Using dev_kfree_skb for tx skbs breaks AQL. This worked until now only
by accident, because a mac80211 issue breaks AQL on drivers with firmware
rate control that report the rate via ieee80211_tx_status_ext as struct
rate_info.
Signed-off-by: Felix Fietkau <[email protected]>
---
drivers/net/wireless/mediatek/mt76/mt7915/init.c | 8 ++++++--
drivers/net/wireless/mediatek/mt76/mt7915/mac.c | 2 +-
2 files changed, 7 insertions(+), 3 deletions(-)
diff --git a/drivers/net/wireless/mediatek/mt76/mt7915/init.c b/drivers/net/wireless/mediatek/mt76/mt7915/init.c
index aadf56e80bae..d7a3b05ab50c 100644
--- a/drivers/net/wireless/mediatek/mt76/mt7915/init.c
+++ b/drivers/net/wireless/mediatek/mt76/mt7915/init.c
@@ -691,8 +691,12 @@ void mt7915_unregister_device(struct mt7915_dev *dev)
spin_lock_bh(&dev->token_lock);
idr_for_each_entry(&dev->token, txwi, id) {
mt7915_txp_skb_unmap(&dev->mt76, txwi);
- if (txwi->skb)
- dev_kfree_skb_any(txwi->skb);
+ if (txwi->skb) {
+ struct ieee80211_hw *hw;
+
+ hw = mt76_tx_status_get_hw(&dev->mt76, txwi->skb);
+ ieee80211_free_txskb(hw, txwi->skb);
+ }
mt76_put_txwi(&dev->mt76, txwi);
}
spin_unlock_bh(&dev->token_lock);
diff --git a/drivers/net/wireless/mediatek/mt76/mt7915/mac.c b/drivers/net/wireless/mediatek/mt76/mt7915/mac.c
index a264e304a3df..5800b2d1fb23 100644
--- a/drivers/net/wireless/mediatek/mt76/mt7915/mac.c
+++ b/drivers/net/wireless/mediatek/mt76/mt7915/mac.c
@@ -844,7 +844,7 @@ mt7915_tx_complete_status(struct mt76_dev *mdev, struct sk_buff *skb,
if (sta || !(info->flags & IEEE80211_TX_CTL_NO_ACK))
mt7915_tx_status(sta, hw, info, NULL);
- dev_kfree_skb(skb);
+ ieee80211_free_txskb(hw, skb);
}
void mt7915_txp_skb_unmap(struct mt76_dev *dev,
--
2.28.0
Felix Fietkau <[email protected]> writes:
> Using dev_kfree_skb for tx skbs breaks AQL. This worked until now only
> by accident, because a mac80211 issue breaks AQL on drivers with firmware
> rate control that report the rate via ieee80211_tx_status_ext as struct
> rate_info.
Ah, nice catch! Is it safe to assume that you also have a fix pending
for the mac80211 issue you're talking about in the commit message? :)
Acked-by: Toke Høiland-Jørgensen <[email protected]>
-Toke
On 2020-08-12 17:19, Toke Høiland-Jørgensen wrote:
> Felix Fietkau <[email protected]> writes:
>
>> Using dev_kfree_skb for tx skbs breaks AQL. This worked until now only
>> by accident, because a mac80211 issue breaks AQL on drivers with firmware
>> rate control that report the rate via ieee80211_tx_status_ext as struct
>> rate_info.
>
> Ah, nice catch! Is it safe to assume that you also have a fix pending
> for the mac80211 issue you're talking about in the commit message? :)
Sure, I will send it soon.
- Felix
Felix Fietkau <[email protected]> wrote:
> Using dev_kfree_skb for tx skbs breaks AQL. This worked until now only
> by accident, because a mac80211 issue breaks AQL on drivers with firmware
> rate control that report the rate via ieee80211_tx_status_ext as struct
> rate_info.
>
> Signed-off-by: Felix Fietkau <[email protected]>
> Acked-by: Toke Høiland-Jørgensen <[email protected]>
Patch applied to wireless-drivers.git, thanks.
b4be5a53ebf4 mt76: mt7915: use ieee80211_free_txskb to free tx skbs
--
https://patchwork.kernel.org/patch/11710923/
https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches