2 * Copyright (c) 1989, 1993
3 * The Regents of the University of California. All rights reserved.
5 * This code is derived from software contributed
6 * to Berkeley by John Heidemann of the UCLA Ficus project.
8 * Source: * @(#)i405_init.c 2.10 92/04/27 UCLA Ficus project
10 * Redistribution and use in source and binary forms, with or without
11 * modification, are permitted provided that the following conditions
13 * 1. Redistributions of source code must retain the above copyright
14 * notice, this list of conditions and the following disclaimer.
15 * 2. Redistributions in binary form must reproduce the above copyright
16 * notice, this list of conditions and the following disclaimer in the
17 * documentation and/or other materials provided with the distribution.
18 * 4. Neither the name of the University nor the names of its contributors
19 * may be used to endorse or promote products derived from this software
20 * without specific prior written permission.
22 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
23 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
24 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
25 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
26 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
27 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
28 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
29 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
30 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
31 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
35 #include <sys/cdefs.h>
36 __FBSDID("$FreeBSD$");
38 #include <sys/param.h>
39 #include <sys/systm.h>
43 #include <sys/event.h>
44 #include <sys/kernel.h>
45 #include <sys/limits.h>
47 #include <sys/malloc.h>
48 #include <sys/mount.h>
49 #include <sys/mutex.h>
50 #include <sys/unistd.h>
51 #include <sys/vnode.h>
55 #include <vm/vm_object.h>
56 #include <vm/vm_extern.h>
58 #include <vm/vm_map.h>
59 #include <vm/vm_page.h>
60 #include <vm/vm_pager.h>
61 #include <vm/vnode_pager.h>
63 static int vop_nolookup(struct vop_lookup_args *);
64 static int vop_nostrategy(struct vop_strategy_args *);
67 * This vnode table stores what we want to do if the filesystem doesn't
68 * implement a particular VOP.
70 * If there is no specific entry here, we will return EOPNOTSUPP.
74 struct vop_vector default_vnodeops = {
76 .vop_bypass = VOP_EOPNOTSUPP,
78 .vop_advlock = VOP_EINVAL,
79 .vop_bmap = vop_stdbmap,
80 .vop_close = VOP_NULL,
81 .vop_fsync = VOP_NULL,
82 .vop_getpages = vop_stdgetpages,
83 .vop_getwritemount = vop_stdgetwritemount,
84 .vop_inactive = VOP_NULL,
85 .vop_ioctl = VOP_ENOTTY,
86 .vop_kqfilter = vop_stdkqfilter,
87 .vop_islocked = vop_stdislocked,
88 .vop_lease = VOP_NULL,
89 .vop_lock = vop_stdlock,
90 .vop_lookup = vop_nolookup,
92 .vop_pathconf = VOP_EINVAL,
93 .vop_poll = vop_nopoll,
94 .vop_putpages = vop_stdputpages,
95 .vop_readlink = VOP_EINVAL,
96 .vop_revoke = VOP_PANIC,
97 .vop_strategy = vop_nostrategy,
98 .vop_unlock = vop_stdunlock,
102 * Series of placeholder functions for various error returns for
107 vop_eopnotsupp(struct vop_generic_args *ap)
110 printf("vop_notsupp[%s]\n", ap->a_desc->vdesc_name);
117 vop_ebadf(struct vop_generic_args *ap)
124 vop_enotty(struct vop_generic_args *ap)
131 vop_einval(struct vop_generic_args *ap)
138 vop_null(struct vop_generic_args *ap)
145 * Helper function to panic on some bad VOPs in some filesystems.
148 vop_panic(struct vop_generic_args *ap)
151 panic("filesystem goof: vop_panic[%s]", ap->a_desc->vdesc_name);
155 * vop_std<something> and vop_no<something> are default functions for use by
156 * filesystems that need the "default reasonable" implementation for a
157 * particular operation.
159 * The documentation for the operations they implement exists (if it exists)
160 * in the VOP_<SOMETHING>(9) manpage (all uppercase).
