DCTCP congestion control algorithm aims to maximise throughput and minimise
latency in data center networks by utilising the proportion of Explicit
Congestion Notification (ECN) marked packets received from capable hardware as a
congestion signal.
Highlights:
Implemented as a mod_cc(4) module.
ECN (Explicit congestion notification) processing is done differently from
RFC3168.
Takes one-sided DCTCP into consideration where only one of the sides is using
DCTCP and other is using standard ECN.
IETF draft: http://tools.ietf.org/html/draft-bensley-tcpm-dctcp-00
Thesis report by Midori Kato: https://eggert.org/students/kato-thesis.pdf
Submitted by: Midori Kato <katoon@sfc.wide.ad.jp> and
Lars Eggert <lars@netapp.com>
with help and modifications from
hiren
Differential Revision: https://reviews.freebsd.org/D604
Reviewed by: gnn
from the FreeBSD network code. The flag is still kept around in the
"sys/mbuf.h" header file, but does no longer have any users. Instead
the "m_pkthdr.rsstype" field in the mbuf structure is now used to
decide the meaning of the "m_pkthdr.flowid" field. To modify the
"m_pkthdr.rsstype" field please use the existing "M_HASHTYPE_XXX"
macros as defined in the "sys/mbuf.h" header file.
This patch introduces new behaviour in the transmit direction.
Previously network drivers checked if "M_FLOWID" was set in "m_flags"
before using the "m_pkthdr.flowid" field. This check has now now been
replaced by checking if "M_HASHTYPE_GET(m)" is different from
"M_HASHTYPE_NONE". In the future more hashtypes will be added, for
example hashtypes for hardware dedicated flows.
"M_HASHTYPE_OPAQUE" indicates that the "m_pkthdr.flowid" value is
valid and has no particular type. This change removes the need for an
"if" statement in TCP transmit code checking for the presence of a
valid flowid value. The "if" statement mentioned above is now a direct
variable assignment which is then later checked by the respective
network drivers like before.
Additional notes:
- The SCTP code changes will be committed as a separate patch.
- Removal of the "M_FLOWID" flag will also be done separately.
- The FreeBSD version has been bumped.
MFC after: 1 month
Sponsored by: Mellanox Technologies
sending not ready data:
o Add new flag to pru_send() flags - PRUS_NOTREADY.
o Add new protocol method pru_ready().
Sponsored by: Nginx, Inc.
Sponsored by: Netflix
sb_cc member of struct sockbuf to a couple of inline functions:
sbavail() and sbused()
Right now they are equal, but once notion of "not ready socket buffer data",
will be checked in, they are going to be different.
Sponsored by: Netflix
Sponsored by: Nginx, Inc.
It isn't safe to keep unreferenced ifaddrs. Use in6ifa_ifwithaddr() to
determine ifaddr corresponding to destination address. Since currently
we keep addresses with embedded scope zone, in6ifa_ifwithaddr is called
with zero zoneid and marked with XXX.
Also remove route and lle lookups from ip6_input. Use in6ifa_ifwithaddr()
instead.
Sponsored by: Yandex LLC
The current TSO limitation feature only takes the total number of
bytes in an mbuf chain into account and does not limit by the number
of mbufs in a chain. Some kinds of hardware is limited by two
factors. One is the fragment length and the second is the fragment
count. Both of these limits need to be taken into account when doing
TSO. Else some kinds of hardware might have to drop completely valid
mbuf chains because they cannot loaded into the given hardware's DMA
engine. The new way of doing TSO limitation has been made backwards
compatible as input from other FreeBSD developers and will use
defaults for values not set.
Reviewed by: adrian, rmacklem
Sponsored by: Mellanox Technologies
MFC after: 1 week
towards blind SYN/RST spoofed attack.
Originally our stack used in-window checks for incoming SYN/RST
as proposed by RFC793. Later, circa 2003 the RST attack was
mitigated using the technique described in P. Watson
"Slipping in the window" paper [1].
After that, the checks were only relaxed for the sake of
compatibility with some buggy TCP stacks. First, r192912
introduced the vulnerability, just fixed by aforementioned SA.
Second, r167310 had slightly relaxed the default RST checks,
instead of utilizing net.inet.tcp.insecure_rst sysctl.
In 2010 a new technique for mitigation of these attacks was
proposed in RFC5961 [2]. The idea is to send a "challenge ACK"
packet to the peer, to verify that packet arrived isn't spoofed.
If peer receives challenge ACK it should regenerate its RST or
SYN with correct sequence number. This should not only protect
against attacks, but also improve communication with broken
stacks, so authors of reverted r167310 and r192912 won't be
disappointed.
[1] http://bandwidthco.com/whitepapers/netforensics/tcpip/TCP Reset Attacks.pdf
[2] http://www.rfc-editor.org/rfc/rfc5961.txt
Changes made:
o Revert r167310.
o Implement "challenge ACK" protection as specificed in RFC5961
against RST attack. On by default.
- Carefully preserve r138098, which handles empty window edge
case, not described by the RFC.
- Update net.inet.tcp.insecure_rst description.
o Implement "challenge ACK" protection as specificed in RFC5961
against SYN attack. On by default.
- Provide net.inet.tcp.insecure_syn sysctl, to turn off
RFC5961 protection.
The changes were tested at Netflix. The tested box didn't show
any anomalies compared to control box, except slightly increased
number of TCP connection in LAST_ACK state.
Reviewed by: rrs
Sponsored by: Netflix
Sponsored by: Nginx, Inc.
packets targeting a listening socket. Permit to reduce TCP input
processing starvation in context of high SYN load (e.g. short-lived TCP
connections or SYN flood).
Submitted by: Julien Charbon <jcharbon@verisign.com>
Reviewed by: adrian, hiren, jhb, Mike Bentkofsky
mixing on stack memory and UMA memory in one linked list.
Thus, rewrite TCP reassembly code in terms of memory usage. The
algorithm remains unchanged.
We actually do not need extra memory to build a reassembly queue.
Arriving mbufs are always packet header mbufs. So we got the length
of data as pkthdr.len. We got m_nextpkt for linkage. And we need
only one pointer to point at the tcphdr, use PH_loc for that.
In tcpcb the t_segq fields becomes mbuf pointer. The t_segqlen
field now counts not packets, but bytes in the queue. This gives
us more precision when comparing to socket buffer limits.
Sponsored by: Netflix
Sponsored by: Nginx, Inc.
By definition, the very first FIN is not a duplicate. Process it normally
and don't feed it to congestion control as though it were a dupe. Don't
prevent CC from seeing later dupe acks while in a half close state.
- The remote host sends a FIN
- in an ACK for a sequence number for which an ACK has already
been received
- There is still unacked data on route to the remote host
- The packet does not contain a window update
The packet may be dropped without processing the FIN flag.
PR: kern/99188
Submitted by: Staffan Ulfberg <staffan@ulfberg.se>
Discussed with: andre
MFC after: never
In its stead use the Solaris / illumos approach of emulating '-' (dash)
in probe names with '__' (two consecutive underscores).
Reviewed by: markj
MFC after: 3 weeks
option, unbreak the lock tracing release semantic by embedding
calls to LOCKSTAT_PROFILE_RELEASE_LOCK() direclty in the inlined
version of the releasing functions for mutex, rwlock and sxlock.
Failing to do so skips the lockstat_probe_func invokation for
unlocking.
- As part of the LOCKSTAT support is inlined in mutex operation, for
kernel compiled without lock debugging options, potentially every
consumer must be compiled including opt_kdtrace.h.
Fix this by moving KDTRACE_HOOKS into opt_global.h and remove the
dependency by opt_kdtrace.h for all files, as now only KDTRACE_FRAMES
is linked there and it is only used as a compile-time stub [0].
[0] immediately shows some new bug as DTRACE-derived support for debug
in sfxge is broken and it was never really tested. As it was not
including correctly opt_kdtrace.h before it was never enabled so it
was kept broken for a while. Fix this by using a protection stub,
leaving sfxge driver authors the responsibility for fixing it
appropriately [1].
Sponsored by: EMC / Isilon storage division
Discussed with: rstone
[0] Reported by: rstone
[1] Discussed with: philip
to this event, adding if_var.h to files that do need it. Also, include
all includes that now are included due to implicit pollution via if_var.h
Sponsored by: Netflix
Sponsored by: Nginx, Inc.
segments thinking it received only one segment. This causes it to enable
the delay the ACK for 100ms to wait for another segment which may never
come because all the data was received already.
Doing delayed ACK for LRO segments is bogus for two reasons: a) it pushes
us further away from acking every other packet; b) it introduces additional
delay in responding to the sender. The latter is especially bad because it
is in the nature of LRO to aggregated all segments of a burst with no more
coming until an ACK is sent back.
Change the delayed ACK logic to detect LRO segments by being larger than
the MSS for this connection and issuing an immediate ACK for them to keep
the ACK clock ticking without interruption.
Reported by: julian, cperciva
Tested by: cperciva
Reviewed by: lstewart
MFC after: 3 days
sbdrop_locked() to cut acked mbufs from the socket buffer. Free this
chain a batch manner after the socket buffer lock is dropped.
This measurably reduces contention on socket buffer.
Sponsored by: Netflix
Sponsored by: Nginx, Inc.
Approved by: re (marius)
dynamic translation so that their arguments match the definitions for
these providers in Solaris and illumos. Thus, existing scripts for these
providers should work unmodified on FreeBSD.
Tested by: gnn, hiren
MFC after: 1 month
structure is used, but they already have equal fields in the struct
newipsecstat, that was introduced with FAST_IPSEC and then was merged
together with old ipsecstat structure.
This fixes kernel stack overflow on some architectures after migration
ipsecstat to PCPU counters.
Reported by: Taku YAMAMOTO, Maciej Milewski
increased the pointer, not the memory it points to.
In collaboration with: kib
Reported & tested by: Ian FREISLICH <ianf clue.co.za>
Sponsored by: Nginx, Inc.
limited in the amount of data they can handle at once.
Drivers can set ifp->if_hw_tsomax before calling ether_ifattach() to
change the limit.
The lowest allowable size is IP_MAXPACKET / 8 (8192 bytes) as anything
less wouldn't be very useful anymore. The upper limit is still at
IP_MAXPACKET (65536 bytes). Raising it requires further auditing of
the IPv4/v6 code path's as the length field in the IP header would
overflow leading to confusion in firewalls and others packet handler on
the real size of the packet.
The placement into "struct ifnet" is a bit hackish but the best place
that was found. When the stack/driver boundary is updated it should
be handled in a better way.
Submitted by: cperciva (earlier version)
Reviewed by: cperciva
Tested by: cperciva
MFC after: 1 week (using spare struct members to preserve ABI)
duplicate ACK make sure we actually have new data to send.
This prevents us from sending unneccessary pure ACKs.
Reported by: Matt Miller <matt@matthewjmiller.net>
Tested by: Matt Miller <matt@matthewjmiller.net>
MFC after: 2 weeks
connections in the accept queue and contiguous new incoming SYNs.
Compared to the original submitters patch I've moved the test
next to the SYN handling to have it together in a logical unit
and reworded the comment explaining the issue.
Submitted by: Matt Miller <matt@matthewjmiller.net>
Submitted by: Juan Mojica <jmojica@gmail.com>
Reviewed by: Matt Miller (changes)
Tested by: pho
MFC after: 1 week
Convert 'struct ipstat' and 'struct tcpstat' to counter(9).
This speeds up IP forwarding at extreme packet rates, and
makes accounting more precise.
Sponsored by: Nginx, Inc.
For TIMEWAIT handling tcp_input may have to jump back for an additional
pass through pcblookup. Prior to this change the fwd_tag had been
discarded after the first lookup, so a new connection attempt delivered
locally via 'ipfw fwd' would fail to find a match.
As of r248886 the tag will be detached and freed when passed to the
socket buffer.
logic (refer to [1] for associated discussion). snd_cwnd and snd_wnd are
unsigned long and on 64 bit hosts, min() will truncate them to 32 bits and could
therefore potentially corrupt the result (although under normal operation,
neither variable should legitmately exceed 32 bits).
[1] http://lists.freebsd.org/pipermail/freebsd-net/2013-January/034297.html
Submitted by: jhb
MFC after: 1 week
but later after processing and freeing the tag, we need to jump back again
to the findpcb label. Since the fwd_tag pointer wasn't NULL we tried to
process and free the tag for second time.
Reported & tested by: Pawel Tyll <ptyll nitronet.pl>
MFC after: 3 days
Instead, add protocol specific mbuf flags M_IP_NEXTHOP and
M_IP6_NEXTHOP. Use them to indicate that the mbuf's chain
contains the PACKET_TAG_IPFORWARD tag. And do a tag lookup
only when this flag is set.
Suggested by: andre
draft-ietf-tcpm-initcwnd-05. It explains why the increased initial
window improves the overall performance of many web services without
risking congestion collapse.
As long as it remains a draft it is placed under a sysctl marking it
as experimental:
net.inet.tcp.experimental.initcwnd10 = 1
When it becomes an official RFC soon the sysctl will be changed to
the RFC number and moved to net.inet.tcp.
This implementation differs from the RFC draft in that it is a bit
more conservative in the case of packet loss on SYN or SYN|ACK because
we haven't reduced the default RTO to 1 second yet. Also the restart
window isn't yet increased as allowed. Both will be adjusted with
upcoming changes.
Is is enabled by default. In Linux it is enabled since kernel 3.0.
MFC after: 2 weeks
especially in the presence of bi-directional data transfers.
snd_wl1 tracks the right edge, including data in the reassembly
queue, of valid incoming data. This makes it like rcv_nxt plus
reassembly. It never goes backwards to prevent older, possibly
reordered segments from updating the window.
snd_wl2 tracks the left edge of sent data. This makes it a duplicate
of snd_una. However joining them right now is difficult due to
separate update dependencies in different places in the code flow.
snd_wnd tracks the current advertized send window by the peer. In
tcp_output() the effective window is calculated by subtracting the
already in-flight data, snd_nxt less snd_una, from it.
ACK's become the main clock of window updates and will always update
the window when the left edge of what we sent is advanced. The ACK
clock is the primary signaling mechanism in ongoing data transfers.
This works reliably even in the presence of reordering, reassembly
and retransmitted segments. The ACK clock is most important because
it determines how much data we are allowed to inject into the network.
Zero window updates get us out of persistence mode are crucial. Here
a segment that neither moves ACK nor SEQ but enlarges WND is accepted.
When the ACK clock is not active (that is we're not or no longer
sending any data) any segment that moves the extended right SEQ edge,
including out-of-order segments, updates the window. This gives us
updates especially during ping-pong transfers where the peer isn't
done consuming the already acknowledged data from the receive buffer
while responding with data.
The SSH protocol is a prime candidate to benefit from the improved
bi-directional window update logic as it has its own windowing
mechanism on top of TCP and is frequently sending back protocol ACK's.
Tcpdump provided by: darrenr
Tested by: darrenr
MFC after: 2 weeks
We've got more cluster sizes for quite some time now and the orginally
imposed limits and the previously codified thoughts on efficiency gains
are no longer true.
MFC after: 2 weeks
reduce the initial CWND to one segment. This reduction got lost
some time ago due to a change in initialization ordering.
Additionally in tcp_timer_rexmt() avoid entering fast recovery when
we're still in TCPS_SYN_SENT state.
MFC after: 2 weeks
on the related functionality in the runtime via the sysctl variable
net.pfil.forward. It is turned off by default.
Sponsored by: Yandex LLC
Discussed with: net@
MFC after: 2 weeks
before passing a packet to protocol input routines.
For several protocols this mean that now protocol needs to
do subtraction itself, and for another half this means that
we do not need to add header length back to the packet.
Make ip_stripoptions() to adjust ip_len, since now we enter
this function with a packet header whose ip_len does represent
length of entire packet, not payload only.
in network byte order. Any host byte order processing is
done in local variables and host byte order values are
never[1] written to a packet.
After this change a packet processed by the stack isn't
modified at all[2] except for TTL.
After this change a network stack hacker doesn't need to
scratch his head trying to figure out what is the byte order
at the given place in the stack.
[1] One exception still remains. The raw sockets convert host
byte order before pass a packet to an application. Probably
this would remain for ages for compatibility.
[2] The ip_input() still subtructs header len from ip->ip_len,
but this is planned to be fixed soon.
Reviewed by: luigi, Maxim Dounin <mdounin mdounin.ru>
Tested by: ray, Olivier Cochard-Labbe <olivier cochard.me>
that can occur when both sides close at the same time.
If that occurs, without this fix the connection enters
FIN1 on both sides and they will forever send FIN|ACK at
each other until the connection times out. This is because
we stopped processing the FIN|ACK and thus did not advance
the sequence and so never ACK'd each others FIN. This
fix adjusts it so we *do* process the FIN properly and
the race goes away ;-)
MFC after: 1 month
- Stateful TCP offload drivers for Terminator 3 and 4 (T3 and T4) ASICs.
These are available as t3_tom and t4_tom modules that augment cxgb(4)
and cxgbe(4) respectively. The cxgb/cxgbe drivers continue to work as
usual with or without these extra features.
- iWARP driver for Terminator 3 ASIC (kernel verbs). T4 iWARP in the
works and will follow soon.
Build-tested with make universe.
30s overview
============
What interfaces support TCP offload? Look for TOE4 and/or TOE6 in the
capabilities of an interface:
# ifconfig -m | grep TOE
Enable/disable TCP offload on an interface (just like any other ifnet
capability):
# ifconfig cxgbe0 toe
# ifconfig cxgbe0 -toe
Which connections are offloaded? Look for toe4 and/or toe6 in the
output of netstat and sockstat:
# netstat -np tcp | grep toe
# sockstat -46c | grep toe
Reviewed by: bz, gnn
Sponsored by: Chelsio communications.
MFC after: ~3 months (after 9.1, and after ensuring MFC is feasible)
headers for TSO but also for generic checksum offloading. Ideally we
would only have one common function shared amongst all drivers, and
perhaps when updating them for IPv6 we should introduce that.
Eventually we should provide the meta information along with mbufs to
avoid (re-)parsing entirely.
To not break IPv6 (checksums and offload) and to be able to MFC the
changes without risking to hurt 3rd party drivers, duplicate the v4
framework, as other OSes have done as well.
Introduce interface capability flags for TX/RX checksum offload with
IPv6, to allow independent toggling (where possible). Add CSUM_*_IPV6
flags for UDP/TCP over IPv6, and reserve further for SCTP, and IPv6
fragmentation. Define CSUM_DELAY_DATA_IPV6 as we do for legacy IP and
add an alias for CSUM_DATA_VALID_IPV6.
This pretty much brings IPv6 handling in line with IPv4.
TSO is still handled in a different way and not via if_hwassist.
Update ifconfig to allow (un)setting of the new capability flags.
Update loopback to announce the new capabilities and if_hwassist flags.
Individual driver updates will have to follow, as will SCTP.
Reported by: gallatin, dim, ..
Reviewed by: gallatin (glanced at?)
MFC after: 3 days
X-MFC with: r235961,235959,235958
Add code to handle pre-checked TCP checksums as indicated by mbuf
flags to save the entire computation for validation if not needed.
In the IPv6 TCP output path only compute the pseudo-header checksum,
set the checksum offset in the mbuf field along the appropriate flag
as done in IPv4.
In tcp_respond() just initialize the IPv6 payload length to 0 as
ip6_output() will properly set it.
Sponsored by: The FreeBSD Foundation
Sponsored by: iXsystems
Reviewed by: gnn (as part of the whole)
MFC After: 3 days
Factor out the tcp_hc_getmtu() call. As the comments say it
applies to both v4 and v6, so only write it once making it easier
to read the protocol family specifc code.
Sponsored by: The FreeBSD Foundation
Sponsored by: iXsystems
Reviewed by: gnn (as part of the whole)
MFC After: 3 days
proposed MTU value from it and update the TCP host cache. Then
tcp_mss_update() is called on the corresponding tcpcb. It finds the
just allocated entry in the TCP host cache and updates MSS on the
tcpcb. And then we do a fast retransmit of what we have in the tcp
send buffer.
This sequence gets broken if the TCP host cache is exausted. In this
case allocation fails, and later called tcp_mss_update() finds nothing
in cache. The fast retransmit is done with not reduced MSS and is
immidiately replied by remote host with new ICMP datagrams and the
cycle repeats. This ping-pong can go up to wirespeed.
To fix this:
- tcp_mss_update() gets new parameter - mtuoffer, that is like
offer, but needs to have min_protoh subtracted.
- tcp_mtudisc() as notification method renamed to tcp_mtudisc_notify().
- tcp_mtudisc() now accepts not a useless error argument, but proposed
MTU value, that is passed to tcp_mss_update() as mtuoffer.
Reported by: az
Reported by: Andrey Zonov <andrey zonov.org>
Reviewed by: andre (previous version of patch)
hz >> 1000 and thus getting outside the timestamp clock frequenceny of
1ms < x < 1s per tick as mandated by RFC1323, leading to connection
resets on idle connections.
Always use a granularity of 1ms using getmicrouptime() making all but
relevant callouts independent of hz.
Use getmicrouptime(), not getmicrotime() as the latter may make a jump
possibly breaking TCP nfsroot mounts having our timestamps move forward
for more than 24.8 days in a second without having been idle for that
long.
PR: kern/61404
Reviewed by: jhb, mav, rrs
Discussed with: silby, lstewart
Sponsored by: Sandvine Incorporated (originally in 2011)
MFC after: 6 weeks
TCP_KEEPCNT, that allow to control initial timeout, idle time, idle
re-send interval and idle send count on a per-socket basis.
Reviewed by: andre, bz, lstewart
than or equal to rcv_adv and fix tcp_twstart() to handle this case by
assuming the last window was zero rather than a negative value.
The code in tcp_input() already safely handled this case. It can happen
due to delayed ACKs along with a remote sender that sends data beyond
the window we previously advertised. If we have room in our socket buffer
for the extra data beyond the advertised window, we will accept it.
However, if the ACK for that segment is delayed, then we will not
effectively fixup rcv_adv to account for that extra data until the
next segment arrives and forces out an ACK. When that next segment
arrives, rcv_nxt will be beyond rcv_adv.
Tested by: pjd
MFC after: 1 week
The SYSCTL_NODE macro defines a list that stores all child-elements of
that node. If there's no SYSCTL_DECL macro anywhere else, there's no
reason why it shouldn't be static.
They seem to be changed unintentionally in r226437, and there were no
any mentions of renaming in commit log message.
Reported by: Anton Yuzhaninov <citrin citrin ru>
long been superseded by the RFC3390 initial CWND sizing.
Also remove the remnants of TCP_METRICS_CWND which used the
TCP hostcache to set the initial CWND in a non-RFC compliant
way.
MFC after: 1 week
inpcb object.
Skip the TCP_SIGNATURE check in that case as it is consistent with the
output path (no TCP_SIGNATURE for outcoming packets in TIMEWAIT state)
and also because for TIMEWAIT state the verify may be less effective.
Sponsored by: Sandvine Incorporated
Reported by: rwatson
No objections by: rwatson
MFC after: 3 days
and the maximum TCP send and receive buffer limits from 256kB
to 2MB.
For sb_max_adj we need to add the cast as already used in the sysctl
handler to not overflow the type doing the maths.
Note that this is just the defaults. They will allow more memory
to be consumed per socket/connection if needed but not change the
default "idle" memory consumption. All values are still tunable
by sysctls.
Suggested by: gnn
Discussed on: arch (Mar and Aug 2011)
MFC after: 3 weeks
Approved by: re (kib)
Distinguish IPv4 and IPv6 addresses and optional port numbers in
user space to set the option for the correct protocol family.
Add support in the kernel for carrying the new IPv6 destination
address and port.
Add support to TCP and UDP for IPv6 and fix UDP IPv4 to not change
the address in the IP header.
Add support for IPv6 forwarding to a non-local destination.
Add a regession test uitilizing VIMAGE to check all 20 possible
combinations I could think of.
Obtained from: David Dolson at Sandvine Incorporated
(original version for ipfw fwd IPv6 support)
Sponsored by: Sandvine Incorporated
PR: bin/117214
MFC after: 4 weeks
Approved by: re (kib)
hash install, etc. For now, these are arguments are unused, but as we add
RSS support, we will want to use hashes extracted from mbufs, rather than
manually calculated hashes of header fields, due to the expensive of the
software version of Toeplitz (and similar hashes).
Add notes that it would be nice to be able to pass mbufs into lookup
routines in pf(4), optimising firewall lookup in the same way, but the
code structure there doesn't facilitate that currently.
(In principle there is no reason this couldn't be MFCed -- the change
extends rather than modifies the KBI. However, it won't be useful without
other previous possibly less MFCable changes.)
Reviewed by: bz
Sponsored by: Juniper Networks, Inc.
- The existing ipi_lock continues to protect the global inpcb list and
inpcb counter. This lock is now relegated to a small number of
allocation and free operations, and occasional operations that walk
all connections (including, awkwardly, certain UDP multicast receive
operations -- something to revisit).
- A new ipi_hash_lock protects the two inpcbinfo hash tables for
looking up connections and bound sockets, manipulated using new
INP_HASH_*() macros. This lock, combined with inpcb locks, protects
the 4-tuple address space.
Unlike the current ipi_lock, ipi_hash_lock follows the individual inpcb
connection locks, so may be acquired while manipulating a connection on
which a lock is already held, avoiding the need to acquire the inpcbinfo
lock preemptively when a binding change might later be required. As a
result, however, lookup operations necessarily go through a reference
acquire while holding the lookup lock, later acquiring an inpcb lock --
if required.
A new function in_pcblookup() looks up connections, and accepts flags
indicating how to return the inpcb. Due to lock order changes, callers
no longer need acquire locks before performing a lookup: the lookup
routine will acquire the ipi_hash_lock as needed. In the future, it will
also be able to use alternative lookup and locking strategies
transparently to callers, such as pcbgroup lookup. New lookup flags are,
supplementing the existing INPLOOKUP_WILDCARD flag:
INPLOOKUP_RLOCKPCB - Acquire a read lock on the returned inpcb
INPLOOKUP_WLOCKPCB - Acquire a write lock on the returned inpcb
Callers must pass exactly one of these flags (for the time being).
Some notes:
- All protocols are updated to work within the new regime; especially,
TCP, UDPv4, and UDPv6. pcbinfo ipi_lock acquisitions are largely
eliminated, and global hash lock hold times are dramatically reduced
compared to previous locking.
