Under unusual circumstances, TCP collapse can split a big GRO TCP packet
while its being used in a splice(socket->pipe) operation.

skb_splice_bits() releases the socket lock before calling
splice_to_pipe().

[ 1081.353685] WARNING: at net/ipv4/tcp.c:1330 tcp_cleanup_rbuf+0x4d/0xfc()
[ 1081.371956] Hardware name: System x3690 X5 -[7148Z68]-
[ 1081.391820] cleanup rbuf bug: copied AD3BCF1 seq AD370AF rcvnxt AD3CF13

To fix this problem, we must eat skbs in tcp_recv_skb().

Remove the inline keyword from tcp_recv_skb() definition since
it has three call sites.
Reported-by: NChristian Becker <c.becker@traviangames.com>
Cc: Willy Tarreau <w@1wt.eu>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Tested-by: NWilly Tarreau <w@1wt.eu>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

commit 02275a2e (tcp: don't abort splice() after small transfers)
added a regression.

[   83.843570] INFO: rcu_sched self-detected stall on CPU
[   83.844575] INFO: rcu_sched detected stalls on CPUs/tasks: { 6} (detected by 0, t=21002 jiffies, g=4457, c=4456, q=13132)
[   83.844582] Task dump for CPU 6:
[   83.844584] netperf         R  running task        0  8966   8952 0x0000000c
[   83.844587]  0000000000000000 0000000000000006 0000000000006c6c 0000000000000000
[   83.844589]  000000000000006c 0000000000000096 ffffffff819ce2bc ffffffffffffff10
[   83.844592]  ffffffff81088679 0000000000000010 0000000000000246 ffff880c4b9ddcd8
[   83.844594] Call Trace:
[   83.844596]  [<ffffffff81088679>] ? vprintk_emit+0x1c9/0x4c0
[   83.844601]  [<ffffffff815ad449>] ? schedule+0x29/0x70
[   83.844606]  [<ffffffff81537bd2>] ? tcp_splice_data_recv+0x42/0x50
[   83.844610]  [<ffffffff8153beaa>] ? tcp_read_sock+0xda/0x260
[   83.844613]  [<ffffffff81537b90>] ? tcp_prequeue_process+0xb0/0xb0
[   83.844615]  [<ffffffff8153c0f0>] ? tcp_splice_read+0xc0/0x250
[   83.844618]  [<ffffffff814dc0c2>] ? sock_splice_read+0x22/0x30
[   83.844622]  [<ffffffff811b820b>] ? do_splice_to+0x7b/0xa0
[   83.844627]  [<ffffffff811ba4bc>] ? sys_splice+0x59c/0x5d0
[   83.844630]  [<ffffffff8119745b>] ? putname+0x2b/0x40
[   83.844633]  [<ffffffff8118bcb4>] ? do_sys_open+0x174/0x1e0
[   83.844636]  [<ffffffff815b6202>] ? system_call_fastpath+0x16/0x1b

if recv_actor() returns 0, we should stop immediately,
because looping wont give a chance to drain the pipe.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Cc: Willy Tarreau <w@1wt.eu>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

TCP coalescing added a regression in splice(socket->pipe) performance,
for some workloads because of the way tcp_read_sock() is implemented.

The reason for this is the break when (offset + 1 != skb->len).

As we released the socket lock, this condition is possible if TCP stack
added a fragment to the skb, which can happen with TCP coalescing.

So let's go back to the beginning of the loop when this happens,
to give a chance to splice more frags per system call.

Doing so fixes the issue and makes GRO 10% faster than LRO
on CPU-bound splice() workloads instead of the opposite.
Signed-off-by: NWilly Tarreau <w@1wt.eu>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Recent network changes allowed high order pages being used
for skb fragments.

This uncovered a bug in do_tcp_sendpages() which was assuming its caller
provided an array of order-0 page pointers.

We only have to deal with a single page in this function, and its order
is irrelevant.
Reported-by: NWilly Tarreau <w@1wt.eu>
Tested-by: NWilly Tarreau <w@1wt.eu>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

As time passed, available memory increased faster than number of
concurrent tcp sockets.

As a result, a machine with 4GB of ram gets a hash table
with 524288 slots, using 8388608 bytes of memory.

Lets change that by a 16x factor (one slot for 128 KB of ram)

Even if a small machine needs a _lot_ of sockets, tcp lookups are now
very efficient, using one cache line per socket.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Allow an unpriviled user who has created a user namespace, and then
created a network namespace to effectively use the new network
namespace, by reducing capable(CAP_NET_ADMIN) and
capable(CAP_NET_RAW) calls to be ns_capable(net->user_ns,
CAP_NET_ADMIN), or capable(net->user_ns, CAP_NET_RAW) calls.

Settings that merely control a single network device are allowed.
Either the network device is a logical network device where
restrictions make no difference or the network device is hardware NIC
that has been explicity moved from the initial network namespace.

In general policy and network stack state changes are allowed
while resource control is left unchanged.

Allow creating raw sockets.
Allow the SIOCSARP ioctl to control the arp cache.
Allow the SIOCSIFFLAG ioctl to allow setting network device flags.
Allow the SIOCSIFADDR ioctl to allow setting a netdevice ipv4 address.
Allow the SIOCSIFBRDADDR ioctl to allow setting a netdevice ipv4 broadcast address.
Allow the SIOCSIFDSTADDR ioctl to allow setting a netdevice ipv4 destination address.
Allow the SIOCSIFNETMASK ioctl to allow setting a netdevice ipv4 netmask.
Allow the SIOCADDRT and SIOCDELRT ioctls to allow adding and deleting ipv4 routes.

Allow the SIOCADDTUNNEL, SIOCCHGTUNNEL and SIOCDELTUNNEL ioctls for
adding, changing and deleting gre tunnels.

Allow the SIOCADDTUNNEL, SIOCCHGTUNNEL and SIOCDELTUNNEL ioctls for
adding, changing and deleting ipip tunnels.

Allow the SIOCADDTUNNEL, SIOCCHGTUNNEL and SIOCDELTUNNEL ioctls for
adding, changing and deleting ipsec virtual tunnel interfaces.

Allow setting the MRT_INIT, MRT_DONE, MRT_ADD_VIF, MRT_DEL_VIF, MRT_ADD_MFC,
MRT_DEL_MFC, MRT_ASSERT, MRT_PIM, MRT_TABLE socket options on multicast routing
sockets.

Allow setting and receiving IPOPT_CIPSO, IP_OPT_SEC, IP_OPT_SID and
arbitrary ip options.

Allow setting IP_SEC_POLICY/IP_XFRM_POLICY ipv4 socket option.
Allow setting the IP_TRANSPARENT ipv4 socket option.
Allow setting the TCP_REPAIR socket option.
Allow setting the TCP_CONGESTION socket option.
Signed-off-by: N"Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Currently if a socket was repaired with a few packet in a write queue,
a kernel bug may be triggered:

kernel BUG at net/ipv4/tcp_output.c:2330!
RIP: 0010:[<ffffffff8155784f>] tcp_retransmit_skb+0x5ff/0x610

According to the initial realization v3.4-rc2-963-gc0e88ff0,
all skb-s should look like already posted. This patch fixes code
according with this sentence.

Here are three points, which were not done in the initial patch:
1. A tcp send head should not be changed
2. Initialize TSO state of a skb
3. Reset the retransmission time

This patch moves logic from tcp_sendmsg to tcp_write_xmit. A packet
passes the ussual way, but isn't sent to network. This patch solves
all described problems and handles tcp_sendpages.

Cc: Pavel Emelyanov <xemul@parallels.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
Cc: James Morris <jmorris@namei.org>
Cc: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>
Cc: Patrick McHardy <kaber@trash.net>
Signed-off-by: NAndrey Vagin <avagin@openvz.org>
Acked-by: NPavel Emelyanov <xemul@parallels.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add a bit TCPI_OPT_SYN_DATA (32) to the socket option TCP_INFO:tcpi_options.
It's set if the data in SYN (sent or received) is acked by SYN-ACK. Server or
client application can use this information to check Fast Open success rate.
Signed-off-by: NYuchung Cheng <ycheng@google.com>
Acked-by: NNeal Cardwell <ncardwell@google.com>
Acked-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

SIOCINQ can use the lock_sock_fast() version to avoid double acquisition
of socket lock.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

tcp_ioctl() tries to take into account if tcp socket received a FIN
to report correct number bytes in receive queue.