164 * Default vop for filesystems that do not support name lookup
168 struct vop_lookup_args /* {
170 struct vnode **a_vpp;
171 struct componentname *a_cnp;
182 * Strategy routine for VFS devices that have none.
184 * BIO_ERROR and B_INVAL must be cleared prior to calling any strategy
185 * routine. Typically this is done for a BIO_READ strategy call.
186 * Typically B_INVAL is assumed to already be clear prior to a write
187 * and should not be cleared manually unless you just made the buffer
188 * invalid. BIO_ERROR should be cleared either way.
192 vop_nostrategy (struct vop_strategy_args *ap)
194 printf("No strategy for buffer at %p\n", ap->a_bp);
195 vprint("vnode", ap->a_vp);
196 ap->a_bp->b_ioflags |= BIO_ERROR;
197 ap->a_bp->b_error = EOPNOTSUPP;
205 * Standard implementation of POSIX pathconf, to get information about limits
207 * Override per filesystem for the case where the filesystem has smaller
212 struct vop_pathconf_args /* {
219 switch (ap->a_name) {
221 *ap->a_retval = NAME_MAX;
224 *ap->a_retval = PATH_MAX;
227 *ap->a_retval = LINK_MAX;
230 *ap->a_retval = MAX_CANON;
233 *ap->a_retval = MAX_INPUT;
236 *ap->a_retval = PIPE_BUF;
238 case _PC_CHOWN_RESTRICTED:
242 *ap->a_retval = _POSIX_VDISABLE;
251 * Standard lock, unlock and islocked functions.
255 struct vop_lock_args /* {
261 struct vnode *vp = ap->a_vp;
263 return (lockmgr(vp->v_vnlock, ap->a_flags, VI_MTX(vp), ap->a_td));
269 struct vop_unlock_args /* {
275 struct vnode *vp = ap->a_vp;
277 return (lockmgr(vp->v_vnlock, ap->a_flags | LK_RELEASE, VI_MTX(vp),
284 struct vop_islocked_args /* {
290 return (lockstatus(ap->a_vp->v_vnlock, ap->a_td));
294 * Return true for select/poll.
298 struct vop_poll_args /* {
301 struct ucred *a_cred;
306 * Return true for read/write. If the user asked for something
307 * special, return POLLNVAL, so that clients have a way of
308 * determining reliably whether or not the extended
309 * functionality is present without hard-coding knowledge
310 * of specific filesystem implementations.
311 * Stay in sync with kern_conf.c::no_poll().
313 if (ap->a_events & ~POLLSTANDARD)
316 return (ap->a_events & (POLLIN | POLLOUT | POLLRDNORM | POLLWRNORM));
320 * Implement poll for local filesystems that support it.
324 struct vop_poll_args /* {
327 struct ucred *a_cred;
331 if (ap->a_events & ~POLLSTANDARD)
332 return (vn_pollrecord(ap->a_vp, ap->a_td, ap->a_events));
333 return (ap->a_events & (POLLIN | POLLOUT | POLLRDNORM | POLLWRNORM));
337 * Return our mount point, as we will take charge of the writes.
340 vop_stdgetwritemount(ap)
341 struct vop_getwritemount_args /* {
343 struct mount **a_mpp;
347 *(ap->a_mpp) = ap->a_vp->v_mount;
351 /* XXX Needs good comment and VOP_BMAP(9) manpage */
354 struct vop_bmap_args /* {
357 struct bufobj **a_bop;
364 if (ap->a_bop != NULL)
365 *ap->a_bop = &ap->a_vp->v_bufobj;
366 if (ap->a_bnp != NULL)
367 *ap->a_bnp = ap->a_bn * btodb(ap->a_vp->v_mount->mnt_stat.f_iosize);
368 if (ap->a_runp != NULL)
370 if (ap->a_runb != NULL)
377 struct vop_fsync_args /* {
379 struct ucred *a_cred;
384 struct vnode *vp = ap->a_vp;
389 int maxretry = 1000; /* large, arbitrarily chosen */
394 * MARK/SCAN initialization to avoid infinite loops.
