Under NET_DMA, data transfer can grind to a halt when userland issues a
large read on a socket with a high RCVLOWAT (i.e., 512 KB for both).
This appears to be because the NET_DMA design queues up lots of memcpy
operations, but doesn't issue or wait for them (and thus free the
associated skbs) until it is time for tcp_recvmesg() to return.
The socket hangs when its TCP window goes to zero before enough data is
available to satisfy the read.

Periodically issue asynchronous memcpy operations, and free skbs for ones
that have completed, to prevent sockets from going into zero-window mode.
Signed-off-by: NSteven J. Magnani <steve@digidescorp.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

From: Alexandra.Kossovsky@oktetlabs.ru

Fixes kernel bugzilla #15541
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch enables fast retransmissions after one dupACK for
TCP if the stream is identified as thin. This will reduce
latencies for thin streams that are not able to trigger fast
retransmissions due to high packet interarrival time. This
mechanism is only active if enabled by iocontrol or syscontrol
and the stream is identified as thin.
Signed-off-by: NAndreas Petlund <apetlund@simula.no>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch will make TCP use only linear timeouts if the
stream is thin. This will help to avoid the very high latencies
that thin stream suffer because of exponential backoff. This
mechanism is only active if enabled by iocontrol or syscontrol
and the stream is identified as thin. A maximum of 6 linear
timeouts is tried before exponential backoff is resumed.
Signed-off-by: NAndreas Petlund <apetlund@simula.no>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add __percpu sparse annotations to net.

These annotations are to make sparse consider percpu variables to be
in a different address space and warn if accessed without going
through percpu accessors.  This patch doesn't affect normal builds.

The macro and type tricks around snmp stats make things a bit
interesting.  DEFINE/DECLARE_SNMP_STAT() macros mark the target field
as __percpu and SNMP_UPD_PO_STATS() macro is updated accordingly.  All
snmp_mib_*() users which used to cast the argument to (void **) are
updated to cast it to (void __percpu **).
Signed-off-by: NTejun Heo <tj@kernel.org>
Acked-by: NDavid S. Miller <davem@davemloft.net>
Cc: Patrick McHardy <kaber@trash.net>
Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Cc: Vlad Yasevich <vladislav.yasevich@hp.com>
Cc: netdev@vger.kernel.org
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Slightly optimize tcp_sendmsg since NETIF_F_SG is used many
times iteratively in the loop. The only other modification is
to change:
			} else if (i == MAX_SKB_FRAGS ||
				   (!i &&
				   !(sk->sk_route_caps & NETIF_F_SG))) {
	to:
			} else if (i == MAX_SKB_FRAGS || !sg) {

The reason why this change is correct: this code (other than
the MAX_SKB_FRAGS case) executes only due to the else part
of: "if (skb_tailroom(skb) > 0) {" - i.e. there was no space
in the skb to put the data inline. Hence SG is false is a
sufficient condition, and there is no way a fragment can be
added to the skb.

Changelog:
	- Added the above explanation for the change
Signed-off-by: NKrishna Kumar <krkumar2@in.ibm.com>
Acked-by: NIlpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Remove unrequired operations in tcp_push()

Changelog:
	Removed a temporary skb variable from tcp_push()
Signed-off-by: NKrishna Kumar <krkumar2@in.ibm.com>
Acked-by: NIlpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

GCC even warns about it, as reported by Andrew Morton:

net/ipv4/tcp.c: In function 'do_tcp_getsockopt':
net/ipv4/tcp.c:2544: warning: comparison is always false due to limited range of data type
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Provide per socket control of the TCP cookie option and SYN/SYNACK data.

This is a straightforward re-implementation of an earlier (year-old)
patch that no longer applies cleanly, with permission of the original
author (Adam Langley):

    http://thread.gmane.org/gmane.linux.network/102586

The principle difference is using a TCP option to carry the cookie nonce,
instead of a user configured offset in the data.

Allocations have been rearranged to avoid requiring GFP_ATOMIC.

Requires:
   net: TCP_MSS_DEFAULT, TCP_MSS_DESIRED
   TCPCT part 1c: sysctl_tcp_cookie_size, socket option TCP_COOKIE_TRANSACTIONS
   TCPCT part 1d: define TCP cookie option, extend existing struct's

Signed-off-by: William.Allen.Simpson@gmail.com
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Define (missing) hash message size for SHA1.

Define hashing size constants specific to TCP cookies.

Add new function: tcp_cookie_generator().

Maintain global secret values for tcp_cookie_generator().

This is a significantly revised implementation of earlier (15-year-old)
Photuris [RFC-2522] code for the KA9Q cooperative multitasking platform.

Linux RCU technique appears to be well-suited to this application, though
neither of the circular queue items are freed.

These functions will also be used in subsequent patches that implement
additional features.

Signed-off-by: William.Allen.Simpson@gmail.com
Acked-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

NFS can reuse its TCP socket after calling tcp_disconnect().

We noticed window scaling was not negotiated in SYN packet of next
connection request.

Fix is to clear tp->window_clamp in tcp_disconnect().
Reported-by: NKrzysztof Oledzki <ole@ans.pl>
Tested-by: NKrzysztof Oledzki <ole@ans.pl>
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The addition of rcv_nxt allows to discern whether the skb
was out of place or tp->copied. Also catch fancy combination
of flags if necessary (sadly we might miss the actual causer
flags as it might have already returned).

Btw, we perhaps would want to forward copied_seq in
somewhere or otherwise we might have some nice loop with
WARN stuff within but where to do that safely I don't
know at this stage until more is known (but it is not
made significantly worse by this patch).
Signed-off-by: NIlpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Commit b6b39e8f (tcp: Try to catch MSG_PEEK bug) added a printk()
to the WARN_ON() that's in tcp.c. This patch changes this combination
to WARN(); the advantage of WARN() is that the printk message shows up
inside the message, so that kerneloops.org will collect the message.

In addition, this gets rid of an extra if() statement.
Signed-off-by: NArjan van de Ven <arjan@linux.intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch tries to print out more information when we hit the
MSG_PEEK bug in tcp_recvmsg.  It's been around since at least
2005 and it's about time that we finally fix it.
Signed-off-by: NHerbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Fix TCP_DEFER_ACCEPT conversion between seconds and
retransmission to match the TCP SYN-ACK retransmission periods
because the time is converted to such retransmissions. The old
algorithm selects one more retransmission in some cases. Allow
up to 255 retransmissions.
Signed-off-by: NJulian Anastasov <ja@ssi.bg>
Acked-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In order to have better cache layouts of struct sock (separate zones
for rx/tx paths), we need this preliminary patch.

Goal is to transfert fields used at lookup time in the first
read-mostly cache line (inside struct sock_common) and move sk_refcnt
to a separate cache line (only written by rx path)

This patch adds inet_ prefix to daddr, rcv_saddr, dport, num, saddr,
sport and id fields. This allows a future patch to define these
fields as macros, like sk_refcnt, without name clashes.
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Storing the mask (size - 1) instead of the size allows fast path to be
a bit faster.
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

tcp_splice_read() doesnt take into account socket's O_NONBLOCK flag

Before this patch :

splice(socket,0,pipe,0,128*1024,SPLICE_F_MOVE);
causes a random endless block (if pipe is full) and
splice(socket,0,pipe,0,128*1024,SPLICE_F_MOVE | SPLICE_F_NONBLOCK);
will return 0 immediately if the TCP buffer is empty.

User application has no way to instruct splice() that socket should be in blocking mode
but pipe in nonblock more.

Many projects cannot use splice(tcp -> pipe) because of this flaw.

http://git.samba.org/?p=samba.git;a=history;f=source3/lib/recvfile.c;h=ea0159642137390a0f7e57a123684e6e63e47581;hb=HEAD
http://lkml.indiana.edu/hypermail/linux/kernel/0807.2/0687.html

Linus introduced  SPLICE_F_NONBLOCK in commit 29e35094
(splice: add SPLICE_F_NONBLOCK flag )

  It doesn't make the splice itself necessarily nonblocking (because the
  actual file descriptors that are spliced from/to may block unless they
  have the O_NONBLOCK flag set), but it makes the splice pipe operations
  nonblocking.

Linus intention was clear : let SPLICE_F_NONBLOCK control the splice pipe mode only

This patch instruct tcp_splice_read() to use the underlying file O_NONBLOCK
flag, as other socket operations do.

Users will then call :

splice(socket,0,pipe,0,128*1024,SPLICE_F_MOVE | SPLICE_F_NONBLOCK );

to block on data coming from socket (if file is in blocking mode),
and not block on pipe output (to avoid deadlock)

First version of this patch was submitted by Octavian Purdila
Reported-by: NVolker Lendecke <vl@samba.org>
Reported-by: NJason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NOctavian Purdila <opurdila@ixiacom.com>
Acked-by: NLinus Torvalds <torvalds@linux-foundation.org>
Acked-by: NJens Axboe <jens.axboe@oracle.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

net/ipv4/tcp.c: In function 'do_tcp_setsockopt':
net/ipv4/tcp.c:2050: warning: comparison of distinct pointer types lacks a cast
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This provides safety against negative optlen at the type
level instead of depending upon (sometimes non-trivial)
checks against this sprinkled all over the the place, in
each and every implementation.

Based upon work done by Arjan van de Ven and feedback
from Linus Torvalds.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Sizing of memory allocations shouldn't depend on the number of physical
pages found in a system, as that generally includes (perhaps a huge amount
of) non-RAM pages.  The amount of what actually is usable as storage
should instead be used as a basis here.

Some of the calculations (i.e.  those not intending to use high memory)
should likely even use (totalram_pages - totalhigh_pages).
Signed-off-by: NJan Beulich <jbeulich@novell.com>
Acked-by: NRusty Russell <rusty@rustcorp.com.au>
Acked-by: NIngo Molnar <mingo@elte.hu>
Cc: Dave Airlie <airlied@linux.ie>
Cc: Kyle McMartin <kyle@mcmartin.ca>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Cc: Hugh Dickins <hugh.dickins@tiscali.co.uk>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Patrick McHardy <kaber@trash.net>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

It was once upon time so that snd_sthresh was a 16-bit quantity.
...That has not been true for long period of time. I run across
some ancient compares which still seem to trust such legacy.
Put all that magic into a single place, I hopefully found all
of them.

Compile tested, though linking of allyesconfig is ridiculous
nowadays it seems.
Signed-off-by: NIlpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This fixed a lockdep warning which appeared when doing stress
memory tests over NFS:

	inconsistent {RECLAIM_FS-ON-W} -> {IN-RECLAIM_FS-W} usage.

	page reclaim => nfs_writepage => tcp_sendmsg => lock sk_lock

	mount_root => nfs_root_data => tcp_close => lock sk_lock =>
			tcp_send_fin => alloc_skb_fclone => page reclaim

David raised a concern that if the allocation fails in tcp_send_fin(), and it's
GFP_ATOMIC, we are going to yield() (which sleeps) and loop endlessly waiting
for the allocation to succeed.

But fact is, the original GFP_KERNEL also sleeps. GFP_ATOMIC+yield() looks
weird, but it is no worse the implicit sleep inside GFP_KERNEL. Both could
loop endlessly under memory pressure.

CC: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
CC: David S. Miller <davem@davemloft.net>
CC: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: NWu Fengguang <fengguang.wu@intel.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Introduce keepalive_probes(tp) helper, and use it, like 
keepalive_time_when(tp) and keepalive_intvl_when(tp)
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Adding memory barrier after the poll_wait function, paired with
receive callbacks. Adding fuctions sock_poll_wait and sk_has_sleeper
to wrap the memory barrier.

Without the memory barrier, following race can happen.
The race fires, when following code paths meet, and the tp->rcv_nxt
and __add_wait_queue updates stay in CPU caches.

CPU1                         CPU2

sys_select                   receive packet
  ...                        ...
  __add_wait_queue           update tp->rcv_nxt
  ...                        ...
  tp->rcv_nxt check          sock_def_readable
  ...                        {
  schedule                      ...
                                if (sk->sk_sleep && waitqueue_active(sk->sk_sleep))
                                        wake_up_interruptible(sk->sk_sleep)
                                ...
                             }

If there was no cache the code would work ok, since the wait_queue and
rcv_nxt are opposit to each other.

Meaning that once tp->rcv_nxt is updated by CPU2, the CPU1 either already
passed the tp->rcv_nxt check and sleeps, or will get the new value for
tp->rcv_nxt and will return with new data mask.
In both cases the process (CPU1) is being added to the wait queue, so the
waitqueue_active (CPU2) call cannot miss and will wake up CPU1.

The bad case is when the __add_wait_queue changes done by CPU1 stay in its
cache, and so does the tp->rcv_nxt update on CPU2 side.  The CPU1 will then
endup calling schedule and sleep forever if there are no more data on the
socket.

