This warning can trigger if there is a mismatch between frames that were
sent with the sta pointer set vs tx status frames reported for the sta address.
This can happen due to race conditions on re-creating stations, or even
in the case of .sta_add/remove being used instead of .sta_state, which can cause
frames to be sent to a station that has not been uploaded yet.
If there is an actual underflow issue, it should show up in the device airtime
warning below, so it is better to remove this one.
Signed-off-by: Felix Fietkau <[email protected]>
---
net/mac80211/sta_info.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/net/mac80211/sta_info.c b/net/mac80211/sta_info.c
index cd8487bc6fc2..af4cc5fb678e 100644
--- a/net/mac80211/sta_info.c
+++ b/net/mac80211/sta_info.c
@@ -1923,9 +1923,7 @@ void ieee80211_sta_update_pending_airtime(struct ieee80211_local *local,
if (sta) {
tx_pending = atomic_sub_return(tx_airtime,
&sta->airtime[ac].aql_tx_pending);
- if (WARN_ONCE(tx_pending < 0,
- "STA %pM AC %d txq pending airtime underflow: %u, %u",
- sta->addr, ac, tx_pending, tx_airtime))
+ if (tx_pending < 0)
atomic_cmpxchg(&sta->airtime[ac].aql_tx_pending,
tx_pending, 0);
}
--
2.24.0
Felix Fietkau <[email protected]> writes:
> This warning can trigger if there is a mismatch between frames that were
> sent with the sta pointer set vs tx status frames reported for the sta address.
>
> This can happen due to race conditions on re-creating stations, or even
> in the case of .sta_add/remove being used instead of .sta_state, which can cause
> frames to be sent to a station that has not been uploaded yet.
>
> If there is an actual underflow issue, it should show up in the device airtime
> warning below, so it is better to remove this one.
Fair enough.
Acked-by: Toke Høiland-Jørgensen <[email protected]>