But its flaky because if the application ate the last skb,
we return 1 instead of 0.

Correct way to detect that FIN was received is to test SOCK_DONE.
Reported-by: NElliot Hughes <enh@google.com>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Tom Herbert <therbert@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We currently use a per socket order-0 page cache for tcp_sendmsg()
operations.

This page is used to build fragments for skbs.

Its done to increase probability of coalescing small write() into
single segments in skbs still in write queue (not yet sent)

But it wastes a lot of memory for applications handling many mostly
idle sockets, since each socket holds one page in sk->sk_sndmsg_page

Its also quite inefficient to build TSO 64KB packets, because we need
about 16 pages per skb on arches where PAGE_SIZE = 4096, so we hit
page allocator more than wanted.

This patch adds a per task frag allocator and uses bigger pages,
if available. An automatic fallback is done in case of memory pressure.

(up to 32768 bytes per frag, thats order-3 pages on x86)

This increases TCP stream performance by 20% on loopback device,
but also benefits on other network devices, since 8x less frags are
mapped on transmit and unmapped on tx completion. Alexander Duyck
mentioned a probable performance win on systems with IOMMU enabled.

Its possible some SG enabled hardware cant cope with bigger fragments,
but their ndo_start_xmit() should already handle this, splitting a
fragment in sub fragments, since some arches have PAGE_SIZE=65536

Successfully tested on various ethernet devices.
(ixgbe, igb, bnx2x, tg3, mellanox mlx4)
Signed-off-by: NEric Dumazet <edumazet@google.com>
Cc: Ben Hutchings <bhutchings@solarflare.com>
Cc: Vijay Subramanian <subramanian.vijay@gmail.com>
Cc: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: NVijay Subramanian <subramanian.vijay@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

rcv_wscale is a symetric parameter with snd_wscale.

Both this parameters are set on a connection handshake.

Without this value a remote window size can not be interpreted correctly,
because a value from a packet should be shifted on rcv_wscale.

And one more thing is that wscale_ok should be set too.

This patch doesn't break a backward compatibility.
If someone uses it in a old scheme, a rcv window
will be restored with the same bug (rcv_wscale = 0).

v2: Save backward compatibility on big-endian system. Before
    the first two bytes were snd_wscale and the second two bytes were
    rcv_wscale. Now snd_wscale is opt_val & 0xFFFF and rcv_wscale >> 16.
    This approach is independent on byte ordering.

Cc: David S. Miller <davem@davemloft.net>
Cc: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
Cc: James Morris <jmorris@namei.org>
Cc: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>
Cc: Patrick McHardy <kaber@trash.net>
CC: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: NAndrew Vagin <avagin@openvz.org>
Acked-by: NPavel Emelyanov <xemul@parallels.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NChristoph Paasch <christoph.paasch@uclouvain.be>
Acked-by: NH.K. Jerry Chu <hkchu@google.com>
Acked-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If recv() syscall is called for a TCP socket so that
  - IOAT DMA is used
  - MSG_WAITALL flag is used
  - requested length is bigger than sk_rcvbuf
  - enough data has already arrived to bring rcv_wnd to zero
then when tcp_recvmsg() gets to calling sk_wait_data(), receive
window can be still zero while sk_async_wait_queue exhausts
enough space to keep it zero. As this queue isn't cleaned until
the tcp_service_net_dma() call, sk_wait_data() cannot receive
any data and blocks forever.

If zero receive window and non-empty sk_async_wait_queue is
detected before calling sk_wait_data(), process the queue first.
Signed-off-by: NMichal Kubecek <mkubecek@suse.cz>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch builds on top of the previous patch to add the support
for TFO listeners. This includes -

1. allocating, properly initializing, and managing the per listener
fastopen_queue structure when TFO is enabled

2. changes to the inet_csk_accept code to support TFO. E.g., the
request_sock can no longer be freed upon accept(), not until 3WHS
finishes

3. allowing a TCP_SYN_RECV socket to properly poll() and sendmsg()
if it's a TFO socket

4. properly closing a TFO listener, and a TFO socket before 3WHS
finishes

5. supporting TCP_FASTOPEN socket option

6. modifying tcp_check_req() to use to check a TFO socket as well
as request_sock

7. supporting TCP's TFO cookie option

8. adding a new SYN-ACK retransmit handler to use the timer directly
off the TFO socket rather than the listener socket. Note that TFO
server side will not retransmit anything other than SYN-ACK until
the 3WHS is completed.

