src/unexcoff.c

   1 /* Copyright (C) 1985-1988, 1992-1994, 2001-2018 Free Software
   2  * Foundation, Inc.
   3
   4 This file is part of GNU Emacs.
   5
   6 GNU Emacs is free software: you can redistribute it and/or modify
   7 it under the terms of the GNU General Public License as published by
   8 the Free Software Foundation, either version 3 of the License, or (at
   9 your option) any later version.
  10
  11 GNU Emacs is distributed in the hope that it will be useful,
  12 but WITHOUT ANY WARRANTY; without even the implied warranty of
  13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14 GNU General Public License for more details.
  15
  16 You should have received a copy of the GNU General Public License
  17 along with GNU Emacs.  If not, see <https://www.gnu.org/licenses/>.  */
  18
  19
  20 /*
  21  * unexcoff.c - Convert a running program into an a.out or COFF file.
  22  *
  23  * ==================================================================
  24  * Note: This file is currently used only by the MSDOS (a.k.a. DJGPP)
  25  * build of Emacs.  If you are not interested in the MSDOS build, you
  26  * are looking at the wrong version of unexec!
  27  * ==================================================================
  28  *
  29  * Author:      Spencer W. Thomas
  30  *              Computer Science Dept.
  31  *              University of Utah
  32  * Date:        Tue Mar  2 1982
  33  * Originally under the name unexec.c.
  34  * Modified heavily since then.
  35  *
  36  * Synopsis:
  37  *      unexec (const char *new_name, const char *old_name);
  38  *
  39  * Takes a snapshot of the program and makes an a.out format file in the
  40  * file named by the string argument new_name.
  41  * If a_name is non-NULL, the symbol table will be taken from the given file.
  42  * On some machines, an existing a_name file is required.
  43  *
  44  * If you make improvements I'd like to get them too.
  45  * harpo!utah-cs!thomas, thomas@Utah-20
  46  *
  47  */
  48
  49 /* Modified to support SysVr3 shared libraries by James Van Artsdalen
  50  * of Dell Computer Corporation.  james@bigtex.cactus.org.
  51  */
  52
  53 #include <config.h>
  54 #include "unexec.h"
  55 #include "lisp.h"
  56
  57 #define PERROR(file) report_error (file, new)
  58
  59 #ifndef CANNOT_DUMP  /* all rest of file!  */
  60
  61 #ifdef HAVE_COFF_H
  62 #include <coff.h>
  63 #ifdef MSDOS
  64 #include <fcntl.h>  /* for O_RDONLY, O_RDWR */
  65 #include <crt0.h>   /* for _crt0_startup_flags and its bits */
  66 #include <sys/exceptn.h>
  67 static int save_djgpp_startup_flags;
  68 #include <libc/atexit.h>
  69 static struct __atexit *save_atexit_ptr;
  70 #define filehdr external_filehdr
  71 #define scnhdr external_scnhdr
  72 #define syment external_syment
  73 #define auxent external_auxent
  74 #define n_numaux e_numaux
  75 #define n_type e_type
  76 struct aouthdr
  77 {
  78   unsigned short        magic;  /* type of file                         */
  79   unsigned short        vstamp; /* version stamp                        */
  80   unsigned long         tsize;  /* text size in bytes, padded to FW bdry*/
  81   unsigned long         dsize;  /* initialized data "  "                */
  82   unsigned long         bsize;  /* uninitialized data "   "             */
  83   unsigned long         entry;  /* entry pt.                            */
  84   unsigned long         text_start;/* base of text used for this file */
  85   unsigned long         data_start;/* base of data used for this file */
  86 };
  87 #endif /* MSDOS */
  88 #else  /* not HAVE_COFF_H */
  89 #include <a.out.h>
  90 #endif /* not HAVE_COFF_H */
  91
  92 /* Define getpagesize if the system does not.
