1 /* 2 * Copyright (c) 1982, 1986, 1989, 1993 3 * The Regents of the University of California. All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 3. All advertising materials mentioning features or use of this software 14 * must display the following acknowledgement: 15 * This product includes software developed by the University of 16 * California, Berkeley and its contributors. 17 * 4. Neither the name of the University nor the names of its contributors 18 * may be used to endorse or promote products derived from this software 19 * without specific prior written permission. 20 * 21 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 24 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 31 * SUCH DAMAGE. 32 * 33 * @(#)ffs_vnops.c 8.15 (Berkeley) 5/14/95 34 * $FreeBSD$ 35 */ 36 37 #include "opt_ufs.h" 38 39 #include <sys/param.h> 40 #include <sys/systm.h> 41 #include <sys/resourcevar.h> 42 #include <sys/signalvar.h> 43 #include <sys/kernel.h> 44 #include <sys/stat.h> 45 #include <sys/bio.h> 46 #include <sys/buf.h> 47 #include <sys/proc.h> 48 #include <sys/mount.h> 49 #include <sys/vnode.h> 50 #include <sys/conf.h> 51 52 #include <machine/limits.h> 53 54 #include <vm/vm.h> 55 #include <vm/vm_page.h> 56 #include <vm/vm_object.h> 57 #include <vm/vm_extern.h> 58 59 #include <ufs/ufs/extattr.h> 60 #include <ufs/ufs/quota.h> 61 #include <ufs/ufs/inode.h> 62 #include <ufs/ufs/ufsmount.h> 63 #include <ufs/ufs/ufs_extern.h> 64 65 #include <ufs/ffs/fs.h> 66 #include <ufs/ffs/ffs_extern.h> 67 68 int ffs_fsync(struct vop_fsync_args *); 69 static int ffs_getpages(struct vop_getpages_args *); 70 static int ffs_read(struct vop_read_args *); 71 static int ffs_write(struct vop_write_args *); 72 73 /* Global vfs data structures for ufs. */ 74 vop_t **ffs_vnodeop_p; 75 static struct vnodeopv_entry_desc ffs_vnodeop_entries[] = { 76 { &vop_default_desc, (vop_t *) ufs_vnoperate }, 77 { &vop_fsync_desc, (vop_t *) ffs_fsync }, 78 { &vop_getpages_desc, (vop_t *) ffs_getpages }, 79 { &vop_read_desc, (vop_t *) ffs_read }, 80 { &vop_reallocblks_desc, (vop_t *) ffs_reallocblks }, 81 { &vop_write_desc, (vop_t *) ffs_write }, 82 { NULL, NULL } 83 }; 84 static struct vnodeopv_desc ffs_vnodeop_opv_desc = 85 { &ffs_vnodeop_p, ffs_vnodeop_entries }; 86 87 vop_t **ffs_specop_p; 88 static struct vnodeopv_entry_desc ffs_specop_entries[] = { 89 { &vop_default_desc, (vop_t *) ufs_vnoperatespec }, 90 { &vop_fsync_desc, (vop_t *) ffs_fsync }, 91 { NULL, NULL } 92 }; 93 static struct vnodeopv_desc ffs_specop_opv_desc = 94 { &ffs_specop_p, ffs_specop_entries }; 95 96 vop_t **ffs_fifoop_p; 97 static struct vnodeopv_entry_desc ffs_fifoop_entries[] = { 98 { &vop_default_desc, (vop_t *) ufs_vnoperatefifo }, 99 { &vop_fsync_desc, (vop_t *) ffs_fsync }, 100 { NULL, NULL } 101 }; 102 static struct vnodeopv_desc ffs_fifoop_opv_desc = 103 { &ffs_fifoop_p, ffs_fifoop_entries }; 104 105 VNODEOP_SET(ffs_vnodeop_opv_desc); 106 VNODEOP_SET(ffs_specop_opv_desc); 107 VNODEOP_SET(ffs_fifoop_opv_desc); 108 109 #include <ufs/ufs/ufs_readwrite.c> 110 111 /* 112 * Synch an open file. 113 */ 114 /* ARGSUSED */ 115 int 116 ffs_fsync(ap) 117 struct vop_fsync_args /* { 118 struct vnode *a_vp; 119 struct ucred *a_cred; 120 int a_waitfor; 121 struct thread *a_td; 122 } */ *ap; 123 { 124 struct vnode *vp = ap->a_vp; 125 struct inode *ip = VTOI(vp); 126 struct buf *bp; 127 struct buf *nbp; 128 int s, error, wait, passes, skipmeta; 129 ufs_lbn_t lbn; 130 131 wait = (ap->a_waitfor == MNT_WAIT); 132 if (vn_isdisk(vp, NULL)) { 133 lbn = INT_MAX; 134 if (vp->v_rdev->si_mountpoint != NULL && 135 (vp->v_rdev->si_mountpoint->mnt_flag & MNT_SOFTDEP)) 136 softdep_fsync_mountdev(vp); 137 } else { 138 lbn = lblkno(ip->i_fs, (ip->i_size + ip->i_fs->fs_bsize - 1)); 139 } 140 141 /* 142 * Flush all dirty buffers associated with a vnode. 