Dave observed number of machines hitting OOM on the UDP send
path. The workload seems to be sending large UDP packets over
loopback. Since loopback has MTU of 64k kernel will try to
allocate an skb with up to 64k of head space. This has a good
chance of failing under memory pressure. What's worse if
the message length is <32k the allocation may trigger an
OOM killer.
This is entirely avoidable, we can use an skb with page frags.
af_unix solves a similar problem by limiting the head
length to SKB_MAX_ALLOC. This seems like a good and simple
approach. It means that UDP messages > 16kB will now
use fragments if underlying device supports SG, if extra
allocator pressure causes regressions in real workloads
we can switch to trying the large allocation first and
falling back.
v4: pre-calculate all the additions to alloclen so
we can be sure it won't go over order-2
Reported-by: Dave Jones <dsj@fb.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Daniel Borkmann says:
====================
pull-request: bpf 2021-06-23
The following pull-request contains BPF updates for your *net* tree.
We've added 14 non-merge commits during the last 6 day(s) which contain
a total of 13 files changed, 137 insertions(+), 64 deletions(-).
Note that when you merge net into net-next, there is a small merge conflict
between 9f2470fbc4 ("skmsg: Improve udp_bpf_recvmsg() accuracy") from bpf
with c49661aa6f ("skmsg: Remove unused parameters of sk_msg_wait_data()")
from net-next. Resolution is to: i) net/ipv4/udp_bpf.c: take udp_msg_wait_data()
and remove err parameter from the function, ii) net/ipv4/tcp_bpf.c: take
tcp_msg_wait_data() and remove err parameter from the function, iii) for
net/core/skmsg.c and include/linux/skmsg.h: remove the sk_msg_wait_data()
implementation and its prototype in header.
The main changes are:
1) Fix BPF poke descriptor adjustments after insn rewrite, from John Fastabend.
2) Fix regression when using BPF_OBJ_GET with non-O_RDWR flags, from Maciej Żenczykowski.
3) Various bug and error handling fixes for UDP-related sock_map, from Cong Wang.
4) Fix patching of vmlinux BTF IDs with correct endianness, from Tony Ambardar.
5) Two fixes for TX descriptor validation in AF_XDP, from Magnus Karlsson.
6) Fix overflow in size calculation for bpf_map_area_alloc(), from Bui Quang Minh.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit adds two stats for the socket migration feature to evaluate the
effectiveness: LINUX_MIB_TCPMIGRATEREQ(SUCCESS|FAILURE).
If the migration fails because of the own_req race in receiving ACK and
sending SYN+ACK paths, we do not increment the failure stat. Then another
CPU is responsible for the req.
Link: https://lore.kernel.org/bpf/CAK6E8=cgFKuGecTzSCSQ8z3YJ_163C0uwO9yRvfDSE7vOe9mJA@mail.gmail.com/
Suggested-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Acked-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Steffen Klassert says:
====================
pull request (net): ipsec 2021-06-23
1) Don't return a mtu smaller than 1280 on IPv6 pmtu discovery.
From Sabrina Dubroca
2) Fix seqcount rcu-read side in xfrm_policy_lookup_bytype
for the PREEMPT_RT case. From Varad Gautam.
3) Remove a repeated declaration of xfrm_parse_spi.
From Shaokun Zhang.
4) IPv4 beet mode can't handle fragments, but IPv6 does.
commit 68dc022d04 ("xfrm: BEET mode doesn't support
fragments for inner packets") handled IPv4 and IPv6
the same way. Relax the check for IPv6 because fragments
are possible here. From Xin Long.
5) Memory allocation failures are not reported for
XFRMA_ENCAP and XFRMA_COADDR in xfrm_state_construct.
Fix this by moving both cases in front of the function.
6) Fix a missing initialization in the xfrm offload fallback
fail case for bonding devices. From Ayush Sawal.
Please pull or let me know if there are problems.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
When doing source address validation, the flowi4 struct used for
fib_lookup should be in the reverse direction to the given skb.
fl4_dport and fl4_sport returned by fib4_rules_early_flow_dissect
should thus be swapped.
Fixes: 5a847a6e14 ("net/ipv4: Initialize proto and ports in flow struct")
Signed-off-by: Miao Wang <shankerwangmiao@gmail.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
sk_psock_verdict_recv() clones the skb and uses the clone
afterward, so udp_read_sock() should free the skb after using
it, regardless of error or not.
This fixes a real kmemleak.
Fixes: d7f571188e ("udp: Implement ->read_sock() for sockmap")
Signed-off-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Jakub Sitnicki <jakub@cloudflare.com>
Link: https://lore.kernel.org/bpf/20210615021342.7416-4-xiyou.wangcong@gmail.com
I tried to reuse sk_msg_wait_data() for different protocols,
but it turns out it can not be simply reused. For example,
UDP actually uses two queues to receive skb:
udp_sk(sk)->reader_queue and sk->sk_receive_queue. So we have
to check both of them to know whether we have received any
packet.
Also, UDP does not lock the sock during BH Rx path, it makes
no sense for its ->recvmsg() to lock the sock. It is always
possible for ->recvmsg() to be called before packets actually
arrive in the receive queue, we just use best effort to make
it accurate here.
Fixes: 1f5be6b3b0 ("udp: Implement udp_bpf_recvmsg() for sockmap")
Signed-off-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Jakub Sitnicki <jakub@cloudflare.com>
Link: https://lore.kernel.org/bpf/20210615021342.7416-2-xiyou.wangcong@gmail.com
Trivial conflicts in net/can/isotp.c and
tools/testing/selftests/net/mptcp/mptcp_connect.sh
scaled_ppm_to_ppb() was moved from drivers/ptp/ptp_clock.c
to include/linux/ptp_clock_kernel.h in -next so re-apply
the fix there.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When constructing ICMP response messages, the kernel will try to pick a
suitable source address for the outgoing packet. However, if no IPv4
addresses are configured on the system at all, this will fail and we end up
producing an ICMP message with a source address of 0.0.0.0. This can happen
on a box routing IPv4 traffic via v6 nexthops, for instance.
Since 0.0.0.0 is not generally routable on the internet, there's a good
chance that such ICMP messages will never make it back to the sender of the
original packet that the ICMP message was sent in response to. This, in
turn, can create connectivity and PMTUd problems for senders. Fortunately,
RFC7600 reserves a dummy address to be used as a source for ICMP
messages (192.0.0.8/32), so let's teach the kernel to substitute that
address as a last resort if the regular source address selection procedure
fails.
Below is a quick example reproducing this issue with network namespaces:
ip netns add ns0
ip l add type veth peer netns ns0
ip l set dev veth0 up
ip a add 10.0.0.1/24 dev veth0
ip a add fc00:dead:cafe:42::1/64 dev veth0
ip r add 10.1.0.0/24 via inet6 fc00:dead:cafe:42::2
ip -n ns0 l set dev veth0 up
ip -n ns0 a add fc00:dead:cafe:42::2/64 dev veth0
ip -n ns0 r add 10.0.0.0/24 via inet6 fc00:dead:cafe:42::1
ip netns exec ns0 sysctl -w net.ipv4.icmp_ratelimit=0
ip netns exec ns0 sysctl -w net.ipv4.ip_forward=1
tcpdump -tpni veth0 -c 2 icmp &
ping -w 1 10.1.0.1 > /dev/null
tcpdump: verbose output suppressed, use -v[v]... for full protocol decode
listening on veth0, link-type EN10MB (Ethernet), snapshot length 262144 bytes
IP 10.0.0.1 > 10.1.0.1: ICMP echo request, id 29, seq 1, length 64
IP 0.0.0.0 > 10.0.0.1: ICMP net 10.1.0.1 unreachable, length 92
2 packets captured
2 packets received by filter
0 packets dropped by kernel
With this patch the above capture changes to:
IP 10.0.0.1 > 10.1.0.1: ICMP echo request, id 31127, seq 1, length 64
IP 192.0.0.8 > 10.0.0.1: ICMP net 10.1.0.1 unreachable, length 92
Fixes: 1da177e4c3 ("Linux-2.6.12-rc2")
Reported-by: Juliusz Chroboczek <jch@irif.fr>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Daniel Borkmann says:
====================
pull-request: bpf-next 2021-06-17
The following pull-request contains BPF updates for your *net-next* tree.
We've added 50 non-merge commits during the last 25 day(s) which contain
a total of 148 files changed, 4779 insertions(+), 1248 deletions(-).
The main changes are:
1) BPF infrastructure to migrate TCP child sockets from a listener to another
in the same reuseport group/map, from Kuniyuki Iwashima.
2) Add a provably sound, faster and more precise algorithm for tnum_mul() as
noted in https://arxiv.org/abs/2105.05398, from Harishankar Vishwanathan.
3) Streamline error reporting changes in libbpf as planned out in the
'libbpf: the road to v1.0' effort, from Andrii Nakryiko.
4) Add broadcast support to xdp_redirect_map(), from Hangbin Liu.
5) Extends bpf_map_lookup_and_delete_elem() functionality to 4 more map
types, that is, {LRU_,PERCPU_,LRU_PERCPU_,}HASH, from Denis Salopek.
6) Support new LLVM relocations in libbpf to make them more linker friendly,
also add a doc to describe the BPF backend relocations, from Yonghong Song.
7) Silence long standing KUBSAN complaints on register-based shifts in
interpreter, from Daniel Borkmann and Eric Biggers.
8) Add dummy PT_REGS macros in libbpf to fail BPF program compilation when
target arch cannot be determined, from Lorenz Bauer.
9) Extend AF_XDP to support large umems with 1M+ pages, from Magnus Karlsson.
10) Fix two minor libbpf tc BPF API issues, from Kumar Kartikeya Dwivedi.
11) Move libbpf BPF_SEQ_PRINTF/BPF_SNPRINTF macros that can be used by BPF
programs to bpf_helpers.h header, from Florent Revest.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
BUG: memory leak
unreferenced object 0xffff888101bc4c00 (size 32):
comm "syz-executor527", pid 360, jiffies 4294807421 (age 19.329s)
hex dump (first 32 bytes):
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
01 00 00 00 00 00 00 00 ac 14 14 bb 00 00 02 00 ................
backtrace:
[<00000000f17c5244>] kmalloc include/linux/slab.h:558 [inline]
[<00000000f17c5244>] kzalloc include/linux/slab.h:688 [inline]
[<00000000f17c5244>] ip_mc_add1_src net/ipv4/igmp.c:1971 [inline]
[<00000000f17c5244>] ip_mc_add_src+0x95f/0xdb0 net/ipv4/igmp.c:2095
[<000000001cb99709>] ip_mc_source+0x84c/0xea0 net/ipv4/igmp.c:2416
[<0000000052cf19ed>] do_ip_setsockopt net/ipv4/ip_sockglue.c:1294 [inline]
[<0000000052cf19ed>] ip_setsockopt+0x114b/0x30c0 net/ipv4/ip_sockglue.c:1423
[<00000000477edfbc>] raw_setsockopt+0x13d/0x170 net/ipv4/raw.c:857
[<00000000e75ca9bb>] __sys_setsockopt+0x158/0x270 net/socket.c:2117
[<00000000bdb993a8>] __do_sys_setsockopt net/socket.c:2128 [inline]
[<00000000bdb993a8>] __se_sys_setsockopt net/socket.c:2125 [inline]
[<00000000bdb993a8>] __x64_sys_setsockopt+0xba/0x150 net/socket.c:2125
[<000000006a1ffdbd>] do_syscall_64+0x40/0x80 arch/x86/entry/common.c:47
[<00000000b11467c4>] entry_SYSCALL_64_after_hwframe+0x44/0xae
In commit 24803f38a5 ("igmp: do not remove igmp souce list info when set
link down"), the ip_mc_clear_src() in ip_mc_destroy_dev() was removed,
because it was also called in igmpv3_clear_delrec().
Rough callgraph:
inetdev_destroy
-> ip_mc_destroy_dev
-> igmpv3_clear_delrec
-> ip_mc_clear_src
-> RCU_INIT_POINTER(dev->ip_ptr, NULL)
However, ip_mc_clear_src() called in igmpv3_clear_delrec() doesn't
release in_dev->mc_list->sources. And RCU_INIT_POINTER() assigns the
NULL to dev->ip_ptr. As a result, in_dev cannot be obtained through
inetdev_by_index() and then in_dev->mc_list->sources cannot be released
by ip_mc_del1_src() in the sock_close. Rough call sequence goes like:
sock_close
-> __sock_release
-> inet_release
-> ip_mc_drop_socket
-> inetdev_by_index
-> ip_mc_leave_src
-> ip_mc_del_src
-> ip_mc_del1_src
So we still need to call ip_mc_clear_src() in ip_mc_destroy_dev() to free
in_dev->mc_list->sources.
Fixes: 24803f38a5 ("igmp: do not remove igmp souce list info ...")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Chengyang Fan <cy.fan@huawei.com>
Acked-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Timewait sockets have included mark since approx 4.18.
Cc: Eric Dumazet <edumazet@google.com>
Cc: Jon Maxwell <jmaxwell37@gmail.com>
Fixes: 0048369055 ("tcp: Add mark for TIMEWAIT sockets")
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Jon Maxwell <jmaxwell37@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch also changes the code to call reuseport_migrate_sock() and
inet_reqsk_clone(), but unlike the other cases, we do not call
inet_reqsk_clone() right after reuseport_migrate_sock().
Currently, in the receive path for TCP_NEW_SYN_RECV sockets, its listener
has three kinds of refcnt:
(A) for listener itself
(B) carried by reuqest_sock
(C) sock_hold() in tcp_v[46]_rcv()
While processing the req, (A) may disappear by close(listener). Also, (B)
can disappear by accept(listener) once we put the req into the accept
queue. So, we have to hold another refcnt (C) for the listener to prevent
use-after-free.
For socket migration, we call reuseport_migrate_sock() to select a listener
with (A) and to increment the new listener's refcnt in tcp_v[46]_rcv().
This refcnt corresponds to (C) and is cleaned up later in tcp_v[46]_rcv().
Thus we have to take another refcnt (B) for the newly cloned request_sock.
In inet_csk_complete_hashdance(), we hold the count (B), clone the req, and
try to put the new req into the accept queue. By migrating req after
winning the "own_req" race, we can avoid such a worst situation:
CPU 1 looks up req1
CPU 2 looks up req1, unhashes it, then CPU 1 loses the race
CPU 3 looks up req2, unhashes it, then CPU 2 loses the race
...
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20210612123224.12525-8-kuniyu@amazon.co.jp
As with the preceding patch, this patch changes reqsk_timer_handler() to
call reuseport_migrate_sock() and inet_reqsk_clone() to migrate in-flight
requests at retransmitting SYN+ACKs. If we can select a new listener and
clone the request, we resume setting the SYN+ACK timer for the new req. If
we can set the timer, we call inet_ehash_insert() to unhash the old req and
put the new req into ehash.
The noteworthy point here is that by unhashing the old req, another CPU
processing it may lose the "own_req" race in tcp_v[46]_syn_recv_sock() and
drop the final ACK packet. However, the new timer will recover this
situation.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20210612123224.12525-7-kuniyu@amazon.co.jp
When we call close() or shutdown() for listening sockets, each child socket
in the accept queue are freed at inet_csk_listen_stop(). If we can get a
new listener by reuseport_migrate_sock() and clone the request by
inet_reqsk_clone(), we try to add it into the new listener's accept queue
by inet_csk_reqsk_queue_add(). If it fails, we have to call __reqsk_free()
to call sock_put() for its listener and free the cloned request.
After putting the full socket into ehash, tcp_v[46]_syn_recv_sock() sets
NULL to ireq_opt/pktopts in struct inet_request_sock, but ipv6_opt can be
non-NULL. So, we have to set NULL to ipv6_opt of the old request to avoid
double free.
Note that we do not update req->rsk_listener and instead clone the req to
migrate because another path may reference the original request. If we
protected it by RCU, we would need to add rcu_read_lock() in many places.
Suggested-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/netdev/20201209030903.hhow5r53l6fmozjn@kafai-mbp.dhcp.thefacebook.com/
Link: https://lore.kernel.org/bpf/20210612123224.12525-6-kuniyu@amazon.co.jp
When we close a listening socket, to migrate its connections to another
listener in the same reuseport group, we have to handle two kinds of child
sockets. One is that a listening socket has a reference to, and the other
is not.
The former is the TCP_ESTABLISHED/TCP_SYN_RECV sockets, and they are in the
accept queue of their listening socket. So we can pop them out and push
them into another listener's queue at close() or shutdown() syscalls. On
the other hand, the latter, the TCP_NEW_SYN_RECV socket is during the
three-way handshake and not in the accept queue. Thus, we cannot access
such sockets at close() or shutdown() syscalls. Accordingly, we have to
migrate immature sockets after their listening socket has been closed.
Currently, if their listening socket has been closed, TCP_NEW_SYN_RECV
sockets are freed at receiving the final ACK or retransmitting SYN+ACKs. At
that time, if we could select a new listener from the same reuseport group,
no connection would be aborted. However, we cannot do that because
reuseport_detach_sock() sets NULL to sk_reuseport_cb and forbids access to
the reuseport group from closed sockets.
This patch allows TCP_CLOSE sockets to remain in the reuseport group and
access it while any child socket references them. The point is that
reuseport_detach_sock() was called twice from inet_unhash() and
sk_destruct(). This patch replaces the first reuseport_detach_sock() with
reuseport_stop_listen_sock(), which checks if the reuseport group is
capable of migration. If capable, it decrements num_socks, moves the socket
backwards in socks[] and increments num_closed_socks. When all connections
are migrated, sk_destruct() calls reuseport_detach_sock() to remove the
socket from socks[], decrement num_closed_socks, and set NULL to
sk_reuseport_cb.
By this change, closed or shutdowned sockets can keep sk_reuseport_cb.
Consequently, calling listen() after shutdown() can cause EADDRINUSE or
EBUSY in inet_csk_bind_conflict() or reuseport_add_sock() which expects
such sockets not to have the reuseport group. Therefore, this patch also
loosens such validation rules so that a socket can listen again if it has a
reuseport group with num_closed_socks more than 0.
When such sockets listen again, we handle them in reuseport_resurrect(). If
there is an existing reuseport group (reuseport_add_sock() path), we move
the socket from the old group to the new one and free the old one if
necessary. If there is no existing group (reuseport_alloc() path), we
allocate a new reuseport group, detach sk from the old one, and free it if
necessary, not to break the current shutdown behaviour:
- we cannot carry over the eBPF prog of shutdowned sockets
- we cannot attach/detach an eBPF prog to/from listening sockets via
shutdowned sockets
Note that when the number of sockets gets over U16_MAX, we try to detach a
closed socket randomly to make room for the new listening socket in
reuseport_grow().
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/bpf/20210612123224.12525-4-kuniyu@amazon.co.jp
This commit adds a new sysctl option: net.ipv4.tcp_migrate_req. If this
option is enabled or eBPF program is attached, we will be able to migrate
child sockets from a listener to another in the same reuseport group after
close() or shutdown() syscalls.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Benjamin Herrenschmidt <benh@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20210612123224.12525-2-kuniyu@amazon.co.jp
Oliver reported a use case where deleting a VRF device can hang
waiting for the refcnt to drop to 0. The root cause is that the dst
is allocated against the VRF device but cached on the loopback
device.
The use case (added to the selftests) has an implicit VRF crossing
due to the ordering of the FIB rules (lookup local is before the
l3mdev rule, but the problem occurs even if the FIB rules are
re-ordered with local after l3mdev because the VRF table does not
have a default route to terminate the lookup). The end result is
is that the FIB lookup returns the loopback device as the nexthop,
but the ingress device is in a VRF. The mismatch causes the dst
alloc against the VRF device but then cached on the loopback.
The fix is to bring the trick used for IPv6 (see ip6_rt_get_dev_rcu):
pick the dst alloc device based the fib lookup result but with checks
that the result has a nexthop device (e.g., not an unreachable or
prohibit entry).
Fixes: f5a0aab84b ("net: ipv4: dst for local input routes should use l3mdev if relevant")
Reported-by: Oliver Herms <oliver.peter.herms@gmail.com>
Signed-off-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Function 'ping_queue_rcv_skb' not always return success, which will
also return fail. If not check the wrong return value of it, lead to function
`ping_rcv` return success.
Signed-off-by: Zheng Yongjun <zhengyongjun3@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Kaustubh reported and diagnosed a panic in udp_lib_lookup().
The root cause is udp_abort() racing with close(). Both
racing functions acquire the socket lock, but udp{v6}_destroy_sock()
release it before performing destructive actions.
We can't easily extend the socket lock scope to avoid the race,
instead use the SOCK_DEAD flag to prevent udp_abort from doing
any action when the critical race happens.