396 TAILQ_FOREACH(bp, &vp->v_bufobj.bo_dirty.bv_hd, b_bobufs) {
397 bp->b_vflags &= ~BV_SCANNED;
402 * Flush all dirty buffers associated with a vnode.
405 TAILQ_FOREACH_SAFE(bp, &vp->v_bufobj.bo_dirty.bv_hd, b_bobufs, nbp) {
406 if ((bp->b_vflags & BV_SCANNED) != 0)
408 bp->b_vflags |= BV_SCANNED;
409 if (BUF_LOCK(bp, LK_EXCLUSIVE | LK_NOWAIT, NULL))
412 KASSERT(bp->b_bufobj == &vp->v_bufobj,
413 ("bp %p wrong b_bufobj %p should be %p",
414 bp, bp->b_bufobj, &vp->v_bufobj));
415 if ((bp->b_flags & B_DELWRI) == 0)
416 panic("fsync: not dirty");
417 if ((vp->v_object != NULL) && (bp->b_flags & B_CLUSTEROK)) {
428 * If synchronous the caller expects us to completely resolve all
429 * dirty buffers in the system. Wait for in-progress I/O to
430 * complete (which could include background bitmap writes), then
431 * retry if dirty blocks still exist.
433 if (ap->a_waitfor == MNT_WAIT) {
435 bufobj_wwait(bo, 0, 0);
436 if (bo->bo_dirty.bv_cnt > 0) {
438 * If we are unable to write any of these buffers
439 * then we fail now rather than trying endlessly
442 TAILQ_FOREACH(bp, &bo->bo_dirty.bv_hd, b_bobufs)
443 if ((error = bp->b_error) == 0)
445 if (error == 0 && --maxretry >= 0)
452 vprint("fsync: giving up on dirty", vp);
457 /* XXX Needs good comment and more info in the manpage (VOP_GETPAGES(9)). */
460 struct vop_getpages_args /* {
465 vm_ooffset_t a_offset;
469 return vnode_pager_generic_getpages(ap->a_vp, ap->a_m,
470 ap->a_count, ap->a_reqpage);
474 vop_stdkqfilter(struct vop_kqfilter_args *ap)
476 return vfs_kqfilter(ap);
479 /* XXX Needs good comment and more info in the manpage (VOP_PUTPAGES(9)). */
482 struct vop_putpages_args /* {
488 vm_ooffset_t a_offset;
492 return vnode_pager_generic_putpages(ap->a_vp, ap->a_m, ap->a_count,
493 ap->a_sync, ap->a_rtvals);
498 * used to fill the vfs function table to get reasonable default return values.
501 vfs_stdroot (mp, flags, vpp, td)
512 vfs_stdstatfs (mp, sbp, td)
522 vfs_stdvptofh (vp, fhp)
531 vfs_stdquotactl (mp, cmds, uid, arg, td)
543 vfs_stdsync(mp, waitfor, td)
548 struct vnode *vp, *nvp;
549 int error, lockreq, allerror = 0;
551 lockreq = LK_EXCLUSIVE | LK_INTERLOCK;
552 if (waitfor != MNT_WAIT)
553 lockreq |= LK_NOWAIT;
555 * Force stale buffer cache information to be flushed.
559 MNT_VNODE_FOREACH(vp, mp, nvp) {
562 if (vp->v_bufobj.bo_dirty.bv_cnt == 0) {
568 if ((error = vget(vp, lockreq, td)) != 0) {
574 error = VOP_FSYNC(vp, waitfor, td);
578 VOP_UNLOCK(vp, 0, td);
587 vfs_stdnosync (mp, waitfor, td)
597 vfs_stdvget (mp, ino, flags, vpp)
608 vfs_stdfhtovp (mp, fhp, vpp)
619 struct vfsconf *vfsp;
627 struct vfsconf *vfsp;
634 vfs_stdextattrctl(mp, cmd, filename_vp, attrnamespace, attrname, td)
637 struct vnode *filename_vp;
639 const char *attrname;
643 if (filename_vp != NULL)
644 VOP_UNLOCK(filename_vp, 0, td);
649 vfs_stdsysctl(mp, op, req)
652 struct sysctl_req *req;
658 /* end of vfs default ops */