Hi,
On Wed, 2015-08-26 at 16:54 +0000, Adam Moore wrote:
> In September, there was a discussion on how we should handle a situation
> where an autoconnect fails due to receipt of a connect-failed error, which
> can be seen in difficult RF environments.
>
> I understand the current recovery mechanism where we let wpa_supplicant
> age off BSSs during a sufficiently long interval between autoscans in
> order to prevent too many conneciton attempts to the AP.
>
> There is one situation this does not cover, which is where you do not have
> any favorite networks, and are performing a connection for the first time.
> In this scenario, the user facing application must periodically scan at a
> rate that will prevent BSS age offs, in order to keep a list populated and
> to support net.connman.Service.Connect calls. However, this prevents
> clean instances of the service from being created, and a user that happens
> to receive a connect-failed on their first attempt will be unable to try
> again without restarting connman. Even if the more aggressive scanning is
> removed, subsequent Connect attempts will fail until the failed BSS is
> aged off. Though this problem resolves itself over time with autoconnect,
> it will be a pretty annoying UX problem for a user who is actively trying
> to connect.
>
> Is there interest in an upstream change to address this? Since Connects
> at this stage will be user initiated, I am less worried about the problem
> of getting locked out of an AP.
This should have been fixed with commits
edc02deab0225b3410179c8937e7cfb4058ced7c from May and
026690ff7c5d8a434b26aaedbe70e3580aa5027a from October last year unless
you have found yet another corner case. What was the ConnMan version you
are using?
Cheers,
Patrik
_______________________________________________
connman mailing list
[email protected]
https://lists.connman.net/mailman/listinfo/connman