The patch also contains an important function
"reqsk_fastopen_remove()" to manage the somewhat complex relation
between a listener, its request_sock, and the corresponding child
socket. See the comment above the function for the detail.
Signed-off-by: NH.K. Jerry Chu <hkchu@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Tom Herbert <therbert@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Cache the device gso_max_segs in sock::sk_gso_max_segs and use it to
limit the size of TSO skbs.  This avoids the need to fall back to
software GSO for local TCP senders.
Signed-off-by: NBen Hutchings <bhutchings@solarflare.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

TCP_USER_TIMEOUT is a TCP level socket option that takes an unsigned int. But
patch "tcp: Add TCP_USER_TIMEOUT socket option"(dca43c75) didn't check the negative
values. If a user assign -1 to it, the socket will set successfully and wait
for 4294967295 miliseconds. This patch add a negative value check to avoid
this issue.
Signed-off-by: NHangbin Liu <liuhangbin@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

sendmsg() (or sendto()) with MSG_FASTOPEN is a combo of connect(2)
and write(2). The application should replace connect() with it to
send data in the opening SYN packet.

For blocking socket, sendmsg() blocks until all the data are buffered
locally and the handshake is completed like connect() call. It
returns similar errno like connect() if the TCP handshake fails.

For non-blocking socket, it returns the number of bytes queued (and
transmitted in the SYN-data packet) if cookie is available. If cookie
is not available, it transmits a data-less SYN packet with Fast Open
cookie request option and returns -EINPROGRESS like connect().

Using MSG_FASTOPEN on connecting or connected socket will result in
simlar errno like repeating connect() calls. Therefore the application
should only use this flag on new sockets.

The buffer size of sendmsg() is independent of the MSS of the connection.
Signed-off-by: NYuchung Cheng <ycheng@google.com>
Acked-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This introduce TSQ (TCP Small Queues)

TSQ goal is to reduce number of TCP packets in xmit queues (qdisc &
device queues), to reduce RTT and cwnd bias, part of the bufferbloat
problem.

sk->sk_wmem_alloc not allowed to grow above a given limit,
allowing no more than ~128KB [1] per tcp socket in qdisc/dev layers at a
given time.

TSO packets are sized/capped to half the limit, so that we have two
TSO packets in flight, allowing better bandwidth use.

As a side effect, setting the limit to 40000 automatically reduces the
standard gso max limit (65536) to 40000/2 : It can help to reduce
latencies of high prio packets, having smaller TSO packets.

This means we divert sock_wfree() to a tcp_wfree() handler, to
queue/send following frames when skb_orphan() [2] is called for the
already queued skbs.

Results on my dev machines (tg3/ixgbe nics) are really impressive,
using standard pfifo_fast, and with or without TSO/GSO.

Without reduction of nominal bandwidth, we have reduction of buffering
per bulk sender :
< 1ms on Gbit (instead of 50ms with TSO)
< 8ms on 100Mbit (instead of 132 ms)

I no longer have 4 MBytes backlogged in qdisc by a single netperf
session, and both side socket autotuning no longer use 4 Mbytes.

As skb destructor cannot restart xmit itself ( as qdisc lock might be
taken at this point ), we delegate the work to a tasklet. We use one
tasklest per cpu for performance reasons.

If tasklet finds a socket owned by the user, it sets TSQ_OWNED flag.
This flag is tested in a new protocol method called from release_sock(),
to eventually send new segments.

