<hurricos>
the coldfire board is literally just a serial-to-network device
<hurricos>
I feel betrayed
<hurricos>
oh, wait ... what?
<hurricos>
whatever's at the other end of this can ping, too. Hmm. Not sure how I traced the output of the CP2102 to the MCU then.
<hurricos>
It's definitely the MCF5272. I can upload an S-Record to it if I want. So uclinux is an option, technically. But nahh
<hurricos>
anyways
<hurricos>
enough fooling around with this. It has a network port. IT's probably fine
<hurricos>
(to be used over the network)
<slingamn>
i'm using the imagebuilder; i want to build an image that always tries to use /dev/sda as the overlay device (it's OK if it fails to boot if /dev/sda is absent). is this possible by hardcoding /etc/fstab or some other /etc files?
<slingamn>
or perhaps by changing the device tree file?
minimal has quit [Quit: Leaving]
danitool has quit [Ping timeout: 480 seconds]
amaumene has joined #openwrt-devel
cbeznea has joined #openwrt-devel
cbeznea has quit [Quit: Leaving.]
cbeznea has joined #openwrt-devel
gladiac has joined #openwrt-devel
gladiac is now known as Guest1510
gladiac has joined #openwrt-devel
Guest1510 has quit [Ping timeout: 480 seconds]
<stintel>
Grommish: it seems after /etc/init.d/network restart things stopped leaking
<Grommish>
stintel: Interesting. I've been pulling free every 5 mins to see the creep, let me check that
<stintel>
after that restart network traffic also dropped
<stintel>
might need to compare that with tcpdump next time I reboot the thing
<Grommish>
stintel: I restarted the network and now eth0 doesn't pickup an ipv4 address.. hmmmm
<stintel>
but I have a config and image that seems to reliably show the leak
<stintel>
so that's something I guess
GNUmoon has quit [Ping timeout: 480 seconds]
<stintel>
rebooted with sqm-scripts disabled, 10MiB less memory usage after boot and no leak so far
<Grommish>
stintel: I'm not running sqm at all
<stintel>
you said you don't use it, are we again chasing a different bug :/
<stintel>
show me tc q output though
<Grommish>
stintel: What do you need?
<stintel>
just show the entire output
<Grommish>
Of the boot?
<stintel>
tc q
<Grommish>
stintel: /bin/ash: tc: not found
<stintel>
ugh
<Grommish>
Not like i can't build a package.. what is tc?
<stintel>
it's part of iproute2 to manipulate "traffic control" settings (qdiscs etc)
<Grommish>
This device isn't hooked to anything but the network via WAN.. so no traffic and can be reflashed at will
<stintel>
reverted to my original changes and leak does not show :/
<Grommish>
stintel: eth0/1/2 are the only ports on the device aside from the console, I'm guessing eth3 is available on the controller, but nothing connected to it
<stintel>
this leak is fucking annoying
<stintel>
now the only other thing is different from last night is that I did a cold boot then
<Grommish>
Well, when I cycled networking, it refuses to pickup a DHCP address until I reboot the device
ecloud has joined #openwrt-devel
<stintel>
really very vague stuff
<Grommish>
I still think it's the octeon-ethernet driver, but I dunno how to prove it
<stintel>
going to try another cold boot to see if it reproduces then
<stintel>
if not that means I just lost my "reliable" repro
<stintel>
and I'll probably just mark octeon source-only and call it quits
goliath has joined #openwrt-devel
ecloud has quit [Ping timeout: 480 seconds]
ecloud has joined #openwrt-devel
<neggles>
stintel: you also gotta remember he's got an octeon3 and you have an octeonii :P
<Grommish>
neggles: Interesting.. When I made the octoen-ethernet as a module, the .ko didn't get loaded into the rom. Any idea if it should have?
