One of the code paths sending deauth/disassoc events ends up calling
this function with rcu_read_lock held, so we must use GFP_ATOMIC in
allocation routines.
Reported-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Jouni Malinen <j@w1.fi>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
        struct sk_buff *msg;
        void *hdr;
 
-       msg = nlmsg_new(NLMSG_GOODSIZE, GFP_KERNEL);
+       msg = nlmsg_new(NLMSG_GOODSIZE, GFP_ATOMIC);
        if (!msg)
                return;
 
                return;
        }
 
-       genlmsg_multicast(msg, 0, nl80211_mlme_mcgrp.id, GFP_KERNEL);
+       genlmsg_multicast(msg, 0, nl80211_mlme_mcgrp.id, GFP_ATOMIC);
        return;
 
  nla_put_failure: