#12749 LOW Future : mwifiex may receive command response on a data port

[Zarro Boogs per Child]

#12749: mwifiex may receive command response on a data port
--------------------+-------------------------------------------------------
 Reporter:  Quozl   |                 Owner:                          
     Type:  defect  |                Status:  new                     
 Priority:  low     |             Milestone:  Future Release          
Component:  kernel  |               Version:  Software Build 13.2.0-13
 Keywords:          |           Next_action:  design                  
 Verified:  0       |   Deployment_affected:                          
Blockedby:          |              Blocking:                          
--------------------+-------------------------------------------------------
 With the fix to #12694, two new symptoms were exposed in high load
 testing, on two different laptops:

 {{{
 [  784.953356] mwifiex_sdio mmc0:0001:1: wrong aggr pkt: type=1 len=146
 max_len=1792
 }}}

 {{{
 [23388.093697] mwifiex_sdio mmc0:0001:1: wrong aggr pkt: type=1 len=146
 max_len=1792
 }}}

 Occurs about once or twice a day.

 In the code, this shows that during deaggregation, a command response
 packet was found in the data port.  The packet is dropped with this
 warning.

 On the other hand, the Open Firmware driver will accept command responses,
 events, or data packets on any of the ports.

 This may be a leading cause of mwifiex_cmd_timeout_func.  Theory is that
 if a command is sent at about the same time a data packet is received, the
 data packet and the command response may be aggregated in that order, and
 the driver will receive the data packet and not the command response.

-- 
Ticket URL: <http://dev.laptop.org/ticket/12749>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system