d0ea47437a
things a bit: - use dpcpu data to track the ifps with packets queued up, - per-cpu locking and driver flags - along with .nh_drainedcpu and NETISR_POLICY_CPU. - Put the mbufs in flight reference count, preventing interfaces from going away, under INVARIANTS as this is a general problem of the stack and should be solved in if.c/netisr but still good to verify the internal queuing logic. - Permit changing the MTU to virtually everythinkg like we do for loopback. Hook epair(4) up to the build. Approved by: re (kib)
219 lines
6.8 KiB
Groff
219 lines
6.8 KiB
Groff
.\"
|
|
.\" Copyright (c) 2009 Robert N. M. Watson
|
|
.\" 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(s), this list of conditions and the following disclaimer as
|
|
.\" the first lines of this file unmodified other than the possible
|
|
.\" addition of one or more copyright notices.
|
|
.\" 2. Redistributions in binary form must reproduce the above copyright
|
|
.\" notice(s), this list of conditions and the following disclaimer in the
|
|
.\" documentation and/or other materials provided with the distribution.
|
|
.\"
|
|
.\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER(S) ``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 HOLDER(S) 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.
|
|
.\"
|
|
.\" $FreeBSD$
|
|
.\"
|
|
.Dd July 26, 2009
|
|
.Dt NETISR 9
|
|
.Os
|
|
.Sh NAME
|
|
.Nm netisr
|
|
.Nd Kernel network dispatch service
|
|
.Sh SYNOPSIS
|
|
.In net/netisr.h
|
|
.Ft void
|
|
.Fn netisr_register "const struct netisr_handler *nhp"
|
|
.Ft void
|
|
.Fn netisr_unregister "const struct netisr_handler *nhp"
|
|
.Ft int
|
|
.Fn netisr_dispatch "u_int proto" "struct mbuf *m"
|
|
.Ft int
|
|
.Fn netisr_dispatch_src "u_int proto" "uintptr_t source" "struct mbuf *m"
|
|
.Ft int
|
|
.Fn netisr_queue "u_int proto" "struct mbuf *m"
|
|
.Ft int
|
|
.Fn netisr_queue_src "u_int proto" "uintptr_t source" "struct mbuf *m"
|
|
.Ft void
|
|
.Fn netisr_clearqdrops "const struct netisr_handler *nhp"
|
|
.Ft void
|
|
.Fn netisr_getqdrops "const struct netisr_handler *nhp" "u_int64_t *qdropsp"
|
|
.Ft void
|
|
.Fn netisr_getqlimit "const struct netisr_handler *nhp" "u_int *qlimitp"
|
|
.Ft int
|
|
.Fn netisr_setqlimit "const struct netisr_handler *nhp" "u_int qlimit"
|
|
.Ft u_int
|
|
.Fn netisr_default_flow2cpu "u_int flowid"
|
|
.Ft u_int
|
|
.Fn netisr_get_cpucount "void"
|
|
.Ft u_int
|
|
.Fn netisr_get_cpuid "u_int cpunumber"
|
|
.Sh DESCRIPTION
|
|
The
|
|
.Nm
|
|
kernel interface suite allows device drivers (and other packet sources) to
|
|
direct packets to protocols for directly dispatched or deferred processing.
|
|
.Ss Protocol registration
|
|
Protocols register and unregister handlers using
|
|
.Fn netisr_register
|
|
and
|
|
.Fn netisr_unregister ,
|
|
and may also manage queue limits and statistics using the
|
|
.Fn netisr_clearqdrops ,
|
|
.Fn netisr_getqdrops ,
|
|
.Fn netisr_getqlimit ,
|
|
and
|
|
.Fn netisr_setqlimit.
|
|
.Pp
|
|
.Nm
|
|
supports multi-processor execution of handlers, and relies on a combination
|
|
of source ordering and protocol-specific ordering and work-placement
|
|
policies to decide how do distribute work across one or more worker
|
|
threads.
|
|
Registering protocols will declare one of three policies:
|
|
.Bl -tag -width NETISR_POLICY_SOURCE
|
|
.It Dv NETISR_POLICY_SOURCE
|
|
.Nm
|
|
should maintain source ordering without advice from the protocol.
|
|
.Nm
|
|
will ignore any flow IDs present on
|
|
.Vt mbuf
|
|
headers for the purposes of work placement.
|
|
.It Dv NETISR_POLICY_FLOW
|
|
.Nm
|
|
should maintain flow ordering as defined by the
|
|
.Vt mbuf
|
|
header flow ID field.
|
|
If the protocol implements
|
|
.Va nh_m2flow ,
|
|
then
|
|
.Nm
|
|
will query the protocol in the evet that the
|
|
.Vt mbuf
|
|
doesn't have a flow ID, falling back on source ordering.
