1/* $OpenBSD: buf.h,v 1.114 2024/02/03 18:51:58 beck 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#include <uvm/uvm_extern.h> 46 47#define NOLIST ((struct buf *)0x87654321) 48 49struct buf; 50struct vnode; 51 52LIST_HEAD(bufhead, buf); 53 54/* 55 * Buffer queues 56 */ 57#define BUFQ_NSCAN_N 128 58#define BUFQ_FIFO 0 59#define BUFQ_NSCAN 1 60#define BUFQ_DEFAULT BUFQ_NSCAN 61#define BUFQ_HOWMANY 2 62 63/* 64 * Write limits for bufq - defines high and low water marks for how 65 * many kva slots are allowed to be consumed to parallelize writes from 66 * the buffer cache from any individual bufq. 67 */ 68#define BUFQ_HI 128 69#define BUFQ_LOW 64 70 71struct bufq_impl; 72 73struct bufq { 74 SLIST_ENTRY(bufq) bufq_entries; 75 struct mutex bufq_mtx; 76 void *bufq_data; 77 u_int bufq_outstanding; 78 u_int bufq_hi; 79 u_int bufq_low; 80 int bufq_waiting; 81 int bufq_stop; 82 int bufq_type; 83 const struct bufq_impl *bufq_impl; 84}; 85 86int bufq_init(struct bufq *, int); 87int bufq_switch(struct bufq *, int); 88void bufq_destroy(struct bufq *); 89 90void bufq_queue(struct bufq *, struct buf *); 91struct buf *bufq_dequeue(struct bufq *); 92void bufq_requeue(struct bufq *, struct buf *); 93int bufq_peek(struct bufq *); 94void bufq_drain(struct bufq *); 95 96void bufq_wait(struct bufq *); 97void bufq_done(struct bufq *, struct buf *); 98void bufq_quiesce(void); 99void bufq_restart(void); 100 101/* fifo */ 102SIMPLEQ_HEAD(bufq_fifo_head, buf); 103struct bufq_fifo { 104 SIMPLEQ_ENTRY(buf) bqf_entries; 105}; 106 107/* nscan */ 108SIMPLEQ_HEAD(bufq_nscan_head, buf); 109struct bufq_nscan { 110 SIMPLEQ_ENTRY(buf) bqf_entries; 111}; 112 113/* bufq link in struct buf */ 114union bufq_data { 115 struct bufq_fifo bufq_data_fifo; 116 struct bufq_nscan bufq_data_nscan; 117}; 118 119/* The buffer header describes an I/O operation in the kernel. */ 120struct buf { 121 RBT_ENTRY(buf) b_rbbufs; /* vnode "hash" tree */ 122 LIST_ENTRY(buf) b_list; /* All allocated buffers. */ 123 LIST_ENTRY(buf) b_vnbufs; /* Buffer's associated vnode. */ 124 TAILQ_ENTRY(buf) b_freelist; /* Free list position if not active. */ 125 int cache; /* which cache are we in */ 126 struct proc *b_proc; /* Associated proc; NULL if kernel. */ 127 volatile long b_flags; /* B_* flags. */ 128 long b_bufsize; /* Allocated buffer size. */ 129 long b_bcount; /* Valid bytes in buffer. */ 130 size_t b_resid; /* Remaining I/O. */ 131 int b_error; /* Errno value. */ 132 dev_t b_dev; /* Device associated with buffer. */ 133 caddr_t b_data; /* associated data */ 134 void *b_saveaddr; /* Original b_data for physio. */ 135 136 TAILQ_ENTRY(buf) b_valist; /* LRU of va to reuse. */ 137 138 union bufq_data b_bufq; 139 struct bufq *b_bq; /* What bufq this buf is on */ 140 141 struct uvm_object *b_pobj; 142 struct uvm_object b_uobj; /* Object containing the pages */ 143 off_t b_poffs; /* Offset within object */ 144 145 daddr_t b_lblkno; /* Logical block number. */ 146 daddr_t b_blkno; /* Underlying physical block number. */ 147 /* Function to call upon completion. 