summaryrefslogtreecommitdiffstats
path: root/net/ipv4
Commit message (Collapse)AuthorAgeFilesLines
...
* | | | net: ipv4: route: perform strict checks also for doit handlersJakub Kicinski2019-01-191-2/+70
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Make RTM_GETROUTE's doit handler use strict checks when NETLINK_F_STRICT_CHK is set. v2: - new patch (DaveA). Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | | net: ipv4: netconf: perform strict checks also for doit handlersJakub Kicinski2019-01-191-4/+39
| |/ / |/| | | | | | | | | | | | | | | | | | | | Make RTM_GETNETCONF's doit handler use strict checks when NETLINK_F_STRICT_CHK is set. Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | tcp: move rx_opt & syn_data_acked init to tcp_disconnect()Eric Dumazet2019-01-172-6/+4
| | | | | | | | | | | | | | | | | | | | | | | | If we make sure all listeners have these fields cleared, then a clone will also inherit zero values. Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | tcp: move tp->rack init to tcp_disconnect()Eric Dumazet2019-01-172-6/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If we make sure all listeners have proper tp->rack value, then a clone will also inherit proper initial value. Note that fresh sockets init tp->rack from tcp_init_sock() Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | tcp: move app_limited init to tcp_disconnect()Eric Dumazet2019-01-172-3/+3
| | | | | | | | | | | | | | | | | | | | | | | | If we make sure all listeners have app_limited set to ~0U, then a clone will also inherit proper initial value. Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | tcp: move retrans_out, sacked_out, tlp_high_seq, last_oow_ack_time init to ↵Eric Dumazet2019-01-172-4/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | tcp_disconnect() If we make sure all listeners have these fields cleared, then a clone will also inherit zero values. Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | tcp: do not clear urg_data in tcp_create_openreq_childEric Dumazet2019-01-171-2/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | All listeners have this field cleared already, since tcp_disconnect() clears it and newly created sockets have also a zero value here. So a clone will inherit a zero value here. Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | tcp: move snd_cwnd & snd_cwnd_cnt init to tcp_disconnect()Eric Dumazet2019-01-172-9/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Passive connections can inherit proper value by cloning, if we make sure all listeners have the proper values there. tcp_disconnect() was setting snd_cwnd to 2, which seems quite obsolete since IW10 adoption. Also remove an obsolete comment. Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | tcp: move mdev_us init to tcp_disconnect()Eric Dumazet2019-01-172-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If we make sure a listener always has its mdev_us field set to TCP_TIMEOUT_INIT, we do not need to rewrite this field after a new clone is created. tcp_disconnect() is very seldom used in real applications. Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | tcp: do not clear srtt_us in tcp_create_openreq_childEric Dumazet2019-01-171-1/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | All listeners have this field cleared already, since tcp_disconnect() clears it and newly created sockets have also a zero value here. So a clone will inherit a zero value here. Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | tcp: do not clear packets_out in tcp_create_openreq_child()Eric Dumazet2019-01-171-1/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | New sockets have this field cleared, and tcp_disconnect() calls tcp_write_queue_purge() which among other things also clear tp->packets_out So a listener is guaranteed to have this field cleared. Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | tcp: move icsk_rto init to tcp_disconnect()Eric Dumazet2019-01-172-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If we make sure a listener always has its icsk_rto field set to TCP_TIMEOUT_INIT, we do not need to rewrite this field after a new clone is created. tcp_disconnect() is very seldom used in real applications. Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | tcp: do not set snd_ssthresh in tcp_create_openreq_child()Eric Dumazet2019-01-171-1/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | New sockets get the field set to TCP_INFINITE_SSTHRESH in tcp_init_sock() In case a socket had this field changed and transitions to TCP_LISTEN state, tcp_disconnect() also makes sure snd_ssthresh is set to TCP_INFINITE_SSTHRESH. So a listener has this field set to TCP_INFINITE_SSTHRESH already. Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | tcp: less aggressive window probing on local congestionYuchung Cheng2019-01-171-15/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously when the sender fails to send (original) data packet or window probes due to congestion in the local host (e.g. throttling in qdisc), it'll retry within an RTO or two up to 500ms. In low-RTT networks such as data-centers, RTO is often far below the default minimum 200ms. Then local host congestion could trigger a retry storm pouring gas to the fire. Worse yet, the probe counter (icsk_probes_out) is not properly updated so the aggressive retry may exceed the system limit (15 rounds) until the packet finally slips through. On such rare events, it's wise to retry more conservatively (500ms) and update the stats properly to reflect these incidents and follow the system limit. Note that this is consistent with the behaviors when a keep-alive probe or RTO retry is dropped due to local congestion. Signed-off-by: Yuchung Cheng <ycheng@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Reviewed-by: Neal Cardwell <ncardwell@google.com> Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | tcp: retry more conservatively on local congestionYuchung Cheng2019-01-171-5/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously when the sender fails to retransmit a data packet on timeout due to congestion in the local host (e.g. throttling in qdisc), it'll retry within an RTO up to 500ms. In low-RTT networks such as data-centers, RTO is often far below the default minimum 200ms (and the cap 500ms). Then local host congestion could trigger a retry storm pouring gas to the fire. Worse yet, the retry counter (icsk_retransmits) is not properly updated so the aggressive retry may exceed the system limit (15 rounds) until the packet finally slips through. On such rare events, it's wise to retry more conservatively (500ms) and update the stats properly to reflect these incidents and follow the system limit. Note that this is consistent with the behavior when a keep-alive probe is dropped due to local congestion. Signed-off-by: Yuchung Cheng <ycheng@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Reviewed-by: Neal Cardwell <ncardwell@google.com> Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | tcp: simplify window probe aborting on USER_TIMEOUTYuchung Cheng2019-01-171-7/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously we use the next unsent skb's timestamp to determine when to abort a socket stalling on window probes. This no longer works as skb timestamp reflects the last instead of the first transmission. Instead we can estimate how long the socket has been stalling with the probe count and the exponential backoff behavior. Signed-off-by: Yuchung Cheng <ycheng@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Reviewed-by: Neal Cardwell <ncardwell@google.com> Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | tcp: create a helper to model exponential backoffYuchung Cheng2019-01-171-13/+18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Create a helper to model TCP exponential backoff for the next patch. This is pure refactor w no behavior change. Signed-off-by: Yuchung Cheng <ycheng@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Reviewed-by: Neal Cardwell <ncardwell@google.com> Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | tcp: properly track retry time on passive Fast OpenYuchung Cheng2019-01-171-0/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch addresses a corner issue on timeout behavior of a passive Fast Open socket. A passive Fast Open server may write and close the socket when it is re-trying SYN-ACK to complete the handshake. After the handshake is completely, the server does not properly stamp the recovery start time (tp->retrans_stamp is 0), and the socket may abort immediately on the very first FIN timeout, instead of retying until it passes the system or user specified limit. Signed-off-by: Yuchung Cheng <ycheng@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Reviewed-by: Neal Cardwell <ncardwell@google.com> Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | tcp: always set retrans_stamp on recoveryYuchung Cheng2019-01-172-25/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously TCP socket's retrans_stamp is not set if the retransmission has failed to send. As a result if a socket is experiencing local issues to retransmit packets, determining when to abort a socket is complicated w/o knowning the starting time of the recovery since retrans_stamp may remain zero. This complication causes sub-optimal behavior that TCP may use the latest, instead of the first, retransmission time to compute the elapsed time of a stalling connection due to local issues. Then TCP may disrecard TCP retries settings and keep retrying until it finally succeed: not a good idea when the local host is already strained. The simple fix is to always timestamp the start of a recovery. It's worth noting that retrans_stamp is also used to compare echo timestamp values to detect spurious recovery. This patch does not break that because retrans_stamp is still later than when the original packet was sent. Signed-off-by: Yuchung Cheng <ycheng@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Reviewed-by: Neal Cardwell <ncardwell@google.com> Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | tcp: always timestamp on every skb transmissionYuchung Cheng2019-01-171-8/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously TCP skbs are not always timestamped if the transmission failed due to memory or other local issues. This makes deciding when to abort a socket tricky and complicated because the first unacknowledged skb's timestamp may be 0 on TCP timeout. The straight-forward fix is to always timestamp skb on every transmission attempt. Also every skb retransmission needs to be flagged properly to avoid RTT under-estimation. This can happen upon receiving an ACK for the original packet and the a previous (spurious) retransmission has failed. It's worth noting that this reverts to the old time-stamping style before commit 8c72c65b426b ("tcp: update skb->skb_mstamp more carefully") which addresses a problem in computing the elapsed time of a stalled window-probing socket. The problem will be addressed differently in the next patches with a simpler approach. Signed-off-by: Yuchung Cheng <ycheng@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Reviewed-by: Neal Cardwell <ncardwell@google.com> Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | tcp: exit if nothing to retransmit on RTO timeoutYuchung Cheng2019-01-171-4/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously TCP only warns if its RTO timer fires and the retransmission queue is empty, but it'll cause null pointer reference later on. It's better to avoid such catastrophic failure and simply exit with a warning. Signed-off-by: Yuchung Cheng <ycheng@google.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Reviewed-by: Neal Cardwell <ncardwell@google.com> Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | net/ipv4/udp_tunnel: prefer SO_BINDTOIFINDEX over SO_BINDTODEVICEDavid Herrmann2019-01-171-12/+3
|/ / | | | | | | | | | | | | | | | | | | The udp-tunnel setup allows binding sockets to a network device. Prefer the new SO_BINDTOIFINDEX to avoid temporarily resolving the device-name just to look it up in the ioctl again. Reviewed-by: Tom Gundersen <teg@jklm.no> Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | net: bpfilter: disallow to remove bpfilter module while being usedTaehee Yoo2019-01-111-6/+16
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The bpfilter.ko module can be removed while functions of the bpfilter.ko are executing. so panic can occurred. in order to protect that, locks can be used. a bpfilter_lock protects routines in the __bpfilter_process_sockopt() but it's not enough because __exit routine can be executed concurrently. Now, the bpfilter_umh can not run in parallel. So, the module do not removed while it's being used and it do not double-create UMH process. The members of the umh_info and the bpfilter_umh_ops are protected by the bpfilter_umh_ops.lock. test commands: while : do iptables -I FORWARD -m string --string ap --algo kmp & modprobe -rv bpfilter & done splat looks like: [ 298.623435] BUG: unable to handle kernel paging request at fffffbfff807440b [ 298.628512] #PF error: [normal kernel read fault] [ 298.633018] PGD 124327067 P4D 124327067 PUD 11c1a3067 PMD 119eb2067 PTE 0 [ 298.638859] Oops: 0000 [#1] SMP DEBUG_PAGEALLOC KASAN PTI [ 298.638859] CPU: 0 PID: 2997 Comm: iptables Not tainted 4.20.0+ #154 [ 298.638859] RIP: 0010:__mutex_lock+0x6b9/0x16a0 [ 298.638859] Code: c0 00 00 e8 89 82 ff ff 80 bd 8f fc ff ff 00 0f 85 d9 05 00 00 48 8b 85 80 fc ff ff 48 bf 00 00 00 00 00 fc ff df 48 c1 e8 03 <80> 3c 38 00 0f 85 1d 0e 00 00 48 8b 85 c8 fc ff ff 49 39 47 58 c6 [ 298.638859] RSP: 0018:ffff88810e7777a0 EFLAGS: 00010202 [ 298.638859] RAX: 1ffffffff807440b RBX: ffff888111bd4d80 RCX: 0000000000000000 [ 298.638859] RDX: 1ffff110235ff806 RSI: ffff888111bd5538 RDI: dffffc0000000000 [ 298.638859] RBP: ffff88810e777b30 R08: 0000000080000002 R09: 0000000000000000 [ 298.638859] R10: 0000000000000000 R11: 0000000000000000 R12: fffffbfff168a42c [ 298.638859] R13: ffff888111bd4d80 R14: ffff8881040e9a05 R15: ffffffffc03a2000 [ 298.638859] FS: 00007f39e3758700(0000) GS:ffff88811ae00000(0000) knlGS:0000000000000000 [ 298.638859] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 298.638859] CR2: fffffbfff807440b CR3: 000000011243e000 CR4: 00000000001006f0 [ 298.638859] Call Trace: [ 298.638859] ? mutex_lock_io_nested+0x1560/0x1560 [ 298.638859] ? kasan_kmalloc+0xa0/0xd0 [ 298.638859] ? kmem_cache_alloc+0x1c2/0x260 [ 298.638859] ? __alloc_file+0x92/0x3c0 [ 298.638859] ? alloc_empty_file+0x43/0x120 [ 298.638859] ? alloc_file_pseudo+0x220/0x330 [ 298.638859] ? sock_alloc_file+0x39/0x160 [ 298.638859] ? __sys_socket+0x113/0x1d0 [ 298.638859] ? __x64_sys_socket+0x6f/0xb0 [ 298.638859] ? do_syscall_64+0x138/0x560 [ 298.638859] ? entry_SYSCALL_64_after_hwframe+0x49/0xbe [ 298.638859] ? __alloc_file+0x92/0x3c0 [ 298.638859] ? init_object+0x6b/0x80 [ 298.638859] ? cyc2ns_read_end+0x10/0x10 [ 298.638859] ? cyc2ns_read_end+0x10/0x10 [ 298.638859] ? hlock_class+0x140/0x140 [ 298.638859] ? sched_clock_local+0xd4/0x140 [ 298.638859] ? sched_clock_local+0xd4/0x140 [ 298.638859] ? check_flags.part.37+0x440/0x440 [ 298.638859] ? __lock_acquire+0x4f90/0x4f90 [ 298.638859] ? set_rq_offline.part.89+0x140/0x140 [ ... ] Fixes: d2ba09c17a06 ("net: add skeleton of bpfilter kernel module") Signed-off-by: Taehee Yoo <ap420073@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | net: bpfilter: restart bpfilter_umh when error occurredTaehee Yoo2019-01-111-1/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The bpfilter_umh will be stopped via __stop_umh() when the bpfilter error occurred. The bpfilter_umh() couldn't start again because there is no restart routine. The section of the bpfilter_umh_{start/end} is no longer .init.rodata because these area should be reused in the restart routine. hence the section name is changed to .bpfilter_umh. The bpfilter_ops->start() is restart callback. it will be called when bpfilter_umh is stopped. The stop bit means bpfilter_umh is stopped. this bit is set by both start and stop routine. Before this patch, Test commands: $ iptables -vnL $ kill -9 <pid of bpfilter_umh> $ iptables -vnL [ 480.045136] bpfilter: write fail -32 $ iptables -vnL All iptables commands will fail. After this patch, Test commands: $ iptables -vnL $ kill -9 <pid of bpfilter_umh> $ iptables -vnL $ iptables -vnL Now, all iptables commands will work. Fixes: d2ba09c17a06 ("net: add skeleton of bpfilter kernel module") Signed-off-by: Taehee Yoo <ap420073@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | net: bpfilter: use cleanup callback to release umh_infoTaehee Yoo2019-01-111-7/+26
| | | | | | | | | | | | | | | | | | | | Now, UMH process is killed, do_exit() calls the umh_info->cleanup callback to release members of the umh_info. This patch makes bpfilter_umh's cleanup routine to use the umh_info->cleanup callback. Signed-off-by: Taehee Yoo <ap420073@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | tcp: change txhash on SYN-data timeoutYuchung Cheng2019-01-101-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously upon SYN timeouts the sender recomputes the txhash to try a different path. However this does not apply on the initial timeout of SYN-data (active Fast Open). Therefore an active IPv6 Fast Open connection may incur one second RTO penalty to take on a new path after the second SYN retransmission uses a new flow label. This patch removes this undesirable behavior so Fast Open changes the flow label just like the regular connections. This also helps avoid falsely disabling Fast Open on the sender which triggers after two consecutive SYN timeouts on Fast Open. Signed-off-by: Yuchung Cheng <ycheng@google.com> Reviewed-by: Neal Cardwell <ncardwell@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | ip: on queued skb use skb_header_pointer instead of pskb_may_pullWillem de Bruijn2019-01-101-7/+5
|/ | | | | | | | | | | | | | | | | Commit 2efd4fca703a ("ip: in cmsg IP(V6)_ORIGDSTADDR call pskb_may_pull") avoided a read beyond the end of the skb linear segment by calling pskb_may_pull. That function can trigger a BUG_ON in pskb_expand_head if the skb is shared, which it is when when peeking. It can also return ENOMEM. Avoid both by switching to safer skb_header_pointer. Fixes: 2efd4fca703a ("ip: in cmsg IP(V6)_ORIGDSTADDR call pskb_may_pull") Reported-by: syzbot <syzkaller@googlegroups.com> Suggested-by: Eric Dumazet <edumazet@google.com> Signed-off-by: Willem de Bruijn <willemb@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* fou: Prevent unbounded recursion in GUE error handler also with UDP-LiteStefano Brivio2019-01-041-1/+2
| | | | | | | | | | | | | | In commit 11789039da53 ("fou: Prevent unbounded recursion in GUE error handler"), I didn't take care of the case where UDP-Lite is encapsulated into UDP or UDP-Lite with GUE. From a syzbot report about a possibly similar issue with GUE on IPv6, I just realised the same thing might happen with a UDP-Lite inner payload. Also skip exception handling for inner UDP-Lite protocol. Fixes: 11789039da53 ("fou: Prevent unbounded recursion in GUE error handler") Signed-off-by: Stefano Brivio <sbrivio@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* netlink: fixup regression in RTM_GETADDRArthur Gautier2019-01-041-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | This commit fixes a regression in AF_INET/RTM_GETADDR and AF_INET6/RTM_GETADDR. Before this commit, the kernel would stop dumping addresses once the first skb was full and end the stream with NLMSG_DONE(-EMSGSIZE). The error shouldn't be sent back to netlink_dump so the callback is kept alive. The userspace is expected to call back with a new empty skb. Changes from V1: - The error is not handled in netlink_dump anymore but rather in inet_dump_ifaddr and inet6_dump_addr directly as suggested by David Ahern. Fixes: d7e38611b81e ("net/ipv4: Put target net when address dump fails due to bad attributes") Fixes: 242afaa6968c ("net/ipv6: Put target net when address dump fails due to bad attributes") Cc: David Ahern <dsahern@gmail.com> Cc: "David S . Miller" <davem@davemloft.net> Cc: netdev@vger.kernel.org Signed-off-by: Arthur Gautier <baloo@gandi.net> Reviewed-by: David Ahern <dsahern@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/netLinus Torvalds2019-01-034-14/+24
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Pull networking fixes from David Miller: "Several fixes here. Basically split down the line between newly introduced regressions and long existing problems: 1) Double free in tipc_enable_bearer(), from Cong Wang. 2) Many fixes to nf_conncount, from Florian Westphal. 3) op->get_regs_len() can throw an error, check it, from Yunsheng Lin. 4) Need to use GFP_ATOMIC in *_add_hash_mac_address() of fsl/fman driver, from Scott Wood. 5) Inifnite loop in fib_empty_table(), from Yue Haibing. 6) Use after free in ax25_fillin_cb(), from Cong Wang. 7) Fix socket locking in nr_find_socket(), also from Cong Wang. 8) Fix WoL wakeup enable in r8169, from Heiner Kallweit. 9) On 32-bit sock->sk_stamp is not thread-safe, from Deepa Dinamani. 10) Fix ptr_ring wrap during queue swap, from Cong Wang. 11) Missing shutdown callback in hinic driver, from Xue Chaojing. 12) Need to return NULL on error from ip6_neigh_lookup(), from Stefano Brivio. 13) BPF out of bounds speculation fixes from Daniel Borkmann" * git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (57 commits) ipv6: Consider sk_bound_dev_if when binding a socket to an address ipv6: Fix dump of specific table with strict checking bpf: add various test cases to selftests bpf: prevent out of bounds speculation on pointer arithmetic bpf: fix check_map_access smin_value test when pointer contains offset bpf: restrict unknown scalars of mixed signed bounds for unprivileged bpf: restrict stack pointer arithmetic for unprivileged bpf: restrict map value pointer arithmetic for unprivileged bpf: enable access to ax register also from verifier rewrite bpf: move tmp variable into ax register in interpreter bpf: move {prev_,}insn_idx into verifier env isdn: fix kernel-infoleak in capi_unlocked_ioctl ipv6: route: Fix return value of ip6_neigh_lookup() on neigh_create() error net/hamradio/6pack: use mod_timer() to rearm timers net-next/hinic:add shutdown callback net: hns3: call hns3_nic_net_open() while doing HNAE3_UP_CLIENT ip: validate header length on virtual device xmit tap: call skb_probe_transport_header after setting skb->dev ptr_ring: wrap back ->producer in __ptr_ring_swap_queue() net: rds: remove unnecessary NULL check ...
| * ip: validate header length on virtual device xmitWillem de Bruijn2019-01-013-12/+18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | KMSAN detected read beyond end of buffer in vti and sit devices when passing truncated packets with PF_PACKET. The issue affects additional ip tunnel devices. Extend commit 76c0ddd8c3a6 ("ip6_tunnel: be careful when accessing the inner header") and commit ccfec9e5cb2d ("ip_tunnel: be careful when accessing the inner header"). Move the check to a separate helper and call at the start of each ndo_start_xmit function in net/ipv4 and net/ipv6. Minor changes: - convert dev_kfree_skb to kfree_skb on error path, as dev_kfree_skb calls consume_skb which is not for error paths. - use pskb_network_may_pull even though that is pedantic here, as the same as pskb_may_pull for devices without llheaders. - do not cache ipv6 hdrs if used only once (unsafe across pskb_may_pull, was more relevant to earlier patch) Reported-by: syzbot <syzkaller@googlegroups.com> Signed-off-by: Willem de Bruijn <willemb@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
