Thread (3 messages) 3 messages, 2 authors, 2022-02-17

Re: [PATCH] drivers: hamradio: 6pack: fix UAF bug caused by mod_timer()

From: Jakub Kicinski <kuba@kernel.org>
Date: 2022-02-16 04:40:10
Also in: linux-hams, lkml

On Wed, 16 Feb 2022 10:35:49 +0800 Duoming Zhou wrote:
Although del_timer_sync() in sixpack_close() waits for the timer handler
to finish its execution and then releases the timer, the mod_timer()
in sp_xmit_on_air() could be called by userspace syscall such as
ax25_sendmsg(), ax25_connect() and ax25_ioctl() and wakes up the timer
again. If the timer uses sp_xmit_on_air() to write data on pty work queue
that already released by unregister_netdev(), the UAF bug will happen.
Do you mean sp->xbuff access? It's released right before the netdev
itself is freed. Checking dev->something is also a UAF.
One of the possible race conditions is shown below:

      (USE)                     |      (FREE)
ax25_sendmsg()                  |
  ax25_queue_xmit()             |
    ...                         |
    sp_encaps()                 |  sixpack_close()
      sp_xmit_on_air()          |    del_timer_sync(&sp->tx_t)
        mod_timer(&sp->tx_t,..) |    ...
        (wait a while)          |    unregister_netdev(sp->dev)) //FREE
Please clarify what is getting freed. 
      sp_xmit_on_air()          |    ...
        pty_write()             |
          queue_work_on() //USE |

The corresponding fail log is shown below:
===============================================================
BUG: KASAN: use-after-free in __run_timers.part.0+0x170/0x470
Write of size 8 at addr ffff88800a652ab8 by task swapper/2/0
...
Call Trace:
  ...
  queue_work_on+0x3f/0x50
  pty_write+0xcd/0xe0pty_write+0xcd/0xe0
  sp_xmit_on_air+0xb2/0x1f0
  call_timer_fn+0x28/0x150
  __run_timers.part.0+0x3c2/0x470
  run_timer_softirq+0x3b/0x80
  __do_softirq+0xf1/0x380
  ...

This patch add condition check in sp_xmit_on_air(). If the
registration status of net_device is not equal to NETREG_REGISTERED,
the sp_xmit_on_air() will not write data to pty work queue and
return instead.
I don't think this works as mentioned above. The question is why the tx
queue is not stopped.
quoted hunk ↗ jump to hunk
diff --git a/drivers/net/hamradio/6pack.c b/drivers/net/hamradio/6pack.c
index b1fc153125d..7ee25e06915 100644
--- a/drivers/net/hamradio/6pack.c
+++ b/drivers/net/hamradio/6pack.c
@@ -141,7 +141,8 @@ static void sp_xmit_on_air(struct timer_list *t)
 	struct sixpack *sp = from_timer(sp, t, tx_t);
 	int actual, when = sp->slottime;
 	static unsigned char random;
-
+	if (sp->dev->reg_state !=  NETREG_REGISTERED)
+		return;
 	random = random * 17 + 41;
 
 	if (((sp->status1 & SIXP_DCD_MASK) == 0) && (random < sp->persistence)) {
  
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help