ext2fs_readwrite.c revision 1.70
1 /* $NetBSD: ext2fs_readwrite.c,v 1.70 2015/03/28 03:53:36 riastradh Exp $ */ 2 3 /*- 4 * Copyright (c) 1993 5 * The Regents of the University of California. All rights reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 3. Neither the name of the University nor the names of its contributors 16 * may be used to endorse or promote products derived from this software 17 * without specific prior written permission. 18 * 19 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 22 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 29 * SUCH DAMAGE. 30 * 31 * @(#)ufs_readwrite.c 8.8 (Berkeley) 8/4/94 32 * Modified for ext2fs by Manuel Bouyer. 33 */ 34 35 /*- 36 * Copyright (c) 1997 Manuel Bouyer. 37 * 38 * Redistribution and use in source and binary forms, with or without 39 * modification, are permitted provided that the following conditions 40 * are met: 41 * 1. Redistributions of source code must retain the above copyright 42 * notice, this list of conditions and the following disclaimer. 43 * 2. Redistributions in binary form must reproduce the above copyright 44 * notice, this list of conditions and the following disclaimer in the 45 * documentation and/or other materials provided with the distribution. 46 * 47 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR 48 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES 49 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. 50 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, 51 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT 52 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 53 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 54 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 55 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF 56 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 57 * 58 * @(#)ufs_readwrite.c 8.8 (Berkeley) 8/4/94 59 * Modified for ext2fs by Manuel Bouyer. 60 */ 61 62 #include <sys/cdefs.h> 63 __KERNEL_RCSID(0, "$NetBSD: ext2fs_readwrite.c,v 1.70 2015/03/28 03:53:36 riastradh Exp $"); 64 65 #include <sys/param.h> 66 #include <sys/systm.h> 67 #include <sys/resourcevar.h> 68 #include <sys/kernel.h> 69 #include <sys/file.h> 70 #include <sys/stat.h> 71 #include <sys/buf.h> 72 #include <sys/proc.h> 73 #include <sys/mount.h> 74 #include <sys/vnode.h> 75 #include <sys/signalvar.h> 76 #include <sys/kauth.h> 77 78 #include <ufs/ufs/inode.h> 79 #include <ufs/ufs/ufsmount.h> 80 #include <ufs/ufs/ufs_extern.h> 81 #include <ufs/ext2fs/ext2fs.h> 82 #include <ufs/ext2fs/ext2fs_extern.h> 83 84 /* 85 * Vnode op for reading. 86 */ 87 /* ARGSUSED */ 88 int 89 ext2fs_read(void *v) 90 { 91 struct vop_read_args /* { 92 struct vnode *a_vp; 93 struct uio *a_uio; 94 int a_ioflag; 95 kauth_cred_t a_cred; 96 } */ *ap = v; 97 struct vnode *vp; 98 struct inode *ip; 99 struct uio *uio; 100 struct ufsmount *ump; 101 vsize_t bytelen; 102 int advice; 103 int error; 104 105 vp = ap->a_vp; 106 ip = VTOI(vp); 107 ump = ip->i_ump; 108 uio = ap->a_uio; 109 error = 0; 110 111 KASSERT(uio->uio_rw == UIO_READ); 112 KASSERT(vp->v_type == VREG || vp->v_type == VDIR); 113 114 /* XXX Eliminate me by refusing directory reads from userland. */ 115 if (vp->v_type == VDIR) 116 return ext2fs_bufrd(vp, uio, ap->a_ioflag, ap->a_cred); 117 118 if ((uint64_t)uio->uio_offset > ump->um_maxfilesize) 119 return (EFBIG); 120 if (uio->uio_resid == 0) 121 return (0); 122 if (uio->uio_offset >= ext2fs_size(ip)) 123 goto out; 124 125 KASSERT(vp->v_type == VREG); 126 advice = IO_ADV_DECODE(ap->a_ioflag); 127 while (uio->uio_resid > 0) { 128 bytelen = MIN(ext2fs_size(ip) - uio->uio_offset, 129 uio->uio_resid); 130 if (bytelen == 0) 131 break; 132 133 error = ubc_uiomove(&vp->v_uobj, uio, bytelen, advice, 134 UBC_READ | UBC_PARTIALOK | UBC_UNMAP_FLAG(vp)); 135 if (error) 136 break; 137 } 138 139 out: 140 if (!(vp->v_mount->mnt_flag & MNT_NOATIME)) { 141 ip->i_flag |= IN_ACCESS; 142 if ((ap->a_ioflag & IO_SYNC) == IO_SYNC) 143 error = ext2fs_update(vp, NULL, NULL, UPDATE_WAIT); 144 } 145 return (error); 146 } 147 148 /* 149 * UFS op for reading via the buffer cache 150 */ 151 int 152 ext2fs_bufrd(struct vnode *vp, struct uio *uio, int ioflag, kauth_cred_t cred) 153 { 154 struct inode *ip; 155 struct ufsmount *ump; 156 struct m_ext2fs *fs; 157 struct buf *bp; 158 off_t bytesinfile; 159 daddr_t lbn, nextlbn; 160 long size, xfersize, blkoffset; 161 int error; 162 163 KASSERT(uio->uio_rw == UIO_READ); 164 KASSERT(VOP_ISLOCKED(vp)); 165 KASSERT(vp->v_type == VDIR || vp->v_type == VLNK); 166 167 ip = VTOI(vp); 168 ump = ip->i_ump; 169 fs = ip->i_e2fs; 170 error = 0; 171 172 KASSERT(vp->v_type != VLNK || 173 ext2fs_size(ip) >= ump->um_maxsymlinklen); 174 KASSERT(vp->v_type != VLNK || ump->um_maxsymlinklen != 0 || 175 ext2fs_nblock(ip) != 0); 176 177 if (uio->uio_offset > ump->um_maxfilesize) 178 return EFBIG; 179 if (uio->uio_resid == 0) 180 return 0; 181 if (uio->uio_offset >= ext2fs_size(ip)) 182 goto out; 183 184 for (error = 0, bp = NULL; uio->uio_resid > 0; bp = NULL) { 185 bytesinfile = ext2fs_size(ip) - uio->uio_offset; 186 if (bytesinfile <= 0) 187 break; 188 lbn = ext2_lblkno(fs, uio->uio_offset); 189 nextlbn = lbn + 1; 190 size = fs->e2fs_bsize; 191 blkoffset = ext2_blkoff(fs, uio->uio_offset); 192 xfersize = fs->e2fs_bsize - blkoffset; 193 if (uio->uio_resid < xfersize) 194 xfersize = uio->uio_resid; 195 if (bytesinfile < xfersize) 196 xfersize = bytesinfile; 197 198 if (ext2_lblktosize(fs, nextlbn) >= ext2fs_size(ip)) 199 error = bread(vp, lbn, size, NOCRED, 0, &bp); 200 else { 201 int nextsize = fs->e2fs_bsize; 202 error = breadn(vp, lbn, 203 size, &nextlbn, &nextsize, 1, NOCRED, 0, &bp); 204 } 205 if (error) 206 break; 207 208 /* 209 * We should only get non-zero b_resid when an I/O error 210 * has occurred, which should cause us to break above. 211 * However, if the short read did not cause an error, 212 * then we want to ensure that we do not uiomove bad 213 * or uninitialized data. 214 */ 215 size -= bp->b_resid; 216 if (size < xfersize) { 217 if (size == 0) 218 break; 219 xfersize = size; 220 } 221 error = uiomove((char *)bp->b_data + blkoffset, xfersize, uio); 222 if (error) 223 break; 224 brelse(bp, 0); 225 } 226 if (bp != NULL) 227 brelse(bp, 0); 228 229 out: 230 if (!(vp->v_mount->mnt_flag & MNT_NOATIME)) { 231 ip->i_flag |= IN_ACCESS; 232 if ((ioflag & IO_SYNC) == IO_SYNC) 233 error = ext2fs_update(vp, NULL, NULL, UPDATE_WAIT); 234 } 235 return (error); 236 } 237 238 /* 239 * Vnode op for writing. 240 */ 241 int 242 ext2fs_write(void *v) 243 { 244 struct vop_write_args /* { 245 struct vnode *a_vp; 246 struct uio *a_uio; 247 int a_ioflag; 248 kauth_cred_t a_cred; 249 } */ *ap = v; 250 struct vnode *vp; 251 struct uio *uio; 252 struct inode *ip; 253 struct m_ext2fs *fs; 254 struct ufsmount *ump; 255 off_t osize; 256 int blkoffset, error, ioflag, resid; 257 vsize_t bytelen; 258 off_t oldoff = 0; /* XXX */ 259 bool async; 260 int extended = 0; 261 int advice; 262 263 ioflag = ap->a_ioflag; 264 advice = IO_ADV_DECODE(ioflag); 265 uio = ap->a_uio; 266 vp = ap->a_vp; 267 ip = VTOI(vp); 268 ump = ip->i_ump; 269 error = 0; 270 271 KASSERT(uio->uio_rw == UIO_WRITE); 272 KASSERT(vp->v_type == VREG); 273 274 if (ioflag & IO_APPEND) 275 uio->uio_offset = ext2fs_size(ip); 276 if ((ip->i_e2fs_flags & EXT2_APPEND) && 277 uio->uio_offset != ext2fs_size(ip)) 278 return (EPERM); 279 280 fs = ip->i_e2fs; 281 if (uio->uio_offset < 0 || 282 (uint64_t)uio->uio_offset + uio->uio_resid > ump->um_maxfilesize) 283 return (EFBIG); 284 if (uio->uio_resid == 0) 285 return (0); 286 287 async = vp->v_mount->mnt_flag & MNT_ASYNC; 288 resid = uio->uio_resid; 289 osize = ext2fs_size(ip); 290 291 KASSERT(vp->v_type == VREG); 292 while (uio->uio_resid > 0) { 293 oldoff = uio->uio_offset; 294 blkoffset = ext2_blkoff(fs, uio->uio_offset); 295 bytelen = MIN(fs->e2fs_bsize - blkoffset, uio->uio_resid); 296 297 if (vp->v_size < oldoff + bytelen) { 298 uvm_vnp_setwritesize(vp, oldoff + bytelen); 299 } 300 error = ufs_balloc_range(vp, uio->uio_offset, bytelen, 301 ap->a_cred, 0); 302 if (error) 303 break; 304 error = ubc_uiomove(&vp->v_uobj, uio, bytelen, advice, 305 UBC_WRITE | UBC_UNMAP_FLAG(vp)); 306 if (error) 307 break; 308 309 /* 310 * update UVM's notion of the size now that we've 311 * copied the data into the vnode's pages. 312 */ 313 314 if (vp->v_size < uio->uio_offset) { 315 uvm_vnp_setsize(vp, uio->uio_offset); 316 extended = 1; 317 } 318 319 /* 320 * flush what we just wrote if necessary. 321 * XXXUBC simplistic async flushing. 322 */ 323 324 if (!async && oldoff >> 16 != uio->uio_offset >> 16) { 325 mutex_enter(vp->v_interlock); 326 error = VOP_PUTPAGES(vp, (oldoff >> 16) << 16, 327 (uio->uio_offset >> 16) << 16, 328 PGO_CLEANIT | PGO_LAZY); 329 } 330 } 331 if (error == 0 && ioflag & IO_SYNC) { 332 mutex_enter(vp->v_interlock); 333 error = VOP_PUTPAGES(vp, trunc_page(oldoff), 334 round_page(ext2_blkroundup(fs, uio->uio_offset)), 335 PGO_CLEANIT | PGO_SYNCIO); 336 } 337 338 /* 339 * If we successfully wrote any data, and we are not the superuser 340 * we clear the setuid and setgid bits as a precaution against 341 * tampering. 342 */ 343 ip->i_flag |= IN_CHANGE | IN_UPDATE; 344 if (vp->v_mount->mnt_flag & MNT_RELATIME) 345 ip->i_flag |= IN_ACCESS; 346 if (resid > uio->uio_resid && ap->a_cred) { 347 if (ip->i_e2fs_mode & ISUID) { 348 if (kauth_authorize_vnode(ap->a_cred, 349 KAUTH_VNODE_RETAIN_SUID, vp, NULL, EPERM) != 0) 350 ip->i_e2fs_mode &= ISUID; 351 } 352 353 if (ip->i_e2fs_mode & ISGID) { 354 if (kauth_authorize_vnode(ap->a_cred, 355 KAUTH_VNODE_RETAIN_SGID, vp, NULL, EPERM) != 0) 356 ip->i_e2fs_mode &= ~ISGID; 357 } 358 } 359 if (resid > uio->uio_resid) 360 VN_KNOTE(vp, NOTE_WRITE | (extended ? NOTE_EXTEND : 0)); 361 if (error) { 362 (void) ext2fs_truncate(vp, osize, ioflag & IO_SYNC, ap->a_cred); 363 uio->uio_offset -= resid - uio->uio_resid; 364 uio->uio_resid = resid; 365 } else if (resid > uio->uio_resid && (ioflag & IO_SYNC) == IO_SYNC) 366 error = ext2fs_update(vp, NULL, NULL, UPDATE_WAIT); 367 KASSERT(vp->v_size == ext2fs_size(ip)); 368 return (error); 369 } 370 371 /* 372 * UFS op for writing via the buffer cache 373 */ 374 int 375 ext2fs_bufwr(struct vnode *vp, struct uio *uio, int ioflag, kauth_cred_t cred) 376 { 377 struct inode *ip; 378 struct ufsmount *ump; 379 struct m_ext2fs *fs; 380 struct buf *bp; 381 int flags; 382 off_t osize; 383 daddr_t lbn; 384 int resid, blkoffset, xfersize; 385 int extended = 0; 386 int error; 387 388 KASSERT(VOP_ISLOCKED(vp) == LK_EXCLUSIVE); 389 KASSERT(vp->v_type == VDIR || vp->v_type == VLNK); 390 KASSERT(vp->v_type != VDIR || ISSET(ioflag, IO_SYNC)); 391 KASSERT(uio->uio_rw == UIO_WRITE); 392 393 ip = VTOI(vp); 394 ump = ip->i_ump; 395 fs = ip->i_e2fs; 396 error = 0; 397 398 if (uio->uio_offset < 0 || 399 uio->uio_resid > ump->um_maxfilesize || 400 uio->uio_offset > (ump->um_maxfilesize - uio->uio_resid)) 401 return EFBIG; 402 if (uio->uio_resid == 0) 403 return 0; 404 405 flags = ioflag & IO_SYNC ? B_SYNC : 0; 406 resid = uio->uio_resid; 407 osize = ext2fs_size(ip); 408 409 for (error = 0; uio->uio_resid > 0;) { 410 lbn = ext2_lblkno(fs, uio->uio_offset); 411 blkoffset = ext2_blkoff(fs, uio->uio_offset); 412 xfersize = MIN(fs->e2fs_bsize - blkoffset, uio->uio_resid); 413 if (xfersize < fs->e2fs_bsize) 414 flags |= B_CLRBUF; 415 else 416 flags &= ~B_CLRBUF; 417 error = ext2fs_balloc(ip, lbn, blkoffset + xfersize, cred, &bp, 418 flags); 419 if (error) 420 break; 421 if (ext2fs_size(ip) < uio->uio_offset + xfersize) { 422 error = ext2fs_setsize(ip, uio->uio_offset + xfersize); 423 if (error) 424 break; 425 } 426 error = uiomove((char *)bp->b_data + blkoffset, xfersize, uio); 427 428 /* 429 * update UVM's notion of the size now that we've 430 * copied the data into the vnode's pages. 431 */ 432 433 if (vp->v_size < uio->uio_offset) { 434 uvm_vnp_setsize(vp, uio->uio_offset); 435 extended = 1; 436 } 437 438 if (ioflag & IO_SYNC) 439 (void)bwrite(bp); 440 else if (xfersize + blkoffset == fs->e2fs_bsize) 441 bawrite(bp); 442 else 443 bdwrite(bp); 444 if (error || xfersize == 0) 445 break; 446 } 447 448 /* 449 * If we successfully wrote any data, and we are not the superuser 450 * we clear the setuid and setgid bits as a precaution against 451 * tampering. 452 */ 453 ip->i_flag |= IN_CHANGE | IN_UPDATE; 454 if (vp->v_mount->mnt_flag & MNT_RELATIME) 455 ip->i_flag |= IN_ACCESS; 456 if (resid > uio->uio_resid && cred) { 457 if (ip->i_e2fs_mode & ISUID) { 458 if (kauth_authorize_vnode(cred, 459 KAUTH_VNODE_RETAIN_SUID, vp, NULL, EPERM) != 0) 460 ip->i_e2fs_mode &= ISUID; 461 } 462 463 if (ip->i_e2fs_mode & ISGID) { 464 if (kauth_authorize_vnode(cred, 465 KAUTH_VNODE_RETAIN_SGID, vp, NULL, EPERM) != 0) 466 ip->i_e2fs_mode &= ~ISGID; 467 } 468 } 469 if (resid > uio->uio_resid) 470 VN_KNOTE(vp, NOTE_WRITE | (extended ? NOTE_EXTEND : 0)); 471 if (error) { 472 (void) ext2fs_truncate(vp, osize, ioflag & IO_SYNC, cred); 473 uio->uio_offset -= resid - uio->uio_resid; 474 uio->uio_resid = resid; 475 } else if (resid > uio->uio_resid && (ioflag & IO_SYNC) == IO_SYNC) 476 error = ext2fs_update(vp, NULL, NULL, UPDATE_WAIT); 477 KASSERT(vp->v_size == ext2fs_size(ip)); 478 return (error); 479 } 480
Indexes created Tue Sep 30 20:09:53 GMT 2025