gdb/parse.c

   1 /* Parse expressions for GDB.
   2
   3    Copyright (C) 1986-2023 Free Software Foundation, Inc.
   4
   5    Modified from expread.y by the Department of Computer Science at the
   6    State University of New York at Buffalo, 1991.
   7
   8    This file is part of GDB.
   9
  10    This program is free software; you can redistribute it and/or modify
  11    it under the terms of the GNU General Public License as published by
  12    the Free Software Foundation; either version 3 of the License, or
  13    (at your option) any later version.
  14
  15    This program is distributed in the hope that it will be useful,
  16    but WITHOUT ANY WARRANTY; without even the implied warranty of
  17    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18    GNU General Public License for more details.
  19
  20    You should have received a copy of the GNU General Public License
  21    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
  22
  23 /* Parse an expression from text in a string,
  24    and return the result as a struct expression pointer.
  25    That structure contains arithmetic operations in reverse polish,
  26    with constants represented by operations that are followed by special data.
  27    See expression.h for the details of the format.
  28    What is important here is that it can be built up sequentially
  29    during the process of parsing; the lower levels of the tree always
  30    come first in the result.  */
  31
  32 #include "defs.h"
  33 #include <ctype.h>
  34 #include "arch-utils.h"
  35 #include "symtab.h"
  36 #include "gdbtypes.h"
  37 #include "frame.h"
  38 #include "expression.h"
  39 #include "value.h"
  40 #include "command.h"
  41 #include "language.h"
  42 #include "parser-defs.h"
  43 #include "gdbcmd.h"
  44 #include "symfile.h"            /* for overlay functions */
  45 #include "inferior.h"
  46 #include "target-float.h"
  47 #include "block.h"
  48 #include "source.h"
  49 #include "objfiles.h"
  50 #include "user-regs.h"
  51 #include <algorithm>
  52 #include "gdbsupport/gdb_optional.h"
  53 #include "c-exp.h"
  54
  55 static unsigned int expressiondebug = 0;
  56 static void
  57 show_expressiondebug (struct ui_file *file, int from_tty,
  58                       struct cmd_list_element *c, const char *value)
  59 {
  60   gdb_printf (file, _("Expression debugging is %s.\n"), value);
  61 }
  62
  63
  64 /* True if an expression parser should set yydebug.  */
  65 bool parser_debug;
  66
  67 static void
  68 show_parserdebug (struct ui_file *file, int from_tty,
  69                   struct cmd_list_element *c, const char *value)
  70 {
  71   gdb_printf (file, _("Parser debugging is %s.\n"), value);
  72 }
  73
  74
  75 static expression_up parse_exp_in_context
  76      (const char **, CORE_ADDR,
  77       const struct block *, int,
  78       bool, innermost_block_tracker *,
  79       std::unique_ptr<expr_completion_base> *);
  80
  81 /* Documented at it's declaration.  */
  82
  83 void
  84 innermost_block_tracker::update (const struct block *b,
  85                                  innermost_block_tracker_types t)
  86 {
  87   if ((m_types & t) != 0
  88       && (m_innermost_block == NULL
  89           || contained_in (b, m_innermost_block)))
  90     m_innermost_block = b;
  91 }
  92
  93 \f
  94
  95 /* Return the type of MSYMBOL, a minimal symbol of OBJFILE.  If
  96    ADDRESS_P is not NULL, set it to the MSYMBOL's resolved
  97    address.  */
  98
  99 type *
 100 find_minsym_type_and_address (minimal_symbol *msymbol,
 101                               struct objfile *objfile,
 102                               CORE_ADDR *address_p)
 103 {
 104   bound_minimal_symbol bound_msym = {msymbol, objfile};
 105   struct obj_section *section = msymbol->obj_section (objfile);
 106   enum minimal_symbol_type type = msymbol->type ();
 107
 108   bool is_tls = (section != NULL
 109                  && section->the_bfd_section->flags & SEC_THREAD_LOCAL);
 110
 111   /* The minimal symbol might point to a function descriptor;
 112      resolve it to the actual code address instead.  */
 113   CORE_ADDR addr;
 114   if (is_tls)
 115     {
 116       /* Addresses of TLS symbols are really offsets into a
 117          per-objfile/per-thread storage block.  */
 118       addr = bound_msym.minsym->value_raw_address ();
 119     }
 120   else if (msymbol_is_function (objfile, msymbol, &addr))
 121     {
 122       if (addr != bound_msym.value_address ())
 123         {
