| 1 | /* $NetBSD: nfs_vfsops.c,v 1.231 2015/11/02 09:57:43 pgoyette Exp $ */ |
| 2 | |
| 3 | /* |
| 4 | * Copyright (c) 1989, 1993, 1995 |
| 5 | * The Regents of the University of California. All rights reserved. |
| 6 | * |
| 7 | * This code is derived from software contributed to Berkeley by |
| 8 | * Rick Macklem at The University of Guelph. |
| 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 | * 3. 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. |
| 21 | * |
| 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 |
| 32 | * SUCH DAMAGE. |
| 33 | * |
| 34 | * @(#)nfs_vfsops.c 8.12 (Berkeley) 5/20/95 |
| 35 | */ |
| 36 | |
| 37 | #include <sys/cdefs.h> |
| 38 | __KERNEL_RCSID(0, "$NetBSD: nfs_vfsops.c,v 1.231 2015/11/02 09:57:43 pgoyette Exp $" ); |
| 39 | |
| 40 | #if defined(_KERNEL_OPT) |
| 41 | #include "opt_nfs.h" |
| 42 | #endif |
| 43 | |
| 44 | #include <sys/param.h> |
| 45 | #include <sys/ioctl.h> |
| 46 | #include <sys/signal.h> |
| 47 | #include <sys/proc.h> |
| 48 | #include <sys/namei.h> |
| 49 | #include <sys/device.h> |
| 50 | #include <sys/vnode.h> |
| 51 | #include <sys/kernel.h> |
| 52 | #include <sys/mount.h> |
| 53 | #include <sys/buf.h> |
| 54 | #include <sys/mbuf.h> |
| 55 | #include <sys/dirent.h> |
| 56 | #include <sys/socket.h> |
| 57 | #include <sys/socketvar.h> |
| 58 | #include <sys/sysctl.h> |
| 59 | #include <sys/systm.h> |
| 60 | #include <sys/timetc.h> |
| 61 | #include <sys/kauth.h> |
| 62 | #include <sys/module.h> |
| 63 | |
| 64 | #include <net/if.h> |
| 65 | #include <net/route.h> |
| 66 | #include <netinet/in.h> |
| 67 | |
| 68 | #include <nfs/rpcv2.h> |
| 69 | #include <nfs/nfsproto.h> |
| 70 | #include <nfs/nfsnode.h> |
| 71 | #include <nfs/nfs.h> |
| 72 | #include <nfs/nfsmount.h> |
| 73 | #include <nfs/xdr_subs.h> |
| 74 | #include <nfs/nfsm_subs.h> |
| 75 | #include <nfs/nfsdiskless.h> |
| 76 | #include <nfs/nfs_var.h> |
| 77 | |
| 78 | MODULE(MODULE_CLASS_VFS, nfs, NULL); |
| 79 | |
| 80 | extern struct nfsstats nfsstats; |
| 81 | extern int nfs_ticks; |
| 82 | |
| 83 | /* |
| 84 | * keep a count of the nfs mounts to generate ficticious drive names |
| 85 | * for the per drive stats. |
| 86 | */ |
| 87 | unsigned int nfs_mount_count = 0; |
| 88 | |
| 89 | int nfs_commitsize; |
| 90 | |
| 91 | /* |
| 92 | * nfs vfs operations. |
| 93 | */ |
| 94 | |
| 95 | extern const struct vnodeopv_desc nfsv2_vnodeop_opv_desc; |
| 96 | extern const struct vnodeopv_desc spec_nfsv2nodeop_opv_desc; |
| 97 | extern const struct vnodeopv_desc fifo_nfsv2nodeop_opv_desc; |
| 98 | |
| 99 | const struct vnodeopv_desc * const nfs_vnodeopv_descs[] = { |
| 100 | &nfsv2_vnodeop_opv_desc, |
| 101 | &spec_nfsv2nodeop_opv_desc, |
| 102 | &fifo_nfsv2nodeop_opv_desc, |
| 103 | NULL, |
| 104 | }; |
| 105 | |
| 106 | struct vfsops nfs_vfsops = { |
| 107 | .vfs_name = MOUNT_NFS, |
| 108 | .vfs_min_mount_data = sizeof (struct nfs_args), |
| 109 | .vfs_mount = nfs_mount, |
| 110 | .vfs_start = nfs_start, |
| 111 | .vfs_unmount = nfs_unmount, |
| 112 | .vfs_root = nfs_root, |
| 113 | .vfs_quotactl = (void *)eopnotsupp, |
| 114 | .vfs_statvfs = nfs_statvfs, |
| 115 | .vfs_sync = nfs_sync, |
| 116 | .vfs_loadvnode = nfs_loadvnode, |
| 117 | .vfs_vget = nfs_vget, |
| 118 | .vfs_fhtovp = nfs_fhtovp, |
| 119 | .vfs_vptofh = nfs_vptofh, |
| 120 | .vfs_init = nfs_vfs_init, |
| 121 | .vfs_done = nfs_vfs_done, |
| 122 | .vfs_mountroot = nfs_mountroot, |
| 123 | .vfs_snapshot = (void *)eopnotsupp, |
| 124 | .vfs_extattrctl = vfs_stdextattrctl, |
| 125 | .vfs_suspendctl = (void *)eopnotsupp, |
| 126 | .vfs_renamelock_enter = genfs_renamelock_enter, |
| 127 | .vfs_renamelock_exit = genfs_renamelock_exit, |
| 128 | .vfs_fsync = (void *)eopnotsupp, |
| 129 | .vfs_opv_descs = nfs_vnodeopv_descs |
| 130 | }; |
| 131 | |
| 132 | extern u_int32_t nfs_procids[NFS_NPROCS]; |
| 133 | extern u_int32_t nfs_prog, nfs_vers; |
| 134 | static struct sysctllog *nfs_clog; |
| 135 | |
| 136 | static int nfs_mount_diskless(struct nfs_dlmount *, const char *, |
| 137 | struct mount **, struct vnode **, struct lwp *); |
| 138 | static void nfs_sysctl_init(void); |
| 139 | static void nfs_sysctl_fini(void); |
| 140 | |
| 141 | static int |
| 142 | nfs_modcmd(modcmd_t cmd, void *arg) |
| 143 | { |
| 144 | int error; |
| 145 | |
| 146 | switch (cmd) { |
| 147 | case MODULE_CMD_INIT: |
| 148 | error = vfs_attach(&nfs_vfsops); |
| 149 | if (error == 0) { |
| 150 | nfs_sysctl_init(); |
| 151 | } |
| 152 | return error; |
| 153 | case MODULE_CMD_FINI: |
| 154 | error = vfs_detach(&nfs_vfsops); |
| 155 | if (error == 0) { |
| 156 | nfs_sysctl_fini(); |
| 157 | } |
| 158 | return error; |
| 159 | default: |
| 160 | return ENOTTY; |
| 161 | } |
| 162 | } |
| 163 | |
| 164 | /* |
| 165 | * nfs statvfs call |
| 166 | */ |
| 167 | int |
| 168 | nfs_statvfs(struct mount *mp, struct statvfs *sbp) |
