If a listen backlog is very big (to avoid syncookies), then
the listener sk->sk_wmem_alloc is the main source of false
sharing, as we need to touch it twice per SYNACK re-transmit
and TX completion.

(One SYN packet takes listener lock once, but up to 6 SYNACK
are generated)

By attaching the skb to the request socket, we remove this
source of contention.

Tested:

 listen(fd, 10485760); // single listener (no SO_REUSEPORT)
 16 RX/TX queue NIC
 Sustain a SYNFLOOD attack of ~320,000 SYN per second,
 Sending ~1,400,000 SYNACK per second.
 Perf profiles now show listener spinlock being next bottleneck.

    20.29%  [kernel]  [k] queued_spin_lock_slowpath
    10.06%  [kernel]  [k] __inet_lookup_established
     5.12%  [kernel]  [k] reqsk_timer_handler
     3.22%  [kernel]  [k] get_next_timer_interrupt
     3.00%  [kernel]  [k] tcp_make_synack
     2.77%  [kernel]  [k] ipt_do_table
     2.70%  [kernel]  [k] run_timer_softirq
     2.50%  [kernel]  [k] ip_finish_output
     2.04%  [kernel]  [k] cascade
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In this patch, we insert request sockets into TCP/DCCP
regular ehash table (where ESTABLISHED and TIMEWAIT sockets
are) instead of using the per listener hash table.

ACK packets find SYN_RECV pseudo sockets without having
to find and lock the listener.

In nominal conditions, this halves pressure on listener lock.

Note that this will allow for SO_REUSEPORT refinements,
so that we can select a listener using cpu/numa affinities instead
of the prior 'consistent hash', since only SYN packets will
apply this selection logic.

We will shrink listen_sock in the following patch to ease
code review.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Cc: Ying Cai <ycai@google.com>
Cc: Willem de Bruijn <willemb@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This is no longer used.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When request sockets are no longer in a per listener hash table
but on regular TCP ehash, we need to access listener uid
through req->rsk_listener

get_openreq6() also gets a const for its request socket argument.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Once listener is lockless, its sk_state can change anytime.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We'll soon have to call tcp_v[46]_inbound_md5_hash() twice.
Also add const attribute to the socket, as it might be the
unlocked listener for SYN packets.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This fixes a typo : We want to store the NAPI id on child socket.
Presumably nobody really uses busy polling, on short lived flows.

Fixes: 3d97379a ("tcp: move sk_mark_napi_id() at the right place")
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

long term plan is to remove struct listen_sock when its hash
table is no longer there.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

qlen_inc & young_inc were protected by listener lock,
while qlen_dec & young_dec were atomic fields.

Everything needs to be atomic for upcoming lockless listener.

Also move qlen/young in request_sock_queue as we'll get rid
of struct listen_sock eventually.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

struct request_sock_queue fields are currently protected
by the listener 'lock' (not a real spinlock)

We need to add a private spinlock instead, so that softirq handlers
creating children do not have to worry with backlog notion
that the listener 'lock' carries.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The fib_table_lookup tracepoint found 2 places where the flowi4_flags is
not initialized.
Signed-off-by: NDavid Ahern <dsa@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Replace calls to vrf_dev_get_rth with l3mdev_get_rtable.
The check on the flow flags is handled in the l3mdev operation.
Signed-off-by: NDavid Ahern <dsa@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Replace calls to vrf_dev_table and friends with l3mdev_fib_table
and kin.
Signed-off-by: NDavid Ahern <dsa@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Replace calls to vrf_master_ifindex_rcu and vrf_master_ifindex with either
l3mdev_master_ifindex_rcu or l3mdev_master_ifindex.

