[PATCH] ath10k: Add peer flush in ath10k_flush for STATION

[PATCH] ath10k: Add peer flush in ath10k_flush for STATION

[Wen Gong]

In the noisy environment, if there are packets in the queue and can't
send out, the suspend timing will be more than 5 seconds due to the wait,
flush the queue to optimize the suspend timing, and let the upper layer to
retry the packets after resume.

Tested with QCA6174 PCI with firmware
WLAN.RM.4.4.1-00109-QCARMSWPZ-1, but this will also affect QCA9377 PCI.
It's not a regression with new firmware releases.

Signed-off-by: Wen Gong <wgong@codeaurora.org>
 drivers/net/wireless/ath/ath10k/mac.c | 15 +++++++++++++--
 1 file changed, 13 insertions(+), 2 deletions(-)

diff --git a/drivers/net/wireless/ath/ath10k/mac.c b/drivers/net/wireless/ath/ath10k/mac.c
index 97548f9..8df3cec 100644
--- a/drivers/net/wireless/ath/ath10k/mac.c
+++ b/drivers/net/wireless/ath/ath10k/mac.c
@@ -6814,9 +6814,20 @@ static void ath10k_flush(struct ieee80211_hw *hw, struct ieee80211_vif *vif,
 			 u32 queues, bool drop)
 {
 	struct ath10k *ar = hw->priv;
-
-	if (drop)
+	struct ath10k_vif *arvif;
+	u32 bitmap;
+
+	if (drop) {
+		if (vif->type == NL80211_IFTYPE_STATION) {
+			bitmap = ~(1 << WMI_MGMT_TID);
+			list_for_each_entry(arvif, &ar->arvifs, list) {
+				if (arvif->vdev_type == WMI_VDEV_TYPE_STA)
+					ath10k_wmi_peer_flush(ar, arvif->vdev_id,
+							      arvif->bssid, bitmap);
+			}
+		}
 		return;
+	}
 
 	mutex_lock(&ar->conf_mutex);
 	ath10k_mac_wait_tx_complete(ar);
-- 
1.9.1

Re: [PATCH] ath10k: Add peer flush in ath10k_flush for STATION

[Kalle Valo]

Wen Gong <wgong@codeaurora.org> writes:

> In the noisy environment, if there are packets in the queue and can't
> send out, the suspend timing will be more than 5 seconds due to the wait,
> flush the queue to optimize the suspend timing, and let the upper layer to
> retry the packets after resume.
>
> Tested with QCA6174 PCI with firmware
> WLAN.RM.4.4.1-00109-QCARMSWPZ-1, but this will also affect QCA9377 PCI.
> It's not a regression with new firmware releases.
>
> Signed-off-by: Wen Gong <wgong@codeaurora.org>
>  drivers/net/wireless/ath/ath10k/mac.c | 15 +++++++++++++--
>  1 file changed, 13 insertions(+), 2 deletions(-)

The patch is slightly corrupted, there should be a "---" line before the
diffstat and that broke my patchwork script. I fixed it now in the
pending branch, but be careful if you manually edit patch files.

-- 
Kalle Valo

Re: [PATCH] ath10k: Add peer flush in ath10k_flush for STATION

[Kalle Valo]

Wen Gong <wgong@codeaurora.org> wrote:

> In the noisy environment, if there are packets in the queue and can't
> send out, the suspend timing will be more than 5 seconds due to the wait,
> flush the queue to optimize the suspend timing, and let the upper layer to
> retry the packets after resume.
> 
> Tested with QCA6174 PCI with firmware
> WLAN.RM.4.4.1-00109-QCARMSWPZ-1, but this will also affect QCA9377 PCI.
> It's not a regression with new firmware releases.
> 
> Signed-off-by: Wen Gong <wgong@codeaurora.org>
> Signed-off-by: Kalle Valo <kvalo@codeaurora.org>

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

9de4162f0999 ath10k: add peer flush in ath10k_flush for STATION

-- 
https://patchwork.kernel.org/patch/10630271/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches

Re: [PATCH] ath10k: Add peer flush in ath10k_flush for STATION

[Brian Norris]

On Mon, Oct 8, 2018 at 2:03 AM Wen Gong <wgong@codeaurora.org> wrote:
>
> In the noisy environment, if there are packets in the queue and can't
> send out, the suspend timing will be more than 5 seconds due to the wait,
> flush the queue to optimize the suspend timing, and let the upper layer to
> retry the packets after resume.
>
> Tested with QCA6174 PCI with firmware
> WLAN.RM.4.4.1-00109-QCARMSWPZ-1, but this will also affect QCA9377 PCI.
> It's not a regression with new firmware releases.
>
> Signed-off-by: Wen Gong <wgong@codeaurora.org>
>  drivers/net/wireless/ath/ath10k/mac.c | 15 +++++++++++++--
>  1 file changed, 13 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/net/wireless/ath/ath10k/mac.c b/drivers/net/wireless/ath/ath10k/mac.c
> index 97548f9..8df3cec 100644
> --- a/drivers/net/wireless/ath/ath10k/mac.c
> +++ b/drivers/net/wireless/ath/ath10k/mac.c
> @@ -6814,9 +6814,20 @@ static void ath10k_flush(struct ieee80211_hw *hw, struct ieee80211_vif *vif,
>                          u32 queues, bool drop)
>  {
>         struct ath10k *ar = hw->priv;
> -
> -       if (drop)
> +       struct ath10k_vif *arvif;
> +       u32 bitmap;
> +
> +       if (drop) {
> +               if (vif->type == NL80211_IFTYPE_STATION) {

'vif' may be NULL. It is trivially easy for me to crash a system on
v4.20-rc1 by suspending the system while not connected.

I'll send a patch to fix this.

Brian

> +                       bitmap = ~(1 << WMI_MGMT_TID);
> +                       list_for_each_entry(arvif, &ar->arvifs, list) {
> +                               if (arvif->vdev_type == WMI_VDEV_TYPE_STA)
> +                                       ath10k_wmi_peer_flush(ar, arvif->vdev_id,
> +                                                             arvif->bssid, bitmap);
> +                       }
> +               }
>                 return;
> +       }
>
>         mutex_lock(&ar->conf_mutex);
>         ath10k_mac_wait_tx_complete(ar);
> --
> 1.9.1
>