00:04 <Habbie> (it built because i put TARGET_LDFLAGS += -latomic in my openwrt package makefile)

06:38 <russell--> [2138539.834730] ath10k_pci 0000:00:00.0: mac flush vdev 1 drop 0 queues 0x2 ar->paused: 0x0 arvif->paused: 0x0

06:52 <russell--> seen on a misbehaving archer c7

13:34 <russell--> svanheule: do you know what is in the jffs and jffs2 partitions. they seem to use 2MB.

13:34 <russell--> [ 0.814821] 0x000000060000-0x000000160000 : "jffs"

13:34 <russell--> [ 0.826734] 0x000000160000-0x000000260000 : "jffs2"

13:36 <svanheule> russell--: no idea. my guess would be logs or config, but I don't remember (if I ever checked)

13:37 <svanheule> you want to use all available storage on the GS1900s?

13:44 <russell--> i was wondering why it was preserved, i guess i could imagine some reasons

14:02 <svanheule> no reason other than "12MB should be enough for a while"

14:03 <svanheule> otherwise we would have to employ mtd-concat to gain those last 2MB, but then the user would lose their stock config when flashing OpenWrt (which isn't that big of a problem in the end)

14:06 <totkeks> I was looking at the podman init.d script. Does anyone know why it uses `append_param command $command` after setting the command? https://github.com/openwrt/packages/blob/master/utils/podman/files/podman.init

14:07 <russell--> svanheule: i'm not counting on going back to vendor firmware :-)

14:08 <russell--> although i suppose for the PoE devices, that could have some value for reverse engineering

14:09 <svanheule> the main problem going forward will remain to allow an easy factory install

14:10 <svanheule> snapshot builds still fit inside the old 6MB partition, but 24.10 builds with LuCI already don't :(

14:18 <schmars[m]> totkeks: other packages use that to append options like --foo=123 to the command. i'm testing podman too and have been wondering about that line. $command doesn't come from procd, so my best guess is the author copied the init script from another package and forgot to remove the append line

14:19 <totkeks> schmars[m], okay. because I have been looking through the wiki and forum for some docs about that and couldn't find any. like use set for the binary and append for the parameters. but this doesn't seem to be the case

14:20 <schmars[m]> nah you can pass everything right into procd_set_param if it's convenient

14:22 <totkeks> schmars[m], are you using procd to run your containers? I'm testing that currently. also opened a forum thread to gather input

14:24 <schmars[m]> yeah, i just do procd_set_param command /usr/bin/podman run -i mycontainer - there might be more to consider, i'm just dabbling around at the moment

14:24 <stintel> you can always add a `logger -t podman-init "command: $command"` in the init script to see if $command actually contains anything

14:25 <schmars[m]> yep true ^

14:25 <stintel> or run the init script with INIT_TRACE=1 but then you'll have a lot of noise

14:26 <totkeks> windows terminal supports copilot ai. it might be able to filter through that and find the relevant parts :D

14:27 <jakllsch> heh

14:38 <totkeks> mh, i tried PROCD_DEBUG=1 and INIT_TRACE=1. neither shows me why it thinks the instance crashes (container itself, not podman)

14:39 <schmars[m]> you can also redirect stdout/stderr to logread

14:40 <schmars[m]> oh and make sure to not wrap the whole command including parameters in quotes, otherwise it'll try to find a binary for the whole thing

14:49 <stintel> /home/build/openwrt/build_dir/host/cmake-3.30.5/Utilities/cmlibarchive/libarchive/archive_random.c:125:9: error: implicit declaration of function 'arc4random_buf' [-Wimplicit-function-declaration]

14:49 <stintel> 125 | arc4random_buf(buf, nbytes);

14:49 <stintel> | ^~~~~~~~~~~~~~

14:50 <stintel> in openwrt-24.10 4cc1da1e44c28aa153956e179f28b8ee38dc3ac6

14:50 <totkeks> schmars[m], got it running now. needed to remove the -d flag, because well, that detaches the running container and procd thinks it died :D

15:05 <stintel> I guess it's related to building on a musl host

15:14 <totkeks> schmars[m], do you have your code somewhere public?

15:31 <schmars[m]> totkeks: https://github.com/freifunk-berlin/bbb-configs/commit/16bfddef04dd84be57ac26d82c1619d3bf67aa57 - just dabbling around, there's not much for your procd questions. i dont think i even tested the init script in there yet

15:31 <schmars[m]> my dabbling is about networking there mainly

15:31 <schmars[m]> and you can pretty much ignore the buildbot dockerfile related stuff :)

15:53 <totkeks> ah, that's a "bit" different from my tiny home setup :D

16:01 <totkeks> looks something like this now: https://nopaste.net/8RFY0RFazh

16:17 <totkeks> is there a naming convention for uci configs in terms of plural or singular (/etc/config/containers vs container)?

16:19 <enyc> is this an usuasy number of bugs for a release in -rc stage, ooi?

17:14 <SlimeyX> you only have one chance to test the automatic self destruct system

17:49 <totkeks> someone knows how to see output from the init.d script before procd takes over the instance? besides INIT_TRACE=1. like just the stdout/stderr of the commands running in the init script

18:09 <SlimeyX> like this ? https://openwrt.org/docs/guide-developer/procd-init-scripts

18:13 <totkeks> well it suggest INIT_TRACE=1 :D didn't realize that stderr and stdout are part of the trace. like if i'd set +x