Home | History | Annotate | Line # | Download | only in ffs
ffs_vnops.c revision 1.104.4.8
      1 /*	$NetBSD: ffs_vnops.c,v 1.104.4.8 2011/07/17 15:36:04 riz Exp $	*/
      2 
      3 /*-
      4  * Copyright (c) 2008, 2009 The NetBSD Foundation, Inc.
      5  * All rights reserved.
      6  *
      7  * This code is derived from software contributed to The NetBSD Foundation
      8  * by Wasabi Systems, Inc, and by Andrew Doran.
      9  *
     10  * Redistribution and use in source and binary forms, with or without
     11  * modification, are permitted provided that the following conditions
     12  * are met:
     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  *
     19  * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
     20  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
     21  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
     22  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
     23  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
     24  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
     25  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
     26  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
     27  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
     28  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
     29  * POSSIBILITY OF SUCH DAMAGE.
     30  */
     31 
     32 /*
     33  * Copyright (c) 1982, 1986, 1989, 1993
     34  *	The Regents of the University of California.  All rights reserved.
     35  *
     36  * Redistribution and use in source and binary forms, with or without
     37  * modification, are permitted provided that the following conditions
     38  * are met:
     39  * 1. Redistributions of source code must retain the above copyright
     40  *    notice, this list of conditions and the following disclaimer.
     41  * 2. Redistributions in binary form must reproduce the above copyright
     42  *    notice, this list of conditions and the following disclaimer in the
     43  *    documentation and/or other materials provided with the distribution.
     44  * 3. Neither the name of the University nor the names of its contributors
     45  *    may be used to endorse or promote products derived from this software
     46  *    without specific prior written permission.
     47  *
     48  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     49  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     50  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     51  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     52  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     53  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     54  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     55  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     56  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     57  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     58  * SUCH DAMAGE.
     59  *
     60  *	@(#)ffs_vnops.c	8.15 (Berkeley) 5/14/95
     61  */
     62 
     63 #include <sys/cdefs.h>
     64 __KERNEL_RCSID(0, "$NetBSD: ffs_vnops.c,v 1.104.4.8 2011/07/17 15:36:04 riz Exp $");
     65 
     66 #if defined(_KERNEL_OPT)
     67 #include "opt_ffs.h"
     68 #include "opt_wapbl.h"
     69 #endif
     70 
     71 #include <sys/param.h>
     72 #include <sys/systm.h>
     73 #include <sys/resourcevar.h>
     74 #include <sys/kernel.h>
     75 #include <sys/file.h>
     76 #include <sys/stat.h>
     77 #include <sys/buf.h>
     78 #include <sys/event.h>
     79 #include <sys/proc.h>
     80 #include <sys/mount.h>
     81 #include <sys/vnode.h>
     82 #include <sys/pool.h>
     83 #include <sys/signalvar.h>
     84 #include <sys/kauth.h>
     85 #include <sys/wapbl.h>
     86 #include <sys/fstrans.h>
     87 
     88 #include <miscfs/fifofs/fifo.h>
     89 #include <miscfs/genfs/genfs.h>
     90 #include <miscfs/specfs/specdev.h>
     91 
     92 #include <ufs/ufs/inode.h>
     93 #include <ufs/ufs/dir.h>
     94 #include <ufs/ufs/ufs_extern.h>
     95 #include <ufs/ufs/ufsmount.h>
     96 #include <ufs/ufs/ufs_wapbl.h>
     97 
     98 #include <ufs/ffs/fs.h>
     99 #include <ufs/ffs/ffs_extern.h>
    100 
    101 #include <uvm/uvm.h>
    102 
    103 /* Global vfs data structures for ufs. */
    104 int (**ffs_vnodeop_p)(void *);
    105 const struct vnodeopv_entry_desc ffs_vnodeop_entries[] = {
    106 	{ &vop_default_desc, vn_default_error },
    107 	{ &vop_lookup_desc, ufs_lookup },		/* lookup */
    108 	{ &vop_create_desc, ufs_create },		/* create */
    109 	{ &vop_whiteout_desc, ufs_whiteout },		/* whiteout */
    110 	{ &vop_mknod_desc, ufs_mknod },			/* mknod */
