#9457 BLOC Not Tri: XO-1.5 OS30 boot scripts are quite right

[Zarro Boogs per Child]

#9457: XO-1.5 OS30 boot scripts are quite right
-------------------------------+--------------------------------------------
           Reporter:  wad      |       Owner:  dsaxena      
               Type:  defect   |      Status:  new          
           Priority:  blocker  |   Milestone:  Not Triaged  
          Component:  kernel   |     Version:  not specified
         Resolution:           |    Keywords:               
        Next_action:  design   |    Verified:  0            
Deployment_affected:           |   Blockedby:               
           Blocking:           |  
-------------------------------+--------------------------------------------

Comment(by wmb at firmworks.com):

 Okay the solution is for OFW to contain a heuristic to guess the kernel's
 numbering.  Specifically, if the internal slot is empty, external is
 mmcblk0, otherwise internal is mmcblk0 and external is mmcblk1.  The only
 way that would fail would be if somebody pulled out the external SD card
 after starting to boot from it.  In which case they deserve what they get.

 A "better" way would be to have sysfs (or udev or whoever invents hot-plug
 directories in /sys) create a generically named tail like
 mmc0/mmc0:ab93/block/myblockdev , which would be a link to
 /dev/mmcblk<whatever> .  But that has problems too, not the least of which
 is the length of the command line.

 Solaris on SPARC gets this right - it asks OFW for the boot device name,
 which is a full path in the unambiguous physical tree, and correlates that
 to its isomorphic physical tree.

-- 
Ticket URL: <http://dev.laptop.org/ticket/9457#comment:6>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system