New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[RFE] IWD support for systemd-networkd #8157
Comments
Is IWD support complete? What remains to be done? |
@cosmojg I guess it hasn't been touched yet. PRs are welcome. Although I have still no idea how this support would look like. The best idea I have is to have a relationship between wireless networks/ESSIDs and the Something like this:
I strongly disregard just implementing iwd core features into systemd-networkd. It should form a healthy relationship, but configuring iwd over systemd-networkd is something that should be discussed. Maybe @poettering himself has some neat ideas for this. The IWD dbus API is still not fixed yet.. so I would propose waiting for a stable iwd release. |
I have already opened an RFE for this: Probably valid not just for IWD but wpa_supplicant too? |
@amishxda should be possible via IWD dbus API. But we would need a new |
I've created #12842 for this RFE. |
@yuwata wont any systemd networkd integraton with iwd conflict with iwd upstream decision of handling iwd configuration on it's own as in the iwd will manage the DHCP/Static IP assignment? Is it not best to be in contact with iwd upstream @holtmann @denkenz etc to ensure that we dont end up with conflicting implementation which ends up misconfiguration the network and or confuse end users and or incorrect network implementations ( which seems to be the case here with BSSID based on Denis responses upstream )? This all boils down with what's the end game here with WIFI support I mean wpa_supplicant is on it's way out in the linux eco system as in it's being replaced with iwd and iwd seems to be intent on being island in the core/baseOS ( read as not part of consolidated core/baseOS platform thus not part of the system/service management framework thou arguably it belongs there ) and wants to manage it's own network configuration so why should there be an implementation of wpa_supplicant/iwd support in any shape or form to begin with, what purpose does it serve and or problem does it solve? |
I think that having all network configuration (wired and wireless) in a single place is plain convenient. Besides it enables things like setting "Metric=" in a coherent way as well as "[Bridge]" without race-condition. I don't think we would differ much from other iwd users like NetworkManager which got their own separate configuration as well. Having said that, it's definitely nice to get in touch with iwd devs to get their input but it shouldn't discourage us from using it the way which is convenient to us regardless of core/baseOS use case. |
I have discussed this whole issue with one of the iwd devs. According to him, there should be no issue. Both ways can co-exist, but it's preferred to let iwd do the ip management, because systemd-networkd has not enough information about roaming etc. systemd-networkd would simply re-do DHCP at reconnect, with ip management in iwd the connection will be more stable (at least from the user point of view). On the other hand the feature is very appreciated for QoS, routing, etc. |
Why was it preferred to let iwd do the ip management? I would think one would probably want to add a switch on either side or both which configures which one is doing the ip management right? ( such "ControlledBy=" configuration switch is needed for more like keepalived etc ) |
it's preferred to let iwd do ip management mostly because of roaming and standards like |
Submission type
Description
IWD, the new Intel Wireless daemon got its first release now[1].
It has full dbus support and other neat features.
I would like to suggest IWD support in
systemd-networkd
.[1] https://www.kernel.org/pub/linux/network/wireless/
The text was updated successfully, but these errors were encountered: