1 /* $NetBSD: crunchide.c,v 1.12 2004/08/24 12:25:26 wiz Exp $ */ 2 3 /* 4 * Copyright (c) 1997 Christopher G. Demetriou. All rights reserved. 5 * Copyright (c) 1994 University of Maryland 6 * All Rights Reserved. 7 * 8 * Permission to use, copy, modify, distribute, and sell this software and its 9 * documentation for any purpose is hereby granted without fee, provided that 10 * the above copyright notice appear in all copies and that both that 11 * copyright notice and this permission notice appear in supporting 12 * documentation, and that the name of U.M. not be used in advertising or 13 * publicity pertaining to distribution of the software without specific, 14 * written prior permission. U.M. makes no representations about the 15 * suitability of this software for any purpose. It is provided "as is" 16 * without express or implied warranty. 17 * 18 * U.M. DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ALL 19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT SHALL U.M. 20 * BE LIABLE FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES 21 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION 22 * OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN 23 * CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. 24 * 25 * Author: James da Silva, Systems Design and Analysis Group 26 * Computer Science Department 27 * University of Maryland at College Park 28 */ 29 30 /* 31 * crunchide.c - tiptoes through an a.out symbol table, hiding all defined 32 * global symbols. Allows the user to supply a "keep list" of symbols 33 * that are not to be hidden. This program relies on the use of the 34 * linker's -dc flag to actually put global bss data into the file's 35 * bss segment (rather than leaving it as undefined "common" data). 36 * 37 * The point of all this is to allow multiple programs to be linked 38 * together without getting multiple-defined errors. 39 * 40 * For example, consider a program "foo.c". It can be linked with a 41 * small stub routine, called "foostub.c", eg: 42 * int foo_main(int argc, char **argv){ return main(argc, argv); } 43 * like so: 44 * cc -c foo.c foostub.c 45 * ld -dc -r foo.o foostub.o -o foo.combined.o 46 * crunchide -k _foo_main foo.combined.o 47 * at this point, foo.combined.o can be linked with another program 48 * and invoked with "foo_main(argc, argv)". foo's main() and any 49 * other globals are hidden and will not conflict with other symbols. 50 * 51 * TODO: 52 * - resolve the theoretical hanging reloc problem (see check_reloc() 53 * below). I have yet to see this problem actually occur in any real 54 * program. In what cases will gcc/gas generate code that needs a 55 * relative reloc from a global symbol, other than PIC? The 56 * solution is to not hide the symbol from the linker in this case, 57 * but to generate some random name for it so that it doesn't link 58 * with anything but holds the place for the reloc. 59 * - arrange that all the BSS segments start at the same address, so 60 * that the final crunched binary BSS size is the max of all the 61 * component programs' BSS sizes, rather than their sum. 62 */ 63 64 #include <sys/cdefs.h> 65 #ifndef lint 66 __RCSID("$NetBSD: crunchide.c,v 1.12 2004/08/24 12:25:26 wiz Exp $"); 67 #endif 68 69 #include <unistd.h> 70 #include <stdio.h> 71 #include <stdlib.h> 72 #include <string.h> 73 #include <fcntl.h> 74 #include <errno.h> 75 #include <a.out.h> 76 #include <sys/types.h> 77 #include <sys/stat.h> 78 79 #include "extern.h" 80 81 void usage(void); 82 83 void add_to_keep_list(char *symbol); 84 void add_file_to_keep_list(char *filename); 85 86 int hide_syms(const char *filename); 87 88 int verbose; 89 90 int 91 main(int