[1] New /proc/sys/net/ipv4/tcp_limit_output_bytes tunable
[2] skb_orphan() is usually called at TX completion time,
  but some drivers call it in their start_xmit() handler.
  These drivers should at least use BQL, or else a single TCP
  session can still fill the whole NIC TX ring, since TSQ will
  have no effect.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Cc: Dave Taht <dave.taht@bufferbloat.net>
Cc: Tom Herbert <therbert@google.com>
Cc: Matt Mathis <mattmathis@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Nandita Dukkipati <nanditad@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NBen Hutchings <bhutchings@solarflare.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Maintain a local hash table of TCP dynamic metrics blobs.

Computed TCP metrics are no longer maintained in the route metrics.

The table uses RCU and an extremely simple hash so that it has low
latency and low overhead.  A simple hash is legitimate because we only
make metrics blobs for fully established connections.

Some tweaking of the default hash table sizes, metric timeouts, and
the hash chain length limit certainly could use some tweaking.  But
the basic design seems sound.

With help from Eric Dumazet and Joe Perches.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

UDP stack needs a minimum hash size value for proper operation and also
uses alloc_large_system_hash() for proper NUMA distribution of its hash
tables and automatic sizing depending on available system memory.

On some low memory situations, udp_table_init() must ignore the
alloc_large_system_hash() result and reallocs a bigger memory area.

As we cannot easily free old hash table, we leak it and kmemleak can
issue a warning.

This patch adds a low limit parameter to alloc_large_system_hash() to
solve this problem.

We then specify UDP_HTABLE_SIZE_MIN for UDP/UDPLite hash table
allocation.
Reported-by: NMark Asselstine <mark.asselstine@windriver.com>
Reported-by: NTim Bird <tim.bird@am.sony.com>
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Cc: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Replace simple_strtoul with kstrtoul in three similar occurrences, all setup
handlers:
* route.c: set_rhash_entries
* tcp.c: set_thash_entries
* udp.c: set_uhash_entries

Also check if the conversion failed.
Signed-off-by: NEldad Zack <eldad@fogrefinery.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Since recent changes on TCP splicing (starting with commits 2f533844
"tcp: allow splice() to build full TSO packets" and 35f9c09f "tcp:
tcp_sendpages() should call tcp_push() once"), I started seeing
massive stalls when forwarding traffic between two sockets using
splice() when pipe buffers were larger than socket buffers.

Latest changes (net: netdev_alloc_skb() use build_skb()) made the
problem even more apparent.

The reason seems to be that if do_tcp_sendpages() fails on out of memory
condition without being able to send at least one byte, tcp_push() is not
called and the buffers cannot be flushed.

After applying the attached patch, I cannot reproduce the stalls at all
and the data rate it perfectly stable and steady under any condition
which previously caused the problem to be permanent.

The issue seems to have been there since before the kernel migrated to
git, which makes me think that the stalls I occasionally experienced
with tux during stress-tests years ago were probably related to the
same issue.

This issue was first encountered on 3.0.31 and 3.2.17, so please backport
to -stable.
Signed-off-by: NWilly Tarreau <w@1wt.eu>
Acked-by: NEric Dumazet <edumazet@google.com>
Cc: <stable@vger.kernel.org>

bool conversions where possible.

__inline__ -> inline

space cleanups
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

bool/const conversions where possible

__inline__ -> inline

space cleanups
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Standardize the net core ratelimited logging functions.

Coalesce formats, align arguments.
Change a printk then vprintk sequence to use printf extension %pV.
Signed-off-by: NJoe Perches <joe@perches.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

It actually works on the input queue and will use its read mem
routines, thus it's better to have in in the tcp_input.c file.
Signed-off-by: NPavel Emelyanov <xemul@parallels.com>
Acked-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Extend tcp coalescing implementing it from tcp_queue_rcv(), the main
receiver function when application is not blocked in recvmsg().

Function tcp_queue_rcv() is moved a bit to allow its call from
tcp_data_queue()

This gives good results especially if GRO could not kick, and if skb
head is a fragment.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Cc: Alexander Duyck <alexander.h.duyck@intel.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Tom Herbert <therbert@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

tcp_adv_win_scale default value is 2, meaning we expect a good citizen
skb to have skb->len / skb->truesize ratio of 75% (3/4)

In 2.6 kernels we (mis)accounted for typical MSS=1460 frame :
1536 + 64 + 256 = 1856 'estimated truesize', and 1856 * 3/4 = 1392.
So these skbs were considered as not bloated.