148 * Will be called at splbio(). */ 149 void (*b_iodone)(struct buf *); 150 struct vnode *b_vp; /* Device vnode. */ 151 int b_dirtyoff; /* Offset in buffer of dirty region. */ 152 int b_dirtyend; /* Offset of end of dirty region. */ 153 int b_validoff; /* Offset in buffer of valid region. */ 154 int b_validend; /* Offset of end of valid region. */ 155}; 156 157TAILQ_HEAD(bufqueue, buf); 158 159struct bufcache { 160 int64_t hotbufpages; 161 int64_t warmbufpages; 162 int64_t cachepages; 163 struct bufqueue hotqueue; 164 struct bufqueue coldqueue; 165 struct bufqueue warmqueue; 166}; 167 168/* Device driver compatibility definitions. */ 169#define b_active b_bcount /* Driver queue head: drive active. */ 170 171/* 172 * These flags are kept in b_flags. 173 */ 174#define B_WRITE 0x00000000 /* Write buffer (pseudo flag). */ 175#define B_AGE 0x00000001 /* Move to age queue when I/O done. */ 176#define B_NEEDCOMMIT 0x00000002 /* Needs committing to stable storage */ 177#define B_ASYNC 0x00000004 /* Start I/O, do not wait. */ 178#define B_BAD 0x00000008 /* Bad block revectoring in progress. */ 179#define B_BUSY 0x00000010 /* I/O in progress. */ 180#define B_CACHE 0x00000020 /* Bread found us in the cache. */ 181#define B_CALL 0x00000040 /* Call b_iodone from biodone. */ 182#define B_DELWRI 0x00000080 /* Delay I/O until buffer reused. */ 183#define B_DONE 0x00000100 /* I/O completed. */ 184#define B_EINTR 0x00000200 /* I/O was interrupted */ 185#define B_ERROR 0x00000400 /* I/O error occurred. */ 186#define B_INVAL 0x00000800 /* Does not contain valid info. */ 187#define B_NOCACHE 0x00001000 /* Do not cache block after use. */ 188#define B_PHYS 0x00002000 /* I/O to user memory. */ 189#define B_RAW 0x00004000 /* Set by physio for raw transfers. */ 190#define B_READ 0x00008000 /* Read buffer. */ 191#define B_WANTED 0x00010000 /* Process wants this buffer. */ 192#define B_WRITEINPROG 0x00020000 /* Write in progress. */ 193#define B_XXX 0x00040000 /* Debugging flag. */ 194#define B_DEFERRED 0x00080000 /* Skipped over for cleaning */ 195#define B_SCANNED 0x00100000 /* Block already pushed during sync */ 196#define B_PDAEMON 0x00200000 /* I/O started by pagedaemon */ 197#define B_RELEASED 0x00400000 /* free this buffer after its kvm */ 198#define B_WARM 0x00800000 /* buffer is or has been on the warm queue */ 199#define B_COLD 0x01000000 /* buffer is on the cold queue */ 200#define B_BC 0x02000000 /* buffer is managed by the cache */ 201#define B_DMA 0x04000000 /* buffer is DMA reachable */ 202 203#define B_BITS "\20\001AGE\002NEEDCOMMIT\003ASYNC\004BAD\005BUSY" \ 204 "\006CACHE\007CALL\010DELWRI\011DONE\012EINTR\013ERROR" \ 205 "\014INVAL\015NOCACHE\016PHYS\017RAW\020READ" \ 206 "\021WANTED\022WRITEINPROG\023XXX(FORMAT)\024DEFERRED" \ 207 "\025SCANNED\026DAEMON\027RELEASED\030WARM\031COLD\032BC\033DMA" 208 209/* 210 * Zero out the buffer's data area. 211 */ 212#define clrbuf(bp) { \ 213 bzero((bp)->b_data, (bp)->b_bcount); \ 214 (bp)->b_resid = 0; \ 215} 216 217 218/* Flags to low-level allocation routines. */ 219#define B_CLRBUF 0x01 /* Request allocated buffer be cleared. */ 220#define B_SYNC 0x02 /* Do all allocations synchronously. */ 221 222struct