mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: pip-izony <eeodqql09@gmail.com>
To: Heikki Krogerus <heikki.krogerus@linux.intel.com>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Saranya Gopal <saranya.gopal@intel.com>,
	Rajaram Regupathy <rajaram.regupathy@intel.com>,
	Benson Leung <bleung@chromium.org>,
	Andrei Kuchynski <akuchynski@chromium.org>,
	Jameson Thies <jthies@google.com>,
	Kyungtae Kim <Kyungtae.Kim@dartmouth.edu>,
	linux-usb@vger.kernel.org, linux-kernel@vger.kernel.org,
	Seungjin Bae <eeodqql09@gmail.com>,
	stable@vger.kernel.org
Subject: [PATCH v2] usb: typec: ucsi: reject GET_PDOS responses longer than requested
Date: Fri,  2 Oct 2026 16:21:02 -0400	[thread overview]
Message-ID: <20261002202102.599998-1-eeodqql09@gmail.com> (raw)
In-Reply-To: <2026092932-arbitrate-snorkel-3fca@gregkh>

From: Seungjin Bae <eeodqql09@gmail.com>

In ucsi_get_pdos(), the number of PDOs is derived from the data length
reported in the CCI register, which is provided by the PPM firmware.

The function requests at most UCSI_MAX_PDOS PDOs on the first read and
PDO_MAX_OBJECTS - UCSI_MAX_PDOS on the second, and ucsi_send_command()
only copies that many bytes into the buffer. However, the returned
length is taken from the 8 bit CCI data length field and is not bounded
by the size of the request.

If a malicious PPM reports a larger length, e.g. 0xFF, each
read is counted as 63 PDOs and ucsi_get_pdos() returns up to 126, beyond
PDO_MAX_OBJECTS and the number of PDOs actually read.

ucsi_get_src_pdos() stores this value in con->num_pdos, and
ucsi_psy_get_voltage_max() and ucsi_psy_get_current_max() use it to
index con->src_pdos[con->num_pdos - 1], resulting in an out-of-bounds
read. This happens without any userspace action, since
ucsi_get_src_pdos() calls ucsi_port_psy_changed() and the resulting
uevent reads every property.

Reject such responses in ucsi_read_pdos(), if the PPM reports more data
than was requested, print an error and return -EPROTO, so that the
invalid count is never used.

Fixes: b04e1747fbcc ("usb: typec: ucsi: Register USB Power Delivery Capabilities")
Cc: stable@vger.kernel.org
Assisted-by: LLM
Signed-off-by: Seungjin Bae <eeodqql09@gmail.com>
---
 v1 -> v2: Print an error and fail instead of limiting the count, as suggested
 by Heikki.

 drivers/usb/typec/ucsi/ucsi.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/drivers/usb/typec/ucsi/ucsi.c b/drivers/usb/typec/ucsi/ucsi.c
index bef3f9b71d71..17beec16fe19 100644
--- a/drivers/usb/typec/ucsi/ucsi.c
+++ b/drivers/usb/typec/ucsi/ucsi.c
@@ -862,6 +862,7 @@ static int ucsi_read_pdos(struct ucsi_connector *con,
 			  u32 *pdos, int offset, int num_pdos)
 {
 	struct ucsi *ucsi = con->ucsi;
+	int len = num_pdos * sizeof(u32);
 	u64 command;
 	int ret;
 
@@ -875,11 +876,17 @@ static int ucsi_read_pdos(struct ucsi_connector *con,
 	command |= UCSI_GET_PDOS_PDO_OFFSET(offset);
 	command |= UCSI_GET_PDOS_NUM_PDOS(num_pdos - 1);
 	command |= is_source(role) ? UCSI_GET_PDOS_SRC_PDOS : 0;
-	ret = ucsi_send_command(ucsi, command, pdos + offset,
-				num_pdos * sizeof(u32));
+	ret = ucsi_send_command(ucsi, command, pdos + offset, len);
 	if (ret < 0 && ret != -ETIMEDOUT)
 		dev_err(ucsi->dev, "UCSI_GET_PDOS failed (%d)\n", ret);
 
+	if (ret > len) {
+		dev_err(ucsi->dev,
+			"PPM returned %d bytes of PDOs, but only %d were requested\n",
+			ret, len);
+		return -EPROTO;
+	}
+
 	return ret;
 }
 
-- 
2.43.0


      reply	other threads:[~2026-10-02 20:22 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-09-27 21:49 [PATCH] usb: typec: ucsi: limit the PDO count to the number of PDOs requested pip-izony
2026-09-28 14:15 ` Heikki Krogerus
2026-09-29 13:32 ` Greg Kroah-Hartman
2026-10-02 20:21   ` pip-izony [this message]

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=20261002202102.599998-1-eeodqql09@gmail.com \
    --to=eeodqql09@gmail.com \
    --cc=Kyungtae.Kim@dartmouth.edu \
    --cc=akuchynski@chromium.org \
    --cc=bleung@chromium.org \
    --cc=gregkh@linuxfoundation.org \
    --cc=heikki.krogerus@linux.intel.com \
    --cc=jthies@google.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-usb@vger.kernel.org \
    --cc=rajaram.regupathy@intel.com \
    --cc=saranya.gopal@intel.com \
    --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®