These flags are TCP specific. While here, make also several LRO
internal functions to pass tcpcb pointer instead of inpcb one.
Reviewed by: rrs
Differential Revision: https://reviews.freebsd.org/D39698
This makes inpcb lighter and allows future cache line optimizations
of tcpcb. The reason why HPTS originally used inpcb is the compressed
TIME-WAIT state (see 0d7445193a), that used to free a tcpcb, while the
associated connection is still on the HPTS ring.
Reviewed by: rrs
Differential Revision: https://reviews.freebsd.org/D39697
The socket argument is superfluous, as a tcpcb always has one and
only one socket.
Reviewed by: rrs
Differential Revision: https://reviews.freebsd.org/D39434
Both BBR and Rack have the ability to log socket options, which is currently disabled. Rack
has an experimental SaD (Sack Attack Detection) algorithm that should be made available. Also
there is a t_maxpeak_rate that needs to be removed (its un-used).
Reviewed by: tuexen, cc
Sponsored by: Netflix Inc
Differential Revision: https://reviews.freebsd.org/D39427
So stack switching as always been a bit of a issue. We currently use a break before make setup which means that
if something goes wrong you have to try to get back to a stack. This patch among a lot of other things changes that so
that it is a make before break. We also expand some of the function blocks in prep for new features in rack that will allow
more controlled pacing. We also add other abilities such as the pathway for a stack to query a previous stack to acquire from
it critical state information so things in flight don't get dropped or mis-handled when switching stacks. We also add the
concept of a timer granularity. This allows an alternate stack to change from the old ticks granularity to microseconds and
of course this even gives us a pathway to go to nanosecond timekeeping if we need to (something for the data center to consider
for sure).
Once all this lands I will then update rack to begin using all these new features.
Reviewed by: tuexen
Sponsored by: Netflix Inc
Differential Revision: https://reviews.freebsd.org/D39210
The TCP stacks have long accessed t_logstate directly, but in order to do tracepoints and the new bbpoints
we need to move to using the new inline functions. This adds them and moves rack to now use
the tcp_tracepoints.
Reviewed by: tuexen, gallatin
Sponsored by: Netflix Inc
Differential Revision: https://reviews.freebsd.org/D38831
tcp_trace was implemented in tcp_debug.c, which was removed recently.
Reviewed by: rscheff@, zlei@
Sponsored by: Netflix, Inc.
Differential Revision: https://reviews.freebsd.org/D38712
Extend the BBLog RTO event to deal with all timers of the base
stack. Also provide information about starting, stopping, and
running off. The expiration of the retransmission timer is
reported as it was done before.
Reviewed by: rscheff@
Sponsored by: Netflix, Inc.
Differential Revision: https://reviews.freebsd.org/D38710
Rearrange the enum tt_which such that TT_REXMIT is 0. This allows
an extension of the BBLog event RTO in a backwards compatible way.
Remove tcptimers, which was only used in trpt, a utility removed
from the source tree recently.
Reviewed by: glebius@, guest-ccui@
Sponsored by: Netflix, Inc.
Differential Revision: https://reviews.freebsd.org/D38547
This allows us to avoid spurious calls to ktls_disable_ifnet()
When we implemented ifnet kTLSe, we set a flag in the tx socket
buffer (SB_TLS_IFNET) to indicate ifnet kTLS. This flag meant that
now, or in the past, ifnet ktls was active on a socket. Later,
I added code to switch ifnet ktls sessions to software in the case
of lossy TCP connections that have a high retransmit rate.
Because TCP was using SB_TLS_IFNET to know if it needed to do math
to calculate the retransmit ratio and potentially call into
ktls_disable_ifnet(), it was doing unneeded work long after
a session was moved to software.
This patch carefully tracks whether or not ifnet ktls is still enabled
on a TCP connection. Because the inp is now embedded in the tcpcb, and
because TCP is the most frequent accessor of this state, it made sense to
move this from the socket buffer flags to the tcpcb. Because we now need
reliable access to the tcbcb, we take a ref on the inp when creating a tx
ktls session.
While here, I noticed that rack/bbr were incorrectly implementing
tfb_hwtls_change(), and applying the change to all pending sends,
when it should apply only to future sends.
This change reduces spurious calls to ktls_disable_ifnet() by 95% or so
in a Netflix CDN environment.
Reviewed by: markj, rrs
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D38380
During tcp session start, various mechanisms need to
track a few initial RTTs before becoming active.
Prevent overflows of the corresponding tracking counter
and reduce the size of tcpcb simultaneously.
Reviewed By: #transport, tuexen, guest-ccui
Sponsored by: NetApp, Inc.
Differential Revision: https://reviews.freebsd.org/D21117
Use only one callout structure per tcpcb that is responsible for handling
all five TCP timeouts. Use locked version of callout, of course. The
callout function tcp_timer_enter() chooses soonest timer and executes it
with lock held. Unless the timer reports that the tcpcb has been freed,
the callout is rescheduled for next soonest timer, if there is any.
With single callout per tcpcb on connection teardown we should be able
to fully stop the callout and immediately free it, avoiding use of
callout_async_drain(). There is one gotcha here: callout_stop() can
actually touch our memory when a rare race condition happens. See
comment above tcp_timer_stop(). Synchronous stop of the callout makes
tcp_discardcb() the single entry point for tcpcb destructor, merging the
tcp_freecb() to the end of the function.
While here, also remove lots of lingering checks in the beginning of
TCP timer functions. With a locked callout they are unnecessary.
While here, clean unused parts of timer KPI for the pluggable TCP stacks.
While here, remove TCPDEBUG from tcp_timer.c, as this allows for more
simplification of TCP timers. The TCPDEBUG is scheduled for removal.
Move the DTrace probes in timers to the beginning of a function, where
a tcpcb is always existing.
Discussed with: rrs, tuexen, rscheff (the TCP part of the diff)
Reviewed by: hselasky, kib, mav (the callout part)
Differential revision: https://reviews.freebsd.org/D37321
For the TCP protocol inpcb storage specify allocation size that would
provide space to most of the data a TCP connection needs, embedding
into struct tcpcb several structures, that previously were allocated
separately.
The most import one is the inpcb itself. With embedding we can provide
strong guarantee that with a valid TCP inpcb the tcpcb is always valid
and vice versa. Also we reduce number of allocs/frees per connection.
The embedded inpcb is placed in the beginning of the struct tcpcb,
since in_pcballoc() requires that. However, later we may want to move
it around for cache line efficiency, and this can be done with a little
effort. The new intotcpcb() macro is ready for such move.
The congestion algorithm data, the TCP timers and osd(9) data are
also embedded into tcpcb, and temprorary struct tcpcb_mem goes away.
There was no extra allocation here, but we went through extra pointer
every time we accessed this data.
One interesting side effect is that now TCP data is allocated from
SMR-protected zone. Potentially this allows the TCP stacks or other
TCP related modules to utilize that for their own synchronization.
Large part of the change was done with sed script:
s/tp->ccv->/tp->t_ccv./g
s/tp->ccv/\&tp->t_ccv/g
s/tp->cc_algo/tp->t_cc/g
s/tp->t_timers->tt_/tp->tt_/g
s/CCV\(ccv, osd\)/\&CCV(ccv, t_osd)/g
Dependency side effect is that code that needs to know struct tcpcb
should also know struct inpcb, that added several <netinet/in_pcb.h>.
Differential revision: https://reviews.freebsd.org/D37127
Have tcpstats (netstat -s) differentiate between received and sent
ECN-marked packets. Also account for IP ECN bits (on TCP packets)
even when the tcp session has not negotiated ECN support.
Event: IETF 115 Hackathon
Reviewed By: glebius, tuexen, #transport
Sponsored by: NetApp, Inc.
Differential Revision: https://reviews.freebsd.org/D37314
This adds the capability for a modular congestion control
to select which variant of ECN-capable-transport it wants to use
when sending out elegible segments. As an initial CC to utilize
this, DCTCP was selected.
Event: IETF 115 Hackathon
Reviewed By: tuexen, #transport
Sponsored by: NetApp, Inc.
Differential Revision: https://reviews.freebsd.org/D24869
Accurate ECN asks to conservatively estimate, when the
ACE counter may have wrapped due to a single ACK covering a larger
number of segments. This is described in Annex A.2 of the
accurate-ecn draft.
Event: IETF 115 Hackathon
Reviewed By: tuexen, #transport
Sponsored by: NetApp, Inc.
Differential Revision: https://reviews.freebsd.org/D37281
It is called only from tcp_input() and always has valid parameter.
Reviewed by: rscheff, tuexen
Differential revision: https://reviews.freebsd.org/D37115
tcp_respond is another function to build a tcp control packet
quickly. With ECN++ and AccECN, both the IP ECN header, and
the TCP ECN flags are supposed to reflect the correct state.
Also ensure that on receiving multiple ECN SYN-ACKs, the
responses triggered will reflect the latest state.
Reviewed By: tuexen, #transport
Sponsored by: NetApp, Inc.
Differential Revision: https://reviews.freebsd.org/D36973
- Use C99 types uintXX_t instead of u_intXX_t.
- Try to make space/tab usage a little bit more consistent.
- Shorten comments to fit into 80 chars.
Not a functional change, just making future changes easier to read.
The memory savings the tcptw brought back in 2003 (see 340c35de6a) no
longer justify the complexity required to maintain it. For longer
explanation please check out the email [1].
Surpisingly through almost 20 years the TCP stack functionality of
handling the TIME_WAIT state with a normal tcpcb did not bitrot. The
existing tcp_input() properly handles a tcpcb in TCPS_TIME_WAIT state,
which is confirmed by the packetdrill tcp-testsuite [2].
This change just removes tcptw and leaves INP_TIMEWAIT. The flag will
be removed in a separate commit. This makes it easier to review and
possibly debug the changes.
[1] https://lists.freebsd.org/archives/freebsd-net/2022-January/001206.html
[2] https://github.com/freebsd-net/tcp-testsuite
Differential revision: https://reviews.freebsd.org/D36398
TCP has an idle-reduce feature that allows a connection to reduce its
cwnd after it has been idle more than an RTT. This feature only works
for a sending side connection. It does this by at output checking the
idle time (t_rcvtime vs ticks) to see if its more than the RTO timeout.
The problem comes if you are a web server. You get a request and
then send out all the data.. then go idle. The next time you would
send is in response to a request from the peer asking for more data.
But the thing is you updated t_rcvtime when the request came in so
you never reduce.
The fix is to do the idle reduce check also on inbound.
Reviewed by: tuexen, rscheff
Sponsored by: Netflix Inc
Differential Revision: https://reviews.freebsd.org/D36721
After decoupling of protosw(9) and IP wire protocols in 78b1fc05b2 for
IPv4 we got vector ip_ctlprotox[] that is executed only and only from
icmp_input() and respectively for IPv6 we got ip6_ctlprotox[] executed
only and only from icmp6_input(). This allows to use protocol specific
argument types in these methods instead of struct sockaddr and void.
Reviewed by: melifaro
Differential revision: https://reviews.freebsd.org/D36727
Rack has had the ability to timeout connections that just sit idle automatically. This
feature of course is off by default and requires the user set it on (though the socket option
has been missing in tcp_usrreq.c). Lets get the progress timeout fully supported in
the base stack as well as rack.
Reviewed by: tuexen
Sponsored by: Netflix Inc
Differential Revision: https://reviews.freebsd.org/D36716
Right now if you use rack with cubic (the new default cc) you will have
improper results. This is because rack uses different variables than
the base stack (or bbr) and thus tcp_compute_pipe() always returns
so that cubic will choose a 30% backoff not the 50% backoff it should
when it is newreno compatibility mode. The fix is to allow a stack (rack)
to override its own compute_pipe.
Reviewed by: tuexen, rscheff
Sponsored by: Netflix Inc
Differential Revision: https://reviews.freebsd.org/D36711
- The soisconnected() call on transition from SYN_RCVD to ESTABLISHED
is also necessary for a half-synchronized connection. Fix that
just setting the flag, when we transfer SYN-SENT -> SYN-RECEIVED.
- Provide a comment that explains at what conditions the call to
soisconnected() is necessary.
- Hence mechanically rename the TF_INCQUEUE flag to TF_SONOTCONN.
- Extend the change to the BBR and RACK stacks.
Note: the interaction between the accept_filter(9) and the socket layer
is not fully consistent, yet. For most accept filters this call to
soisconnected() will not move the connection from the incomplete queue
to the complete. The move would happen only when the filter has received
the desired data, and soisconnected() would be called once again from
sorwakeup(). Ideally, we should mark socket as connected only there,
and leave the soisconnected() from SYN_RCVD->ESTABLISHED only for the
simultaneous open case. However, this doesn't yet work.
Reviewed by: rscheff, tuexen, rrs
Differential revision: https://reviews.freebsd.org/D36641
This call existed since pre-FreeBSD times, and it is hard to understand
why it was there in the first place. After 6f3caa6d81 it definitely
became necessary always and commit message from f1ee30ccd6 confirms that.
Now that 6f3caa6d81 is effectively backed out by 07285bb4c2, the call
appears to be useful only for sockets that landed on the incomplete queue,
e.g. sockets that have accept_filter(9) enabled on them.
Provide a new TCP flag to mark connections that are known to be on the
incomplete queue, and call soisconnected() only for those connections.
Reviewed by: rrs, tuexen
Differential revision: https://reviews.freebsd.org/D36488
The AccECN handshake and TCP header flags are supported,
no support yet for the AccECN option. This minimalistic
implementation is sufficient to support DCTCP while
dramatically cutting the number of ACKs, and provide ECN
response from the receiver to the CC modules.
Reviewed By: #transport, #manpages, rrs, pauamma
Sponsored by: NetApp, Inc.
Differential Revision: https://reviews.freebsd.org/D21011
o Assert that every protosw has pr_attach. Now this structure is
only for socket protocols declarations and nothing else.
o Merge struct pr_usrreqs into struct protosw. This was suggested
in 1996 by wollman@ (see 7b187005d1), and later reiterated
in 2006 by rwatson@ (see 6fbb9cf860).
o Make struct domain hold a variable sized array of protosw pointers.
For most protocols these pointers are initialized statically.
Those domains that may have loadable protocols have spacers. IPv4
and IPv6 have 8 spacers each (andre@ dff3237ee5).
o For inetsw and inet6sw leave a comment noting that many protosw
entries very likely are dead code.
o Refactor pf_proto_[un]register() into protosw_[un]register().
o Isolate pr_*_notsupp() methods into uipc_domain.c
Reviewed by: melifaro
Differential revision: https://reviews.freebsd.org/D36232
The method was called for two different conditions: 1) the VM layer is
low on pages or 2) one of UMA zones of mbuf allocator exhausted.
This change 2) into a new event handler, but all affected network
subsystems modified to subscribe to both, so this change shall not
bring functional changes under different low memory situations.
There were three subsystems still using pr_drain: TCP, SCTP and frag6.
The latter had its protosw entry for the only reason to register its
pr_drain method.
Reviewed by: tuexen, melifaro
Differential revision: https://reviews.freebsd.org/D36164
Modern TCP stacks uses multiple callouts per tcpcb, and a global
callout is ancient artifact. However it is still used to garbage
collect compressed timewait entries.
Reviewed by: melifaro, tuexen
Differential revision: https://reviews.freebsd.org/D36159
The flag SS_NOFDREF is a private flag of the socket layer. It also
is supposed to be read with SOCK_LOCK(), which we don't own here.
Reviewed by: rrs, tuexen
Differential revision: https://reviews.freebsd.org/D35663
In the transport call on 12/3 Gleb asked to move the CC modules towards
using reference counting to prevent folks from unloading a module in use.
It was also agreed that Michael would do a user space utility like tcp_drop
that could be used to move all connections that are using a specific CC
to some other CC.
This is the half I committed to doing, making it so that we maintain a refcount
on a cc module every time a pcb refers to it and decrementing that every
time a pcb no longer uses a cc module. This also helps us simplify the
whole unloading process by getting rid of tcp_ccunload() which munged
through all the tcb's. Instead we mark a module as being removed and
prevent further references to it. We also make sure that if a module is
marked as being removed it cannot be made as the default and also
the opposite of that, if its a default it fails and does not mark it as being
removed.
Reviewed by: Michael Tuexen, Gleb Smirnoff
Sponsored by: Netflix Inc.
Differential Revision: https://reviews.freebsd.org/D33249
The padding in CURRENT shall not shrink. It is
designed that in CURRENT at always stays
the same, and then when a new stable is branched, it
inherits 6 pointer placeholders that can be used
withing this stable/X lifetime to extend the structure.
Reviewed By: tuexen, #transport
Sponsored by: NetApp, Inc.
Differential Revision: https://reviews.freebsd.org/D34269
Reserve couters in the tcps struct in preparation
for AccECN, extend the debugging output for TF2
flags, optimize the syncache flags from individual
bits to a codepoint for the specifc ECN handshake.
This is in preparation of AccECN.
No functional chance except for extended debug
output capabilities.
Reviewed By: #transport, rrs
Sponsored by: NetApp, Inc.
Differential Revision: https://reviews.freebsd.org/D34161
tcp_ctloutput_set() will be used via the sysctl interface in a
upcoming command line tool tcpsso.
Reviewed by: glebius, rscheff
Sponsored by: Netflix, Inc.
Differential Revision: https://reviews.freebsd.org/D34164
In order to consistently provide access to all
(including reserved) TCP header flag bits,
use an accessor function tcp_get_flags and
tcp_set_flags. Also expand any flag variable from
uint8_t / char to uint16_t.
Reviewed By: hselasky, tuexen, glebius, #transport
Sponsored by: NetApp, Inc.
Differential Revision: https://reviews.freebsd.org/D34130