gdb/solib-darwin.c

   1 /* Handle Darwin shared libraries for GDB, the GNU Debugger.
   2
   3    Copyright (C) 2009-2018 Free Software Foundation, Inc.
   4
   5    This file is part of GDB.
   6
   7    This program is free software; you can redistribute it and/or modify
   8    it under the terms of the GNU General Public License as published by
   9    the Free Software Foundation; either version 3 of the License, or
  10    (at your option) any later version.
  11
  12    This program is distributed in the hope that it will be useful,
  13    but WITHOUT ANY WARRANTY; without even the implied warranty of
  14    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15    GNU General Public License for more details.
  16
  17    You should have received a copy of the GNU General Public License
  18    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
  19
  20 #include "defs.h"
  21
  22 #include "symtab.h"
  23 #include "bfd.h"
  24 #include "symfile.h"
  25 #include "objfiles.h"
  26 #include "gdbcore.h"
  27 #include "target.h"
  28 #include "inferior.h"
  29 #include "regcache.h"
  30 #include "gdbthread.h"
  31 #include "gdb_bfd.h"
  32
  33 #include "solist.h"
  34 #include "solib.h"
  35 #include "solib-svr4.h"
  36
  37 #include "bfd-target.h"
  38 #include "elf-bfd.h"
  39 #include "exec.h"
  40 #include "auxv.h"
  41 #include "mach-o.h"
  42 #include "mach-o/external.h"
  43
  44 struct gdb_dyld_image_info
  45 {
  46   /* Base address (which corresponds to the Mach-O header).  */
  47   CORE_ADDR mach_header;
  48   /* Image file path.  */
  49   CORE_ADDR file_path;
  50   /* st.m_time of image file.  */
  51   unsigned long mtime;
  52 };
  53
  54 /* Content of inferior dyld_all_image_infos structure.
  55    See /usr/include/mach-o/dyld_images.h for the documentation.  */
  56 struct gdb_dyld_all_image_infos
  57 {
  58   /* Version (1).  */
  59   unsigned int version;
  60   /* Number of images.  */
  61   unsigned int count;
  62   /* Image description.  */
  63   CORE_ADDR info;
  64   /* Notifier (function called when a library is added or removed).  */
  65   CORE_ADDR notifier;
  66 };
  67
  68 /* Current all_image_infos version.  */
  69 #define DYLD_VERSION_MIN 1
  70 #define DYLD_VERSION_MAX 15
  71
  72 /* Per PSPACE specific data.  */
  73 struct darwin_info
  74 {
  75   /* Address of structure dyld_all_image_infos in inferior.  */
  76   CORE_ADDR all_image_addr;
  77
  78   /* Gdb copy of dyld_all_info_infos.  */
  79   struct gdb_dyld_all_image_infos all_image;
  80 };
  81
  82 /* Per-program-space data key.  */
  83 static const struct program_space_data *solib_darwin_pspace_data;
  84
  85 static void
  86 darwin_pspace_data_cleanup (struct program_space *pspace, void *arg)
  87 {
  88   xfree (arg);
  89 }
  90
  91 /* Get the current darwin data.  If none is found yet, add it now.  This
  92    function always returns a valid object.  */
  93
  94 static struct darwin_info *
  95 get_darwin_info (void)
  96 {
  97   struct darwin_info *info;
  98
  99   info = (struct darwin_info *) program_space_data (current_program_space,
 100                                                     solib_darwin_pspace_data);
 101   if (info != NULL)
 102     return info;
 103
 104   info = XCNEW (struct darwin_info);
 105   set_program_space_data (current_program_space,
 106                           solib_darwin_pspace_data, info);
 107   return info;
 108 }
 109
 110 /* Return non-zero if the version in dyld_all_image is known.  */
 111
 112 static int
 113 darwin_dyld_version_ok (const struct darwin_info *info)
 114 {
 115   return info->all_image.version >= DYLD_VERSION_MIN
 116     && info->all_image.version <= DYLD_VERSION_MAX;
 117 }
 118
 119 /* Read dyld_all_image from inferior.  */
 120
 121 static void
 122 darwin_load_image_infos (struct darwin_info *info)
 123 {
 124   gdb_byte buf[24];
 125   enum bfd_endian byte_order = gdbarch_byte_order (target_gdbarch ());
 126   struct type *ptr_type = builtin_type (target_gdbarch ())->builtin_data_ptr;
