On Mon, Dec 4, 2017 at 4:43 AM, Ying Xue <ying....@windriver.com> wrote:
> On 12/04/2017 11:54 AM, Cong Wang wrote:
>> In tipc_topsrv_kern_subscr() when s->tipc_conn_new() fails
>> we call tipc_close_conn() to clean up, but in this case
>> con->usr_data is NULL, tipc_subscrb_delete() should be skipped.
>>
>> This fixes the folllowing crash:
>>
>> 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: 0 PID: 3085 Comm: syzkaller064164 Not tainted 4.15.0-rc1+ #137
>> Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS
>> Google 01/01/2011
>> task: 00000000c24413a5 task.stack: 000000005e8160b5
>> RIP: 0010:__lock_acquire+0xd55/0x47f0 kernel/locking/lockdep.c:3378
>> RSP: 0018:ffff8801cb5474a8 EFLAGS: 00010002
>> RAX: dffffc0000000000 RBX: 0000000000000000 RCX: 0000000000000000
>> RDX: 0000000000000004 RSI: 0000000000000000 RDI: ffffffff85ecb400
>> RBP: ffff8801cb547830 R08: 0000000000000001 R09: 0000000000000000
>> R10: 0000000000000000 R11: ffffffff87489d60 R12: ffff8801cd2980c0
>> R13: 0000000000000000 R14: 0000000000000001 R15: 0000000000000020
>> FS: 00000000014ee880(0000) GS:ffff8801db400000(0000) knlGS:0000000000000000
>> CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> CR2: 00007ffee2426e40 CR3: 00000001cb85a000 CR4: 00000000001406f0
>> DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
>> DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
>> Call Trace:
>> lock_acquire+0x1d5/0x580 kernel/locking/lockdep.c:4004
>> __raw_spin_lock_bh include/linux/spinlock_api_smp.h:135 [inline]
>> _raw_spin_lock_bh+0x31/0x40 kernel/locking/spinlock.c:175
>> spin_lock_bh include/linux/spinlock.h:320 [inline]
>> tipc_subscrb_subscrp_delete+0x8f/0x470 net/tipc/subscr.c:201
>> tipc_subscrb_delete net/tipc/subscr.c:238 [inline]
>> tipc_subscrb_release_cb+0x17/0x30 net/tipc/subscr.c:316
>> tipc_close_conn+0x171/0x270 net/tipc/server.c:204
>> tipc_topsrv_kern_subscr+0x724/0x810 net/tipc/server.c:514
>> tipc_group_create+0x702/0x9c0 net/tipc/group.c:184
>> tipc_sk_join net/tipc/socket.c:2747 [inline]
>> tipc_setsockopt+0x249/0xc10 net/tipc/socket.c:2861
>> SYSC_setsockopt net/socket.c:1851 [inline]
>> SyS_setsockopt+0x189/0x360 net/socket.c:1830
>> entry_SYSCALL_64_fastpath+0x1f/0x96
>>
>> Reported-by: syzbot <syzkal...@googlegroups.com>
>> Cc: Jon Maloy <jon.ma...@ericsson.com>
>> Cc: Ying Xue <ying....@windriver.com>
>> Signed-off-by: Cong Wang <xiyou.wangc...@gmail.com>
>> ---
>> net/tipc/subscr.c | 2 ++
>> 1 file changed, 2 insertions(+)
>>
>> diff --git a/net/tipc/subscr.c b/net/tipc/subscr.c
>> index 251065dfd8df..2402a019b8ec 100644
>> --- a/net/tipc/subscr.c
>> +++ b/net/tipc/subscr.c
>> @@ -235,6 +235,8 @@ static struct tipc_subscriber *tipc_subscrb_create(int
>> conid)
>>
>> static void tipc_subscrb_delete(struct tipc_subscriber *subscriber)
>> {
>> + if (!subscriber)
>> + return;
>
> The real root cause of this crash is because when it was failed to
> create a connection through s->tipc_conn_new() in
> tipc_topsrv_kern_subscr(), we just should release the connection
> instance with conn_put() rather than calling tipc_close_conn() to close
> the connection. So a proper fix might look like:
Good point! But con->sock should be already NULL since
we use kzalloc() to allocate con, so just replacing tipc_close_conn()
with con_put() should work.
