2014-02-25 06:57:17

by Janusz Dziedzic

Subject: [PATCH] ath10k: skip management frames in HTT path

After we pass frames with INV_PEER to upper layer in patch:

ath10k: pass frames with invalid peer status to upper layer

we could pass some management frames (in case INVALID_PEER
and MGMT_CTRL) twice to upper layer, once via WMI and once
via HTT. Next we could handle assoc request twice. This
patch remove such regression.

Signed-off-by: Janusz Dziedzic <[email protected]>
---
drivers/net/wireless/ath/ath10k/htt_rx.c | 17 ++++++++++++++++-
1 file changed, 16 insertions(+), 1 deletion(-)

diff --git a/drivers/net/wireless/ath/ath10k/htt_rx.c b/drivers/net/wireless/ath/ath10k/htt_rx.c
index c7a5822..316433f 100644
--- a/drivers/net/wireless/ath/ath10k/htt_rx.c
+++ b/drivers/net/wireless/ath/ath10k/htt_rx.c
@@ -852,6 +852,20 @@ static bool ath10k_htt_rx_has_mic_err(struct sk_buff *skb)
return false;
}

+static bool ath10k_htt_rx_is_mgmt(struct sk_buff *skb)
+{
+ struct htt_rx_desc *rxd;
+ u32 flags;
+
+ rxd = (void *)skb->data - sizeof(*rxd);
+ flags = __le32_to_cpu(rxd->attention.flags);
+
+ if (flags & RX_ATTENTION_FLAGS_MGMT_TYPE)
+ return true;
+
+ return false;
+}
+
static int ath10k_htt_rx_get_csum_state(struct sk_buff *skb)
{
struct htt_rx_desc *rxd;
@@ -944,7 +958,8 @@ static void ath10k_htt_rx_handler(struct ath10k_htt *htt,
status = info.status;

/* Skip mgmt frames while we handle this in WMI */
- if (status == HTT_RX_IND_MPDU_STATUS_MGMT_CTRL) {
+ if (status == HTT_RX_IND_MPDU_STATUS_MGMT_CTRL ||
+ ath10k_htt_rx_is_mgmt(msdu_head)) {
ath10k_htt_rx_free_msdu_chain(msdu_head);
continue;
}
--
1.7.9.5

2014-02-26 17:16:06

by Kalle Valo

Subject: Re: [PATCH] ath10k: skip management frames in HTT path

Janusz Dziedzic <[email protected]> writes:

> After we pass frames with INV_PEER to upper layer in patch:
>
> ath10k: pass frames with invalid peer status to upper layer
>
> we could pass some management frames (in case INVALID_PEER
> and MGMT_CTRL) twice to upper layer, once via WMI and once
> via HTT. Next we could handle assoc request twice. This
> patch remove such regression.
>
> Signed-off-by: Janusz Dziedzic <[email protected]>

I saw a conflict and also did some commit log cleanup. Please check:

https://github.com/kvalo/ath/commit/dd0063fba4434d22fb8a6e6d79bf23cc36f6916a

--
Kalle Valo

2014-02-27 16:42:42

by Kalle Valo

Subject: Re: [PATCH] ath10k: skip management frames in HTT path

Janusz Dziedzic <[email protected]> writes:

> After we pass frames with INV_PEER to upper layer in patch:
>
> ath10k: pass frames with invalid peer status to upper layer
>
> we could pass some management frames (in case INVALID_PEER
> and MGMT_CTRL) twice to upper layer, once via WMI and once
> via HTT. Next we could handle assoc request twice. This
> patch remove such regression.
>
> Signed-off-by: Janusz Dziedzic <[email protected]>

Thanks, applied.

--
Kalle Valo

2014-02-27 06:52:56

by Janusz Dziedzic

Subject: Re: [PATCH] ath10k: skip management frames in HTT path

On 26 February 2014 18:16, Kalle Valo <[email protected]> wrote:
> Janusz Dziedzic <[email protected]> writes:
>
>> After we pass frames with INV_PEER to upper layer in patch:
>>
>> ath10k: pass frames with invalid peer status to upper layer
>>
>> we could pass some management frames (in case INVALID_PEER
>> and MGMT_CTRL) twice to upper layer, once via WMI and once
>> via HTT. Next we could handle assoc request twice. This
>> patch remove such regression.
>>
>> Signed-off-by: Janusz Dziedzic <[email protected]>
>
> I saw a conflict and also did some commit log cleanup. Please check:
>
> https://github.com/kvalo/ath/commit/dd0063fba4434d22fb8a6e6d79bf23cc36f6916a
>
looks good.

BR
Janusz