1/*	$NetBSD: crunchide.c,v 1.8 1997/11/01 06:51:45 lukem Exp $	*/
2/*
3 * Copyright (c) 1997 Christopher G. Demetriou.  All rights reserved.
4 * Copyright (c) 1994 University of Maryland
5 * All Rights Reserved.
6 *
7 * Permission to use, copy, modify, distribute, and sell this software and its
8 * documentation for any purpose is hereby granted without fee, provided that
9 * the above copyright notice appear in all copies and that both that
10 * copyright notice and this permission notice appear in supporting
11 * documentation, and that the name of U.M. not be used in advertising or
12 * publicity pertaining to distribution of the software without specific,
13 * written prior permission.  U.M. makes no representations about the
14 * suitability of this software for any purpose.  It is provided "as is"
15 * without express or implied warranty.
16 *
17 * U.M. DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ALL
18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT SHALL U.M.
19 * BE LIABLE FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
20 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION
21 * OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN
22 * CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
23 *
24 * Author: James da Silva, Systems Design and Analysis Group
25 *			   Computer Science Department
26 *			   University of Maryland at College Park
27 */
28/*
29 * crunchide.c - tiptoes through a symbol table, hiding all defined
30 *	global symbols.  Allows the user to supply a "keep list" of symbols
31 *	that are not to be hidden.
32 *
33 * 	The point of all this is to allow multiple programs to be linked
34 *	together without getting multiple-defined errors.
35 *
36 *	For example, consider a program "foo.c".  It can be linked with a
37 *	small stub routine, called "foostub.c", eg:
38 *	    int foo_main(int argc, char **argv){ return main(argc, argv); }
39 *      like so:
40 *	    cc -c foo.c foostub.c
41 *	    ld -r foo.o foostub.o -o foo.combined.o
42 *	    crunchide -k _foo_main foo.combined.o
43 *	at this point, foo.combined.o can be linked with another program
44 * 	and invoked with "foo_main(argc, argv)".  foo's main() and any
45 * 	other globals are hidden and will not conflict with other symbols.
46 *
47 * TODO:
48 *	- resolve the theoretical hanging reloc problem (see check_reloc()
49 *	  below). I have yet to see this problem actually occur in any real
50 *	  program. In what cases will gcc/gas generate code that needs a
51 *	  relative reloc from a global symbol, other than PIC?  The
52 *	  solution is to not hide the symbol from the linker in this case,
53 *	  but to generate some random name for it so that it doesn't link
54 *	  with anything but holds the place for the reloc.
55 *      - arrange that all the BSS segments start at the same address, so
56 *	  that the final crunched binary BSS size is the max of all the
57 *	  component programs' BSS sizes, rather than their sum.
58 */
59
60#include <sys/cdefs.h>
61#ifndef lint
62__RCSID("$NetBSD: crunchide.c,v 1.8 1997/11/01 06:51:45 lukem Exp $");
63#endif
64#include <sys/types.h>
65#include <sys/stat.h>
66#include <sys/errno.h>
67#include <unistd.h>
68#include <stdio.h>
69#include <stdlib.h>
70#include <string.h>
71#include <fcntl.h>
72
73#include "extern.h"
74
75static const char *pname = "crunchide";
76
77static void usage(void) __dead2;
78
79static void add_to_keep_list(char *symbol);
80static void add_file_to_keep_list(char *filename);
81
82static int hide_syms(const char *filename);
83
84static int verbose;
85
86int main(int, char *[]);
87
88int
89main(int argc, char **argv)
