ext2fs_readwrite.c revision 1.8 1 /* $NetBSD: ext2fs_readwrite.c,v 1.8 1998/09/29 10:24:58 bouyer 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 #if defined(_KERNEL) && !defined(_LKM)
41 #include "opt_uvm.h"
42 #endif
43
44 #include <sys/param.h>
45 #include <sys/systm.h>
46 #include <sys/resourcevar.h>
47 #include <sys/kernel.h>
48 #include <sys/file.h>
49 #include <sys/stat.h>
50 #include <sys/buf.h>
51 #include <sys/proc.h>
52 #include <sys/conf.h>
53 #include <sys/mount.h>
54 #include <sys/vnode.h>
55 #include <sys/malloc.h>
56 #include <sys/signalvar.h>
57
58 #include <vm/vm.h>
59 #if defined(UVM)
60 #include <uvm/uvm_extern.h>
61 #endif
62
63 #include <ufs/ufs/quota.h>
64 #include <ufs/ufs/inode.h>
65 #include <ufs/ext2fs/ext2fs.h>
66 #include <ufs/ext2fs/ext2fs_extern.h>
67
68
69 #define doclusterread 0 /* XXX underway */
70 #define doclusterwrite 0
71
72 /*
73 * Vnode op for reading.
74 */
75 /* ARGSUSED */
76 int
77 ext2fs_read(v)
78 void *v;
79 {
80 struct vop_read_args /* {
81 struct vnode *a_vp;
82 struct uio *a_uio;
83 int a_ioflag;
84 struct ucred *a_cred;
85 } */ *ap = v;
86 register struct vnode *vp;
87 register struct inode *ip;
88 register struct uio *uio;
89 register struct m_ext2fs *fs;
90 struct buf *bp;
91 struct timespec ts;
92 ufs_daddr_t lbn, nextlbn;
93 off_t bytesinfile;
94 long size, xfersize, blkoffset;
95 int error;
96
97 vp = ap->a_vp;
98 ip = VTOI(vp);
99 uio = ap->a_uio;
100
101 #ifdef DIAGNOSTIC
102 if (uio->uio_rw != UIO_READ)
103 panic("%s: mode", "ext2fs_read");
104
105 if (vp->v_type == VLNK) {
106 if ((int)ip->i_e2fs_size < vp->v_mount->mnt_maxsymlinklen ||
107 (vp->v_mount->mnt_maxsymlinklen == 0 &&
108 ip->i_e2fs_nblock == 0))
109 panic("%s: short symlink", "ext2fs_read");
110 } else if (vp->v_type != VREG && vp->v_type != VDIR)
111 panic("%s: type %d", "ext2fs_read", vp->v_type);
112 #endif
113 fs = ip->i_e2fs;
114 if ((u_int64_t)uio->uio_offset >
115 ((u_int64_t)0x80000000 * fs->e2fs_bsize - 1))
116 return (EFBIG);
117 if (uio->uio_resid == 0)
118 return (0);
119
120 for (error = 0, bp = NULL; uio->uio_resid > 0; bp = NULL) {
121 if ((bytesinfile = ip->i_e2fs_size - uio->uio_offset) <= 0)
122 break;
123 lbn = lblkno(fs, uio->uio_offset);
124 nextlbn = lbn + 1;
125 size = fs->e2fs_bsize;
126 blkoffset = blkoff(fs, uio->uio_offset);
127 xfersize = fs->e2fs_bsize - blkoffset;
128 if (uio->uio_resid < xfersize)
129 xfersize = uio->uio_resid;
130 if (bytesinfile < xfersize)
131 xfersize = bytesinfile;
132
133 if (lblktosize(fs, nextlbn) >= ip->i_e2fs_size)
134 error = bread(vp, lbn, size, NOCRED, &bp);
135 else if (doclusterread)
136 error = cluster_read(vp,
137 ip->i_e2fs_size, lbn, size, NOCRED, &bp);
138 else if (lbn - 1 == vp->v_lastr) {
139 int nextsize = fs->e2fs_bsize;
140 error = breadn(vp, lbn,
141 size, &nextlbn, &nextsize, 1, NOCRED, &bp);
142 } else
143 error = bread(vp, lbn, size, NOCRED, &bp);
144 if (error)
145 break;
146 vp->v_lastr = lbn;
147
148 /*
149 * We should only get non-zero b_resid when an I/O error
150 * has occurred, which should cause us to break above.
151 * However, if the short read did not cause an error,
152 * then we want to ensure that we do not uiomove bad
153 * or uninitialized data.
154 */
155 size -= bp->b_resid;
156 if (size < xfersize) {
157 if (size == 0)
158 break;
159 xfersize = size;
160 }
161 error = uiomove((char *)bp->b_data + blkoffset, (int)xfersize,
162 uio);
163 if (error)
164 break;
165 brelse(bp);
166 }
167 if (bp != NULL)
168 brelse(bp);
169 if (!(vp->v_mount->mnt_flag & MNT_NOATIME)) {
170 ip->i_flag |= IN_ACCESS;
171 if ((ap->a_ioflag & IO_SYNC) == IO_SYNC) {
172 TIMEVAL_TO_TIMESPEC(&time, &ts);
173 error = VOP_UPDATE(vp, &ts, &ts, 1);
174 }
175 }
176 return (error);
177 }
178
179 /*
180 * Vnode op for writing.
181 */
182 int
183 ext2fs_write(v)
184 void *v;
185 {
186 struct vop_write_args /* {
187 struct vnode *a_vp;
188 struct uio *a_uio;
189 int a_ioflag;
190 struct ucred *a_cred;
191 } */ *ap = v;
192 register struct vnode *vp;
193 register struct uio *uio;
194 register struct inode *ip;
195 register struct m_ext2fs *fs;
196 struct buf *bp;
197 struct proc *p;
198 ufs_daddr_t lbn;
199 off_t osize;
200 int blkoffset, error, flags, ioflag, resid, size, xfersize;
201 struct timespec ts;
202
203 ioflag = ap->a_ioflag;
204 uio = ap->a_uio;
205 vp = ap->a_vp;
206 ip = VTOI(vp);
207
208 #ifdef DIAGNOSTIC
209 if (uio->uio_rw != UIO_WRITE)
210 panic("%s: mode", "ext2fs_write");
211 #endif
212
213 switch (vp->v_type) {
214 case VREG:
215 if (ioflag & IO_APPEND)
216 uio->uio_offset = ip->i_e2fs_size;
217 if ((ip->i_e2fs_flags & EXT2_APPEND) &&
218 uio->uio_offset != ip->i_e2fs_size)
219 return (EPERM);
220 /* FALLTHROUGH */
221 case VLNK:
222 break;
223 case VDIR:
224 if ((ioflag & IO_SYNC) == 0)
225 panic("%s: nonsync dir write", "ext2fs_write");
226 break;
227 default:
228 panic("%s: type", "ext2fs_write");
229 }
230
231 fs = ip->i_e2fs;
232 if (uio->uio_offset < 0 ||
233 (u_int64_t)uio->uio_offset + uio->uio_resid >
234 ((u_int64_t)0x80000000 * fs->e2fs_bsize - 1))
235 return (EFBIG);
236 /*
237 * Maybe this should be above the vnode op call, but so long as
238 * file servers have no limits, I don't think it matters.
239 */
240 p = uio->uio_procp;
241 if (vp->v_type == VREG && p &&
242 uio->uio_offset + uio->uio_resid >
243 p->p_rlimit[RLIMIT_FSIZE].rlim_cur) {
244 psignal(p, SIGXFSZ);
245 return (EFBIG);
246 }
247
248 resid = uio->uio_resid;
249 osize = ip->i_e2fs_size;
250 flags = ioflag & IO_SYNC ? B_SYNC : 0;
251
252 for (error = 0; uio->uio_resid > 0;) {
253 lbn = lblkno(fs, uio->uio_offset);
254 blkoffset = blkoff(fs, uio->uio_offset);
255 xfersize = fs->e2fs_bsize - blkoffset;
256 if (uio->uio_resid < xfersize)
257 xfersize = uio->uio_resid;
258 if (fs->e2fs_bsize > xfersize)
259 flags |= B_CLRBUF;
260 else
261 flags &= ~B_CLRBUF;
262
263 error = ext2fs_balloc(ip,
264 lbn, blkoffset + xfersize, ap->a_cred, &bp, flags);
265 if (error)
266 break;
267 if (uio->uio_offset + xfersize > ip->i_e2fs_size) {
268 ip->i_e2fs_size = uio->uio_offset + xfersize;
269 #if defined(UVM)
270 uvm_vnp_setsize(vp, ip->i_e2fs_size);
271 #else
272 vnode_pager_setsize(vp, ip->i_e2fs_size);
273 #endif
274 }
275 #if defined(UVM)
276 (void)uvm_vnp_uncache(vp);
277 #else
278 (void)vnode_pager_uncache(vp);
279 #endif
280
281 size = fs->e2fs_bsize - bp->b_resid;
282 if (size < xfersize)
283 xfersize = size;
284
285 error =
286 uiomove((char *)bp->b_data + blkoffset, (int)xfersize, uio);
287 if (ioflag & IO_SYNC)
288 (void)bwrite(bp);
289 else if (xfersize + blkoffset == fs->e2fs_bsize)
290 if (doclusterwrite)
291 cluster_write(bp, ip->i_e2fs_size);
292 else
293 bawrite(bp);
294 else
295 bdwrite(bp);
296 if (error || xfersize == 0)
297 break;
298 ip->i_flag |= IN_CHANGE | IN_UPDATE;
299 }
300 /*
301 * If we successfully wrote any data, and we are not the superuser
302 * we clear the setuid and setgid bits as a precaution against
303 * tampering.
304 */
305 if (resid > uio->uio_resid && ap->a_cred && ap->a_cred->cr_uid != 0)
306 ip->i_e2fs_mode &= ~(ISUID | ISGID);
307 if (error) {
308 if (ioflag & IO_UNIT) {
309 (void)VOP_TRUNCATE(vp, osize,
310 ioflag & IO_SYNC, ap->a_cred, uio->uio_procp);
311 uio->uio_offset -= resid - uio->uio_resid;
312 uio->uio_resid = resid;
313 }
314 } else if (resid > uio->uio_resid && (ioflag & IO_SYNC) == IO_SYNC) {
315 TIMEVAL_TO_TIMESPEC(&time, &ts);
316 error = VOP_UPDATE(vp, &ts, &ts, 1);
317 }
318 return (error);
319 }
320