[ath9k-devel] A question (about aggregation) and a hint (about setting MCS "manually")

[ath9k-devel] A question (about aggregation) and a hint (about setting MCS "manually")

[Eduard GV]

Hi list,

First the hint: it is not possible for now to set the MCS from any high
level interface (e.g. iw o debugfs), but it is easy to code in the driver.
Basically give a value between 0 and 15 to rix (variable in
ath_buf_set_rate(), xmit.c).

Now the question: is there a (high-level)way to configure (or just disable)
A-MPDU aggregation? If not, could you point out where in the code should I
start looking?
Please, correct me if I'm wrong: roughly, the max aggregation possible is
chosen provided that 64KB or 4ms frame time is not exceeded (using the
slowest rate in the MRR series)...??

By the way, using A-MPDU, if a sub-frame is received with errors and count >
1 for a rate in MRR, will the hardware retx the whole aggregated frame?

Thank you
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.ath9k.org/pipermail/ath9k-devel/attachments/20110418/f1abde46/attachment.htm 

[ath9k-devel] A question (about aggregation) and a hint (about setting MCS "manually")

[Sujith]

Eduard GV wrote:
> Now the question: is there a (high-level)way to configure (or just disable) A-MPDU aggregation? If not, could you point out where in the code should I start looking?
> Please, correct me if I'm wrong: roughly, the max aggregation possible is chosen provided that 64KB or 4ms frame time is not exceeded (using the slowest rate in the
> MRR series)...??
> 

In ath9k_init_misc(),

sc->sc_flags &= ~SC_OP_TXAGGR
sc->sc_flags &= ~SC_OP_RXAGGR

> By the way, using A-MPDU, if a sub-frame is received with errors and count > 1 for a rate in MRR, will the hardware retx the whole aggregated frame?

Retry for AMPDU subframes is done in the driver.

Sujith