Line data Source code
1 : /* $OpenBSD: buf.h,v 1.107 2018/08/13 15:26:17 visa Exp $ */
2 : /* $NetBSD: buf.h,v 1.25 1997/04/09 21:12:17 mycroft Exp $ */
3 :
4 : /*
5 : * Copyright (c) 1982, 1986, 1989, 1993
6 : * The Regents of the University of California. All rights reserved.
7 : * (c) UNIX System Laboratories, Inc.
8 : * All or some portions of this file are derived from material licensed
9 : * to the University of California by American Telephone and Telegraph
10 : * Co. or Unix System Laboratories, Inc. and are reproduced herein with
11 : * the permission of UNIX System Laboratories, Inc.
12 : *
13 : * Redistribution and use in source and binary forms, with or without
14 : * modification, are permitted provided that the following conditions
15 : * are met:
16 : * 1. Redistributions of source code must retain the above copyright
17 : * notice, this list of conditions and the following disclaimer.
18 : * 2. Redistributions in binary form must reproduce the above copyright
19 : * notice, this list of conditions and the following disclaimer in the
20 : * documentation and/or other materials provided with the distribution.
21 : * 3. Neither the name of the University nor the names of its contributors
22 : * may be used to endorse or promote products derived from this software
23 : * without specific prior written permission.
24 : *
25 : * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
26 : * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
27 : * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
28 : * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
29 : * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
30 : * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
31 : * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
32 : * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
33 : * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
34 : * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
35 : * SUCH DAMAGE.
36 : *
37 : * @(#)buf.h 8.7 (Berkeley) 1/21/94
38 : */
39 :
40 : #ifndef _SYS_BUF_H_
41 : #define _SYS_BUF_H_
42 : #include <sys/queue.h>
43 : #include <sys/tree.h>
44 : #include <sys/mutex.h>
45 :
46 : #define NOLIST ((struct buf *)0x87654321)
47 :
48 : struct buf;
49 : struct vnode;
50 :
51 : LIST_HEAD(bufhead, buf);
52 :
53 : /*
54 : * To avoid including <ufs/ffs/softdep.h>
55 : */
56 :
57 : LIST_HEAD(workhead, worklist);
58 :
59 : /*
60 : * Buffer queues
61 : */
62 : #define BUFQ_NSCAN_N 128
63 : #define BUFQ_FIFO 0
64 : #define BUFQ_NSCAN 1
65 : #define BUFQ_DEFAULT BUFQ_NSCAN
66 : #define BUFQ_HOWMANY 2
67 :
68 : /*
69 : * Write limits for bufq - defines high and low water marks for how
70 : * many kva slots are allowed to be consumed to parallelize writes from
71 : * the buffer cache from any individual bufq.
72 : */
73 : #define BUFQ_HI 128
74 : #define BUFQ_LOW 64
75 :
76 : struct bufq_impl;
77 :
78 : struct bufq {
79 : SLIST_ENTRY(bufq) bufq_entries;
80 : struct mutex bufq_mtx;
81 : void *bufq_data;
82 : u_int bufq_outstanding;
83 : u_int bufq_hi;
84 : u_int bufq_low;
85 : int bufq_waiting;
86 : int bufq_stop;
87 : int bufq_type;
88 : const struct bufq_impl *bufq_impl;
89 : };
90 :
91 : int bufq_init(struct bufq *, int);
92 : int bufq_switch(struct bufq *, int);
93 : void bufq_destroy(struct bufq *);
94 :
95 : void bufq_queue(struct bufq *, struct buf *);
96 : struct buf *bufq_dequeue(struct bufq *);
97 : void bufq_requeue(struct bufq *, struct buf *);
98 : int bufq_peek(struct bufq *);
99 : void bufq_drain(struct bufq *);
100 :
101 : void bufq_wait(struct bufq *);
102 : void bufq_done(struct bufq *, struct buf *);
103 : void bufq_quiesce(void);
104 : void bufq_restart(void);
105 :
106 : /* fifo */
107 : SIMPLEQ_HEAD(bufq_fifo_head, buf);
108 : struct bufq_fifo {
109 : SIMPLEQ_ENTRY(buf) bqf_entries;
110 : };
111 :
112 : /* nscan */
113 : SIMPLEQ_HEAD(bufq_nscan_head, buf);
114 : struct bufq_nscan {
115 : SIMPLEQ_ENTRY(buf) bqf_entries;
116 : };
117 :
118 : /* bufq link in struct buf */
119 : union bufq_data {
120 : struct bufq_fifo bufq_data_fifo;
121 : struct bufq_nscan bufq_data_nscan;
122 : };
123 :
124 : /*
125 : * These are currently used only by the soft dependency code, hence
126 : * are stored once in a global variable. If other subsystems wanted
127 : * to use these hooks, a pointer to a set of bio_ops could be added
128 : * to each buffer.
129 : */
130 : extern struct bio_ops {
131 : void (*io_start)(struct buf *);
132 : void (*io_complete)(struct buf *);
133 : void (*io_deallocate)(struct buf *);
134 : void (*io_movedeps)(struct buf *, struct buf *);
135 : int (*io_countdeps)(struct buf *, int, int);
136 : } bioops;
137 :
138 : /* The buffer header describes an I/O operation in the kernel. */
139 : struct buf {
140 : RBT_ENTRY(buf) b_rbbufs; /* vnode "hash" tree */
141 : LIST_ENTRY(buf) b_list; /* All allocated buffers. */
142 : LIST_ENTRY(buf) b_vnbufs; /* Buffer's associated vnode. */
143 : TAILQ_ENTRY(buf) b_freelist; /* Free list position if not active. */
144 : int cache; /* which cache are we in */
145 : struct proc *b_proc; /* Associated proc; NULL if kernel. */
146 : volatile long b_flags; /* B_* flags. */
147 : long b_bufsize; /* Allocated buffer size. */
148 : long b_bcount; /* Valid bytes in buffer. */
149 : size_t b_resid; /* Remaining I/O. */
150 : int b_error; /* Errno value. */
151 : dev_t b_dev; /* Device associated with buffer. */
152 : caddr_t b_data; /* associated data */
153 : void *b_saveaddr; /* Original b_data for physio. */
154 :
155 : TAILQ_ENTRY(buf) b_valist; /* LRU of va to reuse. */
156 :
157 : union bufq_data b_bufq;
158 : struct bufq *b_bq; /* What bufq this buf is on */
159 :
160 : struct uvm_object *b_pobj; /* Object containing the pages */
161 : off_t b_poffs; /* Offset within object */
162 :
163 : daddr_t b_lblkno; /* Logical block number. */
164 : daddr_t b_blkno; /* Underlying physical block number. */
165 : /* Function to call upon completion.
166 : * Will be called at splbio(). */
167 : void (*b_iodone)(struct buf *);
168 : struct vnode *b_vp; /* Device vnode. */
169 : int b_dirtyoff; /* Offset in buffer of dirty region. */
170 : int b_dirtyend; /* Offset of end of dirty region. */
171 : int b_validoff; /* Offset in buffer of valid region. */
172 : int b_validend; /* Offset of end of valid region. */
173 : struct workhead b_dep; /* List of filesystem dependencies. */
174 : };
175 :
176 : TAILQ_HEAD(bufqueue, buf);
177 :
178 : struct bufcache {
179 : int64_t hotbufpages;
180 : int64_t warmbufpages;
181 : int64_t cachepages;
182 : struct bufqueue hotqueue;
183 : struct bufqueue coldqueue;
184 : struct bufqueue warmqueue;
185 : };
186 :
187 : /* Device driver compatibility definitions. */
188 : #define b_active b_bcount /* Driver queue head: drive active. */
189 :
190 : /*
191 : * These flags are kept in b_flags.
