ALSA: seq: Fix race during FIFO resize

commit 2d7d54002e396c180db0c800c1046f0a3c471597 upstream.

When a new event is queued while processing to resize the FIFO in
snd_seq_fifo_clear(), it may lead to a use-after-free, as the old pool
that is being queued gets removed.  For avoiding this race, we need to
close the pool to be deleted and sync its usage before actually
deleting it.

The issue was spotted by syzkaller.

Reported-by: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Willy Tarreau <w@1wt.eu>

diff --git a/sound/core/seq/seq_fifo.c b/sound/core/seq/seq_fifo.c
index fc2c55b18f49e26882d18f3cde80a0ed6bc8e302..490b697e83ff59ba0bb67bb8291cced5cdf8302b 100644 (file)
--- a/sound/core/seq/seq_fifo.c
+++ b/sound/core/seq/seq_fifo.c
@@ -267,6 +267,10 @@ int snd_seq_fifo_resize(struct snd_seq_fifo *f, int poolsize)
        /* NOTE: overflow flag is not cleared */
        spin_unlock_irqrestore(&f->lock, flags);
 
+       /* close the old pool and wait until all users are gone */
+       snd_seq_pool_mark_closing(oldpool);
+       snd_use_lock_sync(&f->use_lock);
+
        /* release cells in old pool */
        for (cell = oldhead; cell; cell = next) {
                next = cell->next;