sys/kern/vfs_hash.c

   1 /*-
   2  * SPDX-License-Identifier: BSD-2-Clause
   3  *
   4  * Copyright (c) 2005 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 #include <sys/param.h>
  31 #include <sys/systm.h>
  32 #include <sys/kernel.h>
  33 #include <sys/malloc.h>
  34 #include <sys/mount.h>
  35 #include <sys/rwlock.h>
  36 #include <sys/vnode.h>
  37
  38 static MALLOC_DEFINE(M_VFS_HASH, "vfs_hash", "VFS hash table");
  39
  40 static LIST_HEAD(vfs_hash_head, vnode)  *vfs_hash_tbl;
  41 static LIST_HEAD(,vnode)                vfs_hash_side;
  42 static u_long                           vfs_hash_mask;
  43 static struct rwlock __exclusive_cache_line vfs_hash_lock;
  44
  45 static void
  46 vfs_hashinit(void *dummy __unused)
  47 {
  48
  49         vfs_hash_tbl = hashinit(desiredvnodes, M_VFS_HASH, &vfs_hash_mask);
  50         rw_init(&vfs_hash_lock, "vfs hash");
  51         LIST_INIT(&vfs_hash_side);
  52 }
  53
  54 /* Must be SI_ORDER_SECOND so desiredvnodes is available */
  55 SYSINIT(vfs_hash, SI_SUB_VFS, SI_ORDER_SECOND, vfs_hashinit, NULL);
  56
  57 u_int
  58 vfs_hash_index(struct vnode *vp)
  59 {
  60
  61         return (vp->v_hash + vp->v_mount->mnt_hashseed);
  62 }
  63
  64 static struct vfs_hash_head *
  65 vfs_hash_bucket(const struct mount *mp, u_int hash)
  66 {
  67
  68         return (&vfs_hash_tbl[(hash + mp->mnt_hashseed) & vfs_hash_mask]);
  69 }
  70
  71 int
  72 vfs_hash_get(const struct mount *mp, u_int hash, int flags, struct thread *td,
  73     struct vnode **vpp, vfs_hash_cmp_t *fn, void *arg)
  74 {
  75         struct vnode *vp;
  76         enum vgetstate vs;
  77         int error;
  78
  79         while (1) {
  80                 rw_rlock(&vfs_hash_lock);
  81                 LIST_FOREACH(vp, vfs_hash_bucket(mp, hash), v_hashlist) {
  82                         if (vp->v_hash != hash)
  83                                 continue;
  84                         if (vp->v_mount != mp)
  85                                 continue;
  86                         if (fn != NULL && fn(vp, arg))
  87                                 continue;
  88                         vs = vget_prep(vp);
  89                         rw_runlock(&vfs_hash_lock);
  90                         error = vget_finish(vp, flags, vs);
  91                         if (error == ENOENT && (flags & LK_NOWAIT) == 0)
  92                                 break;
  93                         if (error != 0)
  94                                 return (error);
  95                         if (vp->v_hash != hash ||
  96                             (fn != NULL && fn(vp, arg))) {
  97                                 vput(vp);
  98                                 /* Restart the bucket walk. */
  99                                 break;
 100                         }
 101                         *vpp = vp;
 102                         return (0);
 103                 }
 104                 if (vp == NULL) {
 105                         rw_runlock(&vfs_hash_lock);
 106                         *vpp = NULL;
 107                         return (0);
 108                 }
 109         }
 110 }
 111
 112 void
 113 vfs_hash_ref(const struct mount *mp, u_int hash, struct thread *td,
 114     struct vnode **vpp, vfs_hash_cmp_t *fn, void *arg)
 115 {
 116         struct vnode *vp;
 117
 118         while (1) {
 119                 rw_rlock(&vfs_hash_lock);
 120                 LIST_FOREACH(vp, vfs_hash_bucket(mp, hash), v_hashlist) {
 121                         if (vp->v_hash != hash)
 122                                 continue;
 123                         if (vp->v_mount != mp)
 124                                 continue;
 125                         if (fn != NULL && fn(vp, arg))
 126                                 continue;
 127                         vhold(vp);
 128                         rw_runlock(&vfs_hash_lock);
 129                         vref(vp);
 130                         vdrop(vp);
 131                         *vpp = vp;
 132                         return;
 133                 }
 134                 if (vp == NULL) {
