Allow nil FileHandles.

Change DefaultOperations to be for a ReadOnly in-memory filesystem.

The fd is set on creation, and never written. We rely on the kernel
and on fileEntry.wg to synchronize such that RELEASE always happens
after read/write has completed.

This makes the inodes in loopback reflect the underlying inodes if
they don't span mounts

* MutableDirOperations
* XAttrOperations
* LockOperations

add MaxPages to InitOut

Use a []uint32 in Inode to keep track of open files. This reduces the
memory consumption of this feature.

Add a test that tries to (unsuccessfully) trigger file closing on the
while file.GetAttr is under way

This is necessary to provide a file handle to GetAttr.

The protocol is:

  request: read (with buffer 1024)
  response: data (if buffer is big enough), or (data size, ERANGE)

Before, the raw API provided

* GetXAttrSize: called if the kernel sends a 0 sized buffer (which
  never happens

* GetXAttrData: called in other cases. The return value is []byte
  forcing allocation of potentially large buffers

* ListXAttr: return value []byte, so also problematic for allocation.

Now, the raw API mirrors the kernel API more closely, and use recycled
buffers for more efficiency.

This should be tagged with a version bump because it changes the
RawFileSystem API.

Changes to nodefs and paths will be transparent: a Cancel channel was
added to the fuse.Context struct. In addition fuse.Context now
implements the Go context package. This also increases the minimum Go
version to 1.7.

compile on darwin too

When using automatic NodeIds, we can still have a race.

Suggestion by Kirill Smelkov.