Re: [syzbot] WARNING in destroy_conntrack

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]


On 8/9/21 6:38 PM, syzbot wrote:

syzbot found the following issue on:

HEAD commit:    f9be84db09d2 net: bonding: bond_alb: Remove the dependency..
git tree:       net-next
console output:
kernel config:
dashboard link:
compiler:       gcc (Debian 10.2.1-6) 10.2.1 20210110, GNU ld (GNU Binutils for Debian) 2.35.1
syz repro:
C reproducer:

The issue was bisected to:

commit 65038428b2c6c5be79d3f78a6b79c0cdc3a58a41
Author: Pablo Neira Ayuso <pablo@xxxxxxxxxxxxx>
Date:   Tue Mar 17 13:13:46 2020 +0000

     netfilter: nf_tables: allow to specify stateful expression in set definition

bisection log:
final oops:
console output:

IMPORTANT: if you fix the issue, please add the following tag to the commit:
Reported-by: syzbot+a1eb62c681423ee5c0d7@xxxxxxxxxxxxxxxxxxxxxxxxx
Fixes: 65038428b2c6 ("netfilter: nf_tables: allow to specify stateful expression in set definition")

------------[ cut here ]------------
WARNING: CPU: 1 PID: 8989 at net/netfilter/nf_conntrack_core.c:610 destroy_conntrack+0x232/0x2c0 net/netfilter/nf_conntrack_core.c:610
Modules linked in:
CPU: 0 PID: 8989 Comm: syz-executor188 Not tainted 5.14.0-rc4-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
RIP: 0010:destroy_conntrack+0x232/0x2c0 net/netfilter/nf_conntrack_core.c:610
Code: da fc ff eb 90 e8 ae 27 19 fa 48 89 ef e8 c6 53 02 00 48 89 ef e8 ee 1c 5f fa 5b 5d 41 5c 41 5d e9 93 27 19 fa e8 8e 27 19 fa <0f> 0b e9 2f fe ff ff e8 82 27 19 fa 4c 8d a5 e8 00 00 00 48 b8 00
RSP: 0018:ffffc90002d7f080 EFLAGS: 00010293
RAX: 0000000000000000 RBX: 00000000ffffffff RCX: 0000000000000000
RDX: ffff88802eeb8000 RSI: ffffffff875c8632 RDI: 0000000000000003
RBP: ffff888147d35400 R08: 0000000000000000 R09: ffff888147d35403
R10: ffffffff875c8460 R11: 0000000000000000 R12: ffff888147d35400
R13: ffffffff8b31b880 R14: 0000000000000000 R15: 0000000000000001
FS:  00007fb18d49a700(0000) GS:ffff8880b9c00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007f37c403d088 CR3: 00000000182c1000 CR4: 00000000001506f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
  nf_conntrack_destroy+0xab/0x230 net/netfilter/core.c:677
  nf_conntrack_put include/linux/netfilter/nf_conntrack_common.h:34 [inline]
  nf_ct_put include/net/netfilter/nf_conntrack.h:176 [inline]
  nft_ct_tmpl_put_pcpu+0x15e/0x1e0 net/netfilter/nft_ct.c:356

Hm... Calltrace looks similar to and my debug logs say that right before this check

	WARN_ON(atomic_read(&nfct->use) != 0);

nfct->use is zero. Reproducer is threaded, so, I think, 2 netns concurrently doing inc/dec somewhere. This should be fixed by my previous netfilter patch.

Also, with my previous netfilter patch applied I cannot reproduce this bug locally, so

#syz fix: netfilter: nft_ct: protect nft_ct_pcpu_template_refcnt with mutex

With regards,
Pavel Skripkin

[Index of Archives]     [Netfitler Users]     [Berkeley Packet Filter]     [LARTC]     [Bugtraq]     [Yosemite Forum]

  Powered by Linux