Send again as txt.
On 13 November 2015 at 12:17, Janusz Dziedzic <[email protected]> wrote:
>
>
>
> On 13 November 2015 at 11:45, Borja Salazar <[email protected]> wrote:
>>
>> Hi,
>>
>> Let me explain the patch more thoroughly. We are testing MCC and we've found some issues with the queues handlers, the main problem is that when we transmit a multicast/broadcast frame, where hw_queue is 8 (CAB), we check queue status, which is 2(q), and if it is full and we have to stop it, we end stopping the wrong queue, 8, which is not full. From this point onwards stations are unable to connect to the AP.
>>
>> Let me know if something is not clear.
>>
> I am interested if you reintroduce bug that Sujith already fix in commit:
> ath9k: Enable HW queue control only for MCC
>
> While as I understand correctly this patch disable hw queues for non-mcc (also clear IEEE80211_HW_QUEUE_CONTROL) and your
>
> + if (ath9k_is_chanctx_enabled())
> + q = fi->txq;
> + else
> + q = info->hw_queue;
>
> Use again hw_queue for standard non-mcc operation.
>
> Please check this, I am sure while did only simple check :)
>
> BR
> Janusz
>
>
>>
>> Regards,
>>
>> Borja Salazar
>> Firmware team
>> All information in this email is confidential
>>
>> On Fri, Nov 13, 2015 at 11:33 AM, Borja Salazar <[email protected]> wrote:
>>>
>>> When channel context is enabled, we could be
>>> stopping/awakening an incorrect queue.
>>>
>>> Signed-off-by: Borja Salazar <[email protected]>
>>> ---
>>> drivers/net/wireless/ath/ath9k/xmit.c | 22 ++++++++++++----------
>>> 1 file changed, 12 insertions(+), 10 deletions(-)
>>>
>>> diff --git a/drivers/net/wireless/ath/ath9k/xmit.c b/drivers/net/wireless/ath/ath9k/xmit.c
>>> index 3e3dac3..9b17a59 100644
>>> --- a/drivers/net/wireless/ath/ath9k/xmit.c
>>> +++ b/drivers/net/wireless/ath/ath9k/xmit.c
>>> @@ -147,7 +147,12 @@ static void ath_txq_skb_done(struct ath_softc *sc, struct ath_txq *txq,
>>> {
>>> struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb);
>>> struct ath_frame_info *fi = get_frame_info(skb);
>>> - int q = fi->txq;
>>> + int q;
>>> +
>>> + if (ath9k_is_chanctx_enabled())
>>> + q = fi->txq;
>>> + else
>>> + q = info->hw_queue;
>>>
>>> if (q < 0)
>>> return;
>>> @@ -158,10 +163,7 @@ static void ath_txq_skb_done(struct ath_softc *sc, struct ath_txq *txq,
>>>
>>> if (txq->stopped &&
>>> txq->pending_frames < sc->tx.txq_max_pending[q]) {
>>> - if (ath9k_is_chanctx_enabled())
>>> - ieee80211_wake_queue(sc->hw, info->hw_queue);
>>> - else
>>> - ieee80211_wake_queue(sc->hw, q);
>>> + ieee80211_wake_queue(sc->hw, q);
>>> txq->stopped = false;
>>> }
>>> }
>>> @@ -2299,17 +2301,17 @@ int ath_tx_start(struct ieee80211_hw *hw, struct sk_buff *skb,
>>> * info are no longer valid (overwritten by the ath_frame_info data.
>>> */
>>>
>>> - q = skb_get_queue_mapping(skb);
>>> + if (ath9k_is_chanctx_enabled())
>>> + q = skb_get_queue_mapping(skb);
>>> + else
>>> + q = info->hw_queue;
>>>
>>> ath_txq_lock(sc, txq);
>>> if (txq == sc->tx.txq_map[q]) {
>>> fi->txq = q;
>>> if (++txq->pending_frames > sc->tx.txq_max_pending[q] &&
>>> !txq->stopped) {
>>> - if (ath9k_is_chanctx_enabled())
>>> - ieee80211_stop_queue(sc->hw, info->hw_queue);
>>> - else
>>> - ieee80211_stop_queue(sc->hw, q);
>>> + ieee80211_stop_queue(sc->hw, q);
>>> txq->stopped = true;
>>> }
>>> }
>>> --
>>> 2.3.6
>>>
>>
>
arggg, you are right, we have it the right way as a patch for openwrt,
I made a mistake generating kernel patch, will send V3
Borja Salazar
Firmware team
All information in this email is confidential
On Fri, Nov 13, 2015 at 12:19 PM, Janusz Dziedzic
<[email protected]> wrote:
> Send again as txt.