| 169 | { |
| 170 | struct lwp *l = curlwp; |
| 171 | struct vnode *vp; |
| 172 | struct nfs_statfs *sfp; |
| 173 | char *cp; |
| 174 | u_int32_t *tl; |
| 175 | int32_t t1, t2; |
| 176 | char *bpos, *dpos, *cp2; |
| 177 | struct nfsmount *nmp = VFSTONFS(mp); |
| 178 | int error = 0, retattr; |
| 179 | #ifdef NFS_V2_ONLY |
| 180 | const int v3 = 0; |
| 181 | #else |
| 182 | int v3 = (nmp->nm_flag & NFSMNT_NFSV3); |
| 183 | #endif |
| 184 | struct mbuf *mreq, *mrep = NULL, *md, *mb; |
| 185 | kauth_cred_t cred; |
| 186 | u_quad_t tquad; |
| 187 | struct nfsnode *np; |
| 188 | |
| 189 | #ifndef nolint |
| 190 | sfp = (struct nfs_statfs *)0; |
| 191 | #endif |
| 192 | vp = nmp->nm_vnode; |
| 193 | np = VTONFS(vp); |
| 194 | cred = kauth_cred_alloc(); |
| 195 | #ifndef NFS_V2_ONLY |
| 196 | if (v3 && (nmp->nm_iflag & NFSMNT_GOTFSINFO) == 0) |
| 197 | (void)nfs_fsinfo(nmp, vp, cred, l); |
| 198 | #endif |
| 199 | nfsstats.rpccnt[NFSPROC_FSSTAT]++; |
| 200 | nfsm_reqhead(np, NFSPROC_FSSTAT, NFSX_FH(v3)); |
| 201 | nfsm_fhtom(np, v3); |
| 202 | nfsm_request(np, NFSPROC_FSSTAT, l, cred); |
| 203 | if (v3) |
| 204 | nfsm_postop_attr(vp, retattr, 0); |
| 205 | if (error) { |
| 206 | if (mrep != NULL) { |
| 207 | if (mrep->m_next != NULL) |
| 208 | printf("nfs_vfsops: nfs_statvfs would lose buffers\n" ); |
| 209 | m_freem(mrep); |
| 210 | } |
| 211 | goto nfsmout; |
| 212 | } |
| 213 | nfsm_dissect(sfp, struct nfs_statfs *, NFSX_STATFS(v3)); |
| 214 | sbp->f_flag = nmp->nm_flag; |
| 215 | sbp->f_iosize = min(nmp->nm_rsize, nmp->nm_wsize); |
| 216 | if (v3) { |
| 217 | sbp->f_frsize = sbp->f_bsize = NFS_FABLKSIZE; |
| 218 | tquad = fxdr_hyper(&sfp->sf_tbytes); |
| 219 | sbp->f_blocks = ((quad_t)tquad / (quad_t)NFS_FABLKSIZE); |
| 220 | tquad = fxdr_hyper(&sfp->sf_fbytes); |
| 221 | sbp->f_bfree = ((quad_t)tquad / (quad_t)NFS_FABLKSIZE); |
| 222 | tquad = fxdr_hyper(&sfp->sf_abytes); |
| 223 | tquad = ((quad_t)tquad / (quad_t)NFS_FABLKSIZE); |
| 224 | sbp->f_bresvd = sbp->f_bfree - tquad; |
| 225 | sbp->f_bavail = tquad; |
| 226 | /* Handle older NFS servers returning negative values */ |
| 227 | if ((quad_t)sbp->f_bavail < 0) |
| 228 | sbp->f_bavail = 0; |
| 229 | tquad = fxdr_hyper(&sfp->sf_tfiles); |
| 230 | sbp->f_files = tquad; |
| 231 | tquad = fxdr_hyper(&sfp->sf_ffiles); |
| 232 | sbp->f_ffree = tquad; |
| 233 | sbp->f_favail = tquad; |
| 234 | sbp->f_fresvd = 0; |
| 235 | sbp->f_namemax = NFS_MAXNAMLEN; |
| 236 | } else { |
| 237 | sbp->f_bsize = NFS_FABLKSIZE; |
| 238 | sbp->f_frsize = fxdr_unsigned(int32_t, sfp->sf_bsize); |
| 239 | sbp->f_blocks = fxdr_unsigned(int32_t, sfp->sf_blocks); |
| 240 | sbp->f_bfree = fxdr_unsigned(int32_t, sfp->sf_bfree); |
| 241 | sbp->f_bavail = fxdr_unsigned(int32_t, sfp->sf_bavail); |
| 242 | sbp->f_fresvd = 0; |
| 243 | sbp->f_files = 0; |
| 244 | sbp->f_ffree = 0; |
| 245 | sbp->f_favail = 0; |
| 246 | sbp->f_fresvd = 0; |
| 247 | sbp->f_namemax = NFS_MAXNAMLEN; |
| 248 | } |
| 249 | copy_statvfs_info(sbp, mp); |
| 250 | nfsm_reqdone; |
| 251 | kauth_cred_free(cred); |
| 252 | return (error); |
| 253 | } |
| 254 | |
| 255 | #ifndef NFS_V2_ONLY |
| 256 | /* |
| 257 | * nfs version 3 fsinfo rpc call |
| 258 | */ |
| 259 | int |
| 260 | nfs_fsinfo(struct nfsmount *nmp, struct vnode *vp, kauth_cred_t cred, struct lwp *l) |
| 261 | { |
| 262 | struct nfsv3_fsinfo *fsp; |
| 263 | char *cp; |
| 264 | int32_t t1, t2; |
| 265 | u_int32_t *tl, pref, xmax; |
| 266 | char *bpos, *dpos, *cp2; |
| 267 | int error = 0, retattr; |
| 268 | struct mbuf *mreq, *mrep, *md, *mb; |
| 269 | u_int64_t maxfsize; |
| 270 | struct nfsnode *np = VTONFS(vp); |
| 271 | |
| 272 | nfsstats.rpccnt[NFSPROC_FSINFO]++; |
| 273 | nfsm_reqhead(np, NFSPROC_FSINFO, NFSX_FH(1)); |
| 274 | nfsm_fhtom(np, 1); |
| 275 | nfsm_request(np, NFSPROC_FSINFO, l, cred); |
| 276 | nfsm_postop_attr(vp, retattr, 0); |
| 277 | if (!error) { |
| 278 | nfsm_dissect(fsp, struct nfsv3_fsinfo *, NFSX_V3FSINFO); |
| 279 | pref = fxdr_unsigned(u_int32_t, fsp->fs_wtpref); |
| 280 | if ((nmp->nm_flag & NFSMNT_WSIZE) == 0 && |
| 281 | pref < nmp->nm_wsize && pref >= NFS_FABLKSIZE) |
| 282 | nmp->nm_wsize = (pref + NFS_FABLKSIZE - 1) & |
| 283 | ~(NFS_FABLKSIZE - 1); |
| 284 | xmax = fxdr_unsigned(u_int32_t, fsp->fs_wtmax); |
| 285 | if (xmax < nmp->nm_wsize && xmax > 0) { |
| 286 | nmp->nm_wsize = xmax & ~(NFS_FABLKSIZE - 1); |
| 287 | if (nmp->nm_wsize == 0) |
| 288 | nmp->nm_wsize = xmax; |
| 289 | } |
| 290 | pref = fxdr_unsigned(u_int32_t, fsp->fs_rtpref); |
| 291 | if ((nmp->nm_flag & NFSMNT_RSIZE) == 0 && |
| 292 | pref < nmp->nm_rsize && pref >= NFS_FABLKSIZE) |
| 293 | nmp->nm_rsize = (pref + NFS_FABLKSIZE - 1) & |
| 294 | ~(NFS_FABLKSIZE - 1); |
| 295 | xmax = fxdr_unsigned(u_int32_t, fsp->fs_rtmax); |
| 296 | if (xmax < nmp->nm_rsize && xmax > 0) { |
