4.9-stable review patch. If anyone has any objections, please let me know.
------------------
From: Kirill Tkhai ktkhai@virtuozzo.com
[ Upstream commit 7a107c0f55a3b4c6f84a4323df5610360bde1684 ]
I observed the following deadlock between them:
[task 1] [task 2] [task 3] kill_fasync() mm_update_next_owner() copy_process() spin_lock_irqsave(&fa->fa_lock) read_lock(&tasklist_lock) write_lock_irq(&tasklist_lock) send_sigio() <IRQ> ... read_lock(&fown->lock) kill_fasync() ... read_lock(&tasklist_lock) spin_lock_irqsave(&fa->fa_lock) ...
Task 1 can't acquire read locked tasklist_lock, since there is already task 3 expressed its wish to take the lock exclusive. Task 2 holds the read locked lock, but it can't take the spin lock.
Also, there is possible another deadlock (which I haven't observed):
[task 1] [task 2] f_getown() kill_fasync() read_lock(&f_own->lock) spin_lock_irqsave(&fa->fa_lock,) <IRQ> send_sigio() write_lock_irq(&f_own->lock) kill_fasync() read_lock(&fown->lock) spin_lock_irqsave(&fa->fa_lock,)
Actually, we do not need exclusive fa->fa_lock in kill_fasync_rcu(), as it guarantees fa->fa_file->f_owner integrity only. It may seem, that it used to give a task a small possibility to receive two sequential signals, if there are two parallel kill_fasync() callers, and task handles the first signal fastly, but the behaviour won't become different, since there is exclusive sighand lock in do_send_sig_info().
The patch converts fa_lock into rwlock_t, and this fixes two above deadlocks, as rwlock is allowed to be taken from interrupt handler by qrwlock design.
Signed-off-by: Kirill Tkhai ktkhai@virtuozzo.com Signed-off-by: Jeff Layton jlayton@redhat.com Signed-off-by: Sasha Levin alexander.levin@microsoft.com Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- fs/fcntl.c | 15 +++++++-------- include/linux/fs.h | 2 +- 2 files changed, 8 insertions(+), 9 deletions(-)
--- a/fs/fcntl.c +++ b/fs/fcntl.c @@ -588,9 +588,9 @@ int fasync_remove_entry(struct file *fil if (fa->fa_file != filp) continue;
- spin_lock_irq(&fa->fa_lock); + write_lock_irq(&fa->fa_lock); fa->fa_file = NULL; - spin_unlock_irq(&fa->fa_lock); + write_unlock_irq(&fa->fa_lock);
*fp = fa->fa_next; call_rcu(&fa->fa_rcu, fasync_free_rcu); @@ -635,13 +635,13 @@ struct fasync_struct *fasync_insert_entr if (fa->fa_file != filp) continue;
- spin_lock_irq(&fa->fa_lock); + write_lock_irq(&fa->fa_lock); fa->fa_fd = fd; - spin_unlock_irq(&fa->fa_lock); + write_unlock_irq(&fa->fa_lock); goto out; }
- spin_lock_init(&new->fa_lock); + rwlock_init(&new->fa_lock); new->magic = FASYNC_MAGIC; new->fa_file = filp; new->fa_fd = fd; @@ -704,14 +704,13 @@ static void kill_fasync_rcu(struct fasyn { while (fa) { struct fown_struct *fown; - unsigned long flags;
if (fa->magic != FASYNC_MAGIC) { printk(KERN_ERR "kill_fasync: bad magic number in " "fasync_struct!\n"); return; } - spin_lock_irqsave(&fa->fa_lock, flags); + read_lock(&fa->fa_lock); if (fa->fa_file) { fown = &fa->fa_file->f_owner; /* Don't send SIGURG to processes which have not set a @@ -720,7 +719,7 @@ static void kill_fasync_rcu(struct fasyn if (!(sig == SIGURG && fown->signum == 0)) send_sigio(fown, fa->fa_fd, band); } - spin_unlock_irqrestore(&fa->fa_lock, flags); + read_unlock(&fa->fa_lock); fa = rcu_dereference(fa->fa_next); } } --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1277,7 +1277,7 @@ static inline int locks_lock_file_wait(s }
struct fasync_struct { - spinlock_t fa_lock; + rwlock_t fa_lock; int magic; int fa_fd; struct fasync_struct *fa_next; /* singly linked list */