linux-stable.git/fs/fuse, branch v5.4.76

fuse: fix page dereference after free

2020-11-01T11:01:05+00:00

commit d78092e4937de9ce55edcb4ee4c5e3c707be0190 upstream.

After unlock_request() pages from the ap->pages[] array may be put (e.g. by
aborting the connection) and the pages can be freed.

Prevent use after free by grabbing a reference to the page before calling
unlock_request().

The original patch was created by Pradeep P V K.

Reported-by: Pradeep P V K 
Cc: 
Signed-off-by: Miklos Szeredi 
Signed-off-by: Greg Kroah-Hartman

fuse: fix the ->direct_IO() treatment of iov_iter

2020-10-07T06:01:26+00:00

[ Upstream commit 933a3752babcf6513117d5773d2b70782d6ad149 ]

the callers rely upon having any iov_iter_truncate() done inside
->direct_IO() countered by iov_iter_reexpand().

Reported-by: Qian Cai 
Tested-by: Qian Cai 
Signed-off-by: Al Viro 
Signed-off-by: Sasha Levin

fuse: update attr_version counter on fuse_notify_inval_inode()

2020-10-01T11:18:02+00:00

[ Upstream commit 5ddd9ced9aef6cfa76af27d384c17c9e2d610ce8 ]

A GETATTR request can race with FUSE_NOTIFY_INVAL_INODE, resulting in the
attribute cache being updated with stale information after the
invalidation.

Fix this by bumping the attribute version in fuse_reverse_inval_inode().

Reported-by: Krzysztof Rusek 
Signed-off-by: Miklos Szeredi 
Signed-off-by: Sasha Levin

fuse: don't check refcount after stealing page

2020-10-01T11:18:01+00:00

[ Upstream commit 32f98877c57bee6bc27f443a96f49678a2cd6a50 ]

page_count() is unstable.  Unless there has been an RCU grace period
between when the page was removed from the page cache and now, a
speculative reference may exist from the page cache.

Reported-by: Matthew Wilcox 
Signed-off-by: Miklos Szeredi 
Signed-off-by: Sasha Levin

fuse: fix weird page warning

2020-07-29T08:18:28+00:00

commit a5005c3cda6eeb6b95645e6cc32f58dafeffc976 upstream.

When PageWaiters was added, updating this check was missed.

Reported-by: Nikolaus Rath 
Reported-by: Hugh Dickins 
Fixes: 62906027091f ("mm: add PageWaiters indicating tasks are waiting for a page bit")
Signed-off-by: Miklos Szeredi 
Signed-off-by: André Almeida 
Signed-off-by: Sasha Levin

fuse: Fix parameter for FS_IOC_{GET,SET}FLAGS

2020-07-22T07:33:12+00:00

commit 31070f6ccec09f3bd4f1e28cd1e592fa4f3ba0b6 upstream.

The ioctl encoding for this parameter is a long but the documentation says
it should be an int and the kernel drivers expect it to be an int.  If the
fuse driver treats this as a long it might end up scribbling over the stack
of a userspace process that only allocated enough space for an int.

This was previously discussed in [1] and a patch for fuse was proposed in
[2].  From what I can tell the patch in [2] was nacked in favor of adding
new, "fixed" ioctls and using those from userspace.  However there is still
no "fixed" version of these ioctls and the fact is that it's sometimes
infeasible to change all userspace to use the new one.

Handling the ioctls specially in the fuse driver seems like the most
pragmatic way for fuse servers to support them without causing crashes in
userspace applications that call them.

[1]: https://lore.kernel.org/linux-fsdevel/20131126200559.GH20559@hall.aurel32.net/T/
[2]: https://sourceforge.net/p/fuse/mailman/message/31771759/

Signed-off-by: Chirantan Ekbote 
Fixes: 59efec7b9039 ("fuse: implement ioctl support")
Cc: 
Signed-off-by: Miklos Szeredi 
Signed-off-by: Greg Kroah-Hartman

fuse: use ->reconfigure() instead of ->remount_fs()

2020-07-22T07:33:12+00:00

commit 0189a2d367f49729622fdafaef5da73161591859 upstream.

s_op->remount_fs() is only called from legacy_reconfigure(), which is not
used after being converted to the new API.

Convert to using ->reconfigure().  This restores the previous behavior of
syncing the filesystem and rejecting MS_MANDLOCK on remount.

Fixes: c30da2e981a7 ("fuse: convert to use the new mount API")
Cc:  # v5.4
Signed-off-by: Miklos Szeredi 
Signed-off-by: Greg Kroah-Hartman

fuse: ignore 'data' argument of mount(..., MS_REMOUNT)

2020-07-22T07:33:12+00:00

commit e8b20a474cf2c42698d1942f939ff2128819f151 upstream.

The command

  mount -o remount -o unknownoption /mnt/fuse

succeeds on kernel versions prior to v5.4 and fails on kernel version at or
after.  This is because fuse_parse_param() rejects any unrecognised options
in case of FS_CONTEXT_FOR_RECONFIGURE, just as for FS_CONTEXT_FOR_MOUNT.

This causes a regression in case the fuse filesystem is in fstab, since
remount sends all options found there to the kernel; even ones that are
meant for the initial mount and are consumed by the userspace fuse server.

Fix this by ignoring mount options, just as fuse_remount_fs() did prior to
the conversion to the new API.

Reported-by: Stefan Priebe 
Fixes: c30da2e981a7 ("fuse: convert to use the new mount API")
Cc:  # v5.4
Signed-off-by: Miklos Szeredi 
Signed-off-by: Greg Kroah-Hartman

fuse: don't ignore errors from fuse_writepages_fill()

2020-07-22T07:33:03+00:00

[ Upstream commit 7779b047a57f6824a43d0e1f70de2741b7426b9d ]

fuse_writepages() ignores some errors taken from fuse_writepages_fill() I
believe it is a bug: if .writepages is called with WB_SYNC_ALL it should
either guarantee that all data was successfully saved or return error.

Fixes: 26d614df1da9 ("fuse: Implement writepages callback")
Signed-off-by: Vasily Averin 
Signed-off-by: Miklos Szeredi 
Signed-off-by: Sasha Levin

fuse: copy_file_range should truncate cache

2020-06-24T15:50:30+00:00

[ Upstream commit 9b46418c40fe910e6537618f9932a8be78a3dd6c ]

After the copy operation completes the cache is not up-to-date.  Truncate
all pages in the interval that has successfully been copied.

Truncating completely copied dirty pages is okay, since the data has been
overwritten anyway.  Truncating partially copied dirty pages is not okay;
add a comment for now.

Fixes: 88bc7d5097a1 ("fuse: add support for copy_file_range()")
Signed-off-by: Miklos Szeredi 
Signed-off-by: Sasha Levin