No ethernet interface found after suspend-resume

Chris Ball cjb at laptop.org
Fri Apr 30 11:38:04 EDT 2010


Hi,

   > In 2.6.29 - The sdio drivers assumes a card to be removed
   > virtually during suspend and is re-inserted/re-probed during
   > resume. All this happens transperently.
   > 
   > On printing the command logs - I notice that the command 8
   > onwards the libertas does not respond and i get a command timeout
   > and as a result - the card never gets detected again during
   > happens as is expected.

I think I've seen this too.  With your kernel version, you can't keep
an SD device powered across suspend/resume -- the kernel is going to
assume it needs reinit, and then send it commands (e.g. CMD8) that
aren't valid for the part of the SD state machine that the 8686 thinks
it's in.

So, I think your options are to either:

   * allow the MMC layer to turn off the card at suspend (it tries to
     turn off the card power itself) or just reset it during resume,
     as your later mail describes
   * upgrade to 2.6.33, and move to using the preserved-power MMC
     resume code that starts with upstream commit da68c4eb258..

- Chris.
-- 
Chris Ball   <cjb at laptop.org>
One Laptop Per Child



More information about the libertas-dev mailing list