| 1 | /* |
| 2 | * Copyright (c) 1982, 1986, 1989, 1993 |
| 3 | * The Regents of the University of California. All rights reserved. |
| 4 | * (c) UNIX System Laboratories, Inc. |
| 5 | * All or some portions of this file are derived from material licensed |
| 6 | * to the University of California by American Telephone and Telegraph |
| 7 | * Co. or Unix System Laboratories, Inc. and are reproduced herein with |
| 8 | * the permission of UNIX System Laboratories, Inc. |
| 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. All advertising materials mentioning features or use of this software |
| 19 | * must display the following acknowledgement: |
| 20 | * This product includes software developed by the University of |
| 21 | * California, Berkeley and its contributors. |
| 22 | * 4. Neither the name of the University nor the names of its contributors |
| 23 | * may be used to endorse or promote products derived from this software |
| 24 | * without specific prior written permission. |
| 25 | * |
| 26 | * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND |
| 27 | * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
| 28 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE |
| 29 | * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE |
| 30 | * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
| 31 | * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS |
| 32 | * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) |
| 33 | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT |
| 34 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY |
| 35 | * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF |
| 36 | * SUCH DAMAGE. |
| 37 | * |
| 38 | * @(#)vfs_vnops.c 8.2 (Berkeley) 1/21/94 |
| 39 | * $FreeBSD: src/sys/kern/vfs_vnops.c,v 1.87.2.13 2002/12/29 18:19:53 dillon Exp $ |
| 40 | * $DragonFly: src/sys/kern/vfs_vnops.c,v 1.10 2003/07/26 19:42:11 rob Exp $ |
| 41 | */ |
| 42 | |
| 43 | #include <sys/param.h> |
| 44 | #include <sys/systm.h> |
| 45 | #include <sys/fcntl.h> |
| 46 | #include <sys/file.h> |
| 47 | #include <sys/stat.h> |
| 48 | #include <sys/proc.h> |
| 49 | #include <sys/mount.h> |
| 50 | #include <sys/namei.h> |
| 51 | #include <sys/vnode.h> |
| 52 | #include <sys/buf.h> |
| 53 | #include <sys/filio.h> |
| 54 | #include <sys/ttycom.h> |
| 55 | #include <sys/conf.h> |
| 56 | #include <sys/syslog.h> |
| 57 | |
| 58 | static int vn_closefile __P((struct file *fp, struct thread *td)); |
| 59 | static int vn_ioctl __P((struct file *fp, u_long com, caddr_t data, |
| 60 | struct thread *td)); |
| 61 | static int vn_read __P((struct file *fp, struct uio *uio, |
| 62 | struct ucred *cred, int flags, struct thread *td)); |
| 63 | static int vn_poll __P((struct file *fp, int events, struct ucred *cred, |
| 64 | struct thread *td)); |
| 65 | static int vn_kqfilter __P((struct file *fp, struct knote *kn)); |
| 66 | static int vn_statfile __P((struct file *fp, struct stat *sb, struct thread *td)); |
| 67 | static int vn_write __P((struct file *fp, struct uio *uio, |
| 68 | struct ucred *cred, int flags, struct thread *td)); |
| 69 | |
