VFS messaging/interfacing work stage 1/99. This stage replaces the old
[dragonfly.git] / sys / vfs / ufs / ffs_vnops.c
CommitLineData
984263bc
MD
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: src/sys/ufs/ffs/ffs_vnops.c,v 1.64 2000/01/10 12:04:25 phk Exp $
2d3e977e 35 * $DragonFly: src/sys/vfs/ufs/ffs_vnops.c,v 1.9 2004/08/13 17:51:13 dillon Exp $
984263bc
MD
36 */
37
38#include <sys/param.h>
39#include <sys/systm.h>
40#include <sys/resourcevar.h>
41#include <sys/signalvar.h>
42#include <sys/kernel.h>
43#include <sys/stat.h>
44#include <sys/buf.h>
45#include <sys/proc.h>
46#include <sys/mount.h>
47#include <sys/vnode.h>
48#include <sys/conf.h>
49
50#include <machine/limits.h>
51
52#include <vm/vm.h>
53#include <vm/vm_page.h>
54#include <vm/vm_object.h>
55#include <vm/vm_extern.h>
56
3020e3be
MD
57#include <sys/buf2.h>
58
1f2de5d4
MD
59#include "quota.h"
60#include "inode.h"
61#include "ufsmount.h"
62#include "ufs_extern.h"
984263bc 63
1f2de5d4
MD
64#include "fs.h"
65#include "ffs_extern.h"
984263bc 66
a6ee311a
RG
67static int ffs_fsync (struct vop_fsync_args *);
68static int ffs_getpages (struct vop_getpages_args *);
69static int ffs_putpages (struct vop_putpages_args *);
70static int ffs_read (struct vop_read_args *);
71static int ffs_write (struct vop_write_args *);
984263bc
MD
72
73/* Global vfs data structures for ufs. */
2d3e977e 74struct vop_ops *ffs_vnode_vops;
984263bc 75static struct vnodeopv_entry_desc ffs_vnodeop_entries[] = {
2d3e977e
MD
76 { &vop_default_desc, (void *) ufs_vnoperate },
77 { &vop_fsync_desc, (void *) ffs_fsync },
78 { &vop_getpages_desc, (void *) ffs_getpages },
79 { &vop_putpages_desc, (void *) ffs_putpages },
80 { &vop_read_desc, (void *) ffs_read },
81 { &vop_balloc_desc, (void *) ffs_balloc },
82 { &vop_reallocblks_desc, (void *) ffs_reallocblks },
83 { &vop_write_desc, (void *) ffs_write },
984263bc
MD
84 { NULL, NULL }
85};
86static struct vnodeopv_desc ffs_vnodeop_opv_desc =
2d3e977e 87 { &ffs_vnode_vops, ffs_vnodeop_entries };
984263bc 88
2d3e977e 89struct vop_ops *ffs_spec_vops;
984263bc 90static struct vnodeopv_entry_desc ffs_specop_entries[] = {
2d3e977e
MD
91 { &vop_default_desc, (void *) ufs_vnoperatespec },
92 { &vop_fsync_desc, (void *) ffs_fsync },
984263bc
MD
93 { NULL, NULL }
94};
95static struct vnodeopv_desc ffs_specop_opv_desc =
2d3e977e 96 { &ffs_spec_vops, ffs_specop_entries };
984263bc 97
2d3e977e 98struct vop_ops *ffs_fifo_vops;
984263bc 99static struct vnodeopv_entry_desc ffs_fifoop_entries[] = {
2d3e977e
MD
100 { &vop_default_desc, (void *) ufs_vnoperatefifo },
101 { &vop_fsync_desc, (void *) ffs_fsync },
984263bc
MD
102 { NULL, NULL }
103};
104static struct vnodeopv_desc ffs_fifoop_opv_desc =
2d3e977e 105 { &ffs_fifo_vops, ffs_fifoop_entries };
984263bc
MD
106
107VNODEOP_SET(ffs_vnodeop_opv_desc);
108VNODEOP_SET(ffs_specop_opv_desc);
109VNODEOP_SET(ffs_fifoop_opv_desc);
110
1f2de5d4 111#include "ufs_readwrite.c"
984263bc
MD
112
113/*
114 * Synch an open file.
