1 /* $NetBSD: ext2fs_readwrite.c,v 1.16 2001/02/27 04:37:47 chs Exp $ */ 2 3 /*- 4 * Copyright (c) 1997 Manuel Bouyer. 5 * Copyright (c) 1993 6 * The Regents of the University of California. All rights reserved. 7 * 8 * Redistribution and use in source and binary forms, with or without 9 * modification, are permitted provided that the following conditions 10 * are met: 11 * 1. Redistributions of source code must retain the above copyright 12 * notice, this list of conditions and the following disclaimer. 13 * 2. Redistributions in binary form must reproduce the above copyright 14 * notice, this list of conditions and the following disclaimer in the 15 * documentation and/or other materials provided with the distribution. 16 * 3. All advertising materials mentioning features or use of this software 17 * must display the following acknowledgement: 18 * This product includes software developed by the University of 19 * California, Berkeley and its contributors. 20 * 4. Neither the name of the University nor the names of its contributors 21 * may be used to endorse or promote products derived from this software 22 * without specific prior written permission. 23 * 24 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 25 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 26 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 27 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 28 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 29 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 30 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 31 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 32 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 33 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 34 * SUCH DAMAGE. 35 * 36 * @(#)ufs_readwrite.c 8.8 (Berkeley) 8/4/94 37 * Modified for ext2fs by Manuel Bouyer. 38 */ 39 40 #include <sys/param.h> 41 #include <sys/systm.h> 42 #include <sys/resourcevar.h> 43 #include <sys/kernel.h> 44 #include <sys/file.h> 45 #include <sys/stat.h> 46 #include <sys/buf.h> 47 #include <sys/proc.h> 48 #include <sys/conf.h> 49 #include <sys/mount.h> 50 #include <sys/vnode.h> 51 #include <sys/malloc.h> 52 #include <sys/signalvar.h> 53 54 #include <ufs/ufs/quota.h> 55 #include <ufs/ufs/inode.h> 56 #include <ufs/ext2fs/ext2fs.h> 57 #include <ufs/ext2fs/ext2fs_extern.h> 58 59 60 #define doclusterread 0 /* XXX underway */ 61 #define doclusterwrite 0 62 63 /* 64 * Vnode op for reading. 65 */ 66 /* ARGSUSED */ 67 int 68 ext2fs_read(v) 69 void *v; 70 { 71 struct vop_read_args /* { 72 struct vnode *a_vp; 73 struct uio *a_uio; 74 int a_ioflag; 75 struct ucred *a_cred; 76 } */ *ap = v; 77 struct vnode *vp; 78 struct inode *ip; 79 struct uio *uio; 80 struct m_ext2fs *fs; 81 struct buf *bp; 82 void *win; 83 vsize_t bytelen; 84 ufs_daddr_t lbn, nextlbn; 85 off_t bytesinfile; 86 long size, xfersize, blkoffset; 87 int error; 88 89 vp = ap->a_vp; 90 ip = VTOI(vp); 91 uio = ap->a_uio; 92 93 #ifdef DIAGNOSTIC 94 if (uio->uio_rw != UIO_READ) 95 panic("%s: mode", "ext2fs_read"); 96 97 if (vp->v_type == VLNK) { 98 if ((int)ip->i_e2fs_size < vp->v_mount->mnt_maxsymlinklen || 99 (vp->v_mount->mnt_maxsymlinklen == 0 && 100 ip->i_e2fs_nblock == 0)) 101 panic("%s: short symlink", "ext2fs_read"); 102 } else if (vp->v_type != VREG && vp->v_type != VDIR) 103 