| 297 | nmp->nm_rsize = xmax & ~(NFS_FABLKSIZE - 1); |
| 298 | if (nmp->nm_rsize == 0) |
| 299 | nmp->nm_rsize = xmax; |
| 300 | } |
| 301 | pref = fxdr_unsigned(u_int32_t, fsp->fs_dtpref); |
| 302 | if (pref < nmp->nm_readdirsize && pref >= NFS_DIRFRAGSIZ) |
| 303 | nmp->nm_readdirsize = (pref + NFS_DIRFRAGSIZ - 1) & |
| 304 | ~(NFS_DIRFRAGSIZ - 1); |
| 305 | if (xmax < nmp->nm_readdirsize && xmax > 0) { |
| 306 | nmp->nm_readdirsize = xmax & ~(NFS_DIRFRAGSIZ - 1); |
| 307 | if (nmp->nm_readdirsize == 0) |
| 308 | nmp->nm_readdirsize = xmax; |
| 309 | } |
| 310 | /* XXX */ |
| 311 | nmp->nm_maxfilesize = (u_int64_t)0x80000000 * DEV_BSIZE - 1; |
| 312 | maxfsize = fxdr_hyper(&fsp->fs_maxfilesize); |
| 313 | if (maxfsize > 0 && maxfsize < nmp->nm_maxfilesize) |
| 314 | nmp->nm_maxfilesize = maxfsize; |
| 315 | nmp->nm_mountp->mnt_fs_bshift = |
| 316 | ffs(MIN(nmp->nm_rsize, nmp->nm_wsize)) - 1; |
| 317 | nmp->nm_iflag |= NFSMNT_GOTFSINFO; |
| 318 | } |
| 319 | nfsm_reqdone; |
| 320 | return (error); |
| 321 | } |
| 322 | #endif |
| 323 | |
| 324 | /* |
| 325 | * Mount a remote root fs via. NFS. It goes like this: |
| 326 | * - Call nfs_boot_init() to fill in the nfs_diskless struct |
| 327 | * - build the rootfs mount point and call mountnfs() to do the rest. |
| 328 | */ |
| 329 | int |
| 330 | nfs_mountroot(void) |
| 331 | { |
| 332 | struct timespec ts; |
| 333 | struct nfs_diskless *nd; |
| 334 | struct vattr attr; |
| 335 | struct mount *mp; |
| 336 | struct vnode *vp; |
| 337 | struct lwp *l; |
| 338 | long n; |
| 339 | int error; |
| 340 | |
| 341 | l = curlwp; /* XXX */ |
| 342 | |
| 343 | if (device_class(root_device) != DV_IFNET) |
| 344 | return (ENODEV); |
| 345 | |
| 346 | /* |
| 347 | * XXX time must be non-zero when we init the interface or else |
| 348 | * the arp code will wedge. [Fixed now in if_ether.c] |
| 349 | * However, the NFS attribute cache gives false "hits" when the |
| 350 | * current time < nfs_attrtimeo(nmp, np) so keep this in for now. |
| 351 | */ |
| 352 | if (time_second < NFS_MAXATTRTIMO) { |
| 353 | ts.tv_sec = NFS_MAXATTRTIMO; |
| 354 | ts.tv_nsec = 0; |
| 355 | tc_setclock(&ts); |
| 356 | } |
| 357 | |
| 358 | /* |
| 359 | * Call nfs_boot_init() to fill in the nfs_diskless struct. |
| 360 | * Side effect: Finds and configures a network interface. |
| 361 | */ |
| 362 | nd = kmem_zalloc(sizeof(*nd), KM_SLEEP); |
| 363 | error = nfs_boot_init(nd, l); |
| 364 | if (error) { |
| 365 | kmem_free(nd, sizeof(*nd)); |
| 366 | return (error); |
| 367 | } |
| 368 | |
| 369 | /* |
| 370 | * Create the root mount point. |
| 371 | */ |
| 372 | error = nfs_mount_diskless(&nd->nd_root, "/" , &mp, &vp, l); |
| 373 | if (error) |
| 374 | goto out; |
| 375 | printf("root on %s\n" , nd->nd_root.ndm_host); |
| 376 | |
| 377 | /* |
| 378 | * Link it into the mount list. |
| 379 | */ |
| 380 | mountlist_append(mp); |
| 381 | rootvp = vp; |
| 382 | mp->mnt_vnodecovered = NULLVP; |
| 383 | vfs_unbusy(mp, false, NULL); |
| 384 | |
| 385 | /* Get root attributes (for the time). */ |
| 386 | vn_lock(vp, LK_SHARED | LK_RETRY); |
| 387 | error = VOP_GETATTR(vp, &attr, l->l_cred); |
| 388 | VOP_UNLOCK(vp); |
| 389 | if (error) |
| 390 | panic("nfs_mountroot: getattr for root" ); |
| 391 | n = attr.va_atime.tv_sec; |
| 392 | #ifdef DEBUG |
| 393 | printf("root time: 0x%lx\n" , n); |
| 394 | #endif |
| 395 | setrootfstime(n); |
| 396 | |
| 397 | out: |
| 398 | if (error) |
| 399 | nfs_boot_cleanup(nd, l); |
| 400 | kmem_free(nd, sizeof(*nd)); |
| 401 | return (error); |
| 402 | } |
| 403 | |
| 404 | /* |
| 405 | * Internal version of mount system call for diskless setup. |
| 406 | * Separate function because we used to call it twice. |
| 407 | * (once for root and once for swap) |
| 408 | */ |
| 409 | static int |
| 410 | nfs_mount_diskless(struct nfs_dlmount *ndmntp, const char *mntname, struct mount **mpp, struct vnode **vpp, struct lwp *l) |
| 411 | /* mntname: mount point name */ |
| 412 | { |
| 413 | struct mount *mp; |
| 414 | struct mbuf *m; |
| 415 | int error; |
| 416 | |
| 417 | vfs_rootmountalloc(MOUNT_NFS, mntname, &mp); |
| 418 | |
| 419 | mp->mnt_op = &nfs_vfsops; |
| 420 | |
| 421 | /* |
| 422 | * Historical practice expects NFS root file systems to |
| 423 | * be initially mounted r/w. |
| 424 | */ |
| 425 | mp->mnt_flag &= ~MNT_RDONLY; |
| 426 | |
| 427 | /* Get mbuf for server sockaddr. */ |
| 428 | m = m_get(M_WAIT, MT_SONAME); |
| 429 | if (m == NULL) |
| 430 | panic("nfs_mountroot: mget soname for %s" , mntname); |
| 431 | MCLAIM(m, &nfs_mowner); |
| 432 | memcpy(mtod(m, void *), (void *)ndmntp->ndm_args.addr, |
| 433 | (m->m_len = ndmntp->ndm_args.addr->sa_len)); |
| 434 | |
| 435 | error = mountnfs(&ndmntp->ndm_args, mp, m, mntname, |
| 436 | ndmntp->ndm_args.hostname, vpp, l); |
| 437 | if (error) { |
