Hi all,

After merging the wireless-next tree, today's linux-next build
(arm64_defconfig.log /tmp/next/x86_64 allmodconfig) failed like this:

/tmp/next/build/net/mac80211/mlme.c:
In function 'ieee80211_rx_mgmt_assoc_resp':
/tmp/next/build/net/mac80211/mlme.c:7279:17: error: label 'abandon_assoc'
used but not defined
 7279 |                 goto abandon_assoc;
      |                 ^~~~

Caused by a semantic conflict (probably the same two patches as the
merge I posted a minute ago).  I have applied the patch below and can
carry as needed:

diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c
index 60fc407bb1ebb..5a1c022282202 100644
--- a/net/mac80211/mlme.c
+++ b/net/mac80211/mlme.c
@@ -7276,7 +7276,7 @@ static void ieee80211_rx_mgmt_assoc_resp(struct ieee80211_sub_if_data *sdata,
 	else if (!assoc_data->s1g)
 		aid = le16_to_cpu(mgmt->u.assoc_resp.aid);
 	else if (status_code == WLAN_STATUS_SUCCESS)
-		goto abandon_assoc;
+		goto notify_driver;
 
 	/*
 	 * The 5 MSB of the AID field are reserved for a non-S1G STA. For

Hi,

Thanks for the heads-up!

> /tmp/next/build/net/mac80211/mlme.c:
> In function 'ieee80211_rx_mgmt_assoc_resp':
> /tmp/next/build/net/mac80211/mlme.c:7279:17: error: label 'abandon_assoc'
> used but not defined
>  7279 |                 goto abandon_assoc;
>       |                 ^~~~
> 
> Caused by a semantic conflict (probably the same two patches as the
> merge I posted a minute ago).

Yeah. I was _just_ going to reply to that email asking if it actually
built! :-)

> I have applied the patch below and can
> carry as needed:
> 
> diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c
> index 60fc407bb1ebb..5a1c022282202 100644
> --- a/net/mac80211/mlme.c
> +++ b/net/mac80211/mlme.c
> @@ -7276,7 +7276,7 @@ static void ieee80211_rx_mgmt_assoc_resp(struct ieee80211_sub_if_data *sdata,
>  	else if (!assoc_data->s1g)
>  		aid = le16_to_cpu(mgmt->u.assoc_resp.aid);
>  	else if (status_code == WLAN_STATUS_SUCCESS)
> -		goto abandon_assoc;
> +		goto notify_driver;

I think it needs to be "goto destroy_assoc_data" instead for
correctness, but that's just based on what Jeff told me last night.

Sorry for the mess! I didn't realize this was going to happen :(

Anyway, the plan is for me to do a (hopefully correct) merge after
wireless will is merged into net etc., test it, and then provide that
with the pull request of wireless-next into net-next. I think it'll take
a few days though, more likely early next week than late this week.

johannes