90{
91    int ch, errors;
92
93    if(argc > 0) pname = argv[0];
94
95    while ((ch = getopt(argc, argv, "k:f:v")) != -1)
96	switch(ch) {
97	case 'k':
98	    add_to_keep_list(optarg);
99	    break;
100	case 'f':
101	    add_file_to_keep_list(optarg);
102	    break;
103	case 'v':
104	    verbose = 1;
105	    break;
106	default:
107	    usage();
108	}
109
110    argc -= optind;
111    argv += optind;
112
113    if(argc == 0) usage();
114
115    errors = 0;
116    while(argc) {
117	if (hide_syms(*argv))
118		errors = 1;
119	argc--, argv++;
120    }
121
122    return errors;
123}
124
125static void
126usage(void)
127{
128    fprintf(stderr,
129	    "usage: %s [-k <symbol-name>] [-f <keep-list-file>] <files> ...\n",
130	    pname);
131    exit(1);
132}
133
134/* ---------------------------- */
135
136static struct keep {
137    struct keep *next;
138    char *sym;
139} *keep_list;
140
141static void
142add_to_keep_list(char *symbol)
143{
144    struct keep *newp, *prevp, *curp;
145    int cmp;
146
147    cmp = 0;
148
149    for(curp = keep_list, prevp = NULL; curp; prevp = curp, curp = curp->next)
150	if((cmp = strcmp(symbol, curp->sym)) <= 0) break;
151
152    if(curp && cmp == 0)
153	return;	/* already in table */
154
155    newp = (struct keep *) malloc(sizeof(struct keep));
156    if(newp) newp->sym = strdup(symbol);
157    if(newp == NULL || newp->sym == NULL) {
158	fprintf(stderr, "%s: out of memory for keep list\n", pname);
159	exit(1);
160    }
161
162    newp->next = curp;
163    if(prevp) prevp->next = newp;
164    else keep_list = newp;
165}
166
167int
168in_keep_list(const char *symbol)
169{
170    struct keep *curp;
171    int cmp;
172
173    cmp = 0;
174
175    for(curp = keep_list; curp; curp = curp->next)
176	if((cmp = strcmp(symbol, curp->sym)) <= 0) break;
177
178    return curp && cmp == 0;
179}
180
181static void
182add_file_to_keep_list(char *filename)
183{
184    FILE *keepf;
185    char symbol[1024];
186    int len;
187
188    if((keepf = fopen(filename, "r")) == NULL) {
189	perror(filename);
190	usage();
191    }
192
193    while(fgets(symbol, sizeof(symbol), keepf)) {
194	len = strlen(symbol);
195	if(len && symbol[len-1] == '\n')
196	    symbol[len-1] = '\0';
197
198	add_to_keep_list(symbol);
199    }
200    fclose(keepf);
201}
202
203/* ---------------------------- */
204
205static struct {
206	const char *name;
207	int	(*check)(int, const char *);	/* 1 if match, zero if not */
208	int	(*hide)(int, const char *);	/* non-zero if error */
209} exec_formats[] = {
210#ifdef NLIST_ELF32
211	{	"ELF32",	check_elf32,	hide_elf32,	},
212#endif
213#ifdef NLIST_ELF64
214	{	"ELF64",	check_elf64,	hide_elf64,	},
215#endif
216};
217
218static int
219hide_syms(const char *filename)
220{
221	int fd, i, n, rv;
222
223	fd = open(filename, O_RDWR, 0);
224	if (fd == -1) {
225		perror(filename);
226		return 1;
227	}
228
229	rv = 0;
230
231        n = sizeof exec_formats / sizeof exec_formats[0];
232        for (i = 0; i < n; i++) {
233		if (lseek(fd, 0, SEEK_SET) != 0) {
234			perror(filename);
235			goto err;
236		}
237                if ((*exec_formats[i].check)(fd, filename) != 0)
238                        break;
239	}
240	if (i == n) {
241		fprintf(stderr, "%s: unknown executable format\n", filename);
242		goto err;
243	}
244
245	if (verbose)
246		fprintf(stderr, "%s is an %s binary\n", filename,
247		    exec_formats[i].name);
248
249	if (lseek(fd, 0, SEEK_SET) != 0) {
250		perror(filename);
251		goto err;
252	}
253	rv = (*exec_formats[i].hide)(fd, filename);
254
255out:
256	close (fd);
257	return (rv);
258
259err:
260	rv = 1;
261	goto out;
262}
263