<YSC>
Recently I'm configuring VLAN on my ipq4019 device, and I got strange results. After searching I found that this device doesn't have two CPU net port, the eth1 is a VLAN device
<YSC>
But the default switch config hide the VLAN 2 configuration, this will make user confuse
<YSC>
So I suggest change the default switch configuration to reflect the real VLAN config, what do you think?
<ynezz>
tmn505: there was bunch of changes related to drivers/pci/host/pci-aardvark.c introduced in 4.14.256 and 4.14.257 so I'm not sure if target/linux/mvebu/patches-4.14/524-PCI-aardvark-set-host-and-device-to-the-same-MAX-payload-size.patch is still needed
<ynezz>
it's quite interesting to see such changes in LTS kernel
<slh64>
YSC: look into the WIP dsa PR for ipq40xx for a proper permanent solution. messing around with the configuration wouldn't be a good idea, hardware and swconfig switch drivers are quirky, with hardcoded assignments and serious performance degradation if you don't do it properly
<ynezz>
tmn505: 19.07/target/linux/mvebu/patches-4.14/527-PCI-aardvark-allow-to-specify-link-capability.patch of yours seems to be superseded by 62a3dc9b65a2b24800fc4267b8cf590fad135034 if I'm reading it correctly, right?
Tapper has joined #openwrt-devel
<\x>
YSC: i also have ipq4019, only eth0 will be usable for vlans if youre still on 21.xx swcconfig. you can try to use that pr for dsa on it, it works nicely with vlans and now youll have 5 ports usable instead of 4.
<\x>
its been stable so far for me, I recommend you to try it.
<hanetzer>
so. you know how some openwrt capable devices may have poe out, controlled via gpio perhaps? Are there any services/packages to manage that?
<hanetzer>
I was thinking more for like, that ubiquiti switch, to have a kind of 'dashboard' in luci to manage it.
* enyc
meeps
<enyc>
lots of github PRs and messages unanswered on ML heh hrrm
<enyc>
maybe lots of these only get tidied-up close to a point-release i.e. 21.02.2
* Pepes
says that everyone can help somehow (reviewing, testing it, etc.) instead of complaining. (Team Dark Forces!:P)
* russell--
just tried a build for oxnas pogoplug-v3, for some reason it is building packages/devel/binutils, which fails on libctf. setting CONFIG_PACKAGE_libctf=m seems to fix it.
<ynezz>
hauke: did you found out what is causing the breakage on that rpi4?
<ynezz>
has anyone any clue why there is this strange SDK variable check in some packages like for example `ifneq ($(SDK)$(CONFIG_PACKAGE_freeradius3-mod-sql),)`
<ynezz>
I mean, why the hell do I need to build everything when using SDK?
PaulFertser has quit [Quit: leaving]
pmelange has joined #openwrt-devel
pmelange has left #openwrt-devel [#openwrt-devel]
pmelange1 has joined #openwrt-devel
pmelange1 has left #openwrt-devel [#openwrt-devel]
cmonroe has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
cmonroe has joined #openwrt-devel
cmonroe has quit []
cmonroe has joined #openwrt-devel
pmelange has joined #openwrt-devel
pmelange has left #openwrt-devel [#openwrt-devel]
<ynezz>
enyc: thanks for reporting, fixed
floof58_ has joined #openwrt-devel
floof58 has quit [Read error: Connection reset by peer]
Borromini has quit [Ping timeout: 480 seconds]
<hauke>
ynezz: why do you have so many changes in the target/linux/generic/backport-4.14/273-batman-adv-Convert-packet.h-to-uapi-header.patch patch?
<hauke>
I pushed my kernel update too
<hauke>
to my staging tree
dangole has joined #openwrt-devel
Borromini has joined #openwrt-devel
<ynezz>
hauke: yeah, that patch was weird, created with Git, but Git had problems applying it, so probably somehow hand modified as there was stat diff partially deleted (rename net/batman-adv/packet.h => include/uapi/linux/batadv_packet.h) but that header file was actually included etc.
<ynezz>
and I prefer to use Git instead of quilt, so that's how my git format-patch created that patch, so I've now shuffled the content in order to achieve smallest diff, thanks for spoting it
<hauke>
ynezz: the commit looks good to me, I only tested 4.14.256
Borromini has quit [Quit: Lost terminal]
dangole has quit [Quit: Leaving]
<tmn505>
ynezz: remove also target/linux/mvebu/patches-4.14/528-arm64-dts-armada-3720-espressobin-set-max-link-to-ge.patch, after bump it does nothing, thanks.
<hauke>
tmn505: did you test the patch on a mvebu device?
<ynezz>
tmn505: can you provide some reasoning why this should be removed, I'll add that into commit description for the future
<hexa->
is it possible to evacuate an ap and make all hosts roam elsewhere gracefully?
<hexa->
s/hosts/stations/
<hurricos>
hexa-: you'd need to use existing c libraries for sure
<hurricos>
check out usteer / hostapd. What you're asking for doesn't sound like an existing rfc / standard so probably not something that's implemented and exposed via lua / shell executables
<hurricos>
implemented in c* and exposed*, I mean to say.
<jow>
hexa-: ubus call hostapd.xxx del_client on all assoicated clients should work
<hurricos>
... but what do I know :^)
<jow>
give a sufficiently high ban time, thne shutdown radio/stop hostapd
<hexa->
sounds interesting
<hexa->
hurricos: i'm using usteer already :)
<tmn505>
ynezz: hauke: oh it does, sorry, I didn't look carefuly at those commits. Anyway last time I checked (unfortunately it was with different kernel version) wtih the cards I had it worked properly without downgrading speed. I didn't test with Compex WLE900VX since I didn't have it, so better to keepp the patch, thanks.