narmstrong changed the topic of #linux-amlogic to: Amlogic mainline kernel development discussion - our wiki http://linux-meson.com/ - ml linux-amlogic@lists.infradead.org - official channel moved from Freenode - publicly logged on https://libera.irclog.whitequark.org/linux-amlogic
jn has quit [Ping timeout: 260 seconds]
jn has joined #linux-amlogic
jn has joined #linux-amlogic
naoki has joined #linux-amlogic
_whitelogger has joined #linux-amlogic
hexdump01 has joined #linux-amlogic
hexdump0815 has quit [Ping timeout: 258 seconds]
hexdump0815 has joined #linux-amlogic
vagrantc has quit [Quit: leaving]
_whitelogger has joined #linux-amlogic
cyrozap has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
cyrozap has joined #linux-amlogic
ldevulder has joined #linux-amlogic
tdebrouw has joined #linux-amlogic
psydroid3 has joined #linux-amlogic
<narmstrong> f_: yay \o/
<narmstrong> f_: you must make sure altering the defconfig doesn't break builf without the tfa blobs and using u-boot the "old" way
psydroid3 has quit [Ping timeout: 258 seconds]
<f_> No reason it would
<f_> The build process will build a u-boot.bin and a u-boot-meson-with-spl.bin
<f_> If you want to use the tfa blobs you'd use the u-boot.bin
<f_> Else you'd use the u-boot-meson-with-spl.bin
<f_> But if that's broken that's a bug in my POV
psydroid3 has joined #linux-amlogic
djrscally has joined #linux-amlogic
psydroid3 has quit [Ping timeout: 258 seconds]
<f_> narmstrong: if you want I can test regular bl2 to make sure
<f_> But I really doubt something broke
<f_> +there
<narmstrong> I’ll pass it in the CI anyway !
<f_> 22:45 <lvrp16> f_: sick, looks like I can patch lzma on top and have smaller bl33 <= and not just bl33 :) bl31, bl30 also
naoki has quit [Quit: naoki]
<narmstrong> f_: how am I suposed to pull stuff from https://scm.dersco.re/amlogic/u-boot-kii-pro.git/ ? I tried `git fetch https://scm.dersco.re/amlogic/u-boot-kii-pro.git wip/spl-2025.08-3` but it gave me some errors `error: File 637d9ff7815b4e0d47f0679dde77c795a48c0e6c (https://scm.dersco.re/amlogic/u-boot-kii-pro.git/objects/63/7d9ff7815b4e0d47f0679dde77c795a48c0e6c) corrupt
<narmstrong> `
<f_> hmm
<f_> oh maybe I need to properly setup git smart http..
<f_> strange ..
<f_> but this is a giant repo for dumb http I think .. so I'll try to get it running asap
<narmstrong> f_: I think the scaper protection makes this
<f_> hmm no I don't think so .. I'm quite sure I whitelisted git's http UA
<narmstrong> ```
<narmstrong> ```
<narmstrong> warning: ignoring alternate that does not end in 'objects': left: 2.25em; /* (container width - part width)/2 */
<narmstrong> warning: ignoring alternate that does not end in 'objects': box-shadow: 0 3.5em #eee; /* container height - part height */
<narmstrong> warning: ignoring alternate that does not end in 'objects': transform-origin: 50% 2.5em; /* container height / 2 */
<f_> aha
<f_> yeah seems like it
<f_> yep .. I return 201 on the challenge
<narmstrong> seems info/refs only works
<f_> Try now
<f_> sorry the LLM scrapers started hammering my website so I had to put measures :(
<f_> er
<narmstrong> worse
<f_> yeah
<f_> ok I removed them for now
<narmstrong> \o/
<f_> let's see if they come back to hammer my poor cgit again
<narmstrong> let's see if CI likes the changes ;-)
<f_> I like maintaining infra but this is not a very exciting thing to do.. it is what it is I guess
<f_> (this = fighting against LLMs)
<f_> sorry for the inconvenience
<narmstrong> no prob
<f_> Does the CI run checkpatch on this?
<f_> https://scm.dersco.re/amlogic/u-boot-kii-pro.git/log/?h=wip/spl-2025.08-cp <- these kind of pass checkpatch (no errors)
<f_> Ah yes, forgot binman complains when you don't set SCP= and BL31=
<f_> how do I make it not error ...
<f_> Maybe I should disable it for now?
<f_> narmstrong: doing `BINMAN_ALLOW_MISSING=1 make -j8` fixes it .... but I wonder if that'd be an acceptable workaround/solution..
<f_> I'm not really happy with having to add BINMAN_ALLOW_MISSING everytime ..
