Refusing to mount beats a kernel crash.
Signed-off-by: NJoern Engel <joern@logfs.org>

It would probably be better to just accept NULL pointers in
mempool_destroy().  But for the current -rc series let's keep things
simple.

This patch was lost in the cracks for a while.
Kevin Cernekee <cernekee@gmail.com> had to rediscover the problem and
send a similar patch because of it. :(
Signed-off-by: NJoern Engel <joern@logfs.org>

Removing sufficiently large files would create aliases for a large
number of segments.  This in turn results in a large number of journal
entries and an overflow of s_je_array.

Cheap fix is to add a BUG_ON, turning memory corruption into something
annoying, but less dangerous.  Real fix is to count the number of
affected segments and prevent the problem completely.
Signed-off-by: NJoern Engel <joern@logfs.org>

All callers are long gone.
Signed-off-by: NJoern Engel <joern@logfs.org>

do_logfs_journal_wl_pass() must call freeseg(), thereby clear
PagePrivate on all pages of the current journal segment.
Signed-off-by: NJoern Engel <joern@logfs.org>

This prevents unnecessary journal writes.  More importantly it prevents
an oops due to a journal write on failed mount.

Erases for block devices were always just emulated by writing 0xff.
Some time back the write was removed and only the page cache was
changed to 0xff.  Superficialy a good idea with two problems:
1. Touching the page cache isn't necessary either.
2. However, writing out 0xff _is_ necessary for the journal.  As the
   journal is scanned linearly, an old non-overwritten commit entry
   can be used on next mount and cause havoc.

This should fix both aspects.

This is a new flash file system. See
Documentation/filesystems/logfs.txt
Signed-off-by: NJoern Engel <joern@logfs.org>