fixed two diskstore issues, a quasi-deadlock creating problems with I/O speed and a race condition among threads

propagate key eviction to slaves and AOF, like for expires, so that replicas and AOFs are always consistent even in maxmemory mode.

Fixed a theoretical non exploitable security bug reported by @chrisrohlf. In theory if we undefine SDS_ABORT_ON_OOM from sds.c AND modify zmalloc.c in order to don't quit on out of memory (but this would break every other part of Redis), on out of memory there is a possible heap overflow.

INFO commandstats section reset with config resetstat, a new microseconds per call field, and a few other minor things

Now INFO outputs the "INFO default" output if no argument is passed. Also added counts of calls for every command, thanks to an idea by Mike Shaver

Fixed issue #435 and at the same time introduced explicit ping in the master-slave channel that will detect a blocked master or a broken even if apparently connected TCP link.

fixed bgsave_in_progress in INFO when BGSAVEing with diskstore enabled, don't DEBUG FLUSHCACHE when bgsave is in progress.