00:06 <schmars[m]> is APKINDEX.tar.gz a new thing? my feed has a packages.adb which used to be enough for APK iirc

00:08 <Habbie> APKINDEX.tar.gz is apk v2

00:08 <Habbie> openwrt uses v3

00:09 <Habbie> is what i know

00:09 <Habbie> so you seeing APKINDEX is a surprise

00:09 <Habbie> (that's all i got)

00:10 <schmars[m]> oh it expects "/packages.adb" to be part of the repository url in the repositories file now

00:10 <Habbie> which reminds me that https://github.com/openwrt/openwrt/pull/16427 is still sitting there

00:10 <Habbie> schmars[m], ah, i noticed that locally too

00:10 <schmars[m]> all good now

00:10 <Habbie> in apk --repository for local paths

00:11 <schmars[m]> looks like it tries "$url/$pkgarch/APKINDEX.tar.gz" if the url doesn't point to packages.adb

00:11 <Habbie> ah

00:11 <Habbie> wonder if it should try the other one too, or if the expectation really is for people to give the .adb path

00:11 <Habbie> i think alpine itself is still on v2 so these rough edges might be entirely ours for now

00:12 <schmars[m]> actually, my own notes say to have packages.adb in the url ;) so my mistake. but interesting observation

00:12 <Habbie> ok :)

09:19 <f00b4r0> https://github.com/openwrt/openwrt/blob/main/target/linux/generic/hack-6.6/711-net-dsa-mv88e6xxx-disable-ATU-violation.patch - dare I ask "why?" (and also "how?", but I digress)

09:58 <robimarko> f004br0: https://github.com/openwrt/openwrt/commit/920eaab1d8179035d0ae1047e75cf9a50da6a6eb

10:04 <f00b4r0> robimarko: thanks. The context is completely lost in the current patch (which was poorly formatted to begin with). But I wonder if we should still carry this hack, considering that ATU violations handling code has been improved upstream, according to a quick google search

10:04 <f00b4r0> e.g. https://lists.infradead.org/pipermail/linux-arm-kernel/2022-October/780000.html and others

10:37 <robimarko> f00b4r0: probably another case of somebody needing to test it

16:39 <ynezz> grift: thanks for heads up about that busybox login applet fix with SELinux https://github.com/openwrt/openwrt/pull/19080 (I've included there the selinux-policy v2.8.2 there as well)

17:43 <enyc> f00b4r0 nbd : Hrrm; any updates on that silly maybe hostapd-triggered slab-leak kernel-patches suspected, etc? I may be able to try to work out if particular devices or not; is it the sort of thing likely improved in newer kernel / 24.10.2 anyway?

17:43 <enyc> Happy to *try* to help with a proper bug report, if that is needed/best.

21:31 <Moe1978> hello all.

21:33 <Moe1978> I am quite motived (7/10) to get openwrt working on a RBSXTG-5HPacD-SA (mikrotik SXT SA5). This device is specifically not supported and I am aware of what this means. I read several threads at attempts at getting this device working, specifically found out that the RX/TX serial probe points are actually disconnected on the board - only after having soldered onto them and been frustrated in my attempt at dumping boot messages.

21:33 <Moe1978> my question is: is there a sane approach to this for someone that has a solid motivation to get it done.