xref: /minix3/usr.bin/gzip/unxz.c (revision 5a645f22a86f086849945a5dd6acbf59f38c913a)
1 /*	$NetBSD: unxz.c,v 1.5 2011/09/30 01:32:21 christos Exp $	*/
2 
3 /*-
4  * Copyright (c) 2011 The NetBSD Foundation, Inc.
5  * All rights reserved.
6  *
7  * This code is derived from software contributed to The NetBSD Foundation
8  * by Christos Zoulas.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  * 1. Redistributions of source code must retain the above copyright
14  *    notice, this list of conditions and the following disclaimer.
15  * 2. Redistributions in binary form must reproduce the above copyright
16  *    notice, this list of conditions and the following disclaimer in the
17  *    documentation and/or other materials provided with the distribution.
18  * 3. All advertising materials mentioning features or use of this software
19  *    must display the following acknowledgement:
20  *        This product includes software developed by the NetBSD
21  *        Foundation, Inc. and its contributors.
22  * 4. Neither the name of The NetBSD Foundation nor the names of its
23  *    contributors may be used to endorse or promote products derived
24  *    from this software without specific prior written permission.
25  *
26  * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
27  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
28  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
29  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
30  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
31  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
32  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
33  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
34  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
35  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
36  * POSSIBILITY OF SUCH DAMAGE.
37  */
38 #include <sys/cdefs.h>
39 __RCSID("$NetBSD: unxz.c,v 1.5 2011/09/30 01:32:21 christos Exp $");
40 
41 #include <stdarg.h>
42 #include <errno.h>
43 #include <stdio.h>
44 #include <unistd.h>
45 #include <lzma.h>
46 
47 static off_t
unxz(int i,int o,char * pre,size_t prelen,off_t * bytes_in)48 unxz(int i, int o, char *pre, size_t prelen, off_t *bytes_in)
49 {
50 	lzma_stream strm = LZMA_STREAM_INIT;
51 	static const int flags = LZMA_TELL_UNSUPPORTED_CHECK|LZMA_CONCATENATED;
52 	lzma_ret ret;
53 	lzma_action action = LZMA_RUN;
54 	off_t bytes_out, bp;
55 	uint8_t ibuf[BUFSIZ];
56 	uint8_t obuf[BUFSIZ];
57 
58 	if (bytes_in == NULL)
59 		bytes_in = &bp;
60 
61 	strm.next_in = ibuf;
62 	memcpy(ibuf, pre, prelen);
63 	strm.avail_in = read(i, ibuf + prelen, sizeof(ibuf) - prelen);
64 	if (strm.avail_in == (size_t)-1)
65 		maybe_err("read failed");
66 	strm.avail_in += prelen;
67 	*bytes_in = strm.avail_in;
68 
69 	if ((ret = lzma_stream_decoder(&strm, UINT64_MAX, flags)) != LZMA_OK)
70 		maybe_errx("Can't initialize decoder (%d)", ret);
71 
72 	strm.next_out = NULL;
73 	strm.avail_out = 0;
74 	if ((ret = lzma_code(&strm, LZMA_RUN)) != LZMA_OK)
75 		maybe_errx("Can't read headers (%d)", ret);
76 
77 	bytes_out = 0;
78 	strm.next_out = obuf;
79 	strm.avail_out = sizeof(obuf);
80 
81 	for (;;) {
82 		if (strm.avail_in == 0) {
83 			strm.next_in = ibuf;
84 			strm.avail_in = read(i, ibuf, sizeof(ibuf));
85 			switch (strm.avail_in) {
86 			case (size_t)-1:
87 				maybe_err("read failed");
88 				/*NOTREACHED*/
89 			case 0:
90 				action = LZMA_FINISH;
91 				break;
92 			default:
93 				*bytes_in += strm.avail_in;
94 				break;
95 			}
96 		}
97 
98 		ret = lzma_code(&strm, action);
99 
100 		// Write and check write error before checking decoder error.
101 		// This way as much data as possible gets written to output
102 		// even if decoder detected an error.
103 		if (strm.avail_out == 0 || ret != LZMA_OK) {
104 			const size_t write_size = sizeof(obuf) - strm.avail_out;
105 
106 			if (write(o, obuf, write_size) != (ssize_t)write_size)
107 				maybe_err("write failed");
108 
109 			strm.next_out = obuf;
110 			strm.avail_out = sizeof(obuf);
111 			bytes_out += write_size;
112 		}
113 
114 		if (ret != LZMA_OK) {
115 			if (ret == LZMA_STREAM_END) {
116 				// Check that there's no trailing garbage.
117 				if (strm.avail_in != 0 || read(i, ibuf, 1))
118 					ret = LZMA_DATA_ERROR;
119 				else {
120 					lzma_end(&strm);
121 					return bytes_out;
122 				}
123 			}
124 
125 			const char *msg;
126 			switch (ret) {
127 			case LZMA_MEM_ERROR:
128 				msg = strerror(ENOMEM);
129 				break;
130 
131 			case LZMA_FORMAT_ERROR:
132 				msg = "File format not recognized";
133 				break;
134 
135 			case LZMA_OPTIONS_ERROR:
136 				// FIXME: Better message?
137 				msg = "Unsupported compression options";
138 				break;
139 
140 			case LZMA_DATA_ERROR:
141 				msg = "File is corrupt";
142 				break;
143 
144 			case LZMA_BUF_ERROR:
145 				msg = "Unexpected end of input";
146 				break;
147 
148 			case LZMA_MEMLIMIT_ERROR:
149 				msg = "Reached memory limit";
150 				break;
151 
152 			default:
153 				maybe_errx("Unknown error (%d)", ret);
154 				break;
155 			}
156 			maybe_errx("%s", msg);
157 
158 		}
159 	}
160 }
161