gcc/analyzer/sm-file.cc

   1 /* A state machine for detecting misuses of <stdio.h>'s FILE * API.
   2    Copyright (C) 2019-2023 Free Software Foundation, Inc.
   3    Contributed by David Malcolm <dmalcolm@redhat.com>.
   4
   5 This file is part of GCC.
   6
   7 GCC is free software; you can redistribute it and/or modify it
   8 under the terms of the GNU General Public License as published by
   9 the Free Software Foundation; either version 3, or (at your option)
  10 any later version.
  11
  12 GCC is distributed in the hope that it will be useful, but
  13 WITHOUT ANY WARRANTY; without even the implied warranty of
  14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  15 General Public License for more details.
  16
  17 You should have received a copy of the GNU General Public License
  18 along with GCC; see the file COPYING3.  If not see
  19 <http://www.gnu.org/licenses/>.  */
  20
  21 #include "config.h"
  22 #define INCLUDE_MEMORY
  23 #include "system.h"
  24 #include "coretypes.h"
  25 #include "make-unique.h"
  26 #include "tree.h"
  27 #include "function.h"
  28 #include "basic-block.h"
  29 #include "gimple.h"
  30 #include "options.h"
  31 #include "diagnostic-path.h"
  32 #include "diagnostic-metadata.h"
  33 #include "analyzer/analyzer.h"
  34 #include "diagnostic-event-id.h"
  35 #include "analyzer/analyzer-logging.h"
  36 #include "analyzer/sm.h"
  37 #include "analyzer/pending-diagnostic.h"
  38 #include "analyzer/function-set.h"
  39 #include "analyzer/analyzer-selftests.h"
  40 #include "selftest.h"
  41 #include "analyzer/call-string.h"
  42 #include "analyzer/program-point.h"
  43 #include "analyzer/store.h"
  44 #include "analyzer/region-model.h"
  45 #include "analyzer/call-details.h"
  46
  47 #if ENABLE_ANALYZER
  48
  49 namespace ana {
  50
  51 namespace {
  52
  53 /* A state machine for detecting misuses of <stdio.h>'s FILE * API.  */
  54
  55 class fileptr_state_machine : public state_machine
  56 {
  57 public:
  58   fileptr_state_machine (logger *logger);
  59
  60   bool inherited_state_p () const final override { return false; }
  61
  62   state_machine::state_t
  63   get_default_state (const svalue *sval) const final override
  64   {
  65     if (tree cst = sval->maybe_get_constant ())
  66       {
  67         if (zerop (cst))
  68           return m_null;
  69       }
  70     return m_start;
  71   }
  72
  73   bool on_stmt (sm_context *sm_ctxt,
  74                 const supernode *node,
  75                 const gimple *stmt) const final override;
  76
  77   void on_condition (sm_context *sm_ctxt,
  78                      const supernode *node,
  79                      const gimple *stmt,
  80                      const svalue *lhs,
  81                      enum tree_code op,
  82                      const svalue *rhs) const final override;
  83
  84   bool can_purge_p (state_t s) const final override;
  85   std::unique_ptr<pending_diagnostic> on_leak (tree var) const final override;
