evils[m]1 has joined #openwrt-devel
gnustomp[m] has joined #openwrt-devel
bluse-blue[m] has joined #openwrt-devel
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
goliath has quit [Quit: SIGSEGV]
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
fieryeagle954[m] has joined #openwrt-devel
mkg20001 has joined #openwrt-devel
Grommish_ is now known as Grommish
znullptr[m] has joined #openwrt-devel
grid has joined #openwrt-devel
ekathva has joined #openwrt-devel
minimal has quit [Quit: Leaving]
srslypascal is now known as Guest31
srslypascal has joined #openwrt-devel
Guest31 has quit [Ping timeout: 480 seconds]
srslypascal is now known as Guest32
srslypascal has joined #openwrt-devel
Guest32 has quit [Ping timeout: 480 seconds]
srslypascal is now known as Guest33
Grommish has quit [Ping timeout: 480 seconds]
srslypascal has joined #openwrt-devel
Guest33 has quit [Ping timeout: 480 seconds]
Grommish has joined #openwrt-devel
sorinello has quit [Quit: Leaving]
srslypascal is now known as Guest38
srslypascal has joined #openwrt-devel
sorinello has joined #openwrt-devel
Guest38 has quit [Ping timeout: 480 seconds]
<KGB-0> https://tests.reproducible-builds.org/openwrt/openwrt_tegra.html has been updated. (100.0% images and 99.9% packages reproducible in our current test framework.)
Grommish_ has joined #openwrt-devel
Grommish has quit [Ping timeout: 480 seconds]
Grommish_ is now known as Grommish
ekathva has quit [Ping timeout: 480 seconds]
ekathva has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
danitool has joined #openwrt-devel
rua has joined #openwrt-devel
cbeznea has joined #openwrt-devel
<rsalvaterra> I'm really starting to hate the vpnc script (vpnc-scripts package) we have.
<rsalvaterra> Of course, I would hate it even more if I could even configure an OpenConnect VPN correctly, but it doesn't seem netifd is having any of it.
<rsalvaterra> That script is so wrong, in so many different ways.
<rsalvaterra> 1) Assumes the user wants to configure all the routes the VPN provides (case in point, I have a stupid VPN connection which happily routes 10.0.0.0/8 through it).
<rsalvaterra> 2) Assumes there's only a dnsmasq instance running (and what if you don't even have dnsmasq? Surprise, surprise! )
<f00b4r0> wireguard is so easy, /me whispers :)
<rsalvaterra> f00b4r0: I don't control the server. :(
<f00b4r0> change server? ;)
<rsalvaterra> Change $dayjob, more likely. :P
<f00b4r0> ah
<f00b4r0> I feel ya. Been there, done that ;P
<rsalvaterra> SSL VPNs… what the hell.
<f00b4r0> indeed
<rsalvaterra> And until recently it was PPtP.
<Grommish> PPP/Slip was always the new hawtness
<rsalvaterra> Which was actually much better, as it's much easier to configure.
<rsalvaterra> (Corporate security is someone else's problem. :P)
<f00b4r0> pptp, oh dear. Had to deal with that mess too recently. For $customer in Austria, where apparently they still use that for broadband access ;P
<rsalvaterra> Insecure as it is, at least PPtP is IP at the lower layer. SSL VPNs are mostly TCP (and we all now how wonderful it is to have TCP over TCP).
<rsalvaterra> *know
borek has joined #openwrt-devel
borek has quit [Ping timeout: 480 seconds]
borek has joined #openwrt-devel
<f00b4r0> have to run, bbl
f00b4r0 has quit [Quit: p00f]
cbeznea1 has joined #openwrt-devel
cbeznea has quit [Ping timeout: 480 seconds]
<jow> rsalvaterra: so you miss some kind of route filter?
<jow> it would make sense as generic netifd facility as other proto handlers also can announce routes, e.g. dhcp provided classless routes you might want to completely or partially filter for whatever reason
<jow> or are you complaining about the vpnc proto handler staging a default route?
