8.2 Kernel status update

[Chris Ball]

Hi,

   > Dave looked at it and doesn't like the implementation. I can throw
   > it into our joyride along with the needed changes to olpcrd to
   > mount with a root reservation of 4M or so as a stop-gap to a better
   > solution until we have something better.  Note that both the patch
   > and Dave's proposal of having a sysfs attribute for root
   > reservation and other attributes don't really address the slow
   > performance at full issue. They just keep the filesystem from being
   > filled so we don't see the condition.

And it doesn't solve the main user problem, which is that Sugar refuses
to boot if the filesystem is full, because it can't write its logfiles.
Having 4M reserved for root won't change that, since Sugar runs as olpc.

- Chris.
-- 
Chris Ball   <cjb at laptop.org>