Change-Id: I106a8b03d50b647c198e2ba45fb643841aba86b0

Seconds granularity is too coarse to understand what is
happening, so add microseconds. In return, the date
is dropped to keep the lines short.

Old format:
  2016/03/14 18:19:38 Dispatch: OPENDIR, NodeId: 1.

New format:
  18:18:35.988071 Dispatch: OPENDIR, NodeId: 1.

This simplifies testing hard links support (aka ClientInodes)
with external tools like fsstress.

Apart from adding the missing write into clientInodeMap, I have
added a few comments about what happens. Hopefully helps the
next person looking at this code.

Also, make the script much less chatty.

If the file handle is specified we should use it instead of
picking one at random.

Note that the only real change is for ftruncate. For fstat, fchown
and fchmod the kernel currently never passes a file descriptor.

Fixes #95.

This lets you Mount multiple nodes in parallel.

This prevents data races flagged by the race detector.

This fixes the data race in TestDeleteNotify.

time.Unix() returns int64 regardless of target platform but timeval is
defined as int32 values on ubuntu-32bit
NsecToTimeval has a nanosecond problem(negative value) before epoch but
still can use second

It was broken in 1b3f7953, which removed the possibility of
trivially inspecting go-fuse internal inode numbers.

Due to a typo, returnRequest never actually returned the read buffer
to readPool.

Also, we want to put the complete buffer back into the pool - this
caused errors later on.

loopback "tar xzf ../linux-3.0.tar.gz" timings:

before: 35 seconds
after:  10 seconds

go-fuse is currently lagging behind libfuse in terms of small-file
performance due to GC overhead.

In preparation to optimizing the code, this patch makes cpu and mem
profiling easier.

Properly implementing Utimens caused the OSX build to break because
the utimensat system call is not available there.

This patch reverts the OSX implementation of Utimens to using
syscall.Utimes(). This is not actually correct but the best we can do.

Thanks to @spaghetti2514 who reported the issue and tested the patches.

fuse.EntryOut can be corrupted if FS implementation does not set all fileds in the Lookup of rawFS
go-mtpfs displays wrong information(ls -l) for directory(size, Nlink, ..)

The problem was that os.Chtimes() operates on the target of the
symlink and not on the symlink itself.
When tar extracts an archive containing a symlink, it actually wants
to set the times on the symlink. This can also be tested using
"touch -h".

This patch adds a wrapper for the Linux utimensat(2) syscall
that allows passing flags along with a unit test.

It uses the flag AT_SYMLINK_NOFOLLOW to implement
loopbackFileSystem.Utimens() properly.

It also add handling for UTIME_NOW and UTIME_OMIT that was already
present in loopbackFile.Utimens() but missing in loopbackFileSystem.

Fixes issue 81 ( https://github.com/hanwen/go-fuse/issues/81 ).