#12030 NORM 12.1.0: XO-1.5 "irq 22: nobody cared" when going into suspend

[Zarro Boogs per Child]

#12030: XO-1.5 "irq 22: nobody cared" when going into suspend
------------------------------------+---------------------------------------
           Reporter:  dsd           |       Owner:  dsd          
               Type:  defect        |      Status:  new          
           Priority:  normal        |   Milestone:  12.1.0       
          Component:  kernel        |     Version:  not specified
         Resolution:                |    Keywords:               
        Next_action:  add to build  |    Verified:  0            
Deployment_affected:                |   Blockedby:               
           Blocking:                |  
------------------------------------+---------------------------------------
Changes (by dsd):

  * next_action:  diagnose => add to build


Comment:

 Another option here is that the mmc block driver should make sure that all
 requests are finished/stopped in its own suspend handler. I think thats
 probably the most appropriate layer for this synchronization to happen.

 And indeed, mmc_blk_suspend() calls mmc_queue_suspend() which is
 documented to stop the block request queue and wait for any requests to
 complete. "This ensures that we won't suspend while a request is being
 processed"

 mmc_blk_suspend() is not being called in 3.3, so this is our problem. This
 is fixed in upstream commit 32d317c60e56c2a34463b51fc0336cc96b3e1735

 Backported to x86-3.3 and confirmed that mmc_blk_suspend() is now being
 called before the mmc hosts get suspended. Pushed as 1907e4b

-- 
Ticket URL: <http://dev.laptop.org/ticket/12030#comment:8>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system