Unlike on earlier chips, DMA completion on MT7615 does not imply actually having sent out any packets. Since AQL will prevent filling the hardware queues and will only allow more packets to be passed to the driver after tx completion, it makes much more sense to schedule the tx tasklet there. This is also needed for scheduling tx in testmode support Signed-off-by: Felix Fietkau <nbd@xxxxxxxx> --- drivers/net/wireless/mediatek/mt76/mt7615/dma.c | 4 ---- drivers/net/wireless/mediatek/mt76/mt7615/mac.c | 6 ++++++ 2 files changed, 6 insertions(+), 4 deletions(-) diff --git a/drivers/net/wireless/mediatek/mt76/mt7615/dma.c b/drivers/net/wireless/mediatek/mt76/mt7615/dma.c index e5a965df899a..1231a5ddf9ea 100644 --- a/drivers/net/wireless/mediatek/mt76/mt7615/dma.c +++ b/drivers/net/wireless/mediatek/mt76/mt7615/dma.c @@ -122,10 +122,6 @@ static int mt7615_poll_tx(struct napi_struct *napi, int budget) mt7615_tx_cleanup(dev); - rcu_read_lock(); - mt7615_mac_sta_poll(dev); - rcu_read_unlock(); - tasklet_schedule(&dev->mt76.tx_tasklet); return 0; diff --git a/drivers/net/wireless/mediatek/mt76/mt7615/mac.c b/drivers/net/wireless/mediatek/mt76/mt7615/mac.c index d150fac50c00..1dc291e8b766 100644 --- a/drivers/net/wireless/mediatek/mt76/mt7615/mac.c +++ b/drivers/net/wireless/mediatek/mt76/mt7615/mac.c @@ -1399,6 +1399,12 @@ static void mt7615_mac_tx_free(struct mt7615_dev *dev, struct sk_buff *skb) } dev_kfree_skb(skb); + + rcu_read_lock(); + mt7615_mac_sta_poll(dev); + rcu_read_unlock(); + + tasklet_schedule(&dev->mt76.tx_tasklet); } void mt7615_queue_rx_skb(struct mt76_dev *mdev, enum mt76_rxq_id q, -- 2.24.0
