xref: /freebsd/stand/libsa/gzipfs.c (revision f126890a)
1 /*
2  * Copyright (c) 1998 Michael Smith.
3  * All rights reserved.
4  *
5  * Redistribution and use in source and binary forms, with or without
6  * modification, are permitted provided that the following conditions
7  * are met:
8  * 1. Redistributions of source code must retain the above copyright
9  *    notice, this list of conditions and the following disclaimer.
10  * 2. Redistributions in binary form must reproduce the above copyright
11  *    notice, this list of conditions and the following disclaimer in the
12  *    documentation and/or other materials provided with the distribution.
13  *
14  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
15  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
16  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
17  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
18  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
19  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
20  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
21  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
22  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
23  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
24  * SUCH DAMAGE.
25  */
26 
27 #include "stand.h"
28 
29 #include <sys/stat.h>
30 #include <string.h>
31 #include <zlib.h>
32 
33 #define Z_BUFSIZE 2048	/* XXX larger? */
34 
35 struct z_file
36 {
37     int			zf_rawfd;
38     off_t		zf_dataoffset;
39     z_stream		zf_zstream;
40     unsigned char	zf_buf[Z_BUFSIZE];
41     int			zf_endseen;
42 };
43 
44 static int	zf_fill(struct z_file *z);
45 static int	zf_open(const char *path, struct open_file *f);
46 static int	zf_close(struct open_file *f);
47 static int	zf_read(struct open_file *f, void *buf, size_t size, size_t *resid);
48 static off_t	zf_seek(struct open_file *f, off_t offset, int where);
49 static int	zf_stat(struct open_file *f, struct stat *sb);
50 
51 struct fs_ops gzipfs_fsops = {
52 	.fs_name = "zip",
53 	.fo_open = zf_open,
54 	.fo_close = zf_close,
55 	.fo_read = zf_read,
56 	.fo_write = null_write,
57 	.fo_seek = zf_seek,
58 	.fo_stat = zf_stat,
59 	.fo_readdir = null_readdir,
60 };
61 
62 static int
63 zf_fill(struct z_file *zf)
64 {
65     int		result;
66     int		req;
67 
68     req = Z_BUFSIZE - zf->zf_zstream.avail_in;
69     result = 0;
70 
71     /* If we need more */
72     if (req > 0) {
73 	/* move old data to bottom of buffer */
74 	if (req < Z_BUFSIZE)
75 	    bcopy(zf->zf_buf + req, zf->zf_buf, Z_BUFSIZE - req);
76 
77 	/* read to fill buffer and update availibility data */
78 	result = read(zf->zf_rawfd, zf->zf_buf + zf->zf_zstream.avail_in, req);
79 	zf->zf_zstream.next_in = zf->zf_buf;
80 	if (result >= 0)
81 	    zf->zf_zstream.avail_in += result;
82     }
83     return(result);
84 }
85 
86 /*
87  * Adapted from get_byte/check_header in libz
88  *
89  * Returns 0 if the header is OK, nonzero if not.
90  */
91 static int
92 get_byte(struct z_file *zf, off_t *curoffp)
93 {
94     if ((zf->zf_zstream.avail_in == 0) && (zf_fill(zf) == -1))
95 	return(-1);
96     zf->zf_zstream.avail_in--;
97     ++*curoffp;
98     return(*(zf->zf_zstream.next_in)++);
99 }
100 
101 static int gz_magic[2] = {0x1f, 0x8b}; /* gzip magic header */
102 
103 /* gzip flag byte */
104 #define ASCII_FLAG	0x01 /* bit 0 set: file probably ascii text */
105 #define HEAD_CRC	0x02 /* bit 1 set: header CRC present */
106 #define EXTRA_FIELD	0x04 /* bit 2 set: extra field present */
107 #define ORIG_NAME	0x08 /* bit 3 set: original file name present */
108 #define COMMENT		0x10 /* bit 4 set: file comment present */
109 #define RESERVED	0xE0 /* bits 5..7: reserved */
110 
111 static int
112 check_header(struct z_file *zf)
113 {
114     int		method; /* method byte */
115     int		flags;  /* flags byte */
116     uInt	len;
117     int		c;
118 
119     zf->zf_dataoffset = 0;
120     /* Check the gzip magic header */
121     for (len = 0; len < 2; len++) {
122 	c = get_byte(zf, &zf->zf_dataoffset);
123 	if (c != gz_magic[len]) {
124 	    return(1);
125 	}
126     }
127     method = get_byte(zf, &zf->zf_dataoffset);
128     flags = get_byte(zf, &zf->zf_dataoffset);
129     if (method != Z_DEFLATED || (flags & RESERVED) != 0) {
130 	return(1);
131     }
132 
133     /* Discard time, xflags and OS code: */
134     for (len = 0; len < 6; len++) (void)get_byte(zf, &zf->zf_dataoffset);
135 
136     if ((flags & EXTRA_FIELD) != 0) { /* skip the extra field */
137 	len  =  (uInt)get_byte(zf, &zf->zf_dataoffset);
138 	len += ((uInt)get_byte(zf, &zf->zf_dataoffset))<<8;
139 	/* len is garbage if EOF but the loop below will quit anyway */
140 	while (len-- != 0 && get_byte(zf, &zf->zf_dataoffset) != -1) ;
141     }
142     if ((flags & ORIG_NAME) != 0) { /* skip the original file name */
143 	while ((c = get_byte(zf, &zf->zf_dataoffset)) != 0 && c != -1) ;
144     }
145     if ((flags & COMMENT) != 0) {   /* skip the .gz file comment */
146 	while ((c = get_byte(zf, &zf->zf_dataoffset)) != 0 && c != -1) ;
147     }
148     if ((flags & HEAD_CRC) != 0) {  /* skip the header crc */
149 	for (len = 0; len < 2; len++) c = get_byte(zf, &zf->zf_dataoffset);
150     }
151     /* if there's data left, we're in business */
152     return((c == -1) ? 1 : 0);
153 }
154 
155 static int
156 zf_open(const char *fname, struct open_file *f)
157 {
158     static char		*zfname;
159     int			rawfd;
160     struct z_file	*zf;
161     char		*cp;
162     int			error;
163     struct stat		sb;
164 
165     /* Have to be in "just read it" mode */
166     if (f->f_flags != F_READ)
167 	return(EPERM);
168 
169     /* If the name already ends in .gz or .bz2, ignore it */
170     if ((cp = strrchr(fname, '.')) && (!strcmp(cp, ".gz")
171 	    || !strcmp(cp, ".bz2") || !strcmp(cp, ".split")))
172 	return(ENOENT);
173 
174     /* Construct new name */
175     zfname = malloc(strlen(fname) + 4);
176     if (zfname == NULL)
177         return(ENOMEM);
178     sprintf(zfname, "%s.gz", fname);
179 
180     /* Try to open the compressed datafile */
181     rawfd = open(zfname, O_RDONLY);
182     free(zfname);
183     if (rawfd == -1)
184 	return(ENOENT);
185 
186     if (fstat(rawfd, &sb) < 0) {
187 	printf("zf_open: stat failed\n");
188 	close(rawfd);
189 	return(ENOENT);
190     }
191     if (!S_ISREG(sb.st_mode)) {
192 	printf("zf_open: not a file\n");
193 	close(rawfd);
194 	return(EISDIR);			/* best guess */
195     }
196 
197     /* Allocate a z_file structure, populate it */
198     zf = malloc(sizeof(struct z_file));
199     if (zf == NULL)
200         return(ENOMEM);
201     bzero(zf, sizeof(struct z_file));
202     zf->zf_rawfd = rawfd;
203 
204     /* Verify that the file is gzipped */
205     if (check_header(zf)) {
206 	close(zf->zf_rawfd);
207 	free(zf);
208 	return(EFTYPE);
209     }
210 
211     /* Initialise the inflation engine */
212     if ((error = inflateInit2(&(zf->zf_zstream), -15)) != Z_OK) {
213 	printf("zf_open: inflateInit returned %d : %s\n", error, zf->zf_zstream.msg);
214 	close(zf->zf_rawfd);
215 	free(zf);
216 	return(EIO);
217     }
218 
219     /* Looks OK, we'll take it */
220     f->f_fsdata = zf;
221     return(0);
222 }
223 
224 static int
225 zf_close(struct open_file *f)
226 {
227     struct z_file	*zf = (struct z_file *)f->f_fsdata;
228 
229     inflateEnd(&(zf->zf_zstream));
230     close(zf->zf_rawfd);
231     free(zf);
232     return(0);
233 }
234 
235 static int
236 zf_read(struct open_file *f, void *buf, size_t size, size_t *resid)
237 {
238     struct z_file	*zf = (struct z_file *)f->f_fsdata;
239     int			error;
240 
241     zf->zf_zstream.next_out = buf;			/* where and how much */
242     zf->zf_zstream.avail_out = size;
243 
244     while (zf->zf_zstream.avail_out && zf->zf_endseen == 0) {
245 	if ((zf->zf_zstream.avail_in == 0) && (zf_fill(zf) == -1)) {
246 	    printf("zf_read: fill error\n");
247 	    return(EIO);
248 	}
249 	if (zf->zf_zstream.avail_in == 0) {		/* oops, unexpected EOF */
250 	    printf("zf_read: unexpected EOF\n");
251 	    if (zf->zf_zstream.avail_out == size)
252 		return(EIO);
253 	    break;
254 	}
255 
256 	error = inflate(&zf->zf_zstream, Z_SYNC_FLUSH);	/* decompression pass */
257 	if (error == Z_STREAM_END) {			/* EOF, all done */
258 	    zf->zf_endseen = 1;
259 	    break;
260 	}
261 	if (error != Z_OK) {				/* argh, decompression error */
262 	    printf("inflate: %s\n", zf->zf_zstream.msg);
263 	    return(EIO);
264 	}
265     }
266     if (resid != NULL)
267 	*resid = zf->zf_zstream.avail_out;
268     return(0);
269 }
270 
271 static int
272 zf_rewind(struct open_file *f)
273 {
274     struct z_file	*zf = (struct z_file *)f->f_fsdata;
275 
276     if (lseek(zf->zf_rawfd, zf->zf_dataoffset, SEEK_SET) == -1)
277 	return(-1);
278     zf->zf_zstream.avail_in = 0;
279     zf->zf_zstream.next_in = NULL;
280     zf->zf_endseen = 0;
281     (void)inflateReset(&zf->zf_zstream);
282 
283     return(0);
284 }
285 
286 static off_t
287 zf_seek(struct open_file *f, off_t offset, int where)
288 {
289     struct z_file	*zf = (struct z_file *)f->f_fsdata;
290     off_t		target;
291     char		discard[16];
292 
293     switch (where) {
294     case SEEK_SET:
295 	target = offset;
296 	break;
297     case SEEK_CUR:
298 	target = offset + zf->zf_zstream.total_out;
299 	break;
300     default:
301 	errno = EINVAL;
302 	return(-1);
303     }
304 
305     /* rewind if required */
306     if (target < zf->zf_zstream.total_out && zf_rewind(f) != 0)
307 	return(-1);
308 
309     /* skip forwards if required */
310     while (target > zf->zf_zstream.total_out) {
311 	errno = zf_read(f, discard, min(sizeof(discard),
312 	    target - zf->zf_zstream.total_out), NULL);
313 	if (errno)
314 	    return(-1);
315 	/* Break out of loop if end of file has been reached. */
316 	if (zf->zf_endseen)
317 	    break;
318     }
319     /* This is where we are (be honest if we overshot) */
320     return(zf->zf_zstream.total_out);
321 }
322 
323 
324 static int
325 zf_stat(struct open_file *f, struct stat *sb)
326 {
327     struct z_file	*zf = (struct z_file *)f->f_fsdata;
328     int			result;
329 
330     /* stat as normal, but indicate that size is unknown */
331     if ((result = fstat(zf->zf_rawfd, sb)) == 0)
332 	sb->st_size = -1;
333     return(result);
334 }
335 
336 
337 
338