  86
  87   /* State for a FILE * returned from fopen that hasn't been checked for
  88      NULL.
  89      It could be an open stream, or could be NULL.  */
  90   state_t m_unchecked;
  91
  92   /* State for a FILE * that's known to be NULL.  */
  93   state_t m_null;
  94
  95   /* State for a FILE * that's known to be a non-NULL open stream.  */
  96   state_t m_nonnull;
  97
  98   /* State for a FILE * that's had fclose called on it.  */
  99   state_t m_closed;
 100
 101   /* Stop state, for a FILE * we don't want to track any more.  */
 102   state_t m_stop;
 103 };
 104
 105 /* Base class for diagnostics relative to fileptr_state_machine.  */
 106
 107 class file_diagnostic : public pending_diagnostic
 108 {
 109 public:
 110   file_diagnostic (const fileptr_state_machine &sm, tree arg)
 111   : m_sm (sm), m_arg (arg)
 112   {}
 113
 114   bool subclass_equal_p (const pending_diagnostic &base_other) const override
 115   {
 116     return same_tree_p (m_arg, ((const file_diagnostic &)base_other).m_arg);
 117   }
 118
 119   label_text describe_state_change (const evdesc::state_change &change)
 120     override
 121   {
 122     if (change.m_old_state == m_sm.get_start_state ()
 123         && change.m_new_state == m_sm.m_unchecked)
 124       // TODO: verify that it's the fopen stmt, not a copy
 125       return label_text::borrow ("opened here");
 126     if (change.m_old_state == m_sm.m_unchecked
 127         && change.m_new_state == m_sm.m_nonnull)
 128       {
 129         if (change.m_expr)
 130           return change.formatted_print ("assuming %qE is non-NULL",
 131                                          change.m_expr);
 132         else
 133           return change.formatted_print ("assuming FILE * is non-NULL");
 134       }
 135     if (change.m_new_state == m_sm.m_null)
 136       {
 137         if (change.m_expr)
 138           return change.formatted_print ("assuming %qE is NULL",
 139                                          change.m_expr);
 140         else
 141           return change.formatted_print ("assuming FILE * is NULL");
 142       }
 143     return label_text ();
 144   }
 145
 146   diagnostic_event::meaning
 147   get_meaning_for_state_change (const evdesc::state_change &change)
 148     const final override
 149   {
 150     if (change.m_old_state == m_sm.get_start_state ()
 151         && change.m_new_state == m_sm.m_unchecked)
 152       return diagnostic_event::meaning (diagnostic_event::VERB_acquire,
 153                                         diagnostic_event::NOUN_resource);
 154     if (change.m_new_state == m_sm.m_closed)
 155       return diagnostic_event::meaning (diagnostic_event::VERB_release,
 156                                         diagnostic_event::NOUN_resource);
 157     return diagnostic_event::meaning ();
 158   }
 159
 160 protected:
 161   const fileptr_state_machine &m_sm;
 162   tree m_arg;
 163 };
 164
 165 class double_fclose : public file_diagnostic
 166 {
 167 public:
 168   double_fclose (const fileptr_state_machine &sm, tree arg)
 169     : file_diagnostic (sm, arg)
 170   {}
 171
 172   const char *get_kind () const final override { return "double_fclose"; }
 173
 174   int get_controlling_option () const final override
 175   {
 176     return OPT_Wanalyzer_double_fclose;
 177   }
 178
 179   bool emit (rich_location *rich_loc, logger *) final override
 180   {
 181     diagnostic_metadata m;
 182     /* CWE-1341: Multiple Releases of Same Resource or Handle.  */
 183     m.add_cwe (1341);
 184     return warning_meta (rich_loc, m, get_controlling_option (),
 185                          "double %<fclose%> of FILE %qE",
 186                          m_arg);
 187   }
 188
 189   label_text describe_state_change (const evdesc::state_change &change)
 190     override
 191   {
 192     if (change.m_new_state == m_sm.m_closed)
 193       {
 194         m_first_fclose_event = change.m_event_id;
 195         return change.formatted_print ("first %qs here", "fclose");
 196       }
 197     return file_diagnostic::describe_state_change (change);
 198   }
 199
 200   label_text describe_final_event (const evdesc::final_event &ev) final override
 201   {
 202     if (m_first_fclose_event.known_p ())
 203       return ev.formatted_print ("second %qs here; first %qs was at %@",
 204                                  "fclose", "fclose",
 205                                  &m_first_fclose_event);
 206     return ev.formatted_print ("second %qs here", "fclose");
 207   }
 208
 209 private:
 210   diagnostic_event_id_t m_first_fclose_event;
 211 };
 212
 213 class file_leak : public file_diagnostic
 214 {
 215 public:
 216   file_leak (const fileptr_state_machine &sm, tree arg)
 217     : file_diagnostic (sm, arg)
 218   {}
 219
 220   const char *get_kind () const final override { return "file_leak"; }
 221
 222   int get_controlling_option () const final override
 223   {
 224     return OPT_Wanalyzer_file_leak;
 225   }
 226
 227   bool emit (rich_location *rich_loc, logger *) final override
 228   {
 229     diagnostic_metadata m;
 230     /* CWE-775: "Missing Release of File Descriptor or Handle after
 231        Effective Lifetime". */
 232     m.add_cwe (775);
 233     if (m_arg)
 234       return warning_meta (rich_loc, m, get_controlling_option (),
 235                            "leak of FILE %qE",
 236                            m_arg);
 237     else
 238       return warning_meta (rich_loc, m, get_controlling_option (),
 239                            "leak of FILE");
 240   }
 241
 242   label_text describe_state_change (const evdesc::state_change &change)
 243     final override
 244   {
 245     if (change.m_new_state == m_sm.m_unchecked)
 246       {
 247         m_fopen_event = change.m_event_id;
 248         return label_text::borrow ("opened here");
 249       }
 250     return file_diagnostic::describe_state_change (change);
 251   }
 252
 253   label_text describe_final_event (const evdesc::final_event &ev) final override
 254   {
 255     if (m_fopen_event.known_p ())
 256       {
 257         if (ev.m_expr)
 258           return ev.formatted_print ("%qE leaks here; was opened at %@",
 259                                      ev.m_expr, &m_fopen_event);
 260         else
 261           return ev.formatted_print ("leaks here; was opened at %@",
 262                                      &m_fopen_event);
 263       }
 264     else
 265       {
 266         if (ev.m_expr)
 267           return ev.formatted_print ("%qE leaks here", ev.m_expr);
 268         else
 269           return ev.formatted_print ("leaks here");
 270       }
 271   }
 272
 273 private:
 274   diagnostic_event_id_t m_fopen_event;
 275 };
 276
 277 /* fileptr_state_machine's ctor.  */
 278
 279 fileptr_state_machine::fileptr_state_machine (logger *logger)
 280 : state_machine ("file", logger)
 281 {
 282   m_unchecked = add_state ("unchecked");
 283   m_null = add_state ("null");
 284   m_nonnull = add_state ("nonnull");
 285   m_closed = add_state ("closed");
 286   m_stop = add_state ("stop");
 287 }
 288
 289 /* Get a set of functions that are known to take a FILE * that must be open,
 290    and are known to not close it.  */
 291
 292 static function_set
 293 get_file_using_fns ()
 294 {
 295   // TODO: populate this list more fully
 296   static const char * const funcnames[] = {
 297     /* This array must be kept sorted.  */
 298     "__fbufsize",
 299     "__flbf",
 300     "__fpending",
 301     "__fpurge",
 302     "__freadable",
 303     "__freading",
 304     "__fsetlocking",
 305     "__fwritable",
 306     "__fwriting",
 307     "clearerr",
 308     "clearerr_unlocked",
 309     "feof",
 310     "feof_unlocked",
 311     "ferror",
 312     "ferror_unlocked",
 313     "fflush", // safe to call with NULL
 314     "fflush_unlocked",  // safe to call with NULL
 315     "fgetc",
 316     "fgetc_unlocked",
 317     "fgetpos",
 318     "fgets",
 319     "fgets_unlocked",
 320     "fgetwc_unlocked",
 321     "fgetws_unlocked",
 322     "fileno",
 323     "fileno_unlocked",
 324     "fprintf",
 325     "fputc",
 326     "fputc_unlocked",
 327     "fputs",
 328     "fputs_unlocked",
 329     "fputwc_unlocked",
 330     "fputws_unlocked",
 331     "fread_unlocked",
 332     "fseek",
 333     "fsetpos",
 334     "ftell",
 335     "fwrite_unlocked",
 336     "getc",
 337     "getc_unlocked",
 338     "getwc_unlocked",
 339     "putc",
 340     "putc_unlocked",
 341     "rewind",
 342     "setbuf",
 343     "setbuffer",
 344     "setlinebuf",
 345     "setvbuf",
 346     "ungetc",
 347     "vfprintf"
 348   };
 349   const size_t count = ARRAY_SIZE (funcnames);
 350   function_set fs (funcnames, count);
 351   return fs;
 352 }
 353
 354 /* Return true if FNDECL is known to require an open FILE *, and is known
 355    to not close it.  */
 356
 357 static bool
 358 is_file_using_fn_p (tree fndecl)
 359 {
 360   function_set fs = get_file_using_fns ();
 361   if (fs.contains_decl_p (fndecl))
 362     return true;
 363
 364   /* Also support variants of these names prefixed with "_IO_".  */
 365   const char *name = IDENTIFIER_POINTER (DECL_NAME (fndecl));
 366   if (startswith (name, "_IO_") && fs.contains_name_p (name + 4))
 367     return true;
 368
 369   return false;
 370 }
 371
 372 /* Implementation of state_machine::on_stmt vfunc for fileptr_state_machine.  */
 373
 374 bool
 375 fileptr_state_machine::on_stmt (sm_context *sm_ctxt,
 376                                 const supernode *node,
 377                                 const gimple *stmt) const
 378 {
 379   if (const gcall *call = dyn_cast <const gcall *> (stmt))
 380     if (tree callee_fndecl = sm_ctxt->get_fndecl_for_call (call))
 381       {
 382         if (is_named_call_p (callee_fndecl, "fopen", call, 2))
 383           {
 384             tree lhs = gimple_call_lhs (call);
 385             if (lhs)
 386               sm_ctxt->on_transition (node, stmt, lhs, m_start, m_unchecked);
 387             else
 388               {
 389                 /* TODO: report leak.  */
 390               }
 391             return true;
 392           }
 393
 394         if (is_named_call_p (callee_fndecl, "fclose", call, 1))
 395           {
 396             tree arg = gimple_call_arg (call, 0);
 397
 398             sm_ctxt->on_transition (node, stmt, arg, m_start, m_closed);
 399
 400             // TODO: is it safe to call fclose (NULL) ?
 401             sm_ctxt->on_transition (node, stmt, arg, m_unchecked, m_closed);
 402             sm_ctxt->on_transition (node, stmt, arg, m_null, m_closed);
 403
 404             sm_ctxt->on_transition (node, stmt , arg, m_nonnull, m_closed);
 405
 406             if (sm_ctxt->get_state (stmt, arg) == m_closed)
 407               {
 408                 tree diag_arg = sm_ctxt->get_diagnostic_tree (arg);
 409                 sm_ctxt->warn (node, stmt, arg,
 410                                make_unique<double_fclose> (*this, diag_arg));
 411                 sm_ctxt->set_next_state (stmt, arg, m_stop);
 412               }
 413             return true;
 414           }
 415
 416         if (is_file_using_fn_p (callee_fndecl))
 417           {
 418             // TODO: operations on unchecked file
 419             return true;
 420           }
 421         // etc
 422       }
 423
 424   return false;
 425 }
 426
 427 /* Implementation of state_machine::on_condition vfunc for
 428    fileptr_state_machine.
 429    Potentially transition state 'unchecked' to 'nonnull' or to 'null'.  */
 430
 431 void
 432 fileptr_state_machine::on_condition (sm_context *sm_ctxt,
 433                                      const supernode *node,
 434                                      const gimple *stmt,
 435                                      const svalue *lhs,
 436                                      enum tree_code op,
 437                                      const svalue *rhs) const
 438 {
 439   if (!rhs->all_zeroes_p ())
 440     return;
 441
 442   // TODO: has to be a FILE *, specifically
 443   if (!any_pointer_p (lhs))
 444     return;
 445   // TODO: has to be a FILE *, specifically
 446   if (!any_pointer_p (rhs))
 447     return;
 448
 449   if (op == NE_EXPR)
 450     {
 451       log ("got 'ARG != 0' match");
 452       sm_ctxt->on_transition (node, stmt,
 453                               lhs, m_unchecked, m_nonnull);
 454     }
 455   else if (op == EQ_EXPR)
 456     {
 457       log ("got 'ARG == 0' match");
 458       sm_ctxt->on_transition (node, stmt,
 459                               lhs, m_unchecked, m_null);
 460     }
 461 }
 462
 463 /* Implementation of state_machine::can_purge_p vfunc for fileptr_state_machine.
