1 /* $NetBSD: ext2fs_readwrite.c,v 1.18 2001/09/22 22:44:08 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 if (uio->uio_offset >= ip->i_e2fs_size) 112 return (0); 113 114 if (vp->v_type == VREG) { 115 error = 0; 116 while (uio->uio_resid > 0) { 117 bytelen = MIN(ip->i_e2fs_size - uio->uio_offset, 118 uio->uio_resid); 119 120 if (bytelen == 0) { 121 break; 122 } 123 win = ubc_alloc(&vp->v_uobj, uio->uio_offset, 124 &bytelen, UBC_READ); 125 error = uiomove(win, bytelen, uio); 126 ubc_release(win, 0); 127 if (error) { 128 break; 129 } 130 } 131 goto out; 132 } 133 134 for (error = 0, bp = NULL; uio->uio_resid > 0; bp = NULL) { 135 if ((bytesinfile = ip->i_e2fs_size - uio->uio_offset) <= 0) 136 break; 137 lbn = lblkno(fs, uio->uio_offset); 138 nextlbn = lbn + 1; 139 size = fs->e2fs_bsize; 140 blkoffset = blkoff(fs, uio->uio_offset); 141 xfersize = fs->e2fs_bsize - blkoffset; 142 if (uio->uio_resid < xfersize) 143 xfersize = uio->uio_resid; 144 if (bytesinfile < xfersize) 145 xfersize = bytesinfile; 146 147 if (lblktosize(fs, nextlbn) >= ip->i_e2fs_size) 148 error = bread(vp, lbn, size, NOCRED, &bp); 149 else { 150 int nextsize = fs->e2fs_bsize; 151 error = breadn(vp, lbn, 152 size, &nextlbn, &nextsize, 1, NOCRED, &bp); 153 } 154 if (error) 155 break; 156 157 /* 158 * We should only get non-zero b_resid when an I/O error 159 * has occurred, which should cause us to break above. 160 * However, if the short read did not cause an error, 161 * then we want to ensure that we do not uiomove bad 162 * or uninitialized data. 163 */ 164 size -= bp->b_resid; 165 if (size < xfersize) { 166 if (size == 0) 167 break; 168 xfersize = size; 169 } 170 error = uiomove((char *)bp->b_data + blkoffset, xfersize, uio); 171 if (error) 172 break; 173 brelse(bp); 174 } 175 if (bp != NULL) 176 brelse(bp); 177 178 out: 179 if (!(vp->v_mount->mnt_flag & MNT_NOATIME)) { 180 ip->i_flag |= IN_ACCESS; 181 if ((ap->a_ioflag & IO_SYNC) == IO_SYNC) 182 error = VOP_UPDATE(vp, NULL, NULL, UPDATE_WAIT); 183 } 184 return (error); 185 } 186 187 /* 188 * Vnode op for writing. 189 */ 190 int 191 ext2fs_write(v) 192 void *v; 193 { 194 struct vop_write_args /* { 195 struct vnode *a_vp; 196 struct uio *a_uio; 197 int a_ioflag; 198 struct ucred *a_cred; 199 } */ *ap = v; 200 struct vnode *vp; 201 struct uio *uio; 202 struct inode *ip; 203 struct m_ext2fs *fs; 204 struct buf *bp; 205 struct proc *p; 206 ufs_daddr_t lbn; 207 off_t osize; 208 int blkoffset, error, flags, ioflag, resid, xfersize; 209 vsize_t bytelen; 210 void *win; 211 off_t oldoff; 212 213 ioflag = ap->a_ioflag; 214 uio = ap->a_uio; 215 vp = ap->a_vp; 216 ip = VTOI(vp); 217 error = 0; 218 219 #ifdef DIAGNOSTIC 220 if (uio->uio_rw != UIO_WRITE) 221 panic("%s: mode", "ext2fs_write"); 222 #endif 223 224 switch (vp->v_type) { 225 case VREG: 226 if (ioflag & IO_APPEND) 227 uio->uio_offset = ip->i_e2fs_size; 228 if ((ip->i_e2fs_flags & EXT2_APPEND) && 229 uio->uio_offset != ip->i_e2fs_size) 230 return (EPERM); 231 /* FALLTHROUGH */ 232 case VLNK: 233 break; 234 case VDIR: 235 if ((ioflag & IO_SYNC) == 0) 236 panic("%s: nonsync dir write", "ext2fs_write"); 237 break; 238 default: 239 panic("%s: type", "ext2fs_write"); 240 } 241 242 fs = ip->i_e2fs; 243 if (uio->uio_offset < 0 || 244 (u_int64_t)uio->uio_offset + uio->uio_resid > 245 ((u_int64_t)0x80000000 * fs->e2fs_bsize - 1)) 246 return (EFBIG); 247 /* 248 * Maybe this should be above the vnode op call, but so long as 249 * file servers have no limits, I don't think it matters. 250 */ 251 p = uio->uio_procp; 252 if (vp->v_type == VREG && p && 253 uio->uio_offset + uio->uio_resid > 254 p->p_rlimit[RLIMIT_FSIZE].rlim_cur) { 255 psignal(p, SIGXFSZ); 256 return (EFBIG); 257 } 258 259 resid = uio->uio_resid; 260 osize = ip->i_e2fs_size; 261 262 if (vp->v_type == VREG) { 263 while (uio->uio_resid > 0) { 264 oldoff = uio->uio_offset; 265 blkoffset = blkoff(fs, uio->uio_offset); 266 bytelen = MIN(fs->e2fs_bsize - blkoffset, 267 uio->uio_resid); 268 269 error = ext2fs_balloc_range(vp, uio->uio_offset, 270 bytelen, ap->a_cred, 0); 271 if (error) { 272 break; 273 } 274 win = ubc_alloc(&vp->v_uobj, uio->uio_offset, 275 &bytelen, UBC_WRITE); 276 error = uiomove(win, bytelen, uio); 277 ubc_release(win, 0); 278 if (error) { 279 break; 280 } 281 282 /* 283 * flush what we just wrote if necessary. 284 * XXXUBC simplistic async flushing. 285 */ 286 287 if (oldoff >> 16 != uio->uio_offset >> 16) { 288 simple_lock(&vp->v_uobj.vmobjlock); 289 error = vp->v_uobj.pgops->pgo_put( 290 &vp->v_uobj, (oldoff >> 16) << 16, 291 (uio->uio_offset >> 16) << 16, PGO_CLEANIT); 292 } 293 } 294 if (error == 0 && ioflag & IO_SYNC) { 295 simple_lock(&vp->v_uobj.vmobjlock); 296 error = vp->v_uobj.pgops->pgo_put(&vp->v_uobj, oldoff, 297 oldoff + bytelen, PGO_CLEANIT|PGO_SYNCIO); 298 } 299 300 goto out; 301 } 302 303 flags = ioflag & IO_SYNC ? B_SYNC : 0; 304 for (error = 0; uio->uio_resid > 0;) { 305 lbn = lblkno(fs, uio->uio_offset); 306 blkoffset = blkoff(fs, uio->uio_offset); 307 xfersize = MIN(fs->e2fs_bsize - blkoffset, uio->uio_resid); 308 if (xfersize < fs->e2fs_bsize) 309 flags |= B_CLRBUF; 310 else 311 flags &= ~B_CLRBUF; 312 error = ext2fs_balloc(ip, 313 lbn, blkoffset + xfersize, ap->a_cred, &bp, flags); 314 if (error) 315 break; 316 if (ip->i_e2fs_size < uio->uio_offset + xfersize) { 317 ip->i_e2fs_size = uio->uio_offset + xfersize; 318 } 319 error = uiomove((char *)bp->b_data + blkoffset, xfersize, uio); 320 if (ioflag & IO_SYNC) 321 (void)bwrite(bp); 322 else if (xfersize + blkoffset == fs->e2fs_bsize) 323 bawrite(bp); 324 else 325 bdwrite(bp); 326 if (error || xfersize == 0) 327 break; 328 } 329 /* 330 * If we successfully wrote any data, and we are not the superuser 331 * we clear the setuid and setgid bits as a precaution against 332 * tampering. 333 */ 334 out: 335 ip->i_flag |= IN_CHANGE | IN_UPDATE; 336 if (resid > uio->uio_resid && ap->a_cred && ap->a_cred->cr_uid != 0) 337 ip->i_e2fs_mode &= ~(ISUID | ISGID); 338 if (error) { 339 (void) VOP_TRUNCATE(vp, osize, ioflag & IO_SYNC, ap->a_cred, 340 uio->uio_procp); 341 uio->uio_offset -= resid - uio->uio_resid; 342 uio->uio_resid = resid; 343 } else if (resid > uio->uio_resid && (ioflag & IO_SYNC) == IO_SYNC) 344 error = VOP_UPDATE(vp, NULL, NULL, UPDATE_WAIT); 345 KASSERT(vp->v_size == ip->i_e2fs_size); 346 return (error); 347 } 348