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]
Date:   Mon, 11 Jan 2021 14:01:51 +0100
From:   Greg Kroah-Hartman <gregkh@...uxfoundation.org>
To:     linux-kernel@...r.kernel.org
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        stable@...r.kernel.org,
        Anant Thazhemadam <anant.thazhemadam@...il.com>,
        Hans de Goede <hdegoede@...hat.com>,
        Marcel Holtmann <marcel@...tmann.org>
Subject: [PATCH 5.4 47/92] Bluetooth: revert: hci_h5: close serdev device and free hu in h5_close

From: Hans de Goede <hdegoede@...hat.com>

commit 5c3b5796866f85354a5ce76a28f8ffba0dcefc7e upstream.

There have been multiple revisions of the patch fix the h5->rx_skb
leak. Accidentally the first revision (which is buggy) and v5 have
both been merged:

v1 commit 70f259a3f427 ("Bluetooth: hci_h5: close serdev device and free
hu in h5_close");
v5 commit 855af2d74c87 ("Bluetooth: hci_h5: fix memory leak in h5_close")

The correct v5 makes changes slightly higher up in the h5_close()
function, which allowed both versions to get merged without conflict.

The changes from v1 unconditionally frees the h5 data struct, this
is wrong because in the serdev enumeration case the memory is
allocated in h5_serdev_probe() like this:

        h5 = devm_kzalloc(dev, sizeof(*h5), GFP_KERNEL);

So its lifetime is tied to the lifetime of the driver being bound
to the serdev and it is automatically freed when the driver gets
unbound. In the serdev case the same h5 struct is re-used over
h5_close() and h5_open() calls and thus MUST not be free-ed in
h5_close().

The serdev_device_close() added to h5_close() is incorrect in the
same way, serdev_device_close() is called on driver unbound too and
also MUST no be called from h5_close().

This reverts the changes made by merging v1 of the patch, so that
just the changes of the correct v5 remain.

Cc: Anant Thazhemadam <anant.thazhemadam@...il.com>
Signed-off-by: Hans de Goede <hdegoede@...hat.com>
Signed-off-by: Marcel Holtmann <marcel@...tmann.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@...uxfoundation.org>

---
 drivers/bluetooth/hci_h5.c |    8 ++------
 1 file changed, 2 insertions(+), 6 deletions(-)

--- a/drivers/bluetooth/hci_h5.c
+++ b/drivers/bluetooth/hci_h5.c
@@ -250,12 +250,8 @@ static int h5_close(struct hci_uart *hu)
 	if (h5->vnd && h5->vnd->close)
 		h5->vnd->close(h5);
 
-	if (hu->serdev)
-		serdev_device_close(hu->serdev);
-
-	kfree_skb(h5->rx_skb);
-	kfree(h5);
-	h5 = NULL;
+	if (!hu->serdev)
+		kfree(h5);
 
 	return 0;
 }


Powered by blists - more mailing lists