site  contact  subhomenews

iwd, alternative to wpa_supplicant

November 18, 2018 — BarryK

Very busy day, heaps of coding...

Have been working on the SFS infrastructure, should have it "finalized" by 0.9.9, fingers crossed.

This morning I was sorting out how to handle when layers change in a container. In the main filesystem, this is handled reasonably well, that is, detection if, say, the devx sfs is inserted or removed. The init script in the initrd does some checking, a bit more after the switch_root, but that is only for the main f.s., not containers. There are things that need to be fixed when the layers change in a container.

This morning also, integrated NetworkManager so that it is part of the other managers (SNS, Frisbee, etc). That is, the old ones are not thrown out, so if you prefer one of those, they are still available.

And, have added ModemManager package, that works with NetworkManager. Also recompiled NetworkManager to support the 'iwd' backend, which is an alternative to wpa_supplicant.

I have had an ongoing problem with wpa_supplcant and my phone hotspot. This is with SNS, and the issue continues with NetworkManager.

wpa-supplcant is a daemon, that negotiates connecting with a wifi hotspot. In the case of my phone, the phone hotspot is recognized, and wpa_supplicant goes into a loop, reporting "Associating", then after a number of tries, gives up. However, if the phone hotspot has been recently turned on, or I turn it off then back on, wpa_supplicant has no trouble with associating.

It seems that the phone goes into a mode from which wpa_supplicant cannot wake it from. Yet, Win10 on my baby laptop has no difficulty doing so.

iwd is the new kid on the block, shaping up to replace wpa_supplicant. There was no recipe to compile iwd in OpenEmbedded, but I did find a very old one that was posted in 2014 (see link in recipe), which was a good starting point. Ended up with this recipe:

The source is downloaded from git, but the last tag is version 0.12.

Will test it tomorrow.

Tags: easy