With recent truesize fixes, a typical MSS=1460 frame truesize is now the
more precise :
2048 + 256 = 2304. But 2304 * 3/4 = 1728.
So these skb are not good citizen anymore, because 1460 < 1728

(GRO can escape this problem because it build skbs with a too low
truesize.)

This also means tcp advertises a too optimistic window for a given
allocated rcvspace : When receiving frames, sk_rmem_alloc can hit
sk_rcvbuf limit and we call tcp_prune_queue()/tcp_collapse() too often,
especially when application is slow to drain its receive queue or in
case of losses (netperf is fast, scp is slow). This is a major latency
source.

We should adjust the len/truesize ratio to 50% instead of 75%

This patch :

1) changes tcp_adv_win_scale default to 1 instead of 2

2) increase tcp_rmem[2] limit from 4MB to 6MB to take into account
better truesize tracking and to allow autotuning tcp receive window to
reach same value than before. Note that same amount of kernel memory is
consumed compared to 2.6 kernels.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Tom Herbert <therbert@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Acked-by: NNeal Cardwell <ncardwell@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch implements RFC 5827 early retransmit (ER) for TCP.
It reduces DUPACK threshold (dupthresh) if outstanding packets are
less than 4 to recover losses by fast recovery instead of timeout.

While the algorithm is simple, small but frequent network reordering
makes this feature dangerous: the connection repeatedly enter
false recovery and degrade performance. Therefore we implement
a mitigation suggested in the appendix of the RFC that delays
entering fast recovery by a small interval, i.e., RTT/4. Currently
ER is conservative and is disabled for the rest of the connection
after the first reordering event. A large scale web server
experiment on the performance impact of ER is summarized in
section 6 of the paper "Proportional Rate Reduction for TCP”,
IMC 2011. http://conferences.sigcomm.org/imc/2011/docs/p155.pdf

Note that Linux has a similar feature called THIN_DUPACK. The
differences are THIN_DUPACK do not mitigate reorderings and is only
used after slow start. Currently ER is disabled if THIN_DUPACK is
enabled. I would be happy to merge THIN_DUPACK feature with ER if
people think it's a good idea.

ER is enabled by sysctl_tcp_early_retrans:
  0: Disables ER

  1: Reduce dupthresh to packets_out - 1 when outstanding packets < 4.

  2: (Default) reduce dupthresh like mode 1. In addition, delay
     entering fast recovery by RTT/4.

Note: mode 2 is implemented in the third part of this patch series.
Signed-off-by: NYuchung Cheng <ycheng@google.com>
Acked-by: NNeal Cardwell <ncardwell@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Don't pick __u8/__u16 values directly from raw pointers, but instead use
an array of structures of code:value pairs. This is OK, since the buffer
we take options from is not an skb memory, but a user-to-kernel one.

For those options which don't require any value now, require this to be
zero (for potential future extension of this API).

v2: Changed tcp_repair_opt to use two __u32-s as spotted by David Laight.
Signed-off-by: NPavel Emelyanov <xemul@parallels.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This commit moves the (substantial) common code shared between
tcp_v4_init_sock() and tcp_v6_init_sock() to a new address-family
independent function, tcp_init_sock().

Centralizing this functionality should help avoid drift issues,
e.g. where the IPv4 side is updated without a corresponding update to
IPv6. There was already some drift: IPv4 initialized snd_cwnd to
TCP_INIT_CWND, while the IPv6 side was still initializing snd_cwnd to
2 (in this case it should not matter, since snd_cwnd is also
initialized in tcp_init_metrics(), but the general risks and
maintenance overhead remain).

When diffing the old and new code, note that new tcp_init_sock()
function uses the order of steps from the tcp_v4_init_sock()
implementation (the order is slightly different in
tcp_v6_init_sock()).
Signed-off-by: NNeal Cardwell <ncardwell@google.com>
Acked-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

There are options, which are set up on a socket while performing
TCP handshake. Need to resurrect them on a socket while repairing.
A new sockoption accepts a buffer and parses it. The buffer should
be CODE:VALUE sequence of bytes, where CODE is standard option
code and VALUE is the respective value.