143 */ 144 passes = NIADDR + 1; 145 skipmeta = 0; 146 if (wait) 147 skipmeta = 1; 148 s = splbio(); 149 loop: 150 TAILQ_FOREACH(bp, &vp->v_dirtyblkhd, b_vnbufs) 151 bp->b_flags &= ~B_SCANNED; 152 for (bp = TAILQ_FIRST(&vp->v_dirtyblkhd); bp; bp = nbp) { 153 nbp = TAILQ_NEXT(bp, b_vnbufs); 154 /* 155 * Reasons to skip this buffer: it has already been considered 156 * on this pass, this pass is the first time through on a 157 * synchronous flush request and the buffer being considered 158 * is metadata, the buffer has dependencies that will cause 159 * it to be redirtied and it has not already been deferred, 160 * or it is already being written. 161 */ 162 if ((bp->b_flags & B_SCANNED) != 0) 163 continue; 164 bp->b_flags |= B_SCANNED; 165 if ((skipmeta == 1 && bp->b_lblkno < 0)) 166 continue; 167 if (!wait && LIST_FIRST(&bp->b_dep) != NULL && 168 (bp->b_flags & B_DEFERRED) == 0 && 169 buf_countdeps(bp, 0)) { 170 bp->b_flags |= B_DEFERRED; 171 continue; 172 } 173 if (BUF_LOCK(bp, LK_EXCLUSIVE | LK_NOWAIT)) 174 continue; 175 if ((bp->b_flags & B_DELWRI) == 0) 176 panic("ffs_fsync: not dirty"); 177 if (vp != bp->b_vp) 178 panic("ffs_fsync: vp != vp->b_vp"); 179 /* 180 * If this is a synchronous flush request, or it is not a 181 * file or device, start the write on this buffer immediatly. 182 */ 183 if (wait || (vp->v_type != VREG && vp->v_type != VBLK)) { 184 185 /* 186 * On our final pass through, do all I/O synchronously 187 * so that we can find out if our flush is failing 188 * because of write errors. 189 */ 190 if (passes > 0 || !wait) { 191 if ((bp->b_flags & B_CLUSTEROK) && !wait) { 192 BUF_UNLOCK(bp); 193 (void) vfs_bio_awrite(bp); 194 } else { 195 bremfree(bp); 196 splx(s); 197 (void) bawrite(bp); 198 s = splbio(); 199 } 200 } else { 201 bremfree(bp); 202 splx(s); 203 if ((error = bwrite(bp)) != 0) 204 return (error); 205 s = splbio(); 206 } 207 } else if ((vp->v_type == VREG) && (bp->b_lblkno >= lbn)) { 208 /* 209 * If the buffer is for data that has been truncated 210 * off the file, then throw it away. 211 */ 212 bremfree(bp); 213 bp->b_flags |= B_INVAL | B_NOCACHE; 214 splx(s); 215 brelse(bp); 216 s = splbio(); 217 } else { 218 BUF_UNLOCK(bp); 219 vfs_bio_awrite(bp); 220 } 221 /* 222 * Since we may have slept during the I/O, we need 223 * to start from a known point. 224 */ 225 nbp = TAILQ_FIRST(&vp->v_dirtyblkhd); 226 } 227 /* 228 * If we were asked to do this synchronously, then go back for 229 * another pass, this time doing the metadata. 230 */ 231 if (skipmeta) { 232 skipmeta = 0; 233 goto loop; 234 } 235 236 if (wait) { 237 while (vp->v_numoutput) { 238 vp->v_flag |= VBWAIT; 239 (void) tsleep((caddr_t)&vp->v_numoutput, 240 PRIBIO + 4, "ffsfsn", 0); 241 } 242 243 /* 244 * Ensure that any filesystem metatdata associated 245 * with the vnode has been written. 246 */ 247 splx(s); 248 if ((error = softdep_sync_metadata(ap)) != 0) 249 return (error); 250 s = splbio(); 251 252 if (!TAILQ_EMPTY(&vp->v_dirtyblkhd)) { 253 /* 254 * Block devices associated with filesystems may 255 * have new I/O requests posted for them even if 256 * the vnode is locked, so no amount of trying will 257 * get them clean. Thus we give block devices a 258 * good effort, then just give up. For all other file 259 * types, go around and try again until it is clean. 260 */ 261 if (passes > 0) { 262 passes -= 1; 263 goto loop; 264 } 265 #ifdef DIAGNOSTIC 266 if (!vn_isdisk(vp, NULL)) 267 vprint("ffs_fsync: dirty", vp); 268 #endif 269 } 270 } 271 splx(s); 272 return (UFS_UPDATE(vp, wait)); 273 } 274