When tearing down vnet jails we can move an if_bridge out (as
part of the normal vnet_if_return()). This can, when it's clearing out
its list of member interfaces, change its link layer address.
That sends an iflladdr_event, but at that point we've already freed the
AF_INET/AF_INET6 if_afdata pointers.
In other words: when the iflladdr_event callbacks fire we can't assume
that ifp->if_afdata[AF_INET] will be set.
Reviewed by: donner@, melifaro@
MFC after: 1 week
Sponsored by: Orange Business Services
Differential Revision: https://reviews.freebsd.org/D28860
terminating a TCP connection.
If a TCP packet must be retransmitted and the data length has changed in the
retransmitted packet, due to the internal workings of TCP, typically when ACK
packets are lost, then there is a 30% chance that the logic in GetDeltaSeqOut()
will find the correct length, which is the last length received.
This can be explained as follows:
If a "227 Entering Passive Mode" packet must be retransmittet and the length
changes from 51 to 50 bytes, for example, then we have three cases for the
list scan in GetDeltaSeqOut(), depending on how many prior packets were
received modulus N_LINK_TCP_DATA=3:
case 1: index 0: original packet 51
index 1: retransmitted packet 50
index 2: not relevant
case 2: index 0: not relevant
index 1: original packet 51
index 2: retransmitted packet 50
case 3: index 0: retransmitted packet 50
index 1: not relevant
index 2: original packet 51
This patch simply changes the searching order for TCP packets, always starting
at the last received packet instead of any received packet, in
GetDeltaAckIn() and GetDeltaSeqOut().
Else no functional changes.
Discussed with: rscheff@
Submitted by: Andreas Longwitz <longwitz@incore.de>
PR: 230755
MFC after: 1 week
Sponsored by: Mellanox Technologies // NVIDIA Networking
Under some circumstances, PRR may end up with a fully
collapsed cwnd when finalizing the loss recovery.
Reviewed By: #transport, kbowling
Reported by: Liang Tian
MFC after: 1 week
Sponsored by: NetApp, Inc.
Differential Revision: https://reviews.freebsd.org/D28780
TCP_LINGERTIME can be traced back to BSD 4.4 Lite and perhaps beyond, in
exactly the same form that it appears here modulo slightly different
context. It used to be the case that there was a single pr_usrreq
method with requests dispatched to it; these exact two lines appeared in
tcp_usrreq's PRU_ATTACH handling.
The only purpose of this that I can find is to cause surprising behavior
on accepted connections. Newly-created sockets will never hit these
paths as one cannot set SO_LINGER prior to socket(2). If SO_LINGER is
set on a listening socket and inherited, one would expect the timeout to
be inherited rather than changed arbitrarily like this -- noting that
SO_LINGER is nonsense on a listening socket beyond inheritance, since
they cannot be 'connected' by definition.
Neither Illumos nor Linux reset the timer like this based on testing and
inspection of Illumos, and testing of Linux.
Reviewed by: rscheff, tuexen
Differential Revision: https://reviews.freebsd.org/D28265
a further CPU enhancements for compressed acks. These
are acks that are compressed into an mbuf. The transport
has to be aware of how to process these, and an upcoming
update to rack will do so. You need the rack changes
to actually test and validate these since if the transport
does not support mbuf compression, then the old code paths
stay in place. We do in this commit take out the concept
of logging if you don't have a lock (which was quite
dangerous and was only for some early debugging but has
been left in the code).
Sponsored by: Netflix Inc.
Differential Revision: https://reviews.freebsd.org/D28374
It fixes loopback route installation for the interfaces
in the different fibs using the same prefix.
Reviewed By: donner
PR: 189088
Differential Revision: https://reviews.freebsd.org/D28673
MFC after: 1 week
- improved pipe calculation which does not degrade under heavy loss
- engaging in Loss Recovery earlier under adverse conditions
- Rescue Retransmission in case some of the trailing packets of a request got lost
All above changes are toggled with the sysctl "rfc6675_pipe" (disabled by default).
Reviewers: #transport, tuexen, lstewart, slavash, jtl, hselasky, kib, rgrimes, chengc_netapp.com, thj, #manpages, kbowling, #netapp, rscheff
Reviewed By: #transport
Subscribers: imp, melifaro
MFC after: 2 weeks
Sponsored by: NetApp, Inc.
Differential Revision: https://reviews.freebsd.org/D18985
Currently ip6_input() calls in6ifa_ifwithaddr() for
every local packet, in order to check if the target ip
belongs to the local ifa in proper state and increase
its counters.
in6ifa_ifwithaddr() references found ifa.
With epoch changes, both `ip6_input()` and all other current callers
of `in6ifa_ifwithaddr()` do not need this reference
anymore, as epoch provides stability guarantee.
Given that, update `in6ifa_ifwithaddr()` to allow
it to return ifa without referencing it, while preserving
option for getting referenced ifa if so desired.
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D28648
Use ISS for SEG.SEQ when sending a SYN-ACK segment in response to
an SYN segment received in the SYN-SENT state on a socket having
the IPPROTO_TCP level socket option TCP_NOOPT enabled.
Reviewed by: rscheff
Sponsored by: Netflix, Inc.
Differential Revision: https://reviews.freebsd.org/D28656
In error case we can leave `inp' locked, also we need to free
mbuf chain `m' in the same case. Release the lock and use `badunlocked'
label to exit with freed mbuf. Also modify UDP error statistic to
match the IPv6 code.
Remove redundant INP_RUNLOCK() from the `if (last == NULL)' block,
there are no ways to reach this point with locked `inp'.
Obtained from: Yandex LLC
MFC after: 3 days
Sponsored by: Yandex LLC
Historically receive buffer overflows have been ignored and programs
could not tell if they missed messages or messages had been truncated
because of overflows. Since programs historically do not expect to get
receive overflow errors, this behavior is not the default.
This is really really important for programs that use route(4) to keep in sync
with the system. If we loose a message then we need to reload the full system
state, otherwise the behaviour from that point is undefined and can lead
to chasing bogus bug reports.
to be a true RFC 6598 NAT444 setup, where each network segment (e.g. user,
subnet) can have their own dedicated port aliasing ranges.
Reviewed by: donner, kp
Approved by: 0mp (mentor), donner, kp
Differential Revision: https://reviews.freebsd.org/D23450
Improve the handling of INIT chunks in specific szenarios and
report and appropriate error cause.
Thanks to Anatoly Korniltsev for reporting the issue for the
userland stack.
MFC after: 3 days
Originally IFCAP_NOMAP meant that the mbuf has external storage pointer
that points to unmapped address. Then, this was extended to array of
such pointers. Then, such mbufs were augmented with header/trailer.
Basically, extended mbufs are extended, and set of features is subject
to change. The new name should be generic enough to avoid further
renaming.
tree that fix the ratelimit code. There were several bugs
in tcp_ratelimit itself and we needed further work to support
the multiple tag format coming for the joint TLS and Ratelimit dances.
Sponsored by: Netflix Inc.
Differential Revision: https://reviews.freebsd.org/D28357
With clearing of recover_fs in bc7ee8e5bc, div/0
was observed while processing partial_acks.
Suspect that rewind of an erraneous RTO may be
causing this - with the above change, recover_fs
would no longer retained at the last calculated
value, and reset. But CC_RTO_ERR can reenable
IN_RECOVERY(), without setting this again.
Adding a safety net prior to the division in that
function, which I missed in D28114.
Summary:
Wrap lines before column 80 in new prr code checked in recently.
No functional changes.
Reviewers: tuexen, rrs, jtl, mm, kbowling, #transport
Reviewed By: tuexen, mm, #transport
Subscribers: imp, melifaro
Differential Revision: https://reviews.freebsd.org/D28329
* Fix bug with /32 aliases introduced in 81728a538d.
* Explicitly document business logic for IPv4 ifa routes.
* Remove remnants of rtinit()
* Deduplicate ifa->route prefix code by moving it into ia_getrtprefix()
* Deduplicate conditional check for ifa_maintain_loopback_route() by
moving into ia_need_loopback_route()
* Remove now-unused flags argument from in_addprefix().
Reviewed by: donner
PR: 252883
Differential Revision: https://reviews.freebsd.org/D28246
Summary:
When using the base stack in conjunction with RACK, it appears that
infrequently, ++tp->t_dupacks is instantly larger than tcprexmtthresh.
This leaves the recover flightsize (sackhint.recover_fs) uninitialized,
leading to a div/0 panic.
Address this by properly initializing the variable just prior to first
use, if it is not properly initialized.
In order to prevent stale information from a prior recovery to
negatively impact the PRR calculations in this event, also clear
recover_fs once loss recovery is finished.
Finally, improve the readability of the initialization of recover_fs
when t_dupacks == tcprexmtthresh by adjusting the indentation and
using the max(1, snd_nxt - snd_una) macro.
Reviewers: rrs, kbowling, tuexen, jtl, #transport, gnn!, jmg, manu, #manpages
Reviewed By: rrs, kbowling, #transport
Subscribers: bdrewery, andrew, rpokala, ae, emaste, bz, bcran, #linuxkpi, imp, melifaro
Differential Revision: https://reviews.freebsd.org/D28114
There are many casts of this struct to uint32_t, so we also need to ensure
that it is sufficiently aligned to safely perform this cast on architectures
that don't allow unaligned accesses. This fixes lots of -Wcast-align warnings.
Reviewed By: ae
Differential Revision: https://reviews.freebsd.org/D27879
This fixes -Wcast-align warnings caused by the underaligned `struct ip`.
This also silences them in the public functions by changing the function
signature from char * to void *. This is source and binary compatible and
avoids the -Wcast-align warning.
Reviewed By: ae, gbe (manpages)
Differential Revision: https://reviews.freebsd.org/D27882
rtinit[1]() is a function used to add or remove interface address prefix routes,
similar to ifa_maintain_loopback_route().
It was intended to be family-agnostic. There is a problem with this approach
in reality.
1) IPv6 code does not use it for the ifa routes. There is a separate layer,
nd6_prelist_(), providing interface for maintaining interface routes. Its part,
responsible for the actual route table interaction, mimics rtenty() code.
2) rtinit tries to combine multiple actions in the same function: constructing
proper route attributes and handling iterations over multiple fibs, for the
non-zero net.add_addr_allfibs use case. It notably increases the code complexity.
3) dstaddr handling. flags parameter re-uses RTF_ flags. As there is no special flag
for p2p connections, host routes and p2p routes are handled in the same way.
Additionally, mapping IFA flags to RTF flags makes the interface pretty messy.
It make rtinit() to clash with ifa_mainain_loopback_route() for IPV4 interface
aliases.
4) rtinit() is the last customer passing non-masked prefixes to rib_action(),
complicating rib_action() implementation.
5) rtinit() coupled ifa announce/withdrawal notifications, producing "false positive"
ifa messages in certain corner cases.
To address all these points, the following has been done:
* rtinit() has been split into multiple functions:
- Route attribute construction were moved to the per-address-family functions,
dealing with (2), (3) and (4).
- funnction providing net.add_addr_allfibs handling and route rtsock notificaions
is the new routing table inteface.
- rtsock ifa notificaion has been moved out as well. resulting set of funcion are only
responsible for the actual route notifications.
Side effects:
* /32 alias does not result in interface routes (/32 route and "host" route)
* RTF_PINNED is now set for IPv6 prefixes corresponding to the interface addresses
Differential revision: https://reviews.freebsd.org/D28186
When timestamp support has been negotiated, TCP segements received
without a timestamp should be discarded. However, there are broken
TCP implementations (for example, stacks used by Omniswitch 63xx and
64xx models), which send TCP segments without timestamps although
they negotiated timestamp support.
This patch adds a sysctl variable which tolerates such TCP segments
and allows to interoperate with broken stacks.
Reviewed by: jtl@, rscheff@
Differential Revision: https://reviews.freebsd.org/D28142
Sponsored by: Netflix, Inc.
PR: 252449
MFC after: 1 week
A TCP RST segment should be processed even it is missing TCP
timestamps.
Reported by: dmgk@, kevans@
Reviewed by: rscheff@, dmgk@
Sponsored by: Netflix, Inc.
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D28143
Relevant inet/inet6 code has the control over deciding what
the RIB lookup function currently is. With that in mind,
explicitly set it to the current value (rn_match) in the
datapath lookups. This avoids cost on indirect call.
Differential Revision: https://reviews.freebsd.org/D28066
Currently default behaviour is to keep only 1 packet per unresolved entry.
Ability to queue more than one packet was added 10 years ago, in r215207,
though the default value was kep intact.
Things have changed since that time. Systems tend to initiate multiple
connections at once for a variety of reasons.
For example, recent kern/252278 bug report describe happy-eyeball DNS
behaviour sending multiple requests to the DNS server.
The primary driver for upper value for the queue length determination is
memory consumption. Remote actors should not be able to easily exhaust
local memory by sending packets to unresolved arp/ND entries.
For now, bump value to 16 packets, to match Darwin implementation.
The proper approach would be to switch the limit to calculate memory
consumption instead of packet count and limit based on memory.
We should MFC this with a variation of D22447.
