This header file only contains the platform data structure definition,
so move it to the include/linux/platform_data/ directory.
Signed-off-by: NVivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

if_nlmsg_size() overestimates the minimum allocation size of netlink
dump request (when called from rtnl_calcit()) or the size of the
message (when called from rtnl_getlink()). This is because
ext_filter_mask is not supported by rtnl_link_get_af_size() and
rtnl_link_get_size().

The over-estimation is significant when at least one netdev has many
VLANs configured (8 bytes for each configured VLAN).

This patch-set "rightsizes" the protocol specific attribute size
calculation by propagating ext_filter_mask to rtnl_link_get_af_size()
and adding this a argument to get_link_af_size op in rtnl_af_ops.

Bridge module already used filtering aware sizing for notifications.
br_get_link_af_size_filtered() is consistent with the modified
get_link_af_size op so it replaces br_get_link_af_size() in br_af_ops.
br_get_link_af_size() becomes unused and thus removed.
Signed-off-by: NRonen Arad <ronen.arad@intel.com>
Acked-by: NSridhar Samudrala <sridhar.samudrala@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch implements the second half of RACK that uses the the most
recent transmit time among all delivered packets to detect losses.

tcp_rack_mark_lost() is called upon receiving a dubious ACK.
It then checks if an not-yet-sacked packet was sent at least
"reo_wnd" prior to the sent time of the most recently delivered.
If so the packet is deemed lost.

The "reo_wnd" reordering window starts with 1msec for fast loss
detection and changes to min-RTT/4 when reordering is observed.
We found 1msec accommodates well on tiny degree of reordering
(<3 pkts) on faster links. We use min-RTT instead of SRTT because
reordering is more of a path property but SRTT can be inflated by
self-inflicated congestion. The factor of 4 is borrowed from the
delayed early retransmit and seems to work reasonably well.

Since RACK is still experimental, it is now used as a supplemental
loss detection on top of existing algorithms. It is only effective
after the fast recovery starts or after the timeout occurs. The
fast recovery is still triggered by FACK and/or dupack threshold
instead of RACK.

We introduce a new sysctl net.ipv4.tcp_recovery for future
experiments of loss recoveries. For now RACK can be disabled by
setting it to 0.
Signed-off-by: NYuchung Cheng <ycheng@google.com>
Signed-off-by: NNeal Cardwell <ncardwell@google.com>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch is the first half of the RACK loss recovery.

RACK loss recovery uses the notion of time instead
of packet sequence (FACK) or counts (dupthresh). It's inspired by the
previous FACK heuristic in tcp_mark_lost_retrans(): when a limited
transmit (new data packet) is sacked, then current retransmitted
sequence below the newly sacked sequence must been lost,
since at least one round trip time has elapsed.

But it has several limitations:
1) can't detect tail drops since it depends on limited transmit
2) is disabled upon reordering (assumes no reordering)
3) only enabled in fast recovery ut not timeout recovery

RACK (Recently ACK) addresses these limitations with the notion
of time instead: a packet P1 is lost if a later packet P2 is s/acked,
as at least one round trip has passed.

Since RACK cares about the time sequence instead of the data sequence
of packets, it can detect tail drops when later retransmission is
s/acked while FACK or dupthresh can't. For reordering RACK uses a
dynamically adjusted reordering window ("reo_wnd") to reduce false
positives on ever (small) degree of reordering.

This patch implements tcp_advanced_rack() which tracks the
most recent transmission time among the packets that have been
delivered (ACKed or SACKed) in tp->rack.mstamp. This timestamp
is the key to determine which packet has been lost.

Consider an example that the sender sends six packets:
T1: P1 (lost)
T2: P2
T3: P3
T4: P4
T100: sack of P2. rack.mstamp = T2
T101: retransmit P1
T102: sack of P2,P3,P4. rack.mstamp = T4
T205: ACK of P4 since the hole is repaired. rack.mstamp = T101

We need to be careful about spurious retransmission because it may
falsely advance tp->rack.mstamp by an RTT or an RTO, causing RACK
to falsely mark all packets lost, just like a spurious timeout.

We identify spurious retransmission by the ACK's TS echo value.
If TS option is not applicable but the retransmission is acknowledged
less than min-RTT ago, it is likely to be spurious. We refrain from
using the transmission time of these spurious retransmissions.