 464    Don't allow purging of pointers in state 'unchecked' or 'nonnull'
 465    (to avoid false leak reports).  */
 466
 467 bool
 468 fileptr_state_machine::can_purge_p (state_t s) const
 469 {
 470   return s != m_unchecked && s != m_nonnull;
 471 }
 472
 473 /* Implementation of state_machine::on_leak vfunc for
 474    fileptr_state_machine, for complaining about leaks of FILE * in
 475    state 'unchecked' and 'nonnull'.  */
 476
 477 std::unique_ptr<pending_diagnostic>
 478 fileptr_state_machine::on_leak (tree var) const
 479 {
 480   return make_unique<file_leak> (*this, var);
 481 }
 482
 483 } // anonymous namespace
 484
 485 /* Internal interface to this file. */
 486
 487 state_machine *
 488 make_fileptr_state_machine (logger *logger)
 489 {
 490   return new fileptr_state_machine (logger);
 491 }
 492
 493 /* Handler for various stdio-related builtins that merely have external
 494    effects that are out of scope for the analyzer: we only want to model
 495    the effects on the return value.  */
 496
 497 class kf_stdio_output_fn : public pure_known_function_with_default_return
 498 {
 499 public:
 500   bool matches_call_types_p (const call_details &) const final override
 501   {
 502     return true;
 503   }
 504
 505   /* A no-op; we just want the conjured return value.  */
 506 };
 507
 508 /* Handler for "ferror"".  */
 509
 510 class kf_ferror : public pure_known_function_with_default_return
 511 {
 512 public:
 513   bool matches_call_types_p (const call_details &cd) const final override
 514   {
 515     return (cd.num_args () == 1
 516             && cd.arg_is_pointer_p (0));
 517   }
 518
 519   /* No side effects.  */
 520 };
 521
 522 /* Handler for "fileno"".  */
 523
 524 class kf_fileno : public pure_known_function_with_default_return
 525 {
 526 public:
 527   bool matches_call_types_p (const call_details &cd) const final override
 528   {
 529     return (cd.num_args () == 1
 530             && cd.arg_is_pointer_p (0));
 531   }
 532
 533   /* No side effects.  */
 534 };
 535
 536 /* Handler for "fgets" and "fgets_unlocked".  */
 537
 538 class kf_fgets : public known_function
 539 {
 540 public:
 541   bool matches_call_types_p (const call_details &cd) const final override
 542   {
 543     return (cd.num_args () == 3
 544             && cd.arg_is_pointer_p (0)
 545             && cd.arg_is_pointer_p (2));
 546   }
 547
 548   void impl_call_pre (const call_details &cd) const final override
 549   {
 550     /* Ideally we would bifurcate state here between the
 551        error vs no error cases.  */
 552     region_model *model = cd.get_model ();
 553     const svalue *ptr_sval = cd.get_arg_svalue (0);
 554     if (const region *reg = ptr_sval->maybe_get_region ())
 555       {
 556         const region *base_reg = reg->get_base_region ();
 557         const svalue *new_sval = cd.get_or_create_conjured_svalue (base_reg);
 558         model->set_value (base_reg, new_sval, cd.get_ctxt ());
 559       }
 560     cd.set_any_lhs_with_defaults ();
 561   }
 562 };
 563
 564 /* Handler for "fread".
 565      size_t fread(void *restrict buffer, size_t size, size_t count,
 566                   FILE *restrict stream);
 567    See e.g. https://en.cppreference.com/w/c/io/fread
 568    and https://www.man7.org/linux/man-pages/man3/fread.3.html */
 569
 570 class kf_fread : public known_function
 571 {
 572 public:
 573   bool matches_call_types_p (const call_details &cd) const final override
 574   {
 575     return (cd.num_args () == 4
 576             && cd.arg_is_pointer_p (0)
 577             && cd.arg_is_size_p (1)
 578             && cd.arg_is_size_p (2)
 579             && cd.arg_is_pointer_p (3));
 580   }
 581
 582   /* For now, assume that any call to "fread" fully clobbers the buffer