<Grommish>
neggles: ./linux-octeon/linux-5.10.103/drivers/staging/octeon/octeon-ethernet.ko gets created in the build_dir hmmm
<Grommish>
I'm going to copy them into the root.orig and see if that fixes it
<Grommish>
but it seems o be the fact it's in drivers/staging
Strykar_ has joined #openwrt-devel
Strykar has quit [Ping timeout: 480 seconds]
Strykar_ is now known as Strykar
danitool has joined #openwrt-devel
<stintel>
you need to make a kmod package
<f00b4r0>
hmm. I seem to be hitting an interesting cornercase: when starting a service, "/etc/init.d/<service> running" returns true *and* "/etc/init.d/<service> status" is not "running". Is this expected?
<f00b4r0>
the other way around
<f00b4r0>
sorry. running returns false and status is running
<f00b4r0>
maybe it's just a race and each command see a different state while the service starts.
GNUmoon has joined #openwrt-devel
Misanthropos has joined #openwrt-devel
ecloud_ has joined #openwrt-devel
ecloud has quit [Read error: Connection reset by peer]
<stintel>
auch, my gs108t-v3 dropped offline after running "ethtool --show-eee lan1"
\x has quit [Ping timeout: 480 seconds]
<stintel>
and it seems to happen reliably so I created a gh issue
\x has joined #openwrt-devel
<f00b4r0>
hmm. I don't understand why kmod-ipt-coova isn't in the archive although coova-chilli is apparently built with it
rsalvaterra has quit []
rsalvaterra has joined #openwrt-devel
Misanthropos has quit [Ping timeout: 480 seconds]
<Grommish>
stintel: Ah ha! Got it.. Thanks :)
<jow>
f00b4r0: you asked that a while ago already
<jow>
f00b4r0: the reason is likely still the same
<jow>
11:09 < jow> f00b4r0: I suppose due to the deps. kmod-ipt-coova depends on (not selects) coova-chilli
<jow>
11:09 < jow> f00b4r0: since phase1 builds CONFIG_ALL_KMODS=y but does not select coova-chilli, it'll never get enabled
<f00b4r0>
oh right.
<f00b4r0>
I'm getting senile :(
<f00b4r0>
thanks for jogging my memory.
<f00b4r0>
jow: I suppose there's no fix for that, since kmods cannot be built by phase2, correct?
<jow>
they can but they'll never end up on the build hosts
<jow>
erm end up on the download servers
<f00b4r0>
I see. No way around this then.
<jow>
apart from fixing the depdnency
<jow>
there's no technical reason for it to be there, likely a cludge for menuconfig cosmetics
<jow>
are cargo culting
<jow>
*or
<f00b4r0>
oh, I see what you mean
<f00b4r0>
I'll check if the module really does not depend on the package includes, if so I'll submit a patch
<jow>
by convetion, kmods should not depend on userspace anyway
<jow>
if at all then the other way around
<f00b4r0>
indeed
<f00b4r0>
but chilli being a trashcan, I wouldn't assume anything ;P
<f00b4r0>
it looks like it could work indeed
<f00b4r0>
I'll submit a patch to remove the dep
<rmilecki>
I'm working on support for a router with 2 WAN interfaces
<rmilecki>
setup (config) needs to allow to configure them separately (e.g. dhcp vs static) so I need 2 "config interface"s in /etc/config/network
<rmilecki>
(I now have "wan1" and "wan2" interfaces)
<rmilecki>
the problem with /etc/config/firewall which has "wan" and "wan6" hardcoded (package/network/config/firewall/files/firewall.config)
<rmilecki>
how to solve that? :)
<jow>
rmilecki: for now usinga uci-defaults cludge
<rmilecki>
jow: use uci-defaults to modify firewall config?
<jow>
yes
<rmilecki>
jow: do you have any long term different solution in mind?
