freebsd-nq/usr.sbin/cxgbetool/cxgbetool.8

789 lines
23 KiB
Groff
Raw Normal View History

.\" Copyright (c) 2015, 2018 Chelsio Inc
.\" All rights reserved.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions are met:
.\"
.\" 1. Redistributions of source code must retain the above copyright notice,
.\" this list of conditions and the following disclaimer.
.\"
.\" 2. Redistributions in binary form must reproduce the above copyright
.\" notice, this list of conditions and the following disclaimer in the
.\" documentation and/or other materials provided with the distribution.
.\"
.\" 3. Neither the name of the Chelsio Inc nor the names of its
.\" contributors may be used to endorse or promote products derived from
.\" this software without specific prior written permission.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
.\" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
.\" ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
.\" LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
.\" CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
.\" SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
.\" INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
.\" CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
.\" ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
.\" POSSIBILITY OF SUCH DAMAGE.
.\"
.\" * Other names and brands may be claimed as the property of others.
.\"
.\" $FreeBSD$
.\"
.Dd Sep 21, 2018
.Dt CXGBETOOL 8
.Os
.Sh NAME
.Nm cxgbetool
.Nd Userspace companion to
.Xr cxgbe 4
.Sh SYNOPSIS
.Bl -item -compact
.It
.Nm Ar nexus command Op Ar parameter ...
.Pp
.It
.Nm Ar nexus Cm clearstats Ar port_id
.It
.Nm Ar nexus Cm context Bro Cm ingress | egress | fl | cong Brc Ar cntxt_id
.It
.Nm Ar nexus Cm hashfilter mode
.It
.Nm Ar nexus Cm hashfilter Ar filter-specification
.It
.Nm Ar nexus Cm hashfilter Ar idx Cm delete
.It
.Nm Ar nexus Cm hashfilter list
.It
.Nm Ar nexus Cm filter mode
.It
.Nm Ar nexus Cm filter Ar idx Ar filter-specification
.It
.Nm Ar nexus Cm filter Ar idx Cm delete Op Cm prio Bro Cm 0 | 1 Brc
.It
.Nm Ar nexus Cm filter list
.It
.Nm Ar nexus Cm i2c Ar port_id devaddr addr Op Ar len
.It
.Nm Ar nexus Cm loadcfg Ar fw-config.txt
.It
.Nm Ar nexus Cm loadcfg clear
.It
.Nm Ar nexus Cm loadfw Ar fw-image.bin
.It
.Nm Ar nexus Cm memdump Ar addr len
.It
cxgbe(4): Add support for Connection Offload Policy (aka COP). COP allows fine-grained control on whether to offload a TCP connection using t4_tom, and what settings to apply to a connection selected for offload. t4_tom must still be loaded and IFCAP_TOE must still be enabled for full TCP offload to take place on an interface. The difference is that IFCAP_TOE used to be the only knob and would enable TOE for all new connections on the inteface, but now the driver will also consult the COP, if any, before offloading to the hardware TOE. A policy is a plain text file with any number of rules, one per line. Each rule has a "match" part consisting of a socket-type (L = listen, A = active open, P = passive open, D = don't care) and a pcap-filter(7) expression, and a "settings" part that specifies whether to offload the connection or not and the parameters to use if so. The general format of a rule is: [socket-type] expr => settings Example. See cxgbetool(8) for more information. [L] ip && port http => offload [L] port 443 => !offload [L] port ssh => offload [P] src net 192.168/16 && dst port ssh => offload !nagle !timestamp cong newreno [P] dst port ssh => offload !nagle ecn cong tahoe [P] dst port http => offload [A] dst port 443 => offload tls [A] dst net 192.168/16 => offload !timestamp cong highspeed The driver processes the rules for each new listen, active open, or passive open and stops at the first match. There is an implicit rule at the end of every policy that prohibits offload when no rule in the policy matches: [D] all => !offload This is a reworked and expanded version of a patch submitted by Krishnamraju Eraparaju @ Chelsio. Sponsored by: Chelsio Communications
2018-04-14 19:07:56 +00:00
.Nm Ar nexus Cm policy Ar cop.txt
.It
.Nm Ar nexus Cm policy clear
.It
.Nm Ar nexus Bro Cm reg | reg64 Brc Ar addr Ns Op Ar =val
.It
.Nm Ar nexus Cm regdump Op Ar register-block ...
.It
.Nm Ar nexus Cm sched-class Ar sub-command Op Ar param Ar value
.It
.Nm Ar nexus Cm sched-queue Ar port Ar queue Ar class
.It
.Nm Ar nexus Cm stdio
.It
.Nm Ar nexus Cm tcb Ar tid
.El
.Sh DESCRIPTION
.Nm
provides command-line access to features and debug facilities exported by
.Xr cxgbe 4
via private ioctls.
The target nexus device,
.Va t4nex%d
, is always the first argument.
(The parent nexus for an Ethernet port
.Va cxgbe%d
is listed in
.Va dev.cxgbe.%d.%parent
in the
.Xr sysctl 8
MIB).
The rest consists of a command and any parameters required by that command.
.Ss Commands
.Bl -ohang
.It Cm clearstats Ar port_id
Clear all transmit, receive, and error statistics of all queues associated
with a port.
The total number of ports attached to a nexus is listed in
.Va dev.t4nex.%d.nports
and the 0 based
.Ar port_id
identifies a port within this range.
.Pp
.Bl -item -compact
.It
.Cm context ingress Ar ingress_cntxt_id
.It
.Cm context cong Ar ingress_cntxt_id
.It
.Cm context egress Ar egress_cntxt_id
.It
.Cm context fl Ar flm_cntxt_id
.El
Display hardware context for an ingress queue, congestion manager, egress
queue, or freelist manager.
.Bl -tag -width ingress_cntxt_id -compact
.It Ar ingress_cntxt_id
context id of an ingress queue -- the value listed in one of
.Va dev.t4nex.%d.fwq.cntxt_id Ns ,
.Va dev.cxgbe.%d.rxq.%d.cntxt_id Ns ,
or
.Va dev.cxgbe.%d.ofld_rxq.%d.cntxt_id Ns .
.It Ar egress_cntxt_id
context id of an egress queue -- the value listed in one of
.Va dev.t4nex.%d.mgmtq.cntxt_id Ns ,
.Va dev.cxgbe.%d.txq.%d.cntxt_id Ns ,
.Va dev.cxgbe.%d.ctrlq.%d.cntxt_id Ns ,
.Va dev.cxgbe.%d.ofld_txq.%d.cntxt_id Ns ,
.Va dev.cxgbe.%d.rxq.%d.fl.cntxt_id Ns ,
or
.Va dev.cxgbe.%d.ofld_rxq.%d.fl.cntxt_id Ns .
Note that freelists are egress queues too.
.It Ar flm_cntxt_id
context id of a freelist manager.
The FLM context id is displayed in the
egress context dump of a freelist as FLMcontextID.
.El
.Pp
.Bl -item -compact
.It
.Cm hashfilter mode
.It
.Cm filter mode
.El
Display a list of match-criteria available for use in filter rules.
A full list of match-criteria known to the chip is in the table below but not
all can be used together and the firmware sets up the available parameters based
on "filterMode" in the configuration file.
Every filter must conform to the filter mode -- multiple match criteria per
filter are allowed but only from among those in the current setting of the
filter mode.
The filter mode for hash filters is a subset of that for normal TCAM filters and
depends on the "filterMask" setting in the firmware configuration file.
Hash filters do not support masked matches and an exact value for every
parameter in the output of "hashfilter mode" (except ipv4/ipv6) must be provided
when creating a hash filter.
.Pp
(Note that
.Ar mask
defaults to all 1s when not provided explicitly.
Hash filters do not support masked matches.
Also note that many of the items being matched are discrete numeric
values rather than bit fields and should be masked with caution.)
.TS
center expand;
cb cb cbw(40m)
cb l l.
Criteria Usage Matches if ...
_
ipv4 T{
.Cm type ipv4
T} T{
incoming packet is an IPv4 datagram.
T}
_
ipv6 T{
.Cm type ipv6
T} T{
incoming packet is an IPv6 datagram.
T}
_
sip T{
.Cm sip Ar addr Ns Op / Ns Ar mask
T} T{
bitwise and of the source address in an incoming IP datagram with
.Ar mask
equals
.Ar addr Ns .
.Ar addr
can be an IPv4 or IPv6 address.
T}
_
dip T{
.Cm dip Ar addr Ns Op / Ns Ar mask
T} T{
bitwise and of the destination address in an incoming IP datagram with
.Ar mask
equals
.Ar addr Ns .
.Ar addr
can be an IPv4 or IPv6 address.
T}
_
sport T{
.Cm sport Ar port Ns Op : Ns Ar mask
T} T{
bitwise and of the source port in an incoming TCP or UDP datagram with
.Ar mask
equals
.Ar port Ns .
T}
_
dport T{
.Cm dport Ar port Ns Op : Ns Ar mask
T} T{
bitwise and of the destination port in an incoming TCP or UDP datagram with
.Ar mask
equals
.Ar port Ns .
T}
_
fcoe T{
.Cm fcoe Brq 0 | 1
T} T{
incoming frame is Fibre Channel over Ethernet(1) or not(0).
T}
_
iport T{
.Cm iport Ar val Ns Op : Ns Ar mask
T} T{
bitwise and of the ingress port with
.Ar mask
equals
.Ar val Ns .
The ingress port is a 3 bit number that identifies the port on which a
frame arrived.
Physical ports are numbered 0-3 and 4-7 are internal loopback paths
within the chip.
Note that ingress port is not a bit field so it is not always possible
to match an arbitrary subset of ingress ports with a single filter rule.
T}
_
ovlan T{
.Cm ovlan Ar tag Ns Op : Ns Ar mask
T} T{
bitwise and of the 16-bit outer VLAN tag of an incoming frame with
.Ar mask
equals
.Ar tag Ns .
T}
_
vlan T{
.Cm vlan Ar tag Ns Op : Ns Ar mask
T} T{
bitwise and of the 16-bit VLAN tag of an incoming QinQ frame with
.Ar mask
equals
.Ar tag Ns .
The inner VLAN tag is used if the incoming frame is QinQ.
T}
_
tos T{
.Cm tos Ar val Ns Op : Ns Ar mask
T} T{
bitwise and of the 8-bit IP Type of Service/IPv6 Traffic Class in an
incoming packet with
.Ar mask
equals
.Ar val Ns .
T}
_
proto T{
.Cm proto Ar ipproto Ns Op : Ns Ar mask
T} T{
bitwise and of the 8-bit IP protocol in an incoming packet with
.Ar mask
equals
.Ar ipproto Ns .
T}
_
ethtype T{
.Cm ethtype Ar type Ns Op : Ns Ar mask
T} T{
bitwise and of the 16-bit Ethernet type field of an incoming frame with
.Ar mask
equals
.Ar type Ns .
T}
_
macidx T{
.Cm macidx Ar idx Ns Op : Ns Ar mask
T} T{
bitwise and of the MAC Address Match Index of an incoming frame with
.Ar mask
equals
.Ar idx Ns .
The MAC Address Match Index refers to an entry in the MPS TCAM or in the
MPS hash. See
.Cm matchtype
for more information.
T}
_
matchtype T{
.Cm matchtype Ar type Ns Op : Ns Ar mask
T} T{
bitwise and of the Match Type of an incoming frame with
.Ar mask
equals
.Ar idx Ns .
Match Type is one of the following:
.Bl -tag -width "n" -compact
.It 0
destination MAC in incoming frame is a unicast L2 address that is
programmed in the MPS TCAM.
.Cm macidx
can be used to match the index (and thus the MAC address) of the match
in the TCAM.
.It 1
destination MAC in incoming frame is a unicast L2 address that "hit" a
hash entry in the MPS hash table.
.Cm macidx
can be used to match the index of the entry in the MPS hash table.
.It 2
destination MAC in incoming frame is a multicast L2 address that is
programmed in the MPS TCAM.
.Cm macidx
can be used to match the index (and thus the MAC address) of the match
in the TCAM.
.It 3
destination MAC in incoming frame is a multicast L2 address that "hit"
an entry in the MPS hash table.
.It 4
interface on which incoming frame was received is in promiscuous mode
and the destination MAC in the frame is not a broadcast address, and
does not match in the MPS TCAM or the MPS hash either. (The frame would
have been discarded if the interface wasn't in promiscuous mode.)
.It 5
interface on which incoming frame was received is in promiscuous mode
and the destination MAC in the frame is not a broadcast address; it
wasn't looked up in the MPS TCAM or the MPS hash because the chip is
configured to give precedence to promiscuous mode classification.
.It 6
destination MAC in incoming frame is a broadcast address.
.It 7
Not documented. Do not use.
.El
T}
_
frag T{
.Cm frag Brq 0 | 1
T} T{
incoming frame is part of a fragmented IP datagram(1) or not(0).
T}
.TE
.Pp
.Bl -item -compact
.It
.Cm hashfilter Ar filter-specification
.It
.Cm filter Ar idx Ar filter-specification
.El
Program a filter.
.Pp
TCAM filters: The number of available filters is in
dev.<nexus>.<instance>.nfilters.
.Ar idx
must be an unused index between 0 and nfilters - 1.
IPv6 filters consume 4 consecutive entries on T4/T5 and 2 on T6 and
.Ar idx
must be aligned to 4 or 2 in this case.
.Pp
Hash filters: These reside in the card's memory instead of its TCAM and are
enabled with a special configuration file that is selected with
.Cm hw.cxgbe.config_file="hashfilter"
in loader.conf.
There are at least half a million filters available with the sample config
shipped with the driver.
Note that the hardware selects the index for a hashfilter and this index is
displayed when the filter is created.
Hash and TCAM filters can be used together.
.Pp
.Ar filter-specification
consists of one or more matches (see Usage in the table above) to try against an
incoming frame, an action to perform when all matches succeed, and some
additional operational parameters.
Hashfilters require an exact value for the 5-tuple (sip, dip, sport, dport,
proto) and for any other match-criteria listed in "hashfilter mode".
Possible filter actions are
.Cm drop Ns ,
.Cm pass Ns , or
.Cm switch Ns .
.Pp
.Bl -tag -width nat_dport -offset indent -compact
Operational parameters that can be used with all filters:
.It Cm hitcnts
Count filter hits: 0 or 1 (default).
.It Cm prio
Filter has priority over active and server regions of TCAM: 0 (default) or 1.
.El
.Pp
.Bl -tag -width nat_dport -offset indent -compact
Operational parameters that can be used with filters with
.Cm action pass Ns :
.It Cm queue
Context id of an ingress queue to which to deliver the packet.
The context id is available in
.Va dev.cxgbe.%d.rxq.%d.cntxt_id .
By default, packets that hit a filter with action pass are delivered based on
their RSS hash as usual.
Use this to steer them to a particular queue.
.It Cm rpttid
Report the filter tid instead of the RSS hash in the rx descriptor.
0 (default) or 1.
.It Cm tcbhash
Select TCB hash information in rx descriptor.
0 (default) or 1
.El
.Pp
.Bl -tag -width nat_dport -offset indent -compact
Operational parameters that can be used with filters with
.Cm action switch Ns :
.It Cm eport
Egress port number on which to send the packet matching the filter.
0 to dev.<nexus>.<instance>.nports - 1.
.It Cm dmac
Replace packet destination MAC address with the one provided before switching
it out of eport.
.It Cm smac
Replace packet source MAC address with the one provided before switching it
out of eport.
.It Cm swapmac
Swap packet source and destination MAC addresses before switching it out of
eport.
.It Cm vlan
Insert, remove, or rewrite the VLAN tag before switching the packet out of
eport.
.Cm none
removes the tag,
.Cm = Ns Ar tag
replaces the existing tag with the one provided, and
.Cm + Ns Ar tag
inserts the given tag into the frame.
.It Cm nat
Specify the desired NAT mode. Valid NAT modes values are:
.Bl -tag -width dip-dp-sip -compact
.It Cm dip
Perform NAT on destination IP.
.It Cm dip-dp
Perform NAT on destination IP, destination port.
.It Cm dip-dp-sip
Perform NAT on destination IP, destination port, source IP.
.It Cm dip-dp-sp
Perform NAT on destination IP, destination port, source port.
.It Cm sip-sp
Perform NAT on source IP, source port.
.It Cm dip-sip-sp
Perform NAT on destination IP, source IP, source port.
.It Cm all
Perform NAT on all 4-tuple fields.
.El
.It Cm natflag
Perform NAT only on segments which do not have TCP FIN or RST set.
.It Cm natseq
Perform NAT only if incoming segment's sequence number + payload length is less
than this supplied value.
.It Cm nat_dip
Perform NAT using this destination IP.
.It Cm nat_sip
Perform NAT using this source IP.
.It Cm nat_dport
Perform NAT using this destination port.
.It Cm nat_sport
Perform NAT using this source port.
Perform NAT only if incoming segment's sequence number + payload length is less
than this supplied value.
.El
.Pp
.Bl -item -compact
.It
.Cm hashfilter Ar idx Cm delete
.It
.Cm filter Ar idx Cm delete
.El
Delete filter that is at the given index.
.It Cm filter Cm list
List all filters programmed into the hardware.
.It Cm i2c Ar port_id devaddr addr Op Ar len
.It Cm loadcfg Ar fw-config.txt
Install the firmware configuration file contained in
.Ar fw-config.txt
to the card.
Set hw.cxgbe.config_file="flash" in loader.conf to get
.Xr cxgbe 4
to use the on-flash configuration.
.It Cm loadcfg Cm clear
Erase configuration file from the card.
.It Cm loadfw Ar fw-image.bin
Install the firmware contained in
.Ar fw-image.bin
to the card.
.It Cm memdump Ar addr len
Display
.Ar len
bytes of data of the card's memory starting at
.Ar addr Ns .
The card's memory map is available in
.Va dev.t4nex.%d.misc.meminfo Ns .
cxgbe(4): Add support for Connection Offload Policy (aka COP). COP allows fine-grained control on whether to offload a TCP connection using t4_tom, and what settings to apply to a connection selected for offload. t4_tom must still be loaded and IFCAP_TOE must still be enabled for full TCP offload to take place on an interface. The difference is that IFCAP_TOE used to be the only knob and would enable TOE for all new connections on the inteface, but now the driver will also consult the COP, if any, before offloading to the hardware TOE. A policy is a plain text file with any number of rules, one per line. Each rule has a "match" part consisting of a socket-type (L = listen, A = active open, P = passive open, D = don't care) and a pcap-filter(7) expression, and a "settings" part that specifies whether to offload the connection or not and the parameters to use if so. The general format of a rule is: [socket-type] expr => settings Example. See cxgbetool(8) for more information. [L] ip && port http => offload [L] port 443 => !offload [L] port ssh => offload [P] src net 192.168/16 && dst port ssh => offload !nagle !timestamp cong newreno [P] dst port ssh => offload !nagle ecn cong tahoe [P] dst port http => offload [A] dst port 443 => offload tls [A] dst net 192.168/16 => offload !timestamp cong highspeed The driver processes the rules for each new listen, active open, or passive open and stops at the first match. There is an implicit rule at the end of every policy that prohibits offload when no rule in the policy matches: [D] all => !offload This is a reworked and expanded version of a patch submitted by Krishnamraju Eraparaju @ Chelsio. Sponsored by: Chelsio Communications
2018-04-14 19:07:56 +00:00
.It Cm policy Ar cop.txt
Install the Connection Offload Policy (COP) in
.Ar cop.txt Ns .
A COP offers fine-grained control over which connections get offloaded and with
what parameters.
Set
.Cm hw.cxgbe.cop_managed_offloading="1"
in loader.conf to ensure that t4_tom will not offload any connection before a
COP is installed.
Note that t4_tom must be loaded and operational (IFCAP_TOE enabled) as always
for any kind of offload based on the hardware TOE.
.Bl -column -offset indent "COP installed" "cop_managed_offloading" "Behavior"
.It Sy COP installed Ta Sy cop_managed_offloading Ta Sy Behavior
.It NO Ta 0 Ta offload all [Default]
.It NO Ta 1 Ta no offload
.It YES Ta Don't Care Ta Rule based offload
.El
.Pp
The policy file consists of empty lines, comments (lines begining with #) and
any number of rules.
Rules are applied in the order they appear in the file and processing stops at
the first match.
There is an implicit rule that disables offload for connections that do not
match anything in the policy.
.Pp
Each rule consists of a filter part, which determines what connections the
rule applies to, and a settings part, which determines whether matching
cxgbe(4): Add support for Connection Offload Policy (aka COP). COP allows fine-grained control on whether to offload a TCP connection using t4_tom, and what settings to apply to a connection selected for offload. t4_tom must still be loaded and IFCAP_TOE must still be enabled for full TCP offload to take place on an interface. The difference is that IFCAP_TOE used to be the only knob and would enable TOE for all new connections on the inteface, but now the driver will also consult the COP, if any, before offloading to the hardware TOE. A policy is a plain text file with any number of rules, one per line. Each rule has a "match" part consisting of a socket-type (L = listen, A = active open, P = passive open, D = don't care) and a pcap-filter(7) expression, and a "settings" part that specifies whether to offload the connection or not and the parameters to use if so. The general format of a rule is: [socket-type] expr => settings Example. See cxgbetool(8) for more information. [L] ip && port http => offload [L] port 443 => !offload [L] port ssh => offload [P] src net 192.168/16 && dst port ssh => offload !nagle !timestamp cong newreno [P] dst port ssh => offload !nagle ecn cong tahoe [P] dst port http => offload [A] dst port 443 => offload tls [A] dst net 192.168/16 => offload !timestamp cong highspeed The driver processes the rules for each new listen, active open, or passive open and stops at the first match. There is an implicit rule at the end of every policy that prohibits offload when no rule in the policy matches: [D] all => !offload This is a reworked and expanded version of a patch submitted by Krishnamraju Eraparaju @ Chelsio. Sponsored by: Chelsio Communications
2018-04-14 19:07:56 +00:00
connections will be offloaded and, if so, with what settings.
The general form of a rule is
.Bl -ohang -offset indent
.It Cm \&[ Ar socket-type Cm \&] Ar pcap-filter Cm => Ar settings
.Pp
.Ar socket-type
is one of the following.
.Bl -tag -width "X" -compact
.It Sy A
Active open.
Connection is being opened by this host.
.It Sy P
Passive open.
Connection was requested by a peer.
.It Sy L
Listen called on a socket.
Disabling offload in such a rule will prevent a hardware listener from being started.
.It Sy D
Don't care.
Matches all of the above.
.El
.Pp
.Ar pcap-filter
is an expression that follows the
.Xr pcap-filter 7
syntax, or it is the keyword
.Cm all
that matches everything.
.Pp
.Ar settings
determine whether connections matching
.Ar socket-type
and
.Ar pcap-filter
are offloaded and optionally sets some per-connection properties if they are.
A combination of the following is allowed.
.Bl -tag -width "timestamp" -compact
.It Cm offload
Connection should be offloaded.
Use
.Cm !offload
or
.Cm not offload
to disable offload instead.
.It Cm coalesce
Enable rx payload coalescing.
Negate to disable.
.It Cm timestamp
Enable TCP timestamp option.
Negate to disable.
.It Cm sack
Enable TCP Selective Acknowledgements (SACK).
Negate to disable.
.It Cm nagle
Enable Nagle's algorithm.
Negate to disable.
.It Cm ecn
Enable Explicit Congestion Notification (ECN).
Negate to disable.
.It Cm ddp
Use Direct Data Placement (zero copy receive) and zero copy transmit on the
connection to service AIO requests on the socket.
Negate to disable.
.It Cm tls
Set ULP mode to ULP_MODE_TLS.
.It Cm cong Ar algo
Use the specified congestion control algorithm.
.Ar algo
must be one of
.Cm reno Ns , Cm tahoe Ns , Cm newreno Ns , or Cm highspeed Ns .
.It Cm class Ar sc
Bind the connection to the specified tx scheduling class.
Valid range is 0 to 14 (for T4) and 0 to 15 (T5 onwards).
.It Cm rxq Ar qid
Use the specified offload rx queue.
.Ar qid
should be between 0 and nofldrxq for the ifnet.
.It Cm txq Ar qnum
Use the specified offload tx queue.
.Ar qid
should be between 0 and nofldtxq for the ifnet.
.It Cm bind Ar qnum
Shorthand for
.Cm rxq Ar qnum Cm txq Ar qnum Ns .
Use only when nofldrxq is the same as nofldtxq.
.It Cm mss Ar val
Set the advertised TCP MSS in the SYN for this connection to
.Ar val
(in bytes).
The hardware MTU table must already have an entry that is suitable for the MSS.
.El
.Pp
.It Example of a COP.
Note that hardware listener for port 22 will be IPv4 only because the rule
before it will prevent any IPv6 servers other than the first two. Also note
that outgoing connections to 192.168/16 are the only outgoing connections that
will get offloaded.
.Bd -literal
[L] port 80 => offload
[L] port 443 => offload
[L] ip6 => !offload
[L] port 22 => offload
[P] dst port 80 => offload cong highspeed !sack !ecn
[P] dst port 443 => offload tls
[A] dst net 192.168/16 => offload
[A] all => !offload
[D] port 22 => offload !nagle
.Ed
.El
.It Cm policy clear
Remove the Connection Offload Policy (COP) if one is in use.
.It Bro Cm reg | reg64 Brc Ar addr Ns Op Ar =val
.It Cm regdump Op Ar register-block ...
Display contents of device registers.
One or more
.Ar register-block
can be specified to limit the registers displayed.
The default is to display registers for all blocks.
Registers with read side effects are not read during a
.Cm regdump
operation.
.Ar register-block
can be
.Cm sge pci dbg mc ma edc0 edc1 cim tp ulp_rx ulp_tx pmrx pmtx mps cplsw
.Cm smb i2c mi uart pmu sf pl le ncsi xgmac Ns .
.It Cm sched-class config Op Ar param Ar value
Configure optional feature capabilities for the TX scheduler.
.Bl -ohang -offset indent
.It Sy type Ar scheduler-type
Use packet for the packet scheduler.
.It Sy minmax Ar value
A non-zero value will enable "minmax" mode; a zero value will disable "minmax" mode.
.Pp
NOTE: Many (most) of the parameters and constraints are adapter-specific
- for instance the number of channels and classes which are available
whether various modes are implemented, etc.
Consult the adapter documentation for specific information on any limitations.
.El
.It Cm sched-class params Op Ar param Ar value
Configure parameters for a scheduling class.
.Bl -ohang -offset indent
.It Sy type Ar scheduler-type
Use packet for packet scheduler.
.It Sy level Ar scheduler-hierarchy-level
The "level" within the scheduling hierarchy which is being programed:
.Pp
.Bl -tag -width "cl-wrr" -compact -offset indent
.It Sy cl-rl
Class Rate Limiting.
.Pp
.It Sy cl-wrr
Class Weighted Round Robin.
.Pp
2018-04-15 19:23:37 +00:00
.It Sy ch-rl
Channel Rate Limiting.
.El
.It Sy mode Ar scheduler-mode
The mode in which the scheduling class is going to operate:
.Pp
.Bl -tag -width "class" -compact -offset indent
.It Sy class
All of the "flows" bound to the scheduling class will be held to aggregate scheduling constraints.
.Pp
.It Sy flow
Each of the "flows" bound to the scheduling class will be held to the scheduling constraints.
.El
.Pp
E.g. if the scheduling class has a TX bandwidth of 10Mb/s, in
.Cm class
mode, all of the "flows" bound to the class would be limited to an
aggregate bandwidth of 10Mb/s; but in
.Cm flow
mode, each of the "flows" bound to the scheduling class would be limited to 10Mb/s.
.It Sy rate-unit Ar scheduler-rate-unit
The units of the scheduler rate constraints:
.Pp
.Bl -tag -width "bits" -compact -offset indent
.It Sy bits
bit rate in Kb/s.
.Pp
.It Sy pkts
packets/s.
.El
.It Sy rate-mode Ar scheduler-rate-mode
The mode of the scheduler rate constraints:
.Pp
.Bl -tag -width "relative" -compact -offset indent
.It Sy relative
percent of port rate.
.Pp
.It Sy absolute
Kb/s.
.El
.It Sy channel Ar scheduler-channel-index
The scheduling channel to which the scheduling class will be bound.
.It Sy class Ar scheduler-class-index
The scheduling class being programmed.
.It Sy min-rate Ar minimum-rate
The minimum guaranteed rate to which a rate-limiting scheduling class hierarchy will have access.
.It Sy max-rate Ar maximum-rate
The maximum rate for a rate-limiting scheduling class hierarchy.
.It Sy weight Ar round-robin-weight
The weight to be used for a weighted-round-robin scheduling hierarchy.
.It Sy pkt-size Ar average-packet-size
The average packet size will be used to compute scheduler constraints for a rate-limited scheduler class hierarchy.
.Pp
NOTE: Many (most) of the parameters and constraints are adapter-specific - for instance the number of channels and classes which are available,
whether various modes are implemented, etc.
Consult the adapter documentation for specific information on any limitations.
.El
.It Cm sched-queue Ar port queue class
Bind the indicated port's NIC TX
.Ar queue
to the specified TX Scheduler
.Ar class.
If the TX
.Ar queue
is
.Cm all, *
or any negative value, the binding will apply to
all of the TX queues associated with the
.Ar interface.
If the class is
.Cm unbind, clear
or any negative value, the TX queue(s) will be unbound from
any current TX Scheduler Class binding.
.It Cm stdio
Switch to interactive mode.
.It Cm tcb Ar tid
Display contents of the hardware TCB (TCP Control Block) for the
connection identfied by
.Ar tid Ns .
.El
.Sh FILES
/sys/dev/cxgbe/t4_ioctl.h
.Sh AUTHORS
This manual page was written by
.An Navdeep Parhar Aq np@FreeBSD.org .