    111 	{ &vop_open_desc, ufs_open },			/* open */
    112 	{ &vop_close_desc, ufs_close },			/* close */
    113 	{ &vop_access_desc, ufs_access },		/* access */
    114 	{ &vop_getattr_desc, ufs_getattr },		/* getattr */
    115 	{ &vop_setattr_desc, ufs_setattr },		/* setattr */
    116 	{ &vop_read_desc, ffs_read },			/* read */
    117 	{ &vop_write_desc, ffs_write },			/* write */
    118 	{ &vop_ioctl_desc, ufs_ioctl },			/* ioctl */
    119 	{ &vop_fcntl_desc, ufs_fcntl },			/* fcntl */
    120 	{ &vop_poll_desc, ufs_poll },			/* poll */
    121 	{ &vop_kqfilter_desc, genfs_kqfilter },		/* kqfilter */
    122 	{ &vop_revoke_desc, ufs_revoke },		/* revoke */
    123 	{ &vop_mmap_desc, ufs_mmap },			/* mmap */
    124 	{ &vop_fsync_desc, ffs_fsync },			/* fsync */
    125 	{ &vop_seek_desc, ufs_seek },			/* seek */
    126 	{ &vop_remove_desc, ufs_remove },		/* remove */
    127 	{ &vop_link_desc, ufs_link },			/* link */
    128 	{ &vop_rename_desc, ufs_rename },		/* rename */
    129 	{ &vop_mkdir_desc, ufs_mkdir },			/* mkdir */
    130 	{ &vop_rmdir_desc, ufs_rmdir },			/* rmdir */
    131 	{ &vop_symlink_desc, ufs_symlink },		/* symlink */
    132 	{ &vop_readdir_desc, ufs_readdir },		/* readdir */
    133 	{ &vop_readlink_desc, ufs_readlink },		/* readlink */
    134 	{ &vop_abortop_desc, ufs_abortop },		/* abortop */
    135 	{ &vop_inactive_desc, ufs_inactive },		/* inactive */
    136 	{ &vop_reclaim_desc, ffs_reclaim },		/* reclaim */
    137 	{ &vop_lock_desc, ffs_lock },			/* lock */
    138 	{ &vop_unlock_desc, ffs_unlock },		/* unlock */
    139 	{ &vop_bmap_desc, ufs_bmap },			/* bmap */
    140 	{ &vop_strategy_desc, ufs_strategy },		/* strategy */
    141 	{ &vop_print_desc, ufs_print },			/* print */
    142 	{ &vop_islocked_desc, ffs_islocked },		/* islocked */
    143 	{ &vop_pathconf_desc, ufs_pathconf },		/* pathconf */
    144 	{ &vop_advlock_desc, ufs_advlock },		/* advlock */
    145 	{ &vop_bwrite_desc, vn_bwrite },		/* bwrite */
    146 	{ &vop_getpages_desc, genfs_getpages },		/* getpages */
    147 	{ &vop_putpages_desc, genfs_putpages },		/* putpages */
    148 	{ &vop_openextattr_desc, ffs_openextattr },	/* openextattr */
    149 	{ &vop_closeextattr_desc, ffs_closeextattr },	/* closeextattr */
    150 	{ &vop_getextattr_desc, ffs_getextattr },	/* getextattr */
    151 	{ &vop_setextattr_desc, ffs_setextattr },	/* setextattr */
    152 	{ &vop_listextattr_desc, ffs_listextattr },	/* listextattr */
    153 	{ &vop_deleteextattr_desc, ffs_deleteextattr },	/* deleteextattr */
    154 	{ NULL, NULL }
    155 };
    156 const struct vnodeopv_desc ffs_vnodeop_opv_desc =
    157 	{ &ffs_vnodeop_p, ffs_vnodeop_entries };
    158 
    159 int (**ffs_specop_p)(void *);
    160 const struct vnodeopv_entry_desc ffs_specop_entries[] = {
    161 	{ &vop_default_desc, vn_default_error },
    162 	{ &vop_lookup_desc, spec_lookup },		/* lookup */
    163 	{ &vop_create_desc, spec_create },		/* create */
    164 	{ &vop_mknod_desc, spec_mknod },		/* mknod */
    165 	{ &vop_open_desc, spec_open },			/* open */
    166 	{ &vop_close_desc, ufsspec_close },		/* close */
    167 	{ &vop_access_desc, ufs_access },		/* access */
    168 	{ &vop_getattr_desc, ufs_getattr },		/* getattr */
    169 	{ &vop_setattr_desc, ufs_setattr },		/* setattr */
    170 	{ &vop_read_desc, ufsspec_read },		/* read */
    171 	{ &vop_write_desc, ufsspec_write },		/* write */
    172 	{ &vop_ioctl_desc, spec_ioctl },		/* ioctl */
    173 	{ &vop_fcntl_desc, ufs_fcntl },			/* fcntl */
    174 	{ &vop_poll_desc, spec_poll },			/* poll */
    175 	{ &vop_kqfilter_desc, spec_kqfilter },		/* kqfilter */
    176 	{ &vop_revoke_desc, spec_revoke },		/* revoke */
    177 	{ &vop_mmap_desc, spec_mmap },			/* mmap */
    178 	{ &vop_fsync_desc, ffs_fsync },			/* fsync */
    179 	{ &vop_seek_desc, spec_seek },			/* seek */
    180 	{ &vop_remove_desc, spec_remove },		/* remove */
    181 	{ &vop_link_desc, spec_link },			/* link */
    182 	{ &vop_rename_desc, spec_rename },		/* rename */
    183 	{ &vop_mkdir_desc, spec_mkdir },		/* mkdir */
    184 	{ &vop_rmdir_desc, spec_rmdir },		/* rmdir */
    185 	{ &vop_symlink_desc, spec_symlink },		/* symlink */
    186 	{ &vop_readdir_desc, spec_readdir },		/* readdir */
    187 	{ &vop_readlink_desc, spec_readlink },		/* readlink */
    188 	{ &vop_abortop_desc, spec_abortop },		/* abortop */
    189 	{ &vop_inactive_desc, ufs_inactive },		/* inactive */
    190 	{ &vop_reclaim_desc, ffs_reclaim },		/* reclaim */
    191 	{ &vop_lock_desc, ffs_lock },			/* lock */
    192 	{ &vop_unlock_desc, ffs_unlock },		/* unlock */
    193 	{ &vop_bmap_desc, spec_bmap },			/* bmap */
    194 	{ &vop_strategy_desc, spec_strategy },		/* strategy */
