[PATCH] ALSA: line6: Clamp the playback URB size to the OUT endpoint packet size

[PATCH] ALSA: line6: Clamp the playback URB size to the OUT endpoint packet size

[Xiang Mei]

The playback buffer is sized from the OUT endpoint's max_packet_size_out,
but submit_audio_out_urb() takes the length to write into it from
prev_fsize, which audio_in_callback() derived from the IN endpoint.  Both
come from the device's own descriptors and nothing relates them, so a
device declaring a large iso IN and a small iso OUT wMaxPacketSize
overflows the buffer.  usb_submit_urb() rejects the oversized URB, but
only after the write has run.

Attaching the device is the whole trigger: toneport_startup() acquires the
monitor stream and starts both URB streams on its own, so the overflow
happens in the URB completion handler with no local process involved.

Clamp the length to max_packet_size_out.  A device whose OUT endpoint
matches the audio format it announces never reaches the limit.

BUG: KASAN: slab-out-of-bounds in submit_audio_out_urb (sound/usb/line6/playback.c:242)
Write of size 1024 at addr ffff88801b94bc00 by task vhci_rx/183

Call Trace:
 kasan_report (mm/kasan/report.c:595)
 kasan_check_range (mm/kasan/generic.c:186 mm/kasan/generic.c:200)
 __asan_memset (mm/kasan/shadow.c:84)
 submit_audio_out_urb (sound/usb/line6/playback.c:242)
 audio_out_callback (sound/usb/line6/playback.c:354)
 __usb_hcd_giveback_urb (drivers/usb/core/hcd.c:1657)
 usb_hcd_giveback_urb (drivers/usb/core/hcd.c:1741)
 vhci_rx_loop (drivers/usb/usbip/vhci_rx.c:107)
 kthread (kernel/kthread.c:436)
 ret_from_fork (arch/x86/kernel/process.c:158)
 ret_from_fork_asm (arch/x86/entry/entry_64.S:245)

The buggy address belongs to the object at ffff88801b94bc00
 which belongs to the cache kmalloc-512 of size 512

Cc: stable@vger.kernel.org
Fixes: 7a0f55aeeb8f ("ALSA: line6: Support assymetrical in/out configurations")
Reported-by: <co+f595d33a1b0a565b@bugs.sh>
Assisted-by: LLM
Signed-off-by: Xiang Mei <xmei5@asu.edu>
---
 sound/usb/line6/playback.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/sound/usb/line6/playback.c b/sound/usb/line6/playback.c
index 7ebaf125f969..cfb20585c5ea 100644
--- a/sound/usb/line6/playback.c
+++ b/sound/usb/line6/playback.c
@@ -181,6 +181,7 @@ static int submit_audio_out_urb(struct snd_line6_pcm *line6pcm)
 		}
 
 		fsize *= bytes_per_frame;
+		fsize = min(fsize, line6pcm->max_packet_size_out);
 
 		fout->offset = urb_size;
 		fout->length = fsize;
-- 
2.43.0

Re: [PATCH] ALSA: line6: Clamp the playback URB size to the OUT endpoint packet size

[Xiang Mei]

On Fri, Sep 18, 2026 at 2:22 PM Xiang Mei <xmei5@asu.edu> wrote:
>
> The playback buffer is sized from the OUT endpoint's max_packet_size_out,
> but submit_audio_out_urb() takes the length to write into it from
> prev_fsize, which audio_in_callback() derived from the IN endpoint.  Both
> come from the device's own descriptors and nothing relates them, so a
> device declaring a large iso IN and a small iso OUT wMaxPacketSize
> overflows the buffer.  usb_submit_urb() rejects the oversized URB, but
> only after the write has run.
>
> Attaching the device is the whole trigger: toneport_startup() acquires the
> monitor stream and starts both URB streams on its own, so the overflow
> happens in the URB completion handler with no local process involved.
>
> Clamp the length to max_packet_size_out.  A device whose OUT endpoint
> matches the audio format it announces never reaches the limit.
>
> BUG: KASAN: slab-out-of-bounds in submit_audio_out_urb (sound/usb/line6/playback.c:242)
> Write of size 1024 at addr ffff88801b94bc00 by task vhci_rx/183
>
> Call Trace:
>  kasan_report (mm/kasan/report.c:595)
>  kasan_check_range (mm/kasan/generic.c:186 mm/kasan/generic.c:200)
>  __asan_memset (mm/kasan/shadow.c:84)
>  submit_audio_out_urb (sound/usb/line6/playback.c:242)
>  audio_out_callback (sound/usb/line6/playback.c:354)
>  __usb_hcd_giveback_urb (drivers/usb/core/hcd.c:1657)
>  usb_hcd_giveback_urb (drivers/usb/core/hcd.c:1741)
>  vhci_rx_loop (drivers/usb/usbip/vhci_rx.c:107)
>  kthread (kernel/kthread.c:436)
>  ret_from_fork (arch/x86/kernel/process.c:158)
>  ret_from_fork_asm (arch/x86/entry/entry_64.S:245)
>
> The buggy address belongs to the object at ffff88801b94bc00
>  which belongs to the cache kmalloc-512 of size 512
>
> Cc: stable@vger.kernel.org
> Fixes: 7a0f55aeeb8f ("ALSA: line6: Support assymetrical in/out configurations")
> Reported-by: <co+f595d33a1b0a565b@bugs.sh>
> Assisted-by: LLM
> Signed-off-by: Xiang Mei <xmei5@asu.edu>
> ---
>  sound/usb/line6/playback.c | 1 +
>  1 file changed, 1 insertion(+)
>
> diff --git a/sound/usb/line6/playback.c b/sound/usb/line6/playback.c
> index 7ebaf125f969..cfb20585c5ea 100644
> --- a/sound/usb/line6/playback.c
> +++ b/sound/usb/line6/playback.c
> @@ -181,6 +181,7 @@ static int submit_audio_out_urb(struct snd_line6_pcm *line6pcm)
>                 }
>
>                 fsize *= bytes_per_frame;
> +               fsize = min(fsize, line6pcm->max_packet_size_out);
>
>                 fout->offset = urb_size;
>                 fout->length = fsize;
> --
> 2.43.0
>

Please disregard this patch. It duplicates my other submission
addressing the same out-of-bounds write in submit_audio_out_urb().

Sorry for the duplicate submission and noise.

Thanks,
Xiang