 127   int len;
 128
 129   /* If the structure address is not known, don't continue.  */
 130   if (info->all_image_addr == 0)
 131     return;
 132
 133   /* The structure has 4 fields: version (4 bytes), count (4 bytes),
 134      info (pointer) and notifier (pointer).  */
 135   len = 4 + 4 + 2 * TYPE_LENGTH (ptr_type);
 136   gdb_assert (len <= sizeof (buf));
 137   memset (&info->all_image, 0, sizeof (info->all_image));
 138
 139   /* Read structure raw bytes from target.  */
 140   if (target_read_memory (info->all_image_addr, buf, len))
 141     return;
 142
 143   /* Extract the fields.  */
 144   info->all_image.version = extract_unsigned_integer (buf, 4, byte_order);
 145   if (!darwin_dyld_version_ok (info))
 146     return;
 147
 148   info->all_image.count = extract_unsigned_integer (buf + 4, 4, byte_order);
 149   info->all_image.info = extract_typed_address (buf + 8, ptr_type);
 150   info->all_image.notifier = extract_typed_address
 151     (buf + 8 + TYPE_LENGTH (ptr_type), ptr_type);
 152 }
 153
 154 /* Link map info to include in an allocated so_list entry.  */
 155
 156 struct lm_info_darwin : public lm_info_base
 157 {
 158   /* The target location of lm.  */
 159   CORE_ADDR lm_addr = 0;
 160 };
 161
 162 /* Lookup the value for a specific symbol.  */
 163
 164 static CORE_ADDR
 165 lookup_symbol_from_bfd (bfd *abfd, const char *symname)
 166 {
 167   long storage_needed;
 168   asymbol **symbol_table;
 169   unsigned int number_of_symbols;
 170   unsigned int i;
 171   CORE_ADDR symaddr = 0;
 172
 173   storage_needed = bfd_get_symtab_upper_bound (abfd);
 174
 175   if (storage_needed <= 0)
 176     return 0;
 177
 178   symbol_table = (asymbol **) xmalloc (storage_needed);
 179   number_of_symbols = bfd_canonicalize_symtab (abfd, symbol_table);
 180
 181   for (i = 0; i < number_of_symbols; i++)
 182     {
 183       asymbol *sym = symbol_table[i];
 184
 185       if (strcmp (sym->name, symname) == 0
 186           && (sym->section->flags & (SEC_CODE | SEC_DATA)) != 0)
 187         {
 188           /* BFD symbols are section relative.  */
 189           symaddr = sym->value + sym->section->vma;
 190           break;
 191         }
 192     }
 193   xfree (symbol_table);
 194
 195   return symaddr;
 196 }
 197
 198 /* Return program interpreter string.  */
 199
 200 static char *
 201 find_program_interpreter (void)
 202 {
 203   char *buf = NULL;
 204
 205   /* If we have an exec_bfd, get the interpreter from the load commands.  */
 206   if (exec_bfd)
 207     {
 208       bfd_mach_o_load_command *cmd;
 209
 210       if (bfd_mach_o_lookup_command (exec_bfd,
 211                                      BFD_MACH_O_LC_LOAD_DYLINKER, &cmd) == 1)
 212         return cmd->command.dylinker.name_str;
 213     }
 214
 215   /* If we didn't find it, read from memory.
 216      FIXME: todo.  */
 217   return buf;
 218 }
 219
 220 /*  Not used.  I don't see how the main symbol file can be found: the
 221     interpreter name is needed and it is known from the executable file.
 222     Note that darwin-nat.c implements pid_to_exec_file.  */
 223
 224 static int
 225 open_symbol_file_object (int from_tty)
 226 {
 227   return 0;
 228 }
 229
 230 /* Build a list of currently loaded shared objects.  See solib-svr4.c.  */
 231
 232 static struct so_list *
 233 darwin_current_sos (void)
 234 {
 235   struct type *ptr_type = builtin_type (target_gdbarch ())->builtin_data_ptr;
 236   enum bfd_endian byte_order = gdbarch_byte_order (target_gdbarch ());
 237   int ptr_len = TYPE_LENGTH (ptr_type);
 238   unsigned int image_info_size;
 239   struct so_list *head = NULL;
 240   struct so_list *tail = NULL;
 241   int i;
 242   struct darwin_info *info = get_darwin_info ();
 243
 244   /* Be sure image infos are loaded.  */
 245   darwin_load_image_infos (info);
 246
 247   if (!darwin_dyld_version_ok (info))
 248     return NULL;
 249
 250   image_info_size = ptr_len * 3;
 251
 252   /* Read infos for each solib.
 253      The first entry was rumored to be the executable itself, but this is not
 254      true when a large number of shared libraries are used (table expanded ?).
 255      We now check all entries, but discard executable images.  */
 256   for (i = 0; i < info->all_image.count; i++)
 257     {
 258       CORE_ADDR iinfo = info->all_image.info + i * image_info_size;
 259       gdb_byte buf[image_info_size];
 260       CORE_ADDR load_addr;
 261       CORE_ADDR path_addr;
 262       struct mach_o_header_external hdr;
 263       unsigned long hdr_val;
 264       gdb::unique_xmalloc_ptr<char> file_path;
 265       int errcode;
 266
 267       /* Read image info from inferior.  */
 268       if (target_read_memory (iinfo, buf, image_info_size))
 269         break;
 270
 271       load_addr = extract_typed_address (buf, ptr_type);
 272       path_addr = extract_typed_address (buf + ptr_len, ptr_type);
 273
 274       /* Read Mach-O header from memory.  */
 275       if (target_read_memory (load_addr, (gdb_byte *) &hdr, sizeof (hdr) - 4))
 276         break;
 277       /* Discard wrong magic numbers.  Shouldn't happen.  */
 278       hdr_val = extract_unsigned_integer
 279         (hdr.magic, sizeof (hdr.magic), byte_order);
 280       if (hdr_val != BFD_MACH_O_MH_MAGIC && hdr_val != BFD_MACH_O_MH_MAGIC_64)
 281         continue;
 282       /* Discard executable.  Should happen only once.  */
 283       hdr_val = extract_unsigned_integer
 284         (hdr.filetype, sizeof (hdr.filetype), byte_order);
 285       if (hdr_val == BFD_MACH_O_MH_EXECUTE)
 286         continue;
 287
 288       target_read_string (path_addr, &file_path,
 289                           SO_NAME_MAX_PATH_SIZE - 1, &errcode);
 290       if (errcode)
 291         break;
 292
 293       /* Create and fill the new so_list element.  */
 294       gdb::unique_xmalloc_ptr<struct so_list> newobj (XCNEW (struct so_list));
 295
 296       lm_info_darwin *li = new lm_info_darwin;
 297       newobj->lm_info = li;
 298
 299       strncpy (newobj->so_name, file_path.get (), SO_NAME_MAX_PATH_SIZE - 1);
 300       newobj->so_name[SO_NAME_MAX_PATH_SIZE - 1] = '\0';
 301       strcpy (newobj->so_original_name, newobj->so_name);
 302       li->lm_addr = load_addr;
 303
 304       if (head == NULL)
 305         head = newobj.get ();
 306       else
 307         tail->next = newobj.get ();
 308       tail = newobj.release ();
 309     }
 310
 311   return head;
 312 }
 313
 314 /* Check LOAD_ADDR points to a Mach-O executable header.  Return LOAD_ADDR
 315    in case of success, 0 in case of failure.  */
 316
 317 static CORE_ADDR
 318 darwin_validate_exec_header (CORE_ADDR load_addr)
 319 {
 320   enum bfd_endian byte_order = gdbarch_byte_order (target_gdbarch ());
 321   struct mach_o_header_external hdr;
 322   unsigned long hdr_val;
 323
 324   /* Read Mach-O header from memory.  */
 325   if (target_read_memory (load_addr, (gdb_byte *) &hdr, sizeof (hdr) - 4))
 326     return 0;
 327
 328   /* Discard wrong magic numbers.  Shouldn't happen.  */
 329   hdr_val = extract_unsigned_integer
 330     (hdr.magic, sizeof (hdr.magic), byte_order);
 331   if (hdr_val != BFD_MACH_O_MH_MAGIC && hdr_val != BFD_MACH_O_MH_MAGIC_64)
 332     return 0;
 333
 334   /* Check executable.  */
 335   hdr_val = extract_unsigned_integer
 336     (hdr.filetype, sizeof (hdr.filetype), byte_order);
 337   if (hdr_val == BFD_MACH_O_MH_EXECUTE)
 338     return load_addr;
 339
 340   return 0;
 341 }
 342
 343 /* Get the load address of the executable using dyld list of images.