  93    Note that this may depend on symbols defined in a.out.h.  */
  94 #include "getpagesize.h"
  95
  96 #ifndef makedev                 /* Try to detect types.h already loaded */
  97 #include <sys/types.h>
  98 #endif /* makedev */
  99 #include <stdio.h>
 100 #include <errno.h>
 101
 102 #include <sys/file.h>
 103
 104 extern int etext;
 105
 106 static long block_copy_start;           /* Old executable start point */
 107 static struct filehdr f_hdr;            /* File header */
 108 static struct aouthdr f_ohdr;           /* Optional file header (a.out) */
 109 long bias;                      /* Bias to add for growth */
 110 long lnnoptr;                   /* Pointer to line-number info within file */
 111 #define SYMS_START block_copy_start
 112
 113 static long text_scnptr;
 114 static long data_scnptr;
 115
 116 static long coff_offset;
 117
 118 static int pagemask;
 119
 120 /* Correct an int which is the bit pattern of a pointer to a byte
 121    into an int which is the number of a byte.
 122    This is a no-op on ordinary machines, but not on all.  */
 123
 124 #define ADDR_CORRECT(x) ((char *) (x) - (char *) 0)
 125
 126 #include "lisp.h"
 127
 128 static void
 129 report_error (const char *file, int fd)
 130 {
 131   int err = errno;
 132   if (fd)
 133     emacs_close (fd);
 134   report_file_errno ("Cannot unexec", build_string (file), err);
 135 }
 136
 137 #define ERROR0(msg) report_error_1 (new, msg, 0, 0); return -1
 138 #define ERROR1(msg,x) report_error_1 (new, msg, x, 0); return -1
 139 #define ERROR2(msg,x,y) report_error_1 (new, msg, x, y); return -1
 140
 141 static void
 142 report_error_1 (int fd, const char *msg, int a1, int a2)
 143 {
 144   emacs_close (fd);
 145   error (msg, a1, a2);
 146 }
 147 \f
 148 static int make_hdr (int, int, const char *, const char *);
 149 static int copy_text_and_data (int, int);
 150 static int copy_sym (int, int, const char *, const char *);
 151 static void mark_x (const char *);
 152
 153 /* ****************************************************************
 154  * make_hdr
 155  *
 156  * Make the header in the new a.out from the header in core.
 157  * Modify the text and data sizes.
 158  */
 159 static int
 160 make_hdr (int new, int a_out,
 161           const char *a_name, const char *new_name)
 162 {
 163   auto struct scnhdr f_thdr;            /* Text section header */
 164   auto struct scnhdr f_dhdr;            /* Data section header */
 165   auto struct scnhdr f_bhdr;            /* Bss section header */
 166   auto struct scnhdr scntemp;           /* Temporary section header */
 167   register int scns;
 168   unsigned int bss_start;
 169   unsigned int data_start;
 170
 171   pagemask = getpagesize () - 1;
 172
 173   /* Adjust text/data boundary. */
 174   data_start = (int) DATA_START;
 175   data_start = ADDR_CORRECT (data_start);
 176   data_start = data_start & ~pagemask; /* (Down) to page boundary. */
 177
 178   bss_start = ADDR_CORRECT (sbrk (0)) + pagemask;
 179   bss_start &= ~ pagemask;
 180
 181   if (data_start > bss_start)   /* Can't have negative data size. */
 182     {
 183       ERROR2 ("unexec: data_start (%u) can't be greater than bss_start (%u)",
 184               data_start, bss_start);
 185     }
 186
 187   coff_offset = 0L;             /* stays zero, except in DJGPP */
 188
 189   /* Salvage as much info from the existing file as possible */
 190   if (a_out >= 0)
 191     {
 192 #ifdef MSDOS
 193       /* Support the coff-go32-exe format with a prepended stub, since
 194          this is what GCC 2.8.0 and later generates by default in DJGPP.  */
 195       unsigned short mz_header[3];
 196
 197       if (read (a_out, &mz_header, sizeof (mz_header)) != sizeof (mz_header))
 198         {
 199           PERROR (a_name);
 200         }
 201       if (mz_header[0] == 0x5a4d || mz_header[0] == 0x4d5a) /* "MZ" or "ZM" */
 202         {
 203           coff_offset = (long)mz_header[2] * 512L;
 204           if (mz_header[1])
 205             coff_offset += (long)mz_header[1] - 512L;
 206           lseek (a_out, coff_offset, 0);
 207         }
 208       else
 209         lseek (a_out, 0L, 0);
 210 #endif /* MSDOS */
 211       if (read (a_out, &f_hdr, sizeof (f_hdr)) != sizeof (f_hdr))
 212         {
 213           PERROR (a_name);
 214         }
 215       block_copy_start += sizeof (f_hdr);
 216       if (f_hdr.f_opthdr > 0)
 217         {
 218           if (read (a_out, &f_ohdr, sizeof (f_ohdr)) != sizeof (f_ohdr))
 219             {
 220               PERROR (a_name);
 221             }
 222           block_copy_start += sizeof (f_ohdr);
 223         }
 224       /* Loop through section headers, copying them in */
 225       lseek (a_out, coff_offset + sizeof (f_hdr) + f_hdr.f_opthdr, 0);
 226       for (scns = f_hdr.f_nscns; scns > 0; scns--) {
 227         if (read (a_out, &scntemp, sizeof (scntemp)) != sizeof (scntemp))
 228           {
 229             PERROR (a_name);
 230           }
 231         if (scntemp.s_scnptr > 0L)
 232           {
 233             if (block_copy_start < scntemp.s_scnptr + scntemp.s_size)
 234               block_copy_start = scntemp.s_scnptr + scntemp.s_size;
 235           }
 236         if (strcmp (scntemp.s_name, ".text") == 0)
 237           {
 238             f_thdr = scntemp;
 239           }
 240         else if (strcmp (scntemp.s_name, ".data") == 0)
 241           {
 242             f_dhdr = scntemp;
 243           }
 244         else if (strcmp (scntemp.s_name, ".bss") == 0)
 245           {
 246             f_bhdr = scntemp;
 247           }
 248       }
 249     }
 250   else
 251     {
 252       ERROR0 ("can't build a COFF file from scratch yet");
 253     }
 254
 255   /* Now we alter the contents of all the f_*hdr variables
 256      to correspond to what we want to dump.  */
 257
 258   f_hdr.f_flags |= (F_RELFLG | F_EXEC);
 259   f_ohdr.dsize = bss_start - f_ohdr.data_start;
 260   f_ohdr.bsize = 0;
 261   f_thdr.s_size = f_ohdr.tsize;
 262   f_thdr.s_scnptr = sizeof (f_hdr) + sizeof (f_ohdr);
 263   f_thdr.s_scnptr += (f_hdr.f_nscns) * (sizeof (f_thdr));
 264   lnnoptr = f_thdr.s_lnnoptr;
 265   text_scnptr = f_thdr.s_scnptr;
 266   f_dhdr.s_paddr = f_ohdr.data_start;
 267   f_dhdr.s_vaddr = f_ohdr.data_start;
 268   f_dhdr.s_size = f_ohdr.dsize;
 269   f_dhdr.s_scnptr = f_thdr.s_scnptr + f_thdr.s_size;
 270   data_scnptr = f_dhdr.s_scnptr;
 271   f_bhdr.s_paddr = f_ohdr.data_start + f_ohdr.dsize;
 272   f_bhdr.s_vaddr = f_ohdr.data_start + f_ohdr.dsize;
 273   f_bhdr.s_size = f_ohdr.bsize;
 274   f_bhdr.s_scnptr = 0L;
 275   bias = f_dhdr.s_scnptr + f_dhdr.s_size - block_copy_start;
 276
 277   if (f_hdr.f_symptr > 0L)
 278     {
 279       f_hdr.f_symptr += bias;
 280     }
 281
 282   if (f_thdr.s_lnnoptr > 0L)
 283     {
 284       f_thdr.s_lnnoptr += bias;
 285     }
 286
 287   if (write (new, &f_hdr, sizeof (f_hdr)) != sizeof (f_hdr))
 288     {
 289       PERROR (new_name);
 290     }
 291
 292   if (write (new, &f_ohdr, sizeof (f_ohdr)) != sizeof (f_ohdr))
 293     {
 294       PERROR (new_name);
 295     }
 296
 297   if (write (new, &f_thdr, sizeof (f_thdr)) != sizeof (f_thdr))
 298     {
 299       PERROR (new_name);
 300     }
 301
 302   if (write (new, &f_dhdr, sizeof (f_dhdr)) != sizeof (f_dhdr))
 303     {
 304       PERROR (new_name);
 305     }
 306
 307   if (write (new, &f_bhdr, sizeof (f_bhdr)) != sizeof (f_bhdr))
 308     {
 309       PERROR (new_name);
 310     }
 311
 312   return (0);
 313
 314 }
 315 \f
 316 void
 317 write_segment (int new, const char *ptr, const char *end)
 318 {
 319   register int i, nwrite, ret;
 320   /* This is the normal amount to write at once.
