linux/net/netfilter, branch v4.20

netfilter: nf_conncount: use rb_link_node_rcu() instead of rb_link_node()

2018-12-13T00:14:58Z

rbnode in insert_tree() is rcu protected pointer. So, in order to handle this pointer, _rcu function should be used. rb_link_node_rcu() is a rcu version of rb_link_node(). Fixes: 34848d5c896e ("netfilter: nf_conncount: Split insert and traversal") Signed-off-by: Taehee Yoo Signed-off-by: Pablo Neira Ayuso

netfilter: nat: can't use dst_hold on noref dst

2018-12-12T23:58:22Z

The dst entry might already have a zero refcount, waiting on rcu list to be free'd. Using dst_hold() transitions its reference count to 1, and next dst release will try to free it again -- resulting in a double free: WARNING: CPU: 1 PID: 0 at include/net/dst.h:239 nf_xfrm_me_harder+0xe7/0x130 [nf_nat] RIP: 0010:nf_xfrm_me_harder+0xe7/0x130 [nf_nat] Code: 48 8b 5c 24 60 65 48 33 1c 25 28 00 00 00 75 53 48 83 c4 68 5b 5d 41 5c c3 85 c0 74 0d 8d 48 01 f0 0f b1 0a 74 86 85 c0 75 f3 <0f> 0b e9 7b ff ff ff 29 c6 31 d2 b9 20 00 48 00 4c 89 e7 e8 31 27 Call Trace: nf_nat_ipv4_out+0x78/0x90 [nf_nat_ipv4] nf_hook_slow+0x36/0xd0 ip_output+0x9f/0xd0 ip_forward+0x328/0x440 ip_rcv+0x8a/0xb0 Use dst_hold_safe instead and bail out if we cannot take a reference. Fixes: a4c2fd7f7891 ("net: remove DST_NOCACHE flag") Reported-by: Martin Zaharinov Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso

netfilter: ipset: do not call ipset_nest_end after nla_nest_cancel

2018-12-12T23:55:50Z

In the error handling block, nla_nest_cancel(skb, atd) is called to cancel the nest operation. But then, ipset_nest_end(skb, atd) is unexpected called to end the nest operation. This patch calls the ipset_nest_end only on the branch that nla_nest_cancel is not called. Fixes: 45040978c899 ("netfilter: ipset: Fix set:list type crash when flush/dump set in parallel") Signed-off-by: Pan Bian Signed-off-by: Jozsef Kadlecsik Signed-off-by: Pablo Neira Ayuso

netfilter: seqadj: re-load tcp header pointer after possible head reallocation

2018-12-07T09:54:23Z

When adjusting sack block sequence numbers, skb_make_writable() gets called to make sure tcp options are all in the linear area, and buffer is not shared. This can cause tcp header pointer to get reallocated, so we must reaload it to avoid memory corruption. This bug pre-dates git history. Reported-by: Neel Mehta Reported-by: Shane Huntley Reported-by: Heather Adkins Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso

netfilter: nf_tables: fix suspicious RCU usage in nft_chain_stats_replace()

