This is cleanup, as well as allowing switching between I/O modes while the
file is open in the future.
Signed-off-by: Miklos Szeredi <[email protected]>
Nothing preventing copy_file_range to work on files opened with
FOPEN_DIRECT_IO.
Fixes: 88bc7d5097 ("fuse: add support for copy_file_range()")
Signed-off-by: Miklos Szeredi <[email protected]>
The default splice implementation is grossly inefficient and the iter based
ones work just fine, so use those instead. I've measured an 8x speedup for
splice write (with len = 128k).
Signed-off-by: Miklos Szeredi <[email protected]>
The only caller that needs fc->aborted set is fuse_conn_abort_write().
Setting fc->aborted is now racy (fuse_abort_conn() may already be in
progress or finished) but there's no reason to care.
Signed-off-by: Miklos Szeredi <[email protected]>
Auxiliary requests chained on req->misc.write.next may be leaked on
truncate. Free these as well if the parent request was truncated off.
Signed-off-by: Miklos Szeredi <[email protected]>
Don't reuse the queued request, even if it only contains a single page.
This is needed because previous locking changes (spliting out
fiq->waitq.lock from fc->lock) broke the assumption that request will
remain in FR_PENDING at least until the new page contents are copied.
This fix removes a slight optimization for a rare corner case, so we really
shoudln't care.
Reported-by: Kirill Tkhai <[email protected]>
Fixes: fd22d62ed0 ("fuse: no fc->lock for iqueue parts")
Signed-off-by: Miklos Szeredi <[email protected]>
Restructure the function to better separate the locked and the unlocked
parts. Use the "old_req" local variable to mean only the queued request,
and not any auxiliary requests added onto its misc.write.next list. These
changes are in preparation for the following patch.
Also turn BUG_ON instances into WARN_ON and add a header comment explaining
what the function does.
Signed-off-by: Miklos Szeredi <[email protected]>
Call this from fuse_range_is_writeback() and fuse_writepage_in_flight().
Turn a BUG_ON() into a WARN_ON() in the process.
Signed-off-by: Miklos Szeredi <[email protected]>
NR_WRITEBACK_TEMP is accounted on the temporary page in the request, not
the page cache page.
Fixes: 8b284dc472 ("fuse: writepages: handle same page rewrites")
Cc: <[email protected]> # v3.13
Signed-off-by: Miklos Szeredi <[email protected]>
This reverts commit 007ea44892.
The commit broke some selinux-testsuite cases, and it looks like there's no
straightforward fix keeping the direction of this patch, so revert for now.
The original patch was trying to fix the consistency of permission checks, and
not an observed bug. So reverting should be safe.
Signed-off-by: Miklos Szeredi <[email protected]>
fuse_invalidate_attr() now sets fi->inval_mask instead of fi->i_time, hence
we need to check the inval mask in fuse_permission() as well.
Signed-off-by: Miklos Szeredi <[email protected]>
Fixes: 2f1e81965f ("fuse: allow fine grained attr cache invaldation")
Commit ab2257e994 ("fuse: reduce size of struct fuse_inode") moved parts
of fields related to writeback on regular file and to directory caching
into a union. However fuse_fsync_common() called from fuse_dir_fsync()
touches some writeback related fields, resulting in a crash.
Move writeback related parts from fuse_fsync_common() to fuse_fysnc().
Reported-by: Brett Girton <[email protected]>
Tested-by: Brett Girton <[email protected]>
Fixes: ab2257e994 ("fuse: reduce size of struct fuse_inode")
Signed-off-by: Miklos Szeredi <[email protected]>
In current fuse_drop_waiting() implementation it's possible that
fuse_wait_aborted() will not be woken up in the unlikely case that
fuse_abort_conn() + fuse_wait_aborted() runs in between checking
fc->connected and calling atomic_dec(&fc->num_waiting).
Do the atomic_dec_and_test() unconditionally, which also provides the
necessary barrier against reordering with the fc->connected check.
The explicit smp_mb() in fuse_wait_aborted() is not actually needed, since
the spin_unlock() in fuse_abort_conn() provides the necessary RELEASE
barrier after resetting fc->connected. However, this is not a performance
sensitive path, and adding the explicit barrier makes it easier to
document.
Signed-off-by: Miklos Szeredi <[email protected]>
Fixes: b8f95e5d13 ("fuse: umount should wait for all requests")
Cc: <[email protected]> #v4.19
fuse_request_send_notify_reply() may fail if the connection was reset for
some reason (e.g. fs was unmounted). Don't leak request reference in this
case. Besides leaking memory, this resulted in fc->num_waiting not being
decremented and hence fuse_wait_aborted() left in a hanging and unkillable
state.
Fixes: 2d45ba381a ("fuse: add retrieve request")
Fixes: b8f95e5d13 ("fuse: umount should wait for all requests")
Reported-and-tested-by: [email protected]
Signed-off-by: Miklos Szeredi <[email protected]>
Cc: <[email protected]> #v2.6.36
Current behavior is to automatically disable metacopy if redirect_dir is
not enabled and proceed with the mount.
If "metacopy=on" mount option was given, then this behavior can confuse the
user: no mount failure, yet metacopy is disabled.
This patch makes metacopy=on imply redirect_dir=on.
The converse is also true: turning off full redirect with redirect_dir=
{off|follow|nofollow} will disable metacopy.
If both metacopy=on and redirect_dir={off|follow|nofollow} is specified,
then mount will fail, since there's no way to correctly resolve the
conflict.
Reported-by: Daniel Walsh <[email protected]>
Fixes: d5791044d2 ("ovl: Provide a mount option metacopy=on/off...")
Cc: <[email protected]> # v4.19
Signed-off-by: Miklos Szeredi <[email protected]>
Kaixuxia repors that it's possible to crash overlayfs by removing the
whiteout on the upper layer before creating a directory over it. This is a
reproducer:
mkdir lower upper work merge
touch lower/file
mount -t overlay overlay -olowerdir=lower,upperdir=upper,workdir=work merge
rm merge/file
ls -al merge/file
rm upper/file
ls -al merge/
mkdir merge/file
Before commencing with a vfs_rename(..., RENAME_EXCHANGE) verify that the
lookup of "upper" is positive and is a whiteout, and return ESTALE
otherwise.
Reported by: kaixuxia <[email protected]>
Signed-off-by: Miklos Szeredi <[email protected]>
Fixes: e9be9d5e76 ("overlay filesystem")
Cc: <[email protected]> # v3.18
Now that the workdir and tmpfile copy up modes have been untagled, the
functions become simple enough that the helpers can be folded into the
callers.
Add new helpers where there is any duplication remaining: preparing creds
for creating the object.
Signed-off-by: Miklos Szeredi <[email protected]>
Make permission checking more consistent:
- special files don't need any access check on underling fs
- exec permission check doesn't need to be performed on underlying fs
Reported-by: "J. Bruce Fields" <[email protected]>
Signed-off-by: Miklos Szeredi <[email protected]>
If security_inode_copy_up() fails, it should not set new_creds, so no need
for the cleanup (which would've Oops-ed anyway, due to old_creds being
NULL).
Signed-off-by: Miklos Szeredi <[email protected]>
After sending a synchronous READ request from __fuse_direct_read() we only
need to invalidate atime; none of the other attributes should be changed by
a read().
Signed-off-by: Miklos Szeredi <[email protected]>
If 'auto_inval_data' mode is active, then fuse_file_read_iter() will call
fuse_update_attributes(), which will check the attribute validity and send
a GETATTR request if some of the attributes are no longer valid. The page
cache is then invalidated if the size or mtime have changed.
Then, if a READ request was sent and reply received (which is the case if
the data wasn't cached yet, or if the file is opened for O_DIRECT), the
atime attribute is invalidated.
This will result in the next read() also triggering a GETATTR, ...
This can be fixed by only sending GETATTR if the mode or size are invalid,
we don't need to do a refresh if only atime is invalid.
More generally, none of the callers of fuse_update_attributes() need an
up-to-date atime value, so for now just remove STATX_ATIME from the request
mask when attributes are updated for internal use.
Signed-off-by: Miklos Szeredi <[email protected]>
This patch adds the infrastructure for more fine grained attribute
invalidation. Currently only 'atime' is invalidated separately.