The pattern:
    oif = vrf_master_ifindex(dev) ? : dev->ifindex;
is replaced with
    oif = l3mdev_fib_oif(dev);

And remove the now unused vrf macros.
Signed-off-by: NDavid Ahern <dsa@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Rename IFF_VRF_MASTER to IFF_L3MDEV_MASTER and update the name of the
netif_is_vrf and netif_index_is_vrf macros.
Signed-off-by: NDavid Ahern <dsa@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

While auditing TCP stack for upcoming 'lockless' listener changes,
I found I had to change fastopen_init_queue() to properly init the object
before publishing it.

Otherwise an other cpu could try to lock the spinlock before it gets
properly initialized.

Instead of adding appropriate barriers, just remove dynamic memory
allocations :
- Structure is 28 bytes on 64bit arches. Using additional 8 bytes
  for holding a pointer seems overkill.
- Two listeners can share same cache line and performance would suffer.

If we really want to save few bytes, we would instead dynamically allocate
whole struct request_sock_queue in the future.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

tcp_syn_flood_action() will soon be called with unlocked socket.
In order to avoid SYN flood warning being emitted multiple times,
use xchg().
Extend max_qlen_log and synflood_warned fields in struct listen_sock
to u32
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

These functions do not change the listener socket.
Goal is to make sure tcp_conn_request() is not messing with
listener in a racy way.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Some common IPv4/IPv6 code can be factorized.
Also constify cookie_init_sequence() socket argument.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We'll soon no longer hold listener socket lock, these
functions do not modify the socket in any way.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This method does not touch the listener socket.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

socket is not touched, make it const.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The socket points to the (shared) listener.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Factorize code to get tcp header from skb. It makes no sense
to duplicate code in callers.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Once we realize tcp_rcv_synsent_state_process() does not use
its 'len' argument and we get rid of it, then it becomes clear
this argument is no longer used in tcp_rcv_state_process()
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

None of these functions need to change the socket, make it
const.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

err is initialized to -EINVAL when it is declared. It is not reset until
fib_lookup which is well after the 3 users of the martian_source jump. So
resetting err to -EINVAL at martian_source label is not needed.

Removing that line obviates the need for the martian_source_keep_err label
so delete it.
Signed-off-by: NDavid Ahern <dsa@cumulusnetworks.com>
Signed-off-by: NAlexander Duyck <aduyck@mirantis.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch just swaps the ordering of one of the conditional tests in
ip_route_input_mc.  Specifically it swaps the testing for the source
address to see if it is loopback, and the test to see if we allow a
loopback source address.

The reason for swapping these two tests is because it is much faster to
test if an address is loopback than it is to dereference several pointers
to get at the net structure to see if the use of loopback is allowed.
Signed-off-by: NAlexander Duyck <aduyck@mirantis.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch updates ip_check_mc_rcu so that protocol is passed as a u8
instead of a u16.

The motivation is just to avoid any unneeded type transitions since some
systems will require an instruction to zero extend a u8 field to a u16.
Also it makes it a bit more readable as to the fact that protocol is a u8
so there are no byte ordering changes needed to pass it.
Signed-off-by: NAlexander Duyck <aduyck@mirantis.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Don't make ip_route_me_harder guess which network namespace
it is routing in, pass the network namespace in.
Signed-off-by: N"Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

ip6t_SYNPROXY already does this and this is needed so that we have a
struct net that can be passed down into ip_route_me_harder, so
that ip_route_me_harder can stop guessing it's context.

Along the way pass snet into synproxy_send_client_synack as this
is the only caller of synprox_send_tcp that is not passed snet
already.
Signed-off-by: N"Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

The network namespace is needed when routing a packet.
Stop making nf_afinfo.reroute guess which network namespace
is the proper namespace to route the packet in.
Signed-off-by: N"Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

This is needed so struct net can be pushed down into
ip_route_me_harder.
Signed-off-by: N"Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

Application limited streams such as thin streams, that transmit small
amounts of payload in relatively few packets per RTT, can be prevented
from growing the CWND when in congestion avoidance. This leads to
increased sojourn times for data segments in streams that often transmit
time-dependent data.

