xref: /illumos-gate/usr/src/tools/smatch/src/symbol.h (revision c85f09cc92abd00c84e58ec9f0f5d942906cb713)
1 #ifndef SYMBOL_H
2 #define SYMBOL_H
3 /*
4  * Basic symbol and namespace definitions.
5  *
6  * Copyright (C) 2003 Transmeta Corp.
7  *               2003 Linus Torvalds
8  *
9  * Permission is hereby granted, free of charge, to any person obtaining a copy
10  * of this software and associated documentation files (the "Software"), to deal
11  * in the Software without restriction, including without limitation the rights
12  * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
13  * copies of the Software, and to permit persons to whom the Software is
14  * furnished to do so, subject to the following conditions:
15  *
16  * The above copyright notice and this permission notice shall be included in
17  * all copies or substantial portions of the Software.
18  *
19  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
20  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
21  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
22  * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
23  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
24  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
25  * THE SOFTWARE.
26  */
27 
28 #include "token.h"
29 #include "target.h"
30 
31 /*
32  * An identifier with semantic meaning is a "symbol".
33  *
34  * There's a 1:n relationship: each symbol is always
35  * associated with one identifier, while each identifier
36  * can have one or more semantic meanings due to C scope
37  * rules.
38  *
39  * The progression is symbol -> token -> identifier. The
40  * token contains the information on where the symbol was
41  * declared.
42  */
43 enum namespace {
44 	NS_NONE = 0,
45 	NS_MACRO = 1,
46 	NS_TYPEDEF = 2,
47 	NS_STRUCT = 4,  // Also used for unions and enums.
48 	NS_LABEL = 8,
49 	NS_SYMBOL = 16,
50 	NS_ITERATOR = 32,
51 	NS_PREPROCESSOR = 64,
52 	NS_UNDEF = 128,
53 	NS_KEYWORD = 256,
54 };
55 
56 enum type {
57 	SYM_UNINITIALIZED,
58 	SYM_PREPROCESSOR,
59 	SYM_BASETYPE,
60 	SYM_NODE,
61 	SYM_PTR,
62 	SYM_FN,
63 	SYM_ARRAY,
64 	SYM_STRUCT,
65 	SYM_UNION,
66 	SYM_ENUM,
67 	SYM_TYPEDEF,
68 	SYM_TYPEOF,
69 	SYM_MEMBER,
70 	SYM_BITFIELD,
71 	SYM_LABEL,
72 	SYM_RESTRICT,
73 	SYM_FOULED,
74 	SYM_KEYWORD,
75 	SYM_BAD,
76 };
77 
78 enum keyword {
79 	KW_SPECIFIER 	= 1 << 0,
80 	KW_MODIFIER	= 1 << 1,
81 	KW_QUALIFIER	= 1 << 2,
82 	KW_ATTRIBUTE	= 1 << 3,
83 	KW_STATEMENT	= 1 << 4,
84 	KW_ASM		= 1 << 5,
85 	KW_MODE		= 1 << 6,
86 	KW_SHORT	= 1 << 7,
87 	KW_LONG		= 1 << 8,
88 	KW_EXACT	= 1 << 9,
89 };
90 
91 struct context {
92 	struct expression *context;
93 	unsigned int in, out;
94 };
95 
96 extern struct context *alloc_context(void);
97 
98 DECLARE_PTR_LIST(context_list, struct context);
99 
100 struct ctype {
101 	unsigned long modifiers;
102 	unsigned long alignment;
103 	struct context_list *contexts;
104 	struct ident *as;
105 	struct symbol *base_type;
106 };
107 
108 struct decl_state {
109 	struct ctype ctype;
110 	struct ident **ident;
111 	struct symbol_op *mode;
112 	unsigned char prefer_abstract, is_inline, storage_class, is_tls;
113 	unsigned char is_ext_visible;
114 };
115 
116 struct symbol_op {
117 	enum keyword type;
118 	int (*evaluate)(struct expression *);
119 	int (*expand)(struct expression *, int);
120 	int (*args)(struct expression *);
121 
122 	/* keywords */
123 	struct token *(*declarator)(struct token *token, struct decl_state *ctx);
124 	struct token *(*statement)(struct token *token, struct statement *stmt);
125 	struct token *(*toplevel)(struct token *token, struct symbol_list **list);
126 	struct token *(*attribute)(struct token *token, struct symbol *attr, struct decl_state *ctx);
127 	struct symbol *(*to_mode)(struct symbol *);
128 	void          (*asm_modifier)(struct token *token, unsigned long *mods);
