gdb/parse.c

   1 /* Parse expressions for GDB.
   2
   3    Copyright (C) 1986-2022 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 (const char **, CORE_ADDR,
  76                                            const struct block *, int,
  77                                            bool, innermost_block_tracker *,
  78                                            expr_completion_state *);
  79
  80 /* Documented at it's declaration.  */
  81
  82 void
  83 innermost_block_tracker::update (const struct block *b,
  84                                  innermost_block_tracker_types t)
  85 {
  86   if ((m_types & t) != 0
  87       && (m_innermost_block == NULL
  88           || contained_in (b, m_innermost_block)))
  89     m_innermost_block = b;
  90 }
  91
  92 \f
  93
  94 /* Return the type of MSYMBOL, a minimal symbol of OBJFILE.  If
  95    ADDRESS_P is not NULL, set it to the MSYMBOL's resolved
  96    address.  */
  97
  98 type *
  99 find_minsym_type_and_address (minimal_symbol *msymbol,
 100                               struct objfile *objfile,
 101                               CORE_ADDR *address_p)
 102 {
 103   bound_minimal_symbol bound_msym = {msymbol, objfile};
 104   struct obj_section *section = msymbol->obj_section (objfile);
 105   enum minimal_symbol_type type = MSYMBOL_TYPE (msymbol);
 106
 107   bool is_tls = (section != NULL
 108                  && section->the_bfd_section->flags & SEC_THREAD_LOCAL);
 109
 110   /* The minimal symbol might point to a function descriptor;
 111      resolve it to the actual code address instead.  */
 112   CORE_ADDR addr;
 113   if (is_tls)
 114     {
 115       /* Addresses of TLS symbols are really offsets into a
 116          per-objfile/per-thread storage block.  */
 117       addr = MSYMBOL_VALUE_RAW_ADDRESS (bound_msym.minsym);
 118     }
 119   else if (msymbol_is_function (objfile, msymbol, &addr))
 120     {
 121       if (addr != BMSYMBOL_VALUE_ADDRESS (bound_msym))
 122         {
 123           /* This means we resolved a function descriptor, and we now
 124              have an address for a code/text symbol instead of a data
 125              symbol.  */
 126           if (MSYMBOL_TYPE (msymbol) == mst_data_gnu_ifunc)
 127             type = mst_text_gnu_ifunc;
 128           else
 129             type = mst_text;
 130           section = NULL;
 131         }
 132     }
 133   else
 134     addr = BMSYMBOL_VALUE_ADDRESS (bound_msym);
 135
 136   if (overlay_debugging)
 137     addr = symbol_overlayed_address (addr, section);
 138
 139   if (is_tls)
 140     {
 141       /* Skip translation if caller does not need the address.  */
 142       if (address_p != NULL)
 143         *address_p = target_translate_tls_address (objfile, addr);
 144       return objfile_type (objfile)->nodebug_tls_symbol;
 145     }
 146
 147   if (address_p != NULL)
 148     *address_p = addr;
 149
 150   switch (type)
 151     {
 152     case mst_text:
 153     case mst_file_text:
 154     case mst_solib_trampoline:
 155       return objfile_type (objfile)->nodebug_text_symbol;
 156
 157     case mst_text_gnu_ifunc:
 158       return objfile_type (objfile)->nodebug_text_gnu_ifunc_symbol;
 159
 160     case mst_data:
 161     case mst_file_data:
 162     case mst_bss:
 163     case mst_file_bss:
 164       return objfile_type (objfile)->nodebug_data_symbol;
 165
 166     case mst_slot_got_plt:
 167       return objfile_type (objfile)->nodebug_got_plt_symbol;
 168
 169     default:
 170       return objfile_type (objfile)->nodebug_unknown_symbol;
 171     }
 172 }
 173
 174 /* See parser-defs.h.  */
 175
 176 void
 177 parser_state::mark_struct_expression (expr::structop_base_operation *op)
 178 {
 179   gdb_assert (parse_completion
 180               && (m_completion_state.expout_tag_completion_type
 181                   == TYPE_CODE_UNDEF));
 182   m_completion_state.expout_last_op = op;
 183 }
 184
 185 /* Indicate that the current parser invocation is completing a tag.