The second half is implemented in the next patch that marks packet
lost using RACK timestamp.
Signed-off-by: NYuchung Cheng <ycheng@google.com>
Signed-off-by: NNeal Cardwell <ncardwell@google.com>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

a helper to prepare the first main RACK patch.
Signed-off-by: NYuchung Cheng <ycheng@google.com>
Signed-off-by: NNeal Cardwell <ncardwell@google.com>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Remove the existing lost retransmit detection because RACK subsumes
it completely. This also stops the overloading the ack_seq field of
the skb control block.
Signed-off-by: NYuchung Cheng <ycheng@google.com>
Signed-off-by: NNeal Cardwell <ncardwell@google.com>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Kathleen Nichols' algorithm for tracking the minimum RTT of a
data stream over some measurement window. It uses constant space
and constant time per update. Yet it almost always delivers
the same minimum as an implementation that has to keep all
the data in the window. The measurement window is tunable via
sysctl.net.ipv4.tcp_min_rtt_wlen with a default value of 5 minutes.

The algorithm keeps track of the best, 2nd best & 3rd best min
values, maintaining an invariant that the measurement time of
the n'th best >= n-1'th best. It also makes sure that the three
values are widely separated in the time window since that bounds
the worse case error when that data is monotonically increasing
over the window.

Upon getting a new min, we can forget everything earlier because
it has no value - the new min is less than everything else in the
window by definition and it's the most recent. So we restart fresh
on every new min and overwrites the 2nd & 3rd choices. The same
property holds for the 2nd & 3rd best.

Therefore we have to maintain two invariants to maximize the
information in the samples, one on values (1st.v <= 2nd.v <=
3rd.v) and the other on times (now-win <=1st.t <= 2nd.t <= 3rd.t <=
now). These invariants determine the structure of the code

The RTT input to the windowed filter is the minimum RTT measured
from ACK or SACK, or as the last resort from TCP timestamps.

The accessor tcp_min_rtt() returns the minimum RTT seen in the
window. ~0U indicates it is not available. The minimum is 1usec
even if the true RTT is below that.
Signed-off-by: NYuchung Cheng <ycheng@google.com>
Signed-off-by: NNeal Cardwell <ncardwell@google.com>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add missing rule to export mpls iptunnel header needed by iproute2
Signed-off-by: NStephen Hemminger <stephen@networkplumber.org>
Acked-by: NRoopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

At the time of commit fff32699 ("tcp: reflect SYN queue_mapping into
SYNACK packets") we had little ways to cope with SYN floods.

We no longer need to reflect incoming skb queue mappings, and instead
can pick a TX queue based on cpu cooking the SYNACK, with normal XPS
affinities.

Note that all SYNACK retransmits were picking TX queue 0, this no longer
is a win given that SYNACK rtx are now distributed on all cpus.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Greg reported crashes hitting the following check in __sk_backlog_rcv()

	BUG_ON(!sock_flag(sk, SOCK_MEMALLOC));

The pfmemalloc bit is currently checked in sk_filter().

This works correctly for TCP, because sk_filter() is ran in
tcp_v[46]_rcv() before hitting the prequeue or backlog checks.

For UDP or other protocols, this does not work, because the sk_filter()
is ran from sock_queue_rcv_skb(), which might be called _after_ backlog
queuing if socket is owned by user by the time packet is processed by
softirq handler.

Fixes: b4b9e355 ("netvm: set PF_MEMALLOC as appropriate during SKB processing")
Signed-off-by: NEric Dumazet <edumazet@google.com>
Reported-by: NGreg Thelen <gthelen@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

A recent change to the dst_output handling caused a new warning
when the call to NF_HOOK() is the only used of a local variable
passed as 'dev', and CONFIG_NETFILTER is disabled:

net/ipv6/ip6_output.c: In function 'ip6_output':
net/ipv6/ip6_output.c:135:21: warning: unused variable 'dev' [-Wunused-variable]

The reason for this is that the NF_HOOK macro in this case does
not reference the variable at all, and the call to dev_net(dev)
got removed from the ip6_output function. To avoid that warning now
and in the future, this changes the macro into an equivalent
inline function, which tells the compiler that the variable is
passed correctly but still unused.

The dn_forward function apparently had the same problem in
the past and added a local workaround that no longer works
with the inline function. In order to avoid a regression, we
have to also remove the #ifdef from decnet in the same patch.

Fixes: ede2059d ("dst: Pass net into dst->output")
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

We don't care if module is being unloaded anymore since hook unregister
handling will destroy queue entries using that hook.
Signed-off-by: NFlorian Westphal <fw@strlen.de>
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

since commit 8405a8ff ("netfilter: nf_qeueue: Drop queue entries on
nf_unregister_hook") all pending queued entries are discarded.

So we can simply remove all of the owner handling -- when module is
removed it also needs to unregister all its hooks.
Signed-off-by: NFlorian Westphal <fw@strlen.de>
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

This newly introduced netdevice notifier is called before actual change
upper happens. That provides a possibility for notifier handlers to
know upper change will happen and react to it, including possibility to
forbid the change. That is valuable for drivers which can check if the
upper device linkage is supported and forbid that in case it is not.
Signed-off-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Under stress, a close() on a listener can trigger the
WARN_ON(sk->sk_ack_backlog) in inet_csk_listen_stop()

We need to test if listener is still active before queueing
a child in inet_csk_reqsk_queue_add()

Create a common inet_child_forget() helper, and use it
from inet_csk_reqsk_queue_add() and inet_csk_listen_stop()
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Let's reduce the confusion about inet_csk_reqsk_queue_drop() :
In many cases we also need to release reference on request socket,
so add a helper to do this, reducing code size and complexity.

Fixes: 4bdc3d66 ("tcp/dccp: fix behavior of stale SYN_RECV request sockets")
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Similar to the attr usecase, the caller knows if he is holding RTNL and is
in atomic section. So let the called to decide the correct call variant.

This allows drivers to sleep inside their ops and wait for hw to get the
operation status. Then the status is propagated into switchdev core.
This avoids silent errors in drivers.
Signed-off-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When object is used in deferred work, we cannot use pointers in
switchdev object structures because the memory they point at may be already
used by someone else. So rather do local copy of the value.
Signed-off-by: NJiri Pirko <jiri@mellanox.com>
Acked-by: NScott Feldman <sfeldma@gmail.com>
Reviewed-by: NJohn Fastabend <john.r.fastabend@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Caller should know if he can call attr_set directly (when holding RTNL)
or if he has to defer the att_set processing for later.

This also allows drivers to sleep inside attr_set and report operation
status back to switchdev core. Switchdev core then warns if status is
not ok, instead of silent errors happening in drivers.

Benefit from newly introduced switchdev deferred ops infrastructure.
Signed-off-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NJiri Pirko <jiri@mellanox.com>
Reviewed-by: NVivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Introduce infrastructure which will be used internally to defer ops.
Note that the deferred ops are queued up and either are processed by
scheduled work or explicitly by user calling deferred_process function.
Signed-off-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

By design, when no default MAC addresses are set in the Hypervisor for VFs,
the VFs are passed zero-macs. When such a MAC is received by the VF, it
generates a random MAC address and registers that MAC address
with the Hypervisor.

This random mac generation is currently done in the mlx4_en module.
There is a problem, though, if the mlx4_ib module is loaded by a VF before
the mlx4_en module. In this case, for RoCE, mlx4_ib will see the un-replaced
zero-mac and register that zero-mac as part of QP1 initialization.

Having a zero-mac in the port's MAC table creates problems for a
Baseboard Management Console. The BMC occasionally sends packets with a
zero-mac destination MAC. If there is a zero-mac present in the port's
MAC table, the FW will send such BMC packets to the host driver rather than
to the wire, and BMC will stop working.

To address this problem, we move the replacement of zero-mac addresses
with random-mac addresses to procedure mlx4_slave_cap(), which is part of the
driver startup for VFs, and is before activation of mlx4_ib and mlx4_en.
As a result, zero-mac addresses will never be registered in the port MAC table
by the driver.

In addition, when mlx4_en does initialize the net device, it needs to set
the NET_ADDR_RANDOM flag in the netdev structure if the address was
randomly generated. This is done so that udev on the VM does not create
a new device name after each VF probe (VM boot and such). To accomplish this,
we add a per-port flag in mlx4_dev which gets set whenever mlx4_core replaces
a zero-mac with a randomly-generated mac. This flag is examined when mlx4_en
initializes the net-device.

Fix was suggested by Matan Barak <matanb@mellanox.com>
Signed-off-by: NJack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: NOr Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

On device initialization, wait till firmware indicates that that it is done
with initialization before proceeding to initialize the device.

