From: Sinan Turunc <enscrpt@pm.me>
To: Laurent Pinchart <laurent.pinchart@ideasonboard.com>,
Hans de Goede <hansg@kernel.org>
Cc: mchehab@kernel.org, linux-media@vger.kernel.org,
linux-kernel@vger.kernel.org, stable@vger.kernel.org
Subject: [PATCH] media: uvcvideo: Reject metadata packets shorter than their header
Date: Sun, 15 Nov 2026 12:17:18 +0000 [thread overview]
Message-ID: <20261115121701.189-1-enscrpt@pm.me> (raw)
uvc_video_decode_meta() derives the length of a UVC payload header
from the PTS and SCR flag bits in the packet, not from the packet
length, and with V4L2_META_FMT_UVC it copies that many bytes with
memcpy(meta->buf, &mem[2], length - 2). Nothing compares the derived
length with the packet length, so a packet shorter than the header it
claims is read past its end, and the bytes beyond it - whatever the
URB buffer held there before - are copied into the metadata buffer
that userspace reads.
An emulated UVC camera that sends an 8-byte isochronous packet with a
3-byte header and both the PTS and SCR bits set made the copy take 10
bytes from mem[2]: the last four came from past the packet, and in one
run those four were the tail of the previous packet still in the URB
buffer (03 0c, a UVC header) rather than zeros. With this patch the
same device delivers no metadata blocks at all, and the video stream
is unchanged: 25905 buffers, 5958150 bytes over a 30-second window.
The check is placed after the derived length is final and before the
format override, and treats such a packet the way the function already
treats a 2-byte one: it returns.
Fixes: 088ead255245 ("media: uvcvideo: Add a metadata device node")
Cc: stable@vger.kernel.org
Signed-off-by: Sinan Turunc <enscrpt@pm.me>
---
drivers/media/usb/uvc/uvc_video.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/drivers/media/usb/uvc/uvc_video.c b/drivers/media/usb/uvc/uvc_video.c
index fc3536a4399..25c3b8dc891 100644
--- a/drivers/media/usb/uvc/uvc_video.c
+++ b/drivers/media/usb/uvc/uvc_video.c
@@ -1487,6 +1487,10 @@ static void uvc_video_decode_meta(struct uvc_streaming *stream,
if (has_scr)
len_std += 6;
+ /* A packet shorter than the header it declares is malformed. */
+ if (length < len_std)
+ return;
+
if (stream->meta.format == V4L2_META_FMT_UVC)
length = len_std;
--
2.56.0
next reply other threads:[~2026-11-15 12:17 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-15 12:17 Sinan Turunc [this message]
2026-11-16 2:13 ` Laurent Pinchart
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20261115121701.189-1-enscrpt@pm.me \
--to=enscrpt@pm.me \
--cc=hansg@kernel.org \
--cc=laurent.pinchart@ideasonboard.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-media@vger.kernel.org \
--cc=mchehab@kernel.org \
--cc=stable@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®