bennofs_ has joined #nixos-aarch64
bennofs__ has quit [Ping timeout: 246 seconds]
superherointj has joined #nixos-aarch64
chr0ma[m]1 has joined #nixos-aarch64
<matthewcroughan> samueldr: Can't find default configuration "arch/../configs/orangepi_3_defconfig"!
<matthewcroughan> hmm..
<matthewcroughan> wonder if they called it something different
zupo_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<samueldr> CommitDate: Mon Jan 25 21:52:00 2021 +0000
<samueldr> it's not in 2020.01
<samueldr> uh
<samueldr> 2021.01
<matthewcroughan> oh, I see
<matthewcroughan> so it'll be in 03?
<matthewcroughan> how does nixpkgs update the u-boot source repo?
<samueldr> in whatever next release of u-boot will be, since it's on master right now
<samueldr> nixpkgs updates u-boot at every release
<matthewcroughan> it's in v2021.04-rc1
<matthewcroughan> neat
<matthewcroughan> I can still put the PR in and I guess wait until that's stable :D
superherointj has quit [Quit: Leaving]
<matthewcroughan> samueldr: I get this: error: cannot update ref 'refs/heads/v2021.04-rc1': trying to write non-commit object 183e2c01b3280e5ca899cb0f673d788b4d79d807 to branch 'refs/heads/v2021.04-rc1'
<matthewcroughan> any ideas
<matthewcroughan> using fetchGit
<matthewcroughan> I can git clone and checkout u-boot like that and it works fine
rajivr has joined #nixos-aarch64
h0m1 has quit [Quit: WeeChat 3.0]
h0m1 has joined #nixos-aarch64
cole-h has quit [Ping timeout: 258 seconds]
<matthewcroughan> samueldr: nope, still fails
<matthewcroughan> `U-Boot SPL 2021.04-rc1 (Jan 01 1980 - 00:00:00 +0000)`
<matthewcroughan> Gets stuck at "Trying to boot from MMC1"
<matthewcroughan> samueldr: do you know how I could build uboot with nix-build with debug on?
<samueldr> no
<samueldr> I don't even know how to build it with debug on otherwise
<samueldr> I'd have to read the docs
<matthewcroughan> there's extraConfig
<matthewcroughan> what would that do?
<matthewcroughan> `cat $extraConfigPath >> .config`
<matthewcroughan> L*)
<matthewcroughan> L80 *
<samueldr> well
<samueldr> that's what that does
<samueldr> you can add other configuration settings to the build
<matthewcroughan> so it literally puts the path's file's contents into .config for the build~?
<samueldr> in addition to those found in defconfig
<samueldr> passAsFile makes it so an attribute in a derivation is given as a file
<samueldr> so it'll be whatever you set as extraConfig
<matthewcroughan> samueldr: How would I pass that to nix-build that I'm doing?
<samueldr> I don't know
<matthewcroughan> Currerntly I have `nix-build -I ./ -j 24 -v -A pkgsCross.aarch64-multiplatform.ubootOrangePi3`
<matthewcroughan> so what would I need?
<samueldr> check where the call site of the function is, and add it there I guess
<matthewcroughan> Do I now have to do soem sort of override or have nix-build evaluate a nix expr?
<matthewcroughan> god, my typing is all over the place tonight :D
<samueldr> it's okay just to edit the file
<samueldr> you don't need to override it if you're hacking stuff up
<matthewcroughan> yeah but I don't really understand the syntax in the nix file ha
<samueldr> see if other boards use extraConfig
<samueldr> get inspiration from there I guess
<matthewcroughan> passAsFile = [ ./crap ]
<matthewcroughan> ?
<samueldr> what if you search extraConfig in the uboot/default.nix file?
<matthewcroughan> OH!
<matthewcroughan> I see, it's just a multiline config in the board definition
<matthewcroughan> ubootRock64 uses it
<samueldr> it's lines to add to the .config file
<samueldr> .config being similar to the kernel config setup
<samueldr> not sure if it is exactly it, a fork, or inspired from
<matthewcroughan> All I got to go on is this hah
<matthewcroughan> GDI
<matthewcroughan> why doesn't it work hahaha
<matthewcroughan> it's upstream nwo and it still doesn't work!
