[PATCH] ax25: Fix memory leaks caused by ax25_cb_del()

[PATCH] ax25: Fix memory leaks caused by ax25_cb_del()

[Duoming Zhou]

The previous commit d01ffb9eee4a ("ax25: add refcount in ax25_dev to
avoid UAF bugs") and commit feef318c855a ("ax25: fix UAF bugs of
net_device caused by rebinding operation") increase the refcounts of
ax25_dev and net_device in ax25_bind() and decrease the matching refcounts
in ax25_kill_by_device() in order to prevent UAF bugs. But there are memory
leaks.

If we use ax25_bind() to increase the refcounts of ax25_dev and
net_device, then, use ax25_cb_del() invoked by ax25_destroy_socket()
to delete ax25_cb in hlist before calling ax25_kill_by_device(), the
decrements of refcounts in ax25_kill_by_device() will not be executed,
because ax25_cb is deleted from the hlist.

This patch adds two flags in ax25_dev in order to prevent memory leaks.
If we bind successfully, then, use ax25_cb_del() to delete ax25_cb,
the two "test_bit" condition checks in ax25_kill_by_device() could
pass and the refcounts could be decreased properly.

Fixes: d01ffb9eee4a ("ax25: add refcount in ax25_dev to avoid UAF bugs")
Fixes: feef318c855a ("ax25: fix UAF bugs of net_device caused by rebinding operation")
Reported-by: Thomas Osterried <thomas@osterried.de>
Signed-off-by: Duoming Zhou <duoming@zju.edu.cn>
---
 include/net/ax25.h  |  7 +++++--
 net/ax25/af_ax25.c  | 10 ++++++----
 net/ax25/ax25_dev.c |  3 ++-
 3 files changed, 13 insertions(+), 7 deletions(-)

diff --git a/include/net/ax25.h b/include/net/ax25.h
index 8221af1811d..50b3eacada0 100644
--- a/include/net/ax25.h
+++ b/include/net/ax25.h
@@ -157,7 +157,9 @@ enum {
 #define AX25_DEF_PACLEN		256			/* Paclen=256 */
 #define	AX25_DEF_PROTOCOL	AX25_PROTO_STD_SIMPLEX	/* Standard AX.25 */
 #define AX25_DEF_DS_TIMEOUT	180000			/* DAMA timeout 3 minutes */
-
+#define AX25_DEV_INIT    0
+#define AX25_DEV_KILL    1
+#define AX25_DEV_BIND    2
 typedef struct ax25_uid_assoc {
 	struct hlist_node	uid_node;
 	refcount_t		refcount;
@@ -240,8 +242,9 @@ typedef struct ax25_dev {
 	ax25_dama_info		dama;
 #endif
 	refcount_t		refcount;
+	unsigned long   kill_flag;
+	unsigned long   bind_flag;
 } ax25_dev;
-
 typedef struct ax25_cb {
 	struct hlist_node	ax25_node;
 	ax25_address		source_addr, dest_addr;
diff --git a/net/ax25/af_ax25.c b/net/ax25/af_ax25.c
index 6bd09718077..5519448378d 100644
--- a/net/ax25/af_ax25.c
+++ b/net/ax25/af_ax25.c
@@ -86,6 +86,7 @@ static void ax25_kill_by_device(struct net_device *dev)
 again:
 	ax25_for_each(s, &ax25_list) {
 		if (s->ax25_dev == ax25_dev) {
+			set_bit(AX25_DEV_KILL, &ax25_dev->kill_flag);
 			sk = s->sk;
 			if (!sk) {
 				spin_unlock_bh(&ax25_list_lock);
@@ -114,9 +115,12 @@ static void ax25_kill_by_device(struct net_device *dev)
 			goto again;
 		}
 	}
+	if(!test_bit(AX25_DEV_KILL, &ax25_dev->kill_flag) && test_bit(AX25_DEV_BIND, &ax25_dev->bind_flag)) {
+		dev_put_track(ax25_dev->dev, &ax25_dev->dev_tracker);
+		ax25_dev_put(ax25_dev);
+	}
 	spin_unlock_bh(&ax25_list_lock);
 }
-
 /*
  *	Handle device status changes.
  */
@@ -1132,13 +1136,11 @@ static int ax25_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
 done:
 	ax25_cb_add(ax25);
 	sock_reset_flag(sk, SOCK_ZAPPED);
-
+	set_bit(AX25_DEV_BIND, &ax25_dev->bind_flag);
 out:
 	release_sock(sk);
-
 	return err;
 }
-
 /*
  *	FIXME: nonblock behaviour looks like it may have a bug.
  */
diff --git a/net/ax25/ax25_dev.c b/net/ax25/ax25_dev.c
index d2a244e1c26..7c40914e5c9 100644
--- a/net/ax25/ax25_dev.c
+++ b/net/ax25/ax25_dev.c
@@ -77,7 +77,8 @@ void ax25_dev_device_up(struct net_device *dev)
 	ax25_dev->values[AX25_VALUES_PACLEN]	= AX25_DEF_PACLEN;
 	ax25_dev->values[AX25_VALUES_PROTOCOL]  = AX25_DEF_PROTOCOL;
 	ax25_dev->values[AX25_VALUES_DS_TIMEOUT]= AX25_DEF_DS_TIMEOUT;
-
+	ax25_dev->kill_flag = AX25_DEV_INIT;
+	ax25_dev->bind_flag = AX25_DEV_INIT;
 #if defined(CONFIG_AX25_DAMA_SLAVE) || defined(CONFIG_AX25_DAMA_MASTER)
 	ax25_ds_setup_timer(ax25_dev);
 #endif
--
2.17.1