Reviewers: #manpages, #network, bz, emaste
Reviewed By: emaste, gbe(doc), jilles(doc)
MFC after: 1 month
Differential Revision: https://reviews.freebsd.org/D28068
This change introduces framework that allows to dynamically
attach or detach longest prefix match (lpm) lookup algorithms
to speed up datapath route tables lookups.
Framework takes care of handling initial synchronisation,
route subscription, nhop/nhop groups reference and indexing,
dataplane attachments and fib instance algorithm setup/teardown.
Framework features automatic algorithm selection, allowing for
picking the best matching algorithm on-the-fly based on the
amount of routes in the routing table.
Currently framework code is guarded under FIB_ALGO config option.
An idea is to enable it by default in the next couple of weeks.
The following algorithms are provided by default:
IPv4:
* bsearch4 (lockless binary search in a special IP array), tailored for
small-fib (<16 routes)
* radix4_lockless (lockless immutable radix, re-created on every rtable change),
tailored for small-fib (<1000 routes)
* radix4 (base system radix backend)
* dpdk_lpm4 (DPDK DIR24-8-based lookups), lockless datastrucure, optimized
for large-fib (D27412)
IPv6:
* radix6_lockless (lockless immutable radix, re-created on every rtable change),
tailed for small-fib (<1000 routes)
* radix6 (base system radix backend)
* dpdk_lpm6 (DPDK DIR24-8-based lookups), lockless datastrucure, optimized
for large-fib (D27412)
Performance changes:
Micro benchmarks (I7-7660U, single-core lookups, 2048k dst, code in D27604):
IPv4:
8 routes:
radix4: ~20mpps
radix4_lockless: ~24.8mpps
bsearch4: ~69mpps
dpdk_lpm4: ~67 mpps
700k routes:
radix4_lockless: 3.3mpps
dpdk_lpm4: 46mpps
IPv6:
8 routes:
radix6_lockless: ~20mpps
dpdk_lpm6: ~70mpps
100k routes:
radix6_lockless: 13.9mpps
dpdk_lpm6: 57mpps
Forwarding benchmarks:
+ 10-15% IPv4 forwarding performance (small-fib, bsearch4)
+ 25% IPv4 forwarding performance (full-view, dpdk_lpm4)
+ 20% IPv6 forwarding performance (full-view, dpdk_lpm6)
Control:
Framwork adds the following runtime sysctls:
List algos
* net.route.algo.inet.algo_list: bsearch4, radix4_lockless, radix4
* net.route.algo.inet6.algo_list: radix6_lockless, radix6, dpdk_lpm6
Debug level (7=LOG_DEBUG, per-route)
net.route.algo.debug_level: 5
Algo selection (currently only for fib 0):
net.route.algo.inet.algo: bsearch4
net.route.algo.inet6.algo: radix6_lockless
Support for manually changing algos in non-default fib will be added
soon. Some sysctl names will be changed in the near future.
Differential Revision: https://reviews.freebsd.org/D27401
In order to efficiently serve web traffic on a NUMA
machine, one must avoid as many NUMA domain crossings as
possible. With SO_REUSEPORT_LB, a number of workers can share a
listen socket. However, even if a worker sets affinity to a core
or set of cores on a NUMA domain, it will receive connections
associated with all NUMA domains in the system. This will lead to
cross-domain traffic when the server writes to the socket or
calls sendfile(), and memory is allocated on the server's local
NUMA node, but transmitted on the NUMA node associated with the
TCP connection. Similarly, when the server reads from the socket,
he will likely be reading memory allocated on the NUMA domain
associated with the TCP connection.
This change provides a new socket ioctl, TCP_REUSPORT_LB_NUMA. A
server can now tell the kernel to filter traffic so that only
incoming connections associated with the desired NUMA domain are
given to the server. (Of course, in the case where there are no
servers sharing the listen socket on some domain, then as a
fallback, traffic will be hashed as normal to all servers sharing
the listen socket regardless of domain). This allows a server to
deal only with traffic that is local to its NUMA domain, and
avoids cross-domain traffic in most cases.
This patch, and a corresponding small patch to nginx to use
TCP_REUSPORT_LB_NUMA allows us to serve 190Gb/s of kTLS encrypted
https media content from dual-socket Xeons with only 13% (as
measured by pcm.x) cross domain traffic on the memory controller.
Reviewed by: jhb, bz (earlier version), bcr (man page)
Tested by: gonzo
Sponsored by: Netfix
Differential Revision: https://reviews.freebsd.org/D21636