 583      passed in.  This isn't quite correct (e.g. errors, partial reads;
 584      see PR analyzer/108689), but at least stops us falsely complaining
 585      about the buffer being uninitialized.  */
 586   void impl_call_pre (const call_details &cd) const final override
 587   {
 588     region_model *model = cd.get_model ();
 589     const svalue *ptr_sval = cd.get_arg_svalue (0);
 590     if (const region *reg = ptr_sval->maybe_get_region ())
 591       {
 592         const region *base_reg = reg->get_base_region ();
 593         const svalue *new_sval = cd.get_or_create_conjured_svalue (base_reg);
 594         model->set_value (base_reg, new_sval, cd.get_ctxt ());
 595       }
 596     cd.set_any_lhs_with_defaults ();
 597   }
 598 };
 599
 600 /* Handler for "getc"".  */
 601
 602 class kf_getc : public pure_known_function_with_default_return
 603 {
 604 public:
 605   bool matches_call_types_p (const call_details &cd) const final override
 606   {
 607     return (cd.num_args () == 1
 608             && cd.arg_is_pointer_p (0));
 609   }
 610 };
 611
 612 /* Handler for "getchar"".  */
 613
 614 class kf_getchar : public pure_known_function_with_default_return
 615 {
 616 public:
 617   bool matches_call_types_p (const call_details &cd) const final override
 618   {
 619     return cd.num_args () == 0;
 620   }
 621
 622   /* Empty.  No side-effects (tracking stream state is out-of-scope
 623      for the analyzer).  */
 624 };
 625
 626 /* Populate KFM with instances of known functions relating to
 627    stdio streams.  */
 628
 629 void
 630 register_known_file_functions (known_function_manager &kfm)
 631 {
 632   kfm.add (BUILT_IN_FPRINTF, make_unique<kf_stdio_output_fn> ());
 633   kfm.add (BUILT_IN_FPRINTF_UNLOCKED, make_unique<kf_stdio_output_fn> ());
 634   kfm.add (BUILT_IN_FPUTC, make_unique<kf_stdio_output_fn> ());
 635   kfm.add (BUILT_IN_FPUTC_UNLOCKED, make_unique<kf_stdio_output_fn> ());
 636   kfm.add (BUILT_IN_FPUTS, make_unique<kf_stdio_output_fn> ());
 637   kfm.add (BUILT_IN_FPUTS_UNLOCKED, make_unique<kf_stdio_output_fn> ());
 638   kfm.add (BUILT_IN_FWRITE, make_unique<kf_stdio_output_fn> ());
 639   kfm.add (BUILT_IN_FWRITE_UNLOCKED, make_unique<kf_stdio_output_fn> ());
 640   kfm.add (BUILT_IN_PRINTF, make_unique<kf_stdio_output_fn> ());
 641   kfm.add (BUILT_IN_PRINTF_UNLOCKED, make_unique<kf_stdio_output_fn> ());
 642   kfm.add (BUILT_IN_PUTC, make_unique<kf_stdio_output_fn> ());
 643   kfm.add (BUILT_IN_PUTCHAR, make_unique<kf_stdio_output_fn> ());
 644   kfm.add (BUILT_IN_PUTCHAR_UNLOCKED, make_unique<kf_stdio_output_fn> ());
 645   kfm.add (BUILT_IN_PUTC_UNLOCKED, make_unique<kf_stdio_output_fn> ());
 646   kfm.add (BUILT_IN_PUTS, make_unique<kf_stdio_output_fn> ());
 647   kfm.add (BUILT_IN_PUTS_UNLOCKED, make_unique<kf_stdio_output_fn> ());
 648   kfm.add (BUILT_IN_VFPRINTF, make_unique<kf_stdio_output_fn> ());
 649   kfm.add (BUILT_IN_VPRINTF, make_unique<kf_stdio_output_fn> ());
 650
 651   kfm.add ("ferror", make_unique<kf_ferror> ());
 652   kfm.add ("fgets", make_unique<kf_fgets> ());
 653   kfm.add ("fgets_unlocked", make_unique<kf_fgets> ()); // non-standard
 654   kfm.add ("fileno", make_unique<kf_fileno> ());
 655   kfm.add ("fread", make_unique<kf_fread> ());
 656   kfm.add ("getc", make_unique<kf_getc> ());
 657   kfm.add ("getchar", make_unique<kf_getchar> ());
 658 }
 659
 660 #if CHECKING_P
 661
 662 namespace selftest {
 663
 664 /* Run all of the selftests within this file.  */
 665
 666 void
 667 analyzer_sm_file_cc_tests ()
 668 {
 669   function_set fs = get_file_using_fns ();
 670   fs.assert_sorted ();
 671   fs.assert_sane ();
 672 }
 673
 674 } // namespace selftest
 675
 676 #endif /* CHECKING_P */
 677
 678 } // namespace ana
 679
 680 #endif /* #if ENABLE_ANALYZER */