<enyc>
Habbie: many reasons, many small fixes security or otherwise; if a bug still present reporting it against that point, useful as delineator clear version change on those back-and-forth partition routers; ...
n3ph has joined #openwrt-devel
<enyc>
Habbie: tends to be a time for me to re-run my imagebuilder commands, make images etc...
<T-Bone>
nbd: test location looking really good with no visible side effects so far
T-Bone is now known as f00b4r0
<f00b4r0>
nbd: also I'll open a GH issue with my current findings on the filogic VLAN leak, maybe it'll be useful. The gist of it is there is an untagged LLC frame being emitted on non-DSA ports when a new MAC is learnt by the device for a bridge vlan tagged member. This doesn't happen on DSA ports so I suspect some internal DSA thing is leaking out but couldn't find clues in the code yet.
<f00b4r0>
nbd: that's a builtin driver, right? I need to update the whole thing?
<nbd>
yes
<f00b4r0>
ok. that's going to take a lil longer but I'll do it now, next week is going to be a zoo for me
<nbd>
ok, thx
<f00b4r0>
np yw
n3ph_ has quit [Ping timeout: 480 seconds]
n3ph has joined #openwrt-devel
<f00b4r0>
nbd: Works For Me™
<nbd>
yay
<nbd>
will send it upstream now, then add it to master/24.10
<f00b4r0>
thanks a lot
n3ph has quit [Ping timeout: 480 seconds]
<djfe>
<3
<nbd>
f00b4r0: so is that bug the primary cause of the reported roaming failures?
<nbd>
or are there other related issues?
<f00b4r0>
my current analysis is that it is the primary cause yes.
<f00b4r0>
it messes up ARP tables on routers and switches and then hell breaks loose
<f00b4r0>
DHCP seems especially affected: renewals REPLYs where never reaching the clients
<f00b4r0>
I don't want to jinx it but I think that all the major bugs I'm aware of would now be fixed (assuming the slab leak is fixed also, which I haven't had time to confirm yet)
<stintel>
that failing part seems to be $(LIBBPF_BOOTSTRAP): in tools/bpf/bpftool/Makefile but that hasn't been changed since 2023 (at least in 6.6.102 git log)
<nbd>
stintel: pushed some fixes - it should build now
<stintel>
nbd: thanks, testing
aiyion has joined #openwrt-devel
<stintel>
nbd: can confirm it builds now, thanks again
<stintel>
anyone else noticed uhttpd in master being broken (I suspect due to the openssl bump)? Mon Sep 1 01:09:34 2025 daemon.err uhttpd[6651]: Failed to load certificate/key files
<stintel>
removing / regenerating the certs doesn't help
danitool has joined #openwrt-devel
kenny has joined #openwrt-devel
<stintel>
and the EAP615-Wall says no after sysupgrade