On Wed, Oct 11, 2006 at 09:53:23AM -0400, Kevin Flynn wrote: > My understanding is that hibernation is achieved by copying the contents of > (used?) RAM to the system swap disk partition. > > Hypothetical question: What if the total memory (physical + swap) in use > exceeds the size of the swap space? How can hibernation then occur? Before it starts writing out data to the swap partition, the kernel discards a bunch of cached data, so it's never writing out the total amount of physical memory. This said, it's still possible for you to completely fill memory with pages that the kernel can't purge. If you then attempt to suspend, the kernel will try, fail, and then abort if there is insufficient swap space. Dave -- http://www.codemonkey.org.uk