| * ipv4: fib_rules: Fix possible infinite loop in fib_empty_tableYueHaibing2018-12-301-2/+6
| | | | | | | | | | | | | | | | | | | | | | | | gcc warn this: net/ipv4/fib_rules.c:203 fib_empty_table() warn: always true condition '(id <= 4294967295) => (0-u32max <= u32max)' 'id' is u32, which always not greater than RT_TABLE_MAX (0xFFFFFFFF), So add a check to break while wrap around. Signed-off-by: YueHaibing <yuehaibing@huawei.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | mm: convert totalram_pages and totalhigh_pages variables to atomicArun KS2018-12-281-1/+1
|/ | | | | | | | | | | | | | | | | | | | | | | totalram_pages and totalhigh_pages are made static inline function. Main motivation was that managed_page_count_lock handling was complicating things. It was discussed in length here, https://lore.kernel.org/patchwork/patch/995739/#1181785 So it seemes better to remove the lock and convert variables to atomic, with preventing poteintial store-to-read tearing as a bonus. [akpm@linux-foundation.org: coding style fixes] Link: http://lkml.kernel.org/r/1542090790-21750-4-git-send-email-arunks@codeaurora.org Signed-off-by: Arun KS <arunks@codeaurora.org> Suggested-by: Michal Hocko <mhocko@suse.com> Suggested-by: Vlastimil Babka <vbabka@suse.cz> Reviewed-by: Konstantin Khlebnikov <khlebnikov@yandex-team.ru> Reviewed-by: Pavel Tatashin <pasha.tatashin@soleen.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Vlastimil Babka <vbabka@suse.cz> Cc: David Hildenbrand <david@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/netDavid S. Miller2018-12-241-0/+1
|\ | | | | | | | | | | | | Pull in bug fixes before respinning my net-next pull request. Signed-off-by: David S. Miller <davem@davemloft.net>
| * iptunnel: Set tun_flags in the iptunnel_metadata_reply from srcwenxu2018-12-241-0/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | ip l add tun type gretap external ip r a 10.0.0.2 encap ip id 1000 dst 172.168.0.2 key dev tun ip a a 10.0.0.1/24 dev tun The peer arp request to 10.0.0.1 with tunnel_id, but the arp reply only set the tun_id but not the tun_flags with TUNNEL_KEY. The arp reply packet don't contain tun_id field. Signed-off-by: wenxu <wenxu@ucloud.cn> Signed-off-by: David S. Miller <davem@davemloft.net>
* | net: dccp: fix kernel crash on module loadPeter Oskolkov2018-12-241-8/+26
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Patch eedbbb0d98b2 "net: dccp: initialize (addr,port) ..." added calling to inet_hashinfo2_init() from dccp_init(). However, inet_hashinfo2_init() is marked as __init(), and thus the kernel panics when dccp is loaded as module. Removing __init() tag from inet_hashinfo2_init() is not feasible because it calls into __init functions in mm. This patch adds inet_hashinfo2_init_mod() function that can be called after the init phase is done; changes dccp_init() to call the new function; un-marks inet_hashinfo2_init() as exported. Fixes: eedbbb0d98b2 ("net: dccp: initialize (addr,port) ...") Reported-by: kernel test robot <lkp@intel.com> Signed-off-by: Peter Oskolkov <posk@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/netDavid S. Miller2018-12-211-1/+3
|\|
| * tcp: fix a race in inet_diag_dump_icsk()Eric Dumazet2018-12-201-1/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Alexei reported use after frees in inet_diag_dump_icsk() [1] Because we use refcount_set() when various sockets are setup and inserted into ehash, we also need to make sure inet_diag_dump_icsk() wont race with the refcount_set() operations. Jonathan Lemon sent a patch changing net_twsk_hashdance() but other spots would need risky changes. Instead, fix inet_diag_dump_icsk() as this bug came with linux-4.10 only. [1] Quoting Alexei : First something iterating over sockets finds already freed tw socket: refcount_t: increment on 0; use-after-free. WARNING: CPU: 2 PID: 2738 at lib/refcount.c:153 refcount_inc+0x26/0x30 RIP: 0010:refcount_inc+0x26/0x30 RSP: 0018:ffffc90004c8fbc0 EFLAGS: 00010282 RAX: 000000000000002b RBX: 0000000000000000 RCX: 0000000000000000 RDX: ffff88085ee9d680 RSI: ffff88085ee954c8 RDI: ffff88085ee954c8 RBP: ffff88010ecbd2c0 R08: 0000000000000000 R09: 000000000000174c R10: ffffffff81e7c5a0 R11: 0000000000000000 R12: 0000000000000000 R13: ffff8806ba9bf210 R14: ffffffff82304600 R15: ffff88010ecbd328 FS: 00007f81f5a7d700(0000) GS:ffff88085ee80000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007f81e2a95000 CR3: 000000069b2eb006 CR4: 00000000003606e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: inet_diag_dump_icsk+0x2b3/0x4e0 [inet_diag] // sock_hold(sk); in net/ipv4/inet_diag.c:1002 ? kmalloc_large_node+0x37/0x70 ? __kmalloc_node_track_caller+0x1cb/0x260 ? __alloc_skb+0x72/0x1b0 ? __kmalloc_reserve.isra.40+0x2e/0x80 __inet_diag_dump+0x3b/0x80 [inet_diag] netlink_dump+0x116/0x2a0 netlink_recvmsg+0x205/0x3c0 sock_read_iter+0x89/0xd0 __vfs_read+0xf7/0x140 vfs_read+0x8a/0x140 SyS_read+0x3f/0xa0 do_syscall_64+0x5a/0x100 then a minute later twsk timer fires and hits two bad refcnts for this freed socket: refcount_t: decrement hit 0; leaking memory. WARNING: CPU: 31 PID: 0 at lib/refcount.c:228 refcount_dec+0x2e/0x40 Modules linked in: RIP: 0010:refcount_dec+0x2e/0x40 RSP: 0018:ffff88085f5c3ea8 EFLAGS: 00010296 RAX: 000000000000002c RBX: ffff88010ecbd2c0 RCX: 000000000000083f RDX: 0000000000000000 RSI: 00000000000000f6 RDI: 000000000000003f RBP: ffffc90003c77280 R08: 0000000000000000 R09: 00000000000017d3 R10: ffffffff81e7c5a0 R11: 0000000000000000 R12: ffffffff82ad2d80 R13: ffffffff8182de00 R14: ffff88085f5c3ef8 R15: 0000000000000000 FS: 0000000000000000(0000) GS:ffff88085f5c0000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007fbe42685250 CR3: 0000000002209001 CR4: 00000000003606e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: <IRQ> inet_twsk_kill+0x9d/0xc0 // inet_twsk_bind_unhash(tw, hashinfo); call_timer_fn+0x29/0x110 run_timer_softirq+0x36b/0x3a0 refcount_t: underflow; use-after-free. WARNING: CPU: 31 PID: 0 at lib/refcount.c:187 refcount_sub_and_test+0x46/0x50 RIP: 0010:refcount_sub_and_test+0x46/0x50 RSP: 0018:ffff88085f5c3eb8 EFLAGS: 00010296 RAX: 0000000000000026 RBX: ffff88010ecbd2c0 RCX: 000000000000083f RDX: 0000000000000000 RSI: 00000000000000f6 RDI: 000000000000003f RBP: ffff88010ecbd358 R08: 0000000000000000 R09: 000000000000185b R10: ffffffff81e7c5a0 R11: 0000000000000000 R12: ffff88010ecbd358 R13: ffffffff8182de00 R14: ffff88085f5c3ef8 R15: 0000000000000000 FS: 0000000000000000(0000) GS:ffff88085f5c0000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007fbe42685250 CR3: 0000000002209001 CR4: 00000000003606e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: <IRQ> inet_twsk_put+0x12/0x20 // inet_twsk_put(tw); call_timer_fn+0x29/0x110 run_timer_softirq+0x36b/0x3a0 Fixes: 67db3e4bfbc9 ("tcp: no longer hold ehash lock while calling tcp_get_info()") Signed-off-by: Eric Dumazet <edumazet@google.com> Reported-by: Alexei Starovoitov <ast@kernel.org> Cc: Jonathan Lemon <jonathan.lemon@gmail.com> Acked-by: Jonathan Lemon <jonathan.lemon@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | Merge git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf-nextDavid S. Miller2018-12-207-360/+112
|\ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Pablo Neira Ayuso says: ==================== Netfilter updates for net-next The following patchset contains Netfilter updates for net-next: 1) Support for destination MAC in ipset, from Stefano Brivio. 2) Disallow all-zeroes MAC address in ipset, also from Stefano. 3) Add IPSET_CMD_GET_BYNAME and IPSET_CMD_GET_BYINDEX commands, introduce protocol version number 7, from Jozsef Kadlecsik. A follow up patch to fix ip_set_byindex() is also included in this batch. 4) Honor CTA_MARK_MASK from ctnetlink, from Andreas Jaggi. 5) Statify nf_flow_table_iterate(), from Taehee Yoo. 6) Use nf_flow_table_iterate() to simplify garbage collection in nf_flow_table logic, also from Taehee Yoo. 7) Don't use _bh variants of call_rcu(), rcu_barrier() and synchronize_rcu_bh() in Netfilter, from Paul E. McKenney. 8) Remove NFC_* cache definition from the old caching infrastructure. 9) Remove layer 4 port rover in NAT helpers, use random port instead, from Florian Westphal. 10) Use strscpy() in ipset, from Qian Cai. 11) Remove NF_NAT_RANGE_PROTO_RANDOM_FULLY branch now that random port is allocated by default, from Xiaozhou Liu. 12) Ignore NF_NAT_RANGE_PROTO_RANDOM too, from Florian Westphal. 13) Limit port allocation selection routine in NAT to avoid softlockup splats when most ports are in use, from Florian. 14) Remove unused parameters in nf_ct_l4proto_unregister_sysctl() from Yafang Shao. 15) Direct call to nf_nat_l4proto_unique_tuple() instead of indirection, from Florian Westphal. 16) Several patches to remove all layer 4 NAT indirections, remove nf_nat_l4proto struct, from Florian Westphal. 17) Fix RTP/RTCP source port translation when SNAT is in place, from Alin Nastac. 18) Selective rule dump per chain, from Phil Sutter. 19) Revisit CLUSTERIP target, this includes a deadlock fix from netns path, sleep in atomic, remove bogus WARN_ON_ONCE() and disallow mismatching IP address and MAC address. Patchset from Taehee Yoo. 20) Update UDP timeout to stream after 2 seconds, from Florian. 21) Shrink UDP established timeout to 120 seconds like TCP timewait. 22) Sysctl knobs to set GRE timeouts, from Yafang Shao. 23) Move seq_print_acct() to conntrack core file, from Florian. 24) Add enum for conntrack sysctl knobs, also from Florian. 25) Place nf_conntrack_acct, nf_conntrack_helper, nf_conntrack_events and nf_conntrack_timestamp knobs in the core, from Florian Westphal. As a side effect, shrink netns_ct structure by removing obsolete sysctl anchors, also from Florian. ==================== Signed-off-by: David S. Miller <davem@davemloft.net>
| * | netfilter: ipt_CLUSTERIP: check MAC address when duplicate config is setTaehee Yoo2018-12-181-1/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If same destination IP address config is already existing, that config is just used. MAC address also should be same. However, there is no MAC address checking routine. So that MAC address checking routine is added. test commands: %iptables -A INPUT -p tcp -i lo -d 192.168.0.5 --dport 80 \ -j CLUSTERIP --new --hashmode sourceip \ --clustermac 01:00:5e:00:00:20 --total-nodes 2 --local-node 1 %iptables -A INPUT -p tcp -i lo -d 192.168.0.5 --dport 80 \ -j CLUSTERIP --new --hashmode sourceip \ --clustermac 01:00:5e:00:00:21 --total-nodes 2 --local-node 1 After this patch, above commands are disallowed. Signed-off-by: Taehee Yoo <ap420073@gmail.com> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
| * | netfilter: ipt_CLUSTERIP: fix sleep-in-atomic bug in ↵Taehee Yoo2018-12-181-5/+14
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | clusterip_config_entry_put() A proc_remove() can sleep. so that it can't be inside of spin_lock. Hence proc_remove() is moved to outside of spin_lock. and it also adds mutex to sync create and remove of proc entry(config->pde). test commands: SHELL#1 %while :; do iptables -A INPUT -p udp -i enp2s0 -d 192.168.1.100 \ --dport 9000 -j CLUSTERIP --new --hashmode sourceip \ --clustermac 01:00:5e:00:00:21 --total-nodes 3 --local-node 3; \ iptables -F; done SHELL#2 %while :; do echo +1 > /proc/net/ipt_CLUSTERIP/192.168.1.100; \ echo -1 > /proc/net/ipt_CLUSTERIP/192.168.1.100; done [ 2949.569864] BUG: sleeping function called from invalid context at kernel/sched/completion.c:99 [ 2949.579944] in_atomic(): 1, irqs_disabled(): 0, pid: 5472, name: iptables [ 2949.587920] 1 lock held by iptables/5472: [ 2949.592711] #0: 000000008f0ebcf2 (&(&cn->lock)->rlock){+...}, at: refcount_dec_and_lock+0x24/0x50 [ 2949.603307] CPU: 1 PID: 5472 Comm: iptables Tainted: G W 4.19.0-rc5+ #16 [ 2949.604212] Hardware name: To be filled by O.E.M. To be filled by O.E.M./Aptio CRB, BIOS 5.6.5 07/08/2015 [ 2949.604212] Call Trace: [ 2949.604212] dump_stack+0xc9/0x16b [ 2949.604212] ? show_regs_print_info+0x5/0x5 [ 2949.604212] ___might_sleep+0x2eb/0x420 [ 2949.604212] ? set_rq_offline.part.87+0x140/0x140 [ 2949.604212] ? _rcu_barrier_trace+0x400/0x400 [ 2949.604212] wait_for_completion+0x94/0x710 [ 2949.604212] ? wait_for_completion_interruptible+0x780/0x780 [ 2949.604212] ? __kernel_text_address+0xe/0x30 [ 2949.604212] ? __lockdep_init_map+0x10e/0x5c0 [ 2949.604212] ? __lockdep_init_map+0x10e/0x5c0 [ 2949.604212] ? __init_waitqueue_head+0x86/0x130 [ 2949.604212] ? init_wait_entry+0x1a0/0x1a0 [ 2949.604212] proc_entry_rundown+0x208/0x270 [ 2949.604212] ? proc_reg_get_unmapped_area+0x370/0x370 [ 2949.604212] ? __lock_acquire+0x4500/0x4500 [ 2949.604212] ? complete+0x18/0x70 [ 2949.604212] remove_proc_subtree+0x143/0x2a0 [ 2949.708655] ? remove_proc_entry+0x390/0x390 [ 2949.708655] clusterip_tg_destroy+0x27a/0x630 [ipt_CLUSTERIP] [ ... ] Fixes: b3e456fce9f5 ("netfilter: ipt_CLUSTERIP: fix a race condition of proc file creation") Signed-off-by: Taehee Yoo <ap420073@gmail.com> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
| * | netfilter: ipt_CLUSTERIP: remove wrong WARN_ON_ONCE in netns exit routineTaehee Yoo2018-12-181-1/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When network namespace is destroyed, both clusterip_tg_destroy() and clusterip_net_exit() are called. and clusterip_net_exit() is called before clusterip_tg_destroy(). Hence cleanup check code in clusterip_net_exit() doesn't make sense. test commands: %ip netns add vm1 %ip netns exec vm1 bash %ip link set lo up %iptables -A INPUT -p tcp -i lo -d 192.168.0.5 --dport 80 \ -j CLUSTERIP --new --hashmode sourceip \ --clustermac 01:00:5e:00:00:20 --total-nodes 2 --local-node 1 %exit %ip netns del vm1 splat looks like: [ 341.184508] WARNING: CPU: 1 PID: 87 at net/ipv4/netfilter/ipt_CLUSTERIP.c:840 clusterip_net_exit+0x319/0x380 [ipt_CLUSTERIP] [ 341.184850] Modules linked in: ipt_CLUSTERIP nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 xt_tcpudp iptable_filter bpfilter ip_tables x_tables [ 341.184850] CPU: 1 PID: 87 Comm: kworker/u4:2 Not tainted 4.19.0-rc5+ #16 [ 341.227509] Workqueue: netns cleanup_net [ 341.227509] RIP: 0010:clusterip_net_exit+0x319/0x380 [ipt_CLUSTERIP] [ 341.227509] Code: 0f 85 7f fe ff ff 48 c7 c2 80 64 2c c0 be a8 02 00 00 48 c7 c7 a0 63 2c c0 c6 05 18 6e 00 00 01 e8 bc 38 ff f5 e9 5b fe ff ff <0f> 0b e9 33 ff ff ff e8 4b 90 50 f6 e9 2d fe ff ff 48 89 df e8 de [ 341.227509] RSP: 0018:ffff88011086f408 EFLAGS: 00010202 [ 341.227509] RAX: dffffc0000000000 RBX: 1ffff1002210de85 RCX: 0000000000000000 [ 341.227509] RDX: 1ffff1002210de85 RSI: ffff880110813be8 RDI: ffffed002210de58 [ 341.227509] RBP: ffff88011086f4d0 R08: 0000000000000000 R09: 0000000000000000 [ 341.227509] R10: 0000000000000000 R11: 0000000000000000 R12: 1ffff1002210de81 [ 341.227509] R13: ffff880110625a48 R14: ffff880114cec8c8 R15: 0000000000000014 [ 341.227509] FS: 0000000000000000(0000) GS:ffff880116600000(0000) knlGS:0000000000000000 [ 341.227509] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 341.227509] CR2: 00007f11fd38e000 CR3: 000000013ca16000 CR4: 00000000001006e0 [ 341.227509] Call Trace: [ 341.227509] ? __clusterip_config_find+0x460/0x460 [ipt_CLUSTERIP] [ 341.227509] ? default_device_exit+0x1ca/0x270 [ 341.227509] ? remove_proc_entry+0x1cd/0x390 [ 341.227509] ? dev_change_net_namespace+0xd00/0xd00 [ 341.227509] ? __init_waitqueue_head+0x130/0x130 [ 341.227509] ops_exit_list.isra.10+0x94/0x140 [ 341.227509] cleanup_net+0x45b/0x900 [ ... ] Fixes: 613d0776d3fe ("netfilter: exit_net cleanup check added") Signed-off-by: Taehee Yoo <ap420073@gmail.com> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
| * | netfilter: ipt_CLUSTERIP: fix deadlock in netns exit routineTaehee Yoo2018-12-181-68/+87
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When network namespace is destroyed, cleanup_net() is called. cleanup_net() holds pernet_ops_rwsem then calls each ->exit callback. So that clusterip_tg_destroy() is called by cleanup_net(). And clusterip_tg_destroy() calls unregister_netdevice_notifier(). But both cleanup_net() and clusterip_tg_destroy() hold same lock(pernet_ops_rwsem). hence deadlock occurrs. After this patch, only 1 notifier is registered when module is inserted. And all of configs are added to per-net list. test commands: %ip netns add vm1 %ip netns exec vm1 bash %ip link set lo up %iptables -A INPUT -p tcp -i lo -d 192.168.0.5 --dport 80 \ -j CLUSTERIP --new --hashmode sourceip \ --clustermac 01:00:5e:00:00:20 --total-nodes 2 --local-node 1 %exit %ip netns del vm1 splat looks like: [ 341.809674] ============================================ [ 341.809674] WARNING: possible recursive locking detected [ 341.809674] 4.19.0-rc5+ #16 Tainted: G W [ 341.809674] -------------------------------------------- [ 341.809674] kworker/u4:2/87 is trying to acquire lock: [ 341.809674] 000000005da2d519 (pernet_ops_rwsem){++++}, at: unregister_netdevice_notifier+0x8c/0x460 [ 341.809674] [ 341.809674] but task is already holding lock: [ 341.809674] 000000005da2d519 (pernet_ops_rwsem){++++}, at: cleanup_net+0x119/0x900 [ 341.809674] [ 341.809674] other info that might help us debug this: [ 341.809674] Possible unsafe locking scenario: [ 341.809674] [ 341.809674] CPU0 [ 341.809674] ---- [ 341.809674] lock(pernet_ops_rwsem); [ 341.809674] lock(pernet_ops_rwsem); [ 341.809674] [ 341.809674] *** DEADLOCK *** [ 341.809674] [ 341.809674] May be due to missing lock nesting notation [ 341.809674] [ 341.809674] 3 locks held by kworker/u4:2/87: [ 341.809674] #0: 00000000d9df6c92 ((wq_completion)"%s""netns"){+.