Also update initialization segment layout to match driver/firmware
interface definitions.
Signed-off-by: NEli Cohen <eli@mellanox.com>
Signed-off-by: NOr Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch implement the pci_error_handlers for mlx5_core which allow the
driver to recover from PCI error.

Once an error is detected in the PCI, the mlx5_pci_err_detected is called
and it:
1) Marks the device to be in 'Internal Error' state.
2) Dispatches an event to the mlx5_ib to flush all the outstanding cqes
with error.
3) Returns all the on going commands with error.
4) Unloads the driver.

Afterwards, the FW is reset and mlx5_pci_slot_reset is called and it
enables the device and restore it's pci state.

If the later succeeds, mlx5_pci_resume is called, and it loads the SW
stack.
Signed-off-by: NMajd Dibbiny <majd@mellanox.com>
Signed-off-by: NEli Cohen <eli@mellanox.com>
Signed-off-by: NOr Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The detection of a fatal condition has been updated to take into account
the state reported by the device or by detecting an all ones read of the
firmware version which indicates that the device is not accessible.
Signed-off-by: NEli Cohen <eli@mellanox.com>
Signed-off-by: NOr Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

At listen() time, there is a small window where listener is visible with
a zero backlog, triggering a spurious "Possible SYN flooding on port"
message.

Nothing prevents us from setting the correct backlog.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This zeroes the msg so no random stack data ends up getting
sent, it also limits the function to not accepting > 4
i2c msgs.

Cc: stable@vger.kernel.org
Reviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Revert the commit e2ca690b ("ipv4/icmp: redirect messages
can use the ingress daddr as source"), which tried to introduce a more
suitable behaviour for ICMP redirect messages generated by VRRP routers.
However RFC 5798 section 8.1.1 states:

    The IPv4 source address of an ICMP redirect should be the address
    that the end-host used when making its next-hop routing decision.

while said commit used the generating packet destination
address, which do not match the above and in most cases leads to
no redirect packets to be generated.
Signed-off-by: NPaolo Abeni <pabeni@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

struct at91_can_data was used to pass a callback to the driver, allowing it
to switch the transceiver on and off. As all at91 boards are now using DT,
this is not used anymore, remove that structure.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: NMarc Kleine-Budde <mkl@pengutronix.de>

The can subsystem communicates with user space using a bcm_msg_head
header, which contains two timestamps. This is problematic for
multiple reasons:

a) The structure layout is currently incompatible between 64-bit
   user space and 32-bit user space, and cannot work in compat
   mode (other than x32).

b) The timeval structure layout will change in 32-bit user
   space when we fix the y2038 overflow problem by redefining
   time_t to 64-bit, making new 32-bit user space incompatible
   with the current kernel interface.
   Cars last a long time and often use old kernels, so the actual
   users of this code are the most likely ones to migrate to y2038
   safe user space.

This tries to work around part of the problem by changing the
publicly visible user interface in the header, but not the binary
interface. Fortunately, the values passed around in the structure
are relative times and do not actually suffer from the y2038
overflow, so 32-bit is enough here.

We replace the use of 'struct timeval' with a newly defined
'struct bcm_timeval' that uses the exact same binary layout
as before and that still suffers from problem a) but not problem
b).

The downside of this approach is that any user space program
that currently assigns a timeval structure to these members
rather than writing the tv_sec/tv_usec portions individually
will suffer a compile-time error when built with an updated
kernel header. Fixing this error makes it work fine with old
and new headers though.

We could address problem a) by using '__u32' or 'int' members
rather than 'long', but that would have a more significant
downside in also breaking support for all existing 64-bit user
binaries that might be using this interface, which is likely
not acceptable.
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Acked-by: NOliver Hartkopp <socketcan@hartkopp.net>
Cc: linux-can@vger.kernel.org
Cc: linux-api@vger.kernel.org
Signed-off-by: NMarc Kleine-Budde <mkl@pengutronix.de>

Add operations to retrieve cached IPv6 dst entry from l3mdev device
and lookup IPv6 source address.
Signed-off-by: NDavid Ahern <dsa@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The function nf_ct_frag6_gather is called on both the input and the
output paths of the networking stack.  In particular ipv6_defrag which
calls nf_ct_frag6_gather is called from both the the PRE_ROUTING chain
on input and the LOCAL_OUT chain on output.

