One 32bit hole is following skc_refcnt, use it.
skc_incoming_cpu can also be an union for request_sock rcv_wnd.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Everything should now be ready to finally allow SYN
packets processing without holding listener lock.

Tested:

3.5 Mpps SYNFLOOD. Plenty of cpu cycles available.

Next bottleneck is the refcount taken on listener,
that could be avoided if we remove SLAB_DESTROY_BY_RCU
strict semantic for listeners, and use regular RCU.

    13.18%  [kernel]  [k] __inet_lookup_listener
     9.61%  [kernel]  [k] tcp_conn_request
     8.16%  [kernel]  [k] sha_transform
     5.30%  [kernel]  [k] inet_reqsk_alloc
     4.22%  [kernel]  [k] sock_put
     3.74%  [kernel]  [k] tcp_make_synack
     2.88%  [kernel]  [k] ipt_do_table
     2.56%  [kernel]  [k] memcpy_erms
     2.53%  [kernel]  [k] sock_wfree
     2.40%  [kernel]  [k] tcp_v4_rcv
     2.08%  [kernel]  [k] fib_table_lookup
     1.84%  [kernel]  [k] tcp_openreq_init_rwin
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

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>

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>

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>

tcp_v6_md5_do_lookup() now takes a const socket, even if
CONFIG_TCP_MD5SIG is not set.

Fixes: b83e3deb ("tcp: md5: constify tcp_md5_do_lookup() socket argument")
From: Eric Dumazet <edumazet@google.com>
Reported-by: Nkbuild test robot <fengguang.wu@intel.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>

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>

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>

Before changing dccp_v6_request_recv_sock() sock argument
to const, we need to get rid of security_sk_classify_flow(),
and it seems doable by reusing inet6_csk_route_req() helper.

We need to add a proto parameter to inet6_csk_route_req(),
not assume it is TCP.
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>

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>

When TCP new listener is done, these functions will be called
without socket lock being held. Make sure they don't change
anything.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Soon, listener socket spinlock will no longer be held,
add const arguments to tcp_v[46]_init_req() to make clear these
functions can not mess socket fields.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Neal suggested to move sk_txhash init into tcp_create_openreq_child(),
called both from IPv4 and IPv6.

This opportunity was missed in commit 58d607d3 ("tcp: provide
skb->hash to synack packets")
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NNeal Cardwell <ncardwell@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In commit b73c3d0e ("net: Save TX flow hash in sock and set in skbuf
on xmit"), Tom provided a l4 hash to most outgoing TCP packets.

We'd like to provide one as well for SYNACK packets, so that all packets
of a given flow share same txhash, to later enable bonding driver to
also use skb->hash to perform slave selection.

Note that a SYNACK retransmit shuffles the tx hash, as Tom did
in commit 265f94ff ("net: Recompute sk_txhash on negative routing
advice") for established sockets.

This has nice effect making TCP flows resilient to some kind of black
holes, even at connection establish phase.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Cc: Tom Herbert <tom@herbertland.com>
Cc: Mahesh Bandewar <maheshb@google.com>
Acked-by: NTom Herbert <tom@herbertland.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In commit b357a364 ("inet: fix possible panic in
reqsk_queue_unlink()"), I missed fact that tcp_check_req()
can return the listener socket in one case, and that we must
release the request socket refcount or we leak it.

Tested:

 Following packetdrill test template shows the issue

0     socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
+0    setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
+0    bind(3, ..., ...) = 0
+0    listen(3, 1) = 0

+0    < S 0:0(0) win 2920 <mss 1460,sackOK,nop,nop>
+0    > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK>
+.002 < . 1:1(0) ack 21 win 2920
+0    > R 21:21(0)

Fixes: b357a364 ("inet: fix possible panic in reqsk_queue_unlink()")
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch creates sk_set_txhash and eliminates protocol specific
inet_set_txhash and ip6_set_txhash. sk_set_txhash simply sets a
random number instead of performing flow dissection. sk_set_txash
is also allowed to be called multiple times for the same socket,
we'll need this when redoing the hash for negative routing advice.
Signed-off-by: NTom Herbert <tom@herbertland.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

inet_twsk_deschedule() calls are followed by inet_twsk_put().

Only particular case is in inet_twsk_purge() but there is no point
to defer the inet_twsk_put() after re-enabling BH.

Lets rename inet_twsk_deschedule() to inet_twsk_deschedule_put()
and move the inet_twsk_put() inside.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Remove dead code from tcp_v6_connect()
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

For same reasons than in commit 12e25e10 ("tcp: remove redundant
checks"), we can remove redundant checks done for timewait sockets.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

tcp_v4_rcv() checks the following before calling tcp_v4_do_rcv():

if (th->doff < sizeof(struct tcphdr) / 4)
    goto bad_packet;
if (!pskb_may_pull(skb, th->doff * 4))
    goto discard_it;

So following check in tcp_v4_do_rcv() is redundant
and "goto csum_err;" is wrong anyway.

if (skb->len < tcp_hdrlen(skb) || ...)
	goto csum_err;

A second check can be removed after no_tcp_socket label for same reason.