Currently, a connection is considered CWND limited only after having
successfully transmitted at least one packet with new data, while at the
same time failing to transmit some unsent data from the output queue
because the CWND is full. Applications that produce small amounts of
data may be left in a state where it is never considered to be CWND
limited, because all unsent data is successfully transmitted each time
an incoming ACK opens up for more data to be transmitted in the send
window.

Fix by always testing whether the CWND is fully used after successful
packet transmissions, such that a connection is considered CWND limited
whenever the CWND has been filled. This is the correct behavior as
specified in RFC2861 (section 3.1).

Cc: Andreas Petlund <apetlund@simula.no>
Cc: Carsten Griwodz <griff@simula.no>
Cc: Jonas Markussen <jonassm@ifi.uio.no>
Cc: Kenneth Klette Jonassen <kennetkl@ifi.uio.no>
Cc: Mads Johannessen <madsjoh@ifi.uio.no>
Signed-off-by: NBendik Rønning Opstad <bro.devel+kernel@gmail.com>
Acked-by: NEric Dumazet <edumazet@google.com>
Tested-by: NEric Dumazet <edumazet@google.com>
Acked-by: NNeal Cardwell <ncardwell@google.com>
Tested-by: NNeal Cardwell <ncardwell@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We found that a TCP Fast Open passive connection was vulnerable
to reorders, as the exchange might look like

[1] C -> S S <FO ...> <request>
[2] S -> C S. ack request <options>
[3] S -> C . <answer>

packets [2] and [3] can be generated at almost the same time.

If C receives the 3rd packet before the 2nd, it will drop it as
the socket is in SYN_SENT state and expects a SYNACK.

S will have to retransmit the answer.

Current OOO avoidance in linux is defeated because SYNACK
packets are attached to the LISTEN socket, while DATA packets
are attached to the children. They might be sent by different cpus,
and different TX queues might be selected.

It turns out that for TFO, we created a child, which is a
full blown socket in TCP_SYN_RECV state, and we simply can attach
the SYNACK packet to this socket.

This means that at the time tcp_sendmsg() pushes DATA packet,
skb->ooo_okay will be set iff the SYNACK packet had been sent
and TX completed.

This removes the reorder source at the host level.

We also removed the export of tcp_try_fastopen(), as it is no
longer called from IPv6.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NYuchung Cheng <ycheng@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Andrey reported a panic:

[ 7249.865507] BUG: unable to handle kernel pointer dereference at 000000b4
[ 7249.865559] IP: [<c16afeca>] icmp_route_lookup+0xaa/0x320
[ 7249.865598] *pdpt = 0000000030f7f001 *pde = 0000000000000000
[ 7249.865637] Oops: 0000 [#1]
...
[ 7249.866811] CPU: 0 PID: 0 Comm: swapper/0 Not tainted
4.3.0-999-generic #201509220155
[ 7249.866876] Hardware name: MSI MS-7250/MS-7250, BIOS 080014  08/02/2006
[ 7249.866916] task: c1a5ab00 ti: c1a52000 task.ti: c1a52000
[ 7249.866949] EIP: 0060:[<c16afeca>] EFLAGS: 00210246 CPU: 0
[ 7249.866981] EIP is at icmp_route_lookup+0xaa/0x320
[ 7249.867012] EAX: 00000000 EBX: f483ba48 ECX: 00000000 EDX: f2e18a00
[ 7249.867045] ESI: 000000c0 EDI: f483ba70 EBP: f483b9ec ESP: f483b974
[ 7249.867077]  DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068
[ 7249.867108] CR0: 8005003b CR2: 000000b4 CR3: 36ee07c0 CR4: 000006f0
[ 7249.867141] Stack:
[ 7249.867165]  320310ee 00000000 00000042 320310ee 00000000 c1aeca00
f3920240 f0c69180
[ 7249.867268]  f483ba04 f855058b a89b66cd f483ba44 f8962f4b 00000000
e659266c f483ba54
[ 7249.867361]  8004753c f483ba5c f8962f4b f2031140 000003c1 ffbd8fa0
c16b0e00 00000064
[ 7249.867448] Call Trace:
[ 7249.867494]  [<f855058b>] ? e1000_xmit_frame+0x87b/0xdc0 [e1000e]
[ 7249.867534]  [<f8962f4b>] ? tcp_in_window+0xeb/0xb10 [nf_conntrack]
[ 7249.867576]  [<f8962f4b>] ? tcp_in_window+0xeb/0xb10 [nf_conntrack]
[ 7249.867615]  [<c16b0e00>] ? icmp_send+0xa0/0x380
[ 7249.867648]  [<c16b102f>] icmp_send+0x2cf/0x380
[ 7249.867681]  [<f89c8126>] nf_send_unreach+0xa6/0xc0 [nf_reject_ipv4]
[ 7249.867714]  [<f89cd0da>] reject_tg+0x7a/0x9f [ipt_REJECT]
[ 7249.867746]  [<f88c29a7>] ipt_do_table+0x317/0x70c [ip_tables]
[ 7249.867780]  [<f895e0a6>] ? __nf_conntrack_find_get+0x166/0x3b0
[nf_conntrack]
[ 7249.867838]  [<f895eea8>] ? nf_conntrack_in+0x398/0x600 [nf_conntrack]
[ 7249.867889]  [<f84c0035>] iptable_filter_hook+0x35/0x80 [iptable_filter]
[ 7249.867933]  [<c16776a1>] nf_iterate+0x71/0x80
[ 7249.867970]  [<c1677715>] nf_hook_slow+0x65/0xc0
[ 7249.868002]  [<c1681811>] __ip_local_out_sk+0xc1/0xd0
[ 7249.868034]  [<c1680f30>] ? ip_forward_options+0x1a0/0x1a0
[ 7249.868066]  [<c1681836>] ip_local_out_sk+0x16/0x30
[ 7249.868097]  [<c1684054>] ip_send_skb+0x14/0x80
[ 7249.868129]  [<c16840f4>] ip_push_pending_frames+0x34/0x40
[ 7249.868163]  [<c16844a2>] ip_send_unicast_reply+0x282/0x310
[ 7249.868196]  [<c16a0863>] tcp_v4_send_reset+0x1b3/0x380
[ 7249.868227]  [<c16a1b63>] tcp_v4_rcv+0x323/0x990
[ 7249.868257]  [<c16776a1>] ? nf_iterate+0x71/0x80
[ 7249.868289]  [<c167dc2b>] ip_local_deliver_finish+0x8b/0x230
[ 7249.868322]  [<c167df4c>] ip_local_deliver+0x4c/0xa0
[ 7249.868353]  [<c167dba0>] ? ip_rcv_finish+0x390/0x390
[ 7249.868384]  [<c167d88c>] ip_rcv_finish+0x7c/0x390
[ 7249.868415]  [<c167e280>] ip_rcv+0x2e0/0x420
...

Prior to the VRF change the oif was not set in the flow struct, so the
VRF support should really have only added the vrf_master_ifindex lookup.

Fixes: 613d09b3 ("net: Use VRF device index for lookups on TX")
Cc: Andrey Melnikov <temnota.am@gmail.com>
Signed-off-by: NDavid Ahern <dsa@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

SYNACK packets are sent on behalf on unlocked listeners
or fastopen sockets. Mark socket as const to catch future changes
that might break the assumption.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This is done to make sure we do not change listener socket
while sending SYNACK packets while socket lock is not held.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This documents fact that listener lock might not be held
at the time SYNACK are sent.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

listener socket is not locked when tcp_make_synack() is called.

We better make sure no field is written.

There is one exception : Since SYNACK packets are attached to the listener
at this moment (or SYN_RECV child in case of Fast Open),
sock_wmalloc() needs to update sk->sk_wmem_alloc, but this is done using
atomic operations so this is safe.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>