 344    We assume that the dyld info are correct (which is wrong if the target
 345    is stopped at the first instruction).  */
 346
 347 static CORE_ADDR
 348 darwin_read_exec_load_addr_from_dyld (struct darwin_info *info)
 349 {
 350   struct type *ptr_type = builtin_type (target_gdbarch ())->builtin_data_ptr;
 351   int ptr_len = TYPE_LENGTH (ptr_type);
 352   unsigned int image_info_size = ptr_len * 3;
 353   int i;
 354
 355   /* Read infos for each solib.  One of them should be the executable.  */
 356   for (i = 0; i < info->all_image.count; i++)
 357     {
 358       CORE_ADDR iinfo = info->all_image.info + i * image_info_size;
 359       gdb_byte buf[image_info_size];
 360       CORE_ADDR load_addr;
 361
 362       /* Read image info from inferior.  */
 363       if (target_read_memory (iinfo, buf, image_info_size))
 364         break;
 365
 366       load_addr = extract_typed_address (buf, ptr_type);
 367       if (darwin_validate_exec_header (load_addr) == load_addr)
 368         return load_addr;
 369     }
 370
 371   return 0;
 372 }
 373
 374 /* Get the load address of the executable when the PC is at the dyld
 375    entry point using parameter passed by the kernel (at SP). */
 376
 377 static CORE_ADDR
 378 darwin_read_exec_load_addr_at_init (struct darwin_info *info)
 379 {
 380   struct gdbarch *gdbarch = target_gdbarch ();
 381   enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
 382   int addr_size = gdbarch_addr_bit (gdbarch) / 8;
 383   ULONGEST load_ptr_addr;
 384   ULONGEST load_addr;
 385   gdb_byte buf[8];
 386
 387   /* Get SP.  */
 388   if (regcache_cooked_read_unsigned (get_current_regcache (),
 389                                      gdbarch_sp_regnum (gdbarch),
 390                                      &load_ptr_addr) != REG_VALID)
 391     return 0;
 392
 393   /* Read value at SP (image load address).  */
 394   if (target_read_memory (load_ptr_addr, buf, addr_size))
 395     return 0;
 396
 397   load_addr = extract_unsigned_integer (buf, addr_size, byte_order);
 398
 399   return darwin_validate_exec_header (load_addr);
 400 }
 401
 402 /* Return 1 if PC lies in the dynamic symbol resolution code of the
 403    run time loader.  */
 404
 405 static int
 406 darwin_in_dynsym_resolve_code (CORE_ADDR pc)
 407 {
 408   return 0;
 409 }
 410
 411 /* A wrapper for bfd_mach_o_fat_extract that handles reference
 412    counting properly.  This will either return NULL, or return a new
 413    reference to a BFD.  */
 414
 415 static gdb_bfd_ref_ptr
 416 gdb_bfd_mach_o_fat_extract (bfd *abfd, bfd_format format,
 417                             const bfd_arch_info_type *arch)
 418 {
 419   bfd *result = bfd_mach_o_fat_extract (abfd, format, arch);
 420
 421   if (result == NULL)
 422     return NULL;
 423
 424   if (result == abfd)
 425     gdb_bfd_ref (result);
 426   else
 427     gdb_bfd_mark_parent (result, abfd);
 428
 429   return gdb_bfd_ref_ptr (result);
 430 }
 431
 432 /* Extract dyld_all_image_addr when the process was just created, assuming the
 433    current PC is at the entry of the dynamic linker.  */
 434
 435 static void
 436 darwin_solib_get_all_image_info_addr_at_init (struct darwin_info *info)
 437 {
 438   char *interp_name;
 439   CORE_ADDR load_addr = 0;
 440
 441   /* This method doesn't work with an attached process.  */
 442   if (current_inferior ()->attach_flag)
 443     return;
 444
 445   /* Find the program interpreter.  */
 446   interp_name = find_program_interpreter ();
 447   if (!interp_name)
 448     return;
 449
 450   /* Create a bfd for the interpreter.  */
 451   gdb_bfd_ref_ptr dyld_bfd (gdb_bfd_open (interp_name, gnutarget, -1));
 452   if (dyld_bfd != NULL)
 453     {
 454       gdb_bfd_ref_ptr sub
 455         (gdb_bfd_mach_o_fat_extract (dyld_bfd.get (), bfd_object,
 456                                      gdbarch_bfd_arch_info (target_gdbarch ())));