argc, char *argv[]) 92 { 93 int ch, errors; 94 95 setprogname(argv[0]); 96 97 while ((ch = getopt(argc, argv, "k:f:v")) != -1) 98 switch(ch) { 99 case 'k': 100 add_to_keep_list(optarg); 101 break; 102 case 'f': 103 add_file_to_keep_list(optarg); 104 break; 105 case 'v': 106 verbose = 1; 107 break; 108 default: 109 usage(); 110 } 111 112 argc -= optind; 113 argv += optind; 114 115 if (argc == 0) 116 usage(); 117 118 errors = 0; 119 while (argc) { 120 if (hide_syms(*argv)) 121 errors = 1; 122 argc--, argv++; 123 } 124 125 return errors; 126 } 127 128 void 129 usage(void) 130 { 131 fprintf(stderr, 132 "Usage: %s [-k keep-symbol] [-f keep-list-file] object-file\n" 133 "\t\t [object-file ...]\n", 134 getprogname()); 135 exit(1); 136 } 137 138 /* ---------------------------- */ 139 140 struct keep { 141 struct keep *next; 142 char *sym; 143 } *keep_list; 144 145 void 146 add_to_keep_list(char *symbol) 147 { 148 struct keep *newp, *prevp, *curp; 149 int cmp; 150 151 cmp = 0; 152 153 for (curp = keep_list, prevp = NULL; curp; prevp = curp, curp = curp->next) 154 if ((cmp = strcmp(symbol, curp->sym)) <= 0) 155 break; 156 157 if (curp && cmp == 0) 158 return; /* already in table */ 159 160 newp = (struct keep *) malloc(sizeof(struct keep)); 161 if (newp) 162 newp->sym = strdup(symbol); 163 if (newp == NULL || newp->sym == NULL) { 164 fprintf(stderr, "%s: out of memory for keep list\n", getprogname()); 165 exit(1); 166 } 167 168 newp->next = curp; 169 if (prevp) 170 prevp->next = newp; 171 else 172 keep_list = newp; 173 } 174 175 int 176 in_keep_list(const char *symbol) 177 { 178 struct keep *curp; 179 int cmp; 180 181 cmp = 0; 182 183 for (curp = keep_list; curp; curp = curp->next) 184 if((cmp = strcmp(symbol, curp->sym)) <= 0) 185 break; 186 187 return curp && cmp == 0; 188 } 189 190 void 191 add_file_to_keep_list(char *filename) 192 { 193 FILE *keepf; 194 char symbol[1024]; 195 int len; 196 197 if ((keepf = fopen(filename, "r")) == NULL) { 198 perror(filename); 199 usage(); 200 } 201 202 while (fgets(symbol, 1024, keepf)) { 203 len = strlen(symbol); 204 if (len && symbol[len-1] == '\n') 205 symbol[len-1] = '\0'; 206 207 add_to_keep_list(symbol); 208 } 209 fclose(keepf); 210 } 211 212 /* ---------------------------- */ 213 214 struct { 215 const char *name; 216 int (*check)(int, const char *); /* 1 if match, zero if not */ 217 int (*hide)(int, const char *); /* non-zero if error */ 218 } exec_formats[] = { 219 #ifdef NLIST_AOUT 220 { "a.out", check_aout, hide_aout, }, 221 #endif 222 #ifdef NLIST_COFF 223 { "COFF", check_coff, hide_coff, }, 224 #endif 225 #ifdef NLIST_ECOFF 226 { "ECOFF", check_ecoff, hide_ecoff, }, 227 #endif 228 #ifdef NLIST_ELF32 229 { "ELF32", check_elf32, hide_elf32, }, 230 #endif 231 #ifdef NLIST_ELF64 232 { "ELF64", check_elf64, hide_elf64, }, 233 #endif 234 }; 235 236 int 237 hide_syms(const char *filename) 238 { 239 int fd, i, n, rv; 240 241 fd = open(filename, O_RDWR, 0); 242 if (fd == -1) { 243 perror(filename); 244 return 1; 245 } 246 247 rv = 0; 248 249 n = sizeof exec_formats / sizeof exec_formats[0]; 250 for (i = 0; i < n; i++) { 251 if (lseek(fd, 0, SEEK_SET) != 0) { 252 perror(filename); 253 goto err; 254 } 255 if ((*exec_formats[i].check)(fd, filename) != 0) 256 break; 257 } 258 if (i == n) { 259 fprintf(stderr, "%s: unknown executable format\n", filename); 260 goto err; 261 } 262 263 if (verbose) 264 fprintf(stderr, "%s is an %s binary\n", filename, 265 exec_formats[i].name); 266 267 if (lseek(fd, 0, SEEK_SET) != 0) { 268 perror(filename); 269 goto err; 270 } 271 rv = (*exec_formats[i].hide)(fd, filename); 272 273 out: 274 close(fd); 275 return rv; 276 277 err: 278 rv = 1; 279 goto out; 280 } 281