<jow> the latter can be solved by the generic option defaultroute 0 option
<rsalvaterra> jow: To be honest, I didn't even reach that stage. Somehow I can't get OpenConnect to work at all. :/
<rsalvaterra> I would be happy if it just created the tunnel interface, at lest that would mean it's establishing a connection.
<rsalvaterra> I can do it manually, through the terminal, of course.
<rsalvaterra> I do see the vpnc script blindingly applying all the provided routes, though, unless I'm misinterpreting the code.
<rsalvaterra> Anyway, it's not just the default route, I don't want any routes except for the ones I explicitly create.
<jow> yeah, I see
<jow> that's a generic issue though, I believe dhcp would behave similar
<rsalvaterra> I feel *really* tempted to hack my own ad-hoc vpnc script.
<jow> same with the dnsmasq integration
<rsalvaterra> The whole thing reeks of bitrot.
<jow> handling proto handler supplied DNS info should be done in a central place, not per script
_lore_ has quit [Ping timeout: 480 seconds]
<jow> or at the very least it should respect option peerdns 0
<rsalvaterra> Oh, it doesn't? I thought that was generic and not a netifd proto thing.
<rsalvaterra> Even worse…
<jow> it is, but that vpn-script does more, it also amends the dnsmasq config
<jow> these parts should be guarded by option peerdns 0 too
<jow> netifd only takes care of ignoring the provided dns servers on peerdns 0
<rsalvaterra> The script has no business touching the dnsmasq config. If it exists at all.
<rsalvaterra> For example, I run two dnsmasq instances. That would never work.
<jow> but it can't prevent script code from modifying confdigs elsewhere in the system...
<rsalvaterra> Just to make sure we're on the same page, we're talking about the vpnc script in the vpnc-scripts (why plural? There's only one…) package, right?
<rsalvaterra> For my use case, I just need for the script to configure the IP/mask and gateway for the interface.
rua has quit [Ping timeout: 480 seconds]
<aiyion> Can someone point me to the dts file for a "Ubiquiti PicoStation M2" in ath79? I'm not sure if it's just not present, or if I'm missing something.
<aiyion> I found 'ubnt_picostation-m', but that's for the xm device, isn't it?
Slimey has quit [Read error: Connection reset by peer]
rua has joined #openwrt-devel
<aiyion> or are m2 and m5 both variants? of it?
<jow> rsalvaterra: yep, I am looking at that script
Slimey has joined #openwrt-devel
<jow> rsalvaterra: normally how it works is: a proto handler does its proto specifc thing, extracts relevant settings from env vars or whatever means, then hands over all that info tzo netifd and instructs it to setup the iface
<jow> netifd then honours options souch as peerdns, defaultroute etc. to selectively ignore info provided by thew proto handler
_lore_ has joined #openwrt-devel
<jow> but if a proto handler modifies system config directly, not involving netifd, then all those options have no effect, the protohandler has to honour those itself then
<aiyion> Mhm; "
<aiyion> Picostation M2 device uses Bullet M2 firmware image
<aiyion> "
<jow> e.g. the vpnc-script should not touch dnsmasq config at all if option peerdsns 0 is set
<jow> and it should likely not stage any routes if option defaultroute 0 is set
<jow> the latter likely even works, but only for the default route
<rsalvaterra> Any routes? Not just the default route?
<jow> not sure if "option defaultroute 0" inhibits *all* proto handler routes in netifd
<jow> maybe nbd knows offhand
<rsalvaterra> Hm…
<aiyion> That was true for ar71xx; isn't that a feature of ath79, that each device has its own image?
<nbd> defaultroute 0 should only suppress default routes
<nbd> not other proto routes
<jow> I guess we need to introduce an "option peerroutes 0" then
<jow> or an "list routefilter ..."
<rsalvaterra> I'd go with peerroutes 0.
<aiyion> So xm are 32MB ram devices...
<jow> likely sufficient, yeah
<rsalvaterra> If we're filtering specific routes, we might as well just add them manually as needed.
<jow> right
<jow> as long as the "manually add" approach provides some means to interpolate the dynamic gateway and/or device info
<rsalvaterra> What do you mean, interpolate?