| 438 | vfs_unbusy(mp, false, NULL); |
| 439 | vfs_destroy(mp); |
| 440 | printf("nfs_mountroot: mount %s failed: %d\n" , |
| 441 | mntname, error); |
| 442 | } else |
| 443 | *mpp = mp; |
| 444 | |
| 445 | return (error); |
| 446 | } |
| 447 | |
| 448 | void |
| 449 | nfs_decode_args(struct nfsmount *nmp, struct nfs_args *argp, struct lwp *l) |
| 450 | { |
| 451 | int s; |
| 452 | int adjsock; |
| 453 | int maxio; |
| 454 | |
| 455 | s = splsoftnet(); |
| 456 | |
| 457 | /* |
| 458 | * Silently clear NFSMNT_NOCONN if it's a TCP mount, it makes |
| 459 | * no sense in that context. |
| 460 | */ |
| 461 | if (argp->sotype == SOCK_STREAM) |
| 462 | argp->flags &= ~NFSMNT_NOCONN; |
| 463 | |
| 464 | /* |
| 465 | * Cookie translation is not needed for v2, silently ignore it. |
| 466 | */ |
| 467 | if ((argp->flags & (NFSMNT_XLATECOOKIE|NFSMNT_NFSV3)) == |
| 468 | NFSMNT_XLATECOOKIE) |
| 469 | argp->flags &= ~NFSMNT_XLATECOOKIE; |
| 470 | |
| 471 | /* Re-bind if rsrvd port requested and wasn't on one */ |
| 472 | adjsock = !(nmp->nm_flag & NFSMNT_RESVPORT) |
| 473 | && (argp->flags & NFSMNT_RESVPORT); |
| 474 | /* Also re-bind if we're switching to/from a connected UDP socket */ |
| 475 | adjsock |= ((nmp->nm_flag & NFSMNT_NOCONN) != |
| 476 | (argp->flags & NFSMNT_NOCONN)); |
| 477 | |
| 478 | /* Update flags. */ |
| 479 | nmp->nm_flag = argp->flags; |
| 480 | splx(s); |
| 481 | |
| 482 | if ((argp->flags & NFSMNT_TIMEO) && argp->timeo > 0) { |
| 483 | nmp->nm_timeo = (argp->timeo * NFS_HZ + 5) / 10; |
| 484 | if (nmp->nm_timeo < NFS_MINTIMEO) |
| 485 | nmp->nm_timeo = NFS_MINTIMEO; |
| 486 | else if (nmp->nm_timeo > NFS_MAXTIMEO) |
| 487 | nmp->nm_timeo = NFS_MAXTIMEO; |
| 488 | } |
| 489 | |
| 490 | if ((argp->flags & NFSMNT_RETRANS) && argp->retrans > 1) { |
| 491 | nmp->nm_retry = argp->retrans; |
| 492 | if (nmp->nm_retry > NFS_MAXREXMIT) |
| 493 | nmp->nm_retry = NFS_MAXREXMIT; |
| 494 | } |
| 495 | |
| 496 | #ifndef NFS_V2_ONLY |
| 497 | if (argp->flags & NFSMNT_NFSV3) { |
| 498 | if (argp->sotype == SOCK_DGRAM) |
| 499 | maxio = NFS_MAXDGRAMDATA; |
| 500 | else |
| 501 | maxio = NFS_MAXDATA; |
| 502 | } else |
| 503 | #endif |
| 504 | maxio = NFS_V2MAXDATA; |
| 505 | |
| 506 | if ((argp->flags & NFSMNT_WSIZE) && argp->wsize > 0) { |
| 507 | int osize = nmp->nm_wsize; |
| 508 | nmp->nm_wsize = argp->wsize; |
| 509 | /* Round down to multiple of blocksize */ |
| 510 | nmp->nm_wsize &= ~(NFS_FABLKSIZE - 1); |
| 511 | if (nmp->nm_wsize <= 0) |
| 512 | nmp->nm_wsize = NFS_FABLKSIZE; |
| 513 | adjsock |= (nmp->nm_wsize != osize); |
| 514 | } |
| 515 | if (nmp->nm_wsize > maxio) |
| 516 | nmp->nm_wsize = maxio; |
| 517 | if (nmp->nm_wsize > MAXBSIZE) |
| 518 | nmp->nm_wsize = MAXBSIZE; |
| 519 | |
| 520 | if ((argp->flags & NFSMNT_RSIZE) && argp->rsize > 0) { |
| 521 | int osize = nmp->nm_rsize; |
| 522 | nmp->nm_rsize = argp->rsize; |
| 523 | /* Round down to multiple of blocksize */ |
| 524 | nmp->nm_rsize &= ~(NFS_FABLKSIZE - 1); |
| 525 | if (nmp->nm_rsize <= 0) |
| 526 | nmp->nm_rsize = NFS_FABLKSIZE; |
| 527 | adjsock |= (nmp->nm_rsize != osize); |
| 528 | } |
| 529 | if (nmp->nm_rsize > maxio) |
| 530 | nmp->nm_rsize = maxio; |
| 531 | if (nmp->nm_rsize > MAXBSIZE) |
| 532 | nmp->nm_rsize = MAXBSIZE; |
| 533 | |
| 534 | if ((argp->flags & NFSMNT_READDIRSIZE) && argp->readdirsize > 0) { |
| 535 | nmp->nm_readdirsize = argp->readdirsize; |
| 536 | /* Round down to multiple of minimum blocksize */ |
| 537 | nmp->nm_readdirsize &= ~(NFS_DIRFRAGSIZ - 1); |
| 538 | if (nmp->nm_readdirsize < NFS_DIRFRAGSIZ) |
| 539 | nmp->nm_readdirsize = NFS_DIRFRAGSIZ; |
| 540 | /* Bigger than buffer size makes no sense */ |
| 541 | if (nmp->nm_readdirsize > NFS_DIRBLKSIZ) |
| 542 | nmp->nm_readdirsize = NFS_DIRBLKSIZ; |
| 543 | } else if (argp->flags & NFSMNT_RSIZE) |
| 544 | nmp->nm_readdirsize = nmp->nm_rsize; |
| 545 | |
| 546 | if (nmp->nm_readdirsize > maxio) |
| 547 | nmp->nm_readdirsize = maxio; |
| 548 | |
| 549 | if ((argp->flags & NFSMNT_MAXGRPS) && argp->maxgrouplist >= 0 && |
| 550 | argp->maxgrouplist <= NFS_MAXGRPS) |
| 551 | nmp->nm_numgrps = argp->maxgrouplist; |
| 552 | if ((argp->flags & NFSMNT_READAHEAD) && argp->readahead >= 0 && |
| 553 | argp->readahead <= NFS_MAXRAHEAD) |
| 554 | nmp->nm_readahead = argp->readahead; |
| 555 | if ((argp->flags & NFSMNT_DEADTHRESH) && argp->deadthresh >= 1 && |
| 556 | argp->deadthresh <= NFS_NEVERDEAD) |
| 557 | nmp->nm_deadthresh = argp->deadthresh; |
| 558 | |
| 559 | adjsock |= ((nmp->nm_sotype != argp->sotype) || |
| 560 | (nmp->nm_soproto != argp->proto)); |
| 561 | nmp->nm_sotype = argp->sotype; |
| 562 | nmp->nm_soproto = argp->proto; |
| 563 | |
| 564 | if (nmp->nm_so && adjsock) { |
| 565 | nfs_safedisconnect(nmp); |
| 566 | if (nmp->nm_sotype == SOCK_DGRAM) |
| 567 | while (nfs_connect(nmp, (struct nfsreq *)0, l)) { |