Only 4 options should be handled on repaired socket.

To read 3 out of 4 of these options the TCP_INFO sockoption can be
used. An ability to get the last one (the mss_clamp) was added by
the previous patch.

Now the restore. Three of these options -- timestamp_ok, mss_clamp
and snd_wscale -- are just restored on a coket.

The sack_ok flags has 2 issues. First, whether or not to do sacks
at all. This flag is just read and set back. No other sack  info is
saved or restored, since according to the standart and the code
dropping all sack-ed segments is OK, the sender will resubmit them
again, so after the repair we will probably experience a pause in
connection. Next, the fack bit. It's just set back on a socket if
the respective sysctl is set. No collected stats about packets flow
is preserved. As far as I see (plz, correct me if I'm wrong) the
fack-based congestion algorithm survives dropping all of the stats
and repairs itself eventually, probably losing the performance for
that period.
Signed-off-by: NPavel Emelyanov <xemul@openvz.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The mss_clamp is the only connection-time negotiated option which
cannot be obtained from the user space. Make the TCP_MAXSEG sockopt
report one in the repair mode.
Signed-off-by: NPavel Emelyanov <xemul@openvz.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Reading queues under repair mode is done with recvmsg call.
The queue-under-repair set by TCP_REPAIR_QUEUE option is used
to determine which queue should be read. Thus both send and
receive queue can be read with this.

Caller must pass the MSG_PEEK flag.

Writing to queues is done with sendmsg call and yet again --
the repair-queue option can be used to push data into the
receive queue.

When putting an skb into receive queue a zero tcp header is
appented to its head to address the tcp_hdr(skb)->syn and
the ->fin checks by the (after repair) tcp_recvmsg. These
flags flags are both set to zero and that's why.

The fin cannot be met in the queue while reading the source
socket, since the repair only works for closed/established
sockets and queueing fin packet always changes its state.

The syn in the queue denotes that the respective skb's seq
is "off-by-one" as compared to the actual payload lenght. Thus,
at the rcv queue refill we can just drop this flag and set the
skb's sequences to precice values.

When the repair mode is turned off, the write queue seqs are
updated so that the whole queue is considered to be 'already sent,
waiting for ACKs' (write_seq = snd_nxt <= snd_una). From the
protocol POV the send queue looks like it was sent, but the data
between the write_seq and snd_nxt is lost in the network.

This helps to avoid another sockoption for setting the snd_nxt
sequence. Leaving the whole queue in a 'not yet sent' state (as
it will be after sendmsg-s) will not allow to receive any acks
from the peer since the ack_seq will be after the snd_nxt. Thus
even the ack for the window probe will be dropped and the
connection will be 'locked' with the zero peer window.
Signed-off-by: NPavel Emelyanov <xemul@parallels.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This includes (according the the previous description):

* TCP_REPAIR sockoption

This one just puts the socket in/out of the repair mode.
Allowed for CAP_NET_ADMIN and for closed/establised sockets only.
When repair mode is turned off and the socket happens to be in
the established state the window probe is sent to the peer to
'unlock' the connection.

* TCP_REPAIR_QUEUE sockoption

This one sets the queue which we're about to repair. The
'no-queue' is set by default.

* TCP_QUEUE_SEQ socoption

Sets the write_seq/rcv_nxt of a selected repaired queue.
Allowed for TCP_CLOSE-d sockets only. When the socket changes
its state the other seq-s are changed by the kernel according
to the protocol rules (most of the existing code is actually
reused).

* Ability to forcibly bind a socket to a port

The sk->sk_reuse is set to SK_FORCE_REUSE.

* Immediate connect modification

The connect syscall initializes the connection, then directly jumps
to the code which finalizes it.

* Silent close modification

The close just aborts the connection (similar to SO_LINGER with 0
time) but without sending any FIN/RST-s to peer.
Signed-off-by: NPavel Emelyanov <xemul@parallels.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This is just the preparation patch, which makes the needed for
TCP repair code ready for use.
Signed-off-by: NPavel Emelyanov <xemul@parallels.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Use of "unsigned int" is preferred to bare "unsigned" in net tree.
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>