<jow> assuming you e.g. have a dynamic gateway
<jow> and in /e/c/network you want some config route; option interface vpnc; option target 192.168.77.0/24; option gateway $whatever_is_current
rua has quit [Ping timeout: 480 seconds]
<jow> granted, for many VPN types it is fine to omit the gateway
<jow> e.g. for any PPP based ones
<jow> just throw the network route towards the interface
<jow> maybe it's the same for vpnc
<rsalvaterra> Hmm… I see what you mean. Yes, that could be a problem.
rua has joined #openwrt-devel
<jow> for such cases a route-filter kind of thing would be more suitable
<jow> assuming you e.h. generally want all these routes but just one or two not which are conflicting with the local network
sorinello has quit [Ping timeout: 480 seconds]
<jow> but maybe this use case is so specific and esotheric that it is not worth to support
<rsalvaterra> Well, my address is a /32 in my case, so I'm safe, at least. :)
<rsalvaterra> As it's usually a case for point-to-point links.
sorinello has joined #openwrt-devel
<nbd> jow: isn't that what metrics are for
<nbd> ?
<nbd> to make some routes rank higher than others
<jow> nbd: true, was thinking about metrics all the time as well
<rsalvaterra> nbd: That too.
<jow> nbd: which reminds me, I was toying with the idea to introduce defualt metrics for protos
c0sm1cSlug has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
<jow> nbd: to make sure that e.g. static protos rank higher than dhcp or vpn ones
<jow> that would ensure that .e.g. the lan side rmains reachable in case the wan side receives an IP in the same range
<rsalvaterra> You mean users plugging OpenWrt routers to non-bridged ISP routers? :)
<jow> alternatively we could simply slap default metrics for wan and lan into the default network config
<nbd> i'd prefer to put them in the default config instead of magic type based behavior
<jow> fine with me
<jow> nbd: but what would really help is if the default metric would be something higher than 0
<jow> regardless of the proto
<rsalvaterra> I also prefer configurations to be explicit, but it's only a matter of personal preference. I'm fine either way.
<nbd> btw. what do you think about the email i cc'd you on regarding nft flow offload?
<jow> nbd: your final points make total sense. Best effort, intention based configuration
<jow> the current approach is silly
<nbd> thx. i've had so many discussions with pablo and i always find them very tedious
<jow> allegedly kenrel 5.15+ already does some of the automatic bridge/vlan to lower device resolving automatically so they clearly see the same need
<nbd> the resolving is necessary for hw offload, but the current code forces you to be aware of it to put in the right device
<nbd> because it means you can't just put in the topmost device (e.g. bridge or vlan on top of the bridge)
<jow> yeah, and even then you can only decide between offload the bunch of hw capable ones or soft-offload all
<rsalvaterra> nbd: By the way, is hardware flow offloading working on MT7531 switches (MT7622)?
<nbd> it's working for me with firewall3, but not with firewall4 yet
<rsalvaterra> In that case, it's working for me too. :)
<jow> nbd: you mean the best effort approach?
<nbd> jow: use hw offload where possible, fall back to sw offload elsewhere
<rsalvaterra> (At least I see some HW_OFFLOAD entries in /proc/net/nf_conntrack.)
<jow> yeah, right
rua has quit [Quit: Leaving.]
bluew has quit [Quit: Leaving]
c0sm1cSlug has joined #openwrt-devel
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
<nick[m]1234> robimarko: another device can be merged: https://github.com/sartura/openwrt/pull/21
Lynx- has quit [Remote host closed the connection]
Luke-Jr has quit [Ping timeout: 480 seconds]
Luke-Jr has joined #openwrt-devel
rua has joined #openwrt-devel
borek has quit [Quit: borek]
borek has joined #openwrt-devel
borek1 has joined #openwrt-devel
borek has quit [Ping timeout: 480 seconds]
borek1 is now known as borek
robimarko has joined #openwrt-devel
minimal has joined #openwrt-devel
ptudor has quit [Read error: Connection reset by peer]
danitool has joined #openwrt-devel
valku has joined #openwrt-devel
<KGB-0> https://tests.reproducible-builds.org/openwrt/openwrt_mediatek.html has been updated. (100.0% images and 99.9% packages reproducible in our current test framework.)