| 568 | printf("nfs_args: retrying connect\n" ); |
| 569 | kpause("nfscn3" , false, hz, NULL); |
| 570 | } |
| 571 | } |
| 572 | } |
| 573 | |
| 574 | /* |
| 575 | * VFS Operations. |
| 576 | * |
| 577 | * mount system call |
| 578 | * It seems a bit dumb to copyinstr() the host and path here and then |
| 579 | * memcpy() them in mountnfs(), but I wanted to detect errors before |
| 580 | * doing the sockargs() call because sockargs() allocates an mbuf and |
| 581 | * an error after that means that I have to release the mbuf. |
| 582 | */ |
| 583 | /* ARGSUSED */ |
| 584 | int |
| 585 | nfs_mount(struct mount *mp, const char *path, void *data, size_t *data_len) |
| 586 | { |
| 587 | struct lwp *l = curlwp; |
| 588 | int error; |
| 589 | struct nfs_args *args = data; |
| 590 | struct mbuf *nam; |
| 591 | struct nfsmount *nmp = VFSTONFS(mp); |
| 592 | struct sockaddr *sa; |
| 593 | struct vnode *vp; |
| 594 | char *pth, *hst; |
| 595 | size_t len; |
| 596 | u_char *nfh; |
| 597 | |
| 598 | if (args == NULL) |
| 599 | return EINVAL; |
| 600 | if (*data_len < sizeof *args) |
| 601 | return EINVAL; |
| 602 | |
| 603 | if (mp->mnt_flag & MNT_GETARGS) { |
| 604 | |
| 605 | if (nmp == NULL) |
| 606 | return (EIO); |
| 607 | if (args->addr != NULL) { |
| 608 | sa = mtod(nmp->nm_nam, struct sockaddr *); |
| 609 | error = copyout(sa, args->addr, sa->sa_len); |
| 610 | if (error) |
| 611 | return (error); |
| 612 | args->addrlen = sa->sa_len; |
| 613 | } else |
| 614 | args->addrlen = 0; |
| 615 | |
| 616 | args->version = NFS_ARGSVERSION; |
| 617 | args->sotype = nmp->nm_sotype; |
| 618 | args->proto = nmp->nm_soproto; |
| 619 | args->fh = NULL; |
| 620 | args->fhsize = 0; |
| 621 | args->flags = nmp->nm_flag; |
| 622 | args->wsize = nmp->nm_wsize; |
| 623 | args->rsize = nmp->nm_rsize; |
| 624 | args->readdirsize = nmp->nm_readdirsize; |
| 625 | args->timeo = nmp->nm_timeo; |
| 626 | args->retrans = nmp->nm_retry; |
| 627 | args->maxgrouplist = nmp->nm_numgrps; |
| 628 | args->readahead = nmp->nm_readahead; |
| 629 | args->leaseterm = 0; /* dummy */ |
| 630 | args->deadthresh = nmp->nm_deadthresh; |
| 631 | args->hostname = NULL; |
| 632 | *data_len = sizeof *args; |
| 633 | return 0; |
| 634 | } |
| 635 | |
| 636 | if (args->version != NFS_ARGSVERSION) |
| 637 | return (EPROGMISMATCH); |
| 638 | if (args->flags & (NFSMNT_NQNFS|NFSMNT_KERB)) |
| 639 | return (EPROGUNAVAIL); |
| 640 | #ifdef NFS_V2_ONLY |
| 641 | if (args->flags & NFSMNT_NFSV3) |
| 642 | return (EPROGMISMATCH); |
| 643 | #endif |
| 644 | if (mp->mnt_flag & MNT_UPDATE) { |
| 645 | if (nmp == NULL) |
| 646 | return (EIO); |
| 647 | /* |
| 648 | * When doing an update, we can't change from or to |
| 649 | * v3, or change cookie translation |
| 650 | */ |
| 651 | args->flags = (args->flags & ~(NFSMNT_NFSV3|NFSMNT_XLATECOOKIE)) | |
| 652 | (nmp->nm_flag & (NFSMNT_NFSV3|NFSMNT_XLATECOOKIE)); |
| 653 | nfs_decode_args(nmp, args, l); |
| 654 | return (0); |
| 655 | } |
| 656 | if (args->fhsize < 0 || args->fhsize > NFSX_V3FHMAX) |
| 657 | return (EINVAL); |
| 658 | nfh = malloc(NFSX_V3FHMAX, M_TEMP, M_WAITOK); |
| 659 | error = copyin(args->fh, nfh, args->fhsize); |
| 660 | if (error) |
| 661 | goto free_nfh; |
| 662 | pth = malloc(MNAMELEN, M_TEMP, M_WAITOK); |
| 663 | error = copyinstr(path, pth, MNAMELEN - 1, &len); |
| 664 | if (error) |
| 665 | goto free_pth; |
| 666 | memset(&pth[len], 0, MNAMELEN - len); |
| 667 | hst = malloc(MNAMELEN, M_TEMP, M_WAITOK); |
| 668 | error = copyinstr(args->hostname, hst, MNAMELEN - 1, &len); |
| 669 | if (error) |
| 670 | goto free_hst; |
| 671 | memset(&hst[len], 0, MNAMELEN - len); |
| 672 | /* sockargs() call must be after above copyin() calls */ |
| 673 | error = sockargs(&nam, args->addr, args->addrlen, MT_SONAME); |
| 674 | if (error) |
| 675 | goto free_hst; |
| 676 | MCLAIM(nam, &nfs_mowner); |
| 677 | args->fh = nfh; |
| 678 | error = mountnfs(args, mp, nam, pth, hst, &vp, l); |
| 679 | |
| 680 | free_hst: |
| 681 | free(hst, M_TEMP); |
| 682 | free_pth: |
| 683 | free(pth, M_TEMP); |
| 684 | free_nfh: |
| 685 | free(nfh, M_TEMP); |
| 686 | |
| 687 | return (error); |
| 688 | } |
| 689 | |
| 690 | /* |
| 691 | * Common code for mount and mountroot |
| 692 | */ |
| 693 | int |
| 694 | mountnfs(struct nfs_args *argp, struct mount *mp, struct mbuf *nam, const char *pth, const char *hst, struct vnode **vpp, struct lwp *l) |
| 695 | { |
| 696 | struct nfsmount *nmp; |
| 697 | struct nfsnode *np; |
| 698 | struct vnode *vp; |
| 699 | int error; |
| 700 | struct vattr *attrs; |
| 701 | kauth_cred_t cr; |
| 702 | char iosname[IOSTATNAMELEN]; |
| 703 | |
| 704 | /* |
| 705 | * If the number of nfs iothreads to use has never |
| 706 | * been set, create a reasonable number of them. |
| 707 | */ |
| 708 | |
| 709 | if (nfs_niothreads < 0) { |
