Re: [RFC PATCH 1/7] FUSE: Add the fsnotify opcode and in/out structs to FUSE

[Amir Goldstein]

On Tue, Oct 26, 2021 at 5:56 PM Amir Goldstein <amir73il@xxxxxxxxx> wrote:
>
> On Mon, Oct 25, 2021 at 11:47 PM Ioannis Angelakopoulos
> <iangelak@xxxxxxxxxx> wrote:
> >
> > Since fsnotify is the backend for the inotify subsystem all the backend
> > code implementation we add is related to fsnotify.
> >
> > To support an fsnotify request in FUSE and specifically virtiofs we add a
> > new opcode for the FSNOTIFY (51) operation request in the "fuse.h" header.
> >
> > Also add the "fuse_notify_fsnotify_in" and "fuse_notify_fsnotify_out"
> > structs that are responsible for passing the fsnotify/inotify related data
> > to and from the FUSE server.
> >
> > Signed-off-by: Ioannis Angelakopoulos <iangelak@xxxxxxxxxx>
> > ---
> >  include/uapi/linux/fuse.h | 23 ++++++++++++++++++++++-
> >  1 file changed, 22 insertions(+), 1 deletion(-)
> >
> > diff --git a/include/uapi/linux/fuse.h b/include/uapi/linux/fuse.h
> > index 46838551ea84..418b7fc72417 100644
> > --- a/include/uapi/linux/fuse.h
> > +++ b/include/uapi/linux/fuse.h
> > @@ -186,6 +186,9 @@
> >   *  - add FUSE_SYNCFS
> >   *  7.35
> >   *  - add FUSE_NOTIFY_LOCK
> > + *  7.36
> > + *  - add FUSE_HAVE_FSNOTIFY
> > + *  - add fuse_notify_fsnotify_(in,out)
> >   */
> >
> >  #ifndef _LINUX_FUSE_H
> > @@ -221,7 +224,7 @@
> >  #define FUSE_KERNEL_VERSION 7
> >
> >  /** Minor version number of this interface */
> > -#define FUSE_KERNEL_MINOR_VERSION 35
> > +#define FUSE_KERNEL_MINOR_VERSION 36
> >
> >  /** The node ID of the root inode */
> >  #define FUSE_ROOT_ID 1
> > @@ -338,6 +341,7 @@ struct fuse_file_lock {
> >   *                     write/truncate sgid is killed only if file has group
> >   *                     execute permission. (Same as Linux VFS behavior).
> >   * FUSE_SETXATTR_EXT:  Server supports extended struct fuse_setxattr_in
> > + * FUSE_HAVE_FSNOTIFY: remote fsnotify/inotify event subsystem support
> >   */
> >  #define FUSE_ASYNC_READ                (1 << 0)
> >  #define FUSE_POSIX_LOCKS       (1 << 1)
> > @@ -369,6 +373,7 @@ struct fuse_file_lock {
> >  #define FUSE_SUBMOUNTS         (1 << 27)
> >  #define FUSE_HANDLE_KILLPRIV_V2        (1 << 28)
> >  #define FUSE_SETXATTR_EXT      (1 << 29)
> > +#define FUSE_HAVE_FSNOTIFY     (1 << 30)
> >
> >  /**
> >   * CUSE INIT request/reply flags
> > @@ -515,6 +520,7 @@ enum fuse_opcode {
> >         FUSE_SETUPMAPPING       = 48,
> >         FUSE_REMOVEMAPPING      = 49,
> >         FUSE_SYNCFS             = 50,
> > +       FUSE_FSNOTIFY           = 51,
> >
> >         /* CUSE specific operations */
> >         CUSE_INIT               = 4096,
> > @@ -532,6 +538,7 @@ enum fuse_notify_code {
> >         FUSE_NOTIFY_RETRIEVE = 5,
> >         FUSE_NOTIFY_DELETE = 6,
> >         FUSE_NOTIFY_LOCK = 7,
> > +       FUSE_NOTIFY_FSNOTIFY = 8,
> >         FUSE_NOTIFY_CODE_MAX,
> >  };
> >
> > @@ -571,6 +578,20 @@ struct fuse_getattr_in {
> >         uint64_t        fh;
> >  };
> >
> > +struct fuse_notify_fsnotify_out {
> > +       uint64_t inode;
>
> 64bit inode is not a good unique identifier of the object.
> you need to either include the generation in object identifier
> or much better use the object's nfs file handle, the same way
> that fanotify stores object identifiers.
>
> > +       uint64_t mask;
> > +       uint32_t namelen;
> > +       uint32_t cookie;
>
> I object to persisting with the two-events-joined-by-cookie design.
> Any new design should include a single event for rename
> with information about src and dst.
>
> I know this is inconvenient, but we are NOT going to create a "remote inotify"
> interface, we need to create a "remote fsnotify" interface and if server wants
> to use inotify, it will need to join the disjoined MOVE_FROM/TO event into
> a single "remote event", that FUSE will use to call fsnotify_move().
>

TBH, the disjoint vs. joint from/to event is an unfinished business
for fanotify.
So my objection above is more of a strong wish.
But I admit that if existing network protocols already encode the disjoint
from/to events semantics, I may need to fold back on that objection.

Thanks,
Amir.