public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Peter Hurley <peter@hurleysoftware.com>
To: Marcel Holtmann <marcel@holtmann.org>
Cc: Gustavo Padovan <gustavo@padovan.org>,
	Johan Hedberg <johan.hedberg@gmail.com>,
	Gianluca Anzolin <gianluca@sottospazio.it>,
	Alexander Holler <holler@ahsoftware.de>,
	Andrey Vihrov <andrey.vihrov@gmail.com>,
	Sander Eikelenboom <linux@eikelenboom.it>,
	linux-bluetooth@vger.kernel.org, linux-kernel@vger.kernel.org,
	Peter Hurley <peter@hurleysoftware.com>
Subject: [PATCH 15/24] Bluetooth: Fix unsafe RFCOMM device parenting
Date: Sun,  9 Feb 2014 20:59:15 -0500	[thread overview]
Message-ID: <1391997564-1805-16-git-send-email-peter@hurleysoftware.com> (raw)
In-Reply-To: <1391997564-1805-1-git-send-email-peter@hurleysoftware.com>

Accessing the results of hci_conn_hash_lookup_ba() is unsafe without
holding the hci_dev_lock() during the lookup. For example:

CPU 0                             | CPU 1
hci_conn_hash_lookup_ba           | hci_conn_del
  rcu_read_lock                   |   hci_conn_hash_del
  list_for_each_entry_rcu         |     list_del_rcu
    if (.....)                    |       synchronize_rcu
      rcu_read_unlock             |
                                  |   hci_conn_del_sysfs
                                  |   hci_dev_put
                                  |   hci_conn_put
                                  |     put_device (last reference)
                                  |       bt_link_release
                                  |         kfree(conn)
      return p  << just freed     |

Even if a hci_conn reference were taken (via hci_conn_get), would
not guarantee the lifetime of the sysfs device, but only safe
access to the in-memory structure.

Ensure the hci_conn device stays valid while the rfcomm device
is reparented; rename rfcomm_get_device() to rfcomm_reparent_device()
and perform the reparenting within the function while holding the
hci_dev_lock.

Signed-off-by: Peter Hurley <peter@hurleysoftware.com>
---
 net/bluetooth/rfcomm/tty.c | 20 ++++++++++++++------
 1 file changed, 14 insertions(+), 6 deletions(-)

diff --git a/net/bluetooth/rfcomm/tty.c b/net/bluetooth/rfcomm/tty.c
index a58d693..2975bc4 100644
--- a/net/bluetooth/rfcomm/tty.c
+++ b/net/bluetooth/rfcomm/tty.c
@@ -167,20 +167,29 @@ static struct rfcomm_dev *rfcomm_dev_get(int id)
 	return dev;
 }
 
-static struct device *rfcomm_get_device(struct rfcomm_dev *dev)
+static void rfcomm_reparent_device(struct rfcomm_dev *dev)
 {
 	struct hci_dev *hdev;
 	struct hci_conn *conn;
 
 	hdev = hci_get_route(&dev->dst, &dev->src);
 	if (!hdev)
-		return NULL;
+		return;
 
+	/* The lookup results are unsafe to access without the
+	 * hci device lock (FIXME: why is this not documented?)
+	 */
+	hci_dev_lock(hdev);
 	conn = hci_conn_hash_lookup_ba(hdev, ACL_LINK, &dev->dst);
 
-	hci_dev_put(hdev);
+	/* Just because the acl link is in the hash table is no
+	 * guarantee the sysfs device has been added ...
+	 */
+	if (conn && device_is_registered(&conn->dev))
+		device_move(dev->tty_dev, &conn->dev, DPM_ORDER_DEV_AFTER_PARENT);
 
-	return conn ? &conn->dev : NULL;
+	hci_dev_unlock(hdev);
+	hci_dev_put(hdev);
 }
 
 static ssize_t show_address(struct device *tty_dev, struct device_attribute *attr, char *buf)
@@ -589,8 +598,7 @@ static void rfcomm_dev_state_change(struct rfcomm_dlc *dlc, int err)
 
 	dev->err = err;
 	if (dlc->state == BT_CONNECTED) {
-		device_move(dev->tty_dev, rfcomm_get_device(dev),
-			    DPM_ORDER_DEV_AFTER_PARENT);
+		rfcomm_reparent_device(dev);
 
 		wake_up_interruptible(&dev->port.open_wait);
 	} else if (dlc->state == BT_CLOSED)
-- 
1.8.1.2


  parent reply	other threads:[~2014-02-10  2:20 UTC|newest]