pepe2k has joined #openwrt-devel
sorinello_ has joined #openwrt-devel
sorinello has quit [Ping timeout: 480 seconds]
sorinello_ is now known as sorinello
<hauke> /buffer 6
ptudor has joined #openwrt-devel
rua has quit [Ping timeout: 480 seconds]
guidosarducci has quit [Remote host closed the connection]
guidosarducci has joined #openwrt-devel
pepe2k has quit [Remote host closed the connection]
rua has joined #openwrt-devel
guidosarducci has quit [Remote host closed the connection]
guidosarducci has joined #openwrt-devel
robimarko has quit [Quit: Leaving]
Borromini has joined #openwrt-devel
Borromini has quit [Ping timeout: 480 seconds]
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
borek has quit [Quit: borek]
<KGB-1> https://tests.reproducible-builds.org/openwrt/openwrt_kirkwood.html has been updated. (100.0% images and 99.9% packages reproducible in our current test framework.)
csharper2005 has joined #openwrt-devel
m_ has joined #openwrt-devel
m_ is now known as MaxSoniX
<MaxSoniX> hauke: Hi! I would like to clarify about the comment https://github.com/openwrt/openwrt/pull/4770#issuecomment-1126314927. Do I understand correctly that you can add both devices to one PR with a common dtsi? Is it correct to divide into such commits:
<MaxSoniX> 1) Common dtsi
<MaxSoniX> 2) dts PRO + changes in mt7621.mk
<MaxSoniX> 3) dts NBN + changes in mt7621.mk
shibboleth has joined #openwrt-devel
<csharper2005> hauke: and what do you think about small and clear python scripts in $TOPDIR instead of firmware-utils? Is it possible or strictly prohibited?
<csharper2005> * in $TOPDIR/scripts
goliath has joined #openwrt-devel
MaxSoniX has quit [Remote host closed the connection]
srslypascal is now known as Guest94
srslypascal has joined #openwrt-devel
Borromini has joined #openwrt-devel
Guest94 has quit [Ping timeout: 480 seconds]
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
ekathva has quit [Remote host closed the connection]
rua has quit [Ping timeout: 480 seconds]
shibboleth has quit [Quit: shibboleth]
rua has joined #openwrt-devel
Borromini has quit [Quit: Lost terminal]
<hauke> csharper2005: I think there are already some
<hauke> better discuess this on the mailing list
<csharper2005> hauke: Hi! I also have have such scripts here - https://github.com/openwrt/openwrt/pull/4195 ... I've backported the mtd patch and now have 9 commits :)) Can you advise what to do next with this?
<KGB-1> https://tests.reproducible-builds.org/openwrt/openwrt_x86.html has been updated. (100.0% images and 99.9% packages reproducible in our current test framework.)
danitool has quit [Ping timeout: 480 seconds]
GNUmoon has quit [Remote host closed the connection]
GNUmoon has joined #openwrt-devel
valku has quit [Quit: valku]
rua has quit [Ping timeout: 480 seconds]
rua has joined #openwrt-devel
<Zero_Chaos> okay, I'm I'm getting a lot of this:
<Zero_Chaos> Fri May 13 21:30:11 2022 daemon.notice netifd: radio0b (9103): Command failed: Request timed out
<Zero_Chaos> Fri May 13 21:30:11 2022 daemon.notice netifd: radio0b (9103): Device setup failed: HOSTAPD_START_FAILED
<Zero_Chaos> when I run wifi up with 4 phys it seems to timeout one of them (at random) pretty often. but I can just manually run the same "ubus call hostapd config_add" line that was frozen (from ps) and it works fine
<Zero_Chaos> in openwrt-21.02 6 months ago it was fine, but now it's pretty reliably not fine with >3 phys
bluew has joined #openwrt-devel
csharper2005 has quit [Remote host closed the connection]
goliath has quit [Quit: SIGSEGV]