<mangix>
huh so what is this mt7921? wifi chipset or SoC?
<pkgadd>
wifi
<pkgadd>
and bluetooth
PaulFertser has joined #openwrt-devel
<ynezz>
aparcar[m]: fuzzing is ongoing project, you can take a look at libubox, cgi-io, mdnsd, ubus, ucert, uci projects it's in tests/fuzz
ianchi has joined #openwrt-devel
<ynezz>
I've locally some wip in other projects, like for example odhcpd, but this needed a lot of refactoring (and maybe needs even more) so the user controlled and possibly malicious input (packet parsing) stuff could be used for fuzzing
<ynezz>
such "noop" changes/refactoring introduce regressions, so I want to have a possibility to runtime test them first and thus I got stuck at this runtime testing preparation stage for about a year now :p
<ynezz>
anyway, there is no magic involved, it's just a lot of manual work
<ynezz>
and I've a plan to submit that into oss-fuzz as well, once I'm happy with it
<ynezz>
currently we run the fuzzers via daily scheduled CI jobs on GitLab.com
<ynezz>
but it's very limited testing, basically preventing already fixed regressions as the fuzzer runtime is limited to 5 minutes
<ynezz>
ideally there should be some project dedicated to this, with status page for every project (CI, -Wextra, static analyzers, ubsan, valgrind, unit tests, fuzzing) and sort them by priority with network exposed services and shipped by default first
<ynezz>
and document the workflow, thus make it possible for others to join the effort
<aparcar[m]>
ynezz: I just wanted to look into fuzzing for ubus, but if you got it all figured out and already planing to submit it to oss-fuzz please take the lead
<aparcar[m]>
*keep
<ynezz>
it's certainly not all figured out, there is just some basic fuzzing, feel free to submit that to oss-fuzz
<aparcar[m]>
okay
gch981213 has joined #openwrt-devel
opal has quit [Ping timeout: 268 seconds]
jow has joined #openwrt-devel
xes has joined #openwrt-devel
dedeckeh has joined #openwrt-devel
opal has joined #openwrt-devel
ianchi has quit [Quit: Connection closed for inactivity]
schwicht has quit [Read error: Connection reset by peer]
ekathva has quit [Quit: Connection closed]
schwicht has joined #openwrt-devel
Acinonyx has quit [Ping timeout: 260 seconds]
Acinonyx has joined #openwrt-devel
plntyk has joined #openwrt-devel
Tost has joined #openwrt-devel
caiortp has quit [Ping timeout: 264 seconds]
gch9812138 has joined #openwrt-devel
<karlp>
hrm, let me see if I can reproducce it again first :)
gch981213 has quit [Read error: Connection reset by peer]
gch9812138 is now known as gch981213
<karlp>
don't hold your breath though, if it works for you two, I'm ok with going ahread witht aht.
<karlp>
is that ms to wait after an interface cahnge before notifying snmpd?
<stintel>
karlp: yeah before reloading the service
<stintel>
karlp: I was unable to reproduce myself, after I wrote that first comment :/
<stintel>
but I believe I had it twice on my GS108Tv3
<karlp>
I've been away from openwrt work for a while, working on other things,
<stintel>
I've now added the SNMP service to my Icinga monitoring, so next time I'll be notified instantly, instead of just having empty stats
<stintel>
no worries, just wanted to poke you as you're the one that opened the bug
<karlp>
thansk, I'm going to set that up again and leave it rebooting for a bit, see what I can trigger
<stintel>
but thinking about it, such a trigger on interface event, especially on slow devices, wouldn't be surprising if that triggers a race somewhere sometimes
<karlp>
I mean, it still kinda smells like snmpd bug, but...
<karlp>
good luck convincing them :)
<stintel>
:P
<stintel>
yeah, I agree
<stintel>
but if I can't reliably reproduce it I'm not going to bother reporting it
<stintel>
maybe I should get a really slow device and configure many interfaces
Borromini has joined #openwrt-devel
ivanich has quit [Ping timeout: 264 seconds]
decke has quit [Quit: Leaving.]
ianchi has joined #openwrt-devel
Slimey has quit [Quit: AdiIRC is updating to v4.0]
Slimey has joined #openwrt-devel
poljar has joined #openwrt-devel
poljar1 has quit [Ping timeout: 246 seconds]
poljar has quit [Ping timeout: 245 seconds]
poljar has joined #openwrt-devel
danitool_ has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
<plntyk>
Grommish, if some "standard" types are not picked up its sometimes because of source code flags and/or compiler settings ; some "portable" sources use POSIX_SOURCE or some other variants (some include libbsd ...) and afaik musl is more standard "conform"
<Grommish>
plntyk: Is there a C/CXX flag I could add?
<plntyk>
or patch the build system/detection so that those flags are set
<Grommish>
Ok, well, the issue is with the libpcap. those errors were from target_musl/usr/include/pcap/bgf.h and missig from pcap/pcap.h, which is where I put the typedefs that fixed it
<philipp64>
seems like a kludge that a meta-package has to have an empty /install rules section... but then we don't have a way of calling out meta-packages explicitly, so... maybe there's no better way for now?