 457       if (sub != NULL)
 458         dyld_bfd = sub;
 459       else
 460         dyld_bfd.release ();
 461     }
 462   if (dyld_bfd == NULL)
 463     return;
 464
 465   /* We find the dynamic linker's base address by examining
 466      the current pc (which should point at the entry point for the
 467      dynamic linker) and subtracting the offset of the entry point.  */
 468   load_addr = (regcache_read_pc (get_current_regcache ())
 469                - bfd_get_start_address (dyld_bfd.get ()));
 470
 471   /* Now try to set a breakpoint in the dynamic linker.  */
 472   info->all_image_addr =
 473     lookup_symbol_from_bfd (dyld_bfd.get (), "_dyld_all_image_infos");
 474
 475   if (info->all_image_addr == 0)
 476     return;
 477
 478   info->all_image_addr += load_addr;
 479 }
 480
 481 /* Extract dyld_all_image_addr reading it from
 482    TARGET_OBJECT_DARWIN_DYLD_INFO.  */
 483
 484 static void
 485 darwin_solib_read_all_image_info_addr (struct darwin_info *info)
 486 {
 487   gdb_byte buf[8];
 488   LONGEST len;
 489   struct type *ptr_type = builtin_type (target_gdbarch ())->builtin_data_ptr;
 490
 491   /* Sanity check.  */
 492   if (TYPE_LENGTH (ptr_type) > sizeof (buf))
 493     return;
 494
 495   len = target_read (current_top_target (), TARGET_OBJECT_DARWIN_DYLD_INFO,
 496                      NULL, buf, 0, TYPE_LENGTH (ptr_type));
 497   if (len <= 0)
 498     return;
 499
 500   /* The use of BIG endian is intended, as BUF is a raw stream of bytes.  This
 501       makes the support of remote protocol easier.  */
 502   info->all_image_addr = extract_unsigned_integer (buf, len, BFD_ENDIAN_BIG);
 503 }
 504
 505 /* Shared library startup support.  See documentation in solib-svr4.c.  */
 506
 507 static void
 508 darwin_solib_create_inferior_hook (int from_tty)
 509 {
 510   struct darwin_info *info = get_darwin_info ();
 511   CORE_ADDR load_addr;
 512
 513   info->all_image_addr = 0;
 514
 515   darwin_solib_read_all_image_info_addr (info);
 516
 517   if (info->all_image_addr == 0)
 518     darwin_solib_get_all_image_info_addr_at_init (info);
 519
 520   if (info->all_image_addr == 0)
 521     return;
 522
 523   darwin_load_image_infos (info);
 524
 525   if (!darwin_dyld_version_ok (info))
 526     {
 527       warning (_("unhandled dyld version (%d)"), info->all_image.version);
 528       return;
 529     }
 530
 531   /* Add the breakpoint which is hit by dyld when the list of solib is
 532      modified.  */
 533   create_solib_event_breakpoint (target_gdbarch (), info->all_image.notifier);
 534
 535   if (info->all_image.count != 0)
 536     {
 537       /* Possible relocate the main executable (PIE).  */
 538       load_addr = darwin_read_exec_load_addr_from_dyld (info);
 539     }
 540   else
 541     {
 542       /* Possible issue:
 543          Do not break on the notifier if dyld is not initialized (deduced from
 544          count == 0).  In that case, dyld hasn't relocated itself and the
 545          notifier may point to a wrong address.  */
 546
 547       load_addr = darwin_read_exec_load_addr_at_init (info);
 548     }
 549
 550   if (load_addr != 0 && symfile_objfile != NULL)
 551     {
 552       CORE_ADDR vmaddr;
 553
 554       /* Find the base address of the executable.  */
 555       vmaddr = bfd_mach_o_get_base_address (exec_bfd);
 556
 557       /* Relocate.  */
 558       if (vmaddr != load_addr)
 559         objfile_rebase (symfile_objfile, load_addr - vmaddr);
 560     }
 561 }
 562
 563 static void
 564 darwin_clear_solib (void)
 565 {
 566   struct darwin_info *info = get_darwin_info ();
 567
 568   info->all_image_addr = 0;
 569   info->all_image.version = 0;
 570 }
 571
 572 static void
 573 darwin_free_so (struct so_list *so)
 574 {
 575   lm_info_darwin *li = (lm_info_darwin *) so->lm_info;
 576
 577   delete li;
 578 }
 579
 580 /* The section table is built from bfd sections using bfd VMAs.
