Xen virtual network driver has higher latency than a physical NIC.
Having only 128K as limit for TSQ introduced 30% regression in guest
throughput.
This patch raises the limit to 256K. This reduces the regression to 8%.
This buys us more time to work out a proper solution in the long run.
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Cc: David Miller <davem@davemloft.net>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David 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: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/phy/amd-xgbe-phy.c
drivers/net/wireless/iwlwifi/Kconfig
include/net/mac80211.h
iwlwifi/Kconfig and mac80211.h were both trivial overlapping
changes.
The drivers/net/phy/amd-xgbe-phy.c file got removed in 'net-next' and
the bug fix that happened on the 'net' side is already integrated
into the rest of the amd-xgbe driver.
Signed-off-by: David S. Miller <davem@davemloft.net>
Linux 3.17 and earlier are explicitly engineered so that if the app
doesn't specifically request a CC module on a listener before the SYN
arrives, then the child gets the system default CC when the connection
is established. See tcp_init_congestion_control() in 3.17 or earlier,
which says "if no choice made yet assign the current value set as
default". The change ("net: tcp: assign tcp cong_ops when tcp sk is
created") altered these semantics, so that children got their parent
listener's congestion control even if the system default had changed
after the listener was created.
This commit returns to those original semantics from 3.17 and earlier,
since they are the original semantics from 2007 in 4d4d3d1e8 ("[TCP]:
Congestion control initialization."), and some Linux congestion
control workflows depend on that.
In summary, if a listener socket specifically sets TCP_CONGESTION to
"x", or the route locks the CC module to "x", then the child gets
"x". Otherwise the child gets current system default from
net.ipv4.tcp_congestion_control. That's the behavior in 3.17 and
earlier, and this commit restores that.
Fixes: 55d8694fa8 ("net: tcp: assign tcp cong_ops when tcp sk is created")
Cc: Florian Westphal <fw@strlen.de>
Cc: Daniel Borkmann <dborkman@redhat.com>
Cc: Glenn Judd <glenn.judd@morganstanley.com>
Cc: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
We have two problems in UDP stack related to bogus checksums :
1) We return -EAGAIN to application even if receive queue is not empty.
This breaks applications using edge trigger epoll()
2) Under UDP flood, we can loop forever without yielding to other
processes, potentially hanging the host, especially on non SMP.
This patch is an attempt to make things better.
We might in the future add extra support for rt applications
wanting to better control time spent doing a recv() in a hostile
environment. For example we could validate checksums before queuing
packets in socket receive queue.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pablo Neira Ayuso says:
====================
Netfilter updates for net-next
The following patchset contains Netfilter updates for net-next, they are:
1) default CONFIG_NETFILTER_INGRESS to y for easier compile-testing of all
options.
2) Allow to bind a table to net_device. This introduces the internal
NFT_AF_NEEDS_DEV flag to perform a mandatory check for this binding.
This is required by the next patch.
3) Add the 'netdev' table family, this new table allows you to create ingress
filter basechains. This provides access to the existing nf_tables features
from ingress.
4) Kill unused argument from compat_find_calc_{match,target} in ip_tables
and ip6_tables, from Florian Westphal.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
This is similar to b1cb59cf2efe(net: sysctl_net_core: check SNDBUF
and RCVBUF for min length). I don't think too small values can cause
crashes in the case of udp and tcp, but I've seen this set to too
small values which triggered awful performance. It also makes the
setting consistent across all the wmem/rmem sysctls.
Signed-off-by: Sorin Dumitru <sdumitru@ixiacom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Steffen Klassert says:
====================
pull request (net): ipsec 2015-05-28
1) Fix a race in xfrm_state_lookup_byspi, we need to take
the refcount before we release xfrm_state_lock.
From Li RongQing.
2) Fix IV generation on ESN state. We used just the
low order sequence numbers for IV generation on
ESN, as a result the IV can repeat on the same
state. Fix this by using the high order sequence
number bits too and make sure to always initialize
the high order bits with zero. These patches are
serious stable candidates. Fixes from Herbert Xu.
3) Fix the skb->mark handling on vti. We don't
reset skb->mark in skb_scrub_packet anymore,
so vti must care to restore the original
value back after it was used to lookup the
vti policy and state. Fixes from Alexander Duyck.
Please pull or let me know if there are problems.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
The vti6_rcv_cb and vti_rcv_cb calls were leaving the skb->mark modified
after completing the function. This resulted in the original skb->mark
value being lost. Since we only need skb->mark to be set for
xfrm_policy_check we can pull the assignment into the rcv_cb calls and then
just restore the original mark after xfrm_policy_check has been completed.
Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Instead of modifying skb->mark we can simply modify the flowi_mark that is
generated as a result of the xfrm_decode_session. By doing this we don't
need to actually touch the skb->mark and it can be preserved as it passes
out through the tunnel.
Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
After commit 07f4c90062 ("tcp/dccp: try to not exhaust
ip_local_port_range in connect()") it is advised to have an even number
of ports described in /proc/sys/net/ipv4/ip_local_port_range
This means start/end values should have a different parity.
Let's warn sysadmins of this, so that they can update their settings
if they want to.
Suggested-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
__inet_hash_connect() does not use its third argument (port_offset)
if socket was already bound to a source port.
No need to perform useless but expensive md5 computations.
Reported-by: Crestez Dan Leonard <cdleonard@gmail.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A long standing problem on busy servers is the tiny available TCP port
range (/proc/sys/net/ipv4/ip_local_port_range) and the default
sequential allocation of source ports in connect() system call.
If a host is having a lot of active TCP sessions, chances are
very high that all ports are in use by at least one flow,
and subsequent bind(0) attempts fail, or have to scan a big portion of
space to find a slot.
In this patch, I changed the starting point in __inet_hash_connect()
so that we try to favor even [1] ports, leaving odd ports for bind()
users.
We still perform a sequential search, so there is no guarantee, but
if connect() targets are very different, end result is we leave
more ports available to bind(), and we spread them all over the range,
lowering time for both connect() and bind() to find a slot.
This strategy only works well if /proc/sys/net/ipv4/ip_local_port_range
is even, ie if start/end values have different parity.
Therefore, default /proc/sys/net/ipv4/ip_local_port_range was changed to
32768 - 60999 (instead of 32768 - 61000)
There is no change on security aspects here, only some poor hashing
schemes could be eventually impacted by this change.
[1] : The odd/even property depends on ip_local_port_range values parity
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We currently always send fragments without DF bit set.
Thus, given following setup:
mtu1500 - mtu1500:1400 - mtu1400:1280 - mtu1280
A R1 R2 B
Where R1 and R2 run linux with netfilter defragmentation/conntrack
enabled, then if Host A sent a fragmented packet _with_ DF set to B, R1
will respond with icmp too big error if one of these fragments exceeded
1400 bytes.
However, if R1 receives fragment sizes 1200 and 100, it would
forward the reassembled packet without refragmenting, i.e.
R2 will send an icmp error in response to a packet that was never sent,
citing mtu that the original sender never exceeded.
The other minor issue is that a refragmentation on R1 will conceal the
MTU of R2-B since refragmentation does not set DF bit on the fragments.
This modifies ip_fragment so that we track largest fragment size seen
both for DF and non-DF packets, and set frag_max_size to the largest
value.
If the DF fragment size is larger or equal to the non-df one, we will
consider the packet a path mtu probe:
We set DF bit on the reassembled skb and also tag it with a new IPCB flag
to force refragmentation even if skb fits outdev mtu.
We will also set DF bit on each fragment in this case.
Joint work with Hannes Frederic Sowa.
Reported-by: Jesse Gross <jesse@nicira.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
ip_skb_dst_mtu is small inline helper, but its called in several places.
before: 17061 44 0 17105 42d1 net/ipv4/ip_output.o
after: 16805 44 0 16849 41d1 net/ipv4/ip_output.o
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
We used to get this indirectly I supposed, but no longer do.
Either way, an explicit include should have been done in the
first place.
net/ipv4/fib_trie.c: In function '__node_free_rcu':
>> net/ipv4/fib_trie.c:293:3: error: implicit declaration of function 'vfree' [-Werror=implicit-function-declaration]
vfree(n);
^
net/ipv4/fib_trie.c: In function 'tnode_alloc':
>> net/ipv4/fib_trie.c:312:3: error: implicit declaration of function 'vzalloc' [-Werror=implicit-function-declaration]
return vzalloc(size);
^
>> net/ipv4/fib_trie.c:312:3: warning: return makes pointer from integer without a cast
cc1: some warnings being treated as errors
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
By making sure sk->sk_gso_max_segs minimal value is one,
and sysctl_tcp_min_tso_segs minimal value is one as well,
tcp_tso_autosize() will return a non zero value.
We can then revert 843925f33f
("tcp: Do not apply TSO segment limit to non-TSO packets")
and save few cpu cycles in fast path.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
If tcp ehash table is constrained to a very small number of buckets
(eg boot parameter thash_entries=128), then we can crash if spinlock
array has more entries.
While we are at it, un-inline inet_ehash_locks_alloc() and make
following changes :
- Budget 2 cache lines per cpu worth of 'spinlocks'
- Try to kmalloc() the array to avoid extra TLB pressure.
(Most servers at Google allocate 8192 bytes for this hash table)
- Get rid of various #ifdef
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Send icmp pmtu error if we find that the largest fragment of df-skb
exceeded the output path mtu.
The ip output path will still catch this later on but we can avoid the
forward/postrouting hook traversal by rejecting right away.
This is what ipv6 already does.
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Prepare skb_splice_bits to be able to deal with AF_UNIX sockets.
AF_UNIX sockets don't use lock_sock/release_sock and thus we have to
use a callback to make the locking and unlocking configureable.
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/ethernet/cadence/macb.c
drivers/net/phy/phy.c
include/linux/skbuff.h
net/ipv4/tcp.c
net/switchdev/switchdev.c
Switchdev was a case of RTNH_H_{EXTERNAL --> OFFLOAD}
renaming overlapping with net-next changes of various
sorts.
phy.c was a case of two changes, one adding a local
variable to a function whilst the second was removing
one.
tcp.c overlapped a deadlock fix with the addition of new tcp_info
statistic values.
macb.c involved the addition of two zyncq device entries.
skbuff.h involved adding back ipv4_daddr to nf_bridge_info
whilst net-next changes put two other existing members of
that struct into a union.
Signed-off-by: David S. Miller <davem@davemloft.net>
When replacing an IPv4 route, tb_id member of the new fib_alias
structure is not set in the replace code path so that the new route is
ignored.
Fixes: 0ddcf43d5d ("ipv4: FIB Local/MAIN table collapse")
Signed-off-by: Michal Kubecek <mkubecek@suse.cz>
Acked-by: Alexander Duyck <alexander.h.duyck@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pablo Neira Ayuso says:
====================
Netfilter fixes for net
The following patchset contain Netfilter fixes for your net tree, they are:
1) Fix a race in nfnetlink_log and nfnetlink_queue that can lead to a crash.
This problem is due to wrong order in the per-net registration and netlink
socket events. Patch from Francesco Ruggeri.
2) Make sure that counters that userspace pass us are higher than 0 in all the
x_tables frontends. Discovered via Trinity, patch from Dave Jones.
3) Revert a patch for br_netfilter to rely on the conntrack status bits. This
breaks stateless IPv6 NAT transformations. Patch from Florian Westphal.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
ip_error does not check if in_dev is NULL before dereferencing it.
IThe following sequence of calls is possible:
CPU A CPU B
ip_rcv_finish
ip_route_input_noref()
ip_route_input_slow()
inetdev_destroy()
dst_input()
With the result that a network device can be destroyed while processing
an input packet.
A crash was triggered with only unicast packets in flight, and
forwarding enabled on the only network device. The error condition
was created by the removal of the network device.
As such it is likely the that error code was -EHOSTUNREACH, and the
action taken by ip_error (if in_dev had been accessible) would have
been to not increment any counters and to have tried and likely failed
to send an icmp error as the network device is going away.
Therefore handle this weird case by just dropping the packet if
!in_dev. It will result in dropping the packet sooner, and will not
result in an actual change of behavior.
Fixes: 251da41301 ("ipv4: Cache ip_error() routes even when not forwarding.")
Reported-by: Vittorio Gambaletta <linuxbugs@vittgam.net>
Tested-by: Vittorio Gambaletta <linuxbugs@vittgam.net>
Signed-off-by: Vittorio Gambaletta <linuxbugs@vittgam.net>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Taking socket spinlock in tcp_get_info() can deadlock, as
inet_diag_dump_icsk() holds the &hashinfo->ehash_locks[i],
while packet processing can use the reverse locking order.
We could avoid this locking for TCP_LISTEN states, but lockdep would
certainly get confused as all TCP sockets share same lockdep classes.
[ 523.722504] ======================================================
[ 523.728706] [ INFO: possible circular locking dependency detected ]
[ 523.734990] 4.1.0-dbg-DEV #1676 Not tainted
[ 523.739202] -------------------------------------------------------
[ 523.745474] ss/18032 is trying to acquire lock:
[ 523.750002] (slock-AF_INET){+.-...}, at: [<ffffffff81669d44>] tcp_get_info+0x2c4/0x360
[ 523.758129]
[ 523.758129] but task is already holding lock:
[ 523.763968] (&(&hashinfo->ehash_locks[i])->rlock){+.-...}, at: [<ffffffff816bcb75>] inet_diag_dump_icsk+0x1d5/0x6c0
[ 523.774661]
[ 523.774661] which lock already depends on the new lock.
[ 523.774661]
[ 523.782850]
[ 523.782850] the existing dependency chain (in reverse order) is:
[ 523.790326]
-> #1 (&(&hashinfo->ehash_locks[i])->rlock){+.-...}:
[ 523.796599] [<ffffffff811126bb>] lock_acquire+0xbb/0x270
[ 523.802565] [<ffffffff816f5868>] _raw_spin_lock+0x38/0x50
[ 523.808628] [<ffffffff81665af8>] __inet_hash_nolisten+0x78/0x110
[ 523.815273] [<ffffffff816819db>] tcp_v4_syn_recv_sock+0x24b/0x350
[ 523.822067] [<ffffffff81684d41>] tcp_check_req+0x3c1/0x500
[ 523.828199] [<ffffffff81682d09>] tcp_v4_do_rcv+0x239/0x3d0
[ 523.834331] [<ffffffff816842fe>] tcp_v4_rcv+0xa8e/0xc10
[ 523.840202] [<ffffffff81658fa3>] ip_local_deliver_finish+0x133/0x3e0
[ 523.847214] [<ffffffff81659a9a>] ip_local_deliver+0xaa/0xc0
[ 523.853440] [<ffffffff816593b8>] ip_rcv_finish+0x168/0x5c0
[ 523.859624] [<ffffffff81659db7>] ip_rcv+0x307/0x420
Lets use u64_sync infrastructure instead. As a bonus, 64bit
arches get optimized, as these are nop for them.
Fixes: 0df48c26d8 ("tcp: add tcpi_bytes_acked to tcp_info")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David 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: Marcelo Ricardo Leitner <mleitner@redhat.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
inet_csk_get_port() randomization effort tends to spread
sockets on all the available range (ip_local_port_range)
This is unfortunate because SO_REUSEADDR sockets have
less requirements than non SO_REUSEADDR ones.
If an application uses SO_REUSEADDR hint, it is to try to
allow source ports being shared.
So instead of picking a random port number in ip_local_port_range,
lets try first in first half of the range.
This gives more chances to use upper half of the range for the
sockets with strong requirements (not using SO_REUSEADDR)
Note this patch does not add a new sysctl, and only changes
the way we try to pick port number.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Marcelo Ricardo Leitner <mleitner@redhat.com>
Cc: Flavio Leitner <fbl@redhat.com>
Acked-by: Flavio Leitner <fbl@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We no longer need bsocket atomic counter, as inet_csk_get_port()
calls bind_conflict() regardless of its value, after commit
2b05ad33e1 ("tcp: bind() fix autoselection to share ports")
This patch removes overhead of maintaining this counter and
double inet_csk_get_port() calls under pressure.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Marcelo Ricardo Leitner <mleitner@redhat.com>
Cc: Flavio Leitner <fbl@redhat.com>
Acked-by: Flavio Leitner <fbl@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We currently rely on the setting of SOCK_NOSPACE in the write()
path to ensure that we wake up any epoll edge trigger waiters when
acks return to free space in the write queue. However, if we fail
to allocate even a single skb in the write queue, we could end up
waiting indefinitely.
Fix this by explicitly issuing a wakeup when we detect the condition
of an empty write queue and a return value of -EAGAIN. This allows
userspace to re-try as we expect this to be a temporary failure.
I've tested this approach by artificially making
sk_stream_alloc_skb() return NULL periodically. In that case,
epoll edge trigger waiters will hang indefinitely in epoll_wait()
without this patch.
Signed-off-by: Jason Baron <jbaron@akamai.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In commit 8e4d980ac2 ("tcp: fix behavior for epoll edge trigger")
we fixed a possible hang of TCP sockets under memory pressure,
by allowing sk_stream_alloc_skb() to use sk_forced_mem_schedule()
if no packet is in socket write queue.
It turns out there are other cases where we want to force memory
schedule :
tcp_fragment() & tso_fragment() need to split a big TSO packet into
two smaller ones. If we block here because of TCP memory pressure,
we can effectively block TCP socket from sending new data.
If no further ACK is coming, this hang would be definitive, and socket
has no chance to effectively reduce its memory usage.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After improving setsockopt() coverage in trinity, I started triggering
vmalloc failures pretty reliably from this code path:
warn_alloc_failed+0xe9/0x140
__vmalloc_node_range+0x1be/0x270
vzalloc+0x4b/0x50
__do_replace+0x52/0x260 [ip_tables]
do_ipt_set_ctl+0x15d/0x1d0 [ip_tables]
nf_setsockopt+0x65/0x90
ip_setsockopt+0x61/0xa0
raw_setsockopt+0x16/0x60
sock_common_setsockopt+0x14/0x20
SyS_setsockopt+0x71/0xd0
It turns out we don't validate that the num_counters field in the
struct we pass in from userspace is initialized.
The same problem also exists in ebtables, arptables, ipv6, and the
compat variants.
Signed-off-by: Dave Jones <davej@codemonkey.org.uk>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
This work as a follow-up of commit f7b3bec6f5 ("net: allow setting ecn
via routing table") and adds RFC3168 section 6.1.1.1. fallback for outgoing
ECN connections. In other words, this work adds a retry with a non-ECN
setup SYN packet, as suggested from the RFC on the first timeout:
[...] A host that receives no reply to an ECN-setup SYN within the
normal SYN retransmission timeout interval MAY resend the SYN and
any subsequent SYN retransmissions with CWR and ECE cleared. [...]
Schematic client-side view when assuming the server is in tcp_ecn=2 mode,
that is, Linux default since 2009 via commit 255cac91c3 ("tcp: extend
ECN sysctl to allow server-side only ECN"):
1) Normal ECN-capable path:
SYN ECE CWR ----->
<----- SYN ACK ECE
ACK ----->
2) Path with broken middlebox, when client has fallback:
SYN ECE CWR ----X crappy middlebox drops packet
(timeout, rtx)
SYN ----->
<----- SYN ACK
ACK ----->
In case we would not have the fallback implemented, the middlebox drop
point would basically end up as:
SYN ECE CWR ----X crappy middlebox drops packet
(timeout, rtx)
SYN ECE CWR ----X crappy middlebox drops packet
(timeout, rtx)
SYN ECE CWR ----X crappy middlebox drops packet
(timeout, rtx)
In any case, it's rather a smaller percentage of sites where there would
occur such additional setup latency: it was found in end of 2014 that ~56%
of IPv4 and 65% of IPv6 servers of Alexa 1 million list would negotiate
ECN (aka tcp_ecn=2 default), 0.42% of these webservers will fail to connect
when trying to negotiate with ECN (tcp_ecn=1) due to timeouts, which the
fallback would mitigate with a slight latency trade-off. Recent related
paper on this topic:
Brian Trammell, Mirja Kühlewind, Damiano Boppart, Iain Learmonth,
Gorry Fairhurst, and Richard Scheffenegger:
"Enabling Internet-Wide Deployment of Explicit Congestion Notification."
Proc. PAM 2015, New York.
http://ecn.ethz.ch/ecn-pam15.pdf
Thus, when net.ipv4.tcp_ecn=1 is being set, the patch will perform RFC3168,
section 6.1.1.1. fallback on timeout. For users explicitly not wanting this
which can be in DC use case, we add a net.ipv4.tcp_ecn_fallback knob that
allows for disabling the fallback.
tp->ecn_flags are not being cleared in tcp_ecn_clear_syn() on output, but
rather we let tcp_ecn_rcv_synack() take that over on input path in case a
SYN ACK ECE was delayed. Thus a spurious SYN retransmission will not prevent
ECN being negotiated eventually in that case.
Reference: https://www.ietf.org/proceedings/92/slides/slides-92-iccrg-1.pdf
Reference: https://www.ietf.org/proceedings/89/slides/slides-89-tsvarea-1.pdf
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Mirja Kühlewind <mirja.kuehlewind@tik.ee.ethz.ch>
Signed-off-by: Brian Trammell <trammell@tik.ee.ethz.ch>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Dave That <dave.taht@gmail.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After sending the new data packets to probe (step 2), F-RTO may
incorrectly send more probes if the next ACK advances SND_UNA and
does not sack new packet. However F-RTO RFC 5682 probes at most
once. This bug may cause sender to always send new data instead of
repairing holes, inducing longer HoL blocking on the receiver for
the application.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Undo based on TCP timestamps should only happen on ACKs that advance
SND_UNA, according to the Eifel algorithm in RFC 3522:
Section 3.2:
(4) If the value of the Timestamp Echo Reply field of the
acceptable ACK's Timestamps option is smaller than the
value of RetransmitTS, then proceed to step (5),
Section Terminology:
We use the term 'acceptable ACK' as defined in [RFC793]. That is an
ACK that acknowledges previously unacknowledged data.
This is because upon receiving an out-of-order packet, the receiver
returns the last timestamp that advances RCV_NXT, not the current
timestamp of the packet in the DUPACK. Without checking the flag,
the DUPACK will cause tcp_packet_delayed() to return true and
tcp_try_undo_loss() will revert cwnd reduction.
Note that we check the condition in CA_Recovery already by only
calling tcp_try_undo_partial() if FLAG_SND_UNA_ADVANCED is set or
tcp_try_undo_recovery() if snd_una crosses high_seq.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently the getsockopt() requesting the cached contents of the syn
packet headers will fail silently if the caller uses a buffer that is
too small to contain the requested data. Rather than fail silently and
discard the headers, getsockopt() should return an error and report the
required size to hold the data.
Signed-off-by: Eric B Munson <emunson@akamai.com>
Cc: Eric Dumazet <edumazet@google.com>
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: netdev@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
When bridge netfilter re-fragments an IP packet for output, all
packets that can not be re-fragmented to their original input size
should be silently discarded.
However, current bridge netfilter output path generates an ICMP packet
with 'size exceeded MTU' message for such packets, this is a bug.
This patch refactors the ip_fragment() API to allow two separate
use cases. The bridge netfilter user case will not
send ICMP, the routing output will, as before.
Signed-off-by: Andy Zhou <azhou@nicira.com>
Acked-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
users in [IP_DEFRAG_CONNTRACK_BRIDGE_IN, __IP_DEFRAG_CONNTRACK_BR_IN]
should not ICMP message also.
Reported-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Andy Zhou <azhou@nicira.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Improve readability of skip ICMP for de-fragmentation expiration logic.
This change will also make the logic easier to maintain when the
following patches in this series are applied.
Signed-off-by: Andy Zhou <azhou@nicira.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pablo Neira Ayuso says:
====================
Netfilter updates for net-next
The following patchset contains Netfilter updates for net-next. Briefly
speaking, cleanups and minor fixes for ipset from Jozsef Kadlecsik and
Serget Popovich, more incremental updates to make br_netfilter a better
place from Florian Westphal, ARP support to the x_tables mark match /
target from and context Zhang Chunyu and the addition of context to know
that the x_tables runs through nft_compat. More specifically, they are:
1) Fix sparse warning in ipset/ip_set_hash_ipmark.c when fetching the
IPSET_ATTR_MARK netlink attribute, from Jozsef Kadlecsik.
2) Rename STREQ macro to STRNCMP in ipset, also from Jozsef.
3) Use skb->network_header to calculate the transport offset in
ip_set_get_ip{4,6}_port(). From Alexander Drozdov.
4) Reduce memory consumption per element due to size miscalculation,
this patch and follow up patches from Sergey Popovich.
5) Expand nomatch field from 1 bit to 8 bits to allow to simplify
mtype_data_reset_flags(), also from Sergey.
6) Small clean for ipset macro trickery.
7) Fix error reporting when both ip_set_get_hostipaddr4() and
ip_set_get_extensions() from per-set uadt functions.
8) Simplify IPSET_ATTR_PORT netlink attribute validation.
9) Introduce HOST_MASK instead of hardcoded 32 in ipset.
