Re: [PATCH RFC] wifi: mac80211: reject station association if AP is not started

On lip 01, 2026 09:52, syzbot wrote:
> If an interface is changed to AP mode but not started, its channel context
> configuration (chanctx_conf) remains NULL. If a station is then added to
> this interface, the kernel may automatically set the
> NL80211_STA_FLAG_ASSOCIATED flag for compatibility with older userspace
> applications.
> 
> When this flag is set, sta_apply_auth_flags() attempts to initialize rate
> control for the station by calling rate_control_rate_init_all_links(). This
> eventually leads to rate_control_rate_init(), which dereferences the NULL
> chanctx_conf, triggering a WARN_ON:
> 
> WARNING: net/mac80211/rate.c:51 at rate_control_rate_init+0x5a6/0x630
> ...
> Call Trace:
>  rate_control_rate_init_all_links+0xf4/0x190 net/mac80211/rate.c:84
>  sta_apply_auth_flags+0x1bc/0x430 net/mac80211/cfg.c:2152
>  sta_apply_parameters+0x126d/0x1b10 net/mac80211/cfg.c:2618
>  ieee80211_add_station+0x3de/0x700 net/mac80211/cfg.c:2684
>  rdev_add_station+0xfc/0x290 net/wireless/rdev-ops.h:201
>  nl80211_new_station+0x1b4e/0x1fd0 net/wireless/nl80211.c:9505
> 
> Fix this by rejecting the addition or modification of a station to the
> associated state if the AP has not been started (chanctx_conf is NULL).
> Exempt Multi-Link Operation (MLO) interfaces from this check, as they
> handle chanctx_conf per-link rather than globally on the VIF, and
> rate_control_rate_init() already handles them correctly.

Well, the rate_control_rate_init() isn't really handling the MLO devices yet:

	/* SW rate control isn't supported with MLO right now */
	if (WARN_ON(ieee80211_vif_is_mld(&sta->sdata->vif)))
		return;

I think we can tackle this some other time, but maybe at the moment it is wise to remove the "and
rate_control_rate_init() already handles them correctly" from commit message?

> Fixes: 55de908ab292 ("mac80211: use channel contexts")
> Assisted-by: Gemini:gemini-3.1-pro-preview best-expensive syzbot
> Reported-by: [email protected]
> Closes: https://syzkaller.appspot.com/bug?extid=9bdc0c5998ab45b05030
> Link: https://syzkaller.appspot.com/ai_job?id=86851195-2aa0-410f-9fc9-952d95ae035d
> To: "Johannes Berg" <[email protected]>
> To: <[email protected]>
> Cc: <[email protected]>
> 
> ---
> diff --git a/net/mac80211/cfg.c b/net/mac80211/cfg.c
> index 3b58af59f..b45d7923c 100644
> --- a/net/mac80211/cfg.c
> +++ b/net/mac80211/cfg.c
> @@ -2143,6 +2143,10 @@ static int sta_apply_auth_flags(struct ieee80211_local *local,
>  	if (mask & BIT(NL80211_STA_FLAG_ASSOCIATED) &&
>  	    set & BIT(NL80211_STA_FLAG_ASSOCIATED) &&
>  	    !test_sta_flag(sta, WLAN_STA_ASSOC)) {
> +		if (!ieee80211_vif_is_mld(&sta->sdata->vif) &&
> +		    !rcu_access_pointer(sta->sdata->vif.bss_conf.chanctx_conf))
> +			return -EINVAL;
> +

Could you move this check to the if body below (after the comment) where the WLAN_STA_RATE_CONTROL
flag is checked for? I think having it there, right before calling the
rate_control_rate_init_all_links(sta); is more readable and done only when the WLAN_STA_RATE_CONTROL
is not set.

>  		/*
>  		 * When peer becomes associated, init rate control as
>  		 * well. Some drivers require rate control initialized
> 
> 
> base-commit: dc59e4fea9d83f03bad6bddf3fa2e52491777482

-- 
Slawomir Stepien