 581    Relocate these VMAs according to solib info.  */
 582
 583 static void
 584 darwin_relocate_section_addresses (struct so_list *so,
 585                                    struct target_section *sec)
 586 {
 587   lm_info_darwin *li = (lm_info_darwin *) so->lm_info;
 588
 589   sec->addr += li->lm_addr;
 590   sec->endaddr += li->lm_addr;
 591
 592   /* Best effort to set addr_high/addr_low.  This is used only by
 593      'info sharedlibary'.  */
 594   if (so->addr_high == 0)
 595     {
 596       so->addr_low = sec->addr;
 597       so->addr_high = sec->endaddr;
 598     }
 599   if (sec->endaddr > so->addr_high)
 600     so->addr_high = sec->endaddr;
 601   if (sec->addr < so->addr_low)
 602     so->addr_low = sec->addr;
 603 }
 604 \f
 605 static struct block_symbol
 606 darwin_lookup_lib_symbol (struct objfile *objfile,
 607                           const char *name,
 608                           const domain_enum domain)
 609 {
 610   return (struct block_symbol) {NULL, NULL};
 611 }
 612
 613 static gdb_bfd_ref_ptr
 614 darwin_bfd_open (const char *pathname)
 615 {
 616   int found_file;
 617
 618   /* Search for shared library file.  */
 619   gdb::unique_xmalloc_ptr<char> found_pathname
 620     = solib_find (pathname, &found_file);
 621   if (found_pathname == NULL)
 622     perror_with_name (pathname);
 623
 624   /* Open bfd for shared library.  */
 625   gdb_bfd_ref_ptr abfd (solib_bfd_fopen (found_pathname.get (), found_file));
 626
 627   gdb_bfd_ref_ptr res
 628     (gdb_bfd_mach_o_fat_extract (abfd.get (), bfd_object,
 629                                  gdbarch_bfd_arch_info (target_gdbarch ())));
 630   if (res == NULL)
 631     error (_("`%s': not a shared-library: %s"),
 632            bfd_get_filename (abfd.get ()), bfd_errmsg (bfd_get_error ()));
 633
 634   /* The current filename for fat-binary BFDs is a name generated
 635      by BFD, usually a string containing the name of the architecture.
 636      Reset its value to the actual filename.  */
 637   xfree (bfd_get_filename (res.get ()));
 638   res->filename = xstrdup (pathname);
 639
 640   return res;
 641 }
 642
 643 struct target_so_ops darwin_so_ops;
 644
 645 void
 646 _initialize_darwin_solib (void)
 647 {
 648   solib_darwin_pspace_data
 649     = register_program_space_data_with_cleanup (NULL,
 650                                                 darwin_pspace_data_cleanup);
 651
 652   darwin_so_ops.relocate_section_addresses = darwin_relocate_section_addresses;
 653   darwin_so_ops.free_so = darwin_free_so;
 654   darwin_so_ops.clear_solib = darwin_clear_solib;
 655   darwin_so_ops.solib_create_inferior_hook = darwin_solib_create_inferior_hook;
 656   darwin_so_ops.current_sos = darwin_current_sos;
 657   darwin_so_ops.open_symbol_file_object = open_symbol_file_object;
 658   darwin_so_ops.in_dynsym_resolve_code = darwin_in_dynsym_resolve_code;
 659   darwin_so_ops.lookup_lib_global_symbol = darwin_lookup_lib_symbol;
 660   darwin_so_ops.bfd_open = darwin_bfd_open;
 661 }