<apache8080>
does buildLinux support cross compiling? I tried hardcoding kernelArch but that didn't work
bridge[evilred] has quit [Remote host closed the connection]
<Venom668>
Oh my god.
<Venom668>
u-boot on SD card, OS on eMMC made suspend to ram work
<Venom668>
probably u-boot on SPI would work too
<Venom668>
while having both u-boot and os on emmc doesn't work
<Venom668>
Thank you for your work, samueldr
<Venom668>
It's really helpful <3
vikanezrimaya has left #nixos-aarch64 [#nixos-aarch64]
apache8080 has quit [Ping timeout: 246 seconds]
<samueldr>
Venom668: I would hazard a guess that u-boot's location changes nothing, and that it must have been OS on SD card that was at fault
<samueldr>
though... odd!
<samueldr>
if you are able to test eMMC holding both u-boot and OS, that would be helpful, if you're not, then no probme
<samueldr>
no problem*
<Venom668>
I had u-boot and os on emmc for past few days while trying to get s2r working
<Venom668>
meanwhile juggling kernels and uboots
<samueldr>
it seems odd that the location u-boot loaded from would change anything here with what I know
<samueldr>
but I don't know anything that disproves it either
<clever>
samueldr: solved one of the rpi bootloader bugs, linux is now able to print "Uncompressing Linux... done, booting the kernel." but then it goes silent after that
<Venom668>
yes, very weird
<samueldr>
clever: your open bootloader?
<clever>
yeah
<samueldr>
firmware* :)
<clever>
both
<samueldr>
yeah
<clever>
its currently a custom firmware, with a custom bootloader
<clever>
whats the smallest youve gotten tow-boot down to?
<samueldr>
haven't tried to do anything like minimizing
<samueldr>
my focus has been making it the best user experience first
<clever>
i dont really have any size limits right now, so that works out fine
<samueldr>
minimizing the size would probably require removing UX features... and in that case it's more of a U-Boot project :)
<clever>
-rwxr-xr-x 1 clever users 619K May 17 03:11 u-boot.bin
<clever>
thats probably an ok size
<clever>
in theory, if we just add a custom board type (variant of rpi), it could boot on a pi2
h0m1 has quit [Ping timeout: 258 seconds]
<clever>
but the firmware provided DT will be missing
<clever>
so it will be more like a traditional uboot
<samueldr>
which is fine I guess
<clever>
going to try and fix linux first, then we could look into that
<samueldr>
there's never going to be a "right" time to say it, but officially speaking, the place for NixOS on ARM discussion is now https://matrix.to/#/!RjBlCIbsLDzHBIzmaA:nixos.org?via=nixos.org&via=matrix.org&via=nordgedanken.dev
<samueldr>
welp, I don't like that "via" crap at the end
<Mindavi>
They're typically 'foundational' because I'm trying to upstream the work to get sway cross-compiling
<samueldr>
yeah, those are "in-between"
<samueldr>
more into ecosystem than foundational (e.g. stdenv, compilers)
<Mindavi>
Which I did succeed in, but there's some parts that I can't solve without shelling out to qemu
<Mindavi>
Mm, then I'll try to find the right persons for them
Ox4A6F1 has quit [Quit: authenticating]
Ox4A6F1 has joined #nixos-aarch64
superherointj has quit [Remote host closed the connection]
cirno-999 has quit [Quit: WeeChat 2.9]
<clever>
elvishjerricco: was looking over the eeprom commit history, 2021-03-22 added support for [cm4] in the config files, so if versions are mixed around, different things can apply at diff times in the boot
Mindavi has quit [Quit: leaving]
justanotheruser has quit [Ping timeout: 245 seconds]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
superherointj has quit [Quit: Leaving]
apache8080 has joined #nixos-aarch64
<elvishjerricco>
clever: Does that matter if I'm using a config file with no [<pi model>] stuff?
mica[m] has left #nixos-aarch64 ["User left"]
<clever>
elvishjerricco: thats why i was wanting to test with all of those conditions removed
<elvishjerricco>
clever: Yea, I've had those conditions removed in all my tests this week
<elvishjerricco>
I gave up on netboot; it seems like start4.elf has just always been fundamentally flawed and there's nothing I can do about it short of disassembling it and hoping it doesn't need a major rewrite of the netboot code.
<clever>
elvishjerricco: try rpiboot next then
<elvishjerricco>
So back to sd card shuffling and trying to work out why u-boot can't use the usb port
<elvishjerricco>
hm?
<elvishjerricco>
oh
<elvishjerricco>
that's interesting
<clever>
if you have a 3 in the BOOT_ORDER, it can pull files from the rpiboot program
<clever>
over the usb cable
<clever>
but the CM4 will be in gadget mode, so you cant really test the host mode features
<clever>
but you can at least build and get logs
<elvishjerricco>
well the logs I need have to do with using the usb ports in host mode, no?
<clever>
it may still give useful logs
<clever>
also, is it an emmc model?
<elvishjerricco>
no
<clever>
your only option left then is to either open a ticket to complain, or keep swapping cards
<elvishjerricco>
swapping cards isn't so bad
<elvishjerricco>
at least I can do it all from my desk with uart
<elvishjerricco>
before, I had to cross the room and use an annoying tv as a monitor :P
<clever>
you got a button on the RUN header?
<elvishjerricco>
I assume not because I don't know what that is :P
<clever>
there is a header with 3 pins, RUN, GND, and GLOBAL_EN
<clever>
if you take a reset/power button from an old x86 case, and put it between RUN and GND
<clever>
then you can hard reset the pi, by just whacking the button
<clever>
i hit that after putting the card in, after a swap