Re: [PATCH] ax25: Fix memory leaks caused by ax25_cb_del()

[Dan Carpenter]

This is a very frustrating patch because you make a lot of unnecessary
white space changes and you didn't run checkpatch on your patch.

The whole approach feels like the wrong thing...

I have read your commit message, but I don't understand why we can't
just use normal refcounting.  It sounds like there is a layering
violation somewhere?

Even if we go with this approach ->kill_flag and ->bind_flag should be
booleans.  It makes no sense to have a unsigned long where only BIT(2)
can be set.

regards,
dan carpenter

Re: Re: [PATCH] ax25: Fix memory leaks caused by ax25_cb_del()

[周多明]

Hello,

Thank you very much for pointing the wrong places in my patch.

> This is a very frustrating patch because you make a lot of unnecessary
> white space changes and you didn't run checkpatch on your patch.
> 
> The whole approach feels like the wrong thing...

I will fix it.

> I have read your commit message, but I don't understand why we can't
> just use normal refcounting.  It sounds like there is a layering
> violation somewhere?

The root cause of refcount leak is shown below:

     (Thread 1)                          |      (Thread 2)
ax25_bind()                              |
 ...                                     |
 ax25_addr_ax25dev()                     | 
  ax25_dev_hold()   //(1)                |
 ...                                     |
 dev_hold_track()   //(2)                | 
 ...                                     | ax25_destroy_socket()
                                         |  ax25_cb_del()   
                                         |   ... 
                                         |   spin_lock_bh(&ax25_list_lock);
                                         |   hlist_del_init(&ax25->ax25_node); //(3) 
                                         |   spin_unlock_bh(&ax25_list_lock);
                                    
     (thread 3)                                          
ax25_kill_by_device()                    |
 spin_lock_bh(&ax25_list_lock);          |
 ax25_for_each(s, &ax25_list) {          |
  if (s->ax25_dev == ax25_dev)  //(4)    |   
  ...                                    |
  (the following code could not execute) |

Firstly, we use ax25_bind() to increase the refcount of ax25_dev in 
position (1) and increase the refcount of net_device in position (2).

Then, we use ax25_cb_del() invoked by ax25_destroy_socket()
to delete ax25_cb in hlist in position (3) before calling ax25_kill_by_device().
 
Finally, the decrements of refcounts in ax25_kill_by_device() will not be executed,
because no s->ax25_dev equals to ax25_dev in position (4).

My patch adds two flags in ax25_dev in order to prevent reference count leaks. 
If the above condition happens, the two "test_bit" checks in ax25_kill_by_device()
could pass and the refcounts could be decreased properly.

> Even if we go with this approach ->kill_flag and ->bind_flag should be
> booleans.  It makes no sense to have a unsigned long where only BIT(2)
> can be set.

I will change kill_flag and bind_flag to booleans.

Best wishes,
Duoming Zhou