Thread overview: 46+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-02-10  1:59 [PATCH 00/24] rfcomm fixes Peter Hurley
2014-02-10  1:59 ` [PATCH 01/24] Revert "Bluetooth: Remove rfcomm_carrier_raised()" Peter Hurley
2014-02-10  1:59 ` [PATCH 02/24] Revert "Bluetooth: Always wait for a connection on RFCOMM open()" Peter Hurley
2014-02-10  1:59 ` [PATCH 03/24] Revert "Bluetooth: Move rfcomm_get_device() before rfcomm_dev_activate()" Peter Hurley
2014-02-10  1:59 ` [PATCH 04/24] tty: Fix ref counting for port krefs Peter Hurley
2014-02-13 18:36   ` Greg Kroah-Hartman
2014-02-10  1:59 ` [PATCH 05/24] Bluetooth: Fix racy acquire of rfcomm_dev reference Peter Hurley
2014-02-10  1:59 ` [PATCH 06/24] Bluetooth: Exclude released devices from RFCOMMGETDEVLIST ioctl Peter Hurley
2014-02-10  1:59 ` [PATCH 07/24] Bluetooth: Release rfcomm_dev only once Peter Hurley
2014-02-10  1:59 ` [PATCH 08/24] Bluetooth: Fix unreleased rfcomm_dev reference Peter Hurley
2014-02-10  1:59 ` [PATCH 09/24] Bluetooth: Fix RFCOMM tty teardown race Peter Hurley
2014-02-10  1:59 ` [PATCH 10/24] Bluetooth: Verify dlci not in use before rfcomm_dev create Peter Hurley
2014-02-10  1:59 ` [PATCH 11/24] Bluetooth: Simplify RFCOMM session state eval Peter Hurley
2014-02-10  1:59 ` [PATCH 12/24] Bluetooth: Refactor deferred setup test in rfcomm_dlc_close() Peter Hurley
2014-02-10  1:59 ` [PATCH 13/24] Bluetooth: Refactor dlc disconnect logic " Peter Hurley
2014-02-10  1:59 ` [PATCH 14/24] Bluetooth: Directly close dlc for not yet started RFCOMM session Peter Hurley
2014-02-10  1:59 ` Peter Hurley [this message]
2014-02-10  1:59 ` [PATCH 16/24] Bluetooth: Fix RFCOMM parent device for reused dlc Peter Hurley
2014-02-10  1:59 ` [PATCH 17/24] Bluetooth: Rename __rfcomm_dev_get() to __rfcomm_dev_lookup() Peter Hurley
2014-02-10  1:59 ` [PATCH 18/24] Bluetooth: Serialize RFCOMMCREATEDEV and RFCOMMRELEASEDEV ioctls Peter Hurley
2014-02-10  1:59 ` [PATCH 19/24] Bluetooth: Refactor rfcomm_dev_add() Peter Hurley
2014-02-10  1:59 ` [PATCH 20/24] Bluetooth: Cleanup RFCOMM device registration error handling Peter Hurley
2014-02-10  1:59 ` [PATCH 21/24] Bluetooth: Force -EIO from tty read/write if .activate() fails Peter Hurley
2014-02-10  1:59 ` [PATCH 22/24] Bluetooth: Don't fail RFCOMM tty writes Peter Hurley
2014-02-10  1:59 ` [PATCH 23/24] Bluetooth: Refactor write_room() calculation Peter Hurley
2014-02-10  1:59 ` [PATCH 24/24] Bluetooth: Fix " Peter Hurley
2014-02-10 22:09 ` [PATCH 00/24] rfcomm fixes Marcel Holtmann
2014-02-10 23:00   ` Peter Hurley
2014-02-12 22:58     ` Marcel Holtmann
2014-02-13  0:38       ` Peter Hurley
2014-02-13 21:48         ` Alexander Holler
2014-02-12 11:06   ` Sander Eikelenboom
2014-03-03 19:38     ` Sander Eikelenboom
2014-03-10  8:38       ` [RC6 Bell Chime] " Sander Eikelenboom
2014-03-10 15:08         ` John W. Linville
2014-03-11 15:14           ` [RC6 Bell Chime] " Marcel Holtmann
2014-03-14  0:49             ` Sander Eikelenboom
2014-03-14  1:28               ` Marcel Holtmann
2014-03-14  1:29               ` Peter Hurley
2014-03-15 13:51                 ` Sander Eikelenboom
2014-03-15 17:53                   ` Linus Torvalds
2014-03-15 20:45                     ` Peter Hurley
2014-03-15 22:20                       ` Sander Eikelenboom
2014-03-16  0:16                       ` Linus Torvalds
2014-02-13 21:41 ` Alexander Holler
2014-02-14 21:45 ` Marcel Holtmann

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=1391997564-1805-16-git-send-email-peter@hurleysoftware.com \
    --to=peter@hurleysoftware.com \
    --cc=andrey.vihrov@gmail.com \
    --cc=gianluca@sottospazio.it \
    --cc=gustavo@padovan.org \
    --cc=holler@ahsoftware.de \
    --cc=johan.hedberg@gmail.com \
    --cc=linux-bluetooth@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux@eikelenboom.it \
    --cc=marcel@holtmann.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