#9693 HIGH 1.5-fir: Reset when SCI coincides with SMI handler

[Zarro Boogs per Child]

#9693: Reset when SCI coincides with SMI handler
-------------------------------------------+--------------------------------
           Reporter:  wmb at firmworks.com    |       Owner:  wmb at firmworks.com   
               Type:  defect               |      Status:  assigned            
           Priority:  high                 |   Milestone:  1.5-firmware-C1-SMT 
          Component:  ofw - open firmware  |     Version:  Development firmware
         Resolution:                       |    Keywords:                      
        Next_action:  add to build         |    Verified:  0                   
Deployment_affected:                       |   Blockedby:                      
           Blocking:                       |  
-------------------------------------------+--------------------------------
Changes (by wmb at firmworks.com):

  * status:  new => assigned
  * next_action:  diagnose => add to build
  * milestone:  1.5-firmware => 1.5-firmware-C1-SMT


Comment:

 Fixed by svn 1497.  The problem was a bad jmp offset in the return-from-
 SMM code, which caused it to skip the crucial "wbinvd" instruction that
 prepares the caches for the switch out of System Management Mode.  The
 cache problem explains why the problem was so elusive, happening and not
 happening under strange combinations of circumstances.  With the wbinvd
 reinstated (by jumping to it instead of just past it), the system does not
 crash, even in the cases where it used to crash reproducibly.

-- 
Ticket URL: <http://dev.laptop.org/ticket/9693#comment:3>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system