Diagnosed-and-tested-by: Kaustubh Pandey <kapandey@codeaurora.org>
Fixes: 5d77dca828 ("net: diag: support SOCK_DESTROY for UDP sockets")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Both functions are known to be racy when reading inet_num
as we do not want to grab locks for the common case the socket
has been bound already. The race is resolved in inet_autobind()
by reading again inet_num under the socket lock.
syzbot reported:
BUG: KCSAN: data-race in inet_send_prepare / udp_lib_get_port
write to 0xffff88812cba150e of 2 bytes by task 24135 on cpu 0:
udp_lib_get_port+0x4b2/0xe20 net/ipv4/udp.c:308
udp_v6_get_port+0x5e/0x70 net/ipv6/udp.c:89
inet_autobind net/ipv4/af_inet.c:183 [inline]
inet_send_prepare+0xd0/0x210 net/ipv4/af_inet.c:807
inet6_sendmsg+0x29/0x80 net/ipv6/af_inet6.c:639
sock_sendmsg_nosec net/socket.c:654 [inline]
sock_sendmsg net/socket.c:674 [inline]
____sys_sendmsg+0x360/0x4d0 net/socket.c:2350
___sys_sendmsg net/socket.c:2404 [inline]
__sys_sendmmsg+0x315/0x4b0 net/socket.c:2490
__do_sys_sendmmsg net/socket.c:2519 [inline]
__se_sys_sendmmsg net/socket.c:2516 [inline]
__x64_sys_sendmmsg+0x53/0x60 net/socket.c:2516
do_syscall_64+0x4a/0x90 arch/x86/entry/common.c:47
entry_SYSCALL_64_after_hwframe+0x44/0xae
read to 0xffff88812cba150e of 2 bytes by task 24132 on cpu 1:
inet_send_prepare+0x21/0x210 net/ipv4/af_inet.c:806
inet6_sendmsg+0x29/0x80 net/ipv6/af_inet6.c:639
sock_sendmsg_nosec net/socket.c:654 [inline]
sock_sendmsg net/socket.c:674 [inline]
____sys_sendmsg+0x360/0x4d0 net/socket.c:2350
___sys_sendmsg net/socket.c:2404 [inline]
__sys_sendmmsg+0x315/0x4b0 net/socket.c:2490
__do_sys_sendmmsg net/socket.c:2519 [inline]
__se_sys_sendmmsg net/socket.c:2516 [inline]
__x64_sys_sendmmsg+0x53/0x60 net/socket.c:2516
do_syscall_64+0x4a/0x90 arch/x86/entry/common.c:47
entry_SYSCALL_64_after_hwframe+0x44/0xae
value changed: 0x0000 -> 0x9db4
Reported by Kernel Concurrency Sanitizer on:
CPU: 1 PID: 24132 Comm: syz-executor.2 Not tainted 5.13.0-rc4-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Its set but never read. Reduces size of xfrm_type to 64 bytes on 64bit.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
When 'nla_parse_nested_deprecated' failed, it's no need to
BUG() here, return -EINVAL is ok.
Signed-off-by: Zheng Yongjun <zhengyongjun3@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix some spelling mistakes in comments:
Dont ==> Don't
timout ==> timeout
incomming ==> incoming
necesarry ==> necessary
substract ==> subtract
Signed-off-by: Zheng Yongjun <zhengyongjun3@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
MPTCP is builtin, so no need to add EXPORT_SYMBOL()s.
It will be used to support SO_TIMESTAMP(NS) ancillary
messages in the mptcp receive path.
Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Including <linux/in.h> and <netinet/in.h> in the dependencies breaks
compilation of trinity due to multiple definitions. <linux/in.h> is only
used in <linux/icmp.h> to provide the definition of the struct in_addr,
but this can be substituted out by using the datatype __be32.
Signed-off-by: Andreas Roeseler <andreas.a.roeseler@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch aims to improve the situation when reordering and loss are
ocurring in the same flight of packets.
Previously the reordering would first induce a spurious recovery, then
the subsequent ACK may undo the cwnd (based on the timestamps e.g.).
However the current loss recovery does not proceed to invoke
RACK to install a reordering timer. If some packets are also lost, this
may lead to a long RTO-based recovery. An example is
https://groups.google.com/g/bbr-dev/c/OFHADvJbTEI
The solution is to after reverting the recovery, always invoke RACK
to either mount the RACK timer to fast retransmit after the reordering
window, or restarts the recovery if new loss is identified. Hence
it is possible the sender may go from Recovery to Disorder/Open to
Recovery again in one ACK.
Reported-by: mingkun bian <bianmingkun@gmail.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If the user specifies a hostname or domain name as part of the ip=
command-line option, preserve it and don't overwrite it with one
supplied by DHCP/BOOTP.
For instance, ip=::::myhostname::dhcp will use "myhostname" rather than
ignoring and overwriting it.
Fix the comment on ic_bootp_string that suggests it only copies a string
"if not already set"; it doesn't have any such logic.
Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pablo Neira Ayuso says:
====================
Netfilter updates for net-next
The following patchset contains Netfilter updates for net-next:
1) Support for SCTP chunks matching on nf_tables, from Phil Sutter.
2) Skip LDMXCSR, we don't need a valid MXCSR state. From Stefano Brivio.
3) CONFIG_RETPOLINE for nf_tables set lookups, from Florian Westphal.
4) A few Kconfig leading spaces removal, from Juerg Haefliger.
5) Remove spinlock from xt_limit, from Jason Baron.
6) Remove useless initialization in xt_CT, oneliner from Yang Li.
7) Tree-wide replacement of netlink_unicast() by nfnetlink_unicast().
8) Reduce footprint of several structures: xt_action_param,
nft_pktinfo and nf_hook_state, from Florian.
10) Add nft_thoff() and nft_sk() helpers and use them, also from Florian.
11) Fix documentation in nf_tables pipapo avx2, from Florian Westphal.
12) Fix clang-12 fmt string warnings, also from Florian.
====================
When kalloc or kmemdup failed, should return ENOMEM rather than ENOBUF.
Signed-off-by: Zheng Yongjun <zhengyongjun3@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This allows to change storage placement later on without changing readers.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Alexei Starovoitov says:
====================
pull-request: bpf-next 2021-05-19
The following pull-request contains BPF updates for your *net-next* tree.
We've added 43 non-merge commits during the last 11 day(s) which contain
a total of 74 files changed, 3717 insertions(+), 578 deletions(-).
The main changes are:
1) syscall program type, fd array, and light skeleton, from Alexei.
2) Stop emitting static variables in skeleton, from Andrii.
3) Low level tc-bpf api, from Kumar.
4) Reduce verifier kmalloc/kfree churn, from Lorenz.
====================
In-kernel notifications are already sent when the multipath hash policy
itself changes, but not when the multipath hash fields change.
Add these notifications, so that interested listeners (e.g., switch ASIC
drivers) could perform the necessary configuration.
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since cipso_v4_cache_invalidate() has no return value, so drop
related descriptions in its comments.
Fixes: 446fda4f26 ("[NetLabel]: CIPSOv4 engine")
Signed-off-by: Zheng Yejian <zhengyejian1@huawei.com>
Acked-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a new multipath hash policy where the packet fields used for hash
calculation are determined by user space via the
fib_multipath_hash_fields sysctl that was introduced in the previous
patch.
The current set of available packet fields includes both outer and inner
fields, which requires two invocations of the flow dissector. Avoid
unnecessary dissection of the outer or inner flows by skipping
dissection if none of the outer or inner fields are required.
In accordance with the existing policies, when an skb is not available,
packet fields are extracted from the provided flow key. In which case,
only outer fields are considered.
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
A subsequent patch will add a new multipath hash policy where the packet
fields used for multipath hash calculation are determined by user space.
This patch adds a sysctl that allows user space to set these fields.
The packet fields are represented using a bitmask and are common between
IPv4 and IPv6 to allow user space to use the same numbering across both
protocols. For example, to hash based on standard 5-tuple:
# sysctl -w net.ipv4.fib_multipath_hash_fields=0x0037
net.ipv4.fib_multipath_hash_fields = 0x0037
The kernel rejects unknown fields, for example:
# sysctl -w net.ipv4.fib_multipath_hash_fields=0x1000
sysctl: setting key "net.ipv4.fib_multipath_hash_fields": Invalid argument
More fields can be added in the future, if needed.
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
A subsequent patch will add another multipath hash policy where the
multipath hash is calculated directly by the policy specific code and
not outside of the switch statement.
Prepare for this change by moving the multipath hash calculation inside
the switch statement.
No functional changes intended.
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
'err' and 'flags' are not used, we can just get rid of them.
Signed-off-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Song Liu <song@kernel.org>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20210517022348.50555-1-xiyou.wangcong@gmail.com
In preparation to enable -Wimplicit-fallthrough for Clang, fix multiple
warnings by explicitly adding multiple break statements instead of just
letting the code fall through to the next case.
Link: https://github.com/KSPP/linux/issues/115
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Every protocol has the 'netns_ok' member and it is euqal to 1. The
'if (!prot->netns_ok)' always false in inet_add_protocol().
Signed-off-by: Yejune Deng <yejunedeng@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Treat only the highest, not the lowest, IPv4 address within a local
subnet as a broadcast address.
Signed-off-by: Seth David Schoen <schoen@loyalty.org>
Suggested-by: John Gilmore <gnu@toad.com>
Acked-by: Dave Taht <dave.taht@gmail.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a tracepoint for capturing TCP segments with
a bad checksum. This makes it easy to identify
sources of bad frames in the fleet (e.g. machines
with faulty NICs).
It should also help tools like IOvisor's tcpdrop.py
which are used today to get detailed information
about such packets.
We don't have a socket in many cases so we must
open code the address extraction based just on
the skb.
v2: add missing export for ipv6=m
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Making '!=' operation with 0 directly after calling
the function xfrm_parse_spi() is more efficient,
assignment to err is redundant.
Eliminate the following clang_analyzer warning:
net/ipv4/esp4_offload.c:41:7: warning: Although the value stored to
'err' is used in the enclosing expression, the value is never actually
read from 'err'
No functional change, only more efficient.
Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Daniel Borkmann says:
====================
pull-request: bpf 2021-05-11
The following pull-request contains BPF updates for your *net* tree.
We've added 13 non-merge commits during the last 8 day(s) which contain
a total of 21 files changed, 817 insertions(+), 382 deletions(-).
The main changes are:
1) Fix multiple ringbuf bugs in particular to prevent writable mmap of
read-only pages, from Andrii Nakryiko & Thadeu Lima de Souza Cascardo.
2) Fix verifier alu32 known-const subregister bound tracking for bitwise
operations and/or/xor, from Daniel Borkmann.
3) Reject trampoline attachment for functions with variable arguments,
and also add a deny list of other forbidden functions, from Jiri Olsa.
4) Fix nested bpf_bprintf_prepare() calls used by various helpers by
switching to per-CPU buffers, from Florent Revest.
5) Fix kernel compilation with BTF debug info on ppc64 due to pahole
missing TCP-CC functions like cubictcp_init, from Martin KaFai Lau.
6) Add a kconfig entry to provide an option to disallow unprivileged
BPF by default, from Daniel Borkmann.
7) Fix libbpf compilation for older libelf when GELF_ST_VISIBILITY()
macro is not available, from Arnaldo Carvalho de Melo.
8) Migrate test_tc_redirect to test_progs framework as prep work
for upcoming skb_change_head() fix & selftest, from Jussi Maki.
9) Fix a libbpf segfault in add_dummy_ksym_var() if BTF is not
present, from Ian Rogers.
10) Fix tx_only micro-benchmark in xdpsock BPF sample with proper frame
size, from Magnus Karlsson.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
During the discussion in [0]. It was pointed out that static functions
in ppc64 is prefixed with ".". For example, the 'readelf -s vmlinux.ppc':
89326: c000000001383280 24 NOTYPE LOCAL DEFAULT 31 cubictcp_init
89327: c000000000c97c50 168 FUNC LOCAL DEFAULT 2 .cubictcp_init
The one with FUNC type is ".cubictcp_init" instead of "cubictcp_init".