The use of this infrastructure is extended to the statx(2) interface, which
for now means that if only 'atime' is invalid and STATX_ATIME is not
specified in the mask argument, then no GETATTR request will be generated.
Signed-off-by: Miklos Szeredi <[email protected]>
Unprotected naming of local variables within bit_clear_unless() can easily
lead to using the wrong scope.
Noticed this by code review after having hit this issue in set_mask_bits()
Signed-off-by: Miklos Szeredi <[email protected]>
Fixes: 85ad1d13ee ("md: set MD_CHANGE_PENDING in a atomic region")
Cc: Guoqing Jiang <[email protected]>
Unprotected naming of local variables within the set_mask_bits() can easily
lead to using the wrong scope.
Noticed this when "set_mask_bits(&foo->bar, 0, mask)" behaved as no-op.
Signed-off-by: Miklos Szeredi <[email protected]>
Fixes: 00a1a053eb ("ext4: atomically set inode->i_flags in ext4_set_inode_flags()")
Cc: Theodore Ts'o <[email protected]>
Format has a typo: it was meant to be "%.*s", not "%*s". But at some point
callers grew nonprintable values as well, so use "%*pE" instead with a
maximized length.
Reported-by: Amir Goldstein <[email protected]>
Signed-off-by: Miklos Szeredi <[email protected]>
Fixes: 3a1e819b4e ("ovl: store file handle of lower inode on copy up")
Cc: <[email protected]> # v4.12
Writeback caching currently allocates requests with the maximum number of
possible pages, while the actual number of pages per request depends on a
couple of factors that cannot be determined when the request is allocated
(whether page is already under writeback, whether page is contiguous with
previous pages already added to a request).
This patch allows such requests to start with no page allocation (all pages
inline) and grow the page array on demand.
If the max_pages tunable remains the default value, then this will mean
just one allocation that is the same size as before. If the tunable is
larger, then this adds at most 3 additional memory allocations (which is
generously compensated by the improved performance from the larger
request).
Signed-off-by: Miklos Szeredi <[email protected]>
When allocating page array for a request the array for the page pointers
and the array for page descriptors are allocated by two separate kmalloc()
calls. Merge these into one allocation.
Also instead of initializing the request and the page arrays with memset(),
use the zeroing allocation variants.
Reserved requests never carry pages (page array size is zero). Make that
explicit by initializing the page array pointers to NULL and make sure the
assumption remains true by adding a WARN_ON().
Signed-off-by: Miklos Szeredi <[email protected]>
Do this by grouping fields used for cached writes and putting them into a
union with fileds used for cached readdir (with obviously no overlap, since
we don't have hybrid objects).
Signed-off-by: Miklos Szeredi <[email protected]>
Use the internal iversion counter to make sure modifications of the
directory through this filesystem are not missed by the mtime check (due to
mtime granularity).
Signed-off-by: Miklos Szeredi <[email protected]>
Store the modification time of the directory in the cache, obtained before
starting to fill the cache.
When reading the cache, verify that the directory hasn't changed, by
checking if current modification time is the same as the one stored in the
cache.
This only needs to be done when the current file position is at the
beginning of the directory, as mandated by POSIX.
Signed-off-by: Miklos Szeredi <[email protected]>
Allow the cache to be invalidated when page(s) have gone missing. In this
case increment the version of the cache and reset to an empty state.
Add a version number to the directory stream in struct fuse_file as well,
indicating the version of the cache it's supposed to be reading. If the
cache version doesn't match the stream's version, then reset the stream to
the beginning of the cache.
Signed-off-by: Miklos Szeredi <[email protected]>
The cache is only used if it's completed, not while it's still being
filled; this constraint could be lifted later, if it turns out to be
useful.
Introduce state in struct fuse_file that indicates the position within the
cache. After a seek, reset the position to the beginning of the cache and
search the cache for the current position. If the current position is not
found in the cache, then fall back to uncached readdir.
It can also happen that page(s) disappear from the cache, in which case we
must also fall back to uncached readdir.
Signed-off-by: Miklos Szeredi <[email protected]>
This patch just adds the cache filling functions, which are invoked if
FOPEN_CACHE_DIR flag is set in the OPENDIR reply.
Cache reading and cache invalidation are added by subsequent patches.
The directory cache uses the page cache. Directory entries are packed into
a page in the same format as in the READDIR reply. A page only contains
whole entries, the space at the end of the page is cleared. The page is
locked while being modified.
Multiple parallel readdirs on the same directory can fill the cache; the
only constraint is that continuity must be maintained (d_off of last entry
points to position of current entry).
Signed-off-by: Miklos Szeredi <[email protected]>
Add flag returned by OPENDIR request to allow kernel to cache directory
contents in page cache. The effect of FOPEN_CACHE_DIR is twofold:
a) if not already cached, it writes entries into the cache
b) if already cached, it allows reading entries from the cache
The FOPEN_KEEP_CACHE has the same effect as on regular files: unless this
flag is given the cache is cleared upon completion of open.
So FOPEN_KEEP_CACHE and FOPEN_KEEP_CACHE flags should be used together to
make use of the directory caching facility introduced in the following
patches.
The FUSE_AUTO_INVAL_DATA flag returned in INIT reply also has the same
affect on the directory cache as it has on data cache for regular files.
Signed-off-by: Miklos Szeredi <[email protected]>
Using waitqueue_active() is racy. Make sure we issue a wake_up()
unconditionally after storing into fc->blocked. After that it's okay to
optimize with waitqueue_active() since the first wake up provides the
necessary barrier for all waiters, not the just the woken one.
Signed-off-by: Miklos Szeredi <[email protected]>
Fixes: 3c18ef8117 ("fuse: optimize wake_up")
Cc: <[email protected]> # v3.10
Otherwise fuse_dev_do_write() could come in and finish off the request, and
the set_bit(FR_SENT, ...) could trigger the WARN_ON(test_bit(FR_SENT, ...))
in request_end().
Signed-off-by: Miklos Szeredi <[email protected]>
Reported-by: [email protected]
Fixes: 46c34a348b ("fuse: no fc->lock for pqueue parts")
Cc: <[email protected]> # v4.2
ovl_free_fs() dereferences ofs->workbasedir and ofs->upper_mnt in cases when
those might not have been initialized yet.
Fix the initialization order for these fields.
Reported-by: [email protected]
Signed-off-by: Miklos Szeredi <[email protected]>
Cc: <[email protected]> # v4.15
Fixes: 95e6d4177c ("ovl: grab reference to workbasedir early")
Fixes: a9075cdb46 ("ovl: factor out ovl_free_fs() helper")
iput() ends up calling ->evict() on new inode, which is not yet initialized
by owning fs. So use destroy_inode() instead.
Add to sb->s_inodes list only if inode is not in I_CREATING state (meaning
that it wasn't allocated with new_inode(), which already does the
insertion).
Reported-by: Al Viro <[email protected]>
Signed-off-by: Miklos Szeredi <[email protected]>
Fixes: 80ea09a002 ("vfs: factor out inode_insert5()")
Signed-off-by: Al Viro <[email protected]>
syzbot is hitting NULL pointer dereference at process_init_reply().
This is because deactivate_locked_super() is called before response for
initial request is processed.
Fix this by aborting and waiting for all requests (including FUSE_INIT)
before resetting fc->sb.
Original patch by Tetsuo Handa <[email protected]>.
Reported-by: syzbot <[email protected]>
Fixes: e27c9d3877 ("fuse: fuse: add time_gran to INIT_OUT")
Cc: <[email protected]> # v3.19
Signed-off-by: Miklos Szeredi <[email protected]>
fuse_abort_conn() does not guarantee that all async requests have actually
finished aborting (i.e. their ->end() function is called). This could
actually result in still used inodes after umount.
Add a helper to wait until all requests are fully done. This is done by
looking at the "num_waiting" counter. When this counter drops to zero, we
can be sure that no more requests are outstanding.
Fixes: 0d8e84b043 ("fuse: simplify request abort")
Cc: <[email protected]> # v4.2
Signed-off-by: Miklos Szeredi <[email protected]>
fuse_dev_release() assumes that it's the only one referencing the
fpq->processing list, but that's not true, since fuse_abort_conn() can be
doing the same without any serialization between the two.