 124           /* This means we resolved a function descriptor, and we now
 125              have an address for a code/text symbol instead of a data
 126              symbol.  */
 127           if (msymbol->type () == mst_data_gnu_ifunc)
 128             type = mst_text_gnu_ifunc;
 129           else
 130             type = mst_text;
 131           section = NULL;
 132         }
 133     }
 134   else
 135     addr = bound_msym.value_address ();
 136
 137   if (overlay_debugging)
 138     addr = symbol_overlayed_address (addr, section);
 139
 140   if (is_tls)
 141     {
 142       /* Skip translation if caller does not need the address.  */
 143       if (address_p != NULL)
 144         *address_p = target_translate_tls_address (objfile, addr);
 145       return objfile_type (objfile)->nodebug_tls_symbol;
 146     }
 147
 148   if (address_p != NULL)
 149     *address_p = addr;
 150
 151   switch (type)
 152     {
 153     case mst_text:
 154     case mst_file_text:
 155     case mst_solib_trampoline:
 156       return objfile_type (objfile)->nodebug_text_symbol;
 157
 158     case mst_text_gnu_ifunc:
 159       return objfile_type (objfile)->nodebug_text_gnu_ifunc_symbol;
 160
 161     case mst_data:
 162     case mst_file_data:
 163     case mst_bss:
 164     case mst_file_bss:
 165       return objfile_type (objfile)->nodebug_data_symbol;
 166
 167     case mst_slot_got_plt:
 168       return objfile_type (objfile)->nodebug_got_plt_symbol;
 169
 170     default:
 171       return objfile_type (objfile)->nodebug_unknown_symbol;
 172     }
 173 }
 174
 175 bool
 176 expr_complete_tag::complete (struct expression *exp,
 177                              completion_tracker &tracker)
 178 {
 179   collect_symbol_completion_matches_type (tracker, m_name.get (),
 180                                           m_name.get (), m_code);
 181   return true;
 182 }
 183
 184 /* See parser-defs.h.  */
 185
 186 void
 187 parser_state::mark_struct_expression (expr::structop_base_operation *op)
 188 {
 189   gdb_assert (parse_completion && m_completion_state == nullptr);
 190   m_completion_state.reset (new expr_complete_structop (op));
 191 }
 192
 193 /* Indicate that the current parser invocation is completing a tag.
 194    TAG is the type code of the tag, and PTR and LENGTH represent the
 195    start of the tag name.  */
 196
 197 void
 198 parser_state::mark_completion_tag (enum type_code tag, const char *ptr,
 199                                    int length)
 200 {
 201   gdb_assert (parse_completion && m_completion_state == nullptr);
 202   gdb_assert (tag == TYPE_CODE_UNION
 203               || tag == TYPE_CODE_STRUCT
 204               || tag == TYPE_CODE_ENUM);
 205   m_completion_state.reset
 206     (new expr_complete_tag (tag, make_unique_xstrndup (ptr, length)));
 207 }
 208
 209 /* See parser-defs.h.  */
 210
 211 void
 212 parser_state::push_c_string (int kind, struct stoken_vector *vec)
 213 {
 214   std::vector<std::string> data (vec->len);
 215   for (int i = 0; i < vec->len; ++i)
 216     data[i] = std::string (vec->tokens[i].ptr, vec->tokens[i].length);
 217
 218   push_new<expr::c_string_operation> ((enum c_string_type_values) kind,
 219                                       std::move (data));
 220 }
 221
 222 /* See parser-defs.h.  */
 223
 224 void
 225 parser_state::push_symbol (const char *name, block_symbol sym)
 226 {
 227   if (sym.symbol != nullptr)
 228     {
 229       if (symbol_read_needs_frame (sym.symbol))
 230         block_tracker->update (sym);
 231       push_new<expr::var_value_operation> (sym);
 232     }
 233   else
 234     {
 235       struct bound_minimal_symbol msymbol = lookup_bound_minimal_symbol (name);
 236       if (msymbol.minsym != NULL)
 237         push_new<expr::var_msym_value_operation> (msymbol);
 238       else if (!have_full_symbols () && !have_partial_symbols ())
 239         error (_("No symbol table is loaded.  Use the \"file\" command."));
 240       else
 241         error (_("No symbol \"%s\" in current context."), name);
 242     }
 243 }
 244
 245 /* See parser-defs.h.  */
 246
 247 void
 248 parser_state::push_dollar (struct stoken str)
 249 {
 250   struct block_symbol sym;
 251   struct bound_minimal_symbol msym;
 252   struct internalvar *isym = NULL;
 253   std::string copy;
 254
 255   /* Handle the tokens $digits; also $ (short for $0) and $$ (short for $$1)
 256      and $$digits (equivalent to $<-digits> if you could type that).  */
 257
 258   int negate = 0;
 259   int i = 1;
 260   /* Double dollar means negate the number and add -1 as well.
 261      Thus $$ alone means -1.  */
 262   if (str.length >= 2 && str.ptr[1] == '$')
 263     {
 264       negate = 1;
 265       i = 2;
 266     }
 267   if (i == str.length)
 268     {
 269       /* Just dollars (one or two).  */
 270       i = -negate;
 271       goto handle_last;
 272     }
 273   /* Is the rest of the token digits?  */
 274   for (; i < str.length; i++)
 275     if (!(str.ptr[i] >= '0' && str.ptr[i] <= '9'))
 276       break;
 277   if (i == str.length)
 278     {
 279       i = atoi (str.ptr + 1 + negate);
 280       if (negate)
 281         i = -i;
 282       goto handle_last;
 283     }
 284
 285   /* Handle tokens that refer to machine registers:
 286      $ followed by a register name.  */
 287   i = user_reg_map_name_to_regnum (gdbarch (),
 288                                    str.ptr + 1, str.length - 1);
 289   if (i >= 0)
 290     goto handle_register;
 291
 292   /* Any names starting with $ are probably debugger internal variables.  */
 293
 294   copy = copy_name (str);
 295   isym = lookup_only_internalvar (copy.c_str () + 1);
 296   if (isym)
 297     {
 298       push_new<expr::internalvar_operation> (isym);
 299       return;
 300     }
 301
 302   /* On some systems, such as HP-UX and hppa-linux, certain system routines
 303      have names beginning with $ or $$.  Check for those, first.  */
 304
 305   sym = lookup_symbol (copy.c_str (), NULL, VAR_DOMAIN, NULL);
 306   if (sym.symbol)
 307     {
 308       push_new<expr::var_value_operation> (sym);
 309       return;
 310     }
 311   msym = lookup_bound_minimal_symbol (copy.c_str ());
 312   if (msym.minsym)
 313     {
 314       push_new<expr::var_msym_value_operation> (msym);
 315       return;
 316     }
 317
 318   /* Any other names are assumed to be debugger internal variables.  */
 319
 320   push_new<expr::internalvar_operation>
 321     (create_internalvar (copy.c_str () + 1));
 322   return;
 323 handle_last:
 324   push_new<expr::last_operation> (i);
 325   return;
 326 handle_register:
 327   str.length--;
 328   str.ptr++;
 329   push_new<expr::register_operation> (copy_name (str));
 330   block_tracker->update (expression_context_block,
 331                          INNERMOST_BLOCK_FOR_REGISTERS);
 332   return;
 333 }
 334
 335 \f
 336
 337 const char *
 338 find_template_name_end (const char *p)
 339 {
 340   int depth = 1;
 341   int just_seen_right = 0;
 342   int just_seen_colon = 0;
 343   int just_seen_space = 0;
 344
 345   if (!p || (*p != '<'))
 346     return 0;
 347
 348   while (*++p)
 349     {
 350       switch (*p)
 351         {
 352         case '\'':
 353         case '\"':
 354         case '{':
 355         case '}':
 356           /* In future, may want to allow these??  */
 357           return 0;
 358         case '<':
 359           depth++;              /* start nested template */
 360           if (just_seen_colon || just_seen_right || just_seen_space)
 361             return 0;           /* but not after : or :: or > or space */
 362           break;
 363         case '>':
 364           if (just_seen_colon || just_seen_right)
 365             return 0;           /* end a (nested?) template */
 366           just_seen_right = 1;  /* but not after : or :: */
 367           if (--depth == 0)     /* also disallow >>, insist on > > */
 368             return ++p;         /* if outermost ended, return */
 369           break;
 370         case ':':
 371           if (just_seen_space || (just_seen_colon > 1))
 372             return 0;           /* nested class spec coming up */
 373           just_seen_colon++;    /* we allow :: but not :::: */
 374           break;
 375         case ' ':
 376           break;
 377         default:
 378           if (!((*p >= 'a' && *p <= 'z') ||     /* allow token chars */
 379                 (*p >= 'A' && *p <= 'Z') ||
