I should be able to build my rasbperry pi stuff on an non arm system like a x86_64 server?
<clever> asbachb: under normal conditions, the remote builders must have the same cpu type
<clever> asbachb: but you can run qemu-user, to allow running arm binaries normally
<asbachb> clever: I'll check it out.
pbb has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.]
zupo has joined #nixos-aarch64
<ashkitten> i wish matrix irc puppeting translated matrix mentions to irc mentions
<ashkitten> it's sometimes hard to tell who someone is talking to
<samueldr> same, it's weird when we on the IRC side see an almost unrelated name
<clever> yeah
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<samueldr> nice, been looking to see if armv7l was building now, and it does on staging!
<samueldr> (cross-compilation)
<samueldr> staging-next and master both are failing
<ashkitten> it's also very weird when matrix users send only a few lines in one message and it shows up here as a link
<samueldr> yeah
<Valodim> it normally does translate matrix mentions to irc mentions?
zupo has joined #nixos-aarch64
asbachb has quit [Remote host closed the connection]
FRidh has joined #nixos-aarch64
FRidh2 has quit [Ping timeout: 265 seconds]
ryantrinkle has quit [Ping timeout: 265 seconds]
FRidh has quit [Client Quit]
<samueldr> took some time to scrounge up the necessary bits to make a UART cable for the nexus 7 2013 (and likely other nexus devices)
<gchristensen> nice!
WilliButz has quit [Remote host closed the connection]
WilliButz has joined #nixos-aarch64
<samueldr> oh
<samueldr> RangeError: string (FF765300) too big for integer
ryantrinkle has joined #nixos-aarch64
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zupo has joined #nixos-aarch64
pbb has joined #nixos-aarch64
zupo has quit [Ping timeout: 256 seconds]
zupo has joined #nixos-aarch64
<samueldr> hmm, not having another armv7l device to test with I'm not sure what's happening is specific to nexus 7 2013 or has wide reach
<samueldr> I get "Bad file descriptor" from udevadm, and from trying to open device files /dev/input/event1
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Thra11 has quit [Ping timeout: 258 seconds]
<samueldr> while vendor does work, it seems something is broken with the vanilla build of android with robotnix for google-walleye
<samueldr> android spews so much garbage to the console it's hard to know what's important and what's not :(
<danielrf[m]> any chance you could post the logs somewhere?
<danielrf[m]> adb logcat "*:E" will show just the errors
<danielrf[m]> but even on my device it's still a lot of stuff :)
<samueldr> sorry, I was looking in the build system a bit
<samueldr> danielrf[m]: I assume "vanilla" is supposed to boot completely, right?
<samueldr> it looks like zygote fails to start early, but I'm unsure
<danielrf[m]> yep
<samueldr> let me get a serial log
<danielrf[m]> oh, so you can't even get to settings to enable adb then?
<samueldr> should the "pixel" flavour be expected to work for walleye? it doesn't eval
<samueldr> nope, stuck at "android"
<samueldr> and it goes to rescue party
<danielrf[m]> only "vanilla" and "grapheneos" are valid flavors atm
<danielrf[m]> I should probably move pixel out of the flavors subdir
<samueldr> it was unclear to me what it was supposed to be :)
<samueldr> another thing, I'll probably open an issue, I thinkg having buildNumber/source.buildNumber default to "12345" is dangerous, it hides real issues
<samueldr> especially since it looks like it's expected that flavors should set them up
<danielrf[m]> i agree--I'll add a TODO
<danielrf[m]> having the flavors set them was a relatively new addition
<samueldr> that's kinda what I assumed
<samueldr> just ran strings on vendor.img and it does look like a walleye image (peeking just in case)
<danielrf[m]> I'll start a vanilla marlin build again just to see if I can reproduce this
<samueldr> blah, the log has control codes making it hard to gist :/
<samueldr> "signal 9" is a helpful keyword
<danielrf[m]> samueldr: yep. I can reproduce this. I think I know what the issue is. It's a hack I applied for grapheneos
<danielrf[m]> I guess I shouldn't expect anything different for not properly understanding the hack :D
<danielrf[m]> wait---no I can' reproduce this...
<danielrf[m]> haha it booted up finally
<samueldr> I'll queue a build of grapheneos once I'm done with what I'm doing, to see if that works
<samueldr> nix and nixpkgs is so nice, I can go backwards in time in the Mobile NixOS repo and in Nixpkgs, and I *know* I'm actually testing the build like it was in the past
<danielrf[m]> another thing to potentially try is to add variant="eng" which makes debugging a little easier since adb starts automatically and you get root
<samueldr> trying to see if asus-flo ever worked properly or not
vika_nezrimaya has quit [Ping timeout: 250 seconds]