linux-kernel - [PATCH 4.11 07/78] [media] cec: race fix: dont return -ENONET in cec

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20170619151029.677309681@linuxfoundation.org>
Date:   Mon, 19 Jun 2017 23:15:07 +0800
From:   Greg Kroah-Hartman <gregkh@...uxfoundation.org>
To:     linux-kernel@...r.kernel.org
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        stable@...r.kernel.org, Hans Verkuil <hans.verkuil@...co.com>,
        Mauro Carvalho Chehab <mchehab@...pensource.com>
Subject: [PATCH 4.11 07/78] [media] cec: race fix: dont return -ENONET in cec_receive()

4.11-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Hans Verkuil <hans.verkuil@...co.com>

commit b94aac64a4c17c5af92f9b4ba7164c5b384d5c02 upstream.

When calling CEC_RECEIVE do not check if the adapter is configured.
Typically CEC_RECEIVE is called after a select() and if that indicates
that there are messages in the receive queue, then you should always be
able to dequeue a message.

The race condition here is that a message has been received and is
queued, so select() tells userspace that a message is available. But
before the application calls CEC_RECEIVE the adapter is unconfigured
(e.g. the HDMI cable is removed). Now select will always report that
there is a message, but calling CEC_RECEIVE will always return -ENONET
because the adapter is no longer configured and so will never actually
dequeue the message.

There is really no need for this check, and in fact the ENONET error
code was never documented for CEC_RECEIVE. This may have been a left-over
of old code that was never updated.

Signed-off-by: Hans Verkuil <hans.verkuil@...co.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@...pensource.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@...uxfoundation.org>

---
 drivers/media/cec/cec-api.c |    8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)

--- a/drivers/media/cec/cec-api.c
+++ b/drivers/media/cec/cec-api.c
@@ -267,16 +267,10 @@ static long cec_receive(struct cec_adapt
 			bool block, struct cec_msg __user *parg)
 {
 	struct cec_msg msg = {};
-	long err = 0;
+	long err;

 	if (copy_from_user(&msg, parg, sizeof(msg)))
 		return -EFAULT;
-	mutex_lock(&adap->lock);
-	if (!adap->is_configured && fh->mode_follower < CEC_MODE_MONITOR)
-		err = -ENONET;
-	mutex_unlock(&adap->lock);
-	if (err)
-		return err;

 	err = cec_receive_msg(fh, &msg, block);
 	if (err)