Tv

See https://github.com/bazil/fuse/pull/142 / https://github.com/bazil/fuse/pull/142/commits/bd0064274182aed72c9be7d20562eee1ceaa0502 for inspiration. Basically a flat (non-`Node`/`Handle`-based) interface that gets methods called, that looks up the right `Node`/`Handle` and delegates. Downsides: how does this interact with `vmsplice`...

F_OK, R_OK, W_OK, X_OK, and ORed combinations of the later ones. Beware value portability.

Need to return success with `LookupResponse.Node == 0` and `LookupResponse.EntryValid > 0`. See what `saveLookup` does on `n2 == nil`.

fuse.Conn.Ready closes before mount is visible in /proc/mounts

1

This makes tests using `fstestutil.GetMountInfo` racy. ``` $ go test ./... -test.count=100 ... --- FAIL: TestMountOptionFSNameEvilTab (0.00s) options_test.go:75: mount not found ``` I haven't yet verified whether there's some sort...

fsync(2): ``` EROFS, EINVAL fd is bound to a special file (e.g., a pipe, FIFO, or socket) which does not support synchronization. ```

Support ioctl

2

Also understand `InitHasIoctlDir`.

Let FS opt out of automatic Attr filling in Setattr/Lookup/Create/Mkdir/Symlink

6

What's the right mechanism for the FS to convey that it has already filled Attr, and we shouldn't overwrite it? `Attr.Valid != 0`? What are the exact semantics of validity...

``` #define FATTR_CTIME (1

Interrupt request arriving before referred request should get EAGAIN reply

3

See http://fuse.git.sourceforge.net/git/gitweb.cgi?p=fuse/fuse;a=blob;f=doc/kernel.txt (search for INTERRUPT) for the "specification".

The following messages carry a `Fh` field, but are currently only handled by `NodeXxx` interfaces. Give them a `HandleXxx`, and call that when implemented, falling back to `NodeXxx`. - getattr...