ext2fs_readwrite.c revision 1.19 1 /* $NetBSD: ext2fs_readwrite.c,v 1.19 2001/10/26 05:56:08 lukem 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/inode.h>
55 #include <ufs/ext2fs/ext2fs.h>
56 #include <ufs/ext2fs/ext2fs_extern.h>
57
58
59 #define doclusterread 0 /* XXX underway */
60 #define doclusterwrite 0
61
62 /*
63 * Vnode op for reading.
64 */
65 /* ARGSUSED */
66 int
67 ext2fs_read(v)
68 void *v;
69 {
70 struct vop_read_args /* {
71 struct vnode *a_vp;
72 struct uio *a_uio;
73 int a_ioflag;
74 struct ucred *a_cred;
75 } */ *ap = v;
76 struct vnode *vp;
77 struct inode *ip;
78 struct uio *uio;
79 struct m_ext2fs *fs;
80 struct buf *bp;
81 void *win;
82 vsize_t bytelen;
83 ufs_daddr_t lbn, nextlbn;
84 off_t bytesinfile;
85 long size, xfersize, blkoffset;
86 int error;
87
88 vp = ap->a_vp;
89 ip = VTOI(vp);
90 uio = ap->a_uio;
91
92 #ifdef DIAGNOSTIC
93 if (uio->uio_rw != UIO_READ)
94 panic("%s: mode", "ext2fs_read");
95
96 if (vp->v_type == VLNK) {
97 if ((int)ip->i_e2fs_size < vp->v_mount->mnt_maxsymlinklen ||
98 (vp->v_mount->mnt_maxsymlinklen == 0 &&
99 ip->i_e2fs_nblock == 0))
100 panic("%s: short symlink", "ext2fs_read");
101 } else if (vp->v_type != VREG && vp->v_type != VDIR)
102 panic("%s: type %d", "ext2fs_read", vp->v_type);
103 #endif
104 fs = ip->i_e2fs;
105 if ((u_int64_t)uio->uio_offset >
106 ((u_int64_t)0x80000000 * fs->e2fs_bsize - 1))
107 return (EFBIG);
108 if (uio->uio_resid == 0)
109 return (0);
110 if (uio->uio_offset >= ip->i_e2fs_size)
111 return (0);
112
113 if (vp->v_type == VREG) {
114 error = 0;
115 while (uio->uio_resid > 0) {
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_uobj, 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 {
149 int nextsize = fs->e2fs_bsize;
150 error = breadn(vp, lbn,
151 size, &nextlbn, &nextsize, 1, NOCRED, &bp);
152 }
153 if (error)
154 break;
155
156 /*
157 * We should only get non-zero b_resid when an I/O error
158 * has occurred, which should cause us to break above.
159 * However, if the short read did not cause an error,
160 * then we want to ensure that we do not uiomove bad
161 * or uninitialized data.
162 */
163 size -= bp->b_resid;
164 if (size < xfersize) {
165 if (size == 0)
166 break;
167 xfersize = size;
168 }
169 error = uiomove((char *)bp->b_data + blkoffset, xfersize, uio);
170 if (error)
171 break;
172 brelse(bp);
173 }
174 if (bp != NULL)
175 brelse(bp);
176
177 out:
178 if (!(vp->v_mount->mnt_flag & MNT_NOATIME)) {
179 ip->i_flag |= IN_ACCESS;
180 if ((ap->a_ioflag & IO_SYNC) == IO_SYNC)
181 error = VOP_UPDATE(vp, NULL, NULL, UPDATE_WAIT);
182 }
183 return (error);
184 }
185
186 /*
187 * Vnode op for writing.
188 */
189 int
190 ext2fs_write(v)
191 void *v;
192 {
193 struct vop_write_args /* {
194 struct vnode *a_vp;
195 struct uio *a_uio;
196 int a_ioflag;
197 struct ucred *a_cred;
198 } */ *ap = v;
199 struct vnode *vp;
200 struct uio *uio;
201 struct inode *ip;
202 struct m_ext2fs *fs;
203 struct buf *bp;
204 struct proc *p;
205 ufs_daddr_t lbn;
206 off_t osize;
207 int blkoffset, error, flags, ioflag, resid, xfersize;
208 vsize_t bytelen;
209 void *win;
210 off_t oldoff;
211
212 ioflag = ap->a_ioflag;
213 uio = ap->a_uio;
214 vp = ap->a_vp;
215 ip = VTOI(vp);
216 error = 0;
217
218 #ifdef DIAGNOSTIC
219 if (uio->uio_rw != UIO_WRITE)
220 panic("%s: mode", "ext2fs_write");
221 #endif
222
223 switch (vp->v_type) {
224 case VREG:
225 if (ioflag & IO_APPEND)
226 uio->uio_offset = ip->i_e2fs_size;
227 if ((ip->i_e2fs_flags & EXT2_APPEND) &&
228 uio->uio_offset != ip->i_e2fs_size)
229 return (EPERM);
230 /* FALLTHROUGH */
231 case VLNK:
232 break;
233 case VDIR:
234 if ((ioflag & IO_SYNC) == 0)
235 panic("%s: nonsync dir write", "ext2fs_write");
236 break;
237 default:
238 panic("%s: type", "ext2fs_write");
239 }
240
241 fs = ip->i_e2fs;
242 if (uio->uio_offset < 0 ||
243 (u_int64_t)uio->uio_offset + uio->uio_resid >
244 ((u_int64_t)0x80000000 * fs->e2fs_bsize - 1))
245 return (EFBIG);
246 /*
247 * Maybe this should be above the vnode op call, but so long as
248 * file servers have no limits, I don't think it matters.
249 */
250 p = uio->uio_procp;
251 if (vp->v_type == VREG && p &&
252 uio->uio_offset + uio->uio_resid >
253 p->p_rlimit[RLIMIT_FSIZE].rlim_cur) {
254 psignal(p, SIGXFSZ);
255 return (EFBIG);
256 }
257
258 resid = uio->uio_resid;
259 osize = ip->i_e2fs_size;
260
261 if (vp->v_type == VREG) {
262 while (uio->uio_resid > 0) {
263 oldoff = uio->uio_offset;
264 blkoffset = blkoff(fs, uio->uio_offset);
265 bytelen = MIN(fs->e2fs_bsize - blkoffset,
266 uio->uio_resid);
267
268 error = ext2fs_balloc_range(vp, uio->uio_offset,
269 bytelen, ap->a_cred, 0);
270 if (error) {
271 break;
272 }
273 win = ubc_alloc(&vp->v_uobj, uio->uio_offset,
274 &bytelen, UBC_WRITE);
275 error = uiomove(win, bytelen, uio);
276 ubc_release(win, 0);
277 if (error) {
278 break;
279 }
280
281 /*
282 * flush what we just wrote if necessary.
283 * XXXUBC simplistic async flushing.
284 */
285
286 if (oldoff >> 16 != uio->uio_offset >> 16) {
287 simple_lock(&vp->v_uobj.vmobjlock);
288 error = vp->v_uobj.pgops->pgo_put(
289 &vp->v_uobj, (oldoff >> 16) << 16,
290 (uio->uio_offset >> 16) << 16, PGO_CLEANIT);
291 }
292 }
293 if (error == 0 && ioflag & IO_SYNC) {
294 simple_lock(&vp->v_uobj.vmobjlock);
295 error = vp->v_uobj.pgops->pgo_put(&vp->v_uobj, oldoff,
296 oldoff + bytelen, PGO_CLEANIT|PGO_SYNCIO);
297 }
298
299 goto out;
300 }
301
302 flags = ioflag & IO_SYNC ? B_SYNC : 0;
303 for (error = 0; uio->uio_resid > 0;) {
304 lbn = lblkno(fs, uio->uio_offset);
305 blkoffset = blkoff(fs, uio->uio_offset);
306 xfersize = MIN(fs->e2fs_bsize - blkoffset, uio->uio_resid);
307 if (xfersize < fs->e2fs_bsize)
308 flags |= B_CLRBUF;
309 else
310 flags &= ~B_CLRBUF;
311 error = ext2fs_balloc(ip,
312 lbn, blkoffset + xfersize, ap->a_cred, &bp, flags);
313 if (error)
314 break;
315 if (ip->i_e2fs_size < uio->uio_offset + xfersize) {
316 ip->i_e2fs_size = uio->uio_offset + xfersize;
317 }
318 error = uiomove((char *)bp->b_data + blkoffset, xfersize, uio);
319 if (ioflag & IO_SYNC)
320 (void)bwrite(bp);
321 else if (xfersize + blkoffset == fs->e2fs_bsize)
322 bawrite(bp);
323 else
324 bdwrite(bp);
325 if (error || xfersize == 0)
326 break;
327 }
328 /*
329 * If we successfully wrote any data, and we are not the superuser
330 * we clear the setuid and setgid bits as a precaution against
331 * tampering.
332 */
333 out:
334 ip->i_flag |= IN_CHANGE | IN_UPDATE;
335 if (resid > uio->uio_resid && ap->a_cred && ap->a_cred->cr_uid != 0)
336 ip->i_e2fs_mode &= ~(ISUID | ISGID);
337 if (error) {
338 (void) VOP_TRUNCATE(vp, osize, ioflag & IO_SYNC, ap->a_cred,
339 uio->uio_procp);
340 uio->uio_offset -= resid - uio->uio_resid;
341 uio->uio_resid = resid;
342 } else if (resid > uio->uio_resid && (ioflag & IO_SYNC) == IO_SYNC)
343 error = VOP_UPDATE(vp, NULL, NULL, UPDATE_WAIT);
344 KASSERT(vp->v_size == ip->i_e2fs_size);
345 return (error);
346 }
347