 380                 (*p >= '0' && *p <= '9') ||
 381                 (*p == '_') || (*p == ',') ||   /* commas for template args */
 382                 (*p == '&') || (*p == '*') ||   /* pointer and ref types */
 383                 (*p == '(') || (*p == ')') ||   /* function types */
 384                 (*p == '[') || (*p == ']')))    /* array types */
 385             return 0;
 386         }
 387       if (*p != ' ')
 388         just_seen_space = 0;
 389       if (*p != ':')
 390         just_seen_colon = 0;
 391       if (*p != '>')
 392         just_seen_right = 0;
 393     }
 394   return 0;
 395 }
 396 \f
 397
 398 /* Return a null-terminated temporary copy of the name of a string token.
 399
 400    Tokens that refer to names do so with explicit pointer and length,
 401    so they can share the storage that lexptr is parsing.
 402    When it is necessary to pass a name to a function that expects
 403    a null-terminated string, the substring is copied out
 404    into a separate block of storage.  */
 405
 406 std::string
 407 copy_name (struct stoken token)
 408 {
 409   return std::string (token.ptr, token.length);
 410 }
 411 \f
 412
 413 /* Read an expression from the string *STRINGPTR points to,
 414    parse it, and return a pointer to a struct expression that we malloc.
 415    Use block BLOCK as the lexical context for variable names;
 416    if BLOCK is zero, use the block of the selected stack frame.
 417    Meanwhile, advance *STRINGPTR to point after the expression,
 418    at the first nonwhite character that is not part of the expression
 419    (possibly a null character).
 420
 421    If COMMA is nonzero, stop if a comma is reached.  */
 422
 423 expression_up
 424 parse_exp_1 (const char **stringptr, CORE_ADDR pc, const struct block *block,
 425              int comma, innermost_block_tracker *tracker)
 426 {
 427   return parse_exp_in_context (stringptr, pc, block, comma, false,
 428                                tracker, nullptr);
 429 }
 430
 431 /* As for parse_exp_1, except that if VOID_CONTEXT_P, then
 432    no value is expected from the expression.  */
 433
 434 static expression_up
 435 parse_exp_in_context (const char **stringptr, CORE_ADDR pc,
 436                       const struct block *block,
 437                       int comma, bool void_context_p,
 438                       innermost_block_tracker *tracker,
 439                       std::unique_ptr<expr_completion_base> *completer)
 440 {
 441   const struct language_defn *lang = NULL;
 442
 443   if (*stringptr == 0 || **stringptr == 0)
 444     error_no_arg (_("expression to compute"));
 445
 446   const struct block *expression_context_block = block;
 447   CORE_ADDR expression_context_pc = 0;
 448
 449   innermost_block_tracker local_tracker;
 450   if (tracker == nullptr)
 451     tracker = &local_tracker;
 452
 453   /* If no context specified, try using the current frame, if any.  */
 454   if (!expression_context_block)
 455     expression_context_block = get_selected_block (&expression_context_pc);
 456   else if (pc == 0)
 457     expression_context_pc = expression_context_block->entry_pc ();
 458   else
 459     expression_context_pc = pc;
 460
 461   /* Fall back to using the current source static context, if any.  */
 462
 463   if (!expression_context_block)
 464     {
 465       struct symtab_and_line cursal = get_current_source_symtab_and_line ();
 466
 467       if (cursal.symtab)
 468         expression_context_block
 469           = cursal.symtab->compunit ()->blockvector ()->static_block ();
 470
 471       if (expression_context_block)
 472         expression_context_pc = expression_context_block->entry_pc ();
 473     }
 474
 475   if (language_mode == language_mode_auto && block != NULL)
 476     {
 477       /* Find the language associated to the given context block.
 478          Default to the current language if it can not be determined.
 479
 480          Note that using the language corresponding to the current frame
 481          can sometimes give unexpected results.  For instance, this
 482          routine is often called several times during the inferior
 483          startup phase to re-parse breakpoint expressions after
 484          a new shared library has been loaded.  The language associated
 485          to the current frame at this moment is not relevant for
 486          the breakpoint.  Using it would therefore be silly, so it seems
 487          better to rely on the current language rather than relying on
 488          the current frame language to parse the expression.  That's why
 489          we do the following language detection only if the context block
 490          has been specifically provided.  */
 491       struct symbol *func = block_linkage_function (block);
 492
 493       if (func != NULL)
 494         lang = language_def (func->language ());
 495       if (lang == NULL || lang->la_language == language_unknown)
 496         lang = current_language;
 497     }
 498   else
 499     lang = current_language;
 500
 501   /* get_current_arch may reset CURRENT_LANGUAGE via select_frame.
 502      While we need CURRENT_LANGUAGE to be set to LANG (for lookup_symbol
 503      and others called from *.y) ensure CURRENT_LANGUAGE gets restored
 504      to the value matching SELECTED_FRAME as set by get_current_arch.  */
 505
 506   parser_state ps (lang, get_current_arch (), expression_context_block,
 507                    expression_context_pc, comma, *stringptr,
 508                    completer != nullptr, tracker, void_context_p);
 509
 510   scoped_restore_current_language lang_saver;
 511   set_language (lang->la_language);
 512
 513   try
 514     {
 515       lang->parser (&ps);
 516     }
 517   catch (const gdb_exception &except)
 518     {
 519       /* If parsing for completion, allow this to succeed; but if no
 520          expression elements have been written, then there's nothing
 521          to do, so fail.  */
 522       if (! ps.parse_completion || ps.expout->op == nullptr)
 523         throw;
 524     }
 525
 526   expression_up result = ps.release ();
 527   result->op->set_outermost ();
 528
 529   if (expressiondebug)
 530     result->dump (gdb_stdlog);
 531
 532   if (completer != nullptr)
 533     *completer = std::move (ps.m_completion_state);
 534   *stringptr = ps.lexptr;
 535   return result;
 536 }
 537
 538 /* Parse STRING as an expression, and complain if this fails to use up
 539    all of the contents of STRING.  TRACKER, if non-null, will be
 540    updated by the parser.  VOID_CONTEXT_P should be true to indicate
 541    that the expression may be expected to return a value with void
 542    type.  Parsers are free to ignore this, or to use it to help with
 543    overload resolution decisions.  */
 544
 545 expression_up
 546 parse_expression (const char *string, innermost_block_tracker *tracker,
 547                   bool void_context_p)
 548 {
 549   expression_up exp = parse_exp_in_context (&string, 0, nullptr, 0,
 550                                             void_context_p,
 551                                             tracker, nullptr);
 552   if (*string)
 553     error (_("Junk after end of expression."));
 554   return exp;
 555 }
 556
 557 /* Same as parse_expression, but using the given language (LANG)
 558    to parse the expression.  */
 559
 560 expression_up
 561 parse_expression_with_language (const char *string, enum language lang)
 562 {
 563   gdb::optional<scoped_restore_current_language> lang_saver;
 564   if (current_language->la_language != lang)
 565     {
 566       lang_saver.emplace ();
 567       set_language (lang);
 568     }
 569
 570   return parse_expression (string);
 571 }
 572
 573 /* Parse STRING as an expression.  If the parse is marked for
 574    completion, set COMPLETER and return the expression.  In all other
 575    cases, return NULL.  */
 576
 577 expression_up
 578 parse_expression_for_completion
 579      (const char *string,
 580       std::unique_ptr<expr_completion_base> *completer)
 581 {
 582   expression_up exp;
 583
 584   try
 585     {
 586       exp = parse_exp_in_context (&string, 0, 0, 0, false, nullptr, completer);
 587     }
 588   catch (const gdb_exception_error &except)
 589     {
 590       /* Nothing, EXP remains NULL.  */
 591     }
 592
 593   /* If we didn't get a completion result, be sure to also not return
 594      an expression to our caller.  */
 595   if (*completer == nullptr)
 596     return nullptr;
 597
 598   return exp;
 599 }
 600
 601 /* Parse floating point value P of length LEN.
 602    Return false if invalid, true if valid.
 603    The successfully parsed number is stored in DATA in
 604    target format for floating-point type TYPE.
 605
 606    NOTE: This accepts the floating point syntax that sscanf accepts.  */
 607
 608 bool
 609 parse_float (const char *p, int len,
 610              const struct type *type, gdb_byte *data)
 611 {
 612   return target_float_from_string (data, type, std::string (p, len));
 613 }
 614
 615 /* Return true if the number N_SIGN * N fits in a type with TYPE_BITS and
 616    TYPE_SIGNED_P.  N_SIGNED is either 1 or -1.  */
 617
 618 bool
 619 fits_in_type (int n_sign, ULONGEST n, int type_bits, bool type_signed_p)
 620 {
 621   /* Normalize -0.  */
 622   if (n == 0 && n_sign == -1)
 623     n_sign = 1;
 624
 625   if (n_sign == -1 && !type_signed_p)
 626     /* Can't fit a negative number in an unsigned type.  */
 627     return false;
 628
 629   if (type_bits > sizeof (ULONGEST) * 8)
 630     return true;
 631
 632   ULONGEST smax = (ULONGEST)1 << (type_bits - 1);
 633   if (n_sign == -1)
 634     {
 635       /* Negative number, signed type.  */
 636       return (n <= smax);
 637     }
 638   else if (n_sign == 1 && type_signed_p)
 639     {
 640       /* Positive number, signed type.  */
 641       return (n < smax);
 642     }
 643   else if (n_sign == 1 && !type_signed_p)
 644     {
 645       /* Positive number, unsigned type.  */
 646       return ((n >> 1) >> (type_bits - 1)) == 0;
 647     }
 648   else
 649     gdb_assert_not_reached ("");
 650 }
 651 \f
 652 /* This function avoids direct calls to fprintf
 653    in the parser generated debug code.  */
 654 void
 655 parser_fprintf (FILE *x, const char *y, ...)
 656 {
 657   va_list args;
 658
 659   va_start (args, y);
 660   if (x == stderr)
 661     gdb_vprintf (gdb_stderr, y, args);
 662   else
 663     {
 664       gdb_printf (gdb_stderr, " Unknown FILE used.\n");
 665       gdb_vprintf (gdb_stderr, y, args);
 666     }
 667   va_end (args);
 668 }
 669
 670 /* Return rue if EXP uses OBJFILE (and will become dangling when
 671    OBJFILE is unloaded), otherwise return false.  OBJFILE must not be
 672    a separate debug info file.  */
 673
 674 bool
 675 exp_uses_objfile (struct expression *exp, struct objfile *objfile)
 676 {
 677   gdb_assert (objfile->separate_debug_objfile_backlink == NULL);
 678
 679   return exp->op->uses_objfile (objfile);
 680 }
 681
 682 void _initialize_parse ();
 683 void
 684 _initialize_parse ()
 685 {
 686   add_setshow_zuinteger_cmd ("expression", class_maintenance,
 687                              &expressiondebug,
 688                              _("Set expression debugging."),
 689                              _("Show expression debugging."),
 690                              _("When non-zero, the internal representation "
 691                                "of expressions will be printed."),
 692                              NULL,
 693                              show_expressiondebug,
 694                              &setdebuglist, &showdebuglist);
 695   add_setshow_boolean_cmd ("parser", class_maintenance,
 696                             &parser_debug,
 697                            _("Set parser debugging."),
 698                            _("Show parser debugging."),
 699                            _("When non-zero, expression parser "
 700                              "tracing will be enabled."),
 701                             NULL,
 702                             show_parserdebug,
 703                             &setdebuglist, &showdebuglist);
 704 }