Re: [PATCH] ALSA: usb-audio: fix out-of-bounds write in snd_usbmidi_akai_output()

[Takashi Iwai <tiwai@suse.de>]

On Sat, 25 Jul 2026 11:40:13 +0200,
Baul Lee wrote:
> 
> 
> snd_usbmidi_akai_output() computes its fill-loop bound
> 
> buf_end = ep->max_transfer - MAX_AKAI_SYSEX_LEN - 1;
> 
> as a signed int, so a small device-advertised bulk-OUT max_transfer makes
> buf_end negative.  The loop guard then compares the u32
> urb->transfer_buffer_length against that negative int: the usual arithmetic
> conversion turns buf_end into a large unsigned value, so the guard stays
> true and each iteration keeps appending SysEx framing and payload bytes
> past the end of the URB transfer buffer, which is only max_transfer bytes
> long.
> 
> A USB device that advertises a tiny bulk-OUT endpoint can therefore trigger
> an attacker-length- and content-controlled heap out-of-bounds write when a
> process writes to the created /dev/snd/midiC*D* node.
> 
> Perform the comparison in signed arithmetic so that a negative buf_end
> stops the loop instead of wrapping to a huge unsigned bound.
> 
> Fixes: 4434ade8c933 ("ALSA: usb-audio: add support for Akai MPD16")
> Discovered by XBOW, triaged by Baul Lee <baul.lee@xbow.com>
> Reported-by: Federico Kirschbaum <federico.kirschbaum@xbow.com>
> Reported-by: Baul Lee <baul.lee@xbow.com>

First of all, please put your Signed-off-by tag.  It's a legal
requirement for upstreaming.

About the patch:

> ---
>  sound/usb/midi.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/sound/usb/midi.c b/sound/usb/midi.c
> index d87e3f357cf7..cc7df77632a1 100644
> --- a/sound/usb/midi.c
> +++ b/sound/usb/midi.c
> @@ -799,7 +799,7 @@ static void snd_usbmidi_akai_output(struct
> snd_usb_midi_out_endpoint *ep,
>   buf_end = ep->max_transfer - MAX_AKAI_SYSEX_LEN - 1;
> 
>   /* only try adding more data when there's space for at least 1 SysEx */
> - while (urb->transfer_buffer_length < buf_end) {
> + while ((int)urb->transfer_buffer_length < buf_end) {
>   count = snd_rawmidi_transmit_peek(substream,
>    tmp, MAX_AKAI_SYSEX_LEN);
>   if (!count) {
> --

The spaces are malformed due to your mailer, hence the patch can't be
applied cleanly via git-am.  Please check your mailer setup.

At best, test once submitting to yourself and check whether you can
apply the patch cleanly via git-am locally beforehand.

BTW, your patches don't seem reached to linux-sound ML by some reason;
I couldn't find your posts in lore.kernel.org.  I'm not sure what went
wrong.

All points above applied to your two other patches, too.


thanks,

Takashi