    195 	{ &vop_print_desc, ufs_print },			/* print */
    196 	{ &vop_islocked_desc, ffs_islocked },		/* islocked */
    197 	{ &vop_pathconf_desc, spec_pathconf },		/* pathconf */
    198 	{ &vop_advlock_desc, spec_advlock },		/* advlock */
    199 	{ &vop_bwrite_desc, vn_bwrite },		/* bwrite */
    200 	{ &vop_getpages_desc, spec_getpages },		/* getpages */
    201 	{ &vop_putpages_desc, spec_putpages },		/* putpages */
    202 	{ &vop_openextattr_desc, ffs_openextattr },	/* openextattr */
    203 	{ &vop_closeextattr_desc, ffs_closeextattr },	/* closeextattr */
    204 	{ &vop_getextattr_desc, ffs_getextattr },	/* getextattr */
    205 	{ &vop_setextattr_desc, ffs_setextattr },	/* setextattr */
    206 	{ &vop_listextattr_desc, ffs_listextattr },	/* listextattr */
    207 	{ &vop_deleteextattr_desc, ffs_deleteextattr },	/* deleteextattr */
    208 	{ NULL, NULL }
    209 };
    210 const struct vnodeopv_desc ffs_specop_opv_desc =
    211 	{ &ffs_specop_p, ffs_specop_entries };
    212 
    213 int (**ffs_fifoop_p)(void *);
    214 const struct vnodeopv_entry_desc ffs_fifoop_entries[] = {
    215 	{ &vop_default_desc, vn_default_error },
    216 	{ &vop_lookup_desc, fifo_lookup },		/* lookup */
    217 	{ &vop_create_desc, fifo_create },		/* create */
    218 	{ &vop_mknod_desc, fifo_mknod },		/* mknod */
    219 	{ &vop_open_desc, fifo_open },			/* open */
    220 	{ &vop_close_desc, ufsfifo_close },		/* close */
    221 	{ &vop_access_desc, ufs_access },		/* access */
    222 	{ &vop_getattr_desc, ufs_getattr },		/* getattr */
    223 	{ &vop_setattr_desc, ufs_setattr },		/* setattr */
    224 	{ &vop_read_desc, ufsfifo_read },		/* read */
    225 	{ &vop_write_desc, ufsfifo_write },		/* write */
    226 	{ &vop_ioctl_desc, fifo_ioctl },		/* ioctl */
    227 	{ &vop_fcntl_desc, ufs_fcntl },			/* fcntl */
    228 	{ &vop_poll_desc, fifo_poll },			/* poll */
    229 	{ &vop_kqfilter_desc, fifo_kqfilter },		/* kqfilter */
    230 	{ &vop_revoke_desc, fifo_revoke },		/* revoke */
    231 	{ &vop_mmap_desc, fifo_mmap },			/* mmap */
    232 	{ &vop_fsync_desc, ffs_fsync },			/* fsync */
    233 	{ &vop_seek_desc, fifo_seek },			/* seek */
    234 	{ &vop_remove_desc, fifo_remove },		/* remove */
    235 	{ &vop_link_desc, fifo_link },			/* link */
    236 	{ &vop_rename_desc, fifo_rename },		/* rename */
    237 	{ &vop_mkdir_desc, fifo_mkdir },		/* mkdir */
    238 	{ &vop_rmdir_desc, fifo_rmdir },		/* rmdir */
    239 	{ &vop_symlink_desc, fifo_symlink },		/* symlink */
    240 	{ &vop_readdir_desc, fifo_readdir },		/* readdir */
    241 	{ &vop_readlink_desc, fifo_readlink },		/* readlink */
    242 	{ &vop_abortop_desc, fifo_abortop },		/* abortop */
    243 	{ &vop_inactive_desc, ufs_inactive },		/* inactive */
    244 	{ &vop_reclaim_desc, ffs_reclaim },		/* reclaim */
    245 	{ &vop_lock_desc, ffs_lock },			/* lock */
    246 	{ &vop_unlock_desc, ffs_unlock },		/* unlock */
    247 	{ &vop_bmap_desc, fifo_bmap },			/* bmap */
    248 	{ &vop_strategy_desc, fifo_strategy },		/* strategy */
    249 	{ &vop_print_desc, ufs_print },			/* print */
    250 	{ &vop_islocked_desc, ffs_islocked },		/* islocked */
    251 	{ &vop_pathconf_desc, fifo_pathconf },		/* pathconf */
    252 	{ &vop_advlock_desc, fifo_advlock },		/* advlock */
    253 	{ &vop_bwrite_desc, vn_bwrite },		/* bwrite */
    254 	{ &vop_putpages_desc, fifo_putpages }, 		/* putpages */
    255 	{ &vop_openextattr_desc, ffs_openextattr },	/* openextattr */
    256 	{ &vop_closeextattr_desc, ffs_closeextattr },	/* closeextattr */
    257 	{ &vop_getextattr_desc, ffs_getextattr },	/* getextattr */
    258 	{ &vop_setextattr_desc, ffs_setextattr },	/* setextattr */
    259 	{ &vop_listextattr_desc, ffs_listextattr },	/* listextattr */
    260 	{ &vop_deleteextattr_desc, ffs_deleteextattr },	/* deleteextattr */
    261 	{ NULL, NULL }
    262 };
    263 const struct vnodeopv_desc ffs_fifoop_opv_desc =
    264 	{ &ffs_fifoop_p, ffs_fifoop_entries };
    265 
    266 #include <ufs/ufs/ufs_readwrite.c>
    267 
    268 int
    269 ffs_fsync(void *v)
    270 {
    271 	struct vop_fsync_args /* {
    272 		struct vnode *a_vp;
    273 		kauth_cred_t a_cred;
    274 		int a_flags;
    275 		off_t a_offlo;
    276 		off_t a_offhi;
    277 		struct lwp *a_l;
    278 	} */ *ap = v;
    279 	struct buf *bp;
    280 	int num, error, i;
    281 	struct indir ia[NIADDR + 1];
    282 	int bsize;
    283 	daddr_t blk_high;
    284 	struct vnode *vp;
    285 #ifdef WAPBL
    286 	struct mount *mp;
    287 #endif
    288 
    289 	vp = ap->a_vp;
    290 
    291 	fstrans_start(vp->v_mount, FSTRANS_LAZY);
    292 	/*
    293 	 * XXX no easy way to sync a range in a file with softdep.
