IWD 1.0 Released As Intel's Wireless Daemon For Linux Systems

[Guest]

Hey so is it stable enough for a daily use ? This is good to evolve to other thing than the old wpa_supplicant, I hope this will do the job !

I've been using it for several months now on my laptop with an Intel WiFi card. It used to be somewhat unstable and unreliable on earlier versions, but recently it's been working quite well. It does connect significantly faster than wpa_supplicant (after cold boot and even after suspend and resume).

[Guest]

Interesting. I was experimenting with it over the last year or so but it never seemed to work. And I am not even speaking about WPA Enterprise or eduroam but even for my home wifi. Actually tried that like a month ago, it saw the network, connected to it, everything seemed fine but after roughly 10 minutes I had no internet while iwd still claimed the wifi was up.

It must be something in my system, is anyone using iwd on a Lenovo T530 or maybe with dbus-broker? Just so I can figure out if I am looking at a hardware issue or software issue.

If you're using NetworkManager, you'll have to add a config option to get it to use iwd instead of wpa_supplicant - also, NetworkManager or something else seems to have a hard dependency on wpa_supplicant, so disabling it is not enough, you'll have to mask it to prevent interference with iwd.

[Guest]

is there any reason the improvements can't be carried over to wpa_supplicant, instead of making a new thing? reading the other comments here I can tell this'd be a catastrophe if deployed, and I'm sure many aren't sympathetic to development behind closed doors then opening up, especially as wpa_supplicant is just fine

It was developed in the open.

[guildem]

Should be fixed with 1.0-3

Not fixed for me on 1.0-3, did it work for you ?

Never mind, was my fault, works like a charm !

[polarathene]

You really should watch the posted video.

Which posted video? If I click phoronix links in the article, I go back two articles to this one, but the video is no longer available:

Oh, this one in the comments on page 2? That's 44 mins long, I don't have time for that atm with current workload. Is there a specific part you wanted to point me to?

iwd is supposed to handle wifi more efficiently, exactly as you said. For one, scans for networks will not interrupt your network anymore, reconnecting to a wifi network after suspend will be way faster etc.

I wonder if that'd potentially avoid the problem with one of my wifi devices using the MT76 driver. Recently tried using for a day with both 4.19 and 5.3 kernels(it's had mainline support since 4.17), and it was consistently dropping the connection within 90-120 mins of use(downloading steam game at 1MB/s which is the max provided by current ISP, so nothing wrong with the speed).

When the connection fails(UI describes it as "limited connectivity", and internet stops working), it cannot be disconnected. I just get errors logged like "kernel usb 1-2: vendor request req:07 off:1798 failed:-110".. Arch Wiki suggests that some wifi chipsets/drivers don't respond well to scanning for networks while a connection is already established, so it could be due to that, or some other bug in the driver(but I'd think it'd have been discovered and resolved since 4.17?).

Heck, of of there problems atm. is that iwd is so fast that udev does not manage to rename your wifi device during boot up.

Link to the issue for tracking that? That'd be an issue for me as I use udev to assign a custom name for my USB wifi adapters.

[pkunk]

https://iwd.wiki.kernel.org/interface_lifecycle

[polarathene]

If you're using NetworkManager, you'll have to add a config option to get it to use iwd instead of wpa_supplicant - also, NetworkManager or something else seems to have a hard dependency on wpa_supplicant, so disabling it is not enough, you'll have to mask it to prevent interference with iwd.

Information on disabling/masking wpa_supplicant?(systemd unit?) Arch Wiki states you only need to add the config:

https://wiki.archlinux.org/index.php/NetworkManager#Using_iwd_as_the_Wi-Fi_backend

There's also this Gitlab issue with NetworkManager which suggests due to some changes with IWD, NetworkManager requires a patch atm, until that is upstreamed in a future release:

https://gitlab.freedesktop.org/NetworkManager/NetworkManager/issues/101

[Space Heater]

I'm sure many aren't sympathetic to development behind closed doors then opening up, especially as wpa_supplicant is just fine

What's with these low-effort trolls and spreading misinformation on phoronix?

Their repository at https://git.kernel.org/pub/scm/netwo...eless/iwd.git/ and mailing list at https://lists.01.org/hyperkitty/list/[email protected]/ have always been out in the open.

[mikelpr]

What's with these low-effort trolls and spreading misinformation on phoronix?

Their repository at https://git.kernel.org/pub/scm/netwo...eless/iwd.git/ and mailing list at https://lists.01.org/hyperkitty/list/[email protected]/ have always been out in the open.

well that was definitely not my intention and kept asking for more information but maybe I should have instead googled more. anyway, thanks for the links

[Guest]

Information on disabling/masking wpa_supplicant?(systemd unit?) Arch Wiki states you only need to add the config:

https://wiki.archlinux.org/index.php..._Wi-Fi_backend

There's also this Gitlab issue with NetworkManager which suggests due to some changes with IWD, NetworkManager requires a patch atm, until that is upstreamed in a future release:

https://gitlab.freedesktop.org/Netwo...ger/issues/101

I just found that wpa_supplicant would start up even if I disabled it, so I masked it.

I've been using NetworkManager with iwd for several months now, no problems here - I haven't installed 1.0 yet, looks like that's what the latest comments refer to and that's what the patch is for.