+.}, at: process_one_work+0xafe/0x1de0 [ 341.809674] #1: 00000000c2cbcee2 (net_cleanup_work){+.+.}, at: process_one_work+0xb60/0x1de0 [ 341.809674] #2: 000000005da2d519 (pernet_ops_rwsem){++++}, at: cleanup_net+0x119/0x900 [ 341.809674] [ 341.809674] stack backtrace: [ 341.809674] CPU: 1 PID: 87 Comm: kworker/u4:2 Tainted: G W 4.19.0-rc5+ #16 [ 341.809674] Workqueue: netns cleanup_net [ 341.809674] Call Trace: [ ... ] [ 342.070196] down_write+0x93/0x160 [ 342.070196] ? unregister_netdevice_notifier+0x8c/0x460 [ 342.070196] ? down_read+0x1e0/0x1e0 [ 342.070196] ? sched_clock_cpu+0x126/0x170 [ 342.070196] ? find_held_lock+0x39/0x1c0 [ 342.070196] unregister_netdevice_notifier+0x8c/0x460 [ 342.070196] ? register_netdevice_notifier+0x790/0x790 [ 342.070196] ? __local_bh_enable_ip+0xe9/0x1b0 [ 342.070196] ? __local_bh_enable_ip+0xe9/0x1b0 [ 342.070196] ? clusterip_tg_destroy+0x372/0x650 [ipt_CLUSTERIP] [ 342.070196] ? trace_hardirqs_on+0x93/0x210 [ 342.070196] ? __bpf_trace_preemptirq_template+0x10/0x10 [ 342.070196] ? clusterip_tg_destroy+0x372/0x650 [ipt_CLUSTERIP] [ 342.123094] clusterip_tg_destroy+0x3ad/0x650 [ipt_CLUSTERIP] [ 342.123094] ? clusterip_net_init+0x3d0/0x3d0 [ipt_CLUSTERIP] [ 342.123094] ? cleanup_match+0x17d/0x200 [ip_tables] [ 342.123094] ? xt_unregister_table+0x215/0x300 [x_tables] [ 342.123094] ? kfree+0xe2/0x2a0 [ 342.123094] cleanup_entry+0x1d5/0x2f0 [ip_tables] [ 342.123094] ? cleanup_match+0x200/0x200 [ip_tables] [ 342.123094] __ipt_unregister_table+0x9b/0x1a0 [ip_tables] [ 342.123094] iptable_filter_net_exit+0x43/0x80 [iptable_filter] [ 342.123094] ops_exit_list.isra.10+0x94/0x140 [ 342.123094] cleanup_net+0x45b/0x900 [ ... ] Fixes: 202f59afd441 ("netfilter: ipt_CLUSTERIP: do not hold dev") Signed-off-by: Taehee Yoo <ap420073@gmail.com> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
| * | netfilter: nat: remove nf_nat_l4proto structFlorian Westphal2018-12-174-107/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This removes the (now empty) nf_nat_l4proto struct, all its instances and all the no longer needed runtime (un)register functionality. nf_nat_need_gre() can be axed as well: the module that calls it (to load the no-longer-existing nat_gre module) also calls other nat core functions. GRE nat is now always available if kernel is built with it. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
| * | netfilter: nat: remove l4proto->manip_pktFlorian Westphal2018-12-175-71/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This removes the last l4proto indirection, the two callers, the l3proto packet mangling helpers for ipv4 and ipv6, now call the nf_nat_l4proto_manip_pkt() helper. nf_nat_proto_{dccp,tcp,sctp,gre,icmp,icmpv6} are left behind, even though they contain no functionality anymore to not clutter this patch. Next patch will remove the empty files and the nf_nat_l4proto struct. nf_nat_proto_udp.c is renamed to nf_nat_proto.c, as it now contains the other nat manip functionality as well, not just udp and udplite. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
| * | netfilter: nat: remove l4proto->nlattr_to_rangeFlorian Westphal2018-12-172-6/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | all protocols did set this to nf_nat_l4proto_nlattr_to_range, so just call it directly. The important difference is that we'll now also call it for protocols that we don't support (i.e., nf_nat_proto_unknown did not provide .nlattr_to_range). However, there should be no harm, even icmp provided this callback. If we don't implement a specific l4nat for this, nothing would make use of this information, so adding a big switch/case construct listing all supported l4protocols seems a bit pointless. This change leaves a single function pointer in the l4proto struct. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
| * | netfilter: nat: remove l4proto->in_rangeFlorian Westphal2018-12-172-12/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | With exception of icmp, all of the l4 nat protocols set this to nf_nat_l4proto_in_range. Get rid of this and just check the l4proto in the caller. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
| * | netfilter: nat: fold in_range indirection into callerFlorian Westphal2018-12-171-8/+0
| | | | | | | | | | | | | | | | | | | | | | | | No need for indirections here, we only support ipv4 and ipv6 and the called functions are very small. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
| * | netfilter: nat: remove l4proto->unique_tupleFlorian Westphal2018-12-172-71/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | fold remaining users (icmp, icmpv6, gre) into nf_nat_l4proto_unique_tuple. The static-save of old incarnation of resolved key in gre and icmp is removed as well, just use the prandom based offset like the others. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
| * | netfilter: remove NF_NAT_RANGE_PROTO_RANDOM supportFlorian Westphal2018-12-171-7/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Historically this was net_random() based, and was then converted to a hash based algorithm (private boot seed + hash of endpoint addresses) due to concerns of leaking net_random() bits. RANDOM_FULLY mode was added later to avoid problems with hash based mode (see commit 34ce324019e76, "netfilter: nf_nat: add full port randomization support" for details). Just make prandom_u32() the default search starting point and get rid of ->secure_port() altogether. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
OpenPOWER on IntegriCloud