 186    TAG is the type code of the tag, and PTR and LENGTH represent the
 187    start of the tag name.  */
 188
 189 void
 190 parser_state::mark_completion_tag (enum type_code tag, const char *ptr,
 191                                    int length)
 192 {
 193   gdb_assert (parse_completion
 194               && (m_completion_state.expout_tag_completion_type
 195                   == TYPE_CODE_UNDEF)
 196               && m_completion_state.expout_completion_name == NULL
 197               && m_completion_state.expout_last_op == nullptr);
 198   gdb_assert (tag == TYPE_CODE_UNION
 199               || tag == TYPE_CODE_STRUCT
 200               || tag == TYPE_CODE_ENUM);
 201   m_completion_state.expout_tag_completion_type = tag;
 202   m_completion_state.expout_completion_name
 203     = make_unique_xstrndup (ptr, length);
 204 }
 205
 206 /* See parser-defs.h.  */
 207
 208 void
 209 parser_state::push_c_string (int kind, struct stoken_vector *vec)
 210 {
 211   std::vector<std::string> data (vec->len);
 212   for (int i = 0; i < vec->len; ++i)
 213     data[i] = std::string (vec->tokens[i].ptr, vec->tokens[i].length);
 214
 215   push_new<expr::c_string_operation> ((enum c_string_type_values) kind,
 216                                       std::move (data));
 217 }
 218
 219 /* See parser-defs.h.  */
 220
 221 void
 222 parser_state::push_symbol (const char *name, block_symbol sym)
 223 {
 224   if (sym.symbol != nullptr)
 225     {
 226       if (symbol_read_needs_frame (sym.symbol))
 227         block_tracker->update (sym);
 228       push_new<expr::var_value_operation> (sym);
 229     }
 230   else
 231     {
 232       struct bound_minimal_symbol msymbol = lookup_bound_minimal_symbol (name);
 233       if (msymbol.minsym != NULL)
 234         push_new<expr::var_msym_value_operation> (msymbol);
 235       else if (!have_full_symbols () && !have_partial_symbols ())
 236         error (_("No symbol table is loaded.  Use the \"file\" command."));
 237       else
 238         error (_("No symbol \"%s\" in current context."), name);
 239     }
 240 }
 241
 242 /* See parser-defs.h.  */
 243
 244 void
 245 parser_state::push_dollar (struct stoken str)
 246 {
 247   struct block_symbol sym;
 248   struct bound_minimal_symbol msym;
 249   struct internalvar *isym = NULL;
 250   std::string copy;
 251
 252   /* Handle the tokens $digits; also $ (short for $0) and $$ (short for $$1)
 253      and $$digits (equivalent to $<-digits> if you could type that).  */
 254
 255   int negate = 0;
 256   int i = 1;
 257   /* Double dollar means negate the number and add -1 as well.
 258      Thus $$ alone means -1.  */
 259   if (str.length >= 2 && str.ptr[1] == '$')
 260     {
 261       negate = 1;
 262       i = 2;
 263     }
 264   if (i == str.length)
 265     {
 266       /* Just dollars (one or two).  */
 267       i = -negate;
 268       goto handle_last;
 269     }
 270   /* Is the rest of the token digits?  */
 271   for (; i < str.length; i++)
 272     if (!(str.ptr[i] >= '0' && str.ptr[i] <= '9'))
 273       break;
 274   if (i == str.length)
 275     {
 276       i = atoi (str.ptr + 1 + negate);
 277       if (negate)
 278         i = -i;
 279       goto handle_last;
 280     }
 281
 282   /* Handle tokens that refer to machine registers:
 283      $ followed by a register name.  */
 284   i = user_reg_map_name_to_regnum (gdbarch (),
 285                                    str.ptr + 1, str.length - 1);
 286   if (i >= 0)
 287     goto handle_register;
 288
 289   /* Any names starting with $ are probably debugger internal variables.  */
 290
 291   copy = copy_name (str);
 292   isym = lookup_only_internalvar (copy.c_str () + 1);
 293   if (isym)
 294     {
 295       push_new<expr::internalvar_operation> (isym);
 296       return;
 297     }
 298
 299   /* On some systems, such as HP-UX and hppa-linux, certain system routines
 300      have names beginning with $ or $$.  Check for those, first.  */
 301
 302   sym = lookup_symbol (copy.c_str (), NULL, VAR_DOMAIN, NULL);
