<hurricos>
nick[m]1234: there is definitely such a way, it's just very specific ...
<schmars[m]>
<neggles> "have to use full build system..." <- there's actually quite a few people doing that, might be worth to get a config option. some freifunk people are sending batches of 100s of tp-linnks to some factory that swaps the chips
Lynx- has quit [Quit: Going offline, see ya! (www.adiirc.com)]
<schmars[m]>
(sorry about the matrix formatting, i remember from weechat days that it looked terrible in an irc client)
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
rua has quit [Ping timeout: 480 seconds]
goliath has quit [Quit: SIGSEGV]
GNUmoon has quit [Remote host closed the connection]
rua has joined #openwrt-devel
<neggles>
schmars[m]: it would be nice if the imagebuilder had an advanced option to override imagesize limits yeah
<neggles>
nick[m]1234: if you have a boot log i can tell you what the DT should look like
Tapper has quit [Ping timeout: 480 seconds]
<neggles>
stintel: I have DSA kinda-working on the oc200 :D
<neggles>
it works but it has some trouble with the MIB, probably because it's expecting more ports than there are or somesuch, so I get packets but the driver isn't handling the headers right
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
hexagonwin has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
GNUmoon has joined #openwrt-devel
valku has quit [Quit: valku]
ekathva has joined #openwrt-devel
nixuser has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
hexagonwin has quit [Read error: Connection reset by peer]
<neggles>
this puts vlan 1 untagged/default on all ports and 10/20/30/40 tagged on all ports (at least in theory, works on other dsa devices)
<neggles>
aw man
<neggles>
stintel: OC200's default root password is `HWOmada`
<neggles>
took two weeks for hashcat to find that (i'd forgotten it was running)
hexagonwin has quit [Ping timeout: 480 seconds]
hexagonwin has joined #openwrt-devel
<nick[m]1234>
neggles: great! :) later today I can send you a complete bootlog.
<nick[m]1234>
is there any reason our lua version is so old?
hexagonwin_ has joined #openwrt-devel
hexagonwin has quit [Read error: Connection reset by peer]
nixuser has quit [Remote host closed the connection]
nixuser has joined #openwrt-devel
hexagonwin__ has joined #openwrt-devel
hexagonwin_ has quit [Ping timeout: 480 seconds]
hexagonwin__ has quit [Ping timeout: 480 seconds]
srslypascal has quit [Ping timeout: 480 seconds]
srslypascal has joined #openwrt-devel
<neggles>
nick[m]1234: is it old?
<neggles>
I assume it's as new as it needs to be for luci et al.
Borromini has joined #openwrt-devel
<russell-->
neggles: i'm familiar with examples, what i'm missing is what a bridge-vlan is. ubiquiti er-x doesn't need it
<neggles>
russell--: are you familiar with cisco-style switch config?
<russell-->
no
<neggles>
that's ok, it turns out to not be very useful anyway
<russell-->
er-x uses dsa
<neggles>
so a bridge-vlan object is a vlan ID which the bridge knows about / is aware of
<neggles>
if you don't care about using vlan tags, you don't need one
<russell-->
on erx, it's easy to just bridge eth2 and eth3.100, etc
<neggles>
e.g. on an ER-X, say you wanted to use lan1 for WAN, and lan2-5 for LAN, but you don't care about vlan tagging anywhere - just put lan2 thru lan5 into a bridge, use that for lan, then directly bind lan1 to wan
danitool has joined #openwrt-devel
<russell-->
where eth0-eth4 are the switch ports
<neggles>
then s/lan1/eth0, s/lan2-5/eth1-4
<neggles>
russell--: if eth0-4 are the switch ports, what does the cpu port show up as? switch0?
<russell-->
eth0@dsa
<neggles>
ah ok dsa
<russell-->
if i want tagged packets going out a port, i just add eth2.100 to one bridge, maybe eth2.300 to another bridge
<neggles>
yes, so the whole thing with DSA is it removes that
<neggles>
does it in the switch instead of through the CPU
<russell-->
ethN *are* the switch ports!