Calls to poll_wait in following modules were ommited:
	net/bluetooth/af_bluetooth.c
	net/irda/af_irda.c
	net/irda/irnet/irnet_ppp.c
	net/mac80211/rc80211_pid_debugfs.c
	net/phonet/socket.c
	net/rds/af_rds.c
	net/rfkill/core.c
	net/sunrpc/cache.c
	net/sunrpc/rpc_pipe.c
	net/tipc/socket.c
Signed-off-by: NJiri Olsa <jolsa@redhat.com>
Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If a socket starts out on a non-TSO route, and then switches to
a TSO route, then we will tack on data to the tail of the tx queue
even if it started out life as non-TSO.  This is suboptimal because
all of it will then be copied and checksummed unnecessarily.

This patch fixes this by ensuring that skb->ip_summed is set to
CHECKSUM_PARTIAL before appending extra data beyond the MSS.
Signed-off-by: NHerbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NDavid S. Miller <davem@davemloft.net>

There is no need to repeatedly check flush when comparing TCP
options for GRO as it will be false 99% of the time where it
matters.
Signed-off-by: NHerbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

For the overwhelming majority of cases, skb_gro_header's return
value cannot be NULL.  Yet we must check it because of its current
form.  This patch splits it up into multiple functions in order
to avoid this.
Signed-off-by: NHerbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Instead of checking len > mss || len == 0, we can accomplish
both by checking (len - 1) > mss using the unsigned wraparound.
At nearly a million times a second, this might just help.
Signed-off-by: NHerbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The window has already been checked as part of the flag word
so there is no need to check it explicitly.
Signed-off-by: NHerbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Instead of doing two 16-bit operations for the source/destination
ports, we can do one 32-bit operation to take care both.
Signed-off-by: NHerbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Commit 518a09ef (tcp: Fix recvmsg MSG_PEEK influence of
blocking behavior) lets the loop run longer than the race check
did previously expect, so we need to be more careful with this
check and consider the work we have been doing.

I tried my best to deal with urg hole madness too which happens
here:
	if (!sock_flag(sk, SOCK_URGINLINE)) {
		++*seq;
		...
by using additional offset by one but I certainly have very
little interest in testing that part.
Signed-off-by: NIlpo Järvinen <ilpo.jarvinen@helsinki.fi>
Tested-by: NFrans Pop <elendil@planet.nl>
Tested-by: NIan Zimmermann <itz@buug.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

After calling skb_gro_receive skb->len can no longer be relied
on since if the skb was merged using frags, then its pages will
have been removed and the length reduced.

This caused tcp_gro_receive to prematurely end merging which
resulted in suboptimal performance with ixgbe.

The fix is to store skb->len on the stack.
Reported-by: NMark Wagner <mwagner@redhat.com>
Signed-off-by: NHerbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NRami Rosen <ramirose@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch removes an unused parameter (addr_len) from tcp_recv_urg()
method in net/ipv4/tcp.c.
Signed-off-by: NRami Rosen <ramirose@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

It's not too likely to happen, would basically require crafted
packets (must hit the max guard in tcp_bound_to_half_wnd()).
It seems that nothing that bad would happen as there's tcp_mems
and congestion window that prevent runaway at some point from
hurting all too much (I'm not that sure what all those zero
sized segments we would generate do though in write queue).
Preventing it regardless is certainly the best way to go.
Signed-off-by: NIlpo Järvinen <ilpo.jarvinen@helsinki.fi>
Cc: Evgeniy Polyakov <zbr@ioremap.net>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The results is very unlikely change every so often so we
hardly need to divide again after doing that once for a
connection. Yet, if divide still becomes necessary we
detect that and do the right thing and again settle for
non-divide state. Takes the u16 space which was previously
taken by the plain xmit_size_goal.

This should take care part of the tso vs non-tso difference
we found earlier.
Signed-off-by: NIlpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

There's very little need for most of the callsites to get
tp->xmit_goal_size updated. That will cost us divide as is,
so slice the function in two. Also, the only users of the
tp->xmit_goal_size are directly behind tcp_current_mss(),
so there's no need to store that variable into tcp_sock
at all! The drop of xmit_goal_size currently leaves 16-bit
hole and some reorganization would again be necessary to
change that (but I'm aiming to fill that hole with u16
xmit_goal_size_segs to cache the results of the remaining
divide to get that tso on regression).

Bring xmit_goal_size parts into tcp.c
Signed-off-by: NIlpo Järvinen <ilpo.jarvinen@helsinki.fi>
Cc: Evgeniy Polyakov <zbr@ioremap.net>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

copied was assigned zero right before the goto, so if (copied)
cannot ever be true.
Signed-off-by: NIlpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>