ipv6: fix NULL dereference in ip6_route_dev_notify()
authorEric Dumazet <edumazet@google.com>
Tue, 15 Aug 2017 11:09:51 +0000 (04:09 -0700)
committerDavid S. Miller <davem@davemloft.net>
Wed, 16 Aug 2017 00:06:34 +0000 (17:06 -0700)
Based on a syzkaller report [1], I found that a per cpu allocation
failure in snmp6_alloc_dev() would then lead to NULL dereference in
ip6_route_dev_notify().

It seems this is a very old bug, thus no Fixes tag in this submission.

Let's add in6_dev_put_clear() helper, as we will probably use
it elsewhere (once available/present in net-next)

[1]
kasan: CONFIG_KASAN_INLINE enabled
kasan: GPF could be caused by NULL-ptr deref or user memory access
general protection fault: 0000 [#1] SMP KASAN
Dumping ftrace buffer:
   (ftrace buffer empty)
Modules linked in:
CPU: 1 PID: 17294 Comm: syz-executor6 Not tainted 4.13.0-rc2+ #10
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
task: ffff88019f456680 task.stack: ffff8801c6e58000
RIP: 0010:__read_once_size include/linux/compiler.h:250 [inline]
RIP: 0010:atomic_read arch/x86/include/asm/atomic.h:26 [inline]
RIP: 0010:refcount_sub_and_test+0x7d/0x1b0 lib/refcount.c:178
RSP: 0018:ffff8801c6e5f1b0 EFLAGS: 00010202
RAX: 0000000000000037 RBX: dffffc0000000000 RCX: ffffc90005d25000
RDX: ffff8801c6e5f218 RSI: ffffffff82342bbf RDI: 0000000000000001
RBP: ffff8801c6e5f240 R08: 0000000000000001 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000000 R12: 1ffff10038dcbe37
R13: 0000000000000006 R14: 0000000000000001 R15: 00000000000001b8
FS:  00007f21e0429700(0000) GS:ffff8801dc100000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000001ddbc22000 CR3: 00000001d632b000 CR4: 00000000001426e0
DR0: 0000000020000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000600
Call Trace:
 refcount_dec_and_test+0x1a/0x20 lib/refcount.c:211
 in6_dev_put include/net/addrconf.h:335 [inline]
 ip6_route_dev_notify+0x1c9/0x4a0 net/ipv6/route.c:3732
 notifier_call_chain+0x136/0x2c0 kernel/notifier.c:93
 __raw_notifier_call_chain kernel/notifier.c:394 [inline]
 raw_notifier_call_chain+0x2d/0x40 kernel/notifier.c:401
 call_netdevice_notifiers_info+0x51/0x90 net/core/dev.c:1678
 call_netdevice_notifiers net/core/dev.c:1694 [inline]
 rollback_registered_many+0x91c/0xe80 net/core/dev.c:7107
 rollback_registered+0x1be/0x3c0 net/core/dev.c:7149
 register_netdevice+0xbcd/0xee0 net/core/dev.c:7587
 register_netdev+0x1a/0x30 net/core/dev.c:7669
 loopback_net_init+0x76/0x160 drivers/net/loopback.c:214
 ops_init+0x10a/0x570 net/core/net_namespace.c:118
 setup_net+0x313/0x710 net/core/net_namespace.c:294
 copy_net_ns+0x27c/0x580 net/core/net_namespace.c:418
 create_new_namespaces+0x425/0x880 kernel/nsproxy.c:107
 unshare_nsproxy_namespaces+0xae/0x1e0 kernel/nsproxy.c:206
 SYSC_unshare kernel/fork.c:2347 [inline]
 SyS_unshare+0x653/0xfa0 kernel/fork.c:2297
 entry_SYSCALL_64_fastpath+0x1f/0xbe
RIP: 0033:0x4512c9
RSP: 002b:00007f21e0428c08 EFLAGS: 00000216 ORIG_RAX: 0000000000000110
RAX: ffffffffffffffda RBX: 0000000000718150 RCX: 00000000004512c9
RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000062020200
RBP: 0000000000000086 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000216 R12: 00000000004b973d
R13: 00000000ffffffff R14: 000000002001d000 R15: 00000000000002dd
Code: 50 2b 34 82 c7 00 f1 f1 f1 f1 c7 40 04 04 f2 f2 f2 c7 40 08 f3 f3
f3 f3 e8 a1 43 39 ff 4c 89 f8 48 8b 95 70 ff ff ff 48 c1 e8 03 <0f> b6
0c 18 4c 89 f8 83 e0 07 83 c0 03 38 c8 7c 08 84 c9 0f 85
RIP: __read_once_size include/linux/compiler.h:250 [inline] RSP:
ffff8801c6e5f1b0
RIP: atomic_read arch/x86/include/asm/atomic.h:26 [inline] RSP:
ffff8801c6e5f1b0
RIP: refcount_sub_and_test+0x7d/0x1b0 lib/refcount.c:178 RSP:
ffff8801c6e5f1b0
---[ end trace e441d046c6410d31 ]---

Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
include/net/addrconf.h
net/ipv6/route.c

index 6df79e96a780b3bf0c141fee5b43eded5276775c..f44ff2476758e27ad9f9cc20d96d96f16e5ee98c 100644 (file)
@@ -336,6 +336,16 @@ static inline void in6_dev_put(struct inet6_dev *idev)
                in6_dev_finish_destroy(idev);
 }
 
+static inline void in6_dev_put_clear(struct inet6_dev **pidev)
+{
+       struct inet6_dev *idev = *pidev;
+
+       if (idev) {
+               in6_dev_put(idev);
+               *pidev = NULL;
+       }
+}
+
 static inline void __in6_dev_put(struct inet6_dev *idev)
 {
        refcount_dec(&idev->refcnt);
index 99d4727f2b186348cacb4e9a9004e1625671ac5c..94d6a13d47f0e9ec5ff4cbc50b90d4cd9ca3f38a 100644 (file)
@@ -3721,10 +3721,10 @@ static int ip6_route_dev_notify(struct notifier_block *this,
                /* NETDEV_UNREGISTER could be fired for multiple times by
                 * netdev_wait_allrefs(). Make sure we only call this once.
                 */
-               in6_dev_put(net->ipv6.ip6_null_entry->rt6i_idev);
+               in6_dev_put_clear(&net->ipv6.ip6_null_entry->rt6i_idev);
 #ifdef CONFIG_IPV6_MULTIPLE_TABLES
-               in6_dev_put(net->ipv6.ip6_prohibit_entry->rt6i_idev);
-               in6_dev_put(net->ipv6.ip6_blk_hole_entry->rt6i_idev);
+               in6_dev_put_clear(&net->ipv6.ip6_prohibit_entry->rt6i_idev);
+               in6_dev_put_clear(&net->ipv6.ip6_blk_hole_entry->rt6i_idev);
 #endif
        }