<neggles>
yes, but binding to them that way causes all traffic to hairpin via the CPU
<neggles>
this config would give you, eth0 = WAN, direct-bound, eth1-4 = LAN bridge, with vlan 0 untagged on all 4 attached to interface 'lan', vlan 100 tagged on eth1 with interface 'lan.100', vlan 300 tagged on eth2 with interface 'lan.300'
<neggles>
rather than binding to the tagged vlan on the port, you bind to the VLAN on dsa cpu-facing interface, then use the bridge-vlan to configure tagging on ports
<neggles>
(I am not 100% sure that 'dsa' is the correct base device name to use there)
<neggles>
it should be, though
<neggles>
if `ip link` shows `dsa` + `eth0@dsa` thru `eth4@dsa` then it is
<neggles>
I am also not clear on whether the 'dsa' interface itself needs to be in the bridge/bridge-vlan and I don't know if i have a dsa-enabled thing i can test with
<neggles>
oh i have an ER-X right here
<russell-->
dir860l is another mt7621 based device that behaves the same
<Borromini>
neggles: ip link doesn't show anything like that on MT7621
<russell-->
dir860l uses eth0 and wan lan1-lan4 for ports
<neggles>
compare and contrast
<neggles>
that's defconf, no vlan filtering
<neggles>
russell--: just realised ive used the wrong 'option device' for the VLAN interfaces in my earlier example
<neggles>
updated it
<Borromini>
neggles: is there some backward compatibility of will configs auto-migrate?
<neggles>
Borromini: I dunno, but it's purely a name-change
<Borromini>
whenever you modify /etc/config/network e.g.
<Borromini>
ok
<neggles>
converting from the lan1-5@eth0 to eth0-4@dsa is just "decrement number by 1, change 'lan' to 'eth'"
<neggles>
I imagine having eth0 be a port that doesn't really exist was confusing people
<Borromini>
:)
<russell-->
i think the rule is to follow the labelling on the case
<Borromini>
yes, that's how the labels started out afaik.
<neggles>
that would make sense
<russell-->
speaking of erx, i've had a few (apparently power interruption related) squashfs failures relating to unreadable NAND blocks now, latest being today
<russell-->
i think i've seen it on two other occasions
<russell-->
more than that, actually, but in two cases it rendered the device unbootable, on a few other occasions it was less consequential failure and i was able to reflash it remotely
<russell-->
three cases including todays
<Borromini>
is the ER-X NAND?
Tapper has joined #openwrt-devel
<neggles>
Borromini: yep
<neggles>
russell--: that would be because squashfs doesn't have any bad block handling really
<neggles>
nand should be ubifs
* neggles
shakes fist at ubiquiti
<neggles>
(...and everyone else)
Borromini has quit [Ping timeout: 480 seconds]
<russell-->
it's squashfs on ubi
<russell-->
ubinfo -a
<russell-->
it isn't the lack of badblock handling, it's the sudden loss, or possibly just corruption, of an essential block
<russell-->
Count of bad physical eraseblocks: 0
robimarko has joined #openwrt-devel
<russell-->
fwiw, this is what i want to work on gs108t (and which, i think, works on er-x and dir860l): https://paste.debian.net/1237467/
<neggles>
russell--: also, that's... weird, squashfs being read-only and all
<neggles>
how can you corrupt something you don't write to?
<robimarko>
Doesnt mean that the storage media cant get corrupted
<russell-->
i'd prefer if it didn't, but it's hard to reproduce
Borromini has joined #openwrt-devel
<neggles>
i mean i'm sure it's *possible* but i was under the impression that it was practically impossible to damage/corrupt nand/nor just by reading it
<neggles>
hell i've had it happen before but not regularly (and it was on an octeon with a broken driver :P)
<russell-->
empiricism says otherwise
<russell-->
again, seemed have have coincided with a power anomaly
mirko has quit [Ping timeout: 480 seconds]
<russell-->
kind of kicks robustness in the teeth
<nick[m]1234>
Why does iptables-nft selects iptables-legacy?
<mkg20001>
Which is the right group to ask about libuclient? I'm trying to GET json from localhost (olsr jsoninfo), but after calling uloop_run() none of the passed callbacks get executed.
Misanthropos has joined #openwrt-devel
daniopasa_ has joined #openwrt-devel
<daniopasa_>
one problem
<daniopasa_>
Hi all
<daniopasa_>
Javier Marcet
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]