    294 	 */
    295 	if ((ap->a_offlo == 0 && ap->a_offhi == 0) || DOINGSOFTDEP(vp) ||
    296 	    (vp->v_type != VREG)) {
    297 		int flags = ap->a_flags;
    298 		error = ffs_full_fsync(vp, flags);
    299 		goto out;
    300 	}
    301 
    302 	bsize = vp->v_mount->mnt_stat.f_iosize;
    303 	blk_high = ap->a_offhi / bsize;
    304 	if (ap->a_offhi % bsize != 0)
    305 		blk_high++;
    306 
    307 	/*
    308 	 * First, flush all pages in range.
    309 	 */
    310 
    311 	mutex_enter(&vp->v_interlock);
    312 	error = VOP_PUTPAGES(vp, trunc_page(ap->a_offlo),
    313 	    round_page(ap->a_offhi), PGO_CLEANIT |
    314 	    ((ap->a_flags & FSYNC_WAIT) ? PGO_SYNCIO : 0));
    315 	if (error) {
    316 		goto out;
    317 	}
    318 
    319 #ifdef WAPBL
    320 	mp = wapbl_vptomp(vp);
    321 	if (mp->mnt_wapbl) {
    322 		/*
    323 		 * Don't bother writing out metadata if the syncer is
    324 		 * making the request.  We will let the sync vnode
    325 		 * write it out in a single burst through a call to
    326 		 * VFS_SYNC().
    327 		 */
    328 		if ((ap->a_flags & (FSYNC_DATAONLY | FSYNC_LAZY)) != 0) {
    329 			fstrans_done(vp->v_mount);
    330 			return 0;
    331 		}
    332 		error = 0;
    333 		if (vp->v_tag == VT_UFS && VTOI(vp)->i_flag &
    334 		    (IN_ACCESS | IN_CHANGE | IN_UPDATE | IN_MODIFY |
    335 				 IN_MODIFIED | IN_ACCESSED)) {
    336 			error = UFS_WAPBL_BEGIN(mp);
    337 			if (error) {
    338 				fstrans_done(vp->v_mount);
    339 				return error;
    340 			}
    341 			error = ffs_update(vp, NULL, NULL, UPDATE_CLOSE |
    342 			    ((ap->a_flags & FSYNC_WAIT) ? UPDATE_WAIT : 0));
    343 			UFS_WAPBL_END(mp);
    344 		}
    345 		if (error || (ap->a_flags & FSYNC_NOLOG) != 0) {
    346 			fstrans_done(vp->v_mount);
    347 			return error;
    348 		}
    349 		error = wapbl_flush(mp->mnt_wapbl, 0);
    350 		fstrans_done(vp->v_mount);
    351 		return error;
    352 	}
    353 #endif /* WAPBL */
    354 
    355 	/*
    356 	 * Then, flush indirect blocks.
    357 	 */
    358 
    359 	if (blk_high >= NDADDR) {
    360 		error = ufs_getlbns(vp, blk_high, ia, &num);
    361 		if (error)
    362 			goto out;
    363 
    364 		mutex_enter(&bufcache_lock);
    365 		for (i = 0; i < num; i++) {
    366 			if ((bp = incore(vp, ia[i].in_lbn)) == NULL)
    367 				continue;
    368 			if ((bp->b_cflags & BC_BUSY) != 0 ||
    369 			    (bp->b_oflags & BO_DELWRI) == 0)
    370 				continue;
    371 			bp->b_cflags |= BC_BUSY | BC_VFLUSH;
    372 			mutex_exit(&bufcache_lock);
    373 			bawrite(bp);
    374 			mutex_enter(&bufcache_lock);
    375 		}
    376 		mutex_exit(&bufcache_lock);
    377 	}
    378 
    379 	if (ap->a_flags & FSYNC_WAIT) {
    380 		mutex_enter(&vp->v_interlock);
    381 		while (vp->v_numoutput > 0)
    382 			cv_wait(&vp->v_cv, &vp->v_interlock);
    383 		mutex_exit(&vp->v_interlock);
    384 	}
    385 
    386 	error = ffs_update(vp, NULL, NULL, UPDATE_CLOSE |
    387 	    (((ap->a_flags & (FSYNC_WAIT | FSYNC_DATAONLY)) == FSYNC_WAIT)
    388 	    ? UPDATE_WAIT : 0));
    389 
    390 	if (error == 0 && ap->a_flags & FSYNC_CACHE) {
    391 		int l = 0;
    392 		VOP_IOCTL(VTOI(vp)->i_devvp, DIOCCACHESYNC, &l, FWRITE,
    393 			curlwp->l_cred);
    394 	}
    395 
    396 out:
    397 	fstrans_done(vp->v_mount);
    398 	return error;
    399 }
    400 
    401 /*
    402  * Synch an open file.  Called for VOP_FSYNC().
