1 /*-
2  * Copyright (c) 2010,2011 Kai Wang
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/param.h>
28 #include <sys/stat.h>
29 #include <ctype.h>
30 #include <err.h>
31 #include <gelf.h>
32 #include <stdio.h>
33 #include <stdlib.h>
34 #include <string.h>
35 #include <unistd.h>
36 
37 #include "elfcopy.h"
38 
39 ELFTC_VCSID("$Id: binary.c 3757 2019-06-28 01:15:28Z emaste $");
40 
41 /*
42  * Convert ELF object to `binary'. Sections with SHF_ALLOC flag set
43  * are copied to the result binary. The relative offsets for each section
44  * are retained, so the result binary file might contain "holes".
45  */
46 void
47 create_binary(int ifd, int ofd)
48 {
49 	Elf *e;
50 	Elf_Scn *scn;
51 	Elf_Data *d;
52 	Elf64_Addr baseaddr;
53 	GElf_Shdr sh;
54 	off_t baseoff, off;
55 	int elferr;
56 
57 	if ((e = elf_begin(ifd, ELF_C_READ, NULL)) == NULL)
58 		errx(EXIT_FAILURE, "elf_begin() failed: %s",
59 		    elf_errmsg(-1));
60 
61 	baseaddr = 0;
62 	baseoff = 0;
63 	if (lseek(ofd, baseoff, SEEK_SET) < 0)
64 		err(EXIT_FAILURE, "lseek failed");
65 
66 	/*
67 	 * Find base offset in the first iteration.
68 	 */
69 	baseoff = -1;
70 	scn = NULL;
71 	while ((scn = elf_nextscn(e, scn)) != NULL) {
72 		if (gelf_getshdr(scn, &sh) == NULL) {
73 			warnx("gelf_getshdr failed: %s", elf_errmsg(-1));
74 			(void) elf_errno();
75 			continue;
76 		}
77 		if ((sh.sh_flags & SHF_ALLOC) == 0 ||
78 		    sh.sh_type == SHT_NOBITS ||
79 		    sh.sh_size == 0)
80 			continue;
81 		if (baseoff == -1 || (off_t) sh.sh_offset < baseoff) {
82 			baseoff = sh.sh_offset;
83 			baseaddr = sh.sh_addr;
84 		}
85 	}
86 	elferr = elf_errno();
87 	if (elferr != 0)
88 		warnx("elf_nextscn failed: %s", elf_errmsg(elferr));
89 
90 	if (baseoff == -1)
91 		return;
92 
93 	/*
94 	 * Write out sections in the second iteration.
95 	 */
96 	scn = NULL;
97 	while ((scn = elf_nextscn(e, scn)) != NULL) {
98 		if (gelf_getshdr(scn, &sh) == NULL) {
99 			warnx("gelf_getshdr failed: %s", elf_errmsg(-1));
100 			(void) elf_errno();
101 			continue;
102 		}
103 		if ((sh.sh_flags & SHF_ALLOC) == 0 ||
104 		    sh.sh_type == SHT_NOBITS ||
105 		    sh.sh_size == 0)
106 			continue;
107 		(void) elf_errno();
108 		if ((d = elf_rawdata(scn, NULL)) == NULL) {
109 			elferr = elf_errno();
110 			if (elferr != 0)
111 				warnx("elf_rawdata failed: %s", elf_errmsg(-1));
112 			continue;
113 		}
114 		if (d->d_buf == NULL || d->d_size == 0)
115 			continue;
116 
117 		/* lseek to section offset relative to `baseaddr'. */
118 		off = sh.sh_addr - baseaddr;
119 		if (lseek(ofd, off, SEEK_SET) < 0)
120 			err(EXIT_FAILURE, "lseek failed");
121 
122 		/* Write out section contents. */
123 		if (write(ofd, d->d_buf, d->d_size) != (ssize_t) d->d_size)
124 			err(EXIT_FAILURE, "write failed");
125 	}
126 	elferr = elf_errno();
127 	if (elferr != 0)
128 		warnx("elf_nextscn failed: %s", elf_errmsg(elferr));
129 }
130 
131 #define	_SYMBOL_NAMSZ	1024
132 
133 /*
134  * Convert `binary' to ELF object. The input `binary' is converted to
135  * a relocatable (.o) file, a few symbols will also be created to make
136  * it easier to access the binary data in other compilation units.