Fixes: c3696046be ("fuse: separate pqueue for clones")
Cc: <[email protected]> # v4.2
Signed-off-by: Miklos Szeredi <[email protected]>
Refcounting of request is broken when fuse_abort_conn() is called and
request is on the fpq->io list:
- ref is taken too late
- then it is not dropped
Fixes: 0d8e84b043 ("fuse: simplify request abort")
Cc: <[email protected]> # v4.2
Signed-off-by: Miklos Szeredi <[email protected]>
We can now drop description of the ro/rw inconsistency from the
documentation.
Also clarify, that now fully standard compliant behavior can be enabled
with kernel/module/mount options.
Signed-off-by: Miklos Szeredi <[email protected]>
This was provided for debugging the ro/rw inconsistecy. The inconsitency
is now gone so this option is obsolete.
Signed-off-by: Miklos Szeredi <[email protected]>
Opening regular files on overlayfs is now handled via ovl_open(). Remove
the now unused "open_flags" argument from d_op->d_real() and the d_real()
helper.
Signed-off-by: Miklos Szeredi <[email protected]>
This partially reverts commit c568d68341.
Overlayfs files will now automatically get the correct locks, no need to
hack overlay support in VFS.
It is a partial revert, because it leaves the locks_inode() calls in place
and defines locks_inode() to file_inode(). We could revert those as well,
but it would be unnecessary code churn and it makes sense to document that
we are getting the inode for locking purposes.
Don't revert MS_NOREMOTELOCK yet since that has been part of the userspace
API for some time (though not in a useful way). Will try to remove
internal flags later when the dust around the new mount API settles.
Signed-off-by: Miklos Szeredi <[email protected]>
Acked-by: Jeff Layton <[email protected]>
This reverts commit 4d0c5ba2ff.
We now get write access on both overlay and underlying layers so this patch
is no longer needed for correct operation.
Signed-off-by: Miklos Szeredi <[email protected]>
The underlying real file used by overlayfs still contains the overlay path.
This results in mnt_want_write_file() calls by the filesystem getting
freeze protection on the wrong inode (the overlayfs one instead of the real
one).
Fix by using file_inode(file)->i_sb instead of file->f_path.mnt->mnt_sb.
Reported-by: Amir Goldstein <[email protected]>
Signed-off-by: Miklos Szeredi <[email protected]>
Reviewed-by: Christoph Hellwig <[email protected]>
This reverts commit 7c6893e3c9.
Overlayfs no longer relies on the vfs for checking writability of files.
Signed-off-by: Miklos Szeredi <[email protected]>
This reverts commit 954c736f86.
Overlayfs no longer relies on the vfs for checking writability of files.
Signed-off-by: Miklos Szeredi <[email protected]>
Let overlayfs do its thing when opening a file.
This enables stacking and fixes the corner case when a file is opened for
read, modified through a writable open, and data is read from the read-only
file. After this patch the read-only open will not return stale data even
in this case.
Signed-off-by: Miklos Szeredi <[email protected]>
Signed-off-by: Al Viro <[email protected]>
Implement stacked fiemap().
Need to split inode operations for regular file (which has fiemap) and
special file (which doesn't have fiemap).
Signed-off-by: Miklos Szeredi <[email protected]>
In the common case we can just use the real file cached in
file->private_data. There are two exceptions:
1) File has been copied up since open: in this unlikely corner case just
use a throwaway real file for the operation. If ever this becomes a
perfomance problem (very unlikely, since overlayfs has been doing most fine
without correctly handling this case at all), then we can deal with that by
updating the cached real file.
2) File's f_flags have changed since open: no need to reopen the cached
real file, we can just change the flags there as well.
Signed-off-by: Miklos Szeredi <[email protected]>
Implement file operations on a regular overlay file. The underlying file
is opened separately and cached in ->private_data.
It might be worth making an exception for such files when accounting in
nr_file to confirm to userspace expectations. We are only adding a small
overhead (248bytes for the struct file) since the real inode and dentry are
pinned by overlayfs anyway.
This patch doesn't have any effect, since the vfs will use d_real() to find
the real underlying file to open. The patch at the end of the series will
actually enable this functionality.
AV: make it use open_with_fake_path(), don't mess with override_creds
SzM: still need to mess with override_creds() until no fs uses
current_cred() in their open method.
