gdb/go-lang.c

   1 /* Go language support routines for GDB, the GNU debugger.
   2
   3    Copyright (C) 2012-2021 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 /* TODO:
  21    - split stacks
  22    - printing of native types
  23    - goroutines
  24    - lots more
  25    - gccgo mangling needs redoing
  26      It's too hard, for example, to know whether one is looking at a mangled
  27      Go symbol or not, and their are ambiguities, e.g., the demangler may
  28      get passed *any* symbol, including symbols from other languages
  29      and including symbols that are already demangled.
  30      One thought is to at least add an _G prefix.
  31    - 6g mangling isn't supported yet
  32 */
  33
  34 #include "defs.h"
  35 #include "gdb_obstack.h"
  36 #include "block.h"
  37 #include "symtab.h"
  38 #include "language.h"
  39 #include "varobj.h"
  40 #include "go-lang.h"
  41 #include "c-lang.h"
  42 #include "parser-defs.h"
  43 #include "gdbarch.h"
  44
  45 #include <ctype.h>
  46
  47 /* The main function in the main package.  */
  48 static const char GO_MAIN_MAIN[] = "main.main";
  49
  50 /* Function returning the special symbol name used by Go for the main
  51    procedure in the main program if it is found in minimal symbol list.
  52    This function tries to find minimal symbols so that it finds them even
  53    if the program was compiled without debugging information.  */
  54
  55 const char *
  56 go_main_name (void)
  57 {
  58   struct bound_minimal_symbol msym;
  59
  60   msym = lookup_minimal_symbol (GO_MAIN_MAIN, NULL, NULL);
  61   if (msym.minsym != NULL)
  62     return GO_MAIN_MAIN;
  63
  64   /* No known entry procedure found, the main program is probably not Go.  */
  65   return NULL;
  66 }
  67
  68 /* Return non-zero if TYPE is a gccgo string.
  69    We assume CHECK_TYPEDEF has already been done.  */
  70
  71 static int
  72 gccgo_string_p (struct type *type)
  73 {
  74   /* gccgo strings don't necessarily have a name we can use.  */
  75
  76   if (type->num_fields () == 2)
  77     {
  78       struct type *type0 = type->field (0).type ();
  79       struct type *type1 = type->field (1).type ();
  80
  81       type0 = check_typedef (type0);
  82       type1 = check_typedef (type1);
  83
  84       if (type0->code () == TYPE_CODE_PTR
  85           && strcmp (TYPE_FIELD_NAME (type, 0), "__data") == 0
  86           && type1->code () == TYPE_CODE_INT
  87           && strcmp (TYPE_FIELD_NAME (type, 1), "__length") == 0)
  88         {
  89           struct type *target_type = TYPE_TARGET_TYPE (type0);
  90
  91           target_type = check_typedef (target_type);
  92
  93           if (target_type->code () == TYPE_CODE_INT
  94               && TYPE_LENGTH (target_type) == 1
  95               && strcmp (target_type->name (), "uint8") == 0)
  96             return 1;
  97         }
  98     }
  99
 100   return 0;
 101 }
 102
 103 /* Return non-zero if TYPE is a 6g string.
 104    We assume CHECK_TYPEDEF has already been done.  */
 105
 106 static int
 107 sixg_string_p (struct type *type)
 108 {
 109   if (type->num_fields () == 2
 110       && type->name () != NULL
 111       && strcmp (type->name (), "string") == 0)
 112     return 1;
 113
 114   return 0;
 115 }
 116
 117 /* Classify the kind of Go object that TYPE is.
 118    TYPE is a TYPE_CODE_STRUCT, used to represent a Go object.  */
 119
 120 enum go_type
 121 go_classify_struct_type (struct type *type)
 122 {
 123   type = check_typedef (type);
 124
 125   /* Recognize strings as they're useful to be able to print without
 126      pretty-printers.  */
 127   if (gccgo_string_p (type)
 128       || sixg_string_p (type))
 129     return GO_TYPE_STRING;
 130
 131   return GO_TYPE_NONE;
 132 }
 133
 134 /* Subroutine of unpack_mangled_go_symbol to simplify it.
 135    Given "[foo.]bar.baz", store "bar" in *PACKAGEP and "baz" in *OBJECTP.
 136    We stomp on the last '.' to nul-terminate "bar".
 137    The caller is responsible for memory management.  */
 138
 139 static void
 140 unpack_package_and_object (char *buf,
 141                            const char **packagep, const char **objectp)
 142 {
 143   char *last_dot;
 144
 145   last_dot = strrchr (buf, '.');
 146   gdb_assert (last_dot != NULL);
 147   *objectp = last_dot + 1;
 148   *last_dot = '\0';
 149   last_dot = strrchr (buf, '.');
 150   if (last_dot != NULL)
 151     *packagep = last_dot + 1;
 152   else
 153     *packagep = buf;
 154 }
 155
 156 /* Given a mangled Go symbol, find its package name, object name, and
 157    method type (if present).
 158    E.g., for "libgo_net.textproto.String.N33_libgo_net.textproto.ProtocolError"
 159    *PACKAGEP = "textproto"
 160    *OBJECTP = "String"
 161    *METHOD_TYPE_PACKAGEP = "textproto"
 162    *METHOD_TYPE_OBJECTP = "ProtocolError"
 163
 164    Space for the resulting strings is malloc'd in one buffer.
 165    PACKAGEP,OBJECTP,METHOD_TYPE* will (typically) point into this buffer.
 166    [There are a few exceptions, but the caller is still responsible for
 167    freeing the resulting pointer.]
 168    A pointer to this buffer is returned, or NULL if symbol isn't a
 169    mangled Go symbol.
 170    The caller is responsible for freeing the result.
 171
 172    *METHOD_TYPE_IS_POINTERP is set to a boolean indicating if
 173    the method type is a pointer.
 174
 175    There may be value in returning the outer container,
 176    i.e., "net" in the above example, but for now it's not needed.
 177    Plus it's currently not straightforward to compute,
 178    it comes from -fgo-prefix, and there's no algorithm to compute it.
 179
 180    If we ever need to unpack the method type, this routine should work
 181    for that too.  */
 182
 183 static char *
 184 unpack_mangled_go_symbol (const char *mangled_name,
 185                           const char **packagep,
 186                           const char **objectp,
 187                           const char **method_type_packagep,
 188                           const char **method_type_objectp,
 189                           int *method_type_is_pointerp)
 190 {
 191   char *buf;
 192   char *p;
 193   int len = strlen (mangled_name);
 194   /* Pointer to last digit in "N<digit(s)>_".  */
 195   char *saw_digit;
 196   /* Pointer to "N" if valid "N<digit(s)>_" found.  */
 197   char *method_type;
 198   /* Pointer to the first '.'.  */
 199   const char *first_dot;
 200   /* Pointer to the last '.'.  */
 201   const char *last_dot;
 202   /* Non-zero if we saw a pointer indicator.  */
 203   int saw_pointer;
 204
 205   *packagep = *objectp = NULL;
 206   *method_type_packagep = *method_type_objectp = NULL;
 207   *method_type_is_pointerp = 0;
 208
 209   /* main.init is mangled specially.  */
 210   if (strcmp (mangled_name, "__go_init_main") == 0)
 211     {
 212       char *package = xstrdup ("main");
 213
 214       *packagep = package;
 215       *objectp = "init";
 216       return package;
 217     }
 218
 219   /* main.main is mangled specially (missing prefix).  */
 220   if (strcmp (mangled_name, "main.main") == 0)
 221     {
 222       char *package = xstrdup ("main");
 223
 224       *packagep = package;
 225       *objectp = "main";
 226       return package;
 227     }
 228
 229   /* We may get passed, e.g., "main.T.Foo", which is *not* mangled.
 230      Alas it looks exactly like "prefix.package.object."
 231      To cope for now we only recognize the following prefixes:
 232
 233      go: the default
 234      libgo_.*: used by gccgo's runtime
 235
 236      Thus we don't support -fgo-prefix (except as used by the runtime).  */
 237   if (!startswith (mangled_name, "go.")
