#6010 HIGH Update.: The EC SCI mask should be set in "sleep" mode.

[Zarro Boogs per Child]

#6010: The EC SCI mask should be set in "sleep" mode.
---------------------+------------------------------------------------------
  Reporter:  cjb     |       Owner:  dsaxena         
      Type:  defect  |      Status:  new             
  Priority:  high    |   Milestone:  Update.2 (8.2.0)
 Component:  kernel  |     Version:                  
Resolution:          |    Keywords:                  
  Verified:  0       |    Blocking:  5974            
 Blockedby:          |  
---------------------+------------------------------------------------------

Comment(by dsaxena):

 Replying to [comment:4 cjb]:
 > Seems a little obtuse, since it sounds like userspace (OHM) will end up
 storing a mask and then synthesizing it into write-one-file-at-a-time, but
 I'll take what I can get.  :)
 >

 I agree, and my first thought was to go with that option too, but there
 are advantages to having it be one per file, such as being able to do
 quick tests via the command line. I also talked to Greg upstream and this
 is definitely the preferred method:


 {{{
 <dsaxena>       is my understanding correct in that binary data (a bitmask
 in my case) exported/imported via sysfs is frowned upon?
 <gregkh>        dsaxena: why would you want to export binary data that
 way?
 <gregkh>        dsaxena: what would userspace use it for?
 <dsaxena>       well, let me explain the usage...
 <gregkh>        that
 <gregkh>        's a great introduction :)
 <dsaxena>       on OLPC we can tell the system controller what events
 should trigger us to wake up out of suspend
 <dsaxena>       this is done via setting a bit mask, with one bit per
 event
 <dsaxena>       we want a way for OHM to tell the EC driver which events
 we want to ignore and which to actually wake up from
 <dsaxena>       my first thought was have one file in sysfs per event
 <dsaxena>       write 1|0 to enable|disable
 <gregkh>        ok, that sounds fine.
 <dsaxena>       but having a single file which just exporst the mask to
 userspace just seems nicer from userspace programmers point of view
 <gregkh>        from a bash programmers point of view?
 <gregkh>        and what defines those bits?
 <dsaxena>       the hw spec
 <gregkh>        if they are one per file, it's much easier to figure out
 without having to read a hw spec.
 <dsaxena>       true true
 <gregkh>        and you have a chance that they can be used by other
 arches.
 <gregkh>        so i'd say one value per file, like the rest of sysfs :)
 <dsaxena>       ok :)
 }}}

-- 
Ticket URL: <http://dev.laptop.org/ticket/6010#comment:5>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system