After commit c9eeec26e3 ("tcp: TSQ can use a dynamic limit"), several
users reported throughput regressions, notably on mvneta and wifi
adapters.
802.11 AMPDU requires a fair amount of queueing to be effective.
This patch partially reverts the change done in tcp_write_xmit()
so that the minimal amount is sysctl_tcp_limit_output_bytes.
It also remove the use of this sysctl while building skb stored
in write queue, as TSO autosizing does the right thing anyway.
Users with well behaving NICS and correct qdisc (like sch_fq),
can then lower the default sysctl_tcp_limit_output_bytes value from
128KB to 8KB.
This new usage of sysctl_tcp_limit_output_bytes permits each driver
authors to check how their driver performs when/if the value is set
to a minimum of 4KB.
Normally, line rate for a single TCP flow should be possible,
but some drivers rely on timers to perform TX completion and
too long TX completion delays prevent reaching full throughput.
Fixes: c9eeec26e3 ("tcp: TSQ can use a dynamic limit")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Sujith Manoharan <sujith@msujith.org>
Reported-by: Arnaud Ebalard <arno@natisbad.org>
Tested-by: Sujith Manoharan <sujith@msujith.org>
Cc: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Slow start now increases cwnd by 1 if an ACK acknowledges some packets,
regardless the number of packets. Consequently slow start performance
is highly dependent on the degree of the stretch ACKs caused by
receiver or network ACK compression mechanisms (e.g., delayed-ACK,
GRO, etc). But slow start algorithm is to send twice the amount of
packets of packets left so it should process a stretch ACK of degree
N as if N ACKs of degree 1, then exits when cwnd exceeds ssthresh. A
follow up patch will use the remainder of the N (if greater than 1)
to adjust cwnd in the congestion avoidance phase.
In addition this patch retires the experimental limited slow start
(LSS) feature. LSS has multiple drawbacks but questionable benefit. The
fractional cwnd increase in LSS requires a loop in slow start even
though it's rarely used. Configuring such an increase step via a global
sysctl on different BDPS seems hard. Finally and most importantly the
slow start overshoot concern is now better covered by the Hybrid slow
start (hystart) enabled by default.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Applications have started to use Fast Open (e.g., Chrome browser has
such an optional flag) and the feature has gone through several
generations of kernels since 3.7 with many real network tests. It's
time to enable this flag by default for applications to test more
conveniently and extensively.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Document force_mld_version parameter in ip-sysctl.txt.
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Cc: Hannes Frederic Sowa <hannes@stressinduktion.org>
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
After hearing many people over past years complaining against TSO being
bursty or even buggy, we are proud to present automatic sizing of TSO
packets.
One part of the problem is that tcp_tso_should_defer() uses an heuristic
relying on upcoming ACKS instead of a timer, but more generally, having
big TSO packets makes little sense for low rates, as it tends to create
micro bursts on the network, and general consensus is to reduce the
buffering amount.
This patch introduces a per socket sk_pacing_rate, that approximates
the current sending rate, and allows us to size the TSO packets so
that we try to send one packet every ms.
This field could be set by other transports.
Patch has no impact for high speed flows, where having large TSO packets
makes sense to reach line rate.
For other flows, this helps better packet scheduling and ACK clocking.
This patch increases performance of TCP flows in lossy environments.
A new sysctl (tcp_min_tso_segs) is added, to specify the
minimal size of a TSO packet (default being 2).
A follow-up patch will provide a new packet scheduler (FQ), using
sk_pacing_rate as an input to perform optional per flow pacing.
This explains why we chose to set sk_pacing_rate to twice the current
rate, allowing 'slow start' ramp up.
sk_pacing_rate = 2 * cwnd * mss / srtt
v2: Neal Cardwell reported a suspect deferring of last two segments on
initial write of 10 MSS, I had to change tcp_tso_should_defer() to take
into account tp->xmit_size_goal_segs
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Van Jacobson <vanj@google.com>
Cc: Tom Herbert <therbert@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch implements RFC6980: Drop fragmented ndisc packets by
default. If a fragmented ndisc packet is received the user is informed
that it is possible to disable the check.
Cc: Fernando Gont <fernando@gont.com.ar>
Cc: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit cab70040df ("net: igmp:
Reduce Unsolicited report interval to 1s when using IGMPv3") and
2690048c01 ("net: igmp: Allow user-space
configuration of igmp unsolicited report interval") by William Manley made
igmp unsolicited report intervals configurable per interface and corrected
the interval of unsolicited igmpv3 report messages resendings to 1s.
Same needs to be done for IPv6:
MLDv1 (RFC2710 7.10.): 10 seconds
MLDv2 (RFC3810 9.11.): 1 second
Both intervals are configurable via new procfs knobs
mldv1_unsolicited_report_interval and mldv2_unsolicited_report_interval.
(also added .force_mld_version to ipv6_devconf_dflt to bring structs in
line without semantic changes)
v2:
a) Joined documentation update for IPv4 and IPv6 MLD/IGMP
unsolicited_report_interval procfs knobs.
b) incorporate stylistic feedback from William Manley
v3:
a) add new DEVCONF_* values to the end of the enum (thanks to David
Miller)
Cc: Cong Wang <xiyou.wangcong@gmail.com>
Cc: William Manley <william.manley@youview.com>
Cc: Benjamin LaHaise <bcrl@kvack.org>
Cc: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Get rid of the last module parameter for SCTP and make this
configurable via sysctl for SCTP like all the rest of SCTP's
configuration knobs.
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
| If you want to test which effects syncookies have to your
| network connections you can set this knob to 2 to enable
| unconditionally generation of syncookies.
Original idea and first implementation by Eric Dumazet.
Cc: Florian Westphal <fw@strlen.de>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Idea of this patch is to add optional limitation of number of
unsent bytes in TCP sockets, to reduce usage of kernel memory.
TCP receiver might announce a big window, and TCP sender autotuning
might allow a large amount of bytes in write queue, but this has little
performance impact if a large part of this buffering is wasted :
Write queue needs to be large only to deal with large BDP, not
necessarily to cope with scheduling delays (incoming ACKS make room
for the application to queue more bytes)
For most workloads, using a value of 128 KB or less is OK to give
applications enough time to react to POLLOUT events in time
(or being awaken in a blocking sendmsg())
This patch adds two ways to set the limit :
1) Per socket option TCP_NOTSENT_LOWAT
2) A sysctl (/proc/sys/net/ipv4/tcp_notsent_lowat) for sockets
not using TCP_NOTSENT_LOWAT socket option (or setting a zero value)
Default value being UINT_MAX (0xFFFFFFFF), meaning this has no effect.
This changes poll()/select()/epoll() to report POLLOUT
only if number of unsent bytes is below tp->nosent_lowat
Note this might increase number of sendmsg()/sendfile() calls
when using non blocking sockets,
and increase number of context switches for blocking sockets.
Note this is not related to SO_SNDLOWAT (as SO_SNDLOWAT is
defined as :
Specify the minimum number of bytes in the buffer until
the socket layer will pass the data to the protocol)
Tested:
netperf sessions, and watching /proc/net/protocols "memory" column for TCP
With 200 concurrent netperf -t TCP_STREAM sessions, amount of kernel memory
used by TCP buffers shrinks by ~55 % (20567 pages instead of 45458)
lpq83:~# echo -1 >/proc/sys/net/ipv4/tcp_notsent_lowat
lpq83:~# (super_netperf 200 -t TCP_STREAM -H remote -l 90 &); sleep 60 ; grep TCP /proc/net/protocols
TCPv6 1880 2 45458 no 208 yes ipv6 y y y y y y y y y y y y y n y y y y y
TCP 1696 508 45458 no 208 yes kernel y y y y y y y y y y y y y n y y y y y
lpq83:~# echo 131072 >/proc/sys/net/ipv4/tcp_notsent_lowat
lpq83:~# (super_netperf 200 -t TCP_STREAM -H remote -l 90 &); sleep 60 ; grep TCP /proc/net/protocols
TCPv6 1880 2 20567 no 208 yes ipv6 y y y y y y y y y y y y y n y y y y y
TCP 1696 508 20567 no 208 yes kernel y y y y y y y y y y y y y n y y y y y
Using 128KB has no bad effect on the throughput or cpu usage
of a single flow, although there is an increase of context switches.
A bonus is that we hold socket lock for a shorter amount
of time and should improve latencies of ACK processing.
lpq83:~# echo -1 >/proc/sys/net/ipv4/tcp_notsent_lowat
lpq83:~# perf stat -e context-switches ./netperf -H 7.7.7.84 -t omni -l 20 -c -i10,3
OMNI Send TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 7.7.7.84 () port 0 AF_INET : +/-2.500% @ 99% conf.
Local Remote Local Elapsed Throughput Throughput Local Local Remote Remote Local Remote Service
Send Socket Recv Socket Send Time Units CPU CPU CPU CPU Service Service Demand
Size Size Size (sec) Util Util Util Util Demand Demand Units
Final Final % Method % Method
1651584 6291456 16384 20.00 17447.90 10^6bits/s 3.13 S -1.00 U 0.353 -1.000 usec/KB
Performance counter stats for './netperf -H 7.7.7.84 -t omni -l 20 -c -i10,3':
412,514 context-switches
200.034645535 seconds time elapsed
lpq83:~# echo 131072 >/proc/sys/net/ipv4/tcp_notsent_lowat
lpq83:~# perf stat -e context-switches ./netperf -H 7.7.7.84 -t omni -l 20 -c -i10,3
OMNI Send TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 7.7.7.84 () port 0 AF_INET : +/-2.500% @ 99% conf.
Local Remote Local Elapsed Throughput Throughput Local Local Remote Remote Local Remote Service
Send Socket Recv Socket Send Time Units CPU CPU CPU CPU Service Service Demand
Size Size Size (sec) Util Util Util Util Demand Demand Units
Final Final % Method % Method
1593240 6291456 16384 20.00 17321.16 10^6bits/s 3.35 S -1.00 U 0.381 -1.000 usec/KB
Performance counter stats for './netperf -H 7.7.7.84 -t omni -l 20 -c -i10,3':
2,675,818 context-switches
200.029651391 seconds time elapsed
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Acked-By: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/ethernet/freescale/fec_main.c
drivers/net/ethernet/renesas/sh_eth.c
net/ipv4/gre.c
The GRE conflict is between a bug fix (kfree_skb --> kfree_skb_list)
and the splitting of the gre.c code into seperate files.
The FEC conflict was two sets of changes adding ethtool support code
in an "!CONFIG_M5272" CPP protected block.
Finally the sh_eth.c conflict was between one commit add bits set
in the .eesr_err_check mask whilst another commit removed the
.tx_error_check member and assignments.
Signed-off-by: David S. Miller <davem@davemloft.net>
syncookies is on for default since in commit e994b7c901
(tcp: Don't make syn cookies initial setting depend on CONFIG_SYSCTL).
And fix a typo of CONFIG_SYN_COOKIES.
Signed-off-by: Shan Wei <davidshan@tencent.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
commit 6648bd7e0e (ipv4: Add sysctl knob to control
early socket demux) introduced such sysctl, but forgot to add
doc into Documentation/networking/ip-sysctl.txt. This patch adds it.
Basically I grab the doc from the description of commit 41063e9dd1
(ipv4: Early TCP socket demux.) and the above commit.
Cc: Eric Dumazet <edumazet@google.com>
Cc: Alexander Duyck <alexander.h.duyck@intel.com>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Cong Wang <amwang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch fixes icmp_ignore_bogus_error_responses default value to be 1 instead
of FALSE. It is initialized to 1 in icmp_sk_init().
Signed-off-by: Rami Rosen <ramirose@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch implements F-RTO (foward RTO recovery):
When the first retransmission after timeout is acknowledged, F-RTO
sends new data instead of old data. If the next ACK acknowledges
some never-retransmitted data, then the timeout was spurious and the
congestion state is reverted. Otherwise if the next ACK selectively
acknowledges the new data, then the timeout was genuine and the
loss recovery continues. This idea applies to recurring timeouts
as well. While F-RTO sends different data during timeout recovery,
it does not (and should not) change the congestion control.
The implementaion follows the three steps of SACK enhanced algorithm
(section 3) in RFC5682. Step 1 is in tcp_enter_loss(). Step 2 and
3 are in tcp_process_loss(). The basic version is not supported
because SACK enhanced version also works for non-SACK connections.
The new implementation is functionally in parity with the old F-RTO
implementation except the one case where it increases undo events:
In addition to the RFC algorithm, a spurious timeout may be detected
without sending data in step 2, as long as the SACK confirms not
all the original data are dropped. When this happens, the sender
will undo the cwnd and perhaps enter fast recovery instead. This
additional check increases the F-RTO undo events by 5x compared
to the prior implementation on Google Web servers, since the sender
often does not have new data to send for HTTP.
Note F-RTO may detect spurious timeout before Eifel with timestamps
does so.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The patch series refactor the F-RTO feature (RFC4138/5682).
This is to simplify the loss recovery processing. Existing F-RTO
was developed during the experimental stage (RFC4138) and has
many experimental features. It takes a separate code path from
the traditional timeout processing by overloading CA_Disorder
instead of using CA_Loss state. This complicates CA_Disorder state
handling because it's also used for handling dubious ACKs and undos.
While the algorithm in the RFC does not change the congestion control,
the implementation intercepts congestion control in various places
(e.g., frto_cwnd in tcp_ack()).
The new code implements newer F-RTO RFC5682 using CA_Loss processing
path. F-RTO becomes a small extension in the timeout processing
and interfaces with congestion control and Eifel undo modules.
It lets congestion control (module) determines how many to send
independently. F-RTO only chooses what to send in order to detect
spurious retranmission. If timeout is found spurious it invokes
existing Eifel undo algorithms like DSACK or TCP timestamp based
detection.
The first patch removes all F-RTO code except the sysctl_tcp_frto is
left for the new implementation. Since CA_EVENT_FRTO is removed, TCP
westwood now computes ssthresh on regular timeout CA_EVENT_LOSS event.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
TCPCT uses option-number 253, reserved for experimental use and should
not be used in production environments.
Further, TCPCT does not fully implement RFC 6013.
As a nice side-effect, removing TCPCT increases TCP's performance for
very short flows:
Doing an apache-benchmark with -c 100 -n 100000, sending HTTP-requests
for files of 1KB size.
before this patch:
average (among 7 runs) of 20845.5 Requests/Second
after:
average (among 7 runs) of 21403.6 Requests/Second
Signed-off-by: Christoph Paasch <christoph.paasch@uclouvain.be>
Signed-off-by: David S. Miller <davem@davemloft.net>
The default value is 128, not 256
#grep gc_thresh1 net/ -rI
net/decnet/dn_neigh.c: .gc_thresh1 = 128,
net/ipv6/ndisc.c: .gc_thresh1 = 128,
net/ipv4/arp.c: .gc_thresh1 = 128,
Signed-off-by: Li RongQing <roy.qing.li@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch series implement the Tail loss probe (TLP) algorithm described
in http://tools.ietf.org/html/draft-dukkipati-tcpm-tcp-loss-probe-01. The
first patch implements the basic algorithm.
TLP's goal is to reduce tail latency of short transactions. It achieves
this by converting retransmission timeouts (RTOs) occuring due
to tail losses (losses at end of transactions) into fast recovery.
TLP transmits one packet in two round-trips when a connection is in
Open state and isn't receiving any ACKs. The transmitted packet, aka
loss probe, can be either new or a retransmission. When there is tail
loss, the ACK from a loss probe triggers FACK/early-retransmit based
fast recovery, thus avoiding a costly RTO. In the absence of loss,
there is no change in the connection state.
PTO stands for probe timeout. It is a timer event indicating
that an ACK is overdue and triggers a loss probe packet. The PTO value
is set to max(2*SRTT, 10ms) and is adjusted to account for delayed
ACK timer when there is only one oustanding packet.
TLP Algorithm
On transmission of new data in Open state:
-> packets_out > 1: schedule PTO in max(2*SRTT, 10ms).
-> packets_out == 1: schedule PTO in max(2*RTT, 1.5*RTT + 200ms)
-> PTO = min(PTO, RTO)
Conditions for scheduling PTO:
-> Connection is in Open state.
-> Connection is either cwnd limited or no new data to send.
-> Number of probes per tail loss episode is limited to one.
-> Connection is SACK enabled.
When PTO fires:
new_segment_exists:
-> transmit new segment.
-> packets_out++. cwnd remains same.
no_new_packet:
-> retransmit the last segment.
Its ACK triggers FACK or early retransmit based recovery.
ACK path:
-> rearm RTO at start of ACK processing.
-> reschedule PTO if need be.
In addition, the patch includes a small variation to the Early Retransmit
(ER) algorithm, such that ER and TLP together can in principle recover any
N-degree of tail loss through fast recovery. TLP is controlled by the same
sysctl as ER, tcp_early_retrans sysctl.
tcp_early_retrans==0; disables TLP and ER.
==1; enables RFC5827 ER.
==2; delayed ER.
==3; TLP and delayed ER. [DEFAULT]
==4; TLP only.
The TLP patch series have been extensively tested on Google Web servers.
It is most effective for short Web trasactions, where it reduced RTOs by 15%
and improved HTTP response time (average by 6%, 99th percentile by 10%).
The transmitted probes account for <0.5% of the overall transmissions.
Signed-off-by: Nandita Dukkipati <nanditad@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
TCP Appropriate Byte Count was added by me, but later disabled.
There is no point in maintaining it since it is a potential source
of bugs and Linux already implements other better window protection
heuristics.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since we have removed NCE (Neighbour Cache Entry) reference from
routing entries, the only refcnt holders of an NCE are its timer
(if running) and its owner table, in usual cases. As a result,
neigh_periodic_work() purges NCEs over and over again even for
gateways.
It does not make sense to purge entries, if number of them is
very small, so keep them. The minimum number of entries to keep
is specified by gc_thresh1.
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
Documentation/networking/ip-sysctl.txt
drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
Both conflicts were simply overlapping context.
A build fix for qlcnic is in here too, simply removing the added
devinit annotations which no longer exist.
Signed-off-by: David S. Miller <davem@davemloft.net>
Recent commit (commit 7e3a2dc529 doc: make the description of how tcp_ecn
works more explicit and clear ) clarified the behavior of tcp_ecn sysctl
variable but description is inconsistent. When requested by incoming conections,
ECN is enabled with not just tcp_ecn = 2 but also with tcp_ecn = 1.
This patch makes it clear that with tcp_ecn = 1, ECN is enabled when requested
by incoming connections.
Also fix spelling of 'incoming'.
Signed-off-by: Vijay Subramanian <subramanian.vijay@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
I slipped in a new sysctl without proper documentation. I would like to
make up for this now.
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The description for tcp_fin_timeout should be tigher and more clear.
In addition to being tighter, we should make the spelling of the
state name consistent with what utilities report, remove the now
dated reference to 2.2 and put the default in the consistent place.
Signed-off-by: Rick Jones <rick.jones2@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
'secluded' is used to describe places, not suitable here.
Suggested-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: Shan Wei <davidshan@tencent.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Make the description of how tcp_ecn works a bit more explicit and clear.
Signed-off-by: Rick Jones <rick.jones2@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently sctp allows for the optional use of md5 of sha1 hmac algorithms to
generate cookie values when establishing new connections via two build time
config options. Theres no real reason to make this a static selection. We can
add a sysctl that allows for the dynamic selection of these algorithms at run
time, with the default value determined by the corresponding crypto library
availability.
This comes in handy when, for example running a system in FIPS mode, where use
of md5 is disallowed, but SHA1 is permitted.
Note: This new sysctl has no corresponding socket option to select the cookie
hmac algorithm. I chose not to implement that intentionally, as RFC 6458
contains no option for this value, and I opted not to pollute the socket option
namespace.
Change notes:
v2)
* Updated subject to have the proper sctp prefix as per Dave M.
* Replaced deafult selection options with new options that allow
developers to explicitly select available hmac algs at build time
as per suggestion by Vlad Y.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
CC: Vlad Yasevich <vyasevich@gmail.com>
CC: "David S. Miller" <davem@davemloft.net>
CC: netdev@vger.kernel.org
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds all the necessary data structure and support
functions to implement TFO server side. It also documents a number
of flags for the sysctl_tcp_fastopen knob, and adds a few Linux
extension MIBs.
In addition, it includes the following:
1. a new TCP_FASTOPEN socket option an application must call to
supply a max backlog allowed in order to enable TFO on its listener.
2. A number of key data structures:
"fastopen_rsk" in tcp_sock - for a big socket to access its
request_sock for retransmission and ack processing purpose. It is
non-NULL iff 3WHS not completed.
"fastopenq" in request_sock_queue - points to a per Fast Open
listener data structure "fastopen_queue" to keep track of qlen (# of
outstanding Fast Open requests) and max_qlen, among other things.
"listener" in tcp_request_sock - to point to the original listener
for book-keeping purpose, i.e., to maintain qlen against max_qlen
as part of defense against IP spoofing attack.
3. various data structure and functions, many in tcp_fastopen.c, to
support server side Fast Open cookie operations, including
/proc/sys/net/ipv4/tcp_fastopen_key to allow manual rekeying.
Signed-off-by: H.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: David S. Miller <davem@davemloft.net>
Commit 9ad7c049 ("tcp: RFC2988bis + taking RTT sample from 3WHS for
the passive open side") changed the initRTO from 3secs to 1sec in
accordance to RFC6298 (former RFC2988bis). This reduced the time till
the last SYN retransmission packet gets sent from 93secs to 31secs.
RFC1122 is stating that the retransmission should be done for at least 3
minutes, but this seems to be quite high.
"However, the values of R1 and R2 may be different for SYN
and data segments. In particular, R2 for a SYN segment MUST
be set large enough to provide retransmission of the segment
for at least 3 minutes. The application can close the
connection (i.e., give up on the open attempt) sooner, of
course."
This patch increases the value of TCP_SYN_RETRIES to the value of 6,
providing a retransmission window of 63secs.
The comments for SYN and SYNACK retries have also been updated to
describe the current settings. The same goes for the documentation file
"Documentation/networking/ip-sysctl.txt".
Signed-off-by: Alexander Bergmann <alex@linlab.net>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After IP route cache removal, rt_cache_rebuild_count is no longer
used.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
I've seen several attempts recently made to do quick failover of sctp transports
by reducing various retransmit timers and counters. While its possible to
implement a faster failover on multihomed sctp associations, its not
particularly robust, in that it can lead to unneeded retransmits, as well as
false connection failures due to intermittent latency on a network.
Instead, lets implement the new ietf quick failover draft found here:
http://tools.ietf.org/html/draft-nishida-tsvwg-sctp-failover-05
This will let the sctp stack identify transports that have had a small number of
errors, and avoid using them quickly until their reliability can be
re-established. I've tested this out on two virt guests connected via multiple
isolated virt networks and believe its in compliance with the above draft and
works well.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
CC: Vlad Yasevich <vyasevich@gmail.com>
CC: Sridhar Samudrala <sri@us.ibm.com>
CC: "David S. Miller" <davem@davemloft.net>
CC: linux-sctp@vger.kernel.org
CC: joe@perches.com
Acked-by: Vlad Yasevich <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In trusted networks, e.g., intranet, data-center, the client does not
need to use Fast Open cookie to mitigate DoS attacks. In cookie-less
mode, sendmsg() with MSG_FASTOPEN flag will send SYN-data regardless
of cookie availability.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David 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: Yuchung Cheng <ycheng@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Implement the RFC 5691 mitigation against Blind
Reset attack using RST bit.
Idea is to validate incoming RST sequence,
to match RCV.NXT value, instead of previouly accepted
window : (RCV.NXT <= SEG.SEQ < RCV.NXT+RCV.WND)
If sequence is in window but not an exact match, send
a "challenge ACK", so that the other part can resend an
RST with the appropriate sequence.
Add a new sysctl, tcp_challenge_ack_limit, to limit
number of challenge ACK sent per second.
Add a new SNMP counter to count number of challenge acks sent.
(netstat -s | grep TCPChallengeACK)
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Kiran Kumar Kella <kkiran@broadcom.com>
Signed-off-by: David 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: Eric 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: David S. Miller <davem@davemloft.net>
Routing of 127/8 is tradtionally forbidden, we consider
packets from that address block martian when routing and do
not process corresponding ARP requests.
This is a sane default but renders a huge address space
practically unuseable.
The RFC states that no address within the 127/8 block should
ever appear on any network anywhere but it does not forbid
the use of such addresses outside of the loopback device in
particular. For example to address a pool of virtual guests
behind a load balancer.
This patch adds a new interface option 'route_localnet'
enabling routing of the 127/8 address block and processing
of ARP requests on a specific interface.
Note that for the feature to work, the default local route
covering 127/8 dev lo needs to be removed.
Example:
$ sysctl -w net.ipv4.conf.eth0.route_localnet=1
$ ip route del 127.0.0.0/8 dev lo table local
$ ip addr add 127.1.0.1/16 dev eth0
$ ip route flush cache
V2: Fix invalid check to auto flush cache (thanks davem)
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
if net.bridge.bridge-nf-filter-vlan-tagged sysctl is enabled, bridge
netfilter removes the vlan header temporarily and then feeds the packet
to ip(6)tables.
When the new "bridge-nf-pass-vlan-input-device" sysctl is on
(default off), then bridge netfilter will also set the
in-interface to the vlan interface; if such an interface exists.
This is needed to make iptables REDIRECT target work with
"vlan-on-top-of-bridge" setups and to allow use of "iptables -i" to
match the vlan device name.
Also update Documentation with current brnf default settings.
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Bart De Schuymer <bdschuym@pandora.be>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Conflicts:
drivers/net/ethernet/intel/e1000e/param.c
drivers/net/wireless/iwlwifi/iwl-agn-rx.c
drivers/net/wireless/iwlwifi/iwl-trans-pcie-rx.c
drivers/net/wireless/iwlwifi/iwl-trans.h
Resolved the iwlwifi conflict with mainline using 3-way diff posted
by John Linville and Stephen Rothwell. In 'net' we added a bug
fix to make iwlwifi report a more accurate skb->truesize but this
conflicted with RX path changes that happened meanwhile in net-next.
In e1000e a conflict arose in the validation code for settings of
adapter->itr. 'net-next' had more sophisticated logic so that
logic was used.
Signed-off-by: David 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: Eric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Tom Herbert <therbert@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David 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: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
All parameter descriptions in /proc/sys/net/core/* now is separated
two places. So, merge them into Documentation/sysctl/net.txt.
Signed-off-by: Shan Wei <davidshan@tencent.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The explanation of ip_local_port_range in
Documentation/networking/ip-sysctl.txt contains several factual
errors:
- The default value of ip_local_port_range does not depend on the
amount of memory available in the system.
- tcp_tw_recycle is not enabled by default.
- 1024-4999 is not the default value.
- Etc.
Clean up the mess.
Signed-off-by: Fernando Luis Vazquez Cao <fernando@oss.ntt.co.jp>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit c5ed63d66f24(tcp: fix three tcp sysctls tuning),
sysctl_max_syn_backlog is determined by tcp_hashinfo->ehash_mask,
and the minimal value is 128, and it will increase in proportion to the
memory of machine.
The original description for tcp_max_syn_backlog and sysctl_max_syn_backlog
are out of date.
Changelog:
V2: update description for sysctl_max_syn_backlog
Signed-off-by: Weiping Pan <panweiping3@gmail.com>
Reviewed-by: Shan Wei <shanwei88@gmail.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet