1 /* $NetBSD: ext2fs_readwrite.c,v 1.17 2001/09/15 20:36:41 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 bytelen = MIN(ip->i_e2fs_size - uio->uio_offset, 116 uio->uio_resid); 117 118 if (bytelen == 0) { 119 break; 120 } 121 win = ubc_alloc(&vp->v_uobj, uio->uio_offset, 122 &bytelen, UBC_READ); 123 error = uiomove(win, bytelen, uio); 124 ubc_release(win, 0); 125 if (error) { 126 break; 127 } 128 } 129 goto out; 130 } 131 132 for (error = 0, bp = NULL; uio->uio_resid > 0; bp = NULL) { 133 if ((bytesinfile = ip->i_e2fs_size - uio->uio_offset) <= 0) 134 break; 135 lbn = lblkno(fs, uio->uio_offset); 136 nextlbn = lbn + 1; 137 size = fs->e2fs_bsize; 138 blkoffset = blkoff(fs, uio->uio_offset); 139 xfersize = fs->e2fs_bsize - blkoffset; 140 if (uio->uio_resid < xfersize) 141 xfersize = uio->uio_resid; 142 if (bytesinfile < xfersize) 143 xfersize = bytesinfile; 144 145 if (lblktosize(fs, nextlbn) >= ip->i_e2fs_size) 146 error = bread(vp, lbn, size, NOCRED, &bp); 147 else { 148 int nextsize = fs->e2fs_bsize; 149 error = breadn(vp, lbn, 150 size, &nextlbn, &nextsize, 1, NOCRED, &bp); 151 } 152 if (error) 153 break; 154 155 /* 156 * We should only get non-zero b_resid when an I/O error 157 * has occurred, which should cause us to break above. 158 * However, if the short read did not cause an error, 159 * then we want to ensure that we do not uiomove bad 160 * or uninitialized data. 161 */ 162 size -= bp->b_resid; 163 if (size < xfersize) { 164 if (size == 0) 165 break; 166 xfersize = size; 167 } 168 error = uiomove((char *)bp->b_data + blkoffset, xfersize, uio); 169 if (error) 170 break; 171 brelse(bp); 172 } 173 if (bp != NULL) 174 brelse(bp); 175 176 out: 177 if (!(vp->v_mount->mnt_flag & MNT_NOATIME)) { 178 ip->i_flag |= IN_ACCESS; 179 if ((ap->a_ioflag & IO_SYNC) == IO_SYNC) 180 error = VOP_UPDATE(vp, NULL, NULL, UPDATE_WAIT); 181 } 182 return (error); 183 } 184 185 /* 186 * Vnode op for writing. 187 */ 188 int 189 ext2fs_write(v) 190 void *v; 191 { 192 struct vop_write_args /* { 193 struct vnode *a_vp; 194 struct uio *a_uio; 195 int a_ioflag; 196 struct ucred *a_cred; 197 } */ *ap = v; 198 struct vnode *vp; 199 struct uio *uio; 200 struct inode *ip; 201 struct m_ext2fs *fs; 202 struct buf *bp; 203 struct proc *p; 204 ufs_daddr_t lbn; 205 off_t osize; 206 int blkoffset, error, flags, ioflag, resid, xfersize; 207 vsize_t bytelen; 208 void *win; 209 off_t oldoff; 210 211 ioflag = ap->a_ioflag; 212 uio = ap->a_uio; 213 vp = ap->a_vp; 214 ip = VTOI(vp); 215 error = 0; 216 217 #ifdef DIAGNOSTIC 218 if (uio->uio_rw != UIO_WRITE) 219 panic("%s: mode", "ext2fs_write"); 220 #endif 221 222 switch (vp->v_type) { 223 case VREG: 224 if (ioflag & IO_APPEND) 225 uio->uio_offset = ip->i_e2fs_size; 226 if ((ip->i_e2fs_flags & EXT2_APPEND) && 227 uio->uio_offset != ip->i_e2fs_size) 228 return (EPERM); 229 /* FALLTHROUGH */ 230 case VLNK: 231 break; 232 case VDIR: 233 if ((ioflag & IO_SYNC) == 0) 234 panic("%s: nonsync dir write", "ext2fs_write"); 235 break; 236 default: 237 panic("%s: type", "ext2fs_write"); 238 } 239 240 fs = ip->i_e2fs; 241 if (uio->uio_offset < 0 || 242 (u_int64_t)uio->uio_offset + uio->uio_resid > 243 ((u_int64_t)0x80000000 * fs->e2fs_bsize - 1)) 244 return (EFBIG); 245 /* 246 * Maybe this should be above the vnode op call, but so long as 247 * file servers have no limits, I don't think it matters. 248 */ 249 p = uio->uio_procp; 250 if (vp->v_type == VREG && p && 251 uio->uio_offset + uio->uio_resid > 252 p->p_rlimit[RLIMIT_FSIZE].rlim_cur) { 253 psignal(p, SIGXFSZ); 254 return (EFBIG); 255 } 256 257 resid = uio->uio_resid; 258 osize = ip->i_e2fs_size; 259 260 if (vp->v_type == VREG) { 261 while (uio->uio_resid > 0) { 262 oldoff = uio->uio_offset; 263 blkoffset = blkoff(fs, uio->uio_offset); 264 bytelen = MIN(fs->e2fs_bsize - blkoffset, 265 uio->uio_resid); 266 267 error = ext2fs_balloc_range(vp, uio->uio_offset, 268 bytelen, ap->a_cred, 0); 269 if (error) { 270 break; 271 } 272 win = ubc_alloc(&vp->v_uobj, uio->uio_offset, 273 &bytelen, UBC_WRITE); 274 error = uiomove(win, bytelen, uio); 275 ubc_release(win, 0); 276 if (error) { 277 break; 278 } 279 280 /* 281 * flush what we just wrote if necessary. 282 * XXXUBC simplistic async flushing. 283 */ 284 285 if (oldoff >> 16 != uio->uio_offset >> 16) { 286 simple_lock(&vp->v_uobj.vmobjlock); 287 error = vp->v_uobj.pgops->pgo_put( 288 &vp->v_uobj, (oldoff >> 16) << 16, 289 (uio->uio_offset >> 16) << 16, PGO_CLEANIT); 290 } 291 } 292 if (error == 0 && ioflag & IO_SYNC) { 293 simple_lock(&vp->v_uobj.vmobjlock); 294 error = vp->v_uobj.pgops->pgo_put(&vp->v_uobj, oldoff, 295 oldoff + bytelen, PGO_CLEANIT|PGO_SYNCIO); 296 } 297 298 goto out; 299 } 300 301 flags = ioflag & IO_SYNC ? B_SYNC : 0; 302 for (error = 0; uio->uio_resid > 0;) { 303 lbn = lblkno(fs, uio->uio_offset); 304 blkoffset = blkoff(fs, uio->uio_offset); 305 xfersize = MIN(fs->e2fs_bsize - blkoffset, uio->uio_resid); 306 if (xfersize < fs->e2fs_bsize) 307 flags |= B_CLRBUF; 308 else 309 flags &= ~B_CLRBUF; 310 error = ext2fs_balloc(ip, 311 lbn, blkoffset + xfersize, ap->a_cred, &bp, flags); 312 if (error) 313 break; 314 if (ip->i_e2fs_size < uio->uio_offset + xfersize) { 315 ip->i_e2fs_size = uio->uio_offset + xfersize; 316 } 317 error = uiomove((char *)bp->b_data + blkoffset, xfersize, uio); 318 if (ioflag & IO_SYNC) 319 (void)bwrite(bp); 320 else if (xfersize + blkoffset == fs->e2fs_bsize) 321 bawrite(bp); 322 else 323 bdwrite(bp); 324 if (error || xfersize == 0) 325 break; 326 } 327 /* 328 * If we successfully wrote any data, and we are not the superuser 329 * we clear the setuid and setgid bits as a precaution against 330 * tampering. 331 */ 332 out: 333 ip->i_flag |= IN_CHANGE | IN_UPDATE; 334 if (resid > uio->uio_resid && ap->a_cred && ap->a_cred->cr_uid != 0) 335 ip->i_e2fs_mode &= ~(ISUID | ISGID); 336 if (error) { 337 (void) VOP_TRUNCATE(vp, osize, ioflag & IO_SYNC, ap->a_cred, 338 uio->uio_procp); 339 uio->uio_offset -= resid - uio->uio_resid; 340 uio->uio_resid = resid; 341 } else if (resid > uio->uio_resid && (ioflag & IO_SYNC) == IO_SYNC) 342 error = VOP_UPDATE(vp, NULL, NULL, UPDATE_WAIT); 343 KASSERT(vp->v_size == ip->i_e2fs_size); 344 return (error); 345 } 346