    403  */
    404 /* ARGSUSED */
    405 int
    406 ffs_full_fsync(struct vnode *vp, int flags)
    407 {
    408 	struct buf *bp, *nbp;
    409 	int error, passes, skipmeta, inodedeps_only, waitfor, i;
    410 	struct mount *mp;
    411 
    412 	KASSERT(VTOI(vp) != NULL);
    413 	KASSERT(vp->v_tag == VT_UFS);
    414 
    415 	error = 0;
    416 
    417 	mp = vp->v_mount;
    418 	if (vp->v_type == VBLK && vp->v_specmountpoint != NULL) {
    419 		mp = vp->v_specmountpoint;
    420 		if ((mp->mnt_flag & MNT_SOFTDEP) != 0)
    421 			softdep_fsync_mountdev(vp);
    422 	} else {
    423 		mp = vp->v_mount;
    424 	}
    425 
    426 	mutex_enter(&vp->v_interlock);
    427 
    428 	inodedeps_only = DOINGSOFTDEP(vp) && (flags & FSYNC_RECLAIM)
    429 	    && UVM_OBJ_IS_CLEAN(&vp->v_uobj) && LIST_EMPTY(&vp->v_dirtyblkhd);
    430 
    431 	/*
    432 	 * Flush all dirty data associated with the vnode.
    433 	 */
    434 	if (vp->v_type == VREG || vp->v_type == VBLK) {
    435 		int pflags = PGO_ALLPAGES | PGO_CLEANIT;
    436 
    437 		if ((flags & FSYNC_WAIT))
    438 			pflags |= PGO_SYNCIO;
    439 		if (vp->v_type == VREG &&
    440 		    fstrans_getstate(mp) == FSTRANS_SUSPENDING)
    441 			pflags |= PGO_FREE;
    442 		error = VOP_PUTPAGES(vp, 0, 0, pflags);
    443 		if (error)
    444 			return error;
    445 	} else {
    446 		mutex_exit(&vp->v_interlock);
    447 	}
    448 
    449 #ifdef WAPBL
    450 	if (mp && mp->mnt_wapbl) {
    451 		/*
    452 		 * Don't bother writing out metadata if the syncer is
    453 		 * making the request.  We will let the sync vnode
    454 		 * write it out in a single burst through a call to
    455 		 * VFS_SYNC().
    456 		 */
    457 		if ((flags & (FSYNC_DATAONLY | FSYNC_LAZY)) != 0)
    458 			return 0;
    459 
    460 		if ((VTOI(vp)->i_flag & (IN_ACCESS | IN_CHANGE | IN_UPDATE
    461 		    | IN_MODIFY | IN_MODIFIED | IN_ACCESSED)) != 0) {
    462 			error = UFS_WAPBL_BEGIN(mp);
    463 			if (error)
    464 				return error;
    465 			error = ffs_update(vp, NULL, NULL, UPDATE_CLOSE |
    466 			    ((flags & FSYNC_WAIT) ? UPDATE_WAIT : 0));
    467 			UFS_WAPBL_END(mp);
    468 		}
    469 		if (error || (flags & FSYNC_NOLOG) != 0)
    470 			return error;
    471 
    472 		/*
    473 		 * Don't flush the log if the vnode being flushed
    474 		 * contains no dirty buffers that could be in the log.
    475 		 */
    476 		if (!LIST_EMPTY(&vp->v_dirtyblkhd)) {
    477 			error = wapbl_flush(mp->mnt_wapbl, 0);
    478 			if (error)
    479 				return error;
    480 		}
    481 
    482 		if ((flags & FSYNC_WAIT) != 0) {
    483 			mutex_enter(&vp->v_interlock);
    484 			while (vp->v_numoutput != 0)
    485 				cv_wait(&vp->v_cv, &vp->v_interlock);
    486 			mutex_exit(&vp->v_interlock);
    487 		}
    488 
    489 		return error;
    490 	}
    491 #endif /* WAPBL */
    492 
    493 	/*
    494 	 * Write out metadata for non-logging file systems.  This block can
    495 	 * be simplified once softdep goes.
    496 	 */
    497 	passes = NIADDR + 1;
    498 	skipmeta = 0;
    499 	if (flags & FSYNC_WAIT)
    500 		skipmeta = 1;
    501 
    502 loop:
    503 	mutex_enter(&bufcache_lock);
    504 	LIST_FOREACH(bp, &vp->v_dirtyblkhd, b_vnbufs) {
    505 		bp->b_cflags &= ~BC_SCANNED;
    506 	}
    507 	for (bp = LIST_FIRST(&vp->v_dirtyblkhd); bp; bp = nbp) {
    508 		nbp = LIST_NEXT(bp, b_vnbufs);
    509 		if (bp->b_cflags & (BC_BUSY | BC_SCANNED))
    510 			continue;
    511 		if ((bp->b_oflags & BO_DELWRI) == 0)
    512 			panic("ffs_fsync: not dirty");
    513 		if (skipmeta && bp->b_lblkno < 0)
    514 			continue;
    515 		bp->b_cflags |= BC_BUSY | BC_VFLUSH | BC_SCANNED;
    516 		mutex_exit(&bufcache_lock);
    517 		/*
    518 		 * On our final pass through, do all I/O synchronously
    519 		 * so that we can find out if our flush is failing
    520 		 * because of write errors.
    521 		 */
    522 		if (passes > 0 || !(flags & FSYNC_WAIT))
    523 			(void) bawrite(bp);
    524 		else if ((error = bwrite(bp)) != 0)
    525 			return (error);
    526 		/*
    527 		 * Since we unlocked during the I/O, we need
    528 		 * to start from a known point.
    529 		 */
    530 		mutex_enter(&bufcache_lock);
    531 		nbp = LIST_FIRST(&vp->v_dirtyblkhd);
    532 	}
    533 	mutex_exit(&bufcache_lock);
    534 	if (skipmeta) {
    535 		skipmeta = 0;
    536 		goto loop;
    537 	}
    538 
    539 	if ((flags & FSYNC_WAIT) != 0) {
    540 		mutex_enter(&vp->v_interlock);
    541 		while (vp->v_numoutput) {
    542 			cv_wait(&vp->v_cv, &vp->v_interlock);
    543 		}
    544 		mutex_exit(&vp->v_interlock);
    545 
    546 		/*
    547 		 * Ensure that any filesystem metadata associated
    548 		 * with the vnode has been written.
    549 		 */
    550 		if ((error = softdep_sync_metadata(vp)) != 0)
    551 			return (error);
    552 
    553 		if (!LIST_EMPTY(&vp->v_dirtyblkhd)) {
    554 			/*
    555 			* Block devices associated with filesystems may
    556 			* have new I/O requests posted for them even if
    557 			* the vnode is locked, so no amount of trying will
    558 			* get them clean. Thus we give block devices a
    559 			* good effort, then just give up. For all other file
    560 			* types, go around and try again until it is clean.
    561 			*/
    562 			if (passes > 0) {
    563 				passes--;
    564 				goto loop;
    565 			}
    566 #ifdef DIAGNOSTIC
    567 			if (vp->v_type != VBLK)
    568 				vprint("ffs_fsync: dirty", vp);
    569 #endif
    570 		}
    571 	}
    572 
    573 	if (inodedeps_only)
    574 		waitfor = 0;
    575 	else
    576 		waitfor = (flags & FSYNC_WAIT) != 0 ? UPDATE_WAIT : 0;
    577 	error = ffs_update(vp, NULL, NULL, UPDATE_CLOSE | waitfor);
    578 
    579 	if (error == 0 && (flags & FSYNC_CACHE) != 0) {
    580 		(void)VOP_IOCTL(VTOI(vp)->i_devvp, DIOCCACHESYNC, &i, FWRITE,
    581 		    kauth_cred_get());
    582 	}
    583 
    584 	return error;
    585 }
    586 
    587 /*
    588  * Reclaim an inode so that it can be used for other purposes.
    589  */
    590 int
    591 ffs_reclaim(void *v)
    592 {
    593 	struct vop_reclaim_args /* {
    594 		struct vnode *a_vp;
    595 		struct lwp *a_l;
    596 	} */ *ap = v;
    597 	struct vnode *vp = ap->a_vp;
    598 	struct inode *ip = VTOI(vp);
    599 	struct mount *mp = vp->v_mount;
    600 	struct ufsmount *ump = ip->i_ump;
    601 	void *data;
    602 	int error;
    603 
    604 	fstrans_start(mp, FSTRANS_LAZY);
    605 	if ((error = ufs_reclaim(vp)) != 0) {
    606 		fstrans_done(mp);
    607 		return (error);
    608 	}
    609 	if (ip->i_din.ffs1_din != NULL) {
    610 		if (ump->um_fstype == UFS1)
    611 			pool_cache_put(ffs_dinode1_cache, ip->i_din.ffs1_din);
    612 		else
    613 			pool_cache_put(ffs_dinode2_cache, ip->i_din.ffs2_din);
    614 	}
    615 	/*
    616 	 * To interlock with ffs_sync().
    617 	 */
    618 	genfs_node_destroy(vp);
    619 	mutex_enter(&vp->v_interlock);
    620 	data = vp->v_data;
    621 	vp->v_data = NULL;
    622 	mutex_exit(&vp->v_interlock);
    623 
    624 	/*
    625 	 * XXX MFS ends up here, too, to free an inode.  Should we create
    626 	 * XXX a separate pool for MFS inodes?
    627 	 */
    628 	pool_cache_put(ffs_inode_cache, data);
    629 	fstrans_done(mp);
    630 	return (0);
    631 }
    632 
    633 #if 0
    634 int
    635 ffs_getpages(void *v)
    636 {
    637 	struct vop_getpages_args /* {
    638 		struct vnode *a_vp;
    639 		voff_t a_offset;
    640 		struct vm_page **a_m;
    641 		int *a_count;
    642 		int a_centeridx;
    643 		vm_prot_t a_access_type;
    644 		int a_advice;
    645 		int a_flags;
    646 	} */ *ap = v;
    647 	struct vnode *vp = ap->a_vp;
    648 	struct inode *ip = VTOI(vp);
    649 	struct fs *fs = ip->i_fs;
    650 
    651 	/*
    652 	 * don't allow a softdep write to create pages for only part of a block.
    653 	 * the dependency tracking requires that all pages be in memory for
    654 	 * a block involved in a dependency.
    655 	 */
    656 
    657 	if (ap->a_flags & PGO_OVERWRITE &&
    658 	    (blkoff(fs, ap->a_offset) != 0 ||
    659 	     blkoff(fs, *ap->a_count << PAGE_SHIFT) != 0) &&
    660 	    DOINGSOFTDEP(ap->a_vp)) {
    661 		if ((ap->a_flags & PGO_LOCKED) == 0) {
    662 			mutex_exit(&vp->v_interlock);
    663 		}
    664 		return EINVAL;
    665 	}
    666 	return genfs_getpages(v);
    667 }
    668 #endif
    669 
    670 /*
    671  * Return the last logical file offset that should be written for this file
    672  * if we're doing a write that ends at "size".
    673  */
    674 
    675 void
    676 ffs_gop_size(struct vnode *vp, off_t size, off_t *eobp, int flags)
    677 {
    678 	struct inode *ip = VTOI(vp);
    679 	struct fs *fs = ip->i_fs;
    680 	daddr_t olbn, nlbn;
    681 
    682 	olbn = lblkno(fs, ip->i_size);
    683 	nlbn = lblkno(fs, size);
    684 	if (nlbn < NDADDR && olbn <= nlbn) {
    685 		*eobp = fragroundup(fs, size);
    686 	} else {
    687 		*eobp = blkroundup(fs, size);
    688 	}
    689 }
    690 
    691 int
    692 ffs_openextattr(void *v)
    693 {
    694 	struct vop_openextattr_args /* {
    695 		struct vnode *a_vp;
    696 		kauth_cred_t a_cred;
    697 		struct proc *a_p;
    698 	} */ *ap = v;
    699 	struct inode *ip = VTOI(ap->a_vp);
    700 	struct fs *fs = ip->i_fs;
    701 
    702 	/* Not supported for UFS1 file systems. */
    703 	if (fs->fs_magic == FS_UFS1_MAGIC)
    704 		return (EOPNOTSUPP);
    705 
    706 	/* XXX Not implemented for UFS2 file systems. */
    707 	return (EOPNOTSUPP);
    708 }
    709 
    710 int
    711 ffs_closeextattr(void *v)
    712 {
    713 	struct vop_closeextattr_args /* {
    714 		struct vnode *a_vp;
    715 		int a_commit;
    716 		kauth_cred_t a_cred;
    717 		struct proc *a_p;
    718 	} */ *ap = v;
    719 	struct inode *ip = VTOI(ap->a_vp);
    720 	struct fs *fs = ip->i_fs;
    721 
    722 	/* Not supported for UFS1 file systems. */
    723 	if (fs->fs_magic == FS_UFS1_MAGIC)
    724 		return (EOPNOTSUPP);
    725 
    726 	/* XXX Not implemented for UFS2 file systems. */
    727 	return (EOPNOTSUPP);
    728 }
    729 
    730 int
    731 ffs_getextattr(void *v)
    732 {
    733 	struct vop_getextattr_args /* {
    734 		struct vnode *a_vp;
    735 		int a_attrnamespace;
    736 		const char *a_name;
    737 		struct uio *a_uio;
    738 		size_t *a_size;
    739 		kauth_cred_t a_cred;
    740 		struct proc *a_p;
    741 	} */ *ap = v;
    742 	struct vnode *vp = ap->a_vp;
    743 	struct inode *ip = VTOI(vp);
    744 	struct fs *fs = ip->i_fs;
    745 
    746 	if (fs->fs_magic == FS_UFS1_MAGIC) {
    747 #ifdef UFS_EXTATTR
    748 		int error;
    749 
    750 		fstrans_start(vp->v_mount, FSTRANS_SHARED);
    751 		error = ufs_getextattr(ap);
    752 		fstrans_done(vp->v_mount);
    753 		return error;
    754 #else
    755 		return (EOPNOTSUPP);
    756 #endif
    757 	}
    758 
    759 	/* XXX Not implemented for UFS2 file systems. */
    760 	return (EOPNOTSUPP);
    761 }
    762 
    763 int
    764 ffs_setextattr(void *v)
    765 {
    766 	struct vop_setextattr_args /* {
    767 		struct vnode *a_vp;
    768 		int a_attrnamespace;
    769 		const char *a_name;
    770 		struct uio *a_uio;
    771 		kauth_cred_t a_cred;
    772 		struct proc *a_p;
    773 	} */ *ap = v;
    774 	struct vnode *vp = ap->a_vp;
    775 	struct inode *ip = VTOI(vp);
    776 	struct fs *fs = ip->i_fs;
    777 
    778 	if (fs->fs_magic == FS_UFS1_MAGIC) {
    779 #ifdef UFS_EXTATTR
    780 		int error;
    781 
    782 		fstrans_start(vp->v_mount, FSTRANS_SHARED);
    783 		error = ufs_setextattr(ap);
    784 		fstrans_done(vp->v_mount);
    785 		return error;
    786 #else
    787 		return (EOPNOTSUPP);
    788 #endif
    789 	}
    790 
    791 	/* XXX Not implemented for UFS2 file systems. */
    792 	return (EOPNOTSUPP);
    793 }
    794 
    795 int
    796 ffs_listextattr(void *v)
    797 {
    798 	struct vop_listextattr_args /* {
    799 		struct vnode *a_vp;
    800 		int a_attrnamespace;
    801 		struct uio *a_uio;
    802 		size_t *a_size;
    803 		kauth_cred_t a_cred;
    804 		struct proc *a_p;
    805 	} */ *ap = v;
    806 	struct inode *ip = VTOI(ap->a_vp);
    807 	struct fs *fs = ip->i_fs;
    808 
    809 	if (fs->fs_magic == FS_UFS1_MAGIC) {
    810 #ifdef UFS_EXTATTR
    811 		struct vnode *vp = ap->a_vp;
    812 		int error;
    813 
    814 		fstrans_start(vp->v_mount, FSTRANS_SHARED);
    815 		error = ufs_listextattr(ap);
    816 		fstrans_done(vp->v_mount);
    817 		return error;
    818 #else
    819 		return (EOPNOTSUPP);
    820 #endif
    821 	}
    822 
    823 	/* XXX Not implemented for UFS2 file systems. */
    824 	return (EOPNOTSUPP);
    825 }
    826 
    827 int
    828 ffs_deleteextattr(void *v)
    829 {
    830 	struct vop_deleteextattr_args /* {
    831 		struct vnode *a_vp;
    832 		int a_attrnamespace;
    833 		kauth_cred_t a_cred;
    834 		struct proc *a_p;
    835 	} */ *ap = v;
    836 	struct vnode *vp = ap->a_vp;
    837 	struct inode *ip = VTOI(vp);
    838 	struct fs *fs = ip->i_fs;
    839 
    840 	if (fs->fs_magic == FS_UFS1_MAGIC) {
    841 #ifdef UFS_EXTATTR
    842 		int error;
    843 
    844 		fstrans_start(vp->v_mount, FSTRANS_SHARED);
    845 		error = ufs_deleteextattr(ap);
    846 		fstrans_done(vp->v_mount);
    847 		return error;
    848 #else
    849 		return (EOPNOTSUPP);
    850 #endif
    851 	}
    852 
    853 	/* XXX Not implemented for UFS2 file systems. */
    854 	return (EOPNOTSUPP);
    855 }
    856 
    857 /*
    858  * Lock the node.
    859  */
    860 int
    861 ffs_lock(void *v)
    862 {
    863 	struct vop_lock_args /* {
    864 		struct vnode *a_vp;
    865 		int a_flags;
    866 	} */ *ap = v;
    867 	struct vnode *vp = ap->a_vp;
    868 	struct mount *mp = vp->v_mount;
    869 	int flags = ap->a_flags;
    870 
    871 	if ((flags & LK_INTERLOCK) != 0) {
    872 		mutex_exit(&vp->v_interlock);
    873 		flags &= ~LK_INTERLOCK;
    874 	}
    875 
    876 	/*
    877 	 * Fake lock during file system suspension.
    878 	 */
    879 	if ((vp->v_type == VREG || vp->v_type == VDIR) &&
    880 	    fstrans_is_owner(mp) &&
    881 	    fstrans_getstate(mp) == FSTRANS_SUSPENDING) {
    882 		return 0;
    883 	}
    884 
    885 	return (vlockmgr(vp->v_vnlock, flags));
    886 }
    887 
    888 /*
    889  * Unlock the node.
    890  */
    891 int
    892 ffs_unlock(void *v)
    893 {
    894 	struct vop_unlock_args /* {
    895 		struct vnode *a_vp;
    896 		int a_flags;
    897 	} */ *ap = v;
    898 	struct vnode *vp = ap->a_vp;
    899 	struct mount *mp = vp->v_mount;
    900 
    901 	KASSERT(ap->a_flags == 0);
    902 
    903 	/*
    904 	 * Fake unlock during file system suspension.
    905 	 */
    906 	if ((vp->v_type == VREG || vp->v_type == VDIR) &&
    907 	    fstrans_is_owner(mp) &&
    908 	    fstrans_getstate(mp) == FSTRANS_SUSPENDING) {
    909 		return 0;
    910 	}
    911 	return (vlockmgr(vp->v_vnlock, LK_RELEASE));
    912 }
    913 
    914 /*
    915  * Return whether or not the node is locked.
    916  */
    917 int
    918 ffs_islocked(void *v)
    919 {
    920 	struct vop_islocked_args /* {
    921 		struct vnode *a_vp;
    922 	} */ *ap = v;
    923 	struct vnode *vp = ap->a_vp;
    924 
    925 	return (vlockstatus(vp->v_vnlock));
    926 }
    927