<philipp64>
well, not *entirely* empty. a single "\t:" rule...
<philipp64>
speaking of firewall, when does it get refreshed next?
<philipp64>
but there have been commits since then.
<aparcar[m]>
ldir: so nftables is a good thing?
dedeckeh has joined #openwrt-devel
<ldir>
aparcar[m]: I don't have any real experience of it. Apparently it's the 'new' hotness.
<ldir>
I've a personal interest in that something easy to implement in a patch to an iptables helper requires a significant extension to the nftables parser...beyond my skill.
<Grommish>
Woot.. I got Suricata 6.0.2 working.. Thanks agai plntyk!
<ldir>
I'm unclear if packages using iptables (sqm-scripts, miniupnpd, banip, adblock) etc will require re-writes and/or nftables versions.
<olmari>
Nftables has (possibility of) compatibility layer, no idea how 100% it is
danitool has quit [Quit: Cubum autem in duos cubos, aut quadratoquadratum in duos quadratoquadratos]
flokli is now known as FLOKLI
<Borromini>
Grommish: how long since you wanted to get that going? :D
<Borromini>
congrats though :)
<Grommish>
Borromini: Far too long.. but.. Openwrt can add Suricata 6 and rust-lang if they want now ;p
<Borromini>
=)
<Grommish>
also means i can't really put off this ER10x anymore
<Grommish>
meh
<Borromini>
:P
<Borromini>
you got further than the actual owner didn't you
<Borromini>
did you find a dsa driver?
<Grommish>
No, but I found a datasheet
<Grommish>
I wonder if I can change the 8366rb dsa driver with the proper stuff
<Grommish>
I don't know enough to not try ;D
<Borromini>
:P
<Grommish>
and yes, I still have that odd switch issue, but 9 of the 10 ports are working
<Grommish>
I need to figure out the LEDs for the rtl switch now
zkrx has quit [Ping timeout: 246 seconds]
zkrx has joined #openwrt-devel
<Borromini>
:)
<philipp64>
aparcar[m]: ping
<aparcar[m]>
philipp64: pong
<philipp64>
Have a look at PR 4006 on openwrt? Would be nice to have CI/CD coverage of the parent repo... sigh.
<philipp64>
Not just packages.
<aparcar[m]>
philipp64: feel free to write a mail that you want some more CI
<aparcar[m]>
I can build test this but have no clue about ASM
<philipp64>
Well, now might not be an opportune time... CI has a bad taste in everyone's mouth right at the moment.
<aparcar[m]>
does it?
<philipp64>
If the gmpdh stuff builds in strongswan for x86_64 it should be okay...
<philipp64>
We've had to turn off CircleCI because it's broken and no one seems to be owning it right now.
owrt-snap-builds has quit [Quit: buildmaster reconfigured: bot disconnecting]
owrt-snap-builds has joined #openwrt-devel
<aparcar[m]>
well
owrt-2102-builds has quit [Quit: buildmaster reconfigured: bot disconnecting]
owrt-2102-builds has joined #openwrt-devel
owrt-1907-builds has quit [Quit: buildmaster reconfigured: bot disconnecting]
owrt-1907-builds has joined #openwrt-devel
owrt-snap-builds has quit [Remote host closed the connection]
owrt-2102-builds has quit [Remote host closed the connection]
owrt-1907-builds has quit [Remote host closed the connection]
paracyst_ has quit [Read error: Connection reset by peer]
owrt-snap-builds has joined #openwrt-devel
owrt-1907-builds has joined #openwrt-devel
owrt-2102-builds has joined #openwrt-devel
<philipp64>
I love CI/CD. I'd argue for more coverage, like a sanity boot, applying vonfigurations, a ping test... positive and negative testing of the firewall, etc...
<philipp64>
But we're a long way from there.
<aparcar[m]>
Well figure something out
<aparcar[m]>
*we'll
owrt-1907-builds has quit [Remote host closed the connection]
owrt-2102-builds has quit [Read error: Connection reset by peer]
owrt-snap-builds has quit [Remote host closed the connection]
owrt-snap-builds has joined #openwrt-devel
owrt-1907-builds has joined #openwrt-devel
owrt-2102-builds has joined #openwrt-devel
poljar has quit [Remote host closed the connection]
owrt-snap-builds has quit [Remote host closed the connection]
owrt-1907-builds has quit [Remote host closed the connection]
owrt-2102-builds has quit [Remote host closed the connection]
poljar has joined #openwrt-devel
rmilecki has quit [Ping timeout: 246 seconds]
owrt-snap-builds has joined #openwrt-devel
owrt-2102-builds has joined #openwrt-devel
owrt-1907-builds has joined #openwrt-devel
<Hauke>
jow: I looked at the uhhtp problem with the 400 you saw
* russell--
was just looking at his logs, sorted lines by nick, jow (and permutations) is the runaway leader over the last decade-ish (about 7.8%).
<Grommish>
For the package folks. I'm making a Config.in file for options. If these options are set in ./configure as flags, do I have to do an ifeq($(CONFIG_xxx),y) to insert the --whatever in the CONFIGURE_ARGS?