00:02 <stintel> alright so unbricked via ymodem, saveenv on the eap615-wall is apparently enough to kill u-boot

00:04 <m4t> heh i bricked my pogoplug with saveenv once... default environment address overlapped the u-boot img

00:05 <m4t> maybe it worked at some point but u-boot grew in size

00:05 <stintel> maybe I can just change KERNEL_LOADADDR in target/linux/ramips/image/mt7621.mk

00:29 <stintel> ugh

00:34 <owrt-images-builds> Build [#170](https://buildbot.openwrt.org/images/#/builders/160/builds/170) of `openwrt-24.10_mediatek/mt7623` failed: failed Git remote refs fetched (failure)

00:42 <stintel> oh for fucks sake

00:45 <stintel> I'm getting close to reverting 6.12 default for ramips, clearly this was not properly tested

04:57 <owrt-images-builds> Build [#323](https://buildbot.openwrt.org/images/#/builders/52/builds/323) of `main_omap/generic` completed successfully.

04:58 <owrt-images-builds> Build [#172](https://buildbot.openwrt.org/images/#/builders/220/builds/172) of `openwrt-24.10_omap/generic` failed: failed Git remote refs fetched (failure)

05:47 <owrt-images-builds> Build [#317](https://buildbot.openwrt.org/images/#/builders/234/builds/317) of `main_bcm4908/generic` failed: failed Git remote refs fetched (failure)

07:27 <djfe> f00b4r0: 20:15:47 atleast for Mediatek, apparently some qualcomm targets also experience a similar issue regarding roaming (no clue) but it was reported here atleast:

07:27 <djfe> https://github.com/openwrt/openwrt/issues/11650#issuecomment-3240387112

07:29 <djfe> 20:19:38 I'm not entirely sure the slab leak has been fixed, I've one device that reboots daily and still shows a memory leak on almost every reboot: https://grafana.ffac.rocks/d/000000002/node?refresh=1m&orgId=1&var-node=0cea14536095&from=now-7d&to=now&timezone=browser&var-nodetolink=0cea1453636d&var-hostname=ffac-Couvenmuseum-OG1-Festsaal

08:16 <f00b4r0> djfe: your dashboard does not show sufficient granularity to confirm this is a slab leak

08:17 <djfe> true in the past weeks on the new mt76 it was one, the few times I looked, but I'll keep a closer look now at /proc/meminfo

08:18 <f00b4r0> you should exposed "used" mem at least. Then you can deduct slab. But ideally you really should keep track of all used and slab_recl/unrecl

08:20 <f00b4r0> anyhow, I'm off. ttyl

10:30 <owrt-images-builds> Build [#322](https://buildbot.openwrt.org/images/#/builders/58/builds/322) of `main_apm821xx/sata` completed successfully.

12:24 <owrt-images-builds> Build [#316](https://buildbot.openwrt.org/images/#/builders/208/builds/316) of `main_bcm47xx/legacy` failed: failed Git remote refs fetched (failure)

18:27 <soxrok2212> Ansuel: ping

18:43 <soxrok2212> merbanan: perhaps

18:43 <soxrok2212> do you know if the an7581-rfb u-boot would work here on this w1700k

19:02 <merbanan> umm, I have not tested

19:03 <merbanan> what is the an7581-rfb uboot ?

19:09 <soxrok2212> its the only defconfig i see in upstream u-boot for an7581

19:09 <soxrok2212> probably won't

19:10 <merbanan> ok, I have not had the time to test it

19:15 <merbanan> you might need to pack the boot loader also

19:17 <plappermaul> Next two Realtek patches ready

19:17 <plappermaul> https://github.com/openwrt/openwrt/pull/19873

19:18 <plappermaul> https://github.com/openwrt/openwrt/pull/19832

20:02 <LEdoian> Hello, I keep wondering about netifd. When setting an interface to a ip[46]table, the interface_add_addr_rules function creates IP rules for to use that table for the traffic.

20:03 <LEdoian> Is there a way to sidestep that? I want to use my own rules for when that table is used

20:03 <LEdoian> (Apart from writing my own ruleset with lower priority from scratch)

20:08 <LEdoian> (The problem I am solving is that I need a router to be reachable in one network (for management), but not forward packets from other networks to it (they should be routed through a firewall))

20:16 <owrt-images-builds> Build [#318](https://buildbot.openwrt.org/images/#/builders/234/builds/318) of `main_bcm4908/generic` completed successfully.

20:49 <enyc> LEdoian: I suspect this imore of an #openwrt channel question. Each openwrt interface can be set in a firewall zone and you don't need to have input enabled on each zone (access to openwrt itslef from that zone) and you don't have to allow from one zone to nanither. I don't think this into custom-rules-territory myself.

21:00 <LEdoian> I don't think this is a firewall question, because I need to change routing decisions and in netfilter the route hook type only applies to output chain, not forward, and I think I would still need a separate routing tables and rules for that.

21:01 <LEdoian> And I thought about asking here, because I think I know what is happening, I just need netifd not to do that…

22:39 <hauke> nbd: could you please upload https://www.hauke-m.de/files/backports-openwrt/backports-6.12.44.tar.xz to http://mirror2.openwrt.org/sources/

22:41 <owrt-images-builds> Build [#169](https://buildbot.openwrt.org/images/#/builders/129/builds/169) of `openwrt-24.10_ipq806x/generic` failed: failed './scripts/feeds update' (failure)