137  */
138 void
139 create_elf_from_binary(struct elfcopy *ecp, int ifd, const char *ifn)
140 {
141 	char name[_SYMBOL_NAMSZ];
142 	struct section *sec, *sec_temp, *shtab;
143 	struct stat sb;
144 	GElf_Ehdr oeh;
145 	GElf_Shdr sh;
146 	void *content;
147 	uint64_t off, data_start, data_end, data_size;
148 	char *sym_basename, *p;
149 
150 	/* Reset internal section list. */
151 	if (!TAILQ_EMPTY(&ecp->v_sec))
152 		TAILQ_FOREACH_SAFE(sec, &ecp->v_sec, sec_list, sec_temp) {
153 			TAILQ_REMOVE(&ecp->v_sec, sec, sec_list);
154 			free(sec);
155 		}
156 
157 	if (fstat(ifd, &sb) == -1)
158 		err(EXIT_FAILURE, "fstat failed");
159 
160 	/* Read the input binary file to a internal buffer. */
161 	if ((content = malloc(sb.st_size)) == NULL)
162 		err(EXIT_FAILURE, "malloc failed");
163 	if (read(ifd, content, sb.st_size) != sb.st_size)
164 		err(EXIT_FAILURE, "read failed");
165 
166 	/*
167 	 * TODO: copy the input binary to output binary verbatim if -O is not
168 	 * specified.
169 	 */
170 
171 	/* Create EHDR for output .o file. */
172 	if (gelf_newehdr(ecp->eout, ecp->oec) == NULL)
173 		errx(EXIT_FAILURE, "gelf_newehdr failed: %s",
174 		    elf_errmsg(-1));
175 	if (gelf_getehdr(ecp->eout, &oeh) == NULL)
176 		errx(EXIT_FAILURE, "gelf_getehdr() failed: %s",
177 		    elf_errmsg(-1));
178 
179 	/* Initialise e_ident fields. */
180 	oeh.e_ident[EI_CLASS] = ecp->oec;
181 	oeh.e_ident[EI_DATA] = ecp->oed;
182 	/*
183 	 * TODO: Set OSABI according to the OS platform where elfcopy(1)
184 	 * was build. (probably)
185 	 */
186 	oeh.e_ident[EI_OSABI] = ELFOSABI_NONE;
187 	oeh.e_machine = ecp->oem;
188 	oeh.e_type = ET_REL;
189 	oeh.e_entry = 0;
190 
191 	ecp->flags |= RELOCATABLE;
192 
193 	/* Create .shstrtab section */
194 	init_shstrtab(ecp);
195 	ecp->shstrtab->off = 0;
196 
197 	/*
198 	 * Create `.data' section which contains the binary data. The
199 	 * section is inserted immediately after EHDR.
200 	 */
201 	off = gelf_fsize(ecp->eout, ELF_T_EHDR, 1, EV_CURRENT);
202 	if (off == 0)
203 		errx(EXIT_FAILURE, "gelf_fsize() failed: %s", elf_errmsg(-1));
204 	(void) create_external_section(ecp, ".data", NULL, content, sb.st_size,
205 	    off, SHT_PROGBITS, ELF_T_BYTE, SHF_ALLOC | SHF_WRITE, 1, 0, 1);
206 
207 	/* Insert .shstrtab after .data section. */
208 	if ((ecp->shstrtab->os = elf_newscn(ecp->eout)) == NULL)
209 		errx(EXIT_FAILURE, "elf_newscn failed: %s",
210 		    elf_errmsg(-1));
211 	insert_to_sec_list(ecp, ecp->shstrtab, 1);
212 
213 	/* Insert section header table here. */
214 	shtab = insert_shtab(ecp, 1);
215 
216 	/* Count in .symtab and .strtab section headers.  */
217 	shtab->sz += gelf_fsize(ecp->eout, ELF_T_SHDR, 2, EV_CURRENT);
218 
219 	if ((sym_basename = strdup(ifn)) == NULL)
220 		err(1, "strdup");
221 	for (p = sym_basename; *p != '\0'; p++)
222 		if (!isalnum(*p & 0xFF))
223 			*p = '_';
224 #define	_GEN_SYMNAME(S) do {						\
225 	snprintf(name, sizeof(name), "%s%s%s", "_binary_", sym_basename, S); \
226 } while (0)
227 
228 	/*
229 	 * Create symbol table.
230 	 */
231 	create_external_symtab(ecp);
232 	data_start = 0;
233 	data_end = data_start + sb.st_size;
234 	data_size = sb.st_size;
235 	_GEN_SYMNAME("_start");
236 	add_to_symtab(ecp, name, data_start, 0, 1,
237 	    ELF32_ST_INFO(STB_GLOBAL, STT_NOTYPE), 0, 1);
238 	_GEN_SYMNAME("_end");
239 	add_to_symtab(ecp, name, data_end, 0, 1,
240 	    ELF32_ST_INFO(STB_GLOBAL, STT_NOTYPE), 0, 1);
241 	_GEN_SYMNAME("_size");
242 	add_to_symtab(ecp, name, data_size, 0, SHN_ABS,
243 	    ELF32_ST_INFO(STB_GLOBAL, STT_NOTYPE), 0, 1);
244 	finalize_external_symtab(ecp);
245 	create_symtab_data(ecp);
246 #undef	_GEN_SYMNAME
247 	free(sym_basename);
248 
249 	/*
250 	 * Write the underlying ehdr. Note that it should be called
251 	 * before elf_setshstrndx() since it will overwrite e->e_shstrndx.
252 	 */
253 	if (gelf_update_ehdr(ecp->eout, &oeh) == 0)
254 		errx(EXIT_FAILURE, "gelf_update_ehdr() failed: %s",
255 		    elf_errmsg(-1));
256 
257 	/* Update sh_name pointer for each section header entry. */
258 	ecp->flags |= SYMTAB_EXIST;
259 	update_shdr(ecp, 0);
260 
261 	/* Properly set sh_link field of .symtab section. */
262 	if (gelf_getshdr(ecp->symtab->os, &sh) == NULL)
263 		errx(EXIT_FAILURE, "692 gelf_getshdr() failed: %s",
264 		    elf_errmsg(-1));
265 	sh.sh_link = elf_ndxscn(ecp->strtab->os);
266 	if (!gelf_update_shdr(ecp->symtab->os, &sh))
267 		errx(EXIT_FAILURE, "gelf_update_shdr() failed: %s",
268 		    elf_errmsg(-1));
269 
270 	/* Renew oeh to get the updated e_shstrndx. */
271 	if (gelf_getehdr(ecp->eout, &oeh) == NULL)
272 		errx(EXIT_FAILURE, "gelf_getehdr() failed: %s",
273 		    elf_errmsg(-1));
274 
275 	/* Resync section offsets. */
276 	resync_sections(ecp);
277 
278 	/* Store SHDR offset in EHDR. */
279 	oeh.e_shoff = shtab->off;
280 
281 	/* Update ehdr since we modified e_shoff. */
282 	if (gelf_update_ehdr(ecp->eout, &oeh) == 0)
283 		errx(EXIT_FAILURE, "gelf_update_ehdr() failed: %s",
284 		    elf_errmsg(-1));
285 
286 	/* Write out the output elf object. */
287 	if (elf_update(ecp->eout, ELF_C_WRITE) < 0)
288 		errx(EXIT_FAILURE, "elf_update() failed: %s",
289 		    elf_errmsg(-1));
290 
291 	/* Release allocated resource. */
292 	free(content);
293 	free_elf(ecp);
294 }
295