Same tests can be removed in tcp_v6_do_rcv()

Note : short tcp frames are not properly accounted in tcpInErrs MIB,
because pskb_may_pull() failure simply drops incoming skb, we might
fix this in a separate patch.
Signed-off-by: NEric Dumazet  <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Instead of doing the rt6->rt6i_node check whenever we need
to get the route's cookie.  Refactor it into rt6_get_cookie().
It is a prep work to handle FLOWI_FLAG_KNOWN_NH and also
percpu rt6_info later.
Signed-off-by: NMartin KaFai Lau <kafai@fb.com>
Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
Cc: Steffen Klassert <steffen.klassert@secunet.com>
Cc: Julian Anastasov <ja@ssi.bg>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch removes the assumptions that the returned rt is always
a RTF_CACHE entry with the rt6i_dst and rt6i_src containing the
destination and source address.  The dst and src can be recovered from
the calling site.

We may consider to rename (rt6i_dst, rt6i_src) to
(rt6i_key_dst, rt6i_key_src) later.
Signed-off-by: NMartin KaFai Lau <kafai@fb.com>
Reviewed-by: NHannes Frederic Sowa <hannes@stressinduktion.org>
Cc: Steffen Klassert <steffen.klassert@secunet.com>
Cc: Julian Anastasov <ja@ssi.bg>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch tracks the total number of inbound and outbound segments on a
TCP socket. One may use this number to have an idea on connection
quality when compared against the retransmissions.

RFC4898 named these : tcpEStatsPerfSegsIn and tcpEStatsPerfSegsOut

These are a 32bit field each and can be fetched both from TCP_INFO
getsockopt() if one has a handle on a TCP socket, or from inet_diag
netlink facility (iproute2/ss patch will follow)

Note that tp->segs_out was placed near tp->snd_nxt for good data
locality and minimal performance impact, while tp->segs_in was placed
near tp->bytes_received for the same reason.

Join work with Eric Dumazet.

Note that received SYN are accounted on the listener, but sent SYNACK
are not accounted.
Signed-off-by: NMarcelo Ricardo Leitner <mleitner@redhat.com>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

commit 1d13a96c ("ipv6: tcp: fix flowlabel value in ACK messages
send from TIME_WAIT") added the flow label in the last TCP packets.
Unfortunately, it was not casted properly.

This patch replace the buggy shift with be32_to_cpu/cpu_to_be32.

Fixes: 1d13a96c ("ipv6: tcp: fix flowlabel value in ACK messages")
Reported-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NFlorent Fourcot <florent.fourcot@enst-bretagne.fr>
Acked-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

[ 3897.923145] BUG: unable to handle kernel NULL pointer dereference at
 0000000000000080
[ 3897.931025] IP: [<ffffffffa9f27686>] reqsk_timer_handler+0x1a6/0x243

There is a race when reqsk_timer_handler() and tcp_check_req() call
inet_csk_reqsk_queue_unlink() on the same req at the same time.

Before commit fa76ce73 ("inet: get rid of central tcp/dccp listener
timer"), listener spinlock was held and race could not happen.

To solve this bug, we change reqsk_queue_unlink() to not assume req
must be found, and we return a status, to conditionally release a
refcount on the request sock.

This also means tcp_check_req() in non fastopen case might or not
consume req refcount, so tcp_v6_hnd_req() & tcp_v4_hnd_req() have
to properly handle this.

(Same remark for dccp_check_req() and its callers)

inet_csk_reqsk_queue_drop() is now too big to be inlined, as it is
called 4 times in tcp and 3 times in dccp.

Fixes: fa76ce73 ("inet: get rid of central tcp/dccp listener timer")
Signed-off-by: NEric Dumazet <edumazet@google.com>
Reported-by: NYuchung Cheng <ycheng@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Using a timer wheel for timewait sockets was nice ~15 years ago when
memory was expensive and machines had a single processor.

This does not scale, code is ugly and source of huge latencies
(Typically 30 ms have been seen, cpus spinning on death_lock spinlock.)

We can afford to use an extra 64 bytes per timewait sock and spread
timewait load to all cpus to have better behavior.