- The TCP syncache still relies on the pcbinfo lock, something that we
may want to revisit.
- Support for reverting to the FreeBSD 7.x locking strategy in TCP input
is no longer available -- hash lookup locks are now held only very
briefly during inpcb lookup, rather than for potentially extended
periods. However, the pcbinfo ipi_lock will still be acquired if a
connection state might change such that a connection is added or
removed.
- Raw IP sockets continue to use the pcbinfo ipi_lock for protection,
due to maintaining their own hash tables.
- The interface in6_pcblookup_hash_locked() is maintained, which allows
callers to acquire hash locks and perform one or more lookups atomically
with 4-tuple allocation: this is required only for TCPv6, as there is no
in6_pcbconnect_setup(), which there should be.
- UDPv6 locking remains significantly more conservative than UDPv4
locking, which relates to source address selection. This needs
attention, as it likely significantly reduces parallelism in this code
for multithreaded socket use (such as in BIND).
- In the UDPv4 and UDPv6 multicast cases, we need to revisit locking
somewhat, as they relied on ipi_lock to stablise 4-tuple matches, which
is no longer sufficient. A second check once the inpcb lock is held
should do the trick, keeping the general case from requiring the inpcb
lock for every inpcb visited.
- This work reminds us that we need to revisit locking of the v4/v6 flags,
which may be accessed lock-free both before and after this change.
- Right now, a single lock name is used for the pcbhash lock -- this is
undesirable, and probably another argument is required to take care of
this (or a char array name field in the pcbinfo?).
This is not an MFC candidate for 8.x due to its impact on lookup and
locking semantics. It's possible some of these issues could be worked
around with compatibility wrappers, if necessary.
Reviewed by: bz
Sponsored by: Juniper Networks, Inc.
buffer fills up causing the remote sender to enter into persist mode, but
there is still room available in the receive buffer when a window probe
arrives (either due to window scaling, or due to the local application
very slowing draining data from the receive buffer), then the single byte
of data in the window probe is accepted. However, this can cause rcv_nxt
to be greater than rcv_adv. This condition will only last until the next
ACK packet is pushed out via tcp_output(), and since the previous ACK
advertised a zero window, the ACK should be pushed out while the TCP
pcb is write-locked.
During the window while rcv_nxt is greather than rcv_adv, a few places
would compute the remaining receive window via rcv_adv - rcv_nxt.
However, this value was then (uint32_t)-1. On a 64 bit machine this
could expand to a positive 2^32 - 1 when cast to a long. In particular,
when calculating the receive window in tcp_output(), the result would be
that the receive window was computed as 2^32 - 1 resulting in advertising
a far larger window to the remote peer than actually existed.
Fix various places that compute the remaining receive window to either
assert that it is not negative (i.e. rcv_nxt <= rcv_adv), or treat the
window as full if rcv_nxt is greather than rcv_adv.
Reviewed by: bz
MFC after: 1 month
Add some comments at #endifs given more nestedness. To make the compiler
happy, some default initializations were added in accordance with the style
on the files.
Reviewed by: gnn
Sponsored by: The FreeBSD Foundation
Sponsored by: iXsystems
MFC after: 4 days
persist state and the retransmit timer. However, the code that implements
"bad retransmit recovery" only checks t_rxtshift to see if an ACK has been
received in during the first retransmit timeout window. As a result, if
ticks has wrapped over to a negative value and a socket is in the persist
state, it can incorrectly treat an ACK from the remote peer as a
"bad retransmit recovery" and restore saved values such as snd_ssthresh and
snd_cwnd. However, if the socket has never had a retransmit timeout, then
these saved values will be zero, so snd_ssthresh and snd_cwnd will be set
to 0.
If the socket is in fast recovery (this can be caused by excessive
duplicate ACKs such as those fixed by 220794), then each ACK that arrives
triggers either NewReno or SACK partial ACK handling which clamps snd_cwnd
to be no larger than snd_ssthresh. In effect, the socket's send window
is permamently stuck at 0 even though the remote peer is advertising a
much larger window and pending data is only sent via TCP window probes
(so one byte every few seconds).
Fix this by adding a new TCP pcb flag (TF_PREVVALID) that indicates that
the various snd_*_prev fields in the pcb are valid and only perform
"bad retransmit recovery" if this flag is set in the pcb. The flag is set
on the first retransmit timeout that occurs and is cleared on subsequent
retransmit timeouts or when entering the persist state.
Reviewed by: bz
MFC after: 2 weeks
As long as this is a costy function, even when compiled in (along with
the option TCP_SIGNATURE), it can be disabled via the
net.inet.tcp.signature_verify_input sysctl.
Sponsored by: Sandvine Incorporated
Reviewed by: emaste, bz
MFC after: 2 weeks
to the maximum allowed window. Growing the window too large would cause
an underflow in the calculations in tcp_output() to decide if a window
update should be sent which would prevent the persist timer from being
started if data was pending and the other end of the connection advertised
an initial window size of 0.
PR: kern/154006
Submitted by: Stefan `Sec` Zehl sec 42 org
Reviewed by: bz
MFC after: 1 week
hint is 0 when no SACK data is received to update the hint with. This was
accidentally omitted from r216753.
Sponsored by: FreeBSD Foundation
MFC after: 10 weeks
X-MFC with: 216753
access inbound/outbound events and associated data for established TCP
connections. The hooks only run if at least one hook function is registered
for the hook point, ensuring the impact on the stack is effectively nil when
no TCP Khelp modules are loaded. struct tcp_hhook_data is passed as contextual
data to any registered Khelp module hook functions.
- Add an OSD (Object Specific Data) pointer to struct tcpcb to allow Khelp
modules to associate per-connection data with the TCP control block.
- Bump __FreeBSD_version and add a note to UPDATING regarding to ABI changes
introduced by this commit and r216753.
In collaboration with: David Hayes <dahayes at swin edu au> and
Grenville Armitage <garmitage at swin edu au>
Sponsored by: FreeBSD Foundation
Reviewed by: bz, others along the way
MFC after: 3 months
Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details
about the project are available at: http://caia.swin.edu.au/freebsd/5cc/
- Add a KPI and supporting infrastructure to allow modular congestion control
algorithms to be used in the net stack. Algorithms can maintain per-connection
state if required, and connections maintain their own algorithm pointer, which
allows different connections to concurrently use different algorithms. The
TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to
programmatically query or change the congestion control algorithm respectively
from within an application at runtime.
- Integrate the framework with the TCP stack in as least intrusive a manner as
possible. Care was also taken to develop the framework in a way that should
allow integration with other congestion aware transport protocols (e.g. SCTP)
in the future. The hope is that we will one day be able to share a single set
of congestion control algorithm modules between all congestion aware transport
protocols.
- Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack
and use it to decouple the meaning of recovery from a congestion event and
recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based
congestion control protocols don't generally need to recover from packet loss
and need a different way to note a congestion recovery episode within the
stack.
- Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code
and ensures the stack always uses the appropriate mechanisms for recovering
from packet loss during a congestion recovery episode.
- Extract the NewReno congestion control algorithm from the TCP stack and
massage it into module form. NewReno is always built into the kernel and will
remain the default algorithm for the forseeable future. Implementations of
additional different algorithms will become available in the near future.
- Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code
that relies on the size of "struct tcpcb" is required.
Many thanks go to the Cisco University Research Program Fund at Community
Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work
at the Centre for Advanced Internet Architectures, Swinburne University of
Technology is greatly appreciated.
In collaboration with: David Hayes <dahayes at swin edu au> and
Grenville Armitage <garmitage at swin edu au>
Sponsored by: Cisco URP, FreeBSD Foundation
Reviewed by: rpaulo
Tested by: David Hayes (and many others over the years)
MFC after: 3 months
to give way for the pluggable congestion control framework. It is
the task of the congestion control algorithm to set the congestion
window and amount of inflight data without external interference.
In 'struct tcpcb' the variables previously used by the inflight
limiter are renamed to spares to keep the ABI intact and to have
some more space for future extensions.
In 'struct tcp_info' the variable 'tcpi_snd_bwnd' is not removed to
preserve the ABI. It is always set to 0.
In siftr.c in 'struct pkt_node' the variable 'snd_bwnd' is not removed
to preserve the ABI. It is always set to 0.
These unused variable in the various structures may be reused in the
future or garbage collected before the next release or at some other
point when an ABI change happens anyway for other reasons.
No MFC is planned. The inflight bandwidth limiter stays disabled by
default in the other branches but remains available.
sysctl's and remove any side effects.
Both sysctl's share the same backend infrastructure and due to the
way it was implemented enabling net.inet.tcp.log_in_vain would also
cause log_debug output to be generated. This was surprising and
eventually annoying to the user.
The log output backend is kept the same but a little shim is inserted
to properly separate log_in_vain and log_debug and to remove any side
effects.
PR: kern/137317
MFC after: 1 week
