ryantrinkle has joined #nixos-aarch64
wavirc22 has quit [Ping timeout: 268 seconds]
<samueldr> OH
<samueldr> just realized I can re-use a chunk of the work I did for the recovery menu / boot selection to also show the nixos logo and failure states, with the added benefit that I can use SVG, and side-step bugs
<clever> nice
ryantrinkle has quit [Ping timeout: 265 seconds]
h0m1 has quit [Ping timeout: 260 seconds]
h0m1 has joined #nixos-aarch64
mDuff has joined #nixos-aarch64
andi- has quit [Remote host closed the connection]
andi- has joined #nixos-aarch64
ryantrinkle has joined #nixos-aarch64
Acou_Bass has quit [Ping timeout: 265 seconds]
Acou_Bass has joined #nixos-aarch64
lovesegfault has quit [Quit: WeeChat 2.7]
Acou_Bass has quit [Quit: ZNC 1.7.4 - https://znc.in]
lovesegfault has joined #nixos-aarch64
lovesegfault has quit [Quit: WeeChat 2.7]
zupo has joined #nixos-aarch64
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zupo has joined #nixos-aarch64
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
t184256 has left #nixos-aarch64 ["Disconnected: Replaced by new connection"]
t184256 has joined #nixos-aarch64
zupo has joined #nixos-aarch64
orivej has joined #nixos-aarch64
t184256 has left #nixos-aarch64 [#nixos-aarch64]
zupo has quit [Ping timeout: 252 seconds]
zupo has joined #nixos-aarch64
t184256 has joined #nixos-aarch64
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zupo has joined #nixos-aarch64
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
wavirc22 has joined #nixos-aarch64
<srk> "apparently on the pi 4, the Nix folks are moving away from using the bootloader at all, in favor of directly jumping into the kernel."
<srk> with open firmware? :)
<clever> srk: currently dont have the ability to enable the arm on the 4
<srk> would be nice with any pi
* srk is fighting with DT overlays /o\
<clever> srk: for the rpi3, i can boot linux, but the userland segfaults a lot
<srk> how come?
<clever> dont know yet
<clever> i'm thinking of adding a permanent strace to some syscalls...
<srk> interesting anyway
<clever> to see if the problem is on the kernel or user side
<clever> just recently, i got jtag and gdb working, so i can debug the arm side
ryantrinkle has quit [Ping timeout: 265 seconds]
<srk> device tree tooling is really miserable
<srk> can help with debugging, stlink with blackmagic should be able to jtag
<clever> srk: i already figured out 90% of my DT issues without any jtag, just printk and earlycon
<clever> srk: but the rpi2, does have issues where jtag can help, ive got issues where linux itself hangs hard during decompression
<srk> I mean working with dts and creating overlays
<srk> there are overlays in rpi-linux but they aren't compatible with mainline
<srk> previously I've patched dtsi directly but that's not optimal
<clever> currently, i dont even support overlays
<clever> i just hand edit rpi3.dts before i make the sd image
<clever> srk: also, i dont have a proper jtag adapter either, so i'm using some pi on pi action, lol
<srk> sounds fun :) didn't know that you can use it as a jtag
<clever> srk: openocd has sysfsgpio support, so it will just bit-bang jtag over the /sys api
<clever> srk: and there is a recent PR, adding proper rpi gpio, so it could in theory set multiple pins with a single 32bit write, rather then needing a half a dozen syscalls
<srk> via char device or native linux driver?
<srk> sysfsgpio is not even enabled by default now, chardev needs libgpiod
<clever> 2020-02-02 16:27:59 < karlp> speaking of rpi, PaulFertser do you have rights to close this? http://openocd.zylin.com/#/c/2760/1
<clever> 2020-02-02 16:29:11 < karlp> http://openocd.zylin.com/#/c/4394/2 doesn't seem to have much reason not to merge either, though I've not got one to test it out.
<clever> http://openocd.zylin.com/#/c/4394 appears to be the gpio one
<clever> i see the 0x3f addr, which is the arm physical addr of all IO
<clever> and its only a config file, not any code, so its likely sharing existing code from the rpi2 support
<srk> ah, they have their own impl
<clever> the rpi4 will need new code though, because broadcom improved the pullup config
<clever> cant find it in the source though
<clever> found it
<clever> srk: yeah, it directly mmap's /dev/mem, which wont work on recent kernels due to security options, /dev/gpio-mem is a back-door to get around that
<srk> cool pic
<srk> dtc needs DTC_FLAGS='-@' to generate symbols so overlays can be applied to symbols like <&gpio>
<clever> i still need to figure out how overlays actually work
<srk> this is legit overview except for rpi specific stuff https://www.raspberrypi.org/documentation/configuration/device-tree.md
<clever> i know how rpi uses it, just not how to apply overlays myself or write them
<srk> it sucks to write em, you can apply them via hardware.deviceTree.overlays but it needs some tweaks
<srk> and that DTC argument so dtb files have symbols
<clever> srk: i mean how to apply at the c level, at runtime
<clever> srk: this code will open a dtb in ram, and sets a few fields
<srk> ah, u-boot provides fdt apply command
<srk> think it's based on libfdt
<srk> uses libfdt
wavirc22 has quit [Ping timeout: 268 seconds]
ryantrinkle has joined #nixos-aarch64
<clever> srk: just using u-boot is in my plans
mDuff has quit [Ping timeout: 265 seconds]
zupo has joined #nixos-aarch64
orivej has quit [Ping timeout: 265 seconds]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zupo has joined #nixos-aarch64
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<samueldr> <srk> "apparently on the pi 4, the Nix folks are moving away from using the bootloader at all, in favor of directly jumping into the kernel."
