On Tue, Nov 06, 2018 at 12:28:11PM +0100, Johannes Berg wrote: > > WARNING: External email to Quantenna Communications. Please exercise caution! > > > Hi, > > On Tue, 2018-11-06 at 10:45 +0000, Sergey Matyukevich wrote: > > > On Mon, 2018-10-22 at 23:25 +0530, Tamizh chelvam wrote: > > > > Add infrastructure for per TID aggregation/retry count configurations > > > > such as retry count and AMPDU aggregation control(disable/enable). > > > > In some scenario reducing the number of retry count for a specific data > > > > traffic can reduce the latency by proceeding with the next packet > > > > instead of retrying the same packet more time. This will be useful > > > > where the next packet can resume the operation without an issue. > > > > > > Not sure I understand this, how can you expect to control something on a > > > per-packet basis using this? > > > > > > Sergey, looks like your A-MPDU control is already in here per RA/TID, > > > and A-MSDU could be added easily? > > > Thanks for pointing me at this patch series. Indeed, it looks like an > > exact match for proper RA/TID aware implementation of AMPDU control. > > AMSDU can be added following the same approach. > > > > Great. I guess if you could take a look that'd be nice, and perhaps also > see if you could actually implement it? Your driver patch seemed to > imply the firmware only has global control, rather than per RA/TID. > > Also, do you think A-MPDU length control would be something useful? > Perhaps that should be instead of enable/disable (since setting length > to 0 or 1 could easily mean "no A-MPDU") Hello Johannes, I will send a follow-up A-MSDU patch after this patch set lands in your tree. And then qtnfmac driver patches for both A-MPDU/A-MSDU changes. As for A-MPDU chain length control, I don't think we have any practical use-case for this feature at the moment. However Ben Greear suggested one possible use-case: to decrease A-MPDU chain length for voice TID in order to decrease latency. Anyways, current A-MPDU patches can be easily adapted to support it if needed. Regards, Sergey