<matthewcroughan> something has to be wrong with the nixos sd image, right samueldr?
<matthewcroughan> samueldr:
<matthewcroughan> do you think it's possible our cross compilation is causing the issue?
h0m1 has quit [Ping timeout: 258 seconds]
h0m1 has joined #nixos-aarch64
<matthewcroughan> !!!!!!!!!!1
<samueldr> so we need an armTrustedFirmwareAllwinnerA64 and armTrustedFirmwareAllwinnerH6?
<matthewcroughan> Yup!
<samueldr> well, that's what I understand
<matthewcroughan> that's my immediate thought too
<matthewcroughan> I'm soo close
<samueldr> makes sense since the SPL jumps into ATF
<samueldr> the unqualified ATFAllwinner should be put in aliases point to the A64 variant
<samueldr> or maybe these needs to be re-thought
<samueldr> well... not or
<samueldr> later
<matthewcroughan> yes, so for now we can create an `armTrustedFirmwareAllwinnerH6`
<matthewcroughan> in order to save my sanity
<samueldr> please alsomake a PR to rename+alias AllwinnerA64 to help document it properly
<samueldr> but yes, do try first doing the least effort
<samueldr> we don't _know_ yet that it works
<matthewcroughan> yeah, just going to hack on the local nixpkgs
<matthewcroughan> How reading boring stuff all night has made me happy :)
<matthewcroughan> I am so hopeful
<matthewcroughan> I think my heart might explode if this works
<samueldr> keep in mind that there is no guarantees that the kernel will
<samueldr> but if you get to the next steps in u-boot, at least you can start worrying about the kernel
<matthewcroughan> Yeah but now that I've got a taste for what it means to bring up board support, I kinda dig it.
<matthewcroughan> I've applied for access to the BeagleV boards and am likely to get a response because I applied with company email, so would probably dig that too
<samueldr> as you might have learned, at some point it helps to peruse the fine manual
<matthewcroughan> you know those riscv64 boards?
<samueldr> totally other kind of ball game there though
<matthewcroughan> samueldr: yes! But finding that manual only came as a result of me reading the one warning msg from the make log
<matthewcroughan> `Image 'main-section' is missing external blobs and is non-functional: scp`
<matthewcroughan> `Please read the section on SCP firmware in board/sunxi/README.sunxi64`
<samueldr> I don't know what our risc-v setup looks like
<matthewcroughan> if that warning didn't show, I'd not have known
<matthewcroughan> samueldr: non existent, which is why it's so exciting
<samueldr> and it will differ enough, the path is not going to be well-trodden
<matthewcroughan> an opportunity to learn with everyone else
<matthewcroughan> very rare
<matthewcroughan> usually everything has raced ahead away from you
<matthewcroughan> samueldr: IT ABSOLUTELY WORKED
<matthewcroughan> The kernel is very hapy
<matthewcroughan> happy*
<matthewcroughan> samueldr: welp, everything fking works
<samueldr> please keep the swearing for the home :) even when it's elation
<samueldr> good to hear though
<samueldr> one way I'd like to redesign the trusted firmware derivation may be to either make it an attrset or a function, both serving to ask "configuration xx_yy of ATF please"
<samueldr> so `buildArmTrustedFirmware "sun50i_a64"` or `buildArmTrustedFirmware.sun50i_a64`
<samueldr> uh
<samueldr> or `armTrustedFirmware.sun50i_a64`
<samueldr> the idea being to use exactly the config names
<samueldr> hm, not sure it's worth it
<samueldr> might just want to rename the existing attributes, it's already setup like u-boot is
<samueldr> so instead of S905, Gxbb... instead of Allwinner Sun50iA64 and Sun50iH6
<samueldr> that would more closely mirror the u-boot setup, where the attrnames come from the defconfig built
<matthewcroughan> samueldr: yeah, do you mean like how user.user.<name> currently works?