| 70 | struct fileops vnops = { |
| 71 | vn_read, vn_write, vn_ioctl, vn_poll, vn_kqfilter, |
| 72 | vn_statfile, vn_closefile |
| 73 | }; |
| 74 | |
| 75 | /* |
| 76 | * Common code for vnode open operations. |
| 77 | * Check permissions, and call the VOP_OPEN or VOP_CREATE routine. |
| 78 | * |
| 79 | * Note that this does NOT free nameidata for the successful case, |
| 80 | * due to the NDINIT being done elsewhere. |
| 81 | */ |
| 82 | int |
| 83 | vn_open(ndp, fmode, cmode) |
| 84 | struct nameidata *ndp; |
| 85 | int fmode, cmode; |
| 86 | { |
| 87 | struct vnode *vp; |
| 88 | struct thread *td = ndp->ni_cnd.cn_td; |
| 89 | struct ucred *cred = ndp->ni_cnd.cn_cred; |
| 90 | struct vattr vat; |
| 91 | struct vattr *vap = &vat; |
| 92 | int mode, error; |
| 93 | |
| 94 | KKASSERT(cred == td->td_proc->p_ucred); |
| 95 | |
| 96 | if (fmode & O_CREAT) { |
| 97 | ndp->ni_cnd.cn_nameiop = CREATE; |
| 98 | ndp->ni_cnd.cn_flags = LOCKPARENT | LOCKLEAF; |
| 99 | if ((fmode & O_EXCL) == 0 && (fmode & O_NOFOLLOW) == 0) |
| 100 | ndp->ni_cnd.cn_flags |= FOLLOW; |
| 101 | bwillwrite(); |
| 102 | error = namei(ndp); |
| 103 | if (error) |
| 104 | return (error); |
| 105 | if (ndp->ni_vp == NULL) { |
| 106 | VATTR_NULL(vap); |
| 107 | vap->va_type = VREG; |
| 108 | vap->va_mode = cmode; |
| 109 | if (fmode & O_EXCL) |
| 110 | vap->va_vaflags |= VA_EXCLUSIVE; |
| 111 | VOP_LEASE(ndp->ni_dvp, td, cred, LEASE_WRITE); |
| 112 | error = VOP_CREATE(ndp->ni_dvp, &ndp->ni_vp, |
| 113 | &ndp->ni_cnd, vap); |
| 114 | if (error) { |
| 115 | NDFREE(ndp, NDF_ONLY_PNBUF); |
| 116 | vput(ndp->ni_dvp); |
| 117 | return (error); |
| 118 | } |
| 119 | vput(ndp->ni_dvp); |
| 120 | ASSERT_VOP_UNLOCKED(ndp->ni_dvp, "create"); |
| 121 | ASSERT_VOP_LOCKED(ndp->ni_vp, "create"); |
| 122 | fmode &= ~O_TRUNC; |
| 123 | vp = ndp->ni_vp; |
| 124 | } else { |
| 125 | if (ndp->ni_dvp == ndp->ni_vp) |
| 126 | vrele(ndp->ni_dvp); |
| 127 | else |
| 128 | vput(ndp->ni_dvp); |
| 129 | ndp->ni_dvp = NULL; |
| 130 | vp = ndp->ni_vp; |
| 131 | if (fmode & O_EXCL) { |
| 132 | error = EEXIST; |
| 133 | goto bad; |
| 134 | } |
| 135 | fmode &= ~O_CREAT; |
| 136 | } |
| 137 | } else { |
| 138 | ndp->ni_cnd.cn_nameiop = LOOKUP; |
| 139 | ndp->ni_cnd.cn_flags = |
| 140 | ((fmode & O_NOFOLLOW) ? NOFOLLOW : FOLLOW) | LOCKLEAF; |
| 141 | error = namei(ndp); |
| 142 | if (error) |
| 143 | return (error); |
| 144 | vp = ndp->ni_vp; |
| 145 | } |
| 146 | if (vp->v_type == VLNK) { |
| 147 | error = EMLINK; |
| 148 | goto bad; |
| 149 | } |
| 150 | if (vp->v_type == VSOCK) { |
| 151 | error = EOPNOTSUPP; |
| 152 | goto bad; |
| 153 | } |
| 154 | if ((fmode & O_CREAT) == 0) { |
| 155 | mode = 0; |
| 156 | if (fmode & (FWRITE | O_TRUNC)) { |
| 157 | if (vp->v_type == VDIR) { |
| 158 | error = EISDIR; |
| 159 | goto bad; |
| 160 | } |
| 161 | error = vn_writechk(vp); |
| 162 | if (error) |
| 163 | goto bad; |
| 164 | mode |= VWRITE; |
| 165 | } |
| 166 | if (fmode & FREAD) |
| 167 | mode |= VREAD; |
| 168 | if (mode) { |
| 169 | error = VOP_ACCESS(vp, mode, cred, td); |
| 170 | if (error) |
| 171 | goto bad; |
| 172 | } |
| 173 | } |
| 174 | if (fmode & O_TRUNC) { |
| 175 | VOP_UNLOCK(vp, 0, td); /* XXX */ |
| 176 | VOP_LEASE(vp, td, cred, LEASE_WRITE); |
| 177 | vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td); /* XXX */ |
| 178 | VATTR_NULL(vap); |
| 179 | vap->va_size = 0; |
| 180 | error = VOP_SETATTR(vp, vap, cred, td); |
| 181 | if (error) |
| 182 | goto bad; |
| 183 | } |
| 184 | error = VOP_OPEN(vp, fmode, cred, td); |
| 185 | if (error) |
| 186 | goto bad; |
| 187 | /* |
| 188 | * Make sure that a VM object is created for VMIO support. |
| 189 | */ |
| 190 | if (vn_canvmio(vp) == TRUE) { |
| 191 | if ((error = vfs_object_create(vp, td)) != 0) |
| 192 | goto bad; |
| 193 | } |
| 194 | |
| 195 | if (fmode & FWRITE) |
| 196 | vp->v_writecount++; |
| 197 | return (0); |
| 198 | bad: |
| 199 | NDFREE(ndp, NDF_ONLY_PNBUF); |
| 200 | vput(vp); |
| 201 | return (error); |
| 202 | } |
| 203 | |
| 204 | /* |
| 205 | * Check for write permissions on the specified vnode. |
| 206 | * Prototype text segments cannot be written. |
| 207 | */ |
| 208 | int |
| 209 | vn_writechk(vp) |
| 210 | struct vnode *vp; |
| 211 | { |
| 212 | |
| 213 | /* |
| 214 | * If there's shared text associated with |
| 215 | * the vnode, try to free it up once. If |
| 216 | * we fail, we can't allow writing. |
| 217 | */ |
| 218 | if (vp->v_flag & VTEXT) |
| 219 | return (ETXTBSY); |
| 220 | return (0); |
| 221 | } |
| 222 | |
| 223 | /* |
| 224 | * Vnode close call |
| 225 | */ |
| 226 | int |
| 227 | vn_close(struct vnode *vp, int flags, struct thread *td) |
| 228 | { |
| 229 | int error; |
| 230 | |
| 231 | if (flags & FWRITE) |
| 232 | vp->v_writecount--; |
| 233 | error = VOP_CLOSE(vp, flags, td); |
| 234 | vrele(vp); |
| 235 | return (error); |
| 236 | } |
| 237 | |
| 238 | static __inline |
| 239 | int |
| 240 | sequential_heuristic(struct uio *uio, struct file *fp) |
| 241 | { |
| 242 | /* |
| 243 | * Sequential heuristic - detect sequential operation |
| 244 | */ |
| 245 | if ((uio->uio_offset == 0 && fp->f_seqcount > 0) || |
| 246 | uio->uio_offset == fp->f_nextoff) { |
| 247 | int tmpseq = fp->f_seqcount; |
| 248 | /* |
| 249 | * XXX we assume that the filesystem block size is |
| 250 | * the default. Not true, but still gives us a pretty |
| 251 | * good indicator of how sequential the read operations |
| 252 | * are. |
| 253 | */ |
| 254 | tmpseq += (uio->uio_resid + BKVASIZE - 1) / BKVASIZE; |
| 255 | if (tmpseq > IO_SEQMAX) |
| 256 | tmpseq = IO_SEQMAX; |
| 257 | fp->f_seqcount = tmpseq; |
| 258 | return(fp->f_seqcount << IO_SEQSHIFT); |
| 259 | } |
| 260 | |
| 261 | /* |
| 262 | * Not sequential, quick draw-down of seqcount |
| 263 | */ |
| 264 | if (fp->f_seqcount > 1) |
| 265 | fp->f_seqcount = 1; |
| 266 | else |
| 267 | fp->f_seqcount = 0; |
| 268 | return(0); |
| 269 | } |
| 270 | |
| 271 | /* |
| 272 | * Package up an I/O request on a vnode into a uio and do it. |
| 273 | */ |
| 274 | int |
| 275 | vn_rdwr(rw, vp, base, len, offset, segflg, ioflg, cred, aresid, td) |
| 276 | enum uio_rw rw; |
| 277 | struct vnode *vp; |
| 278 | caddr_t base; |
| 279 | int len; |
| 280 | off_t offset; |
| 281 | enum uio_seg segflg; |
| 282 | int ioflg; |
| 283 | struct ucred *cred; |
| 284 | int *aresid; |
| 285 | struct thread *td; |
| 286 | { |
| 287 | struct uio auio; |
| 288 | struct iovec aiov; |
| 289 | int error; |
| 290 | |
| 291 | if ((ioflg & IO_NODELOCKED) == 0) |
| 292 | vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td); |
| 293 | auio.uio_iov = &aiov; |
| 294 | auio.uio_iovcnt = 1; |
| 295 | aiov.iov_base = base; |
| 296 | aiov.iov_len = len; |
| 297 | auio.uio_resid = len; |
| 298 | auio.uio_offset = offset; |
| 299 | auio.uio_segflg = segflg; |
| 300 | auio.uio_rw = rw; |
| 301 | auio.uio_td = td; |
| 302 | if (rw == UIO_READ) { |
| 303 | error = VOP_READ(vp, &auio, ioflg, cred); |
| 304 | } else { |
| 305 | error = VOP_WRITE(vp, &auio, ioflg, cred); |
| 306 | } |
| 307 | if (aresid) |
| 308 | *aresid = auio.uio_resid; |
| 309 | else |
| 310 | if (auio.uio_resid && error == 0) |
| 311 | error = EIO; |
| 312 | if ((ioflg & IO_NODELOCKED) == 0) |
| 313 | VOP_UNLOCK(vp, 0, td); |
| 314 | return (error); |
| 315 | } |
| 316 | |
| 317 | /* |
| 318 | * Package up an I/O request on a vnode into a uio and do it. The I/O |
| 319 | * request is split up into smaller chunks and we try to avoid saturating |
| 320 | * the buffer cache while potentially holding a vnode locked, so we |
| 321 | * check bwillwrite() before calling vn_rdwr(). We also call uio_yield() |
| 322 | * to give other processes a chance to lock the vnode (either other processes |
| 323 | * core'ing the same binary, or unrelated processes scanning the directory). |
| 324 | */ |
| 325 | int |
| 326 | vn_rdwr_inchunks(rw, vp, base, len, offset, segflg, ioflg, cred, aresid, td) |
| 327 | enum uio_rw rw; |
| 328 | struct vnode *vp; |
| 329 | caddr_t base; |
| 330 | int len; |
| 331 | off_t offset; |
| 332 | enum uio_seg segflg; |
| 333 | int ioflg; |
| 334 | struct ucred *cred; |
| 335 | int *aresid; |
| 336 | struct thread *td; |
| 337 | { |
| 338 | int error = 0; |
| 339 | |
| 340 | do { |
| 341 | int chunk = (len > MAXBSIZE) ? MAXBSIZE : len; |
| 342 | |
| 343 | if (rw != UIO_READ && vp->v_type == VREG) |
| 344 | bwillwrite(); |
| 345 | error = vn_rdwr(rw, vp, base, chunk, offset, segflg, |
| 346 | ioflg, cred, aresid, td); |
| 347 | len -= chunk; /* aresid calc already includes length */ |
| 348 | if (error) |
| 349 | break; |
| 350 | offset += chunk; |
| 351 | base += chunk; |
| 352 | uio_yield(); |
| 353 | } while (len); |
| 354 | if (aresid) |
| 355 | *aresid += len; |
| 356 | return (error); |
| 357 | } |
| 358 | |
| 359 | /* |
| 360 | * File table vnode read routine. |
| 361 | */ |
| 362 | static int |
| 363 | vn_read(fp, uio, cred, flags, td) |
| 364 | struct file *fp; |
| 365 | struct uio *uio; |
| 366 | struct ucred *cred; |
| 367 | struct thread *td; |
| 368 | int flags; |
| 369 | { |
| 370 | struct vnode *vp; |
| 371 | int error, ioflag; |
| 372 | |
| 373 | KASSERT(uio->uio_td == td, ("uio_td %p is not td %p", uio->uio_td, td)); |
| 374 | vp = (struct vnode *)fp->f_data; |
| 375 | ioflag = 0; |
| 376 | if (fp->f_flag & FNONBLOCK) |
| 377 | ioflag |= IO_NDELAY; |
| 378 | if (fp->f_flag & O_DIRECT) |
| 379 | ioflag |= IO_DIRECT; |
| 380 | VOP_LEASE(vp, td, cred, LEASE_READ); |
| 381 | vn_lock(vp, LK_SHARED | LK_NOPAUSE | LK_RETRY, td); |
| 382 | if ((flags & FOF_OFFSET) == 0) |
| 383 | uio->uio_offset = fp->f_offset; |
| 384 | |
| 385 | ioflag |= sequential_heuristic(uio, fp); |
| 386 | |
| 387 | error = VOP_READ(vp, uio, ioflag, cred); |
| 388 | if ((flags & FOF_OFFSET) == 0) |
| 389 | fp->f_offset = uio->uio_offset; |
| 390 | fp->f_nextoff = uio->uio_offset; |
| 391 | VOP_UNLOCK(vp, 0, td); |
| 392 | return (error); |
| 393 | } |
| 394 | |
| 395 | /* |
| 396 | * File table vnode write routine. |
| 397 | */ |
| 398 | static int |
| 399 | vn_write(fp, uio, cred, flags, td) |
| 400 | struct file *fp; |
| 401 | struct uio *uio; |
| 402 | struct ucred *cred; |
| 403 | struct thread *td; |
| 404 | int flags; |
| 405 | { |
| 406 | struct vnode *vp; |
| 407 | int error, ioflag; |
| 408 | |
| 409 | KASSERT(uio->uio_td == td, ("uio_procp %p is not p %p", |
| 410 | uio->uio_td, td)); |
| 411 | vp = (struct vnode *)fp->f_data; |
| 412 | if (vp->v_type == VREG) |
| 413 | bwillwrite(); |
| 414 | vp = (struct vnode *)fp->f_data; /* XXX needed? */ |
| 415 | ioflag = IO_UNIT; |
| 416 | if (vp->v_type == VREG && (fp->f_flag & O_APPEND)) |
| 417 | ioflag |= IO_APPEND; |
| 418 | if (fp->f_flag & FNONBLOCK) |
| 419 | ioflag |= IO_NDELAY; |
| 420 | if (fp->f_flag & O_DIRECT) |
| 421 | ioflag |= IO_DIRECT; |
| 422 | if ((fp->f_flag & O_FSYNC) || |
| 423 | (vp->v_mount && (vp->v_mount->mnt_flag & MNT_SYNCHRONOUS))) |
| 424 | ioflag |= IO_SYNC; |
| 425 | VOP_LEASE(vp, td, cred, LEASE_WRITE); |
| 426 | vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td); |
| 427 | if ((flags & FOF_OFFSET) == 0) |
| 428 | uio->uio_offset = fp->f_offset; |
| 429 | ioflag |= sequential_heuristic(uio, fp); |
| 430 | error = VOP_WRITE(vp, uio, ioflag, cred); |
| 431 | if ((flags & FOF_OFFSET) == 0) |
| 432 | fp->f_offset = uio->uio_offset; |
| 433 | fp->f_nextoff = uio->uio_offset; |
| 434 | VOP_UNLOCK(vp, 0, td); |
| 435 | return (error); |
| 436 | } |
| 437 | |
| 438 | /* |
| 439 | * File table vnode stat routine. |
| 440 | */ |
| 441 | static int |
| 442 | vn_statfile(struct file *fp, struct stat *sb, struct thread *td) |
| 443 | { |
| 444 | struct vnode *vp = (struct vnode *)fp->f_data; |
| 445 | |
| 446 | return vn_stat(vp, sb, td); |
| 447 | } |
| 448 | |
| 449 | int |
| 450 | vn_stat(struct vnode *vp, struct stat *sb, struct thread *td) |
| 451 | { |
| 452 | struct vattr vattr; |
| 453 | struct vattr *vap; |
| 454 | int error; |
| 455 | u_short mode; |
| 456 | |
| 457 | vap = &vattr; |
| 458 | error = VOP_GETATTR(vp, vap, td); |
| 459 | if (error) |
| 460 | return (error); |
| 461 | |
| 462 | /* |
| 463 | * Zero the spare stat fields |
| 464 | */ |
| 465 | sb->st_lspare = 0; |
| 466 | sb->st_qspare[0] = 0; |
| 467 | sb->st_qspare[1] = 0; |
| 468 | |
| 469 | /* |
| 470 | * Copy from vattr table |
| 471 | */ |
| 472 | if (vap->va_fsid != VNOVAL) |
| 473 | sb->st_dev = vap->va_fsid; |
| 474 | else |
| 475 | sb->st_dev = vp->v_mount->mnt_stat.f_fsid.val[0]; |
| 476 | sb->st_ino = vap->va_fileid; |
| 477 | mode = vap->va_mode; |
| 478 | switch (vap->va_type) { |
| 479 | case VREG: |
| 480 | mode |= S_IFREG; |
| 481 | break; |
| 482 | case VDIR: |
| 483 | mode |= S_IFDIR; |
| 484 | break; |
| 485 | case VBLK: |
| 486 | mode |= S_IFBLK; |
| 487 | break; |
| 488 | case VCHR: |
| 489 | mode |= S_IFCHR; |
| 490 | break; |
| 491 | case VLNK: |
| 492 | mode |= S_IFLNK; |
| 493 | /* This is a cosmetic change, symlinks do not have a mode. */ |
| 494 | if (vp->v_mount->mnt_flag & MNT_NOSYMFOLLOW) |
| 495 | sb->st_mode &= ~ACCESSPERMS; /* 0000 */ |
| 496 | else |
| 497 | sb->st_mode |= ACCESSPERMS; /* 0777 */ |
| 498 | break; |
| 499 | case VSOCK: |
| 500 | mode |= S_IFSOCK; |
| 501 | break; |
| 502 | case VFIFO: |
| 503 | mode |= S_IFIFO; |
| 504 | break; |
| 505 | default: |
| 506 | return (EBADF); |
| 507 | }; |
| 508 | sb->st_mode = mode; |
| 509 | sb->st_nlink = vap->va_nlink; |
| 510 | sb->st_uid = vap->va_uid; |
| 511 | sb->st_gid = vap->va_gid; |
| 512 | sb->st_rdev = vap->va_rdev; |
| 513 | sb->st_size = vap->va_size; |
| 514 | sb->st_atimespec = vap->va_atime; |
| 515 | sb->st_mtimespec = vap->va_mtime; |
| 516 | sb->st_ctimespec = vap->va_ctime; |
| 517 | |
| 518 | /* |
| 519 | * According to www.opengroup.org, the meaning of st_blksize is |
| 520 | * "a filesystem-specific preferred I/O block size for this |
| 521 | * object. In some filesystem types, this may vary from file |
| 522 | * to file" |
| 523 | * Default to PAGE_SIZE after much discussion. |
| 524 | */ |
| 525 | |
| 526 | if (vap->va_type == VREG) { |
| 527 | sb->st_blksize = vap->va_blocksize; |
| 528 | } else if (vn_isdisk(vp, NULL)) { |
| 529 | sb->st_blksize = vp->v_rdev->si_bsize_best; |
| 530 | if (sb->st_blksize < vp->v_rdev->si_bsize_phys) |
| 531 | sb->st_blksize = vp->v_rdev->si_bsize_phys; |
| 532 | if (sb->st_blksize < BLKDEV_IOSIZE) |
| 533 | sb->st_blksize = BLKDEV_IOSIZE; |
| 534 | } else { |
| 535 | sb->st_blksize = PAGE_SIZE; |
| 536 | } |
| 537 | |
| 538 | sb->st_flags = vap->va_flags; |
| 539 | if (suser(td)) |
| 540 | sb->st_gen = 0; |
| 541 | else |
| 542 | sb->st_gen = vap->va_gen; |
| 543 | |
| 544 | #if (S_BLKSIZE == 512) |
| 545 | /* Optimize this case */ |
| 546 | sb->st_blocks = vap->va_bytes >> 9; |
| 547 | #else |
| 548 | sb->st_blocks = vap->va_bytes / S_BLKSIZE; |
| 549 | #endif |
| 550 | return (0); |
| 551 | } |
| 552 | |
| 553 | /* |
| 554 | * File table vnode ioctl routine. |
| 555 | */ |
| 556 | static int |
| 557 | vn_ioctl(struct file *fp, u_long com, caddr_t data, struct thread *td) |
| 558 | { |
| 559 | struct vnode *vp = ((struct vnode *)fp->f_data); |
| 560 | struct ucred *ucred; |
| 561 | struct vattr vattr; |
| 562 | int error; |
| 563 | |
| 564 | KKASSERT(td->td_proc != NULL); |
| 565 | ucred = td->td_proc->p_ucred; |
| 566 | |
| 567 | switch (vp->v_type) { |
| 568 | case VREG: |
| 569 | case VDIR: |
| 570 | if (com == FIONREAD) { |
| 571 | error = VOP_GETATTR(vp, &vattr, td); |
| 572 | if (error) |
| 573 | return (error); |
| 574 | *(int *)data = vattr.va_size - fp->f_offset; |
| 575 | return (0); |
| 576 | } |
| 577 | if (com == FIONBIO || com == FIOASYNC) /* XXX */ |
| 578 | return (0); /* XXX */ |
| 579 | /* fall into ... */ |
| 580 | default: |
| 581 | #if 0 |
| 582 | return (ENOTTY); |
| 583 | #endif |
| 584 | case VFIFO: |
| 585 | case VCHR: |
| 586 | case VBLK: |
| 587 | if (com == FIODTYPE) { |
| 588 | if (vp->v_type != VCHR && vp->v_type != VBLK) |
| 589 | return (ENOTTY); |
| 590 | *(int *)data = dev_dflags(vp->v_rdev) & D_TYPEMASK; |
| 591 | return (0); |
| 592 | } |
| 593 | error = VOP_IOCTL(vp, com, data, fp->f_flag, ucred, td); |
| 594 | if (error == 0 && com == TIOCSCTTY) { |
| 595 | struct session *sess = td->td_proc->p_session; |
| 596 | |
| 597 | /* Do nothing if reassigning same control tty */ |
| 598 | if (sess->s_ttyvp == vp) |
| 599 | return (0); |
| 600 | |
| 601 | /* Get rid of reference to old control tty */ |
| 602 | if (sess->s_ttyvp) |
| 603 | vrele(sess->s_ttyvp); |
| 604 | |
| 605 | sess->s_ttyvp = vp; |
| 606 | VREF(vp); |
| 607 | } |
| 608 | return (error); |
| 609 | } |
| 610 | } |
| 611 | |
| 612 | /* |
| 613 | * File table vnode poll routine. |
| 614 | */ |
| 615 | static int |
| 616 | vn_poll(struct file *fp, int events, struct ucred *cred, struct thread *td) |
| 617 | { |
| 618 | return (VOP_POLL(((struct vnode *)fp->f_data), events, cred, td)); |
| 619 | } |
| 620 | |
| 621 | /* |
| 622 | * Check that the vnode is still valid, and if so |
| 623 | * acquire requested lock. |
| 624 | */ |
| 625 | int |
| 626 | #ifndef DEBUG_LOCKS |
| 627 | vn_lock(struct vnode *vp, int flags, struct thread *td) |
| 628 | #else |
| 629 | debug_vn_lock(struct vnode *vp, int flags, struct thread *td, |
| 630 | const char *filename, int line) |
| 631 | #endif |
| 632 | { |
| 633 | int error; |
| 634 | |
| 635 | do { |
| 636 | if ((flags & LK_INTERLOCK) == 0) |
| 637 | lwkt_gettoken(&vp->v_interlock); |
| 638 | if ((vp->v_flag & VXLOCK) && vp->v_vxproc != curproc) { |
| 639 | vp->v_flag |= VXWANT; |
| 640 | lwkt_reltoken(&vp->v_interlock); |
| 641 | tsleep((caddr_t)vp, 0, "vn_lock", 0); |
| 642 | error = ENOENT; |
| 643 | } else { |
| 644 | #if 0 |
| 645 | /* this can now occur in normal operation */ |
| 646 | if (vp->v_vxproc != NULL) |
| 647 | log(LOG_INFO, "VXLOCK interlock avoided in vn_lock\n"); |
| 648 | #endif |
| 649 | #ifdef DEBUG_LOCKS |
| 650 | vp->filename = filename; |
| 651 | vp->line = line; |
| 652 | #endif |
| 653 | error = VOP_LOCK(vp, |
| 654 | flags | LK_NOPAUSE | LK_INTERLOCK, td); |
| 655 | if (error == 0) |
| 656 | return (error); |
| 657 | } |
| 658 | flags &= ~LK_INTERLOCK; |
| 659 | } while (flags & LK_RETRY); |
| 660 | return (error); |
| 661 | } |
| 662 | |
| 663 | /* |
| 664 | * File table vnode close routine. |
| 665 | */ |
| 666 | static int |
| 667 | vn_closefile(struct file *fp, struct thread *td) |
| 668 | { |
| 669 | int err; |
| 670 | |
| 671 | fp->f_ops = &badfileops; |
| 672 | err = vn_close(((struct vnode *)fp->f_data), fp->f_flag, td); |
| 673 | return(err); |
| 674 | } |
| 675 | |
| 676 | static int |
| 677 | vn_kqfilter(struct file *fp, struct knote *kn) |
| 678 | { |
| 679 | |
| 680 | return (VOP_KQFILTER(((struct vnode *)fp->f_data), kn)); |
| 681 | } |