freebsd-nq/share/man/man9/buf_ring.9

134 lines
3.7 KiB
Groff
Raw Normal View History

.\" Copyright (c) 2009 Bitgravity Inc
.\" Written by: Kip Macy <kmacy@FreeBSD.org>
.\" 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.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR 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 AUTHOR 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.
.\"
.\" $FreeBSD$
.\"
The drbr(9) API appeared to be so unclear, that most drivers in tree used it incorrectly, which lead to inaccurate overrated if_obytes accounting. The drbr(9) used to update ifnet stats on drbr_enqueue(), which is not accurate since enqueuing doesn't imply successful processing by driver. Dequeuing neither mean that. Most drivers also called drbr_stats_update() which did accounting again, leading to doubled if_obytes statistics. And in case of severe transmitting, when a packet could be several times enqueued and dequeued it could have been accounted several times. o Thus, make drbr(9) API thinner. Now drbr(9) merely chooses between ALTQ queueing or buf_ring(9) queueing. - It doesn't touch the buf_ring stats any more. - It doesn't touch ifnet stats anymore. - drbr_stats_update() no longer exists. o buf_ring(9) handles its stats itself: - It handles br_drops itself. - br_prod_bytes stats are dropped. Rationale: no one ever reads them but update of a common counter on every packet negatively affects performance due to excessive cache invalidation. - buf_ring_enqueue_bytes() reduced to buf_ring_enqueue(), since we no longer account bytes. o Drivers handle their stats theirselves: if_obytes, if_omcasts. o mlx4(4), igb(4), em(4), vxge(4), oce(4) and ixv(4) no longer use drbr_stats_update(), and update ifnet stats theirselves. o bxe(4) was the most correct driver, it didn't call drbr_stats_update(), thus it was the only driver accurate under moderate load. Now it also maintains stats itself. o ixgbe(4) had already taken stats from hardware, so just - drop software stats updating. - take multicast packet count from hardware as well. o mxge(4) just no longer needs NO_SLOW_STATS define. o cxgb(4), cxgbe(4) need no change, since they obtain stats from hardware. Reviewed by: jfv, gnn
2012-09-28 18:28:27 +00:00
.Dd September 27, 2012
.Dt BUF_RING 9
.Os
.Sh NAME
.Nm buf_ring ,
.Nm buf_ring_alloc ,
.Nm buf_ring_free ,
.Nm buf_ring_enqueue ,
.Nm buf_ring_dequeue_mc ,
.Nm buf_ring_dequeue_sc ,
.Nm buf_ring_count ,
.Nm buf_ring_empty ,
.Nm buf_ring_full ,
.Nm buf_ring_peek ,
.Nd multi-producer, {single, multi}-consumer lock-less ring buffer
.Sh SYNOPSIS
.In sys/param.h
.In sys/buf_ring.h
.Ft struct buf_ring *
.Fn buf_ring_alloc "int count" "struct malloc_type *type" "int flags" "struct mtx *sc_lock"
.Ft void
.Fn buf_ring_free "struct buf_ring *br" "struct malloc_type *type"
.Ft int
.Fn buf_ring_enqueue "struct buf_ring *br" "void *buf"
.Ft void *
.Fn buf_ring_dequeue_mc "struct buf_ring *br"
.Ft void *
.Fn buf_ring_dequeue_sc "struct buf_ring *br"
.Ft int
.Fn buf_ring_count "struct buf_ring *br"
.Ft int
.Fn buf_ring_empty "struct buf_ring *br"
.Ft int
.Fn buf_ring_full "struct buf_ring *br"
.Ft void *
.Fn buf_ring_peek "struct buf_ring *br"
.Sh DESCRIPTION
The
.Nm
functions provide a lock-less multi-producer and lock-less multi-consumer as
well as single-consumer ring buffer.
.Pp
The
.Fn buf_ring_alloc
function is used to allocate a buf_ring ring buffer with
.Fa count
slots using malloc_type
.Fa type
and memory flags
.Fa flags .
The single consumer interface is protected by
.Fa sc_lock .
.Pp
The
.Fn buf_ring_free
function is used to free a buf_ring.
The user is responsible for freeing any enqueued items.
.Pp
The
.Fn buf_ring_enqueue
function is used to enqueue a buffer to a buf_ring.
.Pp
The
.Fn buf_ring_dequeue_mc
function is a multi-consumer safe way of dequeueing elements from a buf_ring.
.Pp
The
.Fn buf_ring_dequeue_sc
function is a single-consumer interface to dequeue elements - requiring
the user to serialize accesses with a lock.
.Pp
The
.Fn buf_ring_count
function returns the number of elements in a buf_ring.
.Pp
The
.Fn buf_ring_empty
function returns
.Dv TRUE
if the buf_ring is empty,
.Dv FALSE
otherwise.
.Pp
The
.Fn buf_ring_full
function returns
.Dv TRUE
if no more items can be enqueued,
.Dv FALSE
otherwise.
.Pp
The
.Fn buf_ring_peek
function returns a pointer to the last element in the buf_ring if the
buf_ring is not empty,
.Dv NULL
otherwise.
.Sh RETURN VALUES
The
.Fn buf_ring_enqueue
The drbr(9) API appeared to be so unclear, that most drivers in tree used it incorrectly, which lead to inaccurate overrated if_obytes accounting. The drbr(9) used to update ifnet stats on drbr_enqueue(), which is not accurate since enqueuing doesn't imply successful processing by driver. Dequeuing neither mean that. Most drivers also called drbr_stats_update() which did accounting again, leading to doubled if_obytes statistics. And in case of severe transmitting, when a packet could be several times enqueued and dequeued it could have been accounted several times. o Thus, make drbr(9) API thinner. Now drbr(9) merely chooses between ALTQ queueing or buf_ring(9) queueing. - It doesn't touch the buf_ring stats any more. - It doesn't touch ifnet stats anymore. - drbr_stats_update() no longer exists. o buf_ring(9) handles its stats itself: - It handles br_drops itself. - br_prod_bytes stats are dropped. Rationale: no one ever reads them but update of a common counter on every packet negatively affects performance due to excessive cache invalidation. - buf_ring_enqueue_bytes() reduced to buf_ring_enqueue(), since we no longer account bytes. o Drivers handle their stats theirselves: if_obytes, if_omcasts. o mlx4(4), igb(4), em(4), vxge(4), oce(4) and ixv(4) no longer use drbr_stats_update(), and update ifnet stats theirselves. o bxe(4) was the most correct driver, it didn't call drbr_stats_update(), thus it was the only driver accurate under moderate load. Now it also maintains stats itself. o ixgbe(4) had already taken stats from hardware, so just - drop software stats updating. - take multicast packet count from hardware as well. o mxge(4) just no longer needs NO_SLOW_STATS define. o cxgb(4), cxgbe(4) need no change, since they obtain stats from hardware. Reviewed by: jfv, gnn
2012-09-28 18:28:27 +00:00
function return
.Er ENOBUFS
if there are no available slots in the buf_ring.
.Sh HISTORY
These functions were introduced in
.Fx 8.0 .