2018-12-04T00:37:13Z

basechain->stats is rcu protected data which is updated from nft_chain_stats_replace(). This function is executed from the commit phase which holds the pernet nf_tables commit mutex - not the global nfnetlink subsystem mutex. Test commands to reproduce the problem are: %iptables-nft -I INPUT %iptables-nft -Z %iptables-nft -Z This patch uses RCU calls to handle basechain->stats updates to fix a splat that looks like: [89279.358755] ============================= [89279.363656] WARNING: suspicious RCU usage [89279.368458] 4.20.0-rc2+ #44 Tainted: G W L [89279.374661] ----------------------------- [89279.379542] net/netfilter/nf_tables_api.c:1404 suspicious rcu_dereference_protected() usage! [...] [89279.406556] 1 lock held by iptables-nft/5225: [89279.411728] #0: 00000000bf45a000 (&net->nft.commit_mutex){+.+.}, at: nf_tables_valid_genid+0x1f/0x70 [nf_tables] [89279.424022] stack backtrace: [89279.429236] CPU: 0 PID: 5225 Comm: iptables-nft Tainted: G W L 4.20.0-rc2+ #44 [89279.430135] Call Trace: [89279.430135] dump_stack+0xc9/0x16b [89279.430135] ? show_regs_print_info+0x5/0x5 [89279.430135] ? lockdep_rcu_suspicious+0x117/0x160 [89279.430135] nft_chain_commit_update+0x4ea/0x640 [nf_tables] [89279.430135] ? sched_clock_local+0xd4/0x140 [89279.430135] ? check_flags.part.35+0x440/0x440 [89279.430135] ? __rhashtable_remove_fast.constprop.67+0xec0/0xec0 [nf_tables] [89279.430135] ? sched_clock_cpu+0x126/0x170 [89279.430135] ? find_held_lock+0x39/0x1c0 [89279.430135] ? hlock_class+0x140/0x140 [89279.430135] ? is_bpf_text_address+0x5/0xf0 [89279.430135] ? check_flags.part.35+0x440/0x440 [89279.430135] ? __lock_is_held+0xb4/0x140 [89279.430135] nf_tables_commit+0x2555/0x39c0 [nf_tables] Fixes: f102d66b335a4 ("netfilter: nf_tables: use dedicated mutex to guard transactions") Signed-off-by: Taehee Yoo Signed-off-by: Pablo Neira Ayuso

netfilter: nf_tables: deactivate expressions in rule replecement routine

2018-11-28T09:56:40Z

There is no expression deactivation call from the rule replacement path, hence, chain counter is not decremented. A few steps to reproduce the problem: %nft add table ip filter %nft add chain ip filter c1 %nft add chain ip filter c1 %nft add rule ip filter c1 jump c2 %nft replace rule ip filter c1 handle 3 accept %nft flush ruleset expression means immediate NFT_JUMP to chain c2. Reference count of chain c2 is increased when the rule is added. When rule is deleted or replaced, the reference counter of c2 should be decreased via nft_rule_expr_deactivate() which calls nft_immediate_deactivate(). Splat looks like: [ 214.396453] WARNING: CPU: 1 PID: 21 at net/netfilter/nf_tables_api.c:1432 nf_tables_chain_destroy.isra.38+0x2f9/0x3a0 [nf_tables] [ 214.398983] Modules linked in: nf_tables nfnetlink [ 214.398983] CPU: 1 PID: 21 Comm: kworker/1:1 Not tainted 4.20.0-rc2+ #44 [ 214.398983] Workqueue: events nf_tables_trans_destroy_work [nf_tables] [ 214.398983] RIP: 0010:nf_tables_chain_destroy.isra.38+0x2f9/0x3a0 [nf_tables] [ 214.398983] Code: 00 00 00 fc ff df 48 89 fa 48 c1 ea 03 80 3c 02 00 0f 85 8e 00 00 00 48 8b 7b 58 e8 e1 2c 4e c6 48 89 df e8 d9 2c 4e c6 eb 9a <0f> 0b eb 96 0f 0b e9 7e fe ff ff e8 a7 7e 4e c6 e9 a4 fe ff ff e8 [ 214.398983] RSP: 0018:ffff8881152874e8 EFLAGS: 00010202 [ 214.398983] RAX: 0000000000000001 RBX: ffff88810ef9fc28 RCX: ffff8881152876f0 [ 214.398983] RDX: dffffc0000000000 RSI: 1ffff11022a50ede RDI: ffff88810ef9fc78 [ 214.398983] RBP: 1ffff11022a50e9d R08: 0000000080000000 R09: 0000000000000000 [ 214.398983] R10: 0000000000000000 R11: 0000000000000000 R12: 1ffff11022a50eba [ 214.398983] R13: ffff888114446e08 R14: ffff8881152876f0 R15: ffffed1022a50ed6 [ 214.398983] FS: 0000000000000000(0000) GS:ffff888116400000(0000) knlGS:0000000000000000 [ 214.398983] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 214.398983] CR2: 00007fab9bb5f868 CR3: 000000012aa16000 CR4: 00000000001006e0 [ 214.398983] Call Trace: [ 214.398983] ? nf_tables_table_destroy.isra.37+0x100/0x100 [nf_tables] [ 214.398983] ? __kasan_slab_free+0x145/0x180 [ 214.398983] ? nf_tables_trans_destroy_work+0x439/0x830 [nf_tables] [ 214.398983] ? kfree+0xdb/0x280 [ 214.398983] nf_tables_trans_destroy_work+0x5f5/0x830 [nf_tables] [ ... ] Fixes: bb7b40aecbf7 ("netfilter: nf_tables: bogus EBUSY in chain deletions") Reported by: Christoph Anton Mitterer Link: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=914505 Link: https://bugzilla.kernel.org/show_bug.cgi?id=201791 Signed-off-by: Taehee Yoo Signed-off-by: Pablo Neira Ayuso

netfilter: nf_conncount: remove wrong condition check routine

2018-11-26T23:48:12Z

All lists that reach the tree_nodes_free() function have both zero counter and true dead flag. The reason for this is that lists to be release are selected by nf_conncount_gc_list() which already decrements the list counter and sets on the dead flag. Therefore, this if statement in tree_nodes_free() is unnecessary and wrong. Fixes: 31568ec09ea0 ("netfilter: nf_conncount: fix list_del corruption in conn_free") Signed-off-by: Taehee Yoo Signed-off-by: Pablo Neira Ayuso

netfilter: add missing error handling code for register functions

2018-11-26T23:35:19Z

register_{netdevice/inetaddr/inet6addr}_notifier may return an error value, this patch adds the code to handle these error paths. Signed-off-by: Taehee Yoo Signed-off-by: Pablo Neira Ayuso

netfilter: nfnetlink_cttimeout: fetch timeouts for udplite and gre, too

2018-11-26T09:25:20Z

syzbot was able to trigger the WARN in cttimeout_default_get() by passing UDPLITE as l4protocol. Alias UDPLITE to UDP, both use same timeout values. Furthermore, also fetch GRE timeouts. GRE is a bit more complicated, as it still can be a module and its netns_proto_gre struct layout isn't visible outside of the gre module. Can't move timeouts around, it appears conntrack sysctl unregister assumes net_generic() returns nf_proto_net, so we get crash. Expose layout of netns_proto_gre instead. A followup nf-next patch could make gre tracker be built-in as well if needed, its not that large. Last, make the WARN() mention the missing protocol value in case anything else is missing. Reported-by: syzbot+2fae8fa157dd92618cae@syzkaller.appspotmail.com Fixes: 8866df9264a3 ("netfilter: nfnetlink_cttimeout: pass default timeout policy to obj_to_nlattr") Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso

ipvs: call ip_vs_dst_notifier earlier than ipv6_dev_notf

2018-11-26T09:23:42Z

ip_vs_dst_event is supposed to clean up all dst used in ipvs' destinations when a net dev is going down. But it works only when the dst's dev is the same as the dev from the event. Now with the same priority but late registration, ip_vs_dst_notifier is always called later than ipv6_dev_notf where the dst's dev is set to lo for NETDEV_DOWN event. As the dst's dev lo is not the same as the dev from the event in ip_vs_dst_event, ip_vs_dst_notifier doesn't actually work. Also as these dst have to wait for dest_trash_timer to clean them up. It would cause some non-permanent kernel warnings: unregister_netdevice: waiting for br0 to become free. Usage count = 3 To fix it, call ip_vs_dst_notifier earlier than ipv6_dev_notf by increasing its priority to ADDRCONF_NOTIFY_PRIORITY + 5. Note that for ipv4 route fib_netdev_notifier doesn't set dst's dev to lo in NETDEV_DOWN event, so this fix is only needed when IP_VS_IPV6 is defined. Fixes: 7a4f0761fce3 ("IPVS: init and cleanup restructuring") Reported-by: Li Shuang Signed-off-by: Xin Long Acked-by: Julian Anastasov Acked-by: Simon Horman Signed-off-by: Pablo Neira Ayuso