192 : */
193 : #define B_WRITE 0x00000000 /* Write buffer (pseudo flag). */
194 : #define B_AGE 0x00000001 /* Move to age queue when I/O done. */
195 : #define B_NEEDCOMMIT 0x00000002 /* Needs committing to stable storage */
196 : #define B_ASYNC 0x00000004 /* Start I/O, do not wait. */
197 : #define B_BAD 0x00000008 /* Bad block revectoring in progress. */
198 : #define B_BUSY 0x00000010 /* I/O in progress. */
199 : #define B_CACHE 0x00000020 /* Bread found us in the cache. */
200 : #define B_CALL 0x00000040 /* Call b_iodone from biodone. */
201 : #define B_DELWRI 0x00000080 /* Delay I/O until buffer reused. */
202 : #define B_DONE 0x00000100 /* I/O completed. */
203 : #define B_EINTR 0x00000200 /* I/O was interrupted */
204 : #define B_ERROR 0x00000400 /* I/O error occurred. */
205 : #define B_INVAL 0x00000800 /* Does not contain valid info. */
206 : #define B_NOCACHE 0x00001000 /* Do not cache block after use. */
207 : #define B_PHYS 0x00002000 /* I/O to user memory. */
208 : #define B_RAW 0x00004000 /* Set by physio for raw transfers. */
209 : #define B_READ 0x00008000 /* Read buffer. */
210 : #define B_WANTED 0x00010000 /* Process wants this buffer. */
211 : #define B_WRITEINPROG 0x00020000 /* Write in progress. */
212 : #define B_XXX 0x00040000 /* Debugging flag. */
213 : #define B_DEFERRED 0x00080000 /* Skipped over for cleaning */
214 : #define B_SCANNED 0x00100000 /* Block already pushed during sync */
215 : #define B_PDAEMON 0x00200000 /* I/O started by pagedaemon */
216 : #define B_RELEASED 0x00400000 /* free this buffer after its kvm */
217 : #define B_WARM 0x00800000 /* buffer is or has been on the warm queue */
218 : #define B_COLD 0x01000000 /* buffer is on the cold queue */
219 : #define B_BC 0x02000000 /* buffer is managed by the cache */
220 : #define B_DMA 0x04000000 /* buffer is DMA reachable */
221 :
222 : #define B_BITS "\20\001AGE\002NEEDCOMMIT\003ASYNC\004BAD\005BUSY" \
223 : "\006CACHE\007CALL\010DELWRI\011DONE\012EINTR\013ERROR" \
224 : "\014INVAL\015NOCACHE\016PHYS\017RAW\020READ" \
225 : "\021WANTED\022WRITEINPROG\023XXX(FORMAT)\024DEFERRED" \
226 : "\025SCANNED\026DAEMON\027RELEASED\030WARM\031COLD\032BC\033DMA"
227 :
228 : /*
229 : * Zero out the buffer's data area.
230 : */
231 : #define clrbuf(bp) { \
232 : bzero((bp)->b_data, (u_int)(bp)->b_bcount); \
233 : (bp)->b_resid = 0; \
234 : }
235 :
236 :
237 : /* Flags to low-level allocation routines. */
238 : #define B_CLRBUF 0x01 /* Request allocated buffer be cleared. */
239 : #define B_SYNC 0x02 /* Do all allocations synchronously. */
240 :
241 : struct cluster_info {
242 : daddr_t ci_lastr; /* last read (read-ahead) */
243 : daddr_t ci_lastw; /* last write (write cluster) */
244 : daddr_t ci_cstart; /* start block of cluster */
245 : daddr_t ci_lasta; /* last allocation */
246 : int ci_clen; /* length of current cluster */
247 : int ci_ralen; /* Read-ahead length */
248 : daddr_t ci_maxra; /* last readahead block */
249 : };
250 :
251 : #ifdef _KERNEL
252 : __BEGIN_DECLS
253 : /* Kva slots (of size MAXPHYS) reserved for syncer and cleaner. */
254 : #define RESERVE_SLOTS 4
255 : /* Buffer cache pages reserved for syncer and cleaner. */
256 : #define RESERVE_PAGES (RESERVE_SLOTS * MAXPHYS / PAGE_SIZE)
257 : /* Minimum size of the buffer cache, in pages. */
258 : #define BCACHE_MIN (RESERVE_PAGES * 2)
259 : #define UNCLEAN_PAGES (bcstats.numbufpages - bcstats.numcleanpages)
260 :
261 : extern struct proc *cleanerproc;
262 : extern long bufpages; /* Max number of pages for buffers' data */
263 : extern struct pool bufpool;
264 : extern struct bufhead bufhead;
265 :
266 : void bawrite(struct buf *);
267 : void bdwrite(struct buf *);
268 : void biodone(struct buf *);
269 : int biowait(struct buf *);
270 : int bread(struct vnode *, daddr_t, int, struct buf **);
271 : int breadn(struct vnode *, daddr_t, int, daddr_t *, int *, int,
272 : struct buf **);
273 : void brelse(struct buf *);
274 : #define bremfree bufcache_take
275 : void bufinit(void);
276 : void buf_dirty(struct buf *);
277 : void buf_undirty(struct buf *);
278 : void buf_adjcnt(struct buf *, long);
279 : int bwrite(struct buf *);
280 : struct buf *getblk(struct vnode *, daddr_t, int, int, int);
281 : struct buf *geteblk(size_t);
282 : struct buf *incore(struct vnode *, daddr_t);
283 :
284 : /*
285 : * bufcache functions
286 : */
287 : void bufcache_take(struct buf *);
288 : void bufcache_release(struct buf *);
289 :
290 : int buf_flip_high(struct buf *);
291 : void buf_flip_dma(struct buf *);
292 : struct buf *bufcache_getcleanbuf(int, int);
293 : struct buf *bufcache_getanycleanbuf(void);
294 : struct buf *bufcache_getdirtybuf(void);
295 :
296 : /*
297 : * buf_kvm_init initializes the kvm handling for buffers.
