1/*-
2 * SPDX-License-Identifier: BSD-2-Clause
3 *
4 * Copyright (c) 2004 Poul-Henning Kamp
5 * All rights reserved.
6 *
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
9 * are met:
10 * 1. Redistributions of source code must retain the above copyright
11 *    notice, this list of conditions and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 *    notice, this list of conditions and the following disclaimer in the
14 *    documentation and/or other materials provided with the distribution.
15 *
16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19 * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
20 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26 * SUCH DAMAGE.
27 */
28
29/*
30 * Architectural notes:
31 *
32 * bufobj is a new object which is what buffers hang from in the buffer
33 * cache.
34 *
35 * This used to be vnodes, but we need non-vnode code to be able
36 * to use the buffer cache as well, specifically geom classes like
37 * raid3 and raid5.
38 *
39 * All vnodes will contain a bufobj initially, but down the road we may
40 * want to only allocate bufobjs when they are needed.  There could be a
41 * large number of vnodes in the system which wouldn't need a bufobj during
42 * their lifetime.
43 *
44 * The exact relationship to the vmobject is not determined at this point,
45 * it may in fact be that we find them to be two sides of the same object
46 * once things starts to crystalize.
47 */
48
49#ifndef _SYS_BUFOBJ_H_
50#define _SYS_BUFOBJ_H_
51
52#if defined(_KERNEL) || defined(_KVM_VNODE)
53
54#include <sys/queue.h>
55#include <sys/_lock.h>
56#include <sys/_rwlock.h>
57#include <sys/_pctrie.h>
58
59struct bufobj;
60struct buf_ops;
61
62extern struct buf_ops buf_ops_bio;
63
64TAILQ_HEAD(buflists, buf);
65
66/* A Buffer list & trie */
67struct bufv {
68	struct buflists	bv_hd;		/* Sorted blocklist */
69	struct pctrie	bv_root;	/* Buf trie */
70	int		bv_cnt;		/* Number of buffers */
71};
72
73typedef void b_strategy_t(struct bufobj *, struct buf *);
74typedef int b_write_t(struct buf *);
75typedef int b_sync_t(struct bufobj *, int waitfor);
76typedef void b_bdflush_t(struct bufobj *, struct buf *);
77
78struct buf_ops {
79	const char	*bop_name;
80	b_write_t	*bop_write;
81	b_strategy_t	*bop_strategy;
82	b_sync_t	*bop_sync;
83	b_bdflush_t	*bop_bdflush;
84};
85
86#define BO_STRATEGY(bo, bp)	((bo)->bo_ops->bop_strategy((bo), (bp)))
87#define BO_SYNC(bo, w)		((bo)->bo_ops->bop_sync((bo), (w)))
88#define BO_WRITE(bo, bp)	((bo)->bo_ops->bop_write((bp)))
89#define BO_BDFLUSH(bo, bp)	((bo)->bo_ops->bop_bdflush((bo), (bp)))
90
91/*
92 * Locking notes:
93 * 'S' is sync_mtx
94 * 'v' is the vnode lock which embeds the bufobj.
95 * '-' Constant and unchanging after initialization.
96 */
97struct bufobj {
98	struct rwlock	bo_lock;	/* Lock which protects "i" things */
99	struct buf_ops	*bo_ops;	/* - Buffer operations */
100	struct vm_object *bo_object;	/* v Place to store VM object */
101	LIST_ENTRY(bufobj) bo_synclist;	/* S dirty vnode list */
102	void		*bo_private;	/* private pointer */
103	struct bufv	bo_clean;	/* i Clean buffers */
104	struct bufv	bo_dirty;	/* i Dirty buffers */
105	int		bo_numoutput;	/* i Writes in progress */
106	u_int		bo_flag;	/* i Flags */
107	int		bo_domain;	/* - Clean queue affinity */
108	int		bo_bsize;	/* - Block size for i/o */
109};
110
111/*
112 * XXX BO_ONWORKLST could be replaced with a check for NULL list elements
113 * in v_synclist.
114 */
115#define	BO_ONWORKLST	(1 << 0)	/* On syncer work-list */
116#define	BO_WWAIT	(1 << 1)	/* Wait for output to complete */
117#define	BO_DEAD		(1 << 2)	/* Dead; only with INVARIANTS */
118#define	BO_NOBUFS	(1 << 3)	/* No bufs allowed */
119
120#define	BO_LOCKPTR(bo)		(&(bo)->bo_lock)
121#define	BO_LOCK(bo)		rw_wlock(BO_LOCKPTR((bo)))
122#define	BO_UNLOCK(bo)		rw_wunlock(BO_LOCKPTR((bo)))
123#define	BO_RLOCK(bo)		rw_rlock(BO_LOCKPTR((bo)))
124#define	BO_RUNLOCK(bo)		rw_runlock(BO_LOCKPTR((bo)))
125#define	ASSERT_BO_WLOCKED(bo)	rw_assert(BO_LOCKPTR((bo)), RA_WLOCKED)
126#define	ASSERT_BO_LOCKED(bo)	rw_assert(BO_LOCKPTR((bo)), RA_LOCKED)
127#define	ASSERT_BO_UNLOCKED(bo)	rw_assert(BO_LOCKPTR((bo)), RA_UNLOCKED)
128
129void bufobj_init(struct bufobj *bo, void *priv);
130void bufobj_wdrop(struct bufobj *bo);
131void bufobj_wref(struct bufobj *bo);
132void bufobj_wrefl(struct bufobj *bo);
133int bufobj_invalbuf(struct bufobj *bo, int flags, int slpflag, int slptimeo);
134int bufobj_wwait(struct bufobj *bo, int slpflag, int timeo);
135int bufsync(struct bufobj *bo, int waitfor);
136void bufbdflush(struct bufobj *bo, struct buf *bp);
137
138#endif /* defined(_KERNEL) || defined(_KVM_VNODE) */
139#endif /* _SYS_BUFOBJ_H_ */
140