<jow>
yes, two
<jow>
1) using "option zone wan" on the two ifaces (but needs to be happen globally)
<jow>
2) add support for ifgroups in netifd, then fw4 and use that
<jow>
my personal favorite is 2
<rmilecki>
i could implement 1
<rmilecki>
2 is probalbly too complex for me
<rmilecki>
if you prefer 2 though there is probably no point in wasting time (and dealing with backward compatibility later) on 1
<jow>
1) is already implemented and used by virtual dynamic interfaces to join themselves into firewall zones without having to modify the firewall config
<jow>
I just never tried it in actual uci configured interfaces, could be that it's not recognized there, yet by fw3 and fw4
<jow>
and it would intersect with the list of interfaces in the firewall config
<jow>
but I guess that's tolerable for now
<jow>
(fw will likely complain about missing "wan" on start)
srslypascal has quit [Remote host closed the connection]
srslypascal has joined #openwrt-devel
niyawe has quit [Quit: niyawe]
niyawe has joined #openwrt-devel
<russell-->
what does this mean? tmp/.config-package.in:103464:symbol PACKAGE_shorewall6 depends on PACKAGE_shorewall6
isak has quit [Remote host closed the connection]
isak has joined #openwrt-devel
<Grommish>
stintel: I've managed to successfully kmod the octeon ethernet driver. Now I can yank it and see what happens
<stintel>
great
<Grommish>
It was counter-inuitive at first that the kmod has to be required rather than Moduled in order to be included so it can be turned on and off :D
<russell-->
blocktrron: someone reports (https://github.com/openwrt/openwrt/issues/5095) a unifi ap as being ar7240 (from the boot log), but the device tree says it's ar7241, do both exist in the wild (like for ubiquiti bullet m's)?
Grommish_ has joined #openwrt-devel
Grommish has quit [Ping timeout: 480 seconds]
<blocktrron>
russell--: if I'm informed correctly they are identical apart from some missing io
danitool has joined #openwrt-devel
shibboleth has joined #openwrt-devel
<stintel>
thanks again for your help and guidance
<stintel>
oops
floof58 has quit [Remote host closed the connection]
<russell-->
blocktrron: they are different when it comes to their ethernet port
<russell-->
in ar71xx they were compatible, but not ath79
shibboleth has quit [Remote host closed the connection]
<russell-->
unless something has changed since march 2020 when i found ath79 for bullet-m didn't work on ar7240 hardware
kenny has joined #openwrt-devel
gladiac has quit [Quit: k thx bye]
dave14305 has joined #openwrt-devel
shibboleth has joined #openwrt-devel
dave14305 has quit [Remote host closed the connection]
<russell-->
blocktrron: this is what I heard previously: 2019-02-24 17:53 gch981213: russell--: ar71xx has the ability to auto-detect many things but ath79 doesn't. ar7240 and ar7241 can't share the same dts on arh79 but they can use the same piece of machine code in ar71xx.
cbeznea has quit [Quit: Leaving.]
<russell-->
and then: 2019-03-23 20:40 gch981213: russell--: I don't think there is a way to do so. Maybe we should create two dts/firmware for the two variants.
shibboleth has quit [Quit: shibboleth]
floof58 has quit [Remote host closed the connection]
GNUmoon has quit [Ping timeout: 480 seconds]
floof58 has joined #openwrt-devel
Borromini has joined #openwrt-devel
floof58 has quit [Ping timeout: 480 seconds]
floof58 has joined #openwrt-devel
floof58 has quit [Ping timeout: 480 seconds]
floof58 has joined #openwrt-devel
Borromini has quit [Quit: leaving]
floof58 has quit [Ping timeout: 480 seconds]
floof58 has joined #openwrt-devel
floof58 has quit []
floof58 has joined #openwrt-devel
floof58 has quit []
floof58 has joined #openwrt-devel
Misanthropos has joined #openwrt-devel
philipp64 has quit [Quit: philipp64]
philipp64 has joined #openwrt-devel
AtomiclyCursed has quit [Quit: ZNC 1.8.2 - https://znc.in]