#5481 NORM Never A: channel setting is weird.

[Zarro Boogs per Child]

#5481: channel setting is weird.
-----------------------+----------------------------------------------------
  Reporter:  dwmw2     |       Owner:  ashish        
      Type:  defect    |      Status:  new           
  Priority:  normal    |   Milestone:  Never Assigned
 Component:  wireless  |     Version:                
Resolution:            |    Keywords:                
  Verified:  0         |  
-----------------------+----------------------------------------------------

Comment(by ashish):

 Replying to [comment:9 dwmw2]:
 > Case 2 would cause the 0x0004 result code I'm about to fix the core
 command queuing to wait a little while and retry, in that case. Would you
 care to suggest the amount of time we should wait before the retry? I was
 thinking 250ms?
 >

 0x0004 result code indicates busy, when firmware is busy with other
 command.
 And should be implemented for all the commands.
 Apart from this, there may be one situation with deauth, where firmware
 receives deauth, start processig it and in the middle of the processing
 goes into exception and returns invalid deauth result as part of deauth
 response reason code (0x1) to tell driver that it could not process deauth
 and driver should retry.
 In this case command response of deauth would be success _not_ 0x0004, but
 deauth reason code will be 1. The reason code I am referring is
 struct cmd_ds_802_11_deauthenticate {
         u8 macaddr[6];
         __le16 reasoncode; // <-- set to 1 by the fw, when it fails to act
 };

 I think 250ms wait is good enough, my only worry is scan which under noisy
 enviorment may take more time. Is it possible to adjust this to more if
 current command is scan? But anyway, I think this should work fine.

 > I was going to add logic so that if we retry the same command for more
 than 5 seconds and it's never accepted, we assume the firmware is crashed
 and toggle the RESET line. Does that sound reasonable?

 Yes, I think it's good to have.

-- 
Ticket URL: <http://dev.laptop.org/ticket/5481#comment:10>
One Laptop Per Child <http://dev.laptop.org>
OLPC bug tracking system