<samueldr> no
<samueldr> forget about it, since changing it wouldn't really be worth it
<samueldr> but just an attrset
<matthewcroughan> I'll make a PR and comment that these are concerns inside of it
<matthewcroughan> and make a wiki page tonight on how to get the Orange Pi 3 running
<matthewcroughan> what a fun night :D
<matthewcroughan> Very happy with that find, despite me not knowing much of anything about bootloaders, the manual had some info that presented itself to my eyeballs.
<samueldr> knowing about bootloaders wouldn't have helped
<matthewcroughan> DEBUG=1 didn't really help very much though
<samueldr> that's the "firmware" part of u-boot!
<matthewcroughan> literally did nothing
<samueldr> u-boot acts as if your bios included grub
<samueldr> (in an oversimplified way)
<matthewcroughan> so arm boards do not typically have any BIOS?
<samueldr> on most ARM platforms supported by u-boot, there is no previous bios, the cpu only knows how to load something else that is the bios
<samueldr> well, the firmware, not "bios"
<samueldr> most of those that u-boot supports don't
<samueldr> but it's not an "arm thing"
<samueldr> there are ARM platforms which ship with an UEFI firmware
<samueldr> there are ARM platforms with bespoke initial fimrwares
<samueldr> but it's convenient how many SoCs push that job to a further program
<samueldr> since we have u-boot that is a really great implementation that helps things work allike on many boards
tilpner_ has joined #nixos-aarch64
tilpner has quit [Ping timeout: 272 seconds]
tilpner_ is now known as tilpner
<matthewcroughan> samueldr: my language is lacking
<matthewcroughan> what is `armTrustedFirmwareAllwinner`?
<matthewcroughan> Is that a function?
<samueldr> an attrname in the package set, pointing to a derivation
<samueldr> > pkgs.armTrustedFirmwareAllwinner
<{^_^}> "<derivation /nix/store/nqa6v114phxg0inr87n68my8mckz639h-arm-trusted-firmware-sun50i_a64-2.4.drv>"
<samueldr> > builtins.typeOf pkgs.armTrustedFirmwareAllwinner
<{^_^}> "set"
<samueldr> uh
<matthewcroughan> so its an attrset?samueldr
<samueldr> > lib.isDerivation pkgs.armTrustedFirmwareAllwinner
<matthewcroughan> I'm really thumb mashing my keys tonight
<{^_^}> true
<samueldr> > builtins.typeOf pkgs
<{^_^}> "set"
<samueldr> so the name "armTrustedFirmwareAllwinner" is the attribute name in the pkgs attrset
<samueldr> which points to a derivation
<matthewcroughan> samueldr: whatdya think?
<matthewcroughan> git commit -m "U-Boot: Adds Orange Pi 3 build" -m "This commit also adds a new attrname (attribute name) called armTrustedFirmwareAllwinnerH6 that is necessary to get ATF (Arm Trusted Firmware) working. This attrname sets platform = "sun50i_h6" unlike armTrustedFirmwareAllwinner which ignores the following section of the U-Boot README from Sunxi and assumes that "sun50i_a64" is the only value required for Sunxi
<matthewcroughan> reading the contributing.md now actually, I need to change some stuff
orivej has joined #nixos-aarch64
orivej has quit [Ping timeout: 240 seconds]
cole-h has joined #nixos-aarch64
rajivr has quit [Quit: Connection closed for inactivity]
zupo has joined #nixos-aarch64
cole-h has quit [Ping timeout: 240 seconds]
Darkmatter66 has joined #nixos-aarch64
tilpner has quit [Remote host closed the connection]
tilpner has joined #nixos-aarch64
rajivr has joined #nixos-aarch64
srk has quit [Remote host closed the connection]
srk has joined #nixos-aarch64
Darkmatter66 has quit [Read error: Connection reset by peer]
Darkmatter66 has joined #nixos-aarch64
Darkmatter66 has quit [Read error: Connection reset by peer]
Darkmatter66 has joined #nixos-aarch64
drag0nius has joined #nixos-aarch64
drag0nius has left #nixos-aarch64 [#nixos-aarch64]
Darkmatter66 has quit [Read error: Connection reset by peer]
Darkmatter66 has joined #nixos-aarch64
quinn has quit [Quit: ZNC 1.8.1 - https://znc.in]
Darkmatter66 has quit [Ping timeout: 246 seconds]
Darkmatter66 has joined #nixos-aarch64
Darkmatter66 has quit [Ping timeout: 240 seconds]
orivej has joined #nixos-aarch64
<s1341> samueldr you availabe to answer some cross-compilation questions?
