#7973 BLOC 8.2.0 (: libertas unknown event IDs and hang

[Zarro Boogs per Child]

#7973: libertas unknown event IDs and hang
------------------------+---------------------------------------------------
   Reporter:  cjb       |       Owner:  dsaxena             
       Type:  defect    |      Status:  new                 
   Priority:  blocker   |   Milestone:  8.2.0 (was Update.2)
  Component:  kernel    |     Version:  not specified       
 Resolution:            |    Keywords:  blocks?:8.2.0       
Next_action:  diagnose  |    Verified:  0                   
  Blockedby:            |    Blocking:                      
------------------------+---------------------------------------------------

Comment(by ashish):

 Thanks for posting http://dev.laptop.org/attachment/ticket/7973/libertas-
 debug, it really helped me to get the root of this problem.
 From this log, it seems that firmware receives a deauth command from the
 driver followed by association request (command 0x50). Unfortunately,
 driver does not check for deauth reason code of the deauth command
 response.
 Sometimes, firmware may fail to act on deauth due to some reason, e.g., no
 buffer at that moment, in these cases driver assumes successful deauth
 (after getting deauth command response, however, without verifying reason
 code, which would be failure), which is not correct.

 A fix had been introduced in the firmware to handle this corner case,
 however, I doubt if that was tested fully. The reason being that after
 deauth, driver generally sends scan which restores firmware to a valid
 state and afterwards association request just works fine. From the log
 here, there is no scan after deauth and before assoc. It is because of
 this missing scan, association request failed and command timed out.
 The attached firmware version 0.0.0.p18 (based on 5.110.22.p17) fixes code
 which cleans-up firmware state machine during deauth command response,
 irrespective of whether deauth is successful or not. This also adds
 previous debug events, in case if you observe any failure to debug
 further.

 Could you please verify with this firmware, which I believe should work
 fine?
 Thanks for your support.

-- 
Ticket URL: <http://dev.laptop.org/ticket/7973#comment:34>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system