The addition of a net parameter makes it explicit which network
namespace the packets are being reassembled in, and removes the need
for nf_ct_frag6_gather to guess.
Signed-off-by: N"Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: NPablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The function ip_defrag is called on both the input and the output
paths of the networking stack.  In particular conntrack when it is
tracking outbound packets from the local machine calls ip_defrag.

So add a struct net parameter and stop making ip_defrag guess which
network namespace it needs to defragment packets in.
Signed-off-by: N"Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: NPablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

RTA_ALIGNTO is currently define as 4. It has to be 4U to prevent warning
for RTA_ALIGN and RTA_DATA expansions when -Wconversion gcc option is
enabled.
This follows NLMSG_ALIGNTO definition in <include/uapi/linux/netlink.h>.
Signed-off-by: NRonen Arad <ronen.arad@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch allows configuring how the source address of ICMP
redirect messages is selected; by default the old behaviour is
retained, while setting icmp_redirects_use_orig_daddr force the
usage of the destination address of the packet that caused the
redirect.

The new behaviour fits closely the RFC 5798 section 8.1.1, and fix the
following scenario:

Two machines are set up with VRRP to act as routers out of a subnet,
they have IPs x.x.x.1/24 and x.x.x.2/24, with VRRP holding on to
x.x.x.254/24.

If a host in said subnet needs to get an ICMP redirect from the VRRP
router, i.e. to reach a destination behind a different gateway, the
source IP in the ICMP redirect is chosen as the primary IP on the
interface that the packet arrived at, i.e. x.x.x.1 or x.x.x.2.

The host will then ignore said redirect, due to RFC 1122 section 3.2.2.2,
and will continue to use the wrong next-op.
Signed-off-by: NPaolo Abeni <pabeni@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Reducing tcp_timewait_sock from 280 bytes to 272 bytes
allows SLAB to pack 15 objects per page instead of 14 (on x86)
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

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>

sk->sk_refcnt is dirtied for every TCP/UDP incoming packet.
This is a performance issue if multiple cpus hit a common socket,
or multiple sockets are chained due to SO_REUSEPORT.

By moving sk_refcnt 8 bytes further, first 128 bytes of sockets
are mostly read. As they contain the lookup keys, this has
a considerable performance impact, as cpus can cache them.

These 8 bytes are not wasted, we use them as a place holder
for various fields, depending on the socket type.

Tested:
 SYN flood hitting a 16 RX queues NIC.
 TCP listener using 16 sockets and SO_REUSEPORT
 and SO_INCOMING_CPU for proper siloing.

 Could process 6.0 Mpps SYN instead of 4.2 Mpps

 Kernel profile looked like :
    11.68%  [kernel]  [k] sha_transform
     6.51%  [kernel]  [k] __inet_lookup_listener
     5.07%  [kernel]  [k] __inet_lookup_established
     4.15%  [kernel]  [k] memcpy_erms
     3.46%  [kernel]  [k] ipt_do_table
     2.74%  [kernel]  [k] fib_table_lookup
     2.54%  [kernel]  [k] tcp_make_synack
     2.34%  [kernel]  [k] tcp_conn_request
     2.05%  [kernel]  [k] __netif_receive_skb_core
     2.03%  [kernel]  [k] kmem_cache_alloc
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

SO_INCOMING_CPU as added in commit 2c8c56e1 was a getsockopt() command
to fetch incoming cpu handling a particular TCP flow after accept()

This commits adds setsockopt() support and extends SO_REUSEPORT selection
logic : If a TCP listener or UDP socket has this option set, a packet is
delivered to this socket only if CPU handling the packet matches the specified
one.

This allows to build very efficient TCP servers, using one listener per
RX queue, as the associated TCP listener should only accept flows handled
in softirq by the same cpu.
This provides optimal NUMA behavior and keep cpu caches hot.

Note that __inet_lookup_listener() still has to iterate over the list of
all listeners. Following patch puts sk_refcnt in a different cache line
to let this iteration hit only shared and read mostly cache lines.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

It's useful to allow users to set fwmark for an individual packet,
without changing the socket state. The function this patch adds in
sock layer can be used by the protocols that need such a feature.
Signed-off-by: NEdward Hyunkoo Jee <edjee@google.com>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Cc: Willem de Bruijn <willemb@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>