Re: Commit 0711d638 breaks mwifiex

[Johannes Berg <johannes@sipsolutions.net>]

Hi,

> While working on an issue that marvell module stops connecting to AP,
> bisect reveals that the issue starts to happen from commit 0711d638,
> which uses wdev->ssid_len instead of wdev->current_bss to determine
> if driver's .disconnect() should be called.
> 
> It happens because mwifiex_cfg80211_connect() returns -EALREADY
> when it finds wdev->current_bss is valid:
> 
> if (priv->wdev.current_bss) {
>     [PRINT LOG]
>     return -EALREADY;
> }
> 
> This would make cfg80211_connect() set wdev->ssid_len to 0, and thus
> mwifiex_cfg80211_disconnect() won't be called by
> cfg80211_disconnect().

Hmm, none of this makes much sense to me right now.

Does mwifiex treat this -EALREADY as *keeping* an old connection, or
tearing it down entirely?

Because right now clearly cfg80211 assumes, on the one hand, that no
connection is kept (resetting ssid_len), but on the other hand it got
here with current_bss set - so perhaps we should reject that before in
cfg80211, rather than in mwifiex?

I think your fix is invalid because we then reset ssid_len and still
keep an old connection (current_bss) which will lead to strange nl80211
behaviour when getting interface data etc.

johannes