[v3] ath10k: transmit queued frames after processing rx packets

When running iperf on ath10k SDIO, TX can stop working:

iperf -c 192.168.1.1 -i 1 -t 20 -w 10K
[  3]  0.0- 1.0 sec  2.00 MBytes  16.8 Mbits/sec
[  3]  1.0- 2.0 sec  3.12 MBytes  26.2 Mbits/sec
[  3]  2.0- 3.0 sec  3.25 MBytes  27.3 Mbits/sec
[  3]  3.0- 4.0 sec   655 KBytes  5.36 Mbits/sec
[  3]  4.0- 5.0 sec  0.00 Bytes  0.00 bits/sec
[  3]  5.0- 6.0 sec  0.00 Bytes  0.00 bits/sec
[  3]  6.0- 7.0 sec  0.00 Bytes  0.00 bits/sec
[  3]  7.0- 8.0 sec  0.00 Bytes  0.00 bits/sec
[  3]  8.0- 9.0 sec  0.00 Bytes  0.00 bits/sec
[  3]  9.0-10.0 sec  0.00 Bytes  0.00 bits/sec
[  3]  0.0-10.3 sec  9.01 MBytes  7.32 Mbits/sec

There are frames in the ieee80211_txq and there are frames that have
been removed from from this queue, but haven't yet been sent on the wire
(num_pending_tx).

When num_pending_tx reaches max_num_pending_tx, we will stop the queues
by calling ieee80211_stop_queues().

As frames that have previously been sent for transmission
(num_pending_tx) are completed, we will decrease num_pending_tx and wake
the queues by calling ieee80211_wake_queue(). ieee80211_wake_queue()
does not call wake_tx_queue, so we might still have frames in the
queue at this point.

While the queues were stopped, the socket buffer might have filled up,
and in order for user space to write more, we need to free the frames
in the queue, since they are accounted to the socket. In order to free
them, we first need to transmit them.

This problem cannot be reproduced on low-latency devices, e.g. pci,
since they call ath10k_mac_tx_push_pending() from
ath10k_htt_txrx_compl_task(). ath10k_htt_txrx_compl_task() is not called
on high-latency devices.
Fix the problem by calling ath10k_mac_tx_push_pending(), after
processing rx packets, just like for low-latency devices, also in the
SDIO case. Since we are calling ath10k_mac_tx_push_pending() directly,
we also need to export it.

Signed-off-by: Niklas Cassel <niklas.cassel@linaro.org>
---
Changes since V2:
Moved ath10k_mac_tx_push_pending() call to ath10k_sdio_irq_handler().

 drivers/net/wireless/ath/ath10k/mac.c  | 1 +
 drivers/net/wireless/ath/ath10k/sdio.c | 3 +++
 2 files changed, 4 insertions(+)

On Thu, May 24, 2018 at 12:15:08AM +0200, Niklas Cassel wrote:
> This problem cannot be reproduced on low-latency devices, e.g. pci,
> since they call ath10k_mac_tx_push_pending() from
> ath10k_htt_txrx_compl_task(). ath10k_htt_txrx_compl_task() is not called
> on high-latency devices.
> Fix the problem by calling ath10k_mac_tx_push_pending(), after
> processing rx packets, just like for low-latency devices, also in the
> SDIO case. Since we are calling ath10k_mac_tx_push_pending() directly,
> we also need to export it.
> 

Even if we are now calling ath10k_mac_tx_push_pending each time we process rx packets,
the number of packets we actually queue from ath10k_mac_tx_push_pending are quite few:

From running iperf for 20 seconds:

# grep ath10k_mac_tx_push_txq /sys/kernel/debug/tracing/trace | grep -v ath10k_mac_op_wake_tx_queue | wc -l
233
number of times ath10k_mac_tx_push_txq was called, but not from ath10k_mac_op_wake_tx_queue,
i.e. number of times ath10k_mac_tx_push_txq was called from ath10k_mac_tx_push_pending.


# grep ath10k_mac_tx_push_txq /sys/kernel/debug/tracing/trace | grep -v ath10k_mac_tx_push_pending | wc -l
28415
number of times ath10k_mac_tx_push_txq was called, but not from ath10k_mac_tx_push_pending,
i.e number of times ath10k_mac_tx_push_txq was called from ath10k_mac_op_wake_tx_queue.


Regards,
Niklas

Niklas Cassel <niklas.cassel@linaro.org> wrote:

> When running iperf on ath10k SDIO, TX can stop working:
> 
> iperf -c 192.168.1.1 -i 1 -t 20 -w 10K
> [  3]  0.0- 1.0 sec  2.00 MBytes  16.8 Mbits/sec
> [  3]  1.0- 2.0 sec  3.12 MBytes  26.2 Mbits/sec
> [  3]  2.0- 3.0 sec  3.25 MBytes  27.3 Mbits/sec
> [  3]  3.0- 4.0 sec   655 KBytes  5.36 Mbits/sec
> [  3]  4.0- 5.0 sec  0.00 Bytes  0.00 bits/sec
> [  3]  5.0- 6.0 sec  0.00 Bytes  0.00 bits/sec
> [  3]  6.0- 7.0 sec  0.00 Bytes  0.00 bits/sec
> [  3]  7.0- 8.0 sec  0.00 Bytes  0.00 bits/sec
> [  3]  8.0- 9.0 sec  0.00 Bytes  0.00 bits/sec
> [  3]  9.0-10.0 sec  0.00 Bytes  0.00 bits/sec
> [  3]  0.0-10.3 sec  9.01 MBytes  7.32 Mbits/sec
> 
> There are frames in the ieee80211_txq and there are frames that have
> been removed from from this queue, but haven't yet been sent on the wire
> (num_pending_tx).
> 
> When num_pending_tx reaches max_num_pending_tx, we will stop the queues
> by calling ieee80211_stop_queues().
> 
> As frames that have previously been sent for transmission
> (num_pending_tx) are completed, we will decrease num_pending_tx and wake
> the queues by calling ieee80211_wake_queue(). ieee80211_wake_queue()
> does not call wake_tx_queue, so we might still have frames in the
> queue at this point.
> 
> While the queues were stopped, the socket buffer might have filled up,
> and in order for user space to write more, we need to free the frames
> in the queue, since they are accounted to the socket. In order to free
> them, we first need to transmit them.
> 
> This problem cannot be reproduced on low-latency devices, e.g. pci,
> since they call ath10k_mac_tx_push_pending() from
> ath10k_htt_txrx_compl_task(). ath10k_htt_txrx_compl_task() is not called
> on high-latency devices.
> Fix the problem by calling ath10k_mac_tx_push_pending(), after
> processing rx packets, just like for low-latency devices, also in the
> SDIO case. Since we are calling ath10k_mac_tx_push_pending() directly,
> we also need to export it.
> 
> Signed-off-by: Niklas Cassel <niklas.cassel@linaro.org>
> Signed-off-by: Kalle Valo <kvalo@codeaurora.org>

Patch applied to ath-next branch of ath.git, thanks.

3f04950f32d5 ath10k: transmit queued frames after processing rx packets