Tested:

On following test, /proc/sys/net/ipv4/tcp_tw_recycle is set to 1
on the target (lpaa24)

Before patch :

lpaa23:~# ./super_netperf 200 -H lpaa24 -t TCP_CC -l 60 -- -p0,0
419594

lpaa23:~# ./super_netperf 200 -H lpaa24 -t TCP_CC -l 60 -- -p0,0
437171

While test is running, we can observe 25 or even 33 ms latencies.

lpaa24:~# ping -c 1000 -i 0.02 -qn lpaa23
...
1000 packets transmitted, 1000 received, 0% packet loss, time 20601ms
rtt min/avg/max/mdev = 0.020/0.217/25.771/1.535 ms, pipe 2

lpaa24:~# ping -c 1000 -i 0.02 -qn lpaa23
...
1000 packets transmitted, 1000 received, 0% packet loss, time 20702ms
rtt min/avg/max/mdev = 0.019/0.183/33.761/1.441 ms, pipe 2

After patch :

About 90% increase of throughput :

lpaa23:~# ./super_netperf 200 -H lpaa24 -t TCP_CC -l 60 -- -p0,0
810442

lpaa23:~# ./super_netperf 200 -H lpaa24 -t TCP_CC -l 60 -- -p0,0
800992

And latencies are kept to minimal values during this load, even
if network utilization is 90% higher :

lpaa24:~# ping -c 1000 -i 0.02 -qn lpaa23
...
1000 packets transmitted, 1000 received, 0% packet loss, time 19991ms
rtt min/avg/max/mdev = 0.023/0.064/0.360/0.042 ms
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The ipv6 code uses a mixture of coding styles. In some instances check for NULL
pointer is done as x != NULL and sometimes as x. x is preferred according to
checkpatch and this patch makes the code consistent by adopting the latter
form.

No changes detected by objdiff.
Signed-off-by: NIan Morris <ipm@chirality.org.uk>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The ipv6 code uses a mixture of coding styles. In some instances check for NULL
pointer is done as x == NULL and sometimes as !x. !x is preferred according to
checkpatch and this patch makes the code consistent by adopting the latter
form.

No changes detected by objdiff.
Signed-off-by: NIan Morris <ipm@chirality.org.uk>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

tcp_v6_fill_cb() will be called twice if socket's state changes from
TCP_TIME_WAIT to TCP_LISTEN. That can result in control buffer data
corruption because in the second tcp_v6_fill_cb() call it's not copying
IP6CB(skb) anymore, but 'seq', 'end_seq', etc., so we can get weird and
unpredictable results. Performance loss of up to 1200% has been observed
in LTP/vxlan03 test.

This can be fixed by copying inet6_skb_parm to the beginning of 'cb'
only if xfrm6_policy_check() and tcp_v6_fill_cb() are going to be
called again.

Fixes: 2dc49d16 ("tcp6: don't move IP6CB before xfrm6_policy_check()")
Signed-off-by: NAlexey Kodanev <alexey.kodanev@oracle.com>
Acked-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

ss should display ipv4 mapped request sockets like this :

tcp    SYN-RECV   0      0  ::ffff:192.168.0.1:8080   ::ffff:192.0.2.1:35261

and not like this :

tcp    SYN-RECV   0      0  192.168.0.1:8080   192.0.2.1:35261

We should init ireq->ireq_family based on listener sk_family,
not the actual protocol carried by SYN packet.

This means we can set ireq_family in inet_reqsk_alloc()

Fixes: 3f66b083 ("inet: introduce ireq_family")
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

With request socks convergence, we no longer need
different lookup methods. A request socket can
use generic lookup function.

Add const qualifier to 2nd tcp_v[46]_md5_lookup() parameter.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Since request and established sockets now have same base,
there is no need to pass two pointers to tcp_v4_md5_hash_skb()
or tcp_v6_md5_hash_skb()

Also add a const qualifier to their struct tcp_md5sig_key argument.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

It is guaranteed that both tcp_v4_rcv() and tcp_v6_rcv()
run from rcu read locked sections :

ip_local_deliver_finish() and ip6_input_finish() both
use rcu_read_lock()

Also align tcp_v6_inbound_md5_hash() on tcp_v4_inbound_md5_hash()
by returning a boolean.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

On s390x, gcc 4.8 compiles this part of tcp_v6_early_demux()

        struct dst_entry *dst = sk->sk_rx_dst;

        if (dst)
                dst = dst_check(dst, inet6_sk(sk)->rx_dst_cookie);

to code reading sk->sk_rx_dst twice, once for the test and once for
the argument of ip6_dst_check() (dst_check() is inline). This allows
ip6_dst_check() to be called with null first argument, causing a crash.

Protect sk->sk_rx_dst access by READ_ONCE() both in IPv4 and IPv6
TCP early demux code.

Fixes: 41063e9d ("ipv4: Early TCP socket demux.")
Fixes: c7109986 ("ipv6: Early TCP socket demux")
Signed-off-by: NMichal Kubecek <mkubecek@suse.cz>
Acked-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>