>
>
> On 13 November 2015 at 12:17, Janusz Dziedzic <[email protected]> wrote:
>>
>>
>>
>> On 13 November 2015 at 11:45, Borja Salazar <[email protected]> wrote:
>>>
>>> Hi,
>>>
>>> Let me explain the patch more thoroughly. We are testing MCC and we've found some issues with the queues handlers, the main problem is that when we transmit a multicast/broadcast frame, where hw_queue is 8 (CAB), we check queue status, which is 2(q), and if it is full and we have to stop it, we end stopping the wrong queue, 8, which is not full. From this point onwards stations are unable to connect to the AP.
>>>
>>> Let me know if something is not clear.
>>>
>> I am interested if you reintroduce bug that Sujith already fix in commit:
>> ath9k: Enable HW queue control only for MCC
>>
>> While as I understand correctly this patch disable hw queues for non-mcc (also clear IEEE80211_HW_QUEUE_CONTROL) and your
>>
>> + if (ath9k_is_chanctx_enabled())
>> + q = fi->txq;
>> + else
>> + q = info->hw_queue;
>>
>> Use again hw_queue for standard non-mcc operation.
>>
>> Please check this, I am sure while did only simple check :)
>>
>> BR
>> Janusz
>>
>>
>>>
>>> Regards,
>>>
>>> Borja Salazar
>>> Firmware team
>>> All information in this email is confidential
>>>
>>> On Fri, Nov 13, 2015 at 11:33 AM, Borja Salazar <[email protected]> wrote:
>>>>
>>>> When channel context is enabled, we could be
>>>> stopping/awakening an incorrect queue.
>>>>
>>>> Signed-off-by: Borja Salazar <[email protected]>
>>>> ---
>>>> drivers/net/wireless/ath/ath9k/xmit.c | 22 ++++++++++++----------
>>>> 1 file changed, 12 insertions(+), 10 deletions(-)
>>>>
>>>> diff --git a/drivers/net/wireless/ath/ath9k/xmit.c b/drivers/net/wireless/ath/ath9k/xmit.c
>>>> index 3e3dac3..9b17a59 100644
>>>> --- a/drivers/net/wireless/ath/ath9k/xmit.c
>>>> +++ b/drivers/net/wireless/ath/ath9k/xmit.c
>>>> @@ -147,7 +147,12 @@ static void ath_txq_skb_done(struct ath_softc *sc, struct ath_txq *txq,
>>>> {
>>>> struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb);
>>>> struct ath_frame_info *fi = get_frame_info(skb);
>>>> - int q = fi->txq;
>>>> + int q;
>>>> +
>>>> + if (ath9k_is_chanctx_enabled())
>>>> + q = fi->txq;
>>>> + else
>>>> + q = info->hw_queue;
>>>>
>>>> if (q < 0)
>>>> return;
>>>> @@ -158,10 +163,7 @@ static void ath_txq_skb_done(struct ath_softc *sc, struct ath_txq *txq,
>>>>
>>>> if (txq->stopped &&
>>>> txq->pending_frames < sc->tx.txq_max_pending[q]) {
>>>> - if (ath9k_is_chanctx_enabled())
>>>> - ieee80211_wake_queue(sc->hw, info->hw_queue);
>>>> - else
>>>> - ieee80211_wake_queue(sc->hw, q);
>>>> + ieee80211_wake_queue(sc->hw, q);
>>>> txq->stopped = false;
>>>> }
>>>> }
>>>> @@ -2299,17 +2301,17 @@ int ath_tx_start(struct ieee80211_hw *hw, struct sk_buff *skb,
>>>> * info are no longer valid (overwritten by the ath_frame_info data.
>>>> */
>>>>
>>>> - q = skb_get_queue_mapping(skb);
>>>> + if (ath9k_is_chanctx_enabled())
>>>> + q = skb_get_queue_mapping(skb);
>>>> + else
>>>> + q = info->hw_queue;
>>>>
>>>> ath_txq_lock(sc, txq);
>>>> if (txq == sc->tx.txq_map[q]) {
>>>> fi->txq = q;
>>>> if (++txq->pending_frames > sc->tx.txq_max_pending[q] &&
>>>> !txq->stopped) {
>>>> - if (ath9k_is_chanctx_enabled())
>>>> - ieee80211_stop_queue(sc->hw, info->hw_queue);
>>>> - else
>>>> - ieee80211_stop_queue(sc->hw, q);
>>>> + ieee80211_stop_queue(sc->hw, q);
>>>> txq->stopped = true;
>>>> }
>>>> }
>>>> --
>>>> 2.3.6
>>>>
>>>
>>