 321      It is the size of block that NFS uses.  */
 322   int writesize = 1 << 13;
 323   int pagesize = getpagesize ();
 324   char zeros[1 << 13];
 325
 326   memset (zeros, 0, sizeof (zeros));
 327
 328   for (i = 0; ptr < end;)
 329     {
 330       /* Distance to next multiple of writesize.  */
 331       nwrite = (((int) ptr + writesize) & -writesize) - (int) ptr;
 332       /* But not beyond specified end.  */
 333       if (nwrite > end - ptr) nwrite = end - ptr;
 334       ret = write (new, ptr, nwrite);
 335       /* If write gets a page fault, it means we reached
 336          a gap between the old text segment and the old data segment.
 337          This gap has probably been remapped into part of the text segment.
 338          So write zeros for it.  */
 339       if (ret == -1 && errno == EFAULT)
 340         {
 341           /* Write only a page of zeros at once,
 342              so that we don't overshoot the start
 343              of the valid memory in the old data segment.  */
 344           if (nwrite > pagesize)
 345             nwrite = pagesize;
 346           write (new, zeros, nwrite);
 347         }
 348       i += nwrite;
 349       ptr += nwrite;
 350     }
 351 }
 352 /* ****************************************************************
 353  * copy_text_and_data
 354  *
 355  * Copy the text and data segments from memory to the new a.out
 356  */
 357 static int
 358 copy_text_and_data (int new, int a_out)
 359 {
 360   register char *end;
 361   register char *ptr;
 362
 363 #ifdef MSDOS
 364   /* Dump the original table of exception handlers, not the one
 365      where our exception hooks are registered.  */
 366   __djgpp_exception_toggle ();
 367
 368   /* Switch off startup flags that might have been set at runtime
 369      and which might change the way that dumped Emacs works.  */
 370   save_djgpp_startup_flags = _crt0_startup_flags;
 371   _crt0_startup_flags &= ~(_CRT0_FLAG_NO_LFN | _CRT0_FLAG_NEARPTR);
 372
 373   /* Zero out the 'atexit' chain in the dumped executable, to avoid
 374      calling the atexit functions twice.  (emacs.c:main installs an
 375      atexit function.)  */
 376   save_atexit_ptr = __atexit_ptr;
 377   __atexit_ptr = NULL;
 378 #endif
 379
 380   lseek (new, (long) text_scnptr, 0);
 381   ptr = (char *) f_ohdr.text_start;
 382   end = ptr + f_ohdr.tsize;
 383   write_segment (new, ptr, end);
 384
 385   lseek (new, (long) data_scnptr, 0);
 386   ptr = (char *) f_ohdr.data_start;
 387   end = ptr + f_ohdr.dsize;
 388   write_segment (new, ptr, end);
 389
 390 #ifdef MSDOS
 391   /* Restore our exception hooks.  */
 392   __djgpp_exception_toggle ();
 393
 394   /* Restore the startup flags.  */
 395   _crt0_startup_flags = save_djgpp_startup_flags;
 396
 397   /* Restore the atexit chain.  */
 398   __atexit_ptr = save_atexit_ptr;
 399 #endif
 400
 401
 402   return 0;
 403 }
 404 \f
 405 /* ****************************************************************
 406  * copy_sym
 407  *
 408  * Copy the relocation information and symbol table from the a.out to the new
 409  */
 410 static int
 411 copy_sym (int new, int a_out, const char *a_name, const char *new_name)
 412 {
 413   char page[1024];
 414   int n;
 415
 416   if (a_out < 0)
 417     return 0;
 418
 419   if (SYMS_START == 0L)
 420     return 0;
 421
 422   if (lnnoptr)                  /* if there is line number info */
 423     lseek (a_out, coff_offset + lnnoptr, 0);    /* start copying from there */
 424   else
 425     lseek (a_out, coff_offset + SYMS_START, 0); /* Position a.out to symtab. */
 426
 427   while ((n = read (a_out, page, sizeof page)) > 0)
 428     {
 429       if (write (new, page, n) != n)
 430         {
 431           PERROR (new_name);
 432         }
 433     }
 434   if (n < 0)
 435     {
 436       PERROR (a_name);
 437     }
 438   return 0;
 439 }
 440 \f
 441
 442 /*
 443  *      If the COFF file contains a symbol table and a line number section,
 444  *      then any auxiliary entries that have values for x_lnnoptr must
 445  *      be adjusted by the amount that the line number section has moved
 446  *      in the file (bias computed in make_hdr).  The #@$%&* designers of
 447  *      the auxiliary entry structures used the absolute file offsets for
 448  *      the line number entry rather than an offset from the start of the
 449  *      line number section!