 303   if (sym.symbol)
 304     {
 305       push_new<expr::var_value_operation> (sym);
 306       return;
 307     }
 308   msym = lookup_bound_minimal_symbol (copy.c_str ());
 309   if (msym.minsym)
 310     {
 311       push_new<expr::var_msym_value_operation> (msym);
 312       return;
 313     }
 314
 315   /* Any other names are assumed to be debugger internal variables.  */
 316
 317   push_new<expr::internalvar_operation>
 318     (create_internalvar (copy.c_str () + 1));
 319   return;
 320 handle_last:
 321   push_new<expr::last_operation> (i);
 322   return;
 323 handle_register:
 324   str.length--;
 325   str.ptr++;
 326   push_new<expr::register_operation> (copy_name (str));
 327   block_tracker->update (expression_context_block,
 328                          INNERMOST_BLOCK_FOR_REGISTERS);
 329   return;
 330 }
 331
 332 \f
 333
 334 const char *
 335 find_template_name_end (const char *p)
 336 {
 337   int depth = 1;
 338   int just_seen_right = 0;
 339   int just_seen_colon = 0;
 340   int just_seen_space = 0;
 341
 342   if (!p || (*p != '<'))
 343     return 0;
 344
 345   while (*++p)
 346     {
 347       switch (*p)
 348         {
 349         case '\'':
 350         case '\"':
 351         case '{':
 352         case '}':
 353           /* In future, may want to allow these??  */
 354           return 0;
 355         case '<':
 356           depth++;              /* start nested template */
 357           if (just_seen_colon || just_seen_right || just_seen_space)
 358             return 0;           /* but not after : or :: or > or space */
 359           break;
 360         case '>':
 361           if (just_seen_colon || just_seen_right)
 362             return 0;           /* end a (nested?) template */
 363           just_seen_right = 1;  /* but not after : or :: */
 364           if (--depth == 0)     /* also disallow >>, insist on > > */
 365             return ++p;         /* if outermost ended, return */
 366           break;
 367         case ':':
 368           if (just_seen_space || (just_seen_colon > 1))
 369             return 0;           /* nested class spec coming up */
 370           just_seen_colon++;    /* we allow :: but not :::: */
 371           break;
 372         case ' ':
 373           break;
 374         default:
 375           if (!((*p >= 'a' && *p <= 'z') ||     /* allow token chars */
 376                 (*p >= 'A' && *p <= 'Z') ||
 377                 (*p >= '0' && *p <= '9') ||
 378                 (*p == '_') || (*p == ',') ||   /* commas for template args */
 379                 (*p == '&') || (*p == '*') ||   /* pointer and ref types */
 380                 (*p == '(') || (*p == ')') ||   /* function types */
 381                 (*p == '[') || (*p == ']')))    /* array types */
 382             return 0;
 383         }
 384       if (*p != ' ')
 385         just_seen_space = 0;
 386       if (*p != ':')
 387         just_seen_colon = 0;
 388       if (*p != '>')
 389         just_seen_right = 0;
 390     }
 391   return 0;
 392 }
 393 \f
 394
 395 /* Return a null-terminated temporary copy of the name of a string token.
 396
 397    Tokens that refer to names do so with explicit pointer and length,
 398    so they can share the storage that lexptr is parsing.
 399    When it is necessary to pass a name to a function that expects
 400    a null-terminated string, the substring is copied out
 401    into a separate block of storage.  */
 402
 403 std::string
 404 copy_name (struct stoken token)
 405 {
 406   return std::string (token.ptr, token.length);
 407 }
 408 \f
 409
 410 /* Read an expression from the string *STRINGPTR points to,
 411    parse it, and return a pointer to a struct expression that we malloc.
 412    Use block BLOCK as the lexical context for variable names;
 413    if BLOCK is zero, use the block of the selected stack frame.
 414    Meanwhile, advance *STRINGPTR to point after the expression,
 415    at the first nonwhite character that is not part of the expression
 416    (possibly a null character).
 417
 418    If COMMA is nonzero, stop if a comma is reached.  */
 419
 420 expression_up
 421 parse_exp_1 (const char **stringptr, CORE_ADDR pc, const struct block *block,
 422              int comma, innermost_block_tracker *tracker)
 423 {
 424   return parse_exp_in_context (stringptr, pc, block, comma, false,
 425                                tracker, nullptr);
 426 }
 427
 428 /* As for parse_exp_1, except that if VOID_CONTEXT_P, then
 429    no value is expected from the expression.  */
 430
 431 static expression_up
 432 parse_exp_in_context (const char **stringptr, CORE_ADDR pc,
 433                       const struct block *block,
 434                       int comma, bool void_context_p,
 435                       innermost_block_tracker *tracker,
 436                       expr_completion_state *cstate)
 437 {
 438   const struct language_defn *lang = NULL;
 439
 440   if (*stringptr == 0 || **stringptr == 0)
 441     error_no_arg (_("expression to compute"));
 442
 443   const struct block *expression_context_block = block;
 444   CORE_ADDR expression_context_pc = 0;
 445
 446   innermost_block_tracker local_tracker;
 447   if (tracker == nullptr)
 448     tracker = &local_tracker;
 449
 450   /* If no context specified, try using the current frame, if any.  */
 451   if (!expression_context_block)
 452     expression_context_block = get_selected_block (&expression_context_pc);
 453   else if (pc == 0)
 454     expression_context_pc = BLOCK_ENTRY_PC (expression_context_block);
 455   else
 456     expression_context_pc = pc;
 457
 458   /* Fall back to using the current source static context, if any.  */
 459
 460   if (!expression_context_block)
 461     {
 462       struct symtab_and_line cursal = get_current_source_symtab_and_line ();
 463       if (cursal.symtab)
 464         expression_context_block
 465           = BLOCKVECTOR_BLOCK (cursal.symtab->blockvector (),
 466                                STATIC_BLOCK);
 467       if (expression_context_block)
 468         expression_context_pc = BLOCK_ENTRY_PC (expression_context_block);
 469     }
 470
 471   if (language_mode == language_mode_auto && block != NULL)
 472     {
 473       /* Find the language associated to the given context block.
 474          Default to the current language if it can not be determined.
 475
 476          Note that using the language corresponding to the current frame
 477          can sometimes give unexpected results.  For instance, this
 478          routine is often called several times during the inferior
 479          startup phase to re-parse breakpoint expressions after
 480          a new shared library has been loaded.  The language associated
 481          to the current frame at this moment is not relevant for
 482          the breakpoint.  Using it would therefore be silly, so it seems
 483          better to rely on the current language rather than relying on
 484          the current frame language to parse the expression.  That's why
 485          we do the following language detection only if the context block
 486          has been specifically provided.  */
 487       struct symbol *func = block_linkage_function (block);
 488
 489       if (func != NULL)
 490         lang = language_def (func->language ());
 491       if (lang == NULL || lang->la_language == language_unknown)
 492         lang = current_language;
 493     }
 494   else
 495     lang = current_language;
 496
 497   /* get_current_arch may reset CURRENT_LANGUAGE via select_frame.
 498      While we need CURRENT_LANGUAGE to be set to LANG (for lookup_symbol
 499      and others called from *.y) ensure CURRENT_LANGUAGE gets restored
 500      to the value matching SELECTED_FRAME as set by get_current_arch.  */
 501
 502   parser_state ps (lang, get_current_arch (), expression_context_block,
 503                    expression_context_pc, comma, *stringptr,
 504                    cstate != nullptr, tracker, void_context_p);
 505
 506   scoped_restore_current_language lang_saver;
 507   set_language (lang->la_language);
 508
 509   try
 510     {
 511       lang->parser (&ps);
 512     }
 513   catch (const gdb_exception &except)
 514     {
 515       /* If parsing for completion, allow this to succeed; but if no
 516          expression elements have been written, then there's nothing
 517          to do, so fail.  */
 518       if (! ps.parse_completion || ps.expout->op == nullptr)
 519         throw;
 520     }
 521
 522   expression_up result = ps.release ();
 523   result->op->set_outermost ();
 524
 525   if (expressiondebug)
 526     dump_prefix_expression (result.get (), gdb_stdlog);
 527
 528   if (cstate != nullptr)
 529     *cstate = std::move (ps.m_completion_state);
 530   *stringptr = ps.lexptr;
 531   return result;
 532 }
 533
 534 /* Parse STRING as an expression, and complain if this fails to use up
 535    all of the contents of STRING.  TRACKER, if non-null, will be
 536    updated by the parser.  VOID_CONTEXT_P should be true to indicate
 537    that the expression may be expected to return a value with void
 538    type.  Parsers are free to ignore this, or to use it to help with
 539    overload resolution decisions.  */
 540
 541 expression_up
 542 parse_expression (const char *string, innermost_block_tracker *tracker,
 543                   bool void_context_p)
 544 {
 545   expression_up exp = parse_exp_in_context (&string, 0, nullptr, 0,
 546                                             void_context_p,
 547                                             tracker, nullptr);
 548   if (*string)
 549     error (_("Junk after end of expression."));
 550   return exp;
 551 }
 552
 553 /* Same as parse_expression, but using the given language (LANG)
 554    to parse the expression.  */
 555
 556 expression_up
 557 parse_expression_with_language (const char *string, enum language lang)
 558 {
 559   gdb::optional<scoped_restore_current_language> lang_saver;
 560   if (current_language->la_language != lang)
 561     {
 562       lang_saver.emplace ();
 563       set_language (lang);
 564     }
 565
 566   return parse_expression (string);
 567 }
 568
 569 /* Parse STRING as an expression.  If parsing ends in the middle of a
 570    field reference, return the type of the left-hand-side of the
 571    reference; furthermore, if the parsing ends in the field name,
 572    return the field name in *NAME.  If the parsing ends in the middle
 573    of a field reference, but the reference is somehow invalid, throw
 574    an exception.  In all other cases, return NULL.  */
 575
 576 struct type *
 577 parse_expression_for_completion (const char *string,
 578                                  gdb::unique_xmalloc_ptr<char> *name,
 579                                  enum type_code *code)
 580 {
 581   expression_up exp;
 582   expr_completion_state cstate;
 583
 584   try
 585     {
 586       exp = parse_exp_in_context (&string, 0, 0, 0, false, nullptr, &cstate);
 587     }
 588   catch (const gdb_exception_error &except)
 589     {
 590       /* Nothing, EXP remains NULL.  */
 591     }
 592
 593   if (exp == NULL)
 594     return NULL;
 595
 596   if (cstate.expout_tag_completion_type != TYPE_CODE_UNDEF)
 597     {
 598       *code = cstate.expout_tag_completion_type;
 599       *name = std::move (cstate.expout_completion_name);
 600       return NULL;
 601     }
 602
 603   if (cstate.expout_last_op == nullptr)
 604     return nullptr;
 605
 606   expr::structop_base_operation *op = cstate.expout_last_op;
 607   const std::string &fld = op->get_string ();
 608   *name = make_unique_xstrdup (fld.c_str ());
 609   return value_type (op->evaluate_lhs (exp.get ()));
 610 }
 611
 612 /* Parse floating point value P of length LEN.
 613    Return false if invalid, true if valid.
 614    The successfully parsed number is stored in DATA in
 615    target format for floating-point type TYPE.
 616
 617    NOTE: This accepts the floating point syntax that sscanf accepts.  */
 618
 619 bool
 620 parse_float (const char *p, int len,
 621              const struct type *type, gdb_byte *data)
 622 {
 623   return target_float_from_string (data, type, std::string (p, len));
 624 }
 625 \f
 626 /* This function avoids direct calls to fprintf
 627    in the parser generated debug code.  */
 628 void
 629 parser_fprintf (FILE *x, const char *y, ...)
 630 {
 631   va_list args;
 632
 633   va_start (args, y);
 634   if (x == stderr)
 635     gdb_vprintf (gdb_stderr, y, args);
 636   else
 637     {
 638       gdb_printf (gdb_stderr, " Unknown FILE used.\n");
 639       gdb_vprintf (gdb_stderr, y, args);
 640     }
 641   va_end (args);
 642 }
 643
 644 /* Return rue if EXP uses OBJFILE (and will become dangling when
 645    OBJFILE is unloaded), otherwise return false.  OBJFILE must not be
 646    a separate debug info file.  */
 647
 648 bool
 649 exp_uses_objfile (struct expression *exp, struct objfile *objfile)
 650 {
 651   gdb_assert (objfile->separate_debug_objfile_backlink == NULL);
 652
 653   return exp->op->uses_objfile (objfile);
 654 }
 655
 656 void _initialize_parse ();
 657 void
 658 _initialize_parse ()
 659 {
 660   add_setshow_zuinteger_cmd ("expression", class_maintenance,
 661                              &expressiondebug,
 662                              _("Set expression debugging."),
 663                              _("Show expression debugging."),
 664                              _("When non-zero, the internal representation "
 665                                "of expressions will be printed."),
 666                              NULL,
 667                              show_expressiondebug,
 668                              &setdebuglist, &showdebuglist);
 669   add_setshow_boolean_cmd ("parser", class_maintenance,
 670                             &parser_debug,
 671                            _("Set parser debugging."),
 672                            _("Show parser debugging."),
 673                            _("When non-zero, expression parser "
 674                              "tracing will be enabled."),
 675                             NULL,
 676                             show_parserdebug,
 677                             &setdebuglist, &showdebuglist);
 678 }