<samueldr> uh?
<clever> samueldr: that sounds like a description of how the official firmware boots
<clever> but i think a u-boot would still be better, some kind of menu
<samueldr> a deleted comment from a stackexchange question
<samueldr> https://webcache.googleusercontent.com/search?q=cache:9zxujlZoACYJ:https://raspberrypi.stackexchange.com/questions/107816/applying-dwc2-dtbo-to-bcm2711-rpi-4-b-dtb-at-build-time-triggering-fdt-err-notfo+&cd=1&hl=en&ct=clnk&gl=ca&client=firefox-b-d
<clever> ah
<samueldr> there may be confusion coming from your project, and from our stance in preferring mainline and generic stuff
<clever> not sure how exactly they got that idea though, i cant even get the arm cpu working on the rpi4 currently
<srk> almost solved the overlays \o/ http://ix.io/2aIQ
<samueldr> Internal Server Error \o/
<srk> bit more magic and it's ready for PR
<srk> haha
<srk> these paste services
<srk> samueldr: should I create a PR for resize2fs -M for make-ext4-fs? looks like it currently fails on armv7/aarch64 with current size calculation
<samueldr> it fails how and how does this fix it?
<samueldr> (and yes, probably)
<srk> it's a long story but even with higher fudge factors it won't allow resizing to calculated size
<{^_^}> #62262 (by yegortimoshenko, 35 weeks ago, closed): lib/make-ext4: bump fudge factor to 96 MiB
<samueldr> the actual solution would be to work with the new harness I've been writing
<samueldr> it uses make_ext4fs which is a tool made to make ext4 filesystems
<srk> is it worth it to do it manually?
<samueldr> it's made and used by the android project to make their reproducible filesystem images
<srk> since results are compressed anyway (in most cases)
<srk> ah, ok
<samueldr> though, now that we compress the images, it's less of an issue if we have a bigger image initially
<samueldr> this is what I figured out for ext4
<srk> interesting
<samueldr> for bigger than 256MiB images, the filesystem overhead is stable according to the size
<samueldr> but due to how the structures are made, for smaller sizes it will break in weird ways
<samueldr> ext4 has weird bumps in its graph because of internal structures that makes it hard for smaller sizes to mathematically derive
<samueldr> FAT32's has been verified with the mkfs.vfat code
zupo has joined #nixos-aarch64
<srk> today I was dreaming about Haskell stage-1/2 scripts
<srk> could be quite nice with turtle but hard to build
<samueldr> :)
<samueldr> clever has an haskell init
<srk> cause there's scary stuff like local optionsFiltered="$(IFS=,; for i in $options; do if [ "${i:0:2}" != "x-" ]; then echo -n $i,; fi; done)"
<srk> cool!
<samueldr> the reason I went with mruby is that ruby is a language I'm already well acquainted with
<clever> srk: it even includes a very basic repl, to let you run commands
<clever> mostly just a placeholder, to show that stdio works
<samueldr> but more importantly: I'm able to cross-compile it and statically
<samueldr> I looked at rust, but currently it doesn't cross-compile
<clever> srk: i think the biggest blocker from a 100% haskell initrd, is kernel module loading, modprobe does a lot of fancy stuff, and i wouldnt want to re-implement that in haskell
<samueldr> a future goal will be to use all the knowledge learned from that to rewrite in rust
<srk> clever: fun!
<srk> not sure if you can statically crosscompile haskell heh
<clever> srk: but you can ffi into mount (already done), pivot_root, and excve earily enough
<srk> statically sure but cross
orivej has joined #nixos-aarch64
Ox4A6F has quit [*.net *.split]
goibhniu has quit [*.net *.split]
Ericson2314 has quit [*.net *.split]
Ericson2314 has joined #nixos-aarch64
goibhniu has joined #nixos-aarch64
Ox4A6F has joined #nixos-aarch64
orivej has quit [Ping timeout: 265 seconds]
Ox4A6F has quit [*.net *.split]
goibhniu has quit [*.net *.split]
Ericson2314 has quit [*.net *.split]
Ericson2314 has joined #nixos-aarch64
goibhniu has joined #nixos-aarch64
Ox4A6F has joined #nixos-aarch64
cidkid has joined #nixos-aarch64
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
zarel has quit [Quit: ZNC 1.7.4 - https://znc.in]
zarel has joined #nixos-aarch64
<gchristensen> ,stty
<gchristensen> ,stty = echo "stty rows $(tput lines) cols $(tput cols)"
<{^_^}> stty defined
<cidkid> Is there anywhere that has more indepth mobile-nixos documentation
Acou_Bass has joined #nixos-aarch64
cidkid has quit [Remote host closed the connection]
cidkid has joined #nixos-aarch64
<cidkid> mm with nixOS mobile can I flash the rootfs to the system partition intstead of userdata?
<cidkid> as ssh-initrd fails to ssh properly
<cidkid> more of I get connection refused
mDuff has joined #nixos-aarch64