#10142 NORM 1.5-sof: kernel should cache some EC responses

[Zarro Boogs per Child]

#10142: kernel should cache some EC responses
-------------------------+--------------------------------------------------
 Reporter:  pgf          |                 Owner:  pgf                               
     Type:  enhancement  |                Status:  new                               
 Priority:  normal       |             Milestone:  1.5-software-later                
Component:  kernel       |               Version:  Development source as of this date
 Keywords:               |           Next_action:  design                            
 Verified:  0            |   Deployment_affected:                                    
Blockedby:               |              Blocking:                                    
-------------------------+--------------------------------------------------
 currently when asked for information known by the EC, such as battery
 type, or current charge level, the kernel always passes the requests
 directly to the EC.

 some information (e.g. battery type, manufacturer, design parameters) will
 never change, unless the battery is replaced.  the EC notifies the kernel
 when the battery is removed/replaced.

 other information (e.g. voltage, current, capacity) won't change very
 quickly, but isn't good forever.  such information would need to be
 invalidated after some amount of time (tens of ms?) but this could still
 help rate limit EC traffic in the case of repetitive queries, or
 sequential queries which all depend on the same basic information.

 since EC SCI events will likely be lost during suspend, all cached data
 should be invalidated after waking.  since our goal is to suspend
 frequently, it's not clear how much total EC traffic caching would save.
 it might, however, reduce heavy bursts.

-- 
Ticket URL: <http://dev.laptop.org/ticket/10142>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system