Dmitry Safonov reported that a WARN_ON() assertion can be trigered by
userspace when calling inotify_show_fdinfo() for an overlayfs watched
inode, whose dentry aliases were discarded with drop_caches.

The WARN_ON() assertion in inotify_show_fdinfo() was removed, because
it is possible for encoding file handle to fail for other reason, but
the impact of failing to encode an overlayfs file handle goes beyond
this assertion.

As shown in the LTP test case mentioned in the link below, failure to
encode an overlayfs file handle from a non-aliased inode also leads to
failure to report an fid with FAN_DELETE_SELF fanotify events.

As Dmitry notes in his analyzis of the problem, ovl_encode_fh() fails
if it cannot find an alias for the inode, but this failure can be fixed.
ovl_encode_fh() seldom uses the alias and in the case of non-decodable
file handles, as is often the case with fanotify fid info,
ovl_encode_fh() never needs to use the alias to encode a file handle.

Defer finding an alias until it is actually needed so ovl_encode_fh()
will not fail in the common case of FAN_DELETE_SELF fanotify events.

Fixes: 16aac5ad1fa9 ("ovl: support encoding non-decodable file handles")
Reported-by: Dmitry Safonov <dima@arista.com>
Closes: https://lore.kernel.org/linux-fsdevel/CAOQ4uxiie81voLZZi2zXS1BziXZCM24nXqPAxbu8kxXCUWdwOg@mail.gmail.com/
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Link: https://lore.kernel.org/r/20250105162404.357058-3-amir73il@gmail.com
Signed-off-by: Christian Brauner <brauner@kernel.org>

We want to be able to encode an fid from an inode with no alias.

Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Link: https://lore.kernel.org/r/20250105162404.357058-2-amir73il@gmail.com
Signed-off-by: Christian Brauner <brauner@kernel.org>

Encoding file handles is usually performed by a filesystem >encode_fh()
method that may fail for various reasons.

The legacy users of exportfs_encode_fh(), namely, nfsd and
name_to_handle_at(2) syscall are ready to cope with the possibility
of failure to encode a file handle.

There are a few other users of exportfs_encode_{fh,fid}() that
currently have a WARN_ON() assertion when ->encode_fh() fails.
Relax those assertions because they are wrong.

The second linked bug report states commit 16aac5ad1fa9 ("ovl: support
encoding non-decodable file handles") in v6.6 as the regressing commit,
but this is not accurate.

The aforementioned commit only increases the chances of the assertion
and allows triggering the assertion with the reproducer using overlayfs,
inotify and drop_caches.

Triggering this assertion was always possible with other filesystems and
other reasons of ->encode_fh() failures and more particularly, it was
also possible with the exact same reproducer using overlayfs that is
mounted with options index=on,nfs_export=on also on kernels < v6.6.
Therefore, I am not listing the aforementioned commit as a Fixes commit.

Backport hint: this patch will have a trivial conflict applying to
v6.6.y, and other trivial conflicts applying to stable kernels < v6.6.

Reported-by: syzbot+ec07f6f5ce62b858579f@syzkaller.appspotmail.com
Tested-by: syzbot+ec07f6f5ce62b858579f@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/linux-unionfs/671fd40c.050a0220.4735a.024f.GAE@google.com/
Reported-by: Dmitry Safonov <dima@arista.com>
Closes: https://lore.kernel.org/linux-fsdevel/CAGrbwDTLt6drB9eaUagnQVgdPBmhLfqqxAf3F+Juqy_o6oP8uw@mail.gmail.com/
Cc: stable@vger.kernel.org
Signed-off-by: Amir Goldstein <amir73il@gmail.com>
Link: https://lore.kernel.org/r/20241219115301.465396-1-amir73il@gmail.com
Signed-off-by: Christian Brauner <brauner@kernel.org>

Pull overlayfs updates from Amir Goldstein:

 - Fix a syzbot reported NULL pointer deref with bfs lower layers

 - Fix a copy up failure of large file from lower fuse fs

 - Followup cleanup of backing_file API from Miklos

 - Introduction and use of revert/override_creds_light() helpers, that
   were suggested by Christian as a mitigation to cache line bouncing
   and false sharing of fields in overlayfs creator_cred long lived
   struct cred copy.