 135                         rw_runlock(&vfs_hash_lock);
 136                         *vpp = NULL;
 137                         return;
 138                 }
 139         }
 140 }
 141
 142 void
 143 vfs_hash_remove(struct vnode *vp)
 144 {
 145
 146         rw_wlock(&vfs_hash_lock);
 147         LIST_REMOVE(vp, v_hashlist);
 148         rw_wunlock(&vfs_hash_lock);
 149 }
 150
 151 int
 152 vfs_hash_insert(struct vnode *vp, u_int hash, int flags, struct thread *td,
 153     struct vnode **vpp, vfs_hash_cmp_t *fn, void *arg)
 154 {
 155         struct vnode *vp2;
 156         enum vgetstate vs;
 157         int error;
 158
 159         *vpp = NULL;
 160         while (1) {
 161                 rw_wlock(&vfs_hash_lock);
 162                 LIST_FOREACH(vp2,
 163                     vfs_hash_bucket(vp->v_mount, hash), v_hashlist) {
 164                         if (vp2->v_hash != hash)
 165                                 continue;
 166                         if (vp2->v_mount != vp->v_mount)
 167                                 continue;
 168                         if (fn != NULL && fn(vp2, arg))
 169                                 continue;
 170                         vs = vget_prep(vp2);
 171                         rw_wunlock(&vfs_hash_lock);
 172                         error = vget_finish(vp2, flags, vs);
 173                         if (error == ENOENT && (flags & LK_NOWAIT) == 0)
 174                                 break;
 175                         rw_wlock(&vfs_hash_lock);
 176                         LIST_INSERT_HEAD(&vfs_hash_side, vp, v_hashlist);
 177                         rw_wunlock(&vfs_hash_lock);
 178                         vgone(vp);
 179                         vput(vp);
 180                         if (!error)
 181                                 *vpp = vp2;
 182                         return (error);
 183                 }
 184                 if (vp2 == NULL)
 185                         break;
 186         }
 187         vp->v_hash = hash;
 188         LIST_INSERT_HEAD(vfs_hash_bucket(vp->v_mount, hash), vp, v_hashlist);
 189         rw_wunlock(&vfs_hash_lock);
 190         return (0);
 191 }
 192
 193 void
 194 vfs_hash_rehash(struct vnode *vp, u_int hash)
 195 {
 196         ASSERT_VOP_ELOCKED(vp, "rehash requires excl lock");
 197
 198         rw_wlock(&vfs_hash_lock);
 199         LIST_REMOVE(vp, v_hashlist);
 200         LIST_INSERT_HEAD(vfs_hash_bucket(vp->v_mount, hash), vp, v_hashlist);
 201         vp->v_hash = hash;
 202         rw_wunlock(&vfs_hash_lock);
 203 }
 204
 205 void
 206 vfs_hash_changesize(u_long newmaxvnodes)
 207 {
 208         struct vfs_hash_head *vfs_hash_newtbl, *vfs_hash_oldtbl;
 209         u_long vfs_hash_newmask, vfs_hash_oldmask;
 210         struct vnode *vp;
 211         int i;
 212
 213         vfs_hash_newtbl = hashinit(newmaxvnodes, M_VFS_HASH,
 214                 &vfs_hash_newmask);
 215         /* If same hash table size, nothing to do */
 216         if (vfs_hash_mask == vfs_hash_newmask) {
 217                 free(vfs_hash_newtbl, M_VFS_HASH);
 218                 return;
 219         }
 220         /*
 221          * Move everything from the old hash table to the new table.
 222          * None of the vnodes in the table can be recycled because to
 223          * do so, they have to be removed from the hash table.
 224          */
 225         rw_wlock(&vfs_hash_lock);
 226         vfs_hash_oldtbl = vfs_hash_tbl;
 227         vfs_hash_oldmask = vfs_hash_mask;
 228         vfs_hash_tbl = vfs_hash_newtbl;
 229         vfs_hash_mask = vfs_hash_newmask;
 230         for (i = 0; i <= vfs_hash_oldmask; i++) {
 231                 while ((vp = LIST_FIRST(&vfs_hash_oldtbl[i])) != NULL) {
 232                         LIST_REMOVE(vp, v_hashlist);
 233                         LIST_INSERT_HEAD(
 234                             vfs_hash_bucket(vp->v_mount, vp->v_hash),
 235                             vp, v_hashlist);
 236                 }
 237         }
 238         rw_wunlock(&vfs_hash_lock);
 239         free(vfs_hash_oldtbl, M_VFS_HASH);
 240 }