Signed-off-by: Miklos Szeredi <[email protected]>
Signed-off-by: Al Viro <[email protected]>
Copy i_size of the underlying inode to the overlay inode in ovl_copyattr().
This is in preparation for stacking I/O operations on overlay files.
This patch shouldn't have any observable effect.
Remove stale comment from ovl_setattr() [spotted by Vivek Goyal].
Signed-off-by: Miklos Szeredi <[email protected]>
This reverts commit 31c3a70695.
Re-add functionality dealing with i_writecount on truncate to overlayfs.
This patch shouldn't have any observable effects, since we just re-assert
the writecout that vfs_truncate() already got for us.
This is in preparation for moving overlay functionality out of the VFS.
Signed-off-by: Miklos Szeredi <[email protected]>
On inode creation copy certain inode flags from the underlying real inode
to the overlay inode.
This is in preparation for moving overlay functionality out of the VFS.
Signed-off-by: Miklos Szeredi <[email protected]>
Copy up mtime and ctime to overlay inode after times in real object are
modified. Be careful not to dirty cachelines when not necessary.
This is in preparation for moving overlay functionality out of the VFS.
This patch shouldn't have any observable effect.
Signed-off-by: Miklos Szeredi <[email protected]>
Stacking file operations in overlay will store an extra open file for each
overlay file opened.
The overhead is just that of "struct file" which is about 256bytes, because
overlay already pins an extra dentry and inode when the file is open, which
add up to a much larger overhead.
For fear of breaking working setups, don't start accounting the extra file.
Signed-off-by: Miklos Szeredi <[email protected]>
Extract vfs_dedupe_file_range_one() helper to deal with a single dedup
request.
Signed-off-by: Miklos Szeredi <[email protected]>
Reviewed-by: Christoph Hellwig <[email protected]>
Clean up f_op->dedupe_file_range() interface.
1) Use loff_t for offsets and length instead of u64
2) Order the arguments the same way as {copy|clone}_file_range().
Signed-off-by: Miklos Szeredi <[email protected]>
Reviewed-by: Darrick J. Wong <[email protected]>
syzbot is reporting NULL pointer dereference at fuse_ctl_remove_conn() [1].
Since fc->ctl_ndents is incremented by fuse_ctl_add_conn() when new_inode()
failed, fuse_ctl_remove_conn() reaches an inode-less dentry and tries to
clear d_inode(dentry)->i_private field.
Fix by only adding the dentry to the array after being fully set up.
When tearing down the control directory, do d_invalidate() on it to get rid
of any mounts that might have been added.
[1] https://syzkaller.appspot.com/bug?id=f396d863067238959c91c0b7cfc10b163638cac6
Reported-by: syzbot <[email protected]>
Fixes: bafa96541b ("[PATCH] fuse: add control filesystem")
Cc: <[email protected]> # v2.6.18
Signed-off-by: Miklos Szeredi <[email protected]>
Al Viro suggested to simplify callers of ovl_create_real() by
returning the created dentry (or ERR_PTR) from ovl_create_real().
Suggested-by: Al Viro <[email protected]>
Signed-off-by: Miklos Szeredi <[email protected]>
The description of this flag says "Don't sync attributes with the server".
In other words: always use the attributes cached in the kernel and don't
send network or local messages to refresh the attributes.
Signed-off-by: Miklos Szeredi <[email protected]>
Fuse has an "atomic_o_trunc" mode, where userspace filesystem uses the
O_TRUNC flag in the OPEN request to truncate the file atomically with the
open.
In this mode there's no need to send a SETATTR request to userspace after
the open, so fuse_do_setattr() checks this mode and returns. But this
misses the important step of truncating the pagecache.
Add the missing parts of truncation to the ATTR_OPEN branch.
Reported-by: Chad Austin <[email protected]>
Fixes: 6ff958edbf ("fuse: add atomic open+truncate support")
Signed-off-by: Miklos Szeredi <[email protected]>
Cc: <[email protected]>
Add some hints about overlayfs kernel config options.
Enabling NFS export by default is especially recommended against, as it
incurs a performance penalty even if the filesystem is not actually
exported.
Signed-off-by: Miklos Szeredi <[email protected]>
Miklos Szeredi