#11633 NORM Not Tri: suspend/resume cycle resets ethtool's setting for eth0

[Zarro Boogs per Child]

#11633: suspend/resume cycle resets ethtool's setting for eth0
----------------------------------------------+-----------------------------
           Reporter:  jvonau                  |       Owner:  pgf                                         
               Type:  defect                  |      Status:  new                                         
           Priority:  normal                  |   Milestone:  Not Triaged                                 
          Component:  power manager (powerd)  |     Version:  1.75/1.5/1.0 Software Build os883 aka 11.3.0
         Resolution:                          |    Keywords:                                              
        Next_action:  diagnose                |    Verified:  0                                           
Deployment_affected:                          |   Blockedby:                                              
           Blocking:                          |  
----------------------------------------------+-----------------------------
Changes (by pgf):

  * next_action:  never set => diagnose


Comment:

 yes, i think postresume hooks can use config variables.  but this won't
 fix the bug mentioned in set_wake_on_wlan(), because the very short
 wakeups that are the issue don't run the resume hooks.  and in any case,
 the real issue is that the interface simply isn't available yet -- to fix
 this, we need to wait for eth0 to appear before going back to sleep.

 you raise a very good point though.  i'd forgotten about this particular
 race, and it's certainly possible that you're hitting it.  i'll think
 about this some more.  i think i noticed that on 1.75 the wol settings are
 preserved across s/r.  if so, perhaps we can backport that fix to the x86
 machines.

-- 
Ticket URL: <http://dev.laptop.org/ticket/11633#comment:1>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system