#10014 NORM Not Tri: failure to find root device

[Zarro Boogs per Child]

#10014: failure to find root device
------------------------------------+---------------------------------------
           Reporter:  pgf           |       Owner:               
               Type:  defect        |      Status:  new          
           Priority:  normal        |   Milestone:  Not Triaged  
          Component:  not assigned  |     Version:  not specified
         Resolution:                |    Keywords:               
        Next_action:  never set     |    Verified:  0            
Deployment_affected:                |   Blockedby:               
           Blocking:                |  
------------------------------------+---------------------------------------

Old description:

> machine was running abbreviated runin tests -- 10 minute cycle, no
> battery, memory, or suspend tests.

New description:

 machine was running abbreviated runin tests -- 10 minute cycle, no
 battery, memory, or suspend tests.

--

Comment(by Quozl):

 The kernel output shows errors prior to enumerating the SD card, and that
 it completed roughly 41 seconds after kernel start.

 The init script in the initrd waits for udev to be idle, then waits about
 ten more seconds while loading and unloading the scsi_wait_scan module
 once very half second, and if the root device is not present by the end it
 gives up.  Apparently this is a usable algorithm, but it does depend on
 the devices appearing within a certain time.

 I don't know how quickly the internal SD card should appear, nor if there
 is a way to speed it up in the same way that the initrd hopes to speed
 SCSI devices.

 So this seems like a race?  Or is there something pathological about the
 detection?

-- 
Ticket URL: <http://dev.laptop.org/ticket/10014#comment:1>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system