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<s1341> how do i use a host-package when cross compiling?
zupo has joined #nixos-aarch64
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
dev_mohe has joined #nixos-aarch64
dev_mohe has quit [Client Quit]
orivej has quit [Ping timeout: 272 seconds]
<matthewcroughan> samueldr: damnit, wifi doesn't work :D
<matthewcroughan> samueldr: platform regulatory.0: Direct firmware load for regulatory.db failed with error -2
<veleiro> starting up the pinebook pro again, is there newer cache for linux/firefox?
<matthewcroughan> hmm.. how does nixos do firmware
<matthewcroughan> [ 15.785253] brcmfmac: brcmf_sdio_htclk: HT Avail timeout (1000000): clkctl 0x50
orivej has joined #nixos-aarch64
orivej has quit [Ping timeout: 256 seconds]
orivej has joined #nixos-aarch64
<matthewcroughan> samueldr: [ 12.890339] Bluetooth: hci0: BCM: firmware Patch file not found, tried:
<matthewcroughan> [ 12.902365] Bluetooth: hci0: BCM: 'brcm/BCM.hcd'
<matthewcroughan> [ 12.897126] Bluetooth: hci0: BCM: 'brcm/BCM4345C5.hcd'
<matthewcroughan> samueldr: kernel 5.9 has a better result. In that, USB3.0 works, it does not on 5.4
rajivr has quit [Quit: Connection closed for inactivity]
zupo has joined #nixos-aarch64
zupo has quit [Client Quit]
<veleiro> i've noticed more cache being pulled from the store for aarch64! hydra builds must be better?
<veleiro> i dont have to build nix every time it appears
<Ke> how is this wrong: kernelPackages = pkgs.linuxPackagesFor (pkgs.linux_5_10.override {
<Ke> CONFIG_ROCKCHIP_CDN_DP = pkgs.lib.kernel.no;
<Ke> structuredExtraConfig = {
<Ke> and remaining things to close the syntax
<Ke> the problem is that such config is just not set
<Ke> or unset in this case
zupo has joined #nixos-aarch64
cole-h has joined #nixos-aarch64
Darkmatter66 has joined #nixos-aarch64
<samueldr> Ke: on my end CONFIG_ is outside structedExtraConfig
<samueldr> but let's assume they are in the wright order
<samueldr> I believe CONFIG_ must be dropped from the option names in nixos modules
orivej has quit [Ping timeout: 240 seconds]
<Ke> let me check
<Ke> I do remember something similar indeed
<samueldr> ugh, wright order -> written in the right order... samueldr wright better
<samueldr> or uh, wite right please
<Ke> thanks, it was as you said
<samueldr> veleiro: I stick to my own cheaty methods to get a system with coverage
<samueldr> this rootfs brings in firefox (and nix)
<samueldr> so following its successes rather than the tip of unstable I know firefox will be there
<samueldr> not ideal, but for the time being it's a good workaround
<samueldr> Ke: that fact annoys me, because to me the option is the full CONFIG_*...