panic("%s: type %d", "ext2fs_read", vp->v_type); 104 #endif 105 fs = ip->i_e2fs; 106 if ((u_int64_t)uio->uio_offset > 107 ((u_int64_t)0x80000000 * fs->e2fs_bsize - 1)) 108 return (EFBIG); 109 if (uio->uio_resid == 0) 110 return (0); 111 112 if (vp->v_type == VREG) { 113 error = 0; 114 while (uio->uio_resid > 0) { 115 116 bytelen = MIN(ip->i_e2fs_size - uio->uio_offset, 117 uio->uio_resid); 118 119 if (bytelen == 0) { 120 break; 121 } 122 win = ubc_alloc(&vp->v_uvm.u_obj, uio->uio_offset, 123 &bytelen, UBC_READ); 124 error = uiomove(win, bytelen, uio); 125 ubc_release(win, 0); 126 if (error) { 127 break; 128 } 129 } 130 goto out; 131 } 132 133 for (error = 0, bp = NULL; uio->uio_resid > 0; bp = NULL) { 134 if ((bytesinfile = ip->i_e2fs_size - uio->uio_offset) <= 0) 135 break; 136 lbn = lblkno(fs, uio->uio_offset); 137 nextlbn = lbn + 1; 138 size = fs->e2fs_bsize; 139 blkoffset = blkoff(fs, uio->uio_offset); 140 xfersize = fs->e2fs_bsize - blkoffset; 141 if (uio->uio_resid < xfersize) 142 xfersize = uio->uio_resid; 143 if (bytesinfile < xfersize) 144 xfersize = bytesinfile; 145 146 if (lblktosize(fs, nextlbn) >= ip->i_e2fs_size) 147 error = bread(vp, lbn, size, NOCRED, &bp); 148 else if (doclusterread) 149 error = cluster_read(vp, 150 ip->i_e2fs_size, lbn, size, NOCRED, &bp); 151 else if (lbn - 1 == vp->v_lastr) { 152 int nextsize = fs->e2fs_bsize; 153 error = breadn(vp, lbn, 154 size, &nextlbn, &nextsize, 1, NOCRED, &bp); 155 } else 156 error = bread(vp, lbn, size, NOCRED, &bp); 157 if (error) 158 break; 159 vp->v_lastr = lbn; 160 161 /* 162 * We should only get non-zero b_resid when an I/O error 163 * has occurred, which should cause us to break above. 164 * However, if the short read did not cause an error, 165 * then we want to ensure that we do not uiomove bad 166 * or uninitialized data. 167 */ 168 size -= bp->b_resid; 169 if (size < xfersize) { 170 if (size == 0) 171 break; 172 xfersize = size; 173 } 174 error = uiomove((char *)bp->b_data + blkoffset, xfersize, uio); 175 if (error) 176 break; 177 brelse(bp); 178 } 179 if (bp != NULL) 180 brelse(bp); 181 182 out: 183 if (!(vp->v_mount->mnt_flag & MNT_NOATIME)) { 184 ip->i_flag |= IN_ACCESS; 185 if ((ap->a_ioflag & IO_SYNC) == IO_SYNC) 186 error = VOP_UPDATE(vp, NULL, NULL, UPDATE_WAIT); 187 } 188 return (error); 189 } 190 191 /* 192 * Vnode op for writing. 193 */ 194 int 195 ext2fs_write(v) 196 void *v; 197 { 198 struct vop_write_args /* { 199 struct vnode *a_vp; 200 struct uio *a_uio; 201 int a_ioflag; 202 struct ucred *a_cred; 203 } */ *ap = v; 204 struct vnode *vp; 205 struct uio *uio; 206 struct inode *ip; 207 struct m_ext2fs *fs; 208 struct buf *bp; 209 struct proc *p; 210 ufs_daddr_t lbn; 211 off_t osize; 212 int blkoffset, error, flags, ioflag, resid, xfersize; 213 vsize_t bytelen; 214 void *win; 215 off_t oldoff; 216 boolean_t rv; 217 218 ioflag = ap->a_ioflag; 219 uio = ap->a_uio; 220 vp = ap->a_vp; 221 ip = VTOI(vp); 222 error = 0; 223 224 #ifdef DIAGNOSTIC 225 if (uio->uio_rw != UIO_WRITE) 226 panic("%s: mode", "ext2fs_write"); 227 #endif 228 229 switch (vp->v_type) { 230 case VREG: 231 if (ioflag & IO_APPEND) 232 uio->uio_offset = ip->i_e2fs_size; 233 if ((ip->i_e2fs_flags & EXT2_APPEND) && 234 uio->uio_offset != ip->i_e2fs_size) 235 return (EPERM); 236 /* FALLTHROUGH */ 237 case VLNK: 238 break; 239 case VDIR: 240 if ((ioflag & IO_SYNC) == 0) 241 panic("%s: nonsync dir write", "ext2fs_write"); 242 break; 243 default: 244 panic("%s: type", "ext2fs_write"); 245 } 246 247 fs = ip->i_e2fs; 248 if (uio->uio_offset < 0 || 249 (u_int64_t)uio->uio_offset + uio->uio_resid > 250 ((u_int64_t)0x80000000 * fs->e2fs_bsize - 1)) 251 return (EFBIG); 252 /* 253 * Maybe this should be above the vnode op call, but so long as 254 * file servers have no limits, I don't think it matters. 255 */ 256 p = uio->uio_procp; 257 if (vp->v_type == VREG && p && 258 uio->uio_offset + uio->uio_resid > 259 p->p_rlimit[RLIMIT_FSIZE].rlim_cur) { 260 psignal(p, SIGXFSZ); 261 return (EFBIG); 262 } 263 264 resid = uio->uio_resid; 265 osize = ip->i_e2fs_size; 266 267 if (vp->v_type == VREG) { 268 while (uio->uio_resid > 0) { 269 oldoff = uio->uio_offset; 270 blkoffset = blkoff(fs, uio->uio_offset); 271 bytelen = MIN(fs->e2fs_bsize - blkoffset, 272 uio->uio_resid); 273 274 /* 275 * XXXUBC if file is mapped and this is the last block, 276 * process one page at a time. 277 */ 278 279 error = ext2fs_balloc_range(vp, uio->uio_offset, 280 bytelen, ap->a_cred, 0); 281 if (error) { 282 break; 283 } 284 win = ubc_alloc(&vp->v_uvm.u_obj, uio->uio_offset, 285 &bytelen, UBC_WRITE); 286 error = uiomove(win, bytelen, uio); 287 ubc_release(win, 0); 288 if (error) { 289 break; 290 } 291 292 /* 293 * flush what we just wrote if necessary. 294 * XXXUBC simplistic async flushing. 295 */ 296 297 if (oldoff >> 16 != uio->uio_offset >> 16) { 298 simple_lock(&vp->v_uvm.u_obj.vmobjlock); 299 rv = vp->v_uvm.u_obj.pgops->pgo_flush( 300 &vp->v_uvm.u_obj, (oldoff >> 16) << 16, 301 (uio->uio_offset >> 16) << 16, PGO_CLEANIT); 302 simple_unlock(&vp->v_uvm.u_obj.vmobjlock); 303 } 304 } 305 goto out; 306 } 307 308 flags = ioflag & IO_SYNC ? B_SYNC : 0; 309 for (error = 0; uio->uio_resid > 0;) { 310 lbn = lblkno(fs, uio->uio_offset); 311 blkoffset = blkoff(fs, uio->uio_offset); 312 xfersize = MIN(fs->e2fs_bsize - blkoffset, uio->uio_resid); 313 if (xfersize < fs->e2fs_bsize) 314 flags |= B_CLRBUF; 315 else 316 flags &= ~B_CLRBUF; 317 error = ext2fs_balloc(ip, 318 lbn, blkoffset + xfersize, ap->a_cred, &bp, flags); 319 if (error) 320 break; 321 if (ip->i_e2fs_size < uio->uio_offset + xfersize) { 322 ip->i_e2fs_size = uio->uio_offset + xfersize; 323 } 324 error = uiomove((char *)bp->b_data + blkoffset, xfersize, uio); 325 if (ioflag & IO_SYNC) 326 (void)bwrite(bp); 327 else if (xfersize + blkoffset == fs->e2fs_bsize) 328 if (doclusterwrite) 329 cluster_write(bp, ip->i_e2fs_size); 330 else 331 bawrite(bp); 332 else 333 bdwrite(bp); 334 if (error || xfersize == 0) 335 break; 336 } 337 /* 338 * If we successfully wrote any data, and we are not the superuser 339 * we clear the setuid and setgid bits as a precaution against 340 * tampering. 341 */ 342 out: 343 ip->i_flag |= IN_CHANGE | IN_UPDATE; 344 if (resid > uio->uio_resid && ap->a_cred && ap->a_cred->cr_uid != 0) 345 ip->i_e2fs_mode &= ~(ISUID | ISGID); 346 if (error) { 347 if (ioflag & IO_UNIT) { 348 (void)VOP_TRUNCATE(vp, osize, 349 ioflag & IO_SYNC, ap->a_cred, uio->uio_procp); 350 uio->uio_offset -= resid - uio->uio_resid; 351 uio->uio_resid = resid; 352 } 353 } else if (resid > uio->uio_resid && (ioflag & IO_SYNC) == IO_SYNC) 354 error = VOP_UPDATE(vp, NULL, NULL, UPDATE_WAIT); 355 return (error); 356 } 357