From: Pablo Neira Ayuso <pablo@netfilter.org>
To: netfilter-devel@vger.kernel.org
Cc: davem@davemloft.net, netdev@vger.kernel.org
Subject: [PATCH 01/17] netfilter: ctnetlink: netns exit must wait for callbacks
Date: Mon, 9 Dec 2019 20:26:22 +0100 [thread overview]
Message-ID: <20191209192638.71184-2-pablo@netfilter.org> (raw)
In-Reply-To: <20191209192638.71184-1-pablo@netfilter.org>
From: Florian Westphal <fw@strlen.de>
Curtis Taylor and Jon Maxwell reported and debugged a crash on 3.10
based kernel.
Crash occurs in ctnetlink_conntrack_events because net->nfnl socket is
NULL. The nfnl socket was set to NULL by netns destruction running on
another cpu.
The exiting network namespace calls the relevant destructors in the
following order:
1. ctnetlink_net_exit_batch
This nulls out the event callback pointer in struct netns.
2. nfnetlink_net_exit_batch
This nulls net->nfnl socket and frees it.
3. nf_conntrack_cleanup_net_list
This removes all remaining conntrack entries.
This is order is correct. The only explanation for the crash so ar is:
cpu1: conntrack is dying, eviction occurs:
-> nf_ct_delete()
-> nf_conntrack_event_report \
-> nf_conntrack_eventmask_report
-> notify->fcn() (== ctnetlink_conntrack_events).
cpu1: a. fetches rcu protected pointer to obtain ctnetlink event callback.
b. gets interrupted.
cpu2: runs netns exit handlers:
a runs ctnetlink destructor, event cb pointer set to NULL.
b runs nfnetlink destructor, nfnl socket is closed and set to NULL.
cpu1: c. resumes and trips over NULL net->nfnl.
Problem appears to be that ctnetlink_net_exit_batch only prevents future
callers of nf_conntrack_eventmask_report() from obtaining the callback.
It doesn't wait of other cpus that might have already obtained the
callbacks address.
I don't see anything in upstream kernels that would prevent similar
crash: We need to wait for all cpus to have exited the event callback.
Fixes: 9592a5c01e79dbc59eb56fa ("netfilter: ctnetlink: netns support")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
net/netfilter/nf_conntrack_netlink.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/net/netfilter/nf_conntrack_netlink.c b/net/netfilter/nf_conntrack_netlink.c
index d8d33ef52ce0..6a1c8f1f6171 100644
--- a/net/netfilter/nf_conntrack_netlink.c
+++ b/net/netfilter/nf_conntrack_netlink.c
@@ -3626,6 +3626,9 @@ static void __net_exit ctnetlink_net_exit_batch(struct list_head *net_exit_list)
list_for_each_entry(net, net_exit_list, exit_list)
ctnetlink_net_exit(net);
+
+ /* wait for other cpus until they are done with ctnl_notifiers */
+ synchronize_rcu();
}
static struct pernet_operations ctnetlink_net_ops = {
--
2.11.0
next prev parent reply other threads:[~2019-12-09 19:26 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-12-09 19:26 [PATCH 00/17] Netfilter fixes for net Pablo Neira Ayuso
2019-12-09 19:26 ` Pablo Neira Ayuso [this message]
2019-12-09 19:26 ` [PATCH 02/17] netfilter: nf_flow_table_offload: Fix block setup as TC_SETUP_FT cmd Pablo Neira Ayuso
2019-12-09 19:26 ` [PATCH 03/17] netfilter: nf_flow_table_offload: Fix block_cb tc_setup_type as TC_SETUP_CLSFLOWER Pablo Neira Ayuso
2019-12-09 19:26 ` [PATCH 04/17] netfilter: nf_flow_table_offload: Don't use offset uninitialized in flow_offload_port_{d,s}nat Pablo Neira Ayuso
2019-12-09 19:26 ` [PATCH 05/17] netfilter: conntrack: tell compiler to not inline nf_ct_resolve_clash Pablo Neira Ayuso
2019-12-09 19:26 ` [PATCH 06/17] netfilter: nf_flow_table_offload: add IPv6 match description Pablo Neira Ayuso
2019-12-09 19:26 ` [PATCH 07/17] netfilter: nf_tables_offload: Check for the NETDEV_UNREGISTER event Pablo Neira Ayuso
2019-12-09 19:26 ` [PATCH 08/17] selftests: netfilter: use randomized netns names Pablo Neira Ayuso
2019-12-09 19:26 ` [PATCH 09/17] netfilter: nf_queue: enqueue skbs with NULL dst Pablo Neira Ayuso
2019-12-09 19:26 ` [PATCH 10/17] netfilter: uapi: Avoid undefined left-shift in xt_sctp.h Pablo Neira Ayuso
2019-12-09 19:26 ` [PATCH 11/17] netfilter: nft_set_rbtree: bogus lookup/get on consecutive elements in named sets Pablo Neira Ayuso
2019-12-09 19:26 ` [PATCH 12/17] netfilter: nf_tables: validate NFT_SET_ELEM_INTERVAL_END Pablo Neira Ayuso
2019-12-09 19:26 ` [PATCH 13/17] netfilter: nf_tables: validate NFT_DATA_VALUE after nft_data_init() Pablo Neira Ayuso
2019-12-09 19:26 ` [PATCH 14/17] netfilter: nf_tables: skip module reference count bump on object updates Pablo Neira Ayuso
2019-12-09 19:26 ` [PATCH 15/17] netfilter: nf_tables_offload: return EOPNOTSUPP if rule specifies no actions Pablo Neira Ayuso
2019-12-09 19:26 ` [PATCH 16/17] netfilter: bridge: make sure to pull arp header in br_nf_forward_arp() Pablo Neira Ayuso
2019-12-09 19:26 ` [PATCH 17/17] netfilter: nf_flow_table_offload: Correct memcpy size for flow_overload_mangle() Pablo Neira Ayuso
2019-12-09 22:03 ` [PATCH 00/17] Netfilter fixes for net David Miller
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20191209192638.71184-2-pablo@netfilter.org \
--to=pablo@netfilter.org \
--cc=davem@davemloft.net \
--cc=netdev@vger.kernel.org \
--cc=netfilter-devel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox