Re: [syzbot] KASAN: use-after-free Read in __snd_rawmidi_transmit_peek

From: Takashi Iwai <tiwai@suse.de>
To: syzbot <syzbot+70e777a39907d6d5fd0a@syzkaller.appspotmail.com>
Cc: alsa-devel@alsa-project.org, coding@diwic.se,
	colin.king@intel.com, linux-kernel@vger.kernel.org,
	perex@perex.cz, syzkaller-bugs@googlegroups.com, tiwai@suse.com
Subject: Re: [syzbot] KASAN: use-after-free Read in __snd_rawmidi_transmit_peek
Date: Tue, 19 Apr 2022 12:00:19 +0200	[thread overview]
Message-ID: <s5hee1t1jgs.wl-tiwai@suse.de> (raw)
In-Reply-To: <s5hk0bl1m6c.wl-tiwai@suse.de>

On Tue, 19 Apr 2022 11:01:47 +0200,
Takashi Iwai wrote:
> 
> On Mon, 18 Apr 2022 12:12:23 +0200,
> syzbot wrote:
> > 
> > Hello,
> > 
> > syzbot found the following issue on:
> > 
> > HEAD commit:    a19944809fe9 Merge tag 'hardening-v5.18-rc3' of git://git...
> > git tree:       upstream
> > console output: https://syzkaller.appspot.com/x/log.txt?x=16a40ae0f00000
> > kernel config:  https://syzkaller.appspot.com/x/.config?x=eb177500e563582f
> > dashboard link: https://syzkaller.appspot.com/bug?extid=70e777a39907d6d5fd0a
> > compiler:       gcc (Debian 10.2.1-6) 10.2.1 20210110, GNU ld (GNU Binutils for Debian) 2.35.2
> > syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=1590dfa8f00000
> > 
> > IMPORTANT: if you fix the issue, please add the following tag to the commit:
> > Reported-by: syzbot+70e777a39907d6d5fd0a@syzkaller.appspotmail.com
> > 
> > ==================================================================
> > BUG: KASAN: use-after-free in __snd_rawmidi_transmit_peek+0x261/0x360 sound/core/rawmidi.c:1286
> 
> Looks like a leftover work.  The fix patch is below.

On the second thought, this would be rather in a different code path,
and the fix should be smaller like below.  Let's see.


Takashi

-- 8< --
From: Takashi Iwai <tiwai@suse.de>
Subject: [PATCH] ALSA: usb-audio: Clear MIDI port active flag after draining

When a rawmidi output stream is closed, it calls the drain at first,
then does trigger-off only when the drain returns -ERESTARTSYS as a
fallback.  It implies that each driver should turn off the stream
properly after the drain.  Meanwhile, USB-audio MIDI interface didn't
change the port->active flag after the drain.  This may leave the
output work picking up the port that is closed right now, which
eventually leads to a use-after-free for the already released rawmidi
object.

This patch fixes the bug by properly clearing the port->active flag
after the output drain.

Reported-by: syzbot+70e777a39907d6d5fd0a@syzkaller.appspotmail.com
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/00000000000011555605dceaff03@google.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
---
 sound/usb/midi.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/sound/usb/midi.c b/sound/usb/midi.c
index 2c01649c70f6..7c6ca2b433a5 100644
--- a/sound/usb/midi.c
+++ b/sound/usb/midi.c
@@ -1194,6 +1194,7 @@ static void snd_usbmidi_output_drain(struct snd_rawmidi_substream *substream)
 		} while (drain_urbs && timeout);
 		finish_wait(&ep->drain_wait, &wait);
 	}
+	port->active = 0;
 	spin_unlock_irq(&ep->buffer_lock);
 }
 
-- 
2.34.1