10) Return true/false instead of 0/1 in functions that return boolean
in the ipset code.
11) Validate maximum length of the IPSET_ATTR_COMMENT netlink attribute.
12) Allow to dereference from ext_*() ipset macros.
13) Get rid of incorrect definitions of HKEY_DATALEN.
14) Include linux/netfilter/ipset/ip_set.h in the x_tables set match.
15) Reduce nf_bridge_info size in br_netfilter, from Florian Westphal.
16) Release nf_bridge_info after POSTROUTING since this is only needed
from the physdev match, also from Florian.
17) Reduce size of ipset code by deinlining ip_set_put_extensions(),
from Denys Vlasenko.
18) Oneliner to add ARP support to the x_tables mark match/target, from
Zhang Chunyu.
19) Add context to know if the x_tables extension runs from nft_compat,
to address minor problems with three existing extensions.
20) Correct return value in several seqfile *_show() functions in the
netfilter tree, from Joe Perches.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
commit 1d13a96c74 ("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: 1d13a96c74 ("ipv6: tcp: fix flowlabel value in ACK messages")
Reported-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Florent Fourcot <florent.fourcot@enst-bretagne.fr>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Allowing tcp to use ~19% of physical memory is way too much,
and allowed bugs to be hidden. Add to this that some drivers use a full
page per incoming frame, so real cost can be twice the advertized one.
Reduce tcp_mem by 50 % as a first step to sanity.
tcp_mem[0,1,2] defaults are now 4.68%, 6.25%, 9.37% of physical memory.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
While testing tight tcp_mem settings, I found tcp sessions could be
stuck because we do not allow even one skb to be received on them.
By allowing one skb to be received, we introduce fairness and
eventuallu force memory hogs to release their allocation.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Under memory pressure, tcp_sendmsg() can fail to queue a packet
while no packet is present in write queue. If we return -EAGAIN
with no packet in write queue, no ACK packet will ever come
to raise EPOLLOUT.
We need to allow one skb per TCP socket, and make sure that
tcp sockets can release their forward allocations under pressure.
This is a followup to commit 790ba4566c ("tcp: set SOCK_NOSPACE
under memory pressure")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Introduce an optimized version of sk_under_memory_pressure()
for TCP. Our intent is to use it in fast paths.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We plan to use sk_forced_wmem_schedule() in input path as well,
so make it non static and rename it.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix verbose sparse errors :
make C=2 CF=-D__CHECK_ENDIAN__ net/ipv4/netfilter/ipt_SYNPROXY.o
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
make C=2 CF=-D__CHECK_ENDIAN__ net/ipv4/ipip.o
CHECK net/ipv4/ipip.c
net/ipv4/ipip.c:254:27: warning: incorrect type in assignment (different base types)
net/ipv4/ipip.c:254:27: expected restricted __be32 [addressable] [usertype] o_key
net/ipv4/ipip.c:254:27: got restricted __be16 [addressable] [usertype] i_flags
Fixes: 3b7b514f44 ("ipip: fix a regression in ioctl")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, we have four xtables extensions that cannot be used from the
xt over nft compat layer. The problem is that they need real access to
the full blown xt_entry to validate that the rule comes with the right
dependencies. This check was introduced to overcome the lack of
sufficient userspace dependency validation in iptables.
To resolve this problem, this patch introduces a new field to the
xt_tgchk_param structure that tell us if the extension is run from
nft_compat context.
The three affected extensions are:
1) CLUSTERIP, this target has been superseded by xt_cluster. So just
bail out by returning -EINVAL.
2) TCPMSS. Relax the checking when used from nft_compat. If used with
the wrong configuration, it will corrupt !syn packets by adding TCP
MSS option.
3) ebt_stp. Relax the check to make sure it uses the reserved
destination MAC address for STP.
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Tested-by: Arturo Borrero Gonzalez <arturo.borrero.glez@gmail.com>
Wei Liu
Eric Dumazet
David S. Miller
Neal Cardwell
Sorin Dumitru
Alexander Duyck
Florian Westphal
Hannes Frederic Sowa
Michal Kubeček
Eric W. Biederman
Marcelo Ricardo Leitner
Jason Baron
Dave Jones
Daniel Borkmann
Yuchung Cheng
Eric B Munson
Andy Zhou
Florent Fourcot
Pablo Neira Ayuso