xnor_ has quit [Remote host closed the connection]
xnor_ has joined #openwrt-devel
schwicht has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
schwicht has joined #openwrt-devel
xnor_ has quit [Remote host closed the connection]
xnor_ has joined #openwrt-devel
Danct12 has joined #openwrt-devel
sorinello has joined #openwrt-devel
gromero has quit [Remote host closed the connection]
rua has quit [Quit: Leaving.]
rua has joined #openwrt-devel
GNUmoon has quit [Remote host closed the connection]
GNUmoon has joined #openwrt-devel
xnor_ has quit [Remote host closed the connection]
xnor_ has joined #openwrt-devel
ungeskriptet has quit [Remote host closed the connection]
ungeskriptet has joined #openwrt-devel
goliath has joined #openwrt-devel
xnor_ has quit [Remote host closed the connection]
xnor_ has joined #openwrt-devel
valku has quit [Quit: valku]
rua has quit [Remote host closed the connection]
rua has joined #openwrt-devel
__lore__ has joined #openwrt-devel
_lore_ has quit [Ping timeout: 480 seconds]
n3ph has joined #openwrt-devel
nixuser has quit [Read error: Connection reset by peer]
nixuser has joined #openwrt-devel
caskd has quit [Ping timeout: 480 seconds]
xnor_ has quit [Ping timeout: 480 seconds]
caskd has joined #openwrt-devel
grift has quit [Quit: Bye]
grift has joined #openwrt-devel
<trix>
I wonder how useful writing a more "extreme" sysupgrade, possibly target-specific, would be. by that i'm thinking about multiple partitions (A/B), overwriting the other one, then either flipping a switch or overwriting the bootloader last (so that the likelyhood of brick is narrowed down). assuming size isn't a concern
<trix>
i'm using OpenWrt at work and it is something that is considered, but it is more out of fear of something happening during the sysupgrade process, as it would nuke the device and those devices are upgrading unattended out in the wild.
digitalcircuit has quit [Quit: Signing off from Quassel - see ya!]
<KanjiMonster>
trix: A/B partition is already handled by some platforms/devices (e.g. mvebu/cortexa9 linksys devices)
<trix>
KanjiMonster: thanks, this is quite helpful, digging into it
<trix>
it seems like openwrt wraps around and merely uses the pre-existing A/B partitioing system for those linksys devices, adding custom openwrt support for it
<trix>
this is still very helpful, as it shows how those devices do it, and how it could be approached for other devices, im just worried about the "initial setup"
<Rondom>
trix: we also have some support for it downstream. I think one thing that some implementations in OpenWrt miss is the distinction between, currently running image and image that is set to boot. It would be good to add hooks (probably in platform.sh) to determine and set the "bank" to be booted next and to determine the currently running bank
<Rondom>
sysupgrade already has support for specifying current vs other. it is up to platform.sh scripts to make use of it
ungeskriptet_ has joined #openwrt-devel
ungeskriptet has quit [Ping timeout: 480 seconds]
mrkiko has quit [Quit: [TLS] Client upgrade]
mrkiko has joined #openwrt-devel
<trix>
Rondom: thanks for the info, it sounds promising, is this specific to those mvebu/cortexa9 linksys devices, or more generic? any pointers/links to source code or other?
<Rondom>
there is a parameter for sysupgrade which is passed on by procd as an environment variable to do_stage2 / platform.sh-do_upgrade
<Rondom>
if I remember things correctly. I have not had the time to look at it in detail
<Rondom>
but I remember that basic building blocks are there
<Rondom>
then bootcount-init-script needs to be aware to permanently set the currently booted bank (not the back currently set to boot, the two might be different in some edge cases)
<mrkiko>
I am going around trying to setup mptcp. But can't find where is the coe show reads multipath UCi config and sets things up. I am reading trough https://openwrt.org/docs/guide-user/network/mptcp
<mrkiko>
but can't see the code on-device or on git repo, including netifd
<mrkiko>
efahl: yeah, that's the daemon and I actually installed it, even tough I am not sure I do really need it. but even in that package, I can't see the code actually parsing the config
<mrkiko>
efahl: I mean - if I follow the wiki article in mycurrent situation, I guess my changes won't have any effect
<efahl>
well, all the config file stuff in that wiki article is in the openvpn config file, so openvpn must be the thing parsing it and generating the .conf?