From c2a8ce017ec196b69e182a229e2d66771cc94c83 Mon Sep 17 00:00:00 2001 From: Kenjiro Cho Date: Thu, 12 Jul 2001 12:15:03 +0000 Subject: add altq.9 that describes ALTQ queue macros. --- share/man/man9/altq.9 | 571 ++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 571 insertions(+) create mode 100644 share/man/man9/altq.9 (limited to 'share') diff --git a/share/man/man9/altq.9 b/share/man/man9/altq.9 new file mode 100644 index 00000000000..44d1b92c656 --- /dev/null +++ b/share/man/man9/altq.9 @@ -0,0 +1,571 @@ +.\" $OpenBSD: altq.9,v 1.1 2001/07/12 12:15:02 kjc Exp $ +.\" +.\" Copyright (C) 2001 +.\" Sony Computer Science Laboratories 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. +.\" +.\" THIS SOFTWARE IS PROVIDED BY SONY CSL 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 SONY CSL 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. +.\" +.Dd July 10, 2001 +.Dt ALTQ 9 +.Os KAME +.\" +.Sh NAME +.Nm ALTQ +.Nd kernel interfaces for manipulating output queues on network interfaces +.Sh SYNOPSIS +.Fd #include +.Fd #include +.Fd #include +.Ft void \"macro +.Fn IFQ_ENQUEUE "struct ifaltq *ifq" "struct mbuf *m" "int error" +.Ft void \"macro +.Fn IFQ_DEQUEUE "struct ifaltq *ifq" "struct mbuf *m" +.Ft void \"macro +.Fn IFQ_POLL "struct ifaltq *ifq" "struct mbuf *m" +.Ft void \"macro +.Fn IFQ_PURGE "struct ifaltq *ifq" +.Ft void \"macro +.Fn IFQ_CLASSIFY "struct ifaltq *ifq" "struct mbuf *m" "int af" "struct altq_pktattr *pktattr" +.Ft void \"macro +.Fn IFQ_IS_EMPTY "struct ifaltq *ifq" +.Ft void \"macro +.Fn IFQ_SET_MAXLEN "struct ifaltq *ifq" "int len" +.Ft void \"macro +.Fn IFQ_INC_LEN "struct ifaltq *ifq" +.Ft void \"macro +.Fn IFQ_DEC_LEN "struct ifaltq *ifq" +.Ft void \"macro +.Fn IFQ_INC_DROPS "struct ifaltq *ifq" +.Ft void \"macro +.Fn IFQ_SET_READY "struct ifaltq *ifq" +.Sh DESCRIPTION +The ALTQ system is a framework to manage queueing disciplines on network +interfaces. ALTQ introduces new macros to manipulate output queues. +The output queue macros are used to abstract queue operations and not to +touch the internal fields of the output queue structure. +The macros are independent from the ALTQ implementation, and compatible with the +traditional +.Dv ifqueue +macros for ease of transition. +.Pp +.Fn IFQ_ENQUEUE +enqueues a packet +.Fa m , +to the queue +.Fa ifq . +The underlying queueing discipline may discard the packet. +0 is set to +.Fa error +on success and +.Dv ENOBUFS +is set if the packet is discarded. +.Fa m +will be freed by the device driver on success or by the queueing discipline on +failure so that the caller should not touch +.Fa m +after calling +.Fn IFQ_ENQUEUE . +.Pp +.Fn IFQ_DEQUEUE +dequeues a packet from the queue. The dequeued packet is set to +.Fa m , +or +.Dv NULL +is set if no packet is dequeued. +The caller must always check +.Fa m +since a non-empty queue could return +.Dv NULL +under rate-limiting. +.Pp +.Fn IFQ_POLL +returns the next packet without removing it from the queue. +It is guaranteed by the underlying queueing discipline that +.Fn IFQ_DEQUEUE +immediately after +.Fn IFQ_POLL +returns the same packet. +.Pp +.Fn IFQ_PURGE +discards all the packets in the queue. +The purge operation is needed since a non-work conserving queue cannot be +emptied by a dequeue loop. +.Pp +.Fn IFQ_CLASSIFY +classifies a packet to a scheduling class, and set the result to +.Fa pktattr . +.Pp +.Fn IFQ_IS_EMPTY +can be used to check if the queue is empty. +Note that +.Fn IFQ_DEQUEUE +could still return +.Dv NULL +if the queueing discipline is non-work conserving. +.Pp +.Fn IFQ_SET_MAXLEN +sets the queue length limit to the default FIFO queue. +.Pp +.Fn IFQ_INC_LEN +and +.Fn IFQ_DEC_LEN +increment or decrement the current queue length in packets. +.Pp +.Fn IFQ_INC_DROPS +increments the drop counter and is equal to +.Fn IF_DROP . +It is defined for naming consistency. +.Pp +.Fn IFQ_SET_READY +sets a flag to indicate this driver is converted to use the new macros. +ALTQ can be enabled only on interfaces with this flag. +.Sh COMPATIBILITY +.Ss ifaltq structure +In order to keep compatibility with the existing code, the new +output queue structure +.Dv ifaltq +has the same fields. The traditional +.Fn IF_XXX +macros and the code directly referencing to the fields within +.Dv if_snd +still work with +.Dv ifaltq . +(Once we finish conversions of all the drivers, we no longer need +these fields.) +.Bd -literal + ##old-style## ##new-style## + | + struct ifqueue { | struct ifaltq { + struct mbuf *ifq_head; | struct mbuf *ifq_head; + struct mbuf *ifq_tail; | struct mbuf *ifq_tail; + int ifq_len; | int ifq_len; + int ifq_maxlen; | int ifq_maxlen; + int ifq_drops; | int ifq_drops; + }; | /* altq related fields */ + | ...... + | }; + | +.Ed +The new structure replaces +.Dv struct ifqueue +in +.Dv struct ifnet . +.Bd -literal + ##old-style## ##new-style## + | + struct ifnet { | struct ifnet { + .... | .... + | + struct ifqueue if_snd; | struct ifaltq if_snd; + | + .... | .... + }; | }; + | +.Ed +The (simplified) new +.Fn IFQ_XXX +macros looks like: +.Bd -literal + #ifdef ALTQ + #define IFQ_DEQUEUE(ifq, m) \e + if (ALTQ_IS_ENABLED((ifq)) \e + ALTQ_DEQUEUE((ifq), (m)); \e + else \e + IF_DEQUEUE((ifq), (m)); + #else + #define IFQ_DEQUEUE(ifq, m) IF_DEQUEUE((ifq), (m)); + #endif +.Ed +.Ss Enqueue operation +The semantics of the enqueue operation is changed. In the new style, +the enqueue and packet drop are combined since they cannot be easily +separated in many queueing disciplines. +The new enqueue operation corresponds to the following macro that is +written with the old macros. +.Bd -literal +#define IFQ_ENQUEUE(ifq, m, error) \e +do { \e + if (IF_QFULL((ifq))) { \e + m_freem((m)); \e + (error) = ENOBUFS; \e + IF_DROP(ifq); \e + } else { \e + IF_ENQUEUE((ifq), (m)); \e + (error) = 0; \e + } \e +} while (0) +.Ed +.Pp +.Fn IFQ_ENQUEUE +does the followings: +.Bl -hyphen -compact +.It +queue a packet +.It +drop (and free) a packet if the enqueue operation fails +.El +If the enqueue operation fails, an error +.Dv ENOBUFS +is set to +.Dv error . +.Dv mbuf +is freed by the queueing discipline. +The caller should not touch mbuf after calling +.Fn IFQ_ENQUEUE +so that the caller may need to copy +.Fa m_pkthdr.len +or +.Fa m_flags +field beforehand for statistics. +The caller should not use +.Fn senderr +since mbuf was already freed. +.Pp +The new style +.Fn if_output +looks as follows: +.Bd -literal + ##old-style## ##new-style## + | + int | int + ether_output(ifp, m0, dst, rt0) | ether_output(ifp, m0, dst, rt0) + { | { + ...... | ...... + | + | mflags = m->m_flags; + | len = m->m_pkthdr.len; + s = splimp(); | s = splimp(); + if (IF_QFULL(&ifp->if_snd)) { | IFQ_ENQUEUE(&ifp->if_snd, m, + | error); + IF_DROP(&ifp->if_snd); | if (error != 0) { + splx(s); | splx(s); + senderr(ENOBUFS); | retuen (error); + } | } + IF_ENQUEUE(&ifp->if_snd, m); | + ifp->if_obytes += | ifp->if_obytes += len; + m->m_pkthdr.len; | + if (m->m_flags & M_MCAST) | if (mflags & M_MCAST) + ifp->if_omcasts++; | ifp->if_omcasts++; + | + if ((ifp->if_flags & IFF_OACTIVE) | if ((ifp->if_flags & IFF_OACTIVE) + == 0) | == 0) + (*ifp->if_start)(ifp); | (*ifp->if_start)(ifp); + splx(s); | splx(s); + return (error); | return (error); + | + bad: | bad: + if (m) | if (m) + m_freem(m); | m_freem(m); + return (error); | return (error); + } | } + | +.Ed +.Ss Classifier +The classifier mechanism is currently implemented in +.Fn if_output . +.Dv struct altq_pktattr +is used to store the classifier result, and it is passed to the enqueue +function. +(We will change the method to tag the classifier result to mbuf in the future.) +.Bd -literal +int +ether_output(ifp, m0, dst, rt0) +{ + ...... + struct altq_pktattr pktattr; + + ...... + + /* classify the packet before prepending link-headers */ + IFQ_CLASSIFY(&ifp->if_snd, m, dst->sa_family, &pktattr); + + /* prepend link-level headers */ + ...... + + IFQ_ENQUEUE(&ifp->if_snd, m, &pktattr, error); + + ...... +} +.Ed +.Sh HOW TO CONVERT THE EXISTING DRIVERS +First, make sure the corresponding +.Fn if_output +is already converted to the new style. +.Pp +Look for +.Fa if_snd +in the driver. Probably, you need to make changes to the lines that include +.Fa if_snd . +.Ss Empty check operation +If the code checks +.Fa ifq_head +to see whether the queue is empty or not, use +.Fn IFQ_IS_EMPTY . +.Bd -literal + ##old-style## ##new-style## + | + if (ifp->if_snd.ifq_head != NULL) | if (IFQ_IS_EMPTY(&ifp->if_snd) == 0) + | +.Ed +Note that +.Fn IFQ_POLL +can be used for the same purpose, but +.Fn IFQ_POLL +could be costly for a complex scheduling algorithm since +.Fn IFQ_POLL +needs to run the scheduling algorithm to select the next packet. +On the other hand, +.Fn IFQ_EMPTY +checks only if there is any packet stored in the queue. +Another difference is that even when +.Fn IFQ_EMPTY +is +.Dv FALSE , +.Fn IFQ_DEQUEUE +could still return +.Dv NULL +if the queue is under rate-limiting. +.Ss Dequeue operaion +Replace +.Fn IF_DEQUEUE +by +.Fn IFQ_DEQUEUE . +Always check whether the dequeued mbuf is +.Dv NULL +or not. +Note that even when +.Fn IFQ_IS_EMPTY +is +.Dv FALSE , +.Fn IFQ_DEQUEUE +could return +.Dv NULL +due to rate-limiting. +.Bd -literal + ##old-style## ##new-style## + | + IF_DEQUEUE(&ifp->if_snd, m); | IFQ_DEQUEUE(&ifp->if_snd, m); + | if (m == NULL) + | return; + | +.Ed +A driver is supposed to call +.Fn if_start +from transmission complete interrupts in order to trigger the next dequeue. +.Ss Poll-and-dequeue operation +If the code polls the packet at the head of the queue and actually uses +the packet before dequeueing it, use +.Fn IFQ_POLL +and +IFQ_DEQUEUE . +.Bd -literal + ##old-style## ##new-style## + | + m = ifp->if_snd.ifq_head; | IFQ_POLL(&ifp->if_snd, m); + if (m != NULL) { | if (m != NULL) { + | + /* use m to get resources */ | /* use m to get resources */ + if (something goes wrong) | if (something goes wrong) + return; | return; + | + IF_DEQUEUE(&ifp->if_snd, m); | IFQ_DEQUEUE(&ifp->if_snd, m); + | + /* kick the hardware */ | /* kick the hardware */ + } | } + | +.Ed +It is guaranteed that +.Fn IFQ_DEQUEUE +immediately after +.Fn IFQ_POLL +returns the same packet. +(They need to be guarded by +.Fn splimp +if called from outside of +.Fn if_start . +) +.Ss Eliminating IF_PREPEND +If the code uses +.Fn IF_PREPEND , +you have to eliminate it since the prepend operation is not possible for many +queueing disciplines. A common use of +.Fn IF_PREPEND +is to cancel the previous dequeue operation. +You have to convert the logic into poll-and-dequeue. +.Bd -literal + ##old-style## ##new-style## + | + IF_DEQUEUE(&ifp->if_snd, m); | IFQ_POLL(&ifp->if_snd, m); + if (m != NULL) { | if (m != NULL) { + | + if (something_goes_wrong) { | if (something_goes_wrong) { + IF_PREPEND(&ifp->if_snd, m); | + return; | return; + } | } + | + | /* at this point, the driver + | * is committed to send this + | * packet. + | */ + | IFQ_DEQUEUE(&ifp->if_snd, m); + | + /* kick the hardware */ | /* kick the hardware */ + } | } + | +.Ed +.Ss Purge operation +Use +.Fn IFQ_PURGE +to empty the queue. +Note that a non-work conserving queue cannot be emptied by a dequeue loop. +.Bd -literal + ##old-style## ##new-style## + | + while (ifp->if_snd.ifq_head != NULL) {| IFQ_PURGE(&ifp->if_snd); + IF_DEQUEUE(&ifp->if_snd, m); | + m_freem(m); | + } | + | +.Ed +.Ss Attach routine +Use +.Fn IFQ_SET_MAXLEN +to set a value to +.Fa ifq_maxlen . +Add +.Fn IFQ_SET_READY +to show this driver is converted to the new style. +(this is used to distinguish new-style drivers.) +.Bd -literal + ##old-style## ##new-style## + | + ifp->if_snd.ifq_maxlen = qsize; | IFQ_SET_MAXLEN(&ifp->if_snd, qsize); + | IFQ_SET_READY(&ifp->if_snd); + if_attach(ifp); | if_attach(ifp); + | +.Ed +.Ss Other issues +The new macros for statistics: +.Bd -literal + ##old-style## ##new-style## + | + IF_DROP(&ifp->if_snd); | IFQ_INC_DROPS(&ifp->if_snd); + | + ifp->if_snd.ifq_len++; | IFQ_INC_LEN(&ifp->if_snd); + | + ifp->if_snd.ifq_len--; | IFQ_INC_LEN(&ifp->if_snd); + | +.Ed +Some drivers instruct the hardware to invoke transmission complete +interruts only when it thinks necessary. Rate-limiting breaks its assumption. +.Ss How to convert drivers using multiple ifqueues +Some (pseudo) devices (such as slip) have another +.Dv ifqueue +to prioritize packets. It is possible to eliminate the second queue +since ALTQ provides more flexible mechanisms but the following shows +how to keep the original behavior. +.Bd -literal +struct sl_softc { + struct ifnet sc_if; /* network-visible interface */ + ... + struct ifqueue sc_fastq; /* interactive output queue */ + ... +}; +.Ed +The driver doesn't compile in the new model since it has the following +line ( +.Fa if_snd +is no longer a type of +.Dv struct ifqueue . +) +.Bd -literal + struct ifqueue *ifq = &ifp->if_snd; +.Ed +A simple way is to use the original +.Fn IF_XXX +macros for +.Fa sc_fastq +and use the new +.Fn IFQ_XXX +macros for +.Fa if_snd . +The enqueue operation looks like: +.Bd -literal + ##old-style## ##new-style## + | + struct ifqueue *ifq = &ifp->if_snd; | struct ifqueue *ifq = NULL; + | + if (ip->ip_tos & IPTOS_LOWDELAY) | if ((ip->ip_tos & IPTOS_LOWDELAY) && + ifq = &sc->sc_fastq; | !ALTQ_IS_ENABLED(&sc->sc_if.if_snd)) { + | ifq = &sc->sc_fastq; + if (IF_QFULL(ifq)) { | if (IF_QFULL(ifq)) { + IF_DROP(ifq); | IF_DROP(ifq); + m_freem(m); | m_freem(m); + splx(s); | error = ENOBUFS; + sc->sc_if.if_oerrors++; | } else { + return (ENOBUFS); | IF_ENQUEUE(ifq, m); + } | error = 0; + IF_ENQUEUE(ifq, m); | } + | } else + | IFQ_ENQUEUE(&sc->sc_if.if_snd, + | m, error); + | + | if (error) { + | splx(s); + | sc->sc_if.if_oerrors++; + | return (error); + | } + if ((sc->sc_oqlen = | if ((sc->sc_oqlen = + sc->sc_ttyp->t_outq.c_cc) == 0) | sc->sc_ttyp->t_outq.c_cc) == 0) + slstart(sc->sc_ttyp); | slstart(sc->sc_ttyp); + splx(s); | splx(s); + | +.Ed +The dequeue operations looks like: +.Bd -literal + ##old-style## ##new-style## + | + s = splimp(); | s = splimp(); + IF_DEQUEUE(&sc->sc_fastq, m); | IF_DEQUEUE(&sc->sc_fastq, m); + if (m == NULL) | if (m == NULL) + IF_DEQUEUE(&sc->sc_if.if_snd, m); | IFQ_DEQUEUE(&sc->sc_if.if_snd, m); + splx(s); | splx(s); + | +.Ed +.Sh QUEUEING DISCIPLINES +Queueing disciplines need to maintain +.Fa ifq_len . +(used by +.Fn IFQ_IS_EMPTY +). +Queueing disciplines also need to guarantee the same mbuf is returned if +.Fn IFQ_DEQUEUE +is called immediately after +.Fn IFQ_POLL . +.Sh SEE ALSO +.Xr altq.conf 5 , +.Xr altqd 8 , +.Xr tbrconfig 8 +.Sh HISTORY +The +.Nm +system first appeared in March 1997. -- cgit v1.2.3