Re: KASAN: use-after-free Read in tipc_group_size

On Mon, Jan 08, 2018 at 08:11:35PM +0000, Jon Maloy wrote:

quoted

-----Original Message-----
From: Cong Wang [mailto:<redacted>]
Sent: Monday, January 08, 2018 13:44
To: syzbot <redacted>
Cc: David Miller <davem@davemloft.net>; Jon Maloy
[off-list ref]; LKML [off-list ref]; Linux
Kernel Network Developers [off-list ref]; syzkaller-
bugs@googlegroups.com; tipc-discussion@lists.sourceforge.net; Ying Xue
[off-list ref]
Subject: Re: KASAN: use-after-free Read in tipc_group_size

On Mon, Jan 8, 2018 at 6:58 AM, syzbot
[off-list ref] wrote:

quoted

Hello,

syzkaller hit the following crash on
b2cd1df66037e7c4697c7e40496bf7e4a5e16a2d
git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/maste
r
compiler: gcc (GCC) 7.1.1 20170620
.config is attached
Raw console output is attached.
C reproducer is attached
syzkaller reproducer is attached. See https://goo.gl/kgGztJ for
information about syzkaller reproducers


IMPORTANT: if you fix the bug, please add the following tag to the commit:
Reported-by: <redacted>
It will help syzbot understand when the bug is fixed. See footer for
details.
If you forward the report, please keep this part and the footer.

==========================================================
========

quoted

BUG: KASAN: use-after-free in tipc_group_size+0x40/0x50
net/tipc/group.c:158 Read of size 2 at addr ffff8801c08ba280 by task
syzkaller447710/3513

CPU: 0 PID: 3513 Comm: syzkaller447710 Not tainted 4.15.0-rc7+ #252
Hardware name: Google Google Compute Engine/Google Compute Engine,
BIOS Google 01/01/2011 Call Trace:
 __dump_stack lib/dump_stack.c:17 [inline]
 dump_stack+0x194/0x257 lib/dump_stack.c:53
 print_address_description+0x73/0x250 mm/kasan/report.c:252
kasan_report_error mm/kasan/report.c:351 [inline]
 kasan_report+0x25b/0x340 mm/kasan/report.c:409
 __asan_report_load2_noabort+0x14/0x20 mm/kasan/report.c:428
 tipc_group_size+0x40/0x50 net/tipc/group.c:158
 tipc_poll+0x374/0x4f0 net/tipc/socket.c:739

Seems we have to lock the sock for tipc_group_size() in tipc_poll().

Not quite. I think it is that we initialize 'grp' on the stack before we call sock_poll_wait() and access it after it returns.
This is anyway fixed in patch #9 of the series I just sent to net-next, where the poll() handling for group members is redesigned.

///jon

Last occurred on Jan 16.  Seems to have been fixed by commit 60c2530696320:

#syz fix: tipc: fix race between poll() and setsockopt()

- Eric