#3948 NORM Never A: suspend to RAM doesn't work

[Zarro Boogs per Child]

#3948: suspend to RAM doesn't work
---------------------+------------------------------------------------------
  Reporter:  ssb22   |       Owner:  dilinger      
      Type:  defect  |      Status:  new           
  Priority:  normal  |   Milestone:  Never Assigned
 Component:  distro  |     Version:                
Resolution:          |    Keywords:                
  Verified:  0       |  
---------------------+------------------------------------------------------

Comment(by wmb at firmworks.com):

 It starts to get strange pretty quickly:

 {{{
 [ 28.298988] OLPC board with OpenFirmware?: CL1 Q2C26 Q2C
 [ 28.299128] ofw: revision string is too long!
 [ 28.319259] olpc-ec: timeout waiting for EC to quiesce!
 }}}

 "revision string is too long" means that, when the kernel asked OFW for
 the "model" property in the root node, it got back more than 10 bytes (at
 least according to my reading of the latest revision of olpc.c; I didn't
 check the history to see if this code has changed recently).

 But the actual length of that property is 3 bytes, clearly much less than
 10.

 I loaded Q2C26 on my B4 and got this:

 {{{
 ok dev /
 ok .properties
 ...
 model    B4
 ...
 ok " model" get-property .s
 ff8a495c 3 0
 ok
 }}}

 So the property length really is 3 ('B' '4' '\0').

 The next problem is the "timeout waiting for EC to quiesce!".  I wonder if
 they could be related.  Perhaps EC communication is messed up really early
 (which seems likely if the OS can't talk to it at all), and that causes
 OFW to create a bogus model property.  I shall try to simulate that...

-- 
Ticket URL: <https://dev.laptop.org/ticket/3948#comment:14>
One Laptop Per Child <https://dev.laptop.org>
OLPC bug tracking system