0973c589
CP
115 *
116 * ffs_fsync(struct vnode *a_vp, struct ucred *a_cred, int a_waitfor,
117 * struct proc *a_p)
984263bc
MD
118 */
119/* ARGSUSED */
120static int
0973c589 121ffs_fsync(struct vop_fsync_args *ap)
984263bc
MD
122{
123 struct vnode *vp = ap->a_vp;
124 struct buf *bp;
125 struct buf *nbp;
126 int s, error, wait, passes, skipmeta;
127 daddr_t lbn;
128
129 wait = (ap->a_waitfor == MNT_WAIT);
130 if (vn_isdisk(vp, NULL)) {
131 lbn = INT_MAX;
e4c9c0c8
MD
132 if (vp->v_rdev && vp->v_rdev->si_mountpoint != NULL &&
133 (vp->v_rdev->si_mountpoint->mnt_flag & MNT_SOFTDEP))
984263bc
MD
134 softdep_fsync_mountdev(vp);
135 } else {
136 struct inode *ip;
137 ip = VTOI(vp);
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();
149loop:
150 for (bp = TAILQ_FIRST(&vp->v_dirtyblkhd); bp;
151 bp = TAILQ_NEXT(bp, b_vnbufs))
152 bp->b_flags &= ~B_SCANNED;
153 for (bp = TAILQ_FIRST(&vp->v_dirtyblkhd); bp; bp = nbp) {
154 nbp = TAILQ_NEXT(bp, b_vnbufs);
155 /*
156 * Reasons to skip this buffer: it has already been considered
157 * on this pass, this pass is the first time through on a
158 * synchronous flush request and the buffer being considered
159 * is metadata, the buffer has dependencies that will cause
160 * it to be redirtied and it has not already been deferred,
161 * or it is already being written.
162 */
163 if ((bp->b_flags & B_SCANNED) != 0)
164 continue;
165 bp->b_flags |= B_SCANNED;
166 if ((skipmeta == 1 && bp->b_lblkno < 0))
167 continue;
168 if (!wait && LIST_FIRST(&bp->b_dep) != NULL &&
169 (bp->b_flags & B_DEFERRED) == 0 &&
170 bioops.io_countdeps && (*bioops.io_countdeps)(bp, 0)) {
171 bp->b_flags |= B_DEFERRED;
172 continue;
173 }
174 if (BUF_LOCK(bp, LK_EXCLUSIVE | LK_NOWAIT))
175 continue;
176 if ((bp->b_flags & B_DELWRI) == 0)
177 panic("ffs_fsync: not dirty");
178 if (vp != bp->b_vp)
179 panic("ffs_fsync: vp != vp->b_vp");
180 /*
181 * If this is a synchronous flush request, or it is not a
182 * file or device, start the write on this buffer immediatly.
183 */
184 if (wait || (vp->v_type != VREG && vp->v_type != VBLK)) {
185
186 /*
187 * On our final pass through, do all I/O synchronously
188 * so that we can find out if our flush is failing
189 * because of write errors.
190 */
191 if (passes > 0 || !wait) {
192 if ((bp->b_flags & B_CLUSTEROK) && !wait) {
193 BUF_UNLOCK(bp);
194 (void) vfs_bio_awrite(bp);
195 } else {
196 bremfree(bp);
197 splx(s);
198 (void) bawrite(bp);
199 s = splbio();
200 }
201 } else {
202 bremfree(bp);
203 splx(s);
204 if ((error = bwrite(bp)) != 0)
205 return (error);
206 s = splbio();
207 }
208 } else if ((vp->v_type == VREG) && (bp->b_lblkno >= lbn)) {
209 /*
210 * If the buffer is for data that has been truncated
211 * off the file, then throw it away.
212 */
213 bremfree(bp);
214 bp->b_flags |= B_INVAL | B_NOCACHE;
215 splx(s);
216 brelse(bp);
217 s = splbio();
218 } else {
219 BUF_UNLOCK(bp);
220 vfs_bio_awrite(bp);
221 }
222 /*
223 * Since we may have slept during the I/O, we need
224 * to start from a known point.
225 */
226 nbp = TAILQ_FIRST(&vp->v_dirtyblkhd);
227 }
228 /*
229 * If we were asked to do this synchronously, then go back for
230 * another pass, this time doing the metadata.
231 */
232 if (skipmeta) {
233 skipmeta = 0;
234 goto loop;
235 }
236
237 if (wait) {
238 while (vp->v_numoutput) {
239 vp->v_flag |= VBWAIT;
377d4740 240 (void) tsleep((caddr_t)&vp->v_numoutput, 0, "ffsfsn", 0);
984263bc
MD
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}