 450  *
 451  *      When I figure out how to scan through the symbol table and pick out
 452  *      the auxiliary entries that need adjustment, this routine will
 453  *      be fixed.  As it is now, all such entries are wrong and sdb
 454  *      will complain.   Fred Fish, UniSoft Systems Inc.
 455  */
 456
 457 /* This function is probably very slow.  Instead of reopening the new
 458    file for input and output it should copy from the old to the new
 459    using the two descriptors already open (WRITEDESC and READDESC).
 460    Instead of reading one small structure at a time it should use
 461    a reasonable size buffer.  But I don't have time to work on such
 462    things, so I am installing it as submitted to me.  -- RMS.  */
 463
 464 int
 465 adjust_lnnoptrs (int writedesc, int readdesc, const char *new_name)
 466 {
 467   register int nsyms;
 468   register int new;
 469   struct syment symentry;
 470   union auxent auxentry;
 471
 472   if (!lnnoptr || !f_hdr.f_symptr)
 473     return 0;
 474
 475 #ifdef MSDOS
 476   if ((new = writedesc) < 0)
 477 #else
 478   if ((new = emacs_open (new_name, O_RDWR, 0)) < 0)
 479 #endif
 480     {
 481       PERROR (new_name);
 482       return -1;
 483     }
 484
 485   lseek (new, f_hdr.f_symptr, 0);
 486   for (nsyms = 0; nsyms < f_hdr.f_nsyms; nsyms++)
 487     {
 488       read (new, &symentry, SYMESZ);
 489       if (symentry.n_numaux)
 490         {
 491           read (new, &auxentry, AUXESZ);
 492           nsyms++;
 493           if (ISFCN (symentry.n_type) || symentry.n_type == 0x2400)
 494             {
 495               auxentry.x_sym.x_fcnary.x_fcn.x_lnnoptr += bias;
 496               lseek (new, -AUXESZ, 1);
 497               write (new, &auxentry, AUXESZ);
 498             }
 499         }
 500     }
 501 #ifndef MSDOS
 502   emacs_close (new);
 503 #endif
 504   return 0;
 505 }
 506
 507 /* ****************************************************************
 508  * unexec
 509  *
 510  * driving logic.
 511  */
 512 void
 513 unexec (const char *new_name, const char *a_name)
 514 {
 515   int new = -1, a_out = -1;
 516
 517   if (a_name && (a_out = emacs_open (a_name, O_RDONLY, 0)) < 0)
 518     {
 519       PERROR (a_name);
 520     }
 521   if ((new = emacs_open (new_name, O_WRONLY | O_CREAT | O_TRUNC, 0777)) < 0)
 522     {
 523       PERROR (new_name);
 524     }
 525
 526   if (make_hdr (new, a_out, a_name, new_name) < 0
 527       || copy_text_and_data (new, a_out) < 0
 528       || copy_sym (new, a_out, a_name, new_name) < 0
 529       || adjust_lnnoptrs (new, a_out, new_name) < 0
 530       )
 531     {
 532       emacs_close (new);
 533       return;
 534     }
 535
 536   emacs_close (new);
 537   if (a_out >= 0)
 538     emacs_close (a_out);
 539 }
 540
 541 #endif /* not CANNOT_DUMP */