lockdep complains use of uninitialized spinlock at ieee80211_do_stop() [1], for commit f856373e2f31ffd3 ("wifi: mac80211: do not wake queues on a vif that is being stopped") guards clear_bit() using fq.lock even before fq_init() from ieee80211_txq_setup_flows() initializes this spinlock. According to discussion [2], Toke was not happy with expanding usage of fq.lock. Since __ieee80211_wake_txqs() is called under RCU read lock, we can instead use synchronize_rcu() for flushing ieee80211_wake_txqs().
Showing the changes to net/mac80211 (1 of 1 file). Show all
Unified · Split
| @@ -377,9 +377,8 @@ static void ieee80211_do_stop(struct ieee80211_sub_if_data *sdata, bool going_do · 376 unmodified lines | ||
| 377 | 377 | bool cancel_scan; |
| 378 | 378 | struct cfg80211_nan_func *func; |
| 379 | 379 | |
| 380 | − spin_lock_bh(&local->fq.lock); | |
| 381 | 380 | clear_bit(SDATA_STATE_RUNNING, &sdata->state); |
| 382 | − spin_unlock_bh(&local->fq.lock); | |
| 381 | + synchronize_rcu(); /* flush _ieee80211_wake_txqs() */ | |
| 383 | 382 | |
| 384 | 383 | cancel_scan = rcu_access_pointer(local->scan_sdata) == sdata; |
| 385 | 384 | if (cancel_scan) |