<narmstrong> Yeah weird, how does other board do to just show warnings ?
psydroid2 has joined #linux-amlogic
ldevulder has quit [Ping timeout: 250 seconds]
<narmstrong> BINMAN_ALLOW_MISSING=1 doesn't make buildman not error out
<f_> It does make binman not return a non-zero exit
vagrantc has joined #linux-amlogic
<narmstrong> buildman needs the -M parameter for that
<f_> -M is already set by default I think?
<f_> it's --allow-missing. Then BINMAN_ALLOW_MISSING is for --ignore-missing .. confusing I know
<f_> $ cp -ar /tmp/build/fip/* ${CI_PROJECT_DIR}/results/
<f_> cp: cannot stat '/tmp/build/fip/*': No such file or directory
<narmstrong> I just added -M, it buildman fails it doesn't generates the fips for the legacy bl2
ungeskriptet has quit [Ping timeout: 256 seconds]
ungeskriptet has joined #linux-amlogic
<f_> hmm ... what?
<narmstrong> hmm yes -M is not the solution
<f_> oh I confused binman with buildman xD
<f_> I guess the solution would be a bunch of #ifdef's in the binman node, or disable binman completely
<narmstrong> I think -W is the key
<f_> a workaround .. :/ that doesn't feel right to me
<f_> -W, --ignore-warnings
<f_> Return success even if there are warnings
<f_> let's just disable binman for now and then tell people in the docs ..
<narmstrong> the uboot main CI does that for the world build
<f_> mm, okay so not that bad, but I really don't like having to do -W *and* BINMAN_ALLOW_MISSING=1 .. this basically breaks potencial downstream distros that package u-boot
<f_> potential*
<narmstrong> so there's no reason, -E any warning an error and -W will ignore any other warnings
<f_> yeah right
<f_> but I still don't feel great about needing BINMAN_ALLOW_MISSING=1 for legacy fips..
<narmstrong> I don't care, if you really care then you can either add a separate defconfig or a fragment on top
ungeskriptet has quit [Ping timeout: 250 seconds]
<f_> I mean for my pmOS u-boot package I don't really care all that much, I personally want the amlogic u-boot package gone asap
<f_> in favor of fwupd and friends
<f_> but I don't know about the other distros ..
<f_> maybe I'm thinking too much about it and I should leave it
<narmstrong> seems to crash at "Model: Libre"
<f_> strange
<narmstrong> hmm
<f_> I'll test on my side
<narmstrong> it seems it waits for an SPL prompt `Bad pattern found on SPL console: stop_autoboot_prompt`
<f_> oh it's looking for the "stop autoboot" thing?
<f_> that prompt kinda broke recently
<f_> It would show "Hit any key to stop autoboot: 2" then "Hit any key to stop autoboot: 12" and then "Hit any key to stop autoboot: 02"
<f_> maybe it's crashing because of that?
<narmstrong> ok so it booets perfectly fine, the test suite stops
tdebrouw has quit [Quit: Leaving.]
<narmstrong> so the fragment to enable the SPL would probably be a better idea, or a separate defconfig including the main one and enabling the SPL stuff
ungeskriptet has joined #linux-amlogic
<narmstrong> OK I just need to add env__spl_skipped to the test env, no big deal
<f_> a fragment could work...... but I wonder how it would need to be implemented
<f_> because there are some differences between boards when it comes to SPL/dram stuff
<narmstrong> it's just a .config with the config you want to enable, and just pass it at make with the defconfig: `make libretech-cc_defconfig board/librecomputer/aml-s905-cc/spl.config` something like that
<narmstrong> we do this on qcom
<f_> hmm that could work
<f_> but uh, how about boards which don't have a separate boardfile?
<narmstrong> just add it in the board dir they use with a prefix
<f_> ok that makes sense
<f_> actually this solves the other issue I wanted to think about, that is supporting 2 GB lepotato alongside 1 GB lepotato
<f_> I can just have 2 config fragments, one for the 1gb variant and the other for the 2gb variant
<f_> It was either that or have 2 different defconfigs or have only the 1 GB defconfig and 2 GB users would have only 1 GB of RAM instead of the 2 GB they paid for :P
<narmstrong> aaaand it passes https://gitlab.com/amlogic-foss/amlogic-u-boot-autotest/-/jobs/11262258816 congrats you didn't break the non-spl run !
<f_> yay!
<narmstrong> yeah go go with fragments then
<f_> this would also make my work easier 😂
ungeskriptet has quit [Ping timeout: 256 seconds]
ungeskriptet has joined #linux-amlogic
chewitt has quit [Quit: Zzz..]
<lvrp16> haha learned a new trick for configs
<lvrp16> didn't know you can multi-target like that
<f_> yeah kconfig fragments are quite useful!
psydroid2 has quit [Quit: KVIrc 5.2.6 Quasar http://www.kvirc.net/]
djrscally has quit [Ping timeout: 256 seconds]