diff options
author | Jann Horn <jannh@google.com> | 2024-07-23 17:03:56 +0200 |
---|---|---|
committer | Greg Kroah-Hartman <gregkh@linuxfoundation.org> | 2024-07-27 10:33:44 +0200 |
commit | a561145f3ae973ebf3e0aee41624e92a6c5cb38d (patch) | |
tree | 2faf03865f75e6fca613a16169bc44efbe0d7b62 | |
parent | 7f91bd0f2941fa36449ce1a15faaa64f840d9746 (diff) | |
download | linux-stable-a561145f3ae973ebf3e0aee41624e92a6c5cb38d.tar.gz linux-stable-a561145f3ae973ebf3e0aee41624e92a6c5cb38d.tar.bz2 linux-stable-a561145f3ae973ebf3e0aee41624e92a6c5cb38d.zip |
filelock: Fix fcntl/close race recovery compat path
commit f8138f2ad2f745b9a1c696a05b749eabe44337ea upstream.
When I wrote commit 3cad1bc01041 ("filelock: Remove locks reliably when
fcntl/close race is detected"), I missed that there are two copies of the
code I was patching: The normal version, and the version for 64-bit offsets
on 32-bit kernels.
Thanks to Greg KH for stumbling over this while doing the stable
backport...
Apply exactly the same fix to the compat path for 32-bit kernels.
Fixes: c293621bbf67 ("[PATCH] stale POSIX lock handling")
Cc: stable@kernel.org
Link: https://bugs.chromium.org/p/project-zero/issues/detail?id=2563
Signed-off-by: Jann Horn <jannh@google.com>
Link: https://lore.kernel.org/r/20240723-fs-lock-recover-compatfix-v1-1-148096719529@google.com
Signed-off-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-rw-r--r-- | fs/locks.c | 9 |
1 files changed, 4 insertions, 5 deletions
diff --git a/fs/locks.c b/fs/locks.c index 7957a5aa8496..234ebfa8c070 100644 --- a/fs/locks.c +++ b/fs/locks.c @@ -2427,8 +2427,9 @@ int fcntl_setlk64(unsigned int fd, struct file *filp, unsigned int cmd, error = do_lock_file_wait(filp, cmd, file_lock); /* - * Attempt to detect a close/fcntl race and recover by releasing the - * lock that was just acquired. There is no need to do that when we're + * Detect close/fcntl races and recover by zapping all POSIX locks + * associated with this file and our files_struct, just like on + * filp_flush(). There is no need to do that when we're * unlocking though, or for OFD locks. */ if (!error && file_lock->fl_type != F_UNLCK && @@ -2442,9 +2443,7 @@ int fcntl_setlk64(unsigned int fd, struct file *filp, unsigned int cmd, f = fcheck(fd); spin_unlock(¤t->files->file_lock); if (f != filp) { - file_lock->fl_type = F_UNLCK; - error = do_lock_file_wait(filp, cmd, file_lock); - WARN_ON_ONCE(error); + locks_remove_posix(filp, ¤t->files); error = -EBADF; } } |