298 : * buf_acquire sets the B_BUSY flag and ensures that the buffer is
299 : * mapped in the kvm.
300 : * buf_release clears the B_BUSY flag and allows the buffer to become
301 : * unmapped.
302 : * buf_unmap is for internal use only. Unmaps the buffer from kvm.
303 : */
304 : void buf_mem_init(vsize_t);
305 : void buf_acquire(struct buf *);
306 : void buf_acquire_unmapped(struct buf *);
307 : void buf_acquire_nomap(struct buf *);
308 : void buf_map(struct buf *);
309 : void buf_release(struct buf *);
310 : int buf_dealloc_mem(struct buf *);
311 : void buf_fix_mapping(struct buf *, vsize_t);
312 : void buf_alloc_pages(struct buf *, vsize_t);
313 : void buf_free_pages(struct buf *);
314 :
315 : void minphys(struct buf *bp);
316 : int physio(void (*strategy)(struct buf *), dev_t dev, int flags,
317 : void (*minphys)(struct buf *), struct uio *uio);
318 : void brelvp(struct buf *);
319 : void reassignbuf(struct buf *);
320 : void bgetvp(struct vnode *, struct buf *);
321 :
322 : void buf_replacevnode(struct buf *, struct vnode *);
323 : void buf_daemon(void *);
324 : void buf_replacevnode(struct buf *, struct vnode *);
325 : int bread_cluster(struct vnode *, daddr_t, int, struct buf **);
326 :
327 : #ifdef DEBUG
328 : void buf_print(struct buf *);
329 : #endif
330 :
331 : static __inline void
332 0 : buf_start(struct buf *bp)
333 : {
334 0 : if (bioops.io_start)
335 0 : (*bioops.io_start)(bp);
336 0 : }
337 :
338 : static __inline void
339 0 : buf_complete(struct buf *bp)
340 : {
341 0 : if (bioops.io_complete)
342 0 : (*bioops.io_complete)(bp);
343 0 : }
344 :
345 : static __inline void
346 0 : buf_deallocate(struct buf *bp)
347 : {
348 0 : if (bioops.io_deallocate)
349 0 : (*bioops.io_deallocate)(bp);
350 0 : }
351 :
352 : static __inline void
353 : buf_movedeps(struct buf *bp, struct buf *bp2)
354 : {
355 : if (bioops.io_movedeps)
356 : (*bioops.io_movedeps)(bp, bp2);
357 : }
358 :
359 : static __inline int
360 0 : buf_countdeps(struct buf *bp, int i, int islocked)
361 : {
362 0 : if (bioops.io_countdeps)
363 0 : return ((*bioops.io_countdeps)(bp, i, islocked));
364 : else
365 0 : return (0);
366 0 : }
367 :
368 : void cluster_write(struct buf *, struct cluster_info *, u_quad_t);
369 :
370 : __END_DECLS
371 : #endif /* _KERNEL */
372 : #endif /* !_SYS_BUF_H_ */
|
