superherointj has quit [Quit: Leaving]
<hexa-> rpi4 uboot+mainline (5.11.11) works
<samueldr> hexa-: ooh, so we can remove the pi4 specific image!
<samueldr> finally!
<hexa-> after #112677
<{^_^}> https://github.com/NixOS/nixpkgs/pull/112677 (by colemickens, 8 weeks ago, open): nixos rpi bootloader: install files for raspberry pi 4 (rpi4)
<samueldr> yeah
<hexa-> the serial is a bit messy
<colemickens> that PR is still not about SD image
<hexa-> and I'm not clear on the kmods that need to get loaded
<colemickens> sorry to butt in again
<hexa-> sure, but that is a change we can do on top, right?
<samueldr> still useful for end-users that want to stop using u-boot
* colemickens nods
<hexa-> start using u-boot*
<hexa-> finally usable generations
<samueldr> and colemickens: don't be sorry, you authored it :)
<hexa-> indeed
<hexa-> colemickens++
<{^_^}> colemickens's karma got increased to 60
justanotheruser has quit [Ping timeout: 250 seconds]
<colemickens> oh I guess people want me to do stuff on it too
<hexa-> yep, please :p
<hexa-> maybe in time for 21.05 :p
ryantrinkle has quit [Ping timeout: 252 seconds]
justanotheruser has joined #nixos-aarch64
quinn has quit [Ping timeout: 240 seconds]
quinn has joined #nixos-aarch64
<samueldr> wow
<samueldr> every time I dig out an SBC I find an SD card I had "lost"
<hexa-> colemickens: does dtoverlay=disable-wifi and bluetooth actually work for you on mainline?
<colemickens> no
<hexa-> :)
<samueldr> hexa-: config.txt options for device tree (including dtoverlay) will probably get undone by u-boot
<samueldr> well
<samueldr> undone when loading the dtb file for the generation
<hexa-> what a great time to be alive
<hexa-> (╯°□°)╯︵┻━┻
<samueldr> it doesn't know it needs to do things according to the way the raspberry pi bootloader want
<samueldr> wants*
<hexa-> totally understandable
<hexa-> i fault the rpi for alot of things
<hexa-> let's add that
<samueldr> *additionally*, the device trees are not really compatible
<samueldr> the main issue being the labels
<samueldr> they don't match in the foundation kernels compared to mainline
<samueldr> so even if the overlay was applied, it might not be able to apply it against the right nodes
<samueldr> really the fact the foundation still uses their custom bootloader scheme is purely and simply vendor lock-in
<hexa-> right, makes perfect sense
<hexa-> samueldr++
<{^_^}> samueldr's karma got increased to 332
ryantrinkle has joined #nixos-aarch64
<samueldr> hexa-: do you still have everything around to test u-boot?
<hexa-> samueldr: test u-boot? it's up and running
<samueldr> to test 2021.04
<hexa-> sure, can do
<samueldr> no PR, just replace the version and hash
<hexa-> omw
<samueldr> on my end it's... not going well in U-Boot, but might be my own fault
<hexa-> :)
<hexa-> backing up /boot :)
<hexa-> #nixos-riscv when? https://beagleboard.org/beaglev
<samueldr> okay, found the issue
quinn has quit [Ping timeout: 240 seconds]
<samueldr> hexa-: I assume it will work just fine, but please continue if you were testing
<samueldr> (I'm not testing the Nixpkgs U-Boot build)
quinn has joined #nixos-aarch64
quinn has quit [Client Quit]
<samueldr> I need a better "burn" script
<samueldr> that shows a big red window when I do a dumb
<hexa-> samueldr: looked away for a moment and didn't notice it rebooted already
<hexa-> "Doesn't look like anything to me"
<samueldr> hm?
<hexa-> looks the same as before, just with a new version header
<samueldr> yeah, and it should
<samueldr> anyway I found the thing that caused me issues
quinn has joined #nixos-aarch64
quinn has quit [Client Quit]
quinn has joined #nixos-aarch64
quinn has quit [Client Quit]
h0m1 has quit [Ping timeout: 258 seconds]
h0m1 has joined #nixos-aarch64
quinn has joined #nixos-aarch64
quinn has quit [Ping timeout: 265 seconds]
quinn has joined #nixos-aarch64
hexa[m] has quit [*.net *.split]
AlesHuzik[m] has quit [*.net *.split]
zuh0 has quit [*.net *.split]
gh0st[m] has quit [*.net *.split]
hiroshi[m] has quit [*.net *.split]
angerman has quit [*.net *.split]
hexa- has quit [*.net *.split]
hexa[m] has joined #nixos-aarch64
angerman has joined #nixos-aarch64
quinn has quit [Quit: ZNC 1.8.1 - https://znc.in]
hexa- has joined #nixos-aarch64
gh0st[m] has joined #nixos-aarch64
AlesHuzik[m] has joined #nixos-aarch64
hiroshi[m] has joined #nixos-aarch64
zuh0 has joined #nixos-aarch64
quinn has joined #nixos-aarch64
orivej has joined #nixos-aarch64
alpernebbi has joined #nixos-aarch64
justanotheruser has quit [Ping timeout: 250 seconds]
cole-h has quit [Ping timeout: 252 seconds]
srk has joined #nixos-aarch64
quinn has quit [Ping timeout: 240 seconds]
quinn has joined #nixos-aarch64
h0m1 has quit [Quit: WeeChat 3.1]
h0m1 has joined #nixos-aarch64
h0m1 has quit [Client Quit]
h0m1 has joined #nixos-aarch64
h0m1 has quit [Quit: WeeChat 3.1]
h0m1 has joined #nixos-aarch64
Fuseteam has quit [Quit: Idle for 30+ days]
zupo has joined #nixos-aarch64
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
quinn has quit [Quit: ZNC 1.8.2 - https://znc.in]
quinn has joined #nixos-aarch64
Acou_Bass has quit [Ping timeout: 260 seconds]
Acou_Bass has joined #nixos-aarch64
zupo has joined #nixos-aarch64
<anton[m]> folks, I am stuck with my PR: https://github.com/NixOS/nixpkgs/pull/101454
<{^_^}> #101454 (by arapov, 24 weeks ago, open): uboot: init (firmwareOdroidC2/C4, ubootOdroidC4) Hardkernel's Odroid C4 and Odroid C2 board support, drop Armbian dependency
<anton[m]> Is there anyone able to help with review and merge?
dev_mohe has joined #nixos-aarch64
dev_mohe has quit [Client Quit]
zupo_ has joined #nixos-aarch64
zupo has quit [Ping timeout: 260 seconds]
luxemboye has quit [Remote host closed the connection]
luxemboye has joined #nixos-aarch64
h0m1 has quit [Quit: WeeChat 3.1]
h0m1 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…]
zupo has joined #nixos-aarch64
superherointj has joined #nixos-aarch64
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
luxemboye has quit [Remote host closed the connection]
luxemboye has joined #nixos-aarch64
justanotheruser has joined #nixos-aarch64
<samueldr> I need to actually take time to look into it... but I'm not too thrilled with the prospect of relying on a non-mainline U-Boot after we were able to remove all non-mainline U-Boot uses recently
<samueldr> hm, though I see it's for "firmware" stuff, not U-Boot per se
<samueldr> my lack of confidence / hands-on experience with Amlogic stuff doesn't help
<samueldr> but isn't the S905 series of chip able to be used blob-free?
cole-h has joined #nixos-aarch64
<samueldr> at least, I saw that TF-A (ATF; arm trusted firmware) supports(?) S905 recently
<samueldr> hm, that would still leave some binaries from the OEM :(
cole-h has quit [Quit: Goodbye]
<samueldr> angerman: can you look at this issue? https://github.com/angerman/meson64-tools/issues/2
<{^_^}> angerman/meson64-tools#2 (by TechnologyClassroom, 31 weeks ago, open): Missing LICENSE
jumper149 has joined #nixos-aarch64
cole-h has joined #nixos-aarch64
<samueldr> anton[m]: you won't really like it because it means it won't be merged quickly, but I have a new clearer plan for better support of any meson platforms that I'm looking into implementing on top of your changes
<samueldr> e.g. I don't have any odroid boards to test with, but I do have an amlogic board I could, and it just struck me how to tie this all up together better
<samueldr> and it should allow me to better test the change
<anton[m]> samueldr: I am actually happy, and happy to test whatever you have in mind on the boards I have.
<samueldr> anton[m]: is there a particular reason for going with the OEM u-boot repo for the binaries?
<samueldr> I'm thinking about switching (from armbian) to LibreELEC's which has AFAICT the better collection
<samueldr> should make adding support for other amlogic boards much easier since it's a common source
<anton[m]> I used OEM as those were more stable for me on C2 boards couple years ago. To be honest I don't remember what was the difference when I digged into couple years ago. As long as the boards stable I am ok. My motivation to use OEM is based on better experience.
<samueldr> right
<anton[m]> I'll look at LibreELEC, I don't mind to give it a time to run and see how stable it is
<samueldr> anyway the way I'm thinking of doing it it should be a cinch to use any other source as a replacement
<samueldr> something similar, not sure yet if it'll be a discrete input per component (allowing BL31 from TF-A to be switched in through an override) or a single input for all blobs
<anton[m]> yes, makes sense
<samueldr> oh, and TF-A for S905 is already in your build
<samueldr> [for the C2]
quinn has quit [Ping timeout: 252 seconds]
rajivr has quit [Quit: Connection closed for inactivity]
<samueldr> anton[m]: were you using the `sd_fusing.sh` script?
<samueldr> AFAIUI it's just a convention from hardkernel
<samueldr> for amlogic boards it just writes the binary starting at sector 0x01 (0x00 being the first sector)
<anton[m]> samueldr:
<samueldr> not sure if you just highlighte me or something didn't come through the matrix to irc semantics
<anton[m]> * samueldr: it is convention. I used ready script to put it along with u-boot so that someone could use it to flash the card right away.
<samueldr> right, I'm thinking that none of the U-Boot builds in Nixpkgs provide scripts of the sort
<samueldr> I'm not entirely against, but I don't see it as being really needed in the end
<anton[m]> I feel like it is a good thing. Though I had in mind, after this patches be merged, to make the changes so that it would be possible to generate entire sd_image with the uboot in it
quinn has joined #nixos-aarch64
<anton[m]> for odroid boards
<samueldr> in Nixpkgs?
<samueldr> that's going in the direction I'm trying to bring things about
<samueldr> against
<samueldr> important word to forget to include in the edit :(
<samueldr> let me rephrase
<samueldr> I'm trying to remove as much board-specific "weirdness" as possible from Nixpkgs
<samueldr> I'm still kicking myself from adding temporarily a raspberry pi 4 disk image build
<samueldr> it probably created wrong expectations
<samueldr> oh, I see now that the C2 and C4 builds are *really* different
<anton[m]> hmm... do you mean you would ideally see one image that fits all? (this arm boards are really crazy, even though it became better with aarch64, comparing to arm32 zoo)
<samueldr> kind of
<samueldr> it's already the case in fact
<samueldr> mostly
<anton[m]> I don't think it is possible to remove the weirdness, though make it somewhat sane - perhaps
<samueldr> end-users have to assemble the firmware in the disk image when the firmware is not provided in a dedicated storage (e.g. SPI NOR Flash)
<samueldr> and except for the Raspberry Pi, it is mostly a successful strategy
<samueldr> ... and then I am working on the side on a new strategy that will obsolete even that strategy... hopefully
<anton[m]> cool, do you have this all ideas somewhere in git or written? I'd be happy to follow
<anton[m]> and help, if I can
<samueldr> I'll ping you later either this week or next week
<anton[m]> ok!
<samueldr> anton[m]: you have a C4, right? do you also have a C2?
<anton[m]> samueldr: yeah, I have couple C2s and C4s
<samueldr> good, also helpful to confirm we are not breaking the C2 build with the changes
<samueldr> oh, I see it's in yuor "Things done" :)
<anton[m]> there was a guy with HC1, which is similar to C4, he confirmed the changes working for him either
<samueldr> I wonder if we can make the C2 and C4 build use the same tooling
<anton[m]> nope ... at least not for signing the blobs, I found a weird reverse-engineered (my best guess) signer for C4. meson64-tools package, which has a benefit of being able to compile for aarch64. the proprietary one is x86_64 only.
<anton[m]> meson64-tools does the work, though I'd be happy to have something instead. I don't believe I will ever have time to re-write it myself. :)
<anton[m]> something better written.
quinn has quit [Quit: ZNC 1.8.1 - https://znc.in]
zupo has joined #nixos-aarch64
<samueldr> right, no worries
<samueldr> I found a bit more information
quinn has joined #nixos-aarch64
<samueldr> e.g. I figured out that what matters is the SoC family https://linux-meson.com/doku.php#target_hardware
<samueldr> e.g. C4 uses aml_encrypt_g12a, it has a S905X3, part of "SM1" family, which is quoted as "very close to G12A"
<samueldr> similarly, the board I have is S805X, quoted as "GXL", and it matches with aml_encrypt_gxl being used
<samueldr> and aml_encrypt_gxb for the C2, GXBB in that list
<samueldr> in u-boot docs when a given tool is used, it seems all boards using that same tool use it the same way
<samueldr> this is pretty good for us, since it points towards reusable build instructions much more than I though previously
{^_^} has quit [Remote host closed the connection]
{^_^} has joined #nixos-aarch64
jumper149 has quit [Quit: WeeChat 3.1]
justan0theruser has joined #nixos-aarch64
justanotheruser has quit [Ping timeout: 260 seconds]
alpernebbi has quit [Quit: alpernebbi]
orivej has quit [Ping timeout: 252 seconds]
superherointj has quit [Quit: Leaving]
srk has quit [Ping timeout: 260 seconds]
e has quit [Quit: edk]
zupo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
* samueldr learned maybe more than desirable about amlogic init setup
<samueldr> I could not test against a C2 or C4, having neither, but `xxd` comparisons tell me that the only real difference here at the end is with the specific blobs used
<samueldr> according to the U-Boot documentation, this should be usable with 9 other boards as-is
<samueldr> and some of the preparations here should make it usable for 9 other U-Boot supported boards