#10440 HIGH Not Tri: EC firmware pq2e35 for XO-1 leaves power LED on during suspend

[Zarro Boogs per Child]

#10440: EC firmware pq2e35 for XO-1 leaves power LED on during suspend
-------------------------------------------+--------------------------------
           Reporter:  pgf                  |       Owner:  rsmith                                 
               Type:  defect               |      Status:  new                                    
           Priority:  high                 |   Milestone:  Not Triaged                            
          Component:  embedded controller  |     Version:  1.5/1.0 Software Build os852 aka 10.1.2
         Resolution:                       |    Keywords:                                         
        Next_action:  never set            |    Verified:  0                                      
Deployment_affected:                       |   Blockedby:                                         
           Blocking:                       |  
-------------------------------------------+--------------------------------

Comment(by pgf):

 turns out pq2e44 (which was bundled in OSF Q2E41) will do the same thing
 in response to an OFW 's' command, so it's not a regression.

 the issue is that there's a state variable that's checked in the EC before
 it will go into its "idle" state, which is when it would turn off the LED.
 the state variable isn't set automatically when coming out of suspend, the
 way it should be.  it _is_ set by various other event handlers in the EC,
 so it turns out that if you're _using_ the laptop, it's difficult to
 trigger this condition unless you're talking to the system via a serial
 console (which i do a lot).

 if you're _not_ using the laptop, however, it's easy to reproduce.  simply
 boot the laptop and let it idle suspend without touching it.  (at least,
 this does it for me on q2e41 and os852.)

 even with automatic power management enabled on 802, i can't reproduce the
 symptom.  something ohmd does while suspending or resuming "kicks" the EC
 the right way.  i'll investigate what might be different.

-- 
Ticket URL: <http://dev.laptop.org/ticket/10440#comment:1>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system