[ath9k-devel] Sequence numbers in ath9k tx path

[ath9k-devel] Sequence numbers in ath9k tx path

[David Shwatrz]

Hello,

I have a short question regarding the transmit path:

  I see that the sequence number of the ieee80211 header
(ieee80211_hdr)  is incremented in the driver transmit path:

ath_tx_cabq() in xmit.c:
....
hdr->seq_ctrl |= cpu_to_le16(sc->tx.seq_no);

AFAIK, it is also incremented also  in the tx path of the mlme.c:

in ieee80211_tx_h_sequence(): (net/mac80211/tx.c)
...
if (!ieee80211_is_data_qos(hdr->frame_control)) {
...
		hdr->seq_ctrl = cpu_to_le16(tx->sdata->sequence_number);
		tx->sdata->sequence_number += 0x10;
...

and ieee80211_tx_h_sequence() is called by invoke_tx_handlers() in tx.c.

Why is it so ? or don't we reach both these methods?

Regards,
Andy
....

[ath9k-devel] Sequence numbers in ath9k tx path

[Pavel Roskin]

On Tue, 2009-03-10 at 12:46 +0200, David Shwatrz wrote:
> Hello,
> 
> I have a short question regarding the transmit path:
> 
>   I see that the sequence number of the ieee80211 header
> (ieee80211_hdr)  is incremented in the driver transmit path:
> 
> ath_tx_cabq() in xmit.c:
> ....
> hdr->seq_ctrl |= cpu_to_le16(sc->tx.seq_no);
> 
> AFAIK, it is also incremented also  in the tx path of the mlme.c:

mac80211 fills in the sequence number, but the driver or the hardware
are free to replace it when actually transmitting.  As I understand,
ath9k uses its own sequence numbers.

-- 
Regards,
Pavel Roskin