this is start on documenting bufqs. it covers the bits disk drivers need

to know. the grubby bits the midlayer touches on them will get written up
soon.

ok deraadt@

diff --git a/share/man/man9/Makefile b/share/man/man9/Makefile
index c69a893..5ca494d 100644 (file)
--- a/share/man/man9/Makefile
+++ b/share/man/man9/Makefile
@@ -1,4 +1,4 @@
-#      $OpenBSD: Makefile,v 1.217 2014/07/13 23:49:40 uebayasi Exp $
+#      $OpenBSD: Makefile,v 1.218 2014/07/24 01:18:23 dlg Exp $
 #      $NetBSD: Makefile,v 1.4 1996/01/09 03:23:01 thorpej Exp $
 
 #      Makefile for section 9 (kernel function and variable) manual pages.
@@ -7,7 +7,8 @@ MAN=    aml_evalnode.9 atomic_add_int.9 atomic_cas_uint.9 \
        atomic_dec_int.9 atomic_inc_int.9 atomic_setbits_int.9 \
        atomic_sub_int.9 atomic_swap_uint.9 \
        audio.9 autoconf.9 \
-       bemtoh32.9 bio_register.9 boot.9 buffercache.9 bus_dma.9 bus_space.9 \
+       bemtoh32.9 bio_register.9 boot.9 buffercache.9 bufq_init.9 bus_dma.9 \
+       bus_space.9 \
        copy.9 config_attach.9 crypto.9 delay.9 \
        disk.9 disklabel.9 dma_alloc.9 dohooks.9 dopowerhooks.9 \
        domountroothooks.9 doshutdownhooks.9 dostartuphooks.9 \
@@ -71,6 +72,10 @@ MLINKS+=buffercache.9 bread.9 buffercache.9 breadn.9 \
        buffercache.9 geteblk.9 buffercache.9 incore.9 \
        buffercache.9 allocbuf.9 buffercache.9 brelse.9 \
        buffercache.9 biodone.9 buffercache.9 biowait.9
+MLINKS+=bufq_init.9 bufq_switch.9 bufq_init.9 bufq_destroy.9 \
+       bufq_init.9 bufq_queue.9 bufq_init.9 bufq_dequeue.9 \
+       bufq_init.9 bufq_requeue.9 bufq_init.9 bufq_peek.9 \
+       bufq_init.9 bufq_drain.9
 MLINKS+=bus_dma.9 bus_dmamap_create.9 bus_dma.9 bus_dmamap_destroy.9 \
        bus_dma.9 bus_dmamap_load.9 bus_dma.9 bus_dmamap_load_mbuf.9 \
        bus_dma.9 bus_dmamap_load_uio.9 bus_dma.9 bus_dmamap_load_raw.9 \

diff --git a/share/man/man9/bufq_init.9 b/share/man/man9/bufq_init.9
new file mode 100644 (file)
index 0000000..ae2a183
--- /dev/null
+++ b/share/man/man9/bufq_init.9
@@ -0,0 +1,150 @@
+.\"    $OpenBSD: bufq_init.9,v 1.1 2014/07/24 01:18:23 dlg Exp $
+.\"
+.\" Copyright (c) 2013 David Gwynne <dlg@openbsd.org>
+.\"
+.\" Permission to use, copy, modify, and distribute this software for any
+.\" purpose with or without fee is hereby granted, provided that the above
+.\" copyright notice and this permission notice appear in all copies.
+.\"
+.\" THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
+.\" WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
+.\" MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
+.\" ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
+.\" WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
+.\" ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
+.\" OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
+.\"
+.Dd $Mdocdate: July 24 2014 $
+.Dt BUFQ_INIT 9
+.Os
+.Sh NAME
+.Nm bufq_init ,
+.Nm bufq_switch ,
+.Nm bufq_destroy ,
+.Nm bufq_queue ,
+.Nm bufq_dequeue ,
+.Nm bufq_requeue ,
+.Nm bufq_peek ,
+.Nm bufq_drain
+.\" .Nm bufq_wait ,
+.\" .Nm bufq_done ,
+.\" .Nm bufq_quiesce , 
+.\" .Nm bufq_restart
+.Nd buf queues
+.Sh SYNOPSIS
+.In sys/buf.h
+.Ft int
+.Fn "bufq_init" "struct bufq *bufq" "int type"
+.Ft int
+.Fn "bufq_switch" "struct bufq *bufq" "int type"
+.Ft void
+.Fn "bufq_destroy" "struct bufq *bufq"
+.Ft void
+.Fn "bufq_queue" "struct bufq *bufq" "struct buf *bp"
+.Ft struct buf *
+.Fn "bufq_dequeue" "struct bufq *bufq"
+.Ft void
+.Fn "bufq_requeue" "struct bufq *bufq" "struct buf *bp"
+.Ft int
+.Fn "bufq_peek" "struct bufq *bufq"
+.Ft void
+.Fn "bufq_drain" "struct bufq *bufq"
+.Sh DESCRIPTION
+The bufq API implements queueing and scheduling of I/O operations on disk
+devices.
+It provides multiple scheduling algorithms within the API.
+.Pp
+It is the responsibility of the disk device driver to provide
+pre-allocated bufq structures.
+.Pp
+.Fn bufq_init
+initialiases the
+.Fa bufq
+structure and allocates any state required by the scheduling algorithm
+by the
+.Fa type
+argument.
+.Pp
+.Fn bufq_switch
+can be used to change the scheduler currently used by
+.Fa bufq
+to the algorithm specified by
+.Fa type .
+.Pp
+.Fn bufq_destroy
+frees any state that was used by the scheduler.
+.Pp
+.Fn bufq_queue
+is used to add the buf specified by
+.Fa bp
+to the
+.Fa bufq
+queue.
+.Pp
+.Fn bufq_dequeue
+is used to get the next buf the
+.Fa bufq
+has scheduled to be serviced by a disk.
+The buf will be removed from the queue.
+.Pp
+.Fn bufq_requeue
+can be used to return a previously dequeued buf specified by
+.Fa bp
+to the head of the queue of
+.Fa bufq .
+.Pp
+.Fn bufq_peek
+allows the caller to determine if there are more bufs queued on
+.Fa bufq
+without modifying the list of bufs.
+.Pp
+.Fn bufq_drain
+is a convenience function for devices that have detached.
+It removes all the bufs currently queued on
+.Fa bufq ,
+marks them as failed with an
+.Dv ENXIO
+error, and returns them to the block layer via
+.Xr biodone 9 .
+.Sh CONTEXT
+.Fn bufq_init ,
+.Fn bufq_switch ,
+and
+.Fn bufq_destroy
+can be called during autoconf, or from process context.
+.Pp
+.Nm bufq_queue ,
+.Nm bufq_dequeue ,
+.Nm bufq_requeue ,
+.Nm bufq_peek ,
+and
+.Nm bufq_drain
+can be called during autoconf, from process context, or from interrupt context.
+.Sh RETURN VALUES
+.Fn bufq_init
+and
+.Fn bufq_switch
+will return 0 on success, or an error code as per
+.Xr errno 2 .
+.Pp
+.Fn bufq_dequeue
+returns the next buf that is scheduled to be serviced by the disk.
+.Dv NULL
+is returned If there are no bufs available on the queue.
+.Pp
+.Fn bufq_peek
+returns 1 if there are bufs available to be scheduled on the disk, otherwise 0.
+.Sh SEE ALSO
+.Xr errno 2 ,
+.Xr autoconf 9 ,
+.Xr biodone 9 ,
+.Xr disk 9
+.Sh HISTORY
+The bufq API was written by
+.An Thordur I. Bjornsson
+and
+.An David Gwynne Aq Mt dlg@openbsd.org .
+The bufq API first appeared in
+.Ox 4.8
+and finally succeeded in fully replacing disksort in
+.Ox 5.5 .