linux-stable.git/fs/fuse, branch v6.8.3

fuse: don't unhash root

2024-04-03T13:32:12+00:00

[ Upstream commit b1fe686a765e6c0d71811d825b5a1585a202b777 ]

The root inode is assumed to be always hashed.  Do not unhash the root
inode even if it is marked BAD.

Fixes: 5d069dbe8aaf ("fuse: fix bad inode")
Cc:  # v5.11
Signed-off-by: Miklos Szeredi 
Signed-off-by: Sasha Levin

fuse: fix root lookup with nonzero generation

2024-04-03T13:32:11+00:00

[ Upstream commit 68ca1b49e430f6534d0774a94147a823e3b8b26e ]

The root inode has a fixed nodeid and generation (1, 0).

Prior to the commit 15db16837a35 ("fuse: fix illegal access to inode with
reused nodeid") generation number on lookup was ignored.  After this commit
lookup with the wrong generation number resulted in the inode being
unhashed.  This is correct for non-root inodes, but replacing the root
inode is wrong and results in weird behavior.

Fix by reverting to the old behavior if ignoring the generation for the
root inode, but issuing a warning in dmesg.

Reported-by: Antonio SJ Musumeci 
Closes: https://lore.kernel.org/all/CAOQ4uxhek5ytdN8Yz2tNEOg5ea4NkBb4nk0FGPjPk_9nz-VG3g@mail.gmail.com/
Fixes: 15db16837a35 ("fuse: fix illegal access to inode with reused nodeid")
Cc:  # v5.14
Signed-off-by: Miklos Szeredi 
Signed-off-by: Sasha Levin

fuse: replace remaining make_bad_inode() with fuse_make_bad()

2024-04-03T13:32:11+00:00

[ Upstream commit 82e081aebe4d9c26e196c8260005cc4762b57a5d ]

fuse_do_statx() was added with the wrong helper.

Fixes: d3045530bdd2 ("fuse: implement statx")
Cc:  # v6.6
Signed-off-by: Miklos Szeredi 
Signed-off-by: Sasha Levin

fuse: fix VM_MAYSHARE and direct_io_allow_mmap

2024-04-03T13:32:06+00:00

[ Upstream commit 9511176bbaee0ac60ecc84e7b01cf5972a59ea17 ]

There were multiple issues with direct_io_allow_mmap:

 - fuse_link_write_file() was missing, resulting in warnings in
   fuse_write_file_get() and EIO from msync()

 - "vma->vm_ops = &fuse_file_vm_ops" was not set, but especially
   fuse_page_mkwrite is needed.

The semantics of invalidate_inode_pages2() is so far not clearly defined in
fuse_file_mmap.  It dates back to commit 3121bfe76311 ("fuse: fix
"direct_io" private mmap") Though, as direct_io_allow_mmap is a new
feature, that was for MAP_PRIVATE only.  As invalidate_inode_pages2() is
calling into fuse_launder_folio() and writes out dirty pages, it should be
safe to call invalidate_inode_pages2 for MAP_PRIVATE and MAP_SHARED as
well.

Cc: Hao Xu 
Cc: stable@vger.kernel.org
Fixes: e78662e818f9 ("fuse: add a new fuse init flag to relax restrictions in no cache mode")
Signed-off-by: Bernd Schubert 
Reviewed-by: Amir Goldstein 
Signed-off-by: Miklos Szeredi 
Signed-off-by: Sasha Levin

fuse: fix UAF in rcu pathwalks

2024-02-25T07:10:32+00:00

->permission(), ->get_link() and ->inode_get_acl() might dereference
->s_fs_info (and, in case of ->permission(), ->s_fs_info->fc->user_ns
as well) when called from rcu pathwalk.

Freeing ->s_fs_info->fc is rcu-delayed; we need to make freeing ->s_fs_info
and dropping ->user_ns rcu-delayed too.

Signed-off-by: Al Viro

Merge tag 'vfs-6.8.rw' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs

2024-01-08T19:11:51+00:00

Pull vfs rw updates from Christian Brauner:
 "This contains updates from Amir for read-write backing file helpers
  for stacking filesystems such as overlayfs:

   - Fanotify is currently in the process of introducing pre content
     events. Roughly, a new permission event will be added indicating
     that it is safe to write to the file being accessed. These events
     are used by hierarchical storage managers to e.g., fill the content
     of files on first access.

     During that work we noticed that our current permission checking is
     inconsistent in rw_verify_area() and remap_verify_area().
     Especially in the splice code permission checking is done multiple
     times. For example, one time for the whole range and then again for
     partial ranges inside the iterator.

     In addition, we mostly do permission checking before we call
     file_start_write() except for a few places where we call it after.
     For pre-content events we need such permission checking to be done
     before file_start_write(). So this is a nice reason to clean this
     all up.

     After this series, all permission checking is done before
     file_start_write().

     As part of this cleanup we also massaged the splice code a bit. We
     got rid of a few helpers because we are alredy drowning in special
     read-write helpers. We also cleaned up the return types for splice
     helpers.

   - Introduce generic read-write helpers for backing files. This lifts
     some overlayfs code to common code so it can be used by the FUSE
     passthrough work coming in over the next cycles. Make Amir and
     Miklos the maintainers for this new subsystem of the vfs"

* tag 'vfs-6.8.rw' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs: (30 commits)
  fs: fix __sb_write_started() kerneldoc formatting
  fs: factor out backing_file_mmap() helper
  fs: factor out backing_file_splice_{read,write}() helpers
  fs: factor out backing_file_{read,write}_iter() helpers
  fs: prepare for stackable filesystems backing file helpers
  fsnotify: optionally pass access range in file permission hooks
  fsnotify: assert that file_start_write() is not held in permission hooks
  fsnotify: split fsnotify_perm() into two hooks
  fs: use splice_copy_file_range() inline helper
  splice: return type ssize_t from all helpers
  fs: use do_splice_direct() for nfsd/ksmbd server-side-copy
  fs: move file_start_write() into direct_splice_actor()
  fs: fork splice_file_range() from do_splice_direct()
  fs: create {sb,file}_write_not_started() helpers
  fs: create file_write_started() helper
  fs: create __sb_write_started() helper
  fs: move kiocb_start_write() into vfs_iocb_iter_write()
  fs: move permission hook out of do_iter_read()
  fs: move permission hook out of do_iter_write()
  fs: move file_start_write() into vfs_iter_write()
  ...

fs: use splice_copy_file_range() inline helper

2023-12-12T15:20:02+00:00

generic_copy_file_range() is just a wrapper around splice_file_range(),
which caps the maximum copy length.

The only caller of splice_file_range(), namely __ceph_copy_file_range()
is already ready to cope with short copy.

Move the length capping into splice_file_range() and replace the exported
symbol generic_copy_file_range() with a simple inline helper.

Suggested-by: Christoph Hellwig 
Link: https://lore.kernel.org/linux-fsdevel/20231204083849.GC32438@lst.de/
Reviewed-by: Jan Kara 
Signed-off-by: Amir Goldstein 
Link: https://lore.kernel.org/r/20231212094440.250945-3-amir73il@gmail.com
Signed-off-by: Christian Brauner

fuse: disable FOPEN_PARALLEL_DIRECT_WRITES with FUSE_DIRECT_IO_ALLOW_MMAP

2023-12-04T09:19:32+00:00

The new fuse init flag FUSE_DIRECT_IO_ALLOW_MMAP breaks assumptions made by
FOPEN_PARALLEL_DIRECT_WRITES and causes test generic/095 to hit
BUG_ON(fi->writectr < 0) assertions in fuse_set_nowrite():

generic/095 5s ...
  kernel BUG at fs/fuse/dir.c:1756!
...
  ? fuse_set_nowrite+0x3d/0xdd
  ? do_raw_spin_unlock+0x88/0x8f
  ? _raw_spin_unlock+0x2d/0x43
  ? fuse_range_is_writeback+0x71/0x84
  fuse_sync_writes+0xf/0x19
  fuse_direct_io+0x167/0x5bd
  fuse_direct_write_iter+0xf0/0x146

Auto disable FOPEN_PARALLEL_DIRECT_WRITES when server negotiated
FUSE_DIRECT_IO_ALLOW_MMAP.

Fixes: e78662e818f9 ("fuse: add a new fuse init flag to relax restrictions in no cache mode")
Cc:  # v6.6
Signed-off-by: Amir Goldstein 
Signed-off-by: Miklos Szeredi

fuse: dax: set fc->dax to NULL in fuse_dax_conn_free()

2023-12-04T09:16:53+00:00

fuse_dax_conn_free() will be called when fuse_fill_super_common() fails
after fuse_dax_conn_alloc(). Then deactivate_locked_super() in
virtio_fs_get_tree() will call virtio_kill_sb() to release the discarded
superblock. This will call fuse_dax_conn_free() again in fuse_conn_put(),
resulting in a possible double free.

Fixes: 1dd539577c42 ("virtiofs: add a mount option to enable dax")
Signed-off-by: Hangyu Hua 
Acked-by: Vivek Goyal 
Reviewed-by: Jingbo Xu 
Cc:  # v5.10
Signed-off-by: Miklos Szeredi

fuse: share lookup state between submount and its parent

2023-12-04T09:16:53+00:00

Fuse submounts do not perform a lookup for the nodeid that they inherit
from their parent.  Instead, the code decrements the nlookup on the
submount's fuse_inode when it is instantiated, and no forget is
performed when a submount root is evicted.

Trouble arises when the submount's parent is evicted despite the
submount itself being in use.  In this author's case, the submount was
in a container and deatched from the initial mount namespace via a
MNT_DEATCH operation.  When memory pressure triggered the shrinker, the
inode from the parent was evicted, which triggered enough forgets to
render the submount's nodeid invalid.

Since submounts should still function, even if their parent goes away,
solve this problem by sharing refcounted state between the parent and
its submount.  When all of the references on this shared state reach
zero, it's safe to forget the final lookup of the fuse nodeid.

Signed-off-by: Krister Johansen 
Cc: stable@vger.kernel.org
Fixes: 1866d779d5d2 ("fuse: Allow fuse_fill_super_common() for submounts")
Signed-off-by: Miklos Szeredi