diff options
| author | Al Viro <viro@zeniv.linux.org.uk> | 2020-03-12 18:25:20 -0400 |
|---|---|---|
| committer | Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 2020-03-18 07:14:21 +0100 |
| commit | a8ab0b70979059773f34171c8f05f340f60d47d9 (patch) | |
| tree | 91bf1dbb00b9d2c7d1ed5868d6afc7aa277674bc | |
| parent | a89327c1f7708173beb037386cea0b77357ed52c (diff) | |
| download | linux-stable-a8ab0b70979059773f34171c8f05f340f60d47d9.tar.gz | |
cifs_atomic_open(): fix double-put on late allocation failure
commit d9a9f4849fe0c9d560851ab22a85a666cddfdd24 upstream.
several iterations of ->atomic_open() calling conventions ago, we
used to need fput() if ->atomic_open() failed at some point after
successful finish_open(). Now (since 2016) it's not needed -
struct file carries enough state to make fput() work regardless
of the point in struct file lifecycle and discarding it on
failure exits in open() got unified. Unfortunately, I'd missed
the fact that we had an instance of ->atomic_open() (cifs one)
that used to need that fput(), as well as the stale comment in
finish_open() demanding such late failure handling. Trivially
fixed...
Fixes: fe9ec8291fca "do_last(): take fput() on error after opening to out:"
Cc: stable@kernel.org # v4.7+
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
| -rw-r--r-- | Documentation/filesystems/porting | 7 | ||||
| -rw-r--r-- | fs/cifs/dir.c | 1 | ||||
| -rw-r--r-- | fs/open.c | 3 |
3 files changed, 7 insertions, 4 deletions
diff --git a/Documentation/filesystems/porting b/Documentation/filesystems/porting index 32b5186be412..041b0ded8b44 100644 --- a/Documentation/filesystems/porting +++ b/Documentation/filesystems/porting @@ -627,3 +627,10 @@ in your dentry operations instead. DCACHE_RCUACCESS is gone; having an RCU delay on dentry freeing is the default. DCACHE_NORCU opts out, and only d_alloc_pseudo() has any business doing so. +-- +[mandatory] + + [should've been added in 2016] stale comment in finish_open() + nonwithstanding, failure exits in ->atomic_open() instances should + *NOT* fput() the file, no matter what. Everything is handled by the + caller. diff --git a/fs/cifs/dir.c b/fs/cifs/dir.c index 2fb6fa51fd3c..f6e3c0089825 100644 --- a/fs/cifs/dir.c +++ b/fs/cifs/dir.c @@ -561,7 +561,6 @@ cifs_atomic_open(struct inode *inode, struct dentry *direntry, if (server->ops->close) server->ops->close(xid, tcon, &fid); cifs_del_pending_open(&open); - fput(file); rc = -ENOMEM; } diff --git a/fs/open.c b/fs/open.c index 878478745924..76996f920ebf 100644 --- a/fs/open.c +++ b/fs/open.c @@ -851,9 +851,6 @@ cleanup_file: * the return value of d_splice_alias(), then the caller needs to perform dput() * on it after finish_open(). * - * On successful return @file is a fully instantiated open file. After this, if - * an error occurs in ->atomic_open(), it needs to clean up with fput(). - * * Returns zero on success or -errno if the open failed. */ int finish_open(struct file *file, struct dentry *dentry, |