#12040 NORM Not Tri: XO-1.5 still does not reliably wake up in aggressive runin

[Zarro Boogs per Child]

#12040: XO-1.5 still does not reliably wake up in aggressive runin
--------------------+-------------------------------------------------------
 Reporter:  dsd     |                 Owner:               
     Type:  defect  |                Status:  new          
 Priority:  normal  |             Milestone:  Not Triaged  
Component:  kernel  |               Version:  not specified
 Keywords:          |           Next_action:  never set    
 Verified:  0       |   Deployment_affected:               
Blockedby:          |              Blocking:               
--------------------+-------------------------------------------------------
 Even with the work done in #11867 we are still seeing that the XO-1.5
 doesn't always wake itself up after a suspend cycle in XO-1.5 runin on
 12.1.0 build 18.

 Two suspicions here:
  1. I remember discussion of a XO-1.5 "blind period" where any wakeups
 received immediately after the system sleeps are ignored because the
 system is not able to wake up while suspend voltages are stabilising (or
 something like that). Does this still exist, could it cause this problem?
  2. On x86-3.3 in suspend_enter(), if the RTC or EC interrupt fires after
 dpm_suspend_noirq() then the system lose the wakeup - it'll suspend
 anyway, and it won't wake up from sleep even though there should be an IRQ
 pending. It's hard to measure how long this period is (from when the
 system goes blind until when it enters suspend and can presumably respond
 to wakeup events) but after doing some printk timings I suspect it is
 about 200ms.

-- 
Ticket URL: <http://dev.laptop.org/ticket/12040>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system