129 
130 	int test, set, class;
131 };
132 
133 
134 #define SYM_ATTR_WEAK		0
135 #define SYM_ATTR_NORMAL		1
136 #define SYM_ATTR_STRONG		2
137 
138 struct symbol {
139 	enum type type:8;
140 	enum namespace namespace:9;
141 	unsigned char used:1, attr:2, enum_member:1, bound:1;
142 	struct position pos;		/* Where this symbol was declared */
143 	struct position endpos;		/* Where this symbol ends*/
144 	struct ident *ident;		/* What identifier this symbol is associated with */
145 	struct symbol *next_id;		/* Next semantic symbol that shares this identifier */
146 	struct symbol	*replace;	/* What is this symbol shadowed by in copy-expression */
147 	struct scope	*scope;
148 	union {
149 		struct symbol	*same_symbol;
150 		struct symbol	*next_subobject;
151 	};
152 
153 	struct symbol_op *op;
154 
155 	union {
156 		struct /* NS_MACRO */ {
157 			struct token *expansion;
158 			struct token *arglist;
159 			struct scope *used_in;
160 			void (*expander)(struct token *);
161 		};
162 		struct /* NS_PREPROCESSOR */ {
163 			int (*handler)(struct stream *, struct token **, struct token *);
164 			int normal;
165 		};
166 		struct /* NS_SYMBOL */ {
167 			unsigned long	offset;
168 			int		bit_size;
169 			unsigned int	bit_offset:8,
170 					variadic:1,
171 					initialized:1,
172 					examined:1,
173 					expanding:1,
174 					evaluated:1,
175 					string:1,
176 					designated_init:1,
177 					forced_arg:1,
178 					accessed:1,
179 					builtin:1,
180 					torename:1,
181 					transparent_union:1;
182 			struct expression *array_size;
183 			struct ctype ctype;
184 			struct symbol_list *arguments;
185 			struct statement *stmt;
186 			struct symbol_list *symbol_list;
187 			struct statement *inline_stmt;
188 			struct symbol_list *inline_symbol_list;
189 			struct expression *initializer;
190 			struct entrypoint *ep;
191 			struct symbol *definition;
192 		};
193 	};
194 	union /* backend */ {
195 		struct basic_block *bb_target;	/* label */
196 		void *aux;			/* Auxiliary info, e.g. backend information */
197 		struct {			/* sparse ctags */
198 			char kind;
199 			unsigned char visited:1;
200 		};
201 	};
202 	pseudo_t pseudo;
203 };
204 
205 /* Modifiers */
206 #define MOD_AUTO		0x00000001
207 #define MOD_REGISTER		0x00000002
208 #define MOD_STATIC		0x00000004
209 #define MOD_EXTERN		0x00000008
210 #define MOD_TOPLEVEL		0x00000010	// scoping..
211 #define MOD_TLS			0x00000020
212 #define MOD_ASM_GOTO		MOD_TLS		// never used together
213 #define MOD_INLINE		0x00000040
214 
215 #define MOD_ASSIGNED		0x00000080
216 #define MOD_ADDRESSABLE		0x00000100
217 
218 #define MOD_CONST		0x00000200
219 #define MOD_VOLATILE		0x00000400
220 #define MOD_RESTRICT		0x00000800
221 #define MOD_ATOMIC		0x00001000
222 
223 #define MOD_SIGNED		0x00002000
224 #define MOD_UNSIGNED		0x00004000
225 #define MOD_EXPLICITLY_SIGNED	0x00008000
226 
227 #define MOD_TYPE		0x00010000
228 #define MOD_USERTYPE		0x00020000
229 #define MOD_CHAR		0x00040000
230 #define MOD_SHORT		0x00080000
231 #define MOD_LONG		0x00100000
232 #define MOD_LONGLONG		0x00200000
233 #define MOD_LONGLONGLONG	0x00400000
234 
235 #define MOD_SAFE		0x00800000	// non-null/non-trapping pointer
236 #define MOD_PURE		0x01000000
237 #define MOD_BITWISE		0x02000000
238 #define MOD_NOCAST		0x04000000
239 #define MOD_NODEREF		0x08000000
240 #define MOD_NORETURN		0x10000000
241 #define MOD_EXT_VISIBLE		0x20000000
242 
243 
244 #define MOD_ACCESS	(MOD_ASSIGNED | MOD_ADDRESSABLE)
245 #define MOD_NONLOCAL	(MOD_EXTERN | MOD_TOPLEVEL)
246 #define MOD_STORAGE	(MOD_AUTO | MOD_REGISTER | MOD_STATIC | MOD_EXTERN | MOD_INLINE | MOD_TOPLEVEL)
247 #define MOD_SIGNEDNESS	(MOD_SIGNED | MOD_UNSIGNED | MOD_EXPLICITLY_SIGNED)
248 #define MOD_LONG_ALL	(MOD_LONG | MOD_LONGLONG | MOD_LONGLONGLONG)
249 #define MOD_SPECIFIