xref: /freebsd/stand/libsa/gzipfs.c (revision cddbc3b40812213ff00041f79174cac0be360a2a)
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 <sys/cdefs.h>
28 __FBSDID("$FreeBSD$");
29 
30 #include "stand.h"
31 
32 #include <sys/stat.h>
33 #include <string.h>
34 #include <zlib.h>
35 
36 #define Z_BUFSIZE 2048	/* XXX larger? */
37 
38 struct z_file
39 {
40     int			zf_rawfd;
41     off_t		zf_dataoffset;
42     z_stream		zf_zstream;
43     unsigned char	zf_buf[Z_BUFSIZE];
44     int			zf_endseen;
45 };
46 
47 static int	zf_fill(struct z_file *z);
48 static int	zf_open(const char *path, struct open_file *f);
49 static int	zf_close(struct open_file *f);
50 static int	zf_read(struct open_file *f, void *buf, size_t size, size_t *resid);
51 static off_t	zf_seek(struct open_file *f, off_t offset, int where);
52 static int	zf_stat(struct open_file *f, struct stat *sb);
53 
54 struct fs_ops gzipfs_fsops = {
55     "zip",
56     zf_open,
57     zf_close,
58     zf_read,
59     null_write,
60     zf_seek,
61     zf_stat,
62     null_readdir
63 };
64 
65 static int
66 zf_fill(struct z_file *zf)
67 {
68     int		result;
69     int		req;
70 
71     req = Z_BUFSIZE - zf->zf_zstream.avail_in;
72     result = 0;
73 
74     /* If we need more */
75     if (req > 0) {
76 	/* move old data to bottom of buffer */
77 	if (req < Z_BUFSIZE)
78 	    bcopy(zf->zf_buf + req, zf->zf_buf, Z_BUFSIZE - req);
79 
80 	/* read to fill buffer and update availibility data */
81 	result = read(zf->zf_rawfd, zf->zf_buf + zf->zf_zstream.avail_in, req);
82 	zf->zf_zstream.next_in = zf->zf_buf;
83 	if (result >= 0)
84 	    zf->zf_zstream.avail_in += result;
85     }
86     return(result);
87 }
88 
89 /*
90  * Adapted from get_byte/check_header in libz
91  *
92  * Returns 0 if the header is OK, nonzero if not.
93  */
94 static int
95 get_byte(struct z_file *zf, off_t *curoffp)
96 {
97     if ((zf->zf_zstream.avail_in == 0) && (zf_fill(zf) == -1))
98 	return(-1);
99     zf->zf_zstream.avail_in--;
100     ++*curoffp;
101     return(*(zf->zf_zstream.next_in)++);
102 }
103 
104 static int gz_magic[2] = {0x1f, 0x8b}; /* gzip magic header */
105 
106 /* gzip flag byte */
107 #define ASCII_FLAG	0x01 /* bit 0 set: file probably ascii text */
108 #define HEAD_CRC	0x02 /* bit 1 set: header CRC present */
109 #define EXTRA_FIELD	0x04 /* bit 2 set: extra field present */
110 #define ORIG_NAME	0x08 /* bit 3 set: original file name present */
111 #define COMMENT		0x10 /* bit 4 set: file comment present */
112 #define RESERVED	0xE0 /* bits 5..7: reserved */
113 
114 static int
115 check_header(struct z_file *zf)
116 {
117     int		method; /* method byte */
118     int		flags;  /* flags byte */
119     uInt	len;
120     int		c;
121 
122     zf->zf_dataoffset = 0;
123     /* Check the gzip magic header */
124     for (len = 0; len < 2; len++) {
125 	c = get_byte(zf, &zf->zf_dataoffset);
126 	if (c != gz_magic[len]) {
127 	    return(1);
128 	}
129     }
130     method = get_byte(zf, &zf->zf_dataoffset);
131     flags = get_byte(zf, &zf->zf_dataoffset);
132     if (method != Z_DEFLATED || (flags & RESERVED) != 0) {
133 	return(1);
134     }
135 
136     /* Discard time, xflags and OS code: */
137     for (len = 0; len < 6; len++) (void)get_byte(zf, &zf->zf_dataoffset);
138 
139     if ((flags & EXTRA_FIELD) != 0) { /* skip the extra field */
140 	len  =  (uInt)get_byte(zf, &zf->zf_dataoffset);
141 	len += ((uInt)get_byte(zf, &zf->zf_dataoffset))<<8;
142 	/* len is garbage if EOF but the loop below will quit anyway */
143 	while (len-- != 0 && get_byte(zf, &zf->zf_dataoffset) != -1) ;
144     }
145     if ((flags & ORIG_NAME) != 0) { /* skip the original file name */
146 	while ((c = get_byte(zf, &zf->zf_dataoffset)) != 0 && c != -1) ;
147     }
148     if ((flags & COMMENT) != 0) {   /* skip the .gz file comment */
149 	while ((c = get_byte(zf, &zf->zf_dataoffset)) != 0 && c != -1) ;
150     }
151     if ((flags & HEAD_CRC) != 0) {  /* skip the header crc */
152 	for (len = 0; len < 2; len++) c = get_byte(zf, &zf->zf_dataoffset);
153     }
154     /* if there's data left, we're in business */
155     return((c == -1) ? 1 : 0);
156 }
157 
158 static int
159 zf_open(const char *fname, struct open_file *f)
160 {
161     static char		*zfname;
162     int			rawfd;
163     struct z_file	*zf;
164     char		*cp;
165     int			error;
166     struct stat		sb;
167 
168     /* Have to be in "just read it" mode */
169     if (f->f_flags != F_READ)
170 	return(EPERM);
171 
172     /* If the name already ends in .gz or .bz2, ignore it */
173     if ((cp = strrchr(fname, '.')) && (!strcmp(cp, ".gz")
174 	    || !strcmp(cp, ".bz2") || !strcmp(cp, ".split")))
175 	return(ENOENT);
176 
177     /* Construct new name */
178     zfname = malloc(strlen(fname) + 4);
179     if (zfname == NULL)
180         return(ENOMEM);
181     sprintf(zfname, "%s.gz", fname);
182 
183     /* Try to open the compressed datafile */
184     rawfd = open(zfname, O_RDONLY);
185     free(zfname);
186     if (rawfd == -1)
187 	return(ENOENT);
188 
189     if (fstat(rawfd, &sb) < 0) {
190 	printf("zf_open: stat failed\n");
191 	close(rawfd);
192 	return(ENOENT);
193     }
194     if (!S_ISREG(sb.st_mode)) {
195 	printf("zf_open: not a file\n");
196 	close(rawfd);
197 	return(EISDIR);			/* best guess */
198     }
199 
200     /* Allocate a z_file structure, populate it */
201     zf = malloc(sizeof(struct z_file));
202     if (zf == NULL)
203         return(ENOMEM);
204     bzero(zf, sizeof(struct z_file));
205     zf->zf_rawfd = rawfd;
206 
207     /* Verify that the file is gzipped */
208     if (check_header(zf)) {
209 	close(zf->zf_rawfd);
210 	free(zf);
211 	return(EFTYPE);
212     }
213 
214     /* Initialise the inflation engine */
215     if ((error = inflateInit2(&(zf->zf_zstream), -15)) != Z_OK) {
216 	printf("zf_open: inflateInit returned %d : %s\n", error, zf->zf_zstream.msg);
217 	close(zf->zf_rawfd);
218 	free(zf);
219 	return(EIO);
220     }
221 
222     /* Looks OK, we'll take it */
223     f->f_fsdata = zf;
224     return(0);
225 }
226 
227 static int
228 zf_close(struct open_file *f)
229 {
230     struct z_file	*zf = (struct z_file *)f->f_fsdata;
231 
232     inflateEnd(&(zf->zf_zstream));
233     close(zf->zf_rawfd);
234     free(zf);
235     return(0);
236 }
237 
238 static int
239 zf_read(struct open_file *f, void *buf, size_t size, size_t *resid)
240 {
241     struct z_file	*zf = (struct z_file *)f->f_fsdata;
242     int			error;
243 
244     zf->zf_zstream.next_out = buf;			/* where and how much */
245     zf->zf_zstream.avail_out = size;
246 
247     while (zf->zf_zstream.avail_out && zf->zf_endseen == 0) {
248 	if ((zf->zf_zstream.avail_in == 0) && (zf_fill(zf) == -1)) {
249 	    printf("zf_read: fill error\n");
250 	    return(EIO);
251 	}
252 	if (zf->zf_zstream.avail_in == 0) {		/* oops, unexpected EOF */
253 	    printf("zf_read: unexpected EOF\n");
254 	    if (zf->zf_zstream.avail_out == size)
255 		return(EIO);
256 	    break;
257 	}
258 
259 	error = inflate(&zf->zf_zstream, Z_SYNC_FLUSH);	/* decompression pass */
260 	if (error == Z_STREAM_END) {			/* EOF, all done */
261 	    zf->zf_endseen = 1;
262 	    break;
263 	}
264 	if (error != Z_OK) {				/* argh, decompression error */
265 	    printf("inflate: %s\n", zf->zf_zstream.msg);
266 	    return(EIO);
267 	}
268     }
269     if (resid != NULL)
270 	*resid = zf->zf_zstream.avail_out;
271     return(0);
272 }
273 
274 static int
275 zf_rewind(struct open_file *f)
276 {
277     struct z_file	*zf = (struct z_file *)f->f_fsdata;
278 
279     if (lseek(zf->zf_rawfd, zf->zf_dataoffset, SEEK_SET) == -1)
280 	return(-1);
281     zf->zf_zstream.avail_in = 0;
282     zf->zf_zstream.next_in = NULL;
283     zf->zf_endseen = 0;
284     (void)inflateReset(&zf->zf_zstream);
285 
286     return(0);
287 }
288 
289 static off_t
290 zf_seek(struct open_file *f, off_t offset, int where)
291 {
292     struct z_file	*zf = (struct z_file *)f->f_fsdata;
293     off_t		target;
294     char		discard[16];
295 
296     switch (where) {
297     case SEEK_SET:
298 	target = offset;
299 	break;
300     case SEEK_CUR:
301 	target = offset + zf->zf_zstream.total_out;
302 	break;
303     default:
304 	errno = EINVAL;
305 	return(-1);
306     }
307 
308     /* rewind if required */
309     if (target < zf->zf_zstream.total_out && zf_rewind(f) != 0)
310 	return(-1);
311 
312     /* skip forwards if required */
313     while (target > zf->zf_zstream.total_out) {
314 	errno = zf_read(f, discard, min(sizeof(discard),
315 	    target - zf->zf_zstream.total_out), NULL);
316 	if (errno)
317 	    return(-1);
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