BBB: init timer on startup

It could fix the problem encountered in
https://groups.google.com/forum/?hl=en#!topic/rt-thread-cnusers/tpqZOMsIp8I

TC: if TC is enabled, add the TC dir to CPPPATH

Some component that depend on TC would like to include the tc_comm.h.

[RFC] More tc

We should not use the _tc_stat to determine whether we should loop the
TCs. Use an other variable and introduce tc_loop FinSh command.

Sq lite on armcc

More tc

fix dfs_unmount bug

With system that has RT_TICK_PER_SECOND is 1000, it's easy to get one
tick error.

More tc

Rtt next2

DFS: let mount fail on device open error

more infomation:
http://www.rt-thread.org/phpBB3/topic2965.html

No functional changes.

fix #211

Out side world is difficult to tell whethere there is dead thread
remaining. If rt_thread_idle_excute only do one cleanup, it's hard to
finish cleanups outside the idle thread. So let is loop and do all the
cleanups in one call.

Out side world may be interested in whether the scheduler is locked or
not. Provide an API to tell that.

empty is confused in dfs_filesystem_lookup.

More tc

The tc thread should delay more time than tid2 in order to let it run
the test. The hard-coded "27" is not enough in some cases.

TCs are supposed to exit with the TC_STAT_RUNNING cleared. If it forgot
to do this, do it by ourselves.

dfs: refine code

clean code with pointer rather than index

Fix tc

decrease indentation to make cleaner code
fix mkfs bug when there is no mkfs implementation

The sem that has been created should be deleted. The sem that has been
inited should be detached.

There can be a race condition that the sem has alread been detached
after we did the cleanup. Besides, there is a RT_IPC_CMD_RESET command
in rt_sem_control which is more clear in setting the sem values.