 238       && !startswith (mangled_name, "libgo_"))
 239     return NULL;
 240
 241   /* Quick check for whether a search may be fruitful.  */
 242   /* Ignore anything with @plt, etc. in it.  */
 243   if (strchr (mangled_name, '@') != NULL)
 244     return NULL;
 245   /* It must have at least two dots.  */
 246   first_dot = strchr (mangled_name, '.');
 247   if (first_dot == NULL)
 248     return NULL;
 249   /* Treat "foo.bar" as unmangled.  It can collide with lots of other
 250      languages and it's not clear what the consequences are.
 251      And except for main.main, all gccgo symbols are at least
 252      prefix.package.object.  */
 253   last_dot = strrchr (mangled_name, '.');
 254   if (last_dot == first_dot)
 255     return NULL;
 256
 257   /* More quick checks.  */
 258   if (last_dot[1] == '\0' /* foo. */
 259       || last_dot[-1] == '.') /* foo..bar */
 260     return NULL;
 261
 262   /* At this point we've decided we have a mangled Go symbol.  */
 263
 264   buf = xstrdup (mangled_name);
 265
 266   /* Search backwards looking for "N<digit(s)>".  */
 267   p = buf + len;
 268   saw_digit = method_type = NULL;
 269   saw_pointer = 0;
 270   while (p > buf)
 271     {
 272       int current = *(const unsigned char *) --p;
 273       int current_is_digit = isdigit (current);
 274
 275       if (saw_digit)
 276         {
 277           if (current_is_digit)
 278             continue;
 279           if (current == 'N'
 280               && ((p > buf && p[-1] == '.')
 281                   || (p > buf + 1 && p[-1] == 'p' && p[-2] == '.')))
 282             {
 283               if (atoi (p + 1) == strlen (saw_digit + 2))
 284                 {
 285                   if (p[-1] == '.')
 286                     method_type = p - 1;
 287                   else
 288                     {
 289                       gdb_assert (p[-1] == 'p');
 290                       saw_pointer = 1;
 291                       method_type = p - 2;
 292                     }
 293                   break;
 294                 }
 295             }
 296           /* Not what we're looking for, reset and keep looking.  */
 297           saw_digit = NULL;
 298           saw_pointer = 0;
 299           continue;
 300         }
 301       if (current_is_digit && p[1] == '_')
 302         {
 303           /* Possible start of method "this" [sic] type.  */
 304           saw_digit = p;
 305           continue;
 306         }
 307     }
 308
 309   if (method_type != NULL
 310       /* Ensure not something like "..foo".  */
 311       && (method_type > buf && method_type[-1] != '.'))
 312     {
 313       unpack_package_and_object (saw_digit + 2,
 314                                  method_type_packagep, method_type_objectp);
 315       *method_type = '\0';
 316       *method_type_is_pointerp = saw_pointer;
 317     }
 318
 319   unpack_package_and_object (buf, packagep, objectp);
 320   return buf;
 321 }
 322
 323 /* Implements the la_demangle language_defn routine for language Go.
 324
 325    N.B. This may get passed *any* symbol, including symbols from other
 326    languages and including symbols that are already demangled.
 327    Both of these situations are kinda unfortunate, but that's how things
 328    are today.
 329
 330    N.B. This currently only supports gccgo's mangling.
 331
 332    N.B. gccgo's mangling needs, I think, changing.
 333    This demangler can't work in all situations,
 334    thus not too much effort is currently put into it.  */
 335
 336 char *
 337 go_language::demangle_symbol (const char *mangled_name, int options) const
 338 {
 339   struct obstack tempbuf;
 340   char *result;
 341   char *name_buf;
 342   const char *package_name;
 343   const char *object_name;
 344   const char *method_type_package_name;
 345   const char *method_type_object_name;
 346   int method_type_is_pointer;
 347
 348   if (mangled_name == NULL)
 349     return NULL;
 350
 351   name_buf = unpack_mangled_go_symbol (mangled_name,
 352                                        &package_name, &object_name,
 353                                        &method_type_package_name,
 354                                        &method_type_object_name,
 355                                        &method_type_is_pointer);
 356   if (name_buf == NULL)
 357     return NULL;
 358
 359   obstack_init (&tempbuf);
 360
 361   /* Print methods as they appear in "method expressions".  */
 362   if (method_type_package_name != NULL)
 363     {
 364       /* FIXME: Seems like we should include package_name here somewhere.  */
 365       if (method_type_is_pointer)
 366           obstack_grow_str (&tempbuf, "(*");
 367       obstack_grow_str (&tempbuf, method_type_package_name);
 368       obstack_grow_str (&tempbuf, ".");
 369       obstack_grow_str (&tempbuf, method_type_object_name);
 370       if (method_type_is_pointer)
 371         obstack_grow_str (&tempbuf, ")");
 372       obstack_grow_str (&tempbuf, ".");
 373       obstack_grow_str (&tempbuf, object_name);
 374     }
 375   else
 376     {
 377       obstack_grow_str (&tempbuf, package_name);
 378       obstack_grow_str (&tempbuf, ".");
 379       obstack_grow_str (&tempbuf, object_name);
 380     }
 381   obstack_grow_str0 (&tempbuf, "");
 382
 383   result = xstrdup ((const char *) obstack_finish (&tempbuf));
 384   obstack_free (&tempbuf, NULL);
 385   xfree (name_buf);
 386   return result;
 387 }
 388
 389 /* Given a Go symbol, return its package or NULL if unknown.
 390    Space for the result is malloc'd, caller must free.  */
 391
 392 char *
 393 go_symbol_package_name (const struct symbol *sym)
 394 {
 395   const char *mangled_name = sym->linkage_name ();
 396   const char *package_name;
 397   const char *object_name;
 398   const char *method_type_package_name;
 399   const char *method_type_object_name;
 400   int method_type_is_pointer;
 401   char *name_buf;
 402   char *result;
 403
 404   gdb_assert (sym->language () == language_go);
 405   name_buf = unpack_mangled_go_symbol (mangled_name,
 406                                        &package_name, &object_name,
 407                                        &method_type_package_name,
 408                                        &method_type_object_name,
 409                                        &method_type_is_pointer);
 410   /* Some Go symbols don't have mangled form we interpret (yet).  */
 411   if (name_buf == NULL)
 412     return NULL;
 413   result = xstrdup (package_name);
 414   xfree (name_buf);
 415   return result;
 416 }
 417
 418 /* Return the package that BLOCK is in, or NULL if there isn't one.
 419    Space for the result is malloc'd, caller must free.  */
 420
 421 char *
 422 go_block_package_name (const struct block *block)
 423 {
 424   while (block != NULL)
 425     {
 426       struct symbol *function = BLOCK_FUNCTION (block);
 427
 428       if (function != NULL)
 429         {
 430           char *package_name = go_symbol_package_name (function);
 431
 432           if (package_name != NULL)
 433             return package_name;
 434
 435           /* Stop looking if we find a function without a package name.
 436              We're most likely outside of Go and thus the concept of the
 437              "current" package is gone.  */
 438           return NULL;
 439         }
 440
 441       block = BLOCK_SUPERBLOCK (block);
 442     }
 443
 444   return NULL;
 445 }
 446
 447 /* See language.h.  */
 448
 449 void
 450 go_language::language_arch_info (struct gdbarch *gdbarch,
 451                                  struct language_arch_info *lai) const
 452 {
 453   const struct builtin_go_type *builtin = builtin_go_type (gdbarch);
 454
 455   /* Helper function to allow shorter lines below.  */
 456   auto add  = [&] (struct type * t) -> struct type *
 457   {
 458     lai->add_primitive_type (t);
 459     return t;
 460   };
 461
 462   add (builtin->builtin_void);
 463   add (builtin->builtin_char);
 464   add (builtin->builtin_bool);
 465   add (builtin->builtin_int);
 466   add (builtin->builtin_uint);
 467   add (builtin->builtin_uintptr);
 468   add (builtin->builtin_int8);
 469   add (builtin->builtin_int16);
 470   add (builtin->builtin_int32);
 471   add (builtin->builtin_int64);
 472   add (builtin->builtin_uint8);
 473   add (builtin->builtin_uint16);
 474   add (builtin->builtin_uint32);
 475   add (builtin->builtin_uint64);
 476   add (builtin->builtin_float32);
 477   add (builtin->builtin_float64);
 478   add (builtin->builtin_complex64);
 479   add (builtin->builtin_complex128);
 480
 481   lai->set_string_char_type (builtin->builtin_char);
 482   lai->set_bool_type (builtin->builtin_bool, "bool");
 483 }
 484
 485 /* Single instance of the Go language class.  */
 486
 487 static go_language go_language_defn;
 488
 489 static void *
 490 build_go_types (struct gdbarch *gdbarch)
 491 {
 492   struct builtin_go_type *builtin_go_type
 493     = GDBARCH_OBSTACK_ZALLOC (gdbarch, struct builtin_go_type);
 494
 495   builtin_go_type->builtin_void
 496     = arch_type (gdbarch, TYPE_CODE_VOID, TARGET_CHAR_BIT, "void");
 497   builtin_go_type->builtin_char
 498     = arch_character_type (gdbarch, 8, 1, "char");
 499   builtin_go_type->builtin_bool
 500     = arch_boolean_type (gdbarch, 8, 0, "bool");
 501   builtin_go_type->builtin_int
 502     = arch_integer_type (gdbarch, gdbarch_int_bit (gdbarch), 0, "int");
 503   builtin_go_type->builtin_uint
 504     = arch_integer_type (gdbarch, gdbarch_int_bit (gdbarch), 1, "uint");
 505   builtin_go_type->builtin_uintptr
 506     = arch_integer_type (gdbarch, gdbarch_ptr_bit (gdbarch), 1, "uintptr");
 507   builtin_go_type->builtin_int8
 508     = arch_integer_type (gdbarch, 8, 0, "int8");
 509   builtin_go_type->builtin_int16
 510     = arch_integer_type (gdbarch, 16, 0, "int16");
 511   builtin_go_type->builtin_int32
 512     = arch_integer_type (gdbarch, 32, 0, "int32");
 513   builtin_go_type->builtin_int64
 514     = arch_integer_type (gdbarch, 64, 0, "int64");
 515   builtin_go_type->builtin_uint8
 516     = arch_integer_type (gdbarch, 8, 1, "uint8");
 517   builtin_go_type->builtin_uint16
 518     = arch_integer_type (gdbarch, 16, 1, "uint16");
 519   builtin_go_type->builtin_uint32
 520     = arch_integer_type (gdbarch, 32, 1, "uint32");
 521   builtin_go_type->builtin_uint64
 522     = arch_integer_type (gdbarch, 64, 1, "uint64");
 523   builtin_go_type->builtin_float32
 524     = arch_float_type (gdbarch, 32, "float32", floatformats_ieee_single);
 525   builtin_go_type->builtin_float64
 526     = arch_float_type (gdbarch, 64, "float64", floatformats_ieee_double);
 527   builtin_go_type->builtin_complex64
 528     = init_complex_type ("complex64", builtin_go_type->builtin_float32);
 529   builtin_go_type->builtin_complex128
 530     = init_complex_type ("complex128", builtin_go_type->builtin_float64);
 531
 532   return builtin_go_type;
 533 }
 534
 535 static struct gdbarch_data *go_type_data;
 536
 537 const struct builtin_go_type *
 538 builtin_go_type (struct gdbarch *gdbarch)
 539 {
 540   return (const struct builtin_go_type *) gdbarch_data (gdbarch, go_type_data);
 541 }
 542
 543 void _initialize_go_language ();
 544 void
 545 _initialize_go_language ()
 546 {
 547   go_type_data = gdbarch_data_register_post_init (build_go_types);
 548 }