| 710 | nfs_set_niothreads(NFS_DEFAULT_NIOTHREADS); |
| 711 | } |
| 712 | |
| 713 | if (mp->mnt_flag & MNT_UPDATE) { |
| 714 | nmp = VFSTONFS(mp); |
| 715 | /* update paths, file handles, etc, here XXX */ |
| 716 | m_freem(nam); |
| 717 | return (0); |
| 718 | } else { |
| 719 | nmp = kmem_zalloc(sizeof(*nmp), KM_SLEEP); |
| 720 | mp->mnt_data = nmp; |
| 721 | TAILQ_INIT(&nmp->nm_uidlruhead); |
| 722 | TAILQ_INIT(&nmp->nm_bufq); |
| 723 | rw_init(&nmp->nm_writeverflock); |
| 724 | mutex_init(&nmp->nm_lock, MUTEX_DEFAULT, IPL_NONE); |
| 725 | cv_init(&nmp->nm_rcvcv, "nfsrcv" ); |
| 726 | cv_init(&nmp->nm_sndcv, "nfssnd" ); |
| 727 | cv_init(&nmp->nm_aiocv, "nfsaio" ); |
| 728 | cv_init(&nmp->nm_disconcv, "nfsdis" ); |
| 729 | } |
| 730 | vfs_getnewfsid(mp); |
| 731 | nmp->nm_mountp = mp; |
| 732 | |
| 733 | #ifndef NFS_V2_ONLY |
| 734 | if ((argp->flags & NFSMNT_NFSV3) == 0) |
| 735 | #endif |
| 736 | { |
| 737 | if (argp->fhsize != NFSX_V2FH) { |
| 738 | return EINVAL; |
| 739 | } |
| 740 | } |
| 741 | |
| 742 | /* |
| 743 | * V2 can only handle 32 bit filesizes. For v3, nfs_fsinfo |
| 744 | * will overwrite this. |
| 745 | */ |
| 746 | nmp->nm_maxfilesize = 0xffffffffLL; |
| 747 | |
| 748 | nmp->nm_timeo = NFS_TIMEO; |
| 749 | nmp->nm_retry = NFS_RETRANS; |
| 750 | nmp->nm_wsize = NFS_WSIZE; |
| 751 | nmp->nm_rsize = NFS_RSIZE; |
| 752 | nmp->nm_readdirsize = NFS_READDIRSIZE; |
| 753 | nmp->nm_numgrps = NFS_MAXGRPS; |
| 754 | nmp->nm_readahead = NFS_DEFRAHEAD; |
| 755 | nmp->nm_deadthresh = NFS_DEFDEADTHRESH; |
| 756 | error = set_statvfs_info(pth, UIO_SYSSPACE, hst, UIO_SYSSPACE, |
| 757 | mp->mnt_op->vfs_name, mp, l); |
| 758 | if (error) |
| 759 | goto bad; |
| 760 | nmp->nm_nam = nam; |
| 761 | |
| 762 | /* Set up the sockets and per-host congestion */ |
| 763 | nmp->nm_sotype = argp->sotype; |
| 764 | nmp->nm_soproto = argp->proto; |
| 765 | |
| 766 | nfs_decode_args(nmp, argp, l); |
| 767 | |
| 768 | mp->mnt_fs_bshift = ffs(MIN(nmp->nm_rsize, nmp->nm_wsize)) - 1; |
| 769 | mp->mnt_dev_bshift = DEV_BSHIFT; |
| 770 | |
| 771 | /* |
| 772 | * For Connection based sockets (TCP,...) defer the connect until |
| 773 | * the first request, in case the server is not responding. |
| 774 | */ |
| 775 | if (nmp->nm_sotype == SOCK_DGRAM && |
| 776 | (error = nfs_connect(nmp, (struct nfsreq *)0, l))) |
| 777 | goto bad; |
| 778 | |
| 779 | /* |
| 780 | * This is silly, but it has to be set so that vinifod() works. |
| 781 | * We do not want to do an nfs_statvfs() here since we can get |
| 782 | * stuck on a dead server and we are holding a lock on the mount |
| 783 | * point. |
| 784 | */ |
| 785 | mp->mnt_stat.f_iosize = NFS_MAXDGRAMDATA; |
| 786 | error = nfs_nget(mp, (nfsfh_t *)argp->fh, argp->fhsize, &np); |
| 787 | if (error) |
| 788 | goto bad; |
| 789 | vp = NFSTOV(np); |
| 790 | attrs = malloc(sizeof(struct vattr), M_TEMP, M_WAITOK); |
| 791 | VOP_GETATTR(vp, attrs, l->l_cred); |
| 792 | if ((nmp->nm_flag & NFSMNT_NFSV3) && (vp->v_type == VDIR)) { |
| 793 | cr = kauth_cred_alloc(); |
| 794 | kauth_cred_setuid(cr, attrs->va_uid); |
| 795 | kauth_cred_seteuid(cr, attrs->va_uid); |
| 796 | kauth_cred_setsvuid(cr, attrs->va_uid); |
| 797 | kauth_cred_setgid(cr, attrs->va_gid); |
| 798 | kauth_cred_setegid(cr, attrs->va_gid); |
| 799 | kauth_cred_setsvgid(cr, attrs->va_gid); |
| 800 | nfs_cookieheuristic(vp, &nmp->nm_iflag, l, cr); |
| 801 | kauth_cred_free(cr); |
| 802 | } |
| 803 | free(attrs, M_TEMP); |
| 804 | |
| 805 | /* |
| 806 | * A reference count is needed on the nfsnode representing the |
| 807 | * remote root. If this object is not persistent, then backward |
| 808 | * traversals of the mount point (i.e. "..") will not work if |
| 809 | * the nfsnode gets flushed out of the cache. Ufs does not have |
| 810 | * this problem, because one can identify root inodes by their |
| 811 | * number == UFS_ROOTINO (2). So, just unlock, but no rele. |
| 812 | */ |
| 813 | |
| 814 | nmp->nm_vnode = vp; |
| 815 | if (vp->v_type == VNON) |
| 816 | vp->v_type = VDIR; |
| 817 | vp->v_vflag |= VV_ROOT; |
| 818 | VOP_UNLOCK(vp); |
| 819 | *vpp = vp; |
| 820 | |
| 821 | snprintf(iosname, sizeof(iosname), "nfs%u" , nfs_mount_count++); |
| 822 | nmp->nm_stats = iostat_alloc(IOSTAT_NFS, nmp, iosname); |
| 823 | |
| 824 | return (0); |
| 825 | bad: |
| 826 | nfs_disconnect(nmp); |
| 827 | rw_destroy(&nmp->nm_writeverflock); |
| 828 | mutex_destroy(&nmp->nm_lock); |
| 829 | cv_destroy(&nmp->nm_rcvcv); |
| 830 | cv_destroy(&nmp->nm_sndcv); |
| 831 | cv_destroy(&nmp->nm_aiocv); |
| 832 | cv_destroy(&nmp->nm_disconcv); |
| 833 | kmem_free(nmp, sizeof(*nmp)); |
| 834 | m_freem(nam); |
| 835 | return (error); |
| 836 | } |
| 837 | |
| 838 | /* |
| 839 | * unmount system call |
| 840 | */ |
| 841 | int |
| 842 | nfs_unmount(struct mount *mp, int mntflags) |
| 843 | { |
| 844 | struct nfsmount *nmp = VFSTONFS(mp); |
| 845 | struct vnode *vp; |
| 846 | int error, flags = 0; |
| 847 | |
| 848 | if (mntflags & MNT_FORCE) { |
| 849 | mutex_enter(&nmp->nm_lock); |
| 850 | flags |= FORCECLOSE; |
| 851 | nmp->nm_iflag |= NFSMNT_DISMNTFORCE; |
| 852 | mutex_exit(&nmp->nm_lock); |
| 853 | |
| 854 | } |
| 855 | |
| 856 | /* |
| 857 | * Goes something like this.. |
| 858 | * - Check for activity on the root vnode (other than ourselves). |
| 859 | * - Call vflush() to clear out vnodes for this file system, |
| 860 | * except for the root vnode. |
| 861 | * - Decrement reference on the vnode representing remote root. |
| 862 | * - Close the socket |
| 863 | * - Free up the data structures |
| 864 | */ |
| 865 | /* |
| 866 | * We need to decrement the ref. count on the nfsnode representing |
| 867 | * the remote root. See comment in mountnfs(). |
| 868 | */ |
| 869 | vp = nmp->nm_vnode; |
| 870 | error = vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); |
| 871 | if (error != 0) |
| 872 | goto err; |
| 873 | |
| 874 | if ((mntflags & MNT_FORCE) == 0 && vp->v_usecount > 1) { |
| 875 | VOP_UNLOCK(vp); |
| 876 | error = EBUSY; |
| 877 | goto err; |
| 878 | } |
| 879 | |
| 880 | error = vflush(mp, vp, flags); |
| 881 | if (error) { |
| 882 | VOP_UNLOCK(vp); |
| 883 | goto err; |
| 884 | } |
| 885 | |
| 886 | /* |
| 887 | * We are now committed to the unmount; mark the mount structure |
| 888 | * as doomed so that any sleepers kicked awake by nfs_disconnect |
| 889 | * will go away cleanly. |
| 890 | */ |
| 891 | nmp->nm_iflag |= NFSMNT_DISMNT; |
| 892 | |
| 893 | /* |
| 894 | * No new async I/O will be added, but await for pending |
| 895 | * ones to drain. |
| 896 | */ |
| 897 | while (nfs_iodbusy(nmp)) |
| 898 | kpause("nfsumnt" , false, hz, NULL); |
| 899 | |
| 900 | /* |
| 901 | * Clean up the stats... note that we carefully avoid decrementing |
| 902 | * nfs_mount_count here for good reason - we may not be unmounting |
| 903 | * the last thing mounted. |
| 904 | */ |
| 905 | iostat_free(nmp->nm_stats); |
| 906 | |
| 907 | /* |
| 908 | * There is one reference count to get rid of here |
| 909 | * (see comment in mountnfs()). |
| 910 | */ |
| 911 | VOP_UNLOCK(vp); |
| 912 | vgone(vp); |
| 913 | nfs_disconnect(nmp); |
| 914 | m_freem(nmp->nm_nam); |
| 915 | |
| 916 | rw_destroy(&nmp->nm_writeverflock); |
| 917 | mutex_destroy(&nmp->nm_lock); |
| 918 | cv_destroy(&nmp->nm_rcvcv); |
| 919 | cv_destroy(&nmp->nm_sndcv); |
| 920 | cv_destroy(&nmp->nm_aiocv); |
| 921 | cv_destroy(&nmp->nm_disconcv); |
| 922 | kmem_free(nmp, sizeof(*nmp)); |
| 923 | return (0); |
| 924 | |
| 925 | err: |
| 926 | if (mntflags & MNT_FORCE) { |
| 927 | mutex_enter(&nmp->nm_lock); |
| 928 | nmp->nm_iflag &= ~NFSMNT_DISMNTFORCE; |
| 929 | mutex_exit(&nmp->nm_lock); |
| 930 | } |
| 931 | |
| 932 | return error; |
| 933 | } |
| 934 | |
| 935 | /* |
| 936 | * Return root of a filesystem |
| 937 | */ |
| 938 | int |
| 939 | nfs_root(struct mount *mp, struct vnode **vpp) |
| 940 | { |
| 941 | struct vnode *vp; |
| 942 | struct nfsmount *nmp; |
| 943 | int error; |
| 944 | |
| 945 | nmp = VFSTONFS(mp); |
| 946 | vp = nmp->nm_vnode; |
| 947 | vref(vp); |
| 948 | error = vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); |
| 949 | if (error != 0) { |
| 950 | vrele(vp); |
| 951 | return error; |
| 952 | } |
| 953 | *vpp = vp; |
| 954 | return (0); |
| 955 | } |
| 956 | |
| 957 | extern int syncprt; |
| 958 | |
| 959 | static bool |
| 960 | nfs_sync_selector(void *cl, struct vnode *vp) |
| 961 | { |
| 962 | |
| 963 | return !LIST_EMPTY(&vp->v_dirtyblkhd) || !UVM_OBJ_IS_CLEAN(&vp->v_uobj); |
| 964 | } |
| 965 | |
| 966 | /* |
| 967 | * Flush out the buffer cache |
| 968 | */ |
| 969 | /* ARGSUSED */ |
| 970 | int |
| 971 | nfs_sync(struct mount *mp, int waitfor, kauth_cred_t cred) |
| 972 | { |
| 973 | struct vnode *vp; |
| 974 | struct vnode_iterator *marker; |
| 975 | int error, allerror = 0; |
| 976 | |
| 977 | /* |
| 978 | * Force stale buffer cache information to be flushed. |
| 979 | */ |
| 980 | vfs_vnode_iterator_init(mp, &marker); |
| 981 | while ((vp = vfs_vnode_iterator_next(marker, nfs_sync_selector, |
| 982 | NULL))) |
| 983 | { |
| 984 | error = vn_lock(vp, LK_EXCLUSIVE); |
| 985 | if (error) { |
| 986 | vrele(vp); |
| 987 | continue; |
| 988 | } |
| 989 | error = VOP_FSYNC(vp, cred, |
| 990 | waitfor == MNT_WAIT ? FSYNC_WAIT : 0, 0, 0); |
| 991 | if (error) |
| 992 | allerror = error; |
| 993 | vput(vp); |
| 994 | } |
| 995 | vfs_vnode_iterator_destroy(marker); |
| 996 | return allerror; |
| 997 | } |
| 998 | |
| 999 | /* |
| 1000 | * NFS flat namespace lookup. |
| 1001 | * Currently unsupported. |
| 1002 | */ |
| 1003 | /* ARGSUSED */ |
| 1004 | int |
| 1005 | nfs_vget(struct mount *mp, ino_t ino, struct vnode **vpp) |
| 1006 | { |
| 1007 | |
| 1008 | return (EOPNOTSUPP); |
| 1009 | } |
| 1010 | |
| 1011 | /* |
| 1012 | * Do that sysctl thang... |
| 1013 | */ |
| 1014 | static int |
| 1015 | sysctl_vfs_nfs_iothreads(SYSCTLFN_ARGS) |
| 1016 | { |
| 1017 | struct sysctlnode node; |
| 1018 | int val; |
| 1019 | int error; |
| 1020 | |
| 1021 | val = nfs_niothreads; |
| 1022 | node = *rnode; |
| 1023 | node.sysctl_data = &val; |
| 1024 | error = sysctl_lookup(SYSCTLFN_CALL(&node)); |
| 1025 | if (error || newp == NULL) |
| 1026 | return error; |
| 1027 | |
| 1028 | return nfs_set_niothreads(val); |
| 1029 | } |
| 1030 | |
| 1031 | static void |
| 1032 | nfs_sysctl_init(void) |
| 1033 | { |
| 1034 | |
| 1035 | sysctl_createv(&nfs_clog, 0, NULL, NULL, |
| 1036 | CTLFLAG_PERMANENT, |
| 1037 | CTLTYPE_NODE, "nfs" , |
| 1038 | SYSCTL_DESCR("NFS vfs options" ), |
| 1039 | NULL, 0, NULL, 0, |
| 1040 | CTL_VFS, 2, CTL_EOL); |
| 1041 | /* |
| 1042 | * XXX the "2" above could be dynamic, thereby eliminating one |
| 1043 | * more instance of the "number to vfs" mapping problem, but |
| 1044 | * "2" is the order as taken from sys/mount.h |
| 1045 | */ |
| 1046 | |
| 1047 | sysctl_createv(&nfs_clog, 0, NULL, NULL, |
| 1048 | CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
| 1049 | CTLTYPE_STRUCT, "nfsstats" , |
| 1050 | SYSCTL_DESCR("NFS operation statistics" ), |
| 1051 | NULL, 0, &nfsstats, sizeof(nfsstats), |
| 1052 | CTL_VFS, 2, NFS_NFSSTATS, CTL_EOL); |
| 1053 | sysctl_createv(&nfs_clog, 0, NULL, NULL, |
| 1054 | CTLFLAG_PERMANENT|CTLFLAG_READWRITE, |
| 1055 | CTLTYPE_INT, "iothreads" , |
| 1056 | SYSCTL_DESCR("Number of NFS client processes desired" ), |
| 1057 | sysctl_vfs_nfs_iothreads, 0, NULL, 0, |
| 1058 | CTL_VFS, 2, NFS_IOTHREADS, CTL_EOL); |
| 1059 | } |
| 1060 | |
| 1061 | static void |
| 1062 | nfs_sysctl_fini(void) |
| 1063 | { |
| 1064 | |
| 1065 | sysctl_teardown(&nfs_clog); |
| 1066 | } |
| 1067 | |
| 1068 | /* ARGSUSED */ |
| 1069 | int |
| 1070 | nfs_fhtovp(struct mount *mp, struct fid *fid, struct vnode **vpp) |
| 1071 | { |
| 1072 | size_t fidsize; |
| 1073 | size_t fhsize; |
| 1074 | struct nfsnode *np; |
| 1075 | int error; |
| 1076 | struct vattr va; |
| 1077 | |
| 1078 | fidsize = fid->fid_len; |
| 1079 | if (fidsize < sizeof(*fid)) { |
| 1080 | return EINVAL; |
| 1081 | } |
| 1082 | fhsize = fidsize - sizeof(*fid); |
| 1083 | if ((fhsize % NFSX_UNSIGNED) != 0) { |
| 1084 | return EINVAL; |
| 1085 | } |
| 1086 | if ((VFSTONFS(mp)->nm_flag & NFSMNT_NFSV3) != 0) { |
| 1087 | if (fhsize > NFSX_V3FHMAX || fhsize == 0) { |
| 1088 | return EINVAL; |
| 1089 | } |
| 1090 | } else { |
| 1091 | if (fhsize != NFSX_V2FH) { |
| 1092 | return EINVAL; |
| 1093 | } |
| 1094 | } |
| 1095 | error = nfs_nget(mp, (void *)fid->fid_data, fhsize, &np); |
| 1096 | if (error) { |
| 1097 | return error; |
| 1098 | } |
| 1099 | *vpp = NFSTOV(np); |
| 1100 | error = VOP_GETATTR(*vpp, &va, kauth_cred_get()); |
| 1101 | if (error != 0) { |
| 1102 | vput(*vpp); |
| 1103 | *vpp = NULLVP; |
| 1104 | } |
| 1105 | return error; |
| 1106 | } |
| 1107 | |
| 1108 | /* ARGSUSED */ |
| 1109 | int |
| 1110 | nfs_vptofh(struct vnode *vp, struct fid *buf, size_t *bufsize) |
| 1111 | { |
| 1112 | struct nfsnode *np; |
| 1113 | struct fid *fid; |
| 1114 | size_t fidsize; |
| 1115 | int error = 0; |
| 1116 | |
| 1117 | np = VTONFS(vp); |
| 1118 | fidsize = sizeof(*fid) + np->n_fhsize; |
| 1119 | if (*bufsize < fidsize) { |
| 1120 | error = E2BIG; |
| 1121 | } |
| 1122 | *bufsize = fidsize; |
| 1123 | if (error == 0) { |
| 1124 | struct fid fid_store; |
| 1125 | |
| 1126 | fid = &fid_store; |
| 1127 | memset(fid, 0, sizeof(*fid)); |
| 1128 | fid->fid_len = fidsize; |
| 1129 | memcpy(buf, fid, sizeof(*fid)); |
| 1130 | memcpy(buf->fid_data, np->n_fhp, np->n_fhsize); |
| 1131 | } |
| 1132 | return error; |
| 1133 | } |
| 1134 | |
| 1135 | /* |
| 1136 | * Vfs start routine, a no-op. |
| 1137 | */ |
| 1138 | /* ARGSUSED */ |
| 1139 | int |
| 1140 | nfs_start(struct mount *mp, int flags) |
| 1141 | { |
| 1142 | |
| 1143 | return (0); |
| 1144 | } |
| 1145 | |
| 1146 | /* |
| 1147 | * Called once at VFS init to initialize client-specific data structures. |
| 1148 | */ |
| 1149 | void |
| 1150 | nfs_vfs_init(void) |
| 1151 | { |
| 1152 | |
| 1153 | /* Initialize NFS server / client shared data. */ |
| 1154 | nfs_init(); |
| 1155 | nfs_node_init(); |
| 1156 | |
| 1157 | /* Initialize the kqueue structures */ |
| 1158 | nfs_kqinit(); |
| 1159 | /* Initialize the iod structures */ |
| 1160 | nfs_iodinit(); |
| 1161 | |
| 1162 | nfs_commitsize = uvmexp.npages << (PAGE_SHIFT - 4); |
| 1163 | } |
| 1164 | |
| 1165 | void |
| 1166 | nfs_vfs_done(void) |
| 1167 | { |
| 1168 | |
| 1169 | nfs_node_done(); |
| 1170 | nfs_kqfini(); |
| 1171 | nfs_iodfini(); |
| 1172 | nfs_fini(); |
| 1173 | } |
| 1174 | |