The "." seems to be done by arch/powerpc/include/asm/ppc_asm.h.
This caused that pahole cannot generate the BTF for these tcp-cc kernel
functions because pahole only captures the FUNC type and "cubictcp_init"
is not. It then failed the kernel compilation in ppc64.
This behavior is only reported in ppc64 so far. I tried arm64, s390,
and sparc64 and did not observe this "." prefix and NOTYPE behavior.
Since the kfunc call is only supported in the x86_64 and x86_32 JIT,
this patch limits those tcp-cc functions to x86 only to avoid unnecessary
compilation issue in other ARCHs. In the future, we can examine if it
is better to change all those functions from static to extern.
[0] https://lore.kernel.org/bpf/4e051459-8532-7b61-c815-f3435767f8a0@kernel.org/
Fixes: e78aea8b21 ("bpf: tcp: Put some tcp cong functions in allowlist for bpf-tcp-cc")
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Cc: Michal Suchánek <msuchanek@suse.de>
Cc: Jiri Slaby <jslaby@suse.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: https://lore.kernel.org/bpf/20210508005011.3863757-1-kafai@fb.com
When we call af_ops->set_link_af() we hold a RCU read lock
as we retrieve af_ops from the RCU protected list, but this
is unnecessary because we already hold RTNL lock, which is
the writer lock for protecting rtnl_af_ops, so it is safer
than RCU read lock. Similar for af_ops->validate_link_af().
This was not a problem until we begin to take mutex lock
down the path of ->set_link_af() in __ipv6_dev_mc_dec()
recently. We can just drop the RCU read lock there and
assert RTNL lock.
Reported-and-tested-by: syzbot+7d941e89dd48bcf42573@syzkaller.appspotmail.com
Fixes: 63ed8de4be ("mld: add mc_lock for protecting per-interface mld data")
Tested-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pablo Neira Ayuso says:
====================
Netfilter fixes for net
1) Add SECMARK revision 1 to fix incorrect layout that prevents
from remove rule with this target, from Phil Sutter.
2) Fix pernet exit path spat in arptables, from Florian Westphal.
3) Missing rcu_read_unlock() for unknown nfnetlink callbacks,
reported by syzbot, from Eric Dumazet.
4) Missing check for skb_header_pointer() NULL pointer in
nfnetlink_osf.
5) Remove BUG_ON() after skb_header_pointer() from packet path
in several conntrack helper and the TCP tracker.
6) Fix memleak in the new object error path of userdata.
7) Avoid overflows in nft_hash_buckets(), reported by syzbot,
also from Eric.
8) Avoid overflows in 32bit arches, from Eric.
* git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf:
netfilter: nftables: avoid potential overflows on 32bit arches
netfilter: nftables: avoid overflows in nft_hash_buckets()
netfilter: nftables: Fix a memleak from userdata error path in new objects
netfilter: remove BUG_ON() after skb_header_pointer()
netfilter: nfnetlink_osf: Fix a missing skb_header_pointer() NULL check
netfilter: nfnetlink: add a missing rcu_read_unlock()
netfilter: arptables: use pernet ops struct during unregister
netfilter: xt_SECMARK: add new revision to fix structure layout
====================
Link: https://lore.kernel.org/r/20210507174739.1850-1-pablo@netfilter.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
A prior change (1f466e1f15) introduces separate handling for
->msg_control depending on whether the pointer is a kernel or user
pointer. However, while tcp receive zerocopy is using this field, it
is not properly annotating that the buffer in this case is a user
pointer. This can cause faults when the improper mechanism is used
within put_cmsg().
This patch simply annotates tcp receive zerocopy's use as explicitly
being a user pointer.
Fixes: 7eeba1706e ("tcp: Add receive timestamp support for receive zerocopy.")
Signed-off-by: Arjun Roy <arjunroy@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20210506223530.2266456-1-arjunroy.kdev@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
tcp_set_default_congestion_control() is netns-safe in that it writes
to &net->ipv4.tcp_congestion_control, but it also sets
ca->flags |= TCP_CONG_NON_RESTRICTED which is not namespaced.
This has the unintended side-effect of changing the global
net.ipv4.tcp_allowed_congestion_control sysctl, despite the fact that it
is read-only: 97684f0970 ("net: Make tcp_allowed_congestion_control
readonly in non-init netns")
Resolve this netns "leak" by only allowing the init netns to set the
default algorithm to one that is restricted. This restriction could be
removed if tcp_allowed_congestion_control were namespace-ified in the
future.
This bug was uncovered with
https://github.com/JonathonReinhart/linux-netns-sysctl-verify
Fixes: 6670e15244 ("tcp: Namespace-ify sysctl_tcp_default_congestion_control")
Signed-off-by: Jonathon Reinhart <jonathon.reinhart@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Like with iptables and ebtables, hook unregistration has to use the
pernet ops struct, not the template.
This triggered following splat:
hook not found, pf 3 num 0
WARNING: CPU: 0 PID: 224 at net/netfilter/core.c:480 __nf_unregister_net_hook+0x1eb/0x610 net/netfilter/core.c:480
[..]
nf_unregister_net_hook net/netfilter/core.c:502 [inline]
nf_unregister_net_hooks+0x117/0x160 net/netfilter/core.c:576
arpt_unregister_table_pre_exit+0x67/0x80 net/ipv4/netfilter/arp_tables.c:1565
Fixes: f9006acc8d ("netfilter: arp_tables: pass table pointer via nf_hook_ops")
Reported-by: syzbot+dcccba8a1e41a38cb9df@syzkaller.appspotmail.com
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
The current definitions of constants for PROBE, currently defined only
in the net-next kernel branch, are inconsistent, with
some beginning with ICMP and others with simply EXT. This patch
attempts to standardize the naming conventions of the constants for
PROBE before their release into a stable Kernel, and to update the
relevant definitions in net/ipv4/icmp.c.
Similarly, the definitions for the code field (previously
ICMP_EXT_MAL_QUERY, etc) use the same prefixes as the type field. This
patch adds _CODE_ to the prefix to clarify the distinction of these
constants.
Signed-off-by: Andreas Roeseler <andreas.a.roeseler@gmail.com>
Acked-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20210427153635.2591-1-andreas.a.roeseler@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Pablo Neira Ayuso says:
====================
Netfilter updates for net-next
The following patchset contains Netfilter updates for net-next:
1) The various ip(6)table_foo incarnations are updated to expect
that the table is passed as 'void *priv' argument that netfilter core
passes to the hook functions. This reduces the struct net size by 2
cachelines on x86_64. From Florian Westphal.
2) Add cgroupsv2 support for nftables.
3) Fix bridge log family merge into nf_log_syslog: Missing
unregistration from netns exit path, from Phil Sutter.
4) Add nft_pernet() helper to access nftables pernet area.
5) Add struct nfnl_info to reduce nfnetlink callback footprint and
to facilite future updates. Consolidate nfnetlink callbacks.
6) Add CONFIG_NETFILTER_XTABLES_COMPAT Kconfig knob, also from Florian.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
The compat layer needs to parse untrusted input (the ruleset)
to translate it to a 64bit compatible format.
We had a number of bugs in this department in the past, so allow users
to turn this feature off.
Add CONFIG_NETFILTER_XTABLES_COMPAT kconfig knob and make it default to y
to keep existing behaviour.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Same change as previous patch. Only difference:
no need to handle NULL template_ops parameter, the only caller
(arptable_filter) always passes non-NULL argument.
This removes all remaining accesses to net->ipv4.arptable_filter.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
iptable_x modules rely on 'struct net' to contain a pointer to the
table that should be evaluated.
In order to remove these pointers from struct net, pass them via
the 'priv' pointer in a similar fashion as nf_tables passes the
rule data.
To do that, duplicate the nf_hook_info array passed in from the
iptable_x modules, update the ops->priv pointers of the copy to
refer to the table and then change the hookfn implementations to
just pass the 'priv' argument to the traverser.
After this patch, the xt_table pointers can already be removed
from struct net.
However, changes to struct net result in re-compile of the entire
network stack, so do the removal after arptables and ip6tables
have been converted as well.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
This changes how ip(6)table nat passes the ruleset/table to the
evaluation loop.
At the moment, it will fetch the table from struct net.
This change stores the table in the hook_ops 'priv' argument
instead.
This requires to duplicate the hook_ops for each netns, so
they can store the (per-net) xt_table structure.
The dupliated nat hook_ops get stored in net_generic data area.
They are free'd in the namespace exit path.
This is a pre-requisite to remove the xt_table/ruleset pointers
from struct net.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
No need for these.
There is only one caller, the xtables core, when the table is registered
for the first time with a particular network namespace.
After ->table_init() call, the table is linked into the tables[af] list,
so next call to that function will skip the ->table_init().
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
xtables stores the xt_table structs in the struct net. This isn't
needed anymore, the structures could be passed via the netfilter hook
'private' pointer to the hook functions, which would allow us to remove
those pointers from struct net.
As a first step, reduce the number of accesses to the
net->ipv4.ip6table_{raw,filter,...} pointers.
This allows the tables to get unregistered by name instead of having to
pass the raw address.
The xt_table structure cane looked up by name+address family instead.
This patch is useless as-is (the backends still have the raw pointer
address), but it lowers the bar to remove those.
It also allows to put the 'was table registered in the first place' check
into ip_tables.c rather than have it in each table sub module.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Its the same function as ipt_unregister_table_exit.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
When I changed defrag hooks to no longer get registered by default I
intentionally made it so that registration can only be un-done by unloading
the nf_defrag_ipv4/6 module.
In hindsight this was too conservative; there is no reason to keep defrag
on while there is no feature dependency anymore.
Moreover, this won't work if user isn't allowed to remove nf_defrag module.
This adds the disable() functions for both ipv4 and ipv6 and calls them
from conntrack, TPROXY and the xtables socket module.
ipvs isn't converted here, it will behave as before this patch and
will need module removal.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Alexei Starovoitov says:
====================
pull-request: bpf-next 2021-04-23
The following pull-request contains BPF updates for your *net-next* tree.
We've added 69 non-merge commits during the last 22 day(s) which contain
a total of 69 files changed, 3141 insertions(+), 866 deletions(-).
The main changes are:
1) Add BPF static linker support for extern resolution of global, from Andrii.
2) Refine retval for bpf_get_task_stack helper, from Dave.
3) Add a bpf_snprintf helper, from Florent.
4) A bunch of miscellaneous improvements from many developers.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Pablo Neira Ayuso says:
====================
Netfilter updates for net-next
The following patchset contains Netfilter updates for net-next:
1) Add vlan match and pop actions to the flowtable offload,
patches from wenxu.
2) Reduce size of the netns_ct structure, which itself is
embedded in struct net Make netns_ct a read-mostly structure.
Patches from Florian Westphal.
3) Add FLOW_OFFLOAD_XMIT_UNSPEC to skip dst check from garbage
collector path, as required by the tc CT action. From Roi Dayan.
4) VLAN offload fixes for nftables: Allow for matching on both s-vlan
and c-vlan selectors. Fix match of VLAN id due to incorrect
byteorder. Add a new routine to properly populate flow dissector
ethertypes.
5) Missing keys in ip{6}_route_me_harder() results in incorrect
routes. This includes an update for selftest infra. Patches
from Ido Schimmel.
6) Add counter hardware offload support through FLOW_CLS_STATS.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, a multi-part nexthop dump is restarted based on the number of
nexthops that have been dumped so far. This can result in a lot of
nexthops not being dumped when nexthops are simultaneously deleted:
# ip nexthop | wc -l
65536
# ip nexthop flush
Dump was interrupted and may be inconsistent.
Flushed 36040 nexthops
# ip nexthop | wc -l
29496
Instead, restart the dump based on the nexthop identifier (fixed number)
of the last successfully dumped nexthop:
# ip nexthop | wc -l
65536
# ip nexthop flush
Dump was interrupted and may be inconsistent.
Flushed 65536 nexthops
# ip nexthop | wc -l
0
Reported-by: Maksym Yaremchuk <maksymy@nvidia.com>
Tested-by: Maksym Yaremchuk <maksymy@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jianwen reported that IPv6 Interoperability tests are failing in an
IPsec case where one of the links between the IPsec peers has an MTU
of 1280. The peer generates a packet larger than this MTU, the router
replies with a "Packet too big" message indicating an MTU of 1280.
When the peer tries to send another large packet, xfrm_state_mtu
returns 1280 - ipsec_overhead, which causes ip6_setup_cork to fail
with EINVAL.
We can fix this by forcing xfrm_state_mtu to return IPV6_MIN_MTU when
IPv6 is used. After going through IPsec, the packet will then be
fragmented to obey the actual network's PMTU, just before leaving the
host.
Currently, TFC padding is capped to PMTU - overhead to avoid
fragementation: after padding and encapsulation, we still fit within
the PMTU. That behavior is preserved in this patch.
Fixes: 91657eafb6 ("xfrm: take net hdr len into account for esp payload size calculation")
Reported-by: Jianwen Ji <jiji@redhat.com>
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Netfilter tries to reroute mangled packets as a different route might
need to be used following the mangling. When this happens, netfilter
does not populate the IP protocol, the source port and the destination
port in the flow key. Therefore, FIB rules that match on these fields
are ignored and packets can be misrouted.
Solve this by dissecting the outer flow and populating the flow key
before rerouting the packet. Note that flow dissection only happens when
FIB rules that match on these fields are installed, so in the common
case there should not be a penalty.
Reported-by: Michal Soltys <msoltyspl@yandex.pl>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
- keep the ZC code, drop the code related to reinit
net/bridge/netfilter/ebtables.c
- fix build after move to net_generic
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Steffen Klassert says:
====================
pull request (net-next): ipsec-next 2021-04-14
Not much this time:
1) Simplification of some variable calculations in esp4 and esp6.
From Jiapeng Chong and Junlin Yang.
2) Fix a clang Wformat warning in esp6 and ah6.
From Arnd Bergmann.
Please pull or let me know if there are problems.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, tcp_allowed_congestion_control is global and writable;
writing to it in any net namespace will leak into all other net
namespaces.
tcp_available_congestion_control and tcp_allowed_congestion_control are
the only sysctls in ipv4_net_table (the per-netns sysctl table) with a
NULL data pointer; their handlers (proc_tcp_available_congestion_control
and proc_allowed_congestion_control) have no other way of referencing a
struct net. Thus, they operate globally.
Because ipv4_net_table does not use designated initializers, there is no
easy way to fix up this one "bad" table entry. However, the data pointer
updating logic shouldn't be applied to NULL pointers anyway, so we
instead force these entries to be read-only.
These sysctls used to exist in ipv4_table (init-net only), but they were
moved to the per-net ipv4_net_table, presumably without realizing that
tcp_allowed_congestion_control was writable and thus introduced a leak.
Because the intent of that commit was only to know (i.e. read) "which
congestion algorithms are available or allowed", this read-only solution
should be sufficient.
The logic added in recent commit
31c4d2f160eb: ("net: Ensure net namespace isolation of sysctls")
does not and cannot check for NULL data pointers, because
other table entries (e.g. /proc/sys/net/netfilter/nf_log/) have
.data=NULL but use other methods (.extra2) to access the struct net.
Fixes: 9cb8e048e5 ("net/ipv4/sysctl: show tcp_{allowed, available}_congestion_control in non-initial netns")
Signed-off-by: Jonathon Reinhart <jonathon.reinhart@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The current icmp_rcv function drops all unknown ICMP types, including
ICMP_EXT_ECHOREPLY (type 43). In order to parse Extended Echo Reply messages, we have
to pass these packets to the ping_rcv function, which does not do any
other filtering and passes the packet to the designated socket.
Pass incoming RFC 8335 ICMP Extended Echo Reply packets to the ping_rcv
handler instead of discarding the packet.
Signed-off-by: Andreas Roeseler <andreas.a.roeseler@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pablo Neira Ayuso says:
====================
Netfilter fixes for net
The following patchset contains Netfilter fixes for net:
1) Fix NAT IPv6 offload in the flowtable.
2) icmpv6 is printed as unknown in /proc/net/nf_conntrack.
3) Use div64_u64() in nft_limit, from Eric Dumazet.
4) Use pre_exit to unregister ebtables and arptables hooks,
from Florian Westphal.
5) Fix out-of-bound memset in x_tables compat match/target,
also from Florian.
6) Clone set elements expression to ensure proper initialization.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
xt_compat_match/target_from_user doesn't check that zeroing the area
to start of next rule won't write past end of allocated ruleset blob.
Remove this code and zero the entire blob beforehand.
Reported-by: syzbot+cfc0247ac173f597aaaa@syzkaller.appspotmail.com
Reported-by: Andy Nguyen <theflow@google.com>
Fixes: 9fa492cdc1 ("[NETFILTER]: x_tables: simplify compat API")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Same problem that also existed in iptables/ip(6)tables, when
arptable_filter is removed there is no longer a wait period before the
table/ruleset is free'd.
Unregister the hook in pre_exit, then remove the table in the exit
function.
This used to work correctly because the old nf_hook_unregister API
did unconditional synchronize_net.
The per-net hook unregister function uses call_rcu instead.
Fixes: b9e69e1273 ("netfilter: xtables: don't hook tables by default")
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Conflicts:
MAINTAINERS
- keep Chandrasekar
drivers/net/ethernet/mellanox/mlx5/core/en_main.c
- simple fix + trust the code re-added to param.c in -next is fine
include/linux/bpf.h
- trivial
include/linux/ethtool.h
- trivial, fix kdoc while at it
include/linux/skmsg.h
- move to relevant place in tcp.c, comment re-wrapped
net/core/skmsg.c
- add the sk = sk // sk = NULL around calls
net/tipc/crypto.c
- trivial
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Setting iftoken can fail for several different reasons but there
and there was no report to user as to the cause. Add netlink
extended errors to the processing of the request.
This requires adding additional argument through rtnl_af_ops
set_link_af callback.
Reported-by: Hongren Zheng <li@zenithal.me>
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pablo Neira Ayuso says:
====================
Netfilter updates for net-next
The following batch contains Netfilter/IPVS updates for your net-next tree:
1) Simplify log infrastructure modularity: Merge ipv4, ipv6, bridge,
netdev and ARP families to nf_log_syslog.c. Add module softdeps.
This fixes a rare deadlock condition that might occur when log
module autoload is required. From Florian Westphal.
2) Moves part of netfilter related pernet data from struct net to
net_generic() infrastructure. All of these users can be modules,
so if they are not loaded there is no need to waste space. Size
reduction is 7 cachelines on x86_64, also from Florian.
2) Update nftables audit support to report events once per table,
to get it aligned with iptables. From Richard Guy Briggs.
3) Check for stale routes from the flowtable garbage collector path.
This is fixing IPv6 which breaks due missing check for the dst_cookie.
4) Add a nfnl_fill_hdr() function to simplify netlink + nfnetlink
headers setup.
5) Remove documentation on several statified functions.
6) Remove printk on netns creation for the FTP IPVS tracker,
from Florian Westphal.
7) Remove unnecessary nf_tables_destroy_list_lock spinlock
initialization, from Yang Yingliang.
7) Remove a duplicated forward declaration in ipset,
from Wan Jiabing.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Userspace sends tcp connection (sock) destroy on network switch
i.e switching the default network of the device between multiple
networks(Cellular/Wifi/Ethernet).
Kernel though doesn't send reset for the connections in SYN-SENT state
and these connections continue to remain.
Even as per RFC 793, there is no hard rule to not send RST on ABORT in
this state.
Modify tcp_abort and tcp_disconnect behavior to send RST for connections
in syn-sent state to avoid lingering connections on network switch.
Signed-off-by: Manoj Basapathi <manojbm@codeaurora.org>
Signed-off-by: Sauvik Saha <ssaha@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
These comments in udp_bpf_update_proto() are copied from the
original TCP code and apparently do not apply to UDP. Just
remove them.
Reported-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20210403052715.13854-1-xiyou.wangcong@gmail.com
This allows followup patch to remove the defrag_ipv4 member from struct
net. It also allows to auto-remove the hooks later on by adding a
_disable() function. This will be done later in a follow patch series.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
The MPTCP reset option allows to carry a mptcp-specific error code that
provides more information on the nature of a connection reset.
Reset option data received gets stored in the subflow context so it can
be sent to userspace via the 'subflow closed' netlink event.
When a subflow is closed, the desired error code that should be sent to
the peer is also placed in the subflow context structure.
If a reset is sent before subflow establishment could complete, e.g. on
HMAC failure during an MP_JOIN operation, the mptcp skb extension is
used to store the reset information.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Alexei Starovoitov says:
====================
pull-request: bpf-next 2021-04-01
The following pull-request contains BPF updates for your *net-next* tree.
We've added 68 non-merge commits during the last 7 day(s) which contain
a total of 70 files changed, 2944 insertions(+), 1139 deletions(-).
The main changes are:
1) UDP support for sockmap, from Cong.
2) Verifier merge conflict resolution fix, from Daniel.
3) xsk selftests enhancements, from Maciej.
4) Unstable helpers aka kernel func calling, from Martin.
5) Batches ops for LPM map, from Pedro.
6) Fix race in bpf_get_local_storage, from Yonghong.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Support for UDP_GRO was added in the past but the implementation for
getsockopt was missed which did lead to an error when we tried to
retrieve the setting for UDP_GRO. This patch adds the missing switch
case for UDP_GRO
Fixes: e20cf8d3f1 ("udp: implement GRO for plain UDP sockets.")
Signed-off-by: Norman Maurer <norman_maurer@apple.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
We have to implement udp_bpf_recvmsg() to replace the ->recvmsg()
to retrieve skmsg from ingress_msg.
Signed-off-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20210331023237.41094-14-xiyou.wangcong@gmail.com
Although these two functions are only used by TCP, they are not
specific to TCP at all, both operate on skmsg and ingress_msg,
so fit in net/core/skmsg.c very well.
And we will need them for non-TCP, so rename and move them to
skmsg.c and export them to modules.
Signed-off-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20210331023237.41094-13-xiyou.wangcong@gmail.com
This is similar to tcp_read_sock(), except we do not need
to worry about connections, we just need to retrieve skb
from UDP receive queue.
Note, the return value of ->read_sock() is unused in
sk_psock_verdict_data_ready(), and UDP still does not
support splice() due to lack of ->splice_read(), so users
can not reach udp_read_sock() directly.
Signed-off-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20210331023237.41094-12-xiyou.wangcong@gmail.com
Currently sockmap calls into each protocol to update the struct
proto and replace it. This certainly won't work when the protocol
is implemented as a module, for example, AF_UNIX.
Introduce a new ops sk->sk_prot->psock_update_sk_prot(), so each
protocol can implement its own way to replace the struct proto.
This also helps get rid of symbol dependencies on CONFIG_INET.
Signed-off-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20210331023237.41094-11-xiyou.wangcong@gmail.com
Currently we rely on lock_sock to protect ingress_msg,
it is too big for this, we can actually just use a spinlock
to protect this list like protecting other skb queues.
__tcp_bpf_recvmsg() is still special because of peeking,
it still has to use lock_sock.
Signed-off-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Jakub Sitnicki <jakub@cloudflare.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20210331023237.41094-3-xiyou.wangcong@gmail.com
Steffen Klassert says:
====================
pull request (net): ipsec 2021-03-31
1) Fix ipv4 pmtu checks for xfrm anf vti interfaces.
From Eyal Birger.
2) There are situations where the socket passed to
xfrm_output_resume() is not the same as the one
attached to the skb. Use the socket passed to
xfrm_output_resume() to avoid lookup failures
when xfrm is used with VRFs.
From Evan Nimmo.
3) Make the xfrm_state_hash_generation sequence counter per
network namespace because but its write serialization
lock is also per network namespace. Write protection
is insufficient otherwise.
From Ahmed S. Darwish.
4) Fixup sctp featue flags when used with esp offload.
From Xin Long.
5) xfrm BEET mode doesn't support fragments for inner packets.
This is a limitation of the protocol, so no fix possible.
Warn at least to notify the user about that situation.
From Xin Long.
6) Fix NULL pointer dereference on policy lookup when
namespaces are uses in combination with esp offload.
7) Fix incorrect transformation on esp offload when
packets get segmented at layer 3.
8) Fix some user triggered usages of WARN_ONCE in
the xfrm compat layer.
From Dmitry Safonov.
Please pull or let me know if there are problems.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
sysctl_icmp_echo_enable_probe is an u8.
ipv4_net_table entry should use
.maxlen = sizeof(u8).
.proc_handler = proc_dou8vec_minmax,
Fixes: f1b8fa9fa5 ("net: add sysctl for enabling RFC 8335 PROBE messages")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Andreas Roeseler <andreas.a.roeseler@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently the UDP protocol delivers GSO_FRAGLIST packets to
the sockets without the expected segmentation.
This change addresses the issue introducing and maintaining
a couple of new fields to explicitly accept SKB_GSO_UDP_L4
or GSO_FRAGLIST packets. Additionally updates udp_unexpected_gso()
accordingly.
UDP sockets enabling UDP_GRO stil keep accept_udp_fraglist
zeroed.
v1 -> v2:
- use 2 bits instead of a whole GSO bitmask (Willem)
Fixes: 9fd1ff5d2a ("udp: Support UDP fraglist GRO/GSO.")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After the previous patch, the stack can do L4 UDP aggregation
on top of a UDP tunnel.
In such scenario, udp{4,6}_gro_complete will be called twice. This function
will enter its is_flist branch immediately, even though that is only
correct on the second call, as GSO_FRAGLIST is only relevant for the
inner packet.
Instead, we need to try first UDP tunnel-based aggregation, if the GRO
packet requires that.
This patch changes udp{4,6}_gro_complete to skip the frag list processing
when while encap_mark == 1, identifying processing of the outer tunnel
header.
Additionally, clears the field in udp_gro_complete() so that we can enter
the frag list path on the next round, for the inner header.
v1 -> v2:
- hopefully clarified the commit message
Reviewed-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If NETIF_F_GRO_FRAGLIST or NETIF_F_GRO_UDP_FWD are enabled, and there
are UDP tunnels available in the system, udp_gro_receive() could end-up
doing L4 aggregation (either SKB_GSO_UDP_L4 or SKB_GSO_FRAGLIST) at
the outer UDP tunnel level for packets effectively carrying and UDP
tunnel header.
That could cause inner protocol corruption. If e.g. the relevant
packets carry a vxlan header, different vxlan ids will be ignored/
aggregated to the same GSO packet. Inner headers will be ignored, too,
so that e.g. TCP over vxlan push packets will be held in the GRO
engine till the next flush, etc.
Just skip the SKB_GSO_UDP_L4 and SKB_GSO_FRAGLIST code path if the
current packet could land in a UDP tunnel, and let udp_gro_receive()
do GRO via udp_sk(sk)->gro_receive.
The check implemented in this patch is broader than what is strictly
needed, as the existing UDP tunnel could be e.g. configured on top of
a different device: we could end-up skipping GRO at-all for some packets.
Anyhow, that is a very thin corner case and covering it will add quite
a bit of complexity.
v1 -> v2:
- hopefully clarify the commit message
Fixes: 9fd1ff5d2a ("udp: Support UDP fraglist GRO/GSO.")
Fixes: 36707061d6 ("udp: allow forwarding of plain (non-fraglisted) UDP GRO packets")
Reviewed-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When UDP packets generated locally by a socket with UDP_SEGMENT
traverse the following path:
UDP tunnel(xmit) -> veth (segmentation) -> veth (gro) ->
UDP tunnel (rx) -> UDP socket (no UDP_GRO)
ip_summed will be set to CHECKSUM_PARTIAL at creation time and
such checksum mode will be preserved in the above path up to the
UDP tunnel receive code where we have:
__iptunnel_pull_header() -> skb_pull_rcsum() ->
skb_postpull_rcsum() -> __skb_postpull_rcsum()
The latter will convert the skb to CHECKSUM_NONE.
The UDP GSO packet will be later segmented as part of the rx socket
receive operation, and will present a CHECKSUM_NONE after segmentation.
Additionally the segmented packets UDP CB still refers to the original
GSO packet len. Overall that causes unexpected/wrong csum validation
errors later in the UDP receive path.
We could possibly address the issue with some additional checks and
csum mangling in the UDP tunnel code. Since the issue affects only
this UDP receive slow path, let's set a suitable csum status there.
Note that SKB_GSO_UDP_L4 or SKB_GSO_FRAGLIST packets lacking an UDP
encapsulation present a valid checksum when landing to udp_queue_rcv_skb(),
as the UDP checksum has been validated by the GRO engine.
v2 -> v3:
- even more verbose commit message and comments
v1 -> v2:
- restrict the csum update to the packets strictly needing them
- hopefully clarify the commit message and code comments
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
similar to previous change: nf_log_syslog now covers ARP logging
as well, the old nf_log_arp module is removed.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Netfilter has multiple log modules:
nf_log_arp
nf_log_bridge
nf_log_ipv4
nf_log_ipv6
nf_log_netdev
nfnetlink_log
nf_log_common
With the exception of nfnetlink_log (packet is sent to userspace for
dissection/logging), all of them log to the kernel ringbuffer.
This is the first part of a series to merge all modules except
nfnetlink_log into a single module: nf_log_syslog.
This allows to reduce code. After the series, only two log modules remain:
nfnetlink_log and nf_log_syslog. The latter provides the same
functionality as the old per-af log modules.
This renames nf_log_ipv4 to nf_log_syslog.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Modify the icmp_rcv function to check PROBE messages and call icmp_echo
if a PROBE request is detected.
Modify the existing icmp_echo function to respond ot both ping and PROBE
requests.
This was tested using a custom modification to the iputils package and
wireshark. It supports IPV4 probing by name, ifindex, and probing by
both IPV4 and IPV6 addresses. It currently does not support responding
to probes off the proxy node (see RFC 8335 Section 2).
The modification to the iputils package is still in development and can
be found here: https://github.com/Juniper-Clinic-2020/iputils.git. It
supports full sending functionality of PROBE requests, but currently
does not parse the response messages, which is why Wireshark is required
to verify the sent and recieved PROBE messages. The modification adds
the ``-e'' flag to the command which allows the user to specify the
interface identifier to query the probed host. An example usage would be
<./ping -4 -e 1 [destination]> to send a PROBE request of ifindex 1 to the
destination node.
Signed-off-by: Andreas Roeseler <andreas.a.roeseler@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Modify the ping_supported function to support PROBE message types. This
allows tools such as the ping command in the iputils package to be
modified to send PROBE requests through the existing framework for
sending ping requests.
Signed-off-by: Andreas Roeseler <andreas.a.roeseler@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Section 8 of RFC 8335 specifies potential security concerns of
responding to PROBE requests, and states that nodes that support PROBE
functionality MUST be able to enable/disable responses and that
responses MUST be disabled by default
Signed-off-by: Andreas Roeseler <andreas.a.roeseler@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
pahole currently only generates the btf_id for external function and
ftrace-able function. Some functions in the bpf_tcp_ca_kfunc_ids
are static (e.g. cubictcp_init). Thus, unless CONFIG_DYNAMIC_FTRACE
is set, btf_ids for those functions will not be generated and the
compilation fails during resolve_btfids.
This patch limits those functions to CONFIG_DYNAMIC_FTRACE. I will
address the pahole generation in a followup and then remove the
CONFIG_DYNAMIC_FTRACE limitation.
Fixes: e78aea8b21 ("bpf: tcp: Put some tcp cong functions in allowlist for bpf-tcp-cc")
Reported-by: Cong Wang <xiyou.wangcong@gmail.com>
Reported-by: Lorenz Bauer <lmb@cloudflare.com>
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20210329221357.834438-1-kafai@fb.com
tcp_min_tso_segs is now stored in u8, so max value is 255.
255 limit is enforced by proc_dou8vec_minmax().
We can therefore remove the gso_max_segs variable.
Fixes: 47996b489bdc ("tcp: convert elligible sysctls to u8")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 94579ac3f6 ("xfrm: Fix double ESP trailer insertion in IPsec
crypto offload.") added a XFRM_XMIT flag to avoid duplicate ESP trailer
insertion on HW offload. This flag is set on the secpath that is shared
amongst segments. This lead to a situation where some segments are
not transformed correctly when segmentation happens at layer 3.
Fix this by using private skb extensions for segmented and hw offloaded
ESP packets.
Fixes: 94579ac3f6 ("xfrm: Fix double ESP trailer insertion in IPsec crypto offload.")
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
There is a typo in the bbr function, s/even/event/.
This patch fixes it.
Fixes: e78aea8b21 ("bpf: tcp: Put some tcp cong functions in allowlist for bpf-tcp-cc")
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20210329003213.2274210-1-kafai@fb.com
After resilient next-hop groups have been added recently, there are two
types of multipath next-hop groups: the legacy "mpath", and the new
"resilient". Calling the legacy next-hop group type "mpath" is unfortunate,
because that describes the fact that a packet could be forwarded in one of
several paths, which is also true for the resilient next-hop groups.
Therefore, to make the naming clearer, rename various artifacts to reflect
the assumptions made. Therefore as of this patch:
- The flag for multipath groups is nh_grp_entry::is_multipath. This
includes the legacy and resilient groups, as well as any future group
types that behave as multipath groups.
Functions that assume this have "mpath" in the name.
- The flag for legacy multipath groups is nh_grp_entry::hash_threshold.
Functions that assume this have "hthr" in the name.
- The flag for resilient groups is nh_grp_entry::resilient.
Functions that assume this have "res" in the name.
Besides the above, struct nh_grp_entry::mpath was renamed to ::hthr as
well.
UAPI artifacts were obviously left intact.
Suggested-by: David Ahern <dsahern@gmail.com>
Signed-off-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
s/readibility/readability/
s/insufficent/insufficient/
Signed-off-by: Bhaskar Chowdhury <unixbhaskar@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, building the bpf-next source with the CONFIG_BPF_SYSCALL
enabled is causing a compilation error:
"net/ipv4/bpf_tcp_ca.c:209:28: error: expected identifier or '(' before
',' token"
Fix this by removing an unnecessary comma.
Fixes: e78aea8b21 ("bpf: tcp: Put some tcp cong functions in allowlist for bpf-tcp-cc")
Reported-by: syzbot+0b74d8ec3bf0cc4e4209@syzkaller.appspotmail.com
Signed-off-by: Atul Gopinathan <atulgopinathan@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20210328120515.113895-1-atulgopinathan@gmail.com
This patch puts some tcp cong helper functions, tcp_slow_start()
and tcp_cong_avoid_ai(), into the allowlist for the bpf-tcp-cc
program.
A few tcp cc implementation functions are also put into the
allowlist. A potential use case is the bpf-tcp-cc implementation
may only want to override a subset of a tcp_congestion_ops. For others,
the bpf-tcp-cc can directly call the kernel counter parts instead of
re-implementing (or copy-and-pasting) them to the bpf program.
They will only be available to the bpf-tcp-cc typed program.
The allowlist functions are not bounded to a fixed ABI contract.
When any of them has changed, the bpf-tcp-cc program has to be changed
like any in-tree/out-of-tree kernel tcp-cc implementations do also.
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20210325015201.1546345-1-kafai@fb.com
The cubic functions in tcp_cubic.c are using the bictcp prefix as
in tcp_bic.c. This patch gives it the proper name cubictcp
because the later patch will allow the bpf prog to directly
call the cubictcp implementation. Renaming them will avoid
the name collision when trying to find the intended
one to call during bpf prog load time.
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20210325015155.1545532-1-kafai@fb.com
Many tcp sysctls are either bools or small ints that can fit into u8.
Reducing space taken by sysctls can save few cache line misses
when sending/receiving data while cpu caches are empty,
for example after cpu idle period.
This is hard to measure with typical network performance tests,
but after this patch, struct netns_ipv4 has shrunk
by three cache lines.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
For these sysctls, their dedicated helpers have
to use proc_dou8vec_minmax().
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This sysctl uses ip_fwd_update_priority() helper,
so the conversion needs to change it.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
These sysctls that can fit in one byte instead of one int
are converted to save space and thus reduce cache line misses.
- icmp_echo_ignore_all, icmp_echo_ignore_broadcasts,
- icmp_ignore_bogus_error_responses, icmp_errors_use_inbound_ifaddr
- tcp_ecn, tcp_ecn_fallback
- ip_default_ttl, ip_no_pmtu_disc, ip_fwd_use_pmtu
- ip_nonlocal_bind, ip_autobind_reuse
- ip_dynaddr, ip_early_demux, raw_l3mdev_accept
- nexthop_compat_mode, fwmark_reflect
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Modify "accomodate" to "accommodate" in net/ipv4/esp4.c.
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Lu Wei <luwei32@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Alexei Starovoitov says:
====================
pull-request: bpf-next 2021-03-24
The following pull-request contains BPF updates for your *net-next* tree.
We've added 37 non-merge commits during the last 15 day(s) which contain
a total of 65 files changed, 3200 insertions(+), 738 deletions(-).
The main changes are:
1) Static linking of multiple BPF ELF files, from Andrii.
2) Move drop error path to devmap for XDP_REDIRECT, from Lorenzo.
3) Spelling fixes from various folks.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
In commit 73f156a6e8 ("inetpeer: get rid of ip_id_count")
I used a very small hash table that could be abused
by patient attackers to reveal sensitive information.
Switch to a dynamic sizing, depending on RAM size.
Typical big hosts will now use 128x more storage (2 MB)
to get a similar increase in security and reduction
of hash collisions.
As a bonus, use of alloc_large_system_hash() spreads
allocated memory among all NUMA nodes.
Fixes: 73f156a6e8 ("inetpeer: get rid of ip_id_count")
Reported-by: Amit Klein <aksecurity@gmail.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Willy Tarreau <w@1wt.eu>
Signed-off-by: David S. Miller <davem@davemloft.net>
ic_close_dev contains a generalization of the logic to not close a
network interface if it's the host port for a DSA switch. This logic is
disguised behind an iteration through the lowers of ic_dev in
ic_close_dev.
When no interface for ipconfig can be found, ic_dev is NULL, and
ic_close_dev:
- dereferences a NULL pointer when assigning selected_dev
- would attempt to search through the lower interfaces of a NULL
net_device pointer
So we should protect against that case.
The "lower_dev" iterator variable was shortened to "lower" in order to
keep the 80 character limit.
Fixes: f68cbaed67 ("net: ipconfig: avoid use-after-free in ic_close_devs")
Fixes: 46acf7bdbc ("Revert "net: ipv4: handle DSA enabled master network devices"")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Tested-by: Heiko Thiery <heiko.thiery@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Now in esp4/6_gso_segment(), before calling inner proto .gso_segment,
NETIF_F_CSUM_MASK bits are deleted, as HW won't be able to do the
csum for inner proto due to the packet encrypted already.
So the UDP/TCP packet has to do the checksum on its own .gso_segment.
But SCTP is using CRC checksum, and for that NETIF_F_SCTP_CRC should
be deleted to make SCTP do the csum in own .gso_segment as well.
In Xiumei's testing with SCTP over IPsec/veth, the packets are kept
dropping due to the wrong CRC checksum.
Reported-by: Xiumei Mu <xmu@redhat.com>
Fixes: 7862b4058b ("esp: Add gso handlers for esp4 and esp6")
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Pablo Neira Ayuso says:
====================
Netfilter fixes for net
1) Several patches to testore use of memory barriers instead of RCU to
ensure consistent access to ruleset, from Mark Tomlinson.
2) Fix dump of expectation via ctnetlink, from Florian Westphal.
3) GRE helper works for IPv6, from Ludovic Senecaux.
4) Set error on unsupported flowtable flags.
5) Use delayed instead of deferrable workqueue in the flowtable,
from Yinjun Zhang.
6) Fix spurious EEXIST in case of add-after-delete flowtable in
the same batch.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
The selftest failed to compile with clang-built bpf-next.
Adding LLVM=1 to your vmlinux and selftest build will use clang.
The error message is:
progs/test_sk_storage_tracing.c:38:18: error: use of undeclared identifier 'BPF_TCP_CLOSE'
if (newstate == BPF_TCP_CLOSE)
^
1 error generated.
make: *** [Makefile:423: /bpf-next/tools/testing/selftests/bpf/test_sk_storage_tracing.o] Error 1
The reason for the failure is that BPF_TCP_CLOSE, a value of
an anonymous enum defined in uapi bpf.h, is not defined in
vmlinux.h. gcc does not have this problem. Since vmlinux.h
is derived from BTF which is derived from vmlinux DWARF,
that means gcc-produced vmlinux DWARF has BPF_TCP_CLOSE
while llvm-produced vmlinux DWARF does not have.
BPF_TCP_CLOSE is referenced in net/ipv4/tcp.c as
BUILD_BUG_ON((int)BPF_TCP_CLOSE != (int)TCP_CLOSE);
The following test mimics the above BUILD_BUG_ON, preprocessed
with clang compiler, and shows gcc DWARF contains BPF_TCP_CLOSE while
llvm DWARF does not.
$ cat t.c
enum {
BPF_TCP_ESTABLISHED = 1,
BPF_TCP_CLOSE = 7,
};
enum {
TCP_ESTABLISHED = 1,
TCP_CLOSE = 7,
};
int test() {
do {
extern void __compiletime_assert_767(void) ;
if ((int)BPF_TCP_CLOSE != (int)TCP_CLOSE) __compiletime_assert_767();
} while (0);
return 0;
}
$ clang t.c -O2 -c -g && llvm-dwarfdump t.o | grep BPF_TCP_CLOSE
$ gcc t.c -O2 -c -g && llvm-dwarfdump t.o | grep BPF_TCP_CLOSE
DW_AT_name ("BPF_TCP_CLOSE")
Further checking clang code find clang actually tried to
evaluate condition at compile time. If it is definitely
true/false, it will perform optimization and the whole if condition
will be removed before generating IR/debuginfo.
This patch explicited add an expression after the
above mentioned BUILD_BUG_ON in net/ipv4/tcp.c like
(void)BPF_TCP_ESTABLISHED
to enable generation of debuginfo for the anonymous
enum which also includes BPF_TCP_CLOSE.
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20210317174132.589276-1-yhs@fb.com
proc_creat_seq() that directly take a struct seq_operations,
and deal with network namespaces in ->open.
Signed-off-by: Yejune Deng <yejune.deng@gmail.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently tcp_check_req can be called with obsolete req socket for which big
socket have been already created (because of CPU race or early demux
assigning req socket to multiple packets in gro batch).
Commit e0f9759f53 ("tcp: try to keep packet if SYN_RCV race
is lost") added retry in case when tcp_check_req is called for PSH|ACK packet.
But if client sends RST+ACK immediatly after connection being
established (it is performing healthcheck, for example) retry does not
occur. In that case tcp_check_req tries to close req socket,
leaving big socket active.
Fixes: e0f9759f53 ("tcp: try to keep packet if SYN_RCV race is lost")
Signed-off-by: Alexander Ovechkin <ovov@yandex-team.ru>
Reported-by: Oleg Senin <olegsenin@yandex-team.ru>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This reverts commit cc00bcaa58.
This (and the preceding) patch basically re-implemented the RCU
mechanisms of patch 784544739a. That patch was replaced because of the
performance problems that it created when replacing tables. Now, we have
the same issue: the call to synchronize_rcu() makes replacing tables
slower by as much as an order of magnitude.
Prior to using RCU a script calling "iptables" approx. 200 times was
taking 1.16s. With RCU this increased to 11.59s.
Revert these patches and fix the issue in a different way.
Signed-off-by: Mark Tomlinson <mark.tomlinson@alliedtelesis.co.nz>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
This reverts commit 443d6e86f8.
This (and the following) patch basically re-implemented the RCU
mechanisms of patch 784544739a. That patch was replaced because of the
performance problems that it created when replacing tables. Now, we have
the same issue: the call to synchronize_rcu() makes replacing tables
slower by as much as an order of magnitude.
Revert these patches and fix the issue in a different way.
Signed-off-by: Mark Tomlinson <mark.tomlinson@alliedtelesis.co.nz>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
All comment lines inside the comment block have been aligned.
Every line of comment starts with a * (uniformity in code).
Signed-off-by: Shubhankar Kuranagatti <shubhankarvk@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix the following coccicheck warnings:
./net/ipv4/esp4.c:757:16-18: WARNING !A || A && B is equivalent to !A || B.
Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
TSQ provides a nice way to avoid bufferbloat on individual socket,
including retransmit packets. We can get rid of the old
heuristic:
/* Do not sent more than we queued. 1/4 is reserved for possible
* copying overhead: fragmentation, tunneling, mangling etc.
*/
if (refcount_read(&sk->sk_wmem_alloc) >
min_t(u32, sk->sk_wmem_queued + (sk->sk_wmem_queued >> 2),
sk->sk_sndbuf))
return -EAGAIN;
This heuristic was giving false positives according to Jakub,
whenever TX completions are delayed above RTT. (Ack packets
are processed by TCP stack before clones are orphaned/freed)
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Jakub Kicinski <kuba@kernel.org>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jakub reported Data included in a Fastopen SYN that had to be
retransmit would have to wait for an RTO if TX completions are slow,
even with prior fix.
This is because tcp_rcv_fastopen_synack() does not use standard
rtx logic, meaning TSQ handler exits early in tcp_tsq_write()
because tp->lost_out == tp->retrans_out
Lets make tcp_rcv_fastopen_synack() use standard rtx logic,
by using tcp_mark_skb_lost() on the skb thats needs to be
sent again.
Not this raised a warning in tcp_fastretrans_alert() during my tests
since we consider the data not being aknowledged
by the receiver does not mean packet was lost on the network.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Jakub Kicinski <kuba@kernel.org>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jakub and Neil reported an increase of RTO timers whenever
TX completions are delayed a bit more (by increasing
NIC TX coalescing parameters)
Main issue is that TCP stack has a logic preventing a packet
being retransmit if the prior clone has not yet been
orphaned or freed.
This logic came with commit 1f3279ae0c ("tcp: avoid
retransmits of TCP packets hanging in host queues")
Thankfully, in the case skb_still_in_host_queue() detects
the initial clone is still in flight, it can use TSQ logic
that will eventually retry later, at the moment the clone
is freed or orphaned.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Neil Spring <ntspring@fb.com>
Reported-by: Jakub Kicinski <kuba@kernel.org>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Now that all the code is in place, stop rejecting requests to create
resilient next-hop groups.
Signed-off-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Nexthop replacements et.al. are notified through netlink, but if a delayed
work migrates buckets on the background, userspace will stay oblivious.
Notify these as RTM_NEWNEXTHOPBUCKET events.
Signed-off-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Allow getting (but not setting) individual buckets to inspect the next hop
mapped therein, idle time, and flags.
Signed-off-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a dump handler for resilient next hop buckets. When next-hop group ID
is given, it walks buckets of that group, otherwise it walks buckets of all
groups. It then dumps the buckets whose next hops match the given filtering
criteria.
Signed-off-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Implement the netlink messages that allow creation and dumping of resilient
nexthop groups.
Signed-off-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The kernel periodically checks the idle time of nexthop buckets to
determine if they are idle and can be re-populated with a new nexthop.
When the resilient nexthop group is offloaded to hardware, the kernel
will not see activity on nexthop buckets unless it is reported from
hardware.
Add a function that can be periodically called by device drivers to
report activity on nexthop buckets after querying it from the underlying
device.
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a function that can be called by device drivers to set "offload" or
"trap" indication on nexthop buckets following nexthop notifications and
other changes such as a neighbour becoming invalid.
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Implement the following notifications towards drivers:
- NEXTHOP_EVENT_REPLACE, when a resilient nexthop group is created.
- NEXTHOP_EVENT_BUCKET_REPLACE any time there is a change in assignment of
next hops to hash table buckets. That includes replacements, deletions,
and delayed upkeep cycles. Some bucket notifications can be vetoed by the
driver, to make it possible to propagate bucket busy-ness flags from the
HW back to the algorithm. Some are however forced, e.g. if a next hop is
deleted, all buckets that use this next hop simply must be migrated,
whether the HW wishes so or not.
- NEXTHOP_EVENT_RES_TABLE_PRE_REPLACE, before a resilient nexthop group is
replaced. Usually the driver will get the bucket notifications as well,
and could veto those. But in some cases, a bucket may not be migrated
immediately, but during delayed upkeep, and that is too late to roll the
transaction back. This notification allows the driver to take a look and
veto the new proposed group up front, before anything is committed.
Signed-off-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
At this moment, there is only one type of next-hop group: an mpath group,
which implements the hash-threshold algorithm.
To select a next hop, hash-threshold algorithm first assigns a range of
hashes to each next hop in the group, and then selects the next hop by
comparing the SKB hash with the individual ranges. When a next hop is
removed from the group, the ranges are recomputed, which leads to
reassignment of parts of hash space from one next hop to another. While
there will usually be some overlap between the previous and the new
distribution, some traffic flows change the next hop that they resolve to.
That causes problems e.g. as established TCP connections are reset, because
the traffic is forwarded to a server that is not familiar with the
connection.
Resilient hashing is a technique to address the above problem. Resilient
next-hop group has another layer of indirection between the group itself
and its constituent next hops: a hash table. The selection algorithm uses a
straightforward modulo operation to choose a hash bucket, and then reads
the next hop that this bucket contains, and forwards traffic there.
This indirection brings an important feature. In the hash-threshold
algorithm, the range of hashes associated with a next hop must be
continuous. With a hash table, mapping between the hash table buckets and
the individual next hops is arbitrary. Therefore when a next hop is deleted
the buckets that held it are simply reassigned to other next hops. When
weights of next hops in a group are altered, it may be possible to choose a
subset of buckets that are currently not used for forwarding traffic, and
use those to satisfy the new next-hop distribution demands, keeping the
"busy" buckets intact. This way, established flows are ideally kept being
forwarded to the same endpoints through the same paths as before the
next-hop group change.
In a nutshell, the algorithm works as follows. Each next hop has a number
of buckets that it wants to have, according to its weight and the number of
buckets in the hash table. In case of an event that might cause bucket
allocation change, the numbers for individual next hops are updated,
similarly to how ranges are updated for mpath group next hops. Following
that, a new "upkeep" algorithm runs, and for idle buckets that belong to a
next hop that is currently occupying more buckets than it wants (it is
"overweight"), it migrates the buckets to one of the next hops that has
fewer buckets than it wants (it is "underweight"). If, after this, there
are still underweight next hops, another upkeep run is scheduled to a
future time.
Chances are there are not enough "idle" buckets to satisfy the new demands.
The algorithm has knobs to select both what it means for a bucket to be
idle, and for whether and when to forcefully migrate buckets if there keeps
being an insufficient number of idle buckets.
There are three users of the resilient data structures.
- The forwarding code accesses them under RCU, and does not modify them
except for updating the time a selected bucket was last used.
- Netlink code, running under RTNL, which may modify the data.
- The delayed upkeep code, which may modify the data. This runs unlocked,
and mutual exclusion between the RTNL code and the delayed upkeep is
maintained by canceling the delayed work synchronously before the RTNL
code touches anything. Later it restarts the delayed work if necessary.
The RTNL code has to implement next-hop group replacement, next hop
removal, etc. For removal, the mpath code uses a neat trick of having a
backup next hop group structure, doing the necessary changes offline, and
then RCU-swapping them in. However, the hash tables for resilient hashing
are about an order of magnitude larger than the groups themselves (the size
might be e.g. 4K entries), and it was felt that keeping two of them is an
overkill. Both the primary next-hop group and the spare therefore use the
same resilient table, and writers are careful to keep all references valid
for the forwarding code. The hash table references next-hop group entries
from the next-hop group that is currently in the primary role (i.e. not
spare). During the transition from primary to spare, the table references a
mix of both the primary group and the spare. When a next hop is deleted,
the corresponding buckets are not set to NULL, but instead marked as empty,
so that the pointer is valid and can be used by the forwarding code. The
buckets are then migrated to a new next-hop group entry during upkeep. The
only times that the hash table is invalid is the very beginning and very
end of its lifetime. Between those points, it is always kept valid.
This patch introduces the core support code itself. It does not handle
notifications towards drivers, which are kept as if the group were an mpath
one. It does not handle netlink either. The only bit currently exposed to
user space is the new next-hop group type, and that is currently bounced.
There is therefore no way to actually access this code.
Signed-off-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
- RTM_NEWNEXTHOP et.al. that handle resilient groups will have a new nested
attribute, NHA_RES_GROUP, whose elements are attributes NHA_RES_GROUP_*.
- RTM_NEWNEXTHOPBUCKET et.al. is a suite of new messages that will
currently serve only for dumping of individual buckets of resilient next
hop groups. For nexthop group buckets, these messages will carry a nested
attribute NHA_RES_BUCKET, whose elements are attributes NHA_RES_BUCKET_*.
There are several reasons why a new suite of messages is created for
nexthop buckets instead of overloading the information on the existing
RTM_{NEW,DEL,GET}NEXTHOP messages.
First, a nexthop group can contain a large number of nexthop buckets (4k
is not unheard of). This imposes limits on the amount of information that
can be encoded for each nexthop bucket given a netlink message is limited
to 64k bytes.
Second, while RTM_NEWNEXTHOPBUCKET is only used for notifications at
this point, in the future it can be extended to provide user space with
control over nexthop buckets configuration.
- The new group type is NEXTHOP_GRP_TYPE_RES. Note that nexthop code is
adjusted to bounce groups with that type for now.
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
With the introduction of resilient nexthop groups, there will be two types
of multipath groups: the current hash-threshold "mpath" ones, and resilient
groups. Both are multipath, but to determine the fact, the system needs to
consider two flags. This might prove costly in the datapath. Therefore,
introduce a new flag, that should be set for next-hop groups that have more
than one nexthop, and should be considered multipath.
Signed-off-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>