 - Store up to two backing file references (upper and lower) in an
   ovl_file container instead of storing a single backing file in
   file->private_data.

   This is used to avoid the practice of opening a short lived backing
   file for the duration of some file operations and to avoid the
   specialized use of FDPUT_FPUT in such occasions, that was getting in
   the way of Al's fd_file() conversions.

* tag 'ovl-update-6.13' of git://git.kernel.org/pub/scm/linux/kernel/git/overlayfs/vfs:
  ovl: Filter invalid inodes with missing lookup function
  ovl: convert ovl_real_fdget() callers to ovl_real_file()
  ovl: convert ovl_real_fdget_path() callers to ovl_real_file_path()
  ovl: store upper real file in ovl_file struct
  ovl: allocate a container struct ovl_file for ovl private context
  ovl: do not open non-data lower file for fsync
  ovl: Optimize override/revert creds
  ovl: pass an explicit reference of creators creds to callers
  ovl: use wrapper ovl_revert_creds()
  fs/backing-file: Convert to revert/override_creds_light()
  cred: Add a light version of override/revert_creds()
  backing-file: clean up the API
  ovl: properly handle large files in ovl_security_fileattr

Add a check to the ovl_dentry_weird() function to prevent the
processing of directory inodes that lack the lookup function.
This is important because such inodes can cause errors in overlayfs
when passed to the lowerstack.

Reported-by: syzbot+a8c9d476508bd14a90e5@syzkaller.appspotmail.com
Link: https://syzkaller.appspot.com/bug?extid=a8c9d476508bd14a90e5
Suggested-by: Miklos Szeredi <miklos@szeredi.hu>
Link: https://lore.kernel.org/linux-unionfs/CAJfpegvx-oS9XGuwpJx=Xe28_jzWx5eRo1y900_ZzWY+=gGzUg@mail.gmail.com/
Signed-off-by: Vasiliy Kovalev <kovalev@altlinux.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Amir Goldstein <amir73il@gmail.com>

Pull statx updates from Al Viro:
 "Sanitize struct filename and lookup flags handling in statx and
  friends"

* tag 'pull-statx' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  libfs: kill empty_dir_getattr()
  fs: Simplify getattr interface function checking AT_GETATTR_NOSEC flag
  fs/stat.c: switch to CLASS(fd_raw)
  kill getname_statx_lookup_flags()
  io_statx_prep(): use getname_uflags()

Pull overlayfs updates from Christian Brauner:
 "Make overlayfs support specifying layers through file descriptors.

  Currently overlayfs only allows specifying layers through path names.
  This is inconvenient for users that want to assemble an overlayfs
  mount purely based on file descriptors:

  This enables user to specify both:

    fsconfig(fd_overlay, FSCONFIG_SET_FD, "upperdir+", NULL, fd_upper);
    fsconfig(fd_overlay, FSCONFIG_SET_FD, "workdir+",  NULL, fd_work);
    fsconfig(fd_overlay, FSCONFIG_SET_FD, "lowerdir+", NULL, fd_lower1);
    fsconfig(fd_overlay, FSCONFIG_SET_FD, "lowerdir+", NULL, fd_lower2);

  in addition to:

    fsconfig(fd_overlay, FSCONFIG_SET_STRING, "upperdir+", "/upper",  0);
    fsconfig(fd_overlay, FSCONFIG_SET_STRING, "workdir+",  "/work",   0);
    fsconfig(fd_overlay, FSCONFIG_SET_STRING, "lowerdir+", "/lower1", 0);
    fsconfig(fd_overlay, FSCONFIG_SET_STRING, "lowerdir+", "/lower2", 0);

  There's also a large set of new overlayfs selftests to test new
  features and some older properties"

* tag 'vfs-6.13.ovl' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:
  selftests: add test for specifying 500 lower layers
  selftests: add overlayfs fd mounting selftests
  selftests: use shared header
  Documentation,ovl: document new file descriptor based layers
  ovl: specify layers via file descriptors
  fs: add helper to use mount option as path or fd

Pull vfs file updates from Christian Brauner:
 "This contains changes the changes for files for this cycle:

   - Introduce a new reference counting mechanism for files.

     As atomic_inc_not_zero() is implemented with a try_cmpxchg() loop
     it has O(N^2) behaviour under contention with N concurrent
     operations and it is in a hot path in __fget_files_rcu().

     The rcuref infrastructures remedies this problem by using an
     unconditional increment relying on safe- and dead zones to make
     this work and requiring rcu protection for the data structure in
     question. This not just scales better it also introduces overflow
     protection.

     However, in contrast to generic rcuref, files require a memory
     barrier and thus cannot rely on *_relaxed() atomic operations and
     also require to be built on atomic_long_t as having massive amounts
     of reference isn't unheard of even if it is just an attack.

     This adds a file specific variant instead of making this a generic
     library.

     This has been tested by various people and it gives consistent
     improvement up to 3-5% on workloads with loads of threads.

   - Add a fastpath for find_next_zero_bit(). Skip 2-levels searching
     via find_next_zero_bit() when there is a free slot in the word that
     contains the next fd. This improves pts/blogbench-1.1.0 read by 8%
     and write by 4% on Intel ICX 160.

   - Conditionally clear full_fds_bits since it's very likely that a bit
     in full_fds_bits has been cleared during __clear_open_fds(). This
     improves pts/blogbench-1.1.0 read up to 13%, and write up to 5% on
     Intel ICX 160.

   - Get rid of all lookup_*_fdget_rcu() variants. They were used to
     lookup files without taking a reference count. That became invalid
     once files were switched to SLAB_TYPESAFE_BY_RCU and now we're
     always taking a reference count. Switch to an already existing
     helper and remove the legacy variants.

   - Remove pointless includes of <linux/fdtable.h>.

   - Avoid cmpxchg() in close_files() as nobody else has a reference to
     the files_struct at that point.

   - Move close_range() into fs/file.c and fold __close_range() into it.

   - Cleanup calling conventions of alloc_fdtable() and expand_files().

   - Merge __{set,clear}_close_on_exec() into one.

   - Make __set_open_fd() set cloexec as well instead of doing it in two
     separate steps"

* tag 'vfs-6.13.file' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:
  selftests: add file SLAB_TYPESAFE_BY_RCU recycling stressor
  fs: port files to file_ref
  fs: add file_ref
  expand_files(): simplify calling conventions
  make __set_open_fd() set cloexec state as well
  fs: protect backing files with rcu
  file.c: merge __{set,clear}_close_on_exec()
  alloc_fdtable(): change calling conventions.
  fs/file.c: add fast path in find_next_fd()
  fs/file.c: conditionally clear full_fds
  fs/file.c: remove sanity_check and add likely/unlikely in alloc_fd()
  move close_range(2) into fs/file.c, fold __close_range() into it
  close_files(): don't bother with xchg()
  remove pointless includes of <linux/fdtable.h>
  get rid of ...lookup...fdget_rcu() family

Stop using struct fd to return a real file from ovl_real_fdget(),
because we no longer return a temporary file object and the callers
always get a borrowed file reference.

Rename the helper to ovl_real_file(), return a borrowed reference of
the real file that is referenced from the overlayfs file or an error.

Signed-off-by: Amir Goldstein <amir73il@gmail.com>

Stop using struct fd to return a real file from ovl_real_fdget_path(),
because we no longer return a temporary file object and the callers
always get a borrowed file reference.

Rename the helper to ovl_real_file_path(), return a borrowed reference
of the real file that is referenced from the overlayfs file or an error.

Signed-off-by: Amir Goldstein <amir73il@gmail.com>