cluster_info { 223 daddr_t ci_lastr; /* last read (read-ahead) */ 224 daddr_t ci_lastw; /* last write (write cluster) */ 225 daddr_t ci_cstart; /* start block of cluster */ 226 daddr_t ci_lasta; /* last allocation */ 227 int ci_clen; /* length of current cluster */ 228 int ci_ralen; /* Read-ahead length */ 229 daddr_t ci_maxra; /* last readahead block */ 230}; 231 232#ifdef _KERNEL 233__BEGIN_DECLS 234/* Kva slots (of size MAXPHYS) reserved for syncer and cleaner. */ 235#define RESERVE_SLOTS 4 236/* Buffer cache pages reserved for syncer and cleaner. */ 237#define RESERVE_PAGES (RESERVE_SLOTS * MAXPHYS / PAGE_SIZE) 238/* Minimum size of the buffer cache, in pages. */ 239#define BCACHE_MIN (RESERVE_PAGES * 2) 240#define UNCLEAN_PAGES (bcstats.numbufpages - bcstats.numcleanpages) 241 242extern struct proc *cleanerproc; 243extern long bufpages; /* Max number of pages for buffers' data */ 244extern struct pool bufpool; 245extern struct bufhead bufhead; 246 247void bawrite(struct buf *); 248void bdwrite(struct buf *); 249void biodone(struct buf *); 250int biowait(struct buf *); 251int bread(struct vnode *, daddr_t, int, struct buf **); 252int breadn(struct vnode *, daddr_t, int, daddr_t *, int *, int, 253 struct buf **); 254void brelse(struct buf *); 255#define bremfree bufcache_take 256void bufinit(void); 257void buf_dirty(struct buf *); 258void buf_undirty(struct buf *); 259void buf_adjcnt(struct buf *, long); 260int bwrite(struct buf *); 261struct buf *getblk(struct vnode *, daddr_t, int, int, uint64_t); 262struct buf *geteblk(size_t); 263struct buf *incore(struct vnode *, daddr_t); 264 265/* 266 * bufcache functions 267 */ 268void bufcache_take(struct buf *); 269void bufcache_release(struct buf *); 270 271int buf_flip_high(struct buf *); 272void buf_flip_dma(struct buf *); 273struct buf *bufcache_getcleanbuf(int, int); 274struct buf *bufcache_getdirtybuf(void); 275 276/* 277 * buf_kvm_init initializes the kvm handling for buffers. 278 * buf_acquire sets the B_BUSY flag and ensures that the buffer is 279 * mapped in the kvm. 280 * buf_release clears the B_BUSY flag and allows the buffer to become 281 * unmapped. 282 * buf_unmap is for internal use only. Unmaps the buffer from kvm. 283 */ 284void buf_mem_init(vsize_t); 285void buf_acquire(struct buf *); 286void buf_acquire_nomap(struct buf *); 287void buf_map(struct buf *); 288void buf_release(struct buf *); 289int buf_dealloc_mem(struct buf *); 290void buf_fix_mapping(struct buf *, vsize_t); 291void buf_alloc_pages(struct buf *, vsize_t); 292void buf_free_pages(struct buf *); 293 294void minphys(struct buf *bp); 295int physio(void (*strategy)(struct buf *), dev_t dev, int flags, 296 void (*minphys)(struct buf *), struct uio *uio); 297void brelvp(struct buf *); 298void reassignbuf(struct buf *); 299void bgetvp(struct vnode *, struct buf *); 300 301void buf_replacevnode(struct buf *, struct vnode *); 302void buf_daemon(void *); 303void buf_replacevnode(struct buf *, struct vnode *); 304int bread_cluster(struct vnode *, daddr_t, int, struct buf **); 305 306__END_DECLS 307#endif /* _KERNEL */ 308#endif /* !_SYS_BUF_H_ */ 309