<samueldr> ... but! CONFIG_* is the exported name of the config in the kernel tree, the actual definitions are without CONFIG_
<samueldr> so I guess it's more right even if it doesn't feel like it :/
<Ke> my stance is always that language is as it's used, though if language is mostly used with false assumptions, I would not go to incoherently define things
<samueldr> maybe my assumption (can't say about yours) is that everyone sees `CONFIG_*
<samueldr> ` as the full name
<samueldr> but that may be wrong :)
<Ke> this is like bash removing extra newlines in certain places for convenience, though again it's very convenient
<Ke> I agree that I did actually expect CONFIG_*
<samueldr> totally understandable
<samueldr> I wonder if linux has any CONFIG_CONFIG_*
<samueldr> almost CONFIG_CONFIGFS_FS=m
<samueldr> but that's really not an issue
<samueldr> we should be able to warn or error on CONFIG_ presence
<samueldr> until such an option is added
<sphalerite> Ugh, I didn't realise way back when that I effectively forfeit my consumer rights when buying from the US.
<sphalerite> Kind of pissed that I "cancelled" my librem5 order back in April last year and still haven't got my money back
<samueldr> yeeeowch
<samueldr> >> Per our policy, we can not provide refund for products that are in the pre-order status.
<samueldr> updated march 5th last year I think
<samueldr> not that it makes it right
<samueldr> but that'll make it more annoying
<samueldr> hm, no entries in the wayback machine
<sphalerite> Yeah I'm pretty sure it's not legal either, I just have no recourse since I'm not a US citizen/don't live in the US
<sphalerite> Ah well. At least I can spread some bad PR.
* sphalerite tweets
tilpner_ has joined #nixos-aarch64
tilpner has quit [Ping timeout: 265 seconds]
tilpner_ is now known as tilpner
<Ke> I did get a refund
<Ke> not sure when I cancelled
<Ke> I have kind of mixed feelings about puri.sm in general
<Ke> I guess I probably canceled earlier
<Ke> by that time I think it was more than a year late
<Ke> if they did ship in order, I could have received the phone probably by now
<Ke> I think I was in first 200 orders
<Ke> anyway, I can stil order it, if they can deliver
<Ke> meanwhile I still have already 2 pinephones
<Ke> I would kind of get nixos running on it, but I absolutely need the phone stuff there
zupo has quit [Ping timeout: 258 seconds]
zupo has joined #nixos-aarch64
ashkitten has joined #nixos-aarch64
nicoo has joined #nixos-aarch64
nicoo has quit [Changing host]
<veleiro> purism has been a deceptive company since day one
<veleiro> they promised things that they couldnt possibly fulfill to get more money
<veleiro> ignorance with free software and technology in general too
<hexa-> yes, but
<hexa-> I don't have the big picture, because I didn't order one, but there have been some achievements that purism developers reached
<hexa-> like libhandy
<veleiro> i feel phone/touch-related design is a bad direction in general, we shouldnt try to be like Google or Apple but whatever, not really into libhandy
<veleiro> they want to be Apple that's for sure
<samueldr> I'd like to try hacking up libinput a bit some day to implement a new input scheme
<samueldr> rather than using "absolute direct input" with the touchscreen
<samueldr> mais the input layer aware of "direct input zones", where you'd e.g. have the virtual keyboard
<samueldr> but everything else acts as a touchpad
<samueldr> so you can move the cursor by swiping
<samueldr> and tap to click
<veleiro> yep or just stop making touch-centric devices and bring back pdas
<samueldr> I think it would be a net gain in accessibilty
<samueldr> because even a fully abled person, in a busy bus on a bad road will mis-touch things
<samueldr> with a touchpad-like input it gets much harder to mis-touch things
<samueldr> but it needs some hacking to allow zones to be input directly, otherwise keyboard input will be hard
<samueldr> (yes, actual physical keyboard would be easier, but keep in mind my main goal will always be to work with what you already have)
<samueldr> using what you have for longer will have a stronger environmental impact than switching, hopefully
<veleiro> yeah getting away from consumer-driven design like destroy to open and repair ability in new designs as well
<veleiro> the psion series 5 type clamshell keyboard for the pinephone sounds awesome
<veleiro> Latch it to me arm!
quinn has joined #nixos-aarch64
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]