|
|
.It NETISR_POLICY_CPU
|
|
.Nm
|
|
will entirely delegate all work placement decisions to the protocol,
|
|
querying
|
|
.Va nh_m2cpuid
|
|
for each packet.
|
|
.El
|
|
.Pp
|
|
Registration is declared using
|
|
.Vt "struct netisr_handler" ,
|
|
whose fields are defined as follows:
|
|
.Bl -tag -width "netisr_handler_t nh_handler"
|
|
.It Vt "const char *" Va nh_name
|
|
Unique character string name of the protocol, which may be included in
|
|
.Xr 2 sysctl
|
|
MIB names, so should not contain whitespace.
|
|
.It Vt netisr_handler_t Va nh_handler
|
|
Protocol handler function that will be invoked on each packet received for
|
|
the protocol.
|
|
.It Vt netisr_m2flow_t Va nh_m2flow
|
|
Optional protocol function to generate a flow ID and set
|
|
.Dv M_FLOWID
|
|
for packets that do not enter
|
|
.Nm
|
|
with
|
|
.Dv M_FLOWID
|
|
defined.
|
|
Will be used only with
|
|
.Dv NETISR_POLICY_FLOW .
|
|
.It Vt netisr_m2cpuid_t Va nh_m2cpuid
|
|
Protocol function to determine what CPU a packet should be processed on.
|
|
Will be used only with
|
|
.Dv NETISR_POLICY_CPU .
|
|
.It Vt netisr_drainedcpu_t Va nh_drainedcpu
|
|
Optional callback function that will be invoked when a per-CPU queue
|
|
was drained.
|
|
It will never fire for directly dispatched packets.
|
|
Unless fully understood, this special-purpose function should not be used.
|
|
.\" In case you intend to use this please send 50 chocolate bars to each
|
|
.\" of rwatson and bz and wait for an answer.
|
|
.It Vt u_int Va nh_proto
|
|
Protocol number used by both protocols to identify themselves to
|
|
.Nm ,
|
|
and by packet sources to select what handler will be used to process
|
|
packets.
|
|
A table of supported protocol numbers appears below.
|
|
For implementation reasons, protocol numbers great than 15 are currently
|
|
unsupported.
|
|
.It Vt u_int Va nh_qlimit
|
|
The maximum per-CPU queue depth for the protocol; due to internal
|
|
implementation details, the effective queue depth may be as much as twice
|
|
this number.
|
|
.It Vt u_int Va nh_policy
|
|
The ordering and work placement policy for the protocol, as described
|
|
earlier.
|
|
.El
|
|
.Ss Packet source interface
|
|
Packet sources, such as network interfaces, may request protocol processing
|
|
using the
|
|
.Fn netisr_dispatch
|
|
and
|
|
.Fn netisr_queue
|
|
interfaces.
|
|
Both accept a protocol number and
|
|
.Vt mbuf
|
|
argument, but while
|
|
.Fn netisr_queue
|
|
will always execute the protocol handler asynchonously in a deferred
|
|
context,
|
|
.Fn netisr_dispatch
|
|
will optionally direct dispatch if permitted by global and per-protocol
|
|
policy.
|
|
.Pp
|
|
In order to provide additional load balancing and flow information,
|
|
packet sources may also specify an opaque source identifier, which in
|
|
practice might be a network interface number or socket pointer, using
|
|
the
|
|
.Fn netisr_dispatch_src
|
|
and
|
|
.Fn netisr_queue_src
|
|
variants.
|
|
.Ss Protocol number constants
|
|
The follow protocol numbers are currently defined:
|
|
.Bl -tag -width NETISR_ATALK1
|
|
.It Dv NETISR_IP
|
|
IPv4
|
|
.It Dv NETISR_IGMP
|
|
IGMPv3 loopback
|
|
.It Dv NETISR_ROUTE
|
|
Routing socket loopback
|
|
.It Dv NETISR_AARP
|
|
Appletalk AARP
|
|
.It Dv NETISR_ATALK1
|
|
Appletalk phase 1
|
|
.It Dv NETISR_ATALK2
|
|
Appletalk phase 2
|
|
.It Dv NETISR_ARP
|
|
ARP
|
|
.It Dv NETISR_IPX
|
|
IPX/SPX
|
|
.It Dv NETISR_IPV6
|
|
IPv6
|
|
.It Dv NETISR_NATM
|
|
ATM
|
|
.It Dv NETISR_EPAIR
|
|
.Xr epair 4
|
|
.El
|
|
.Sh AUTHORS
|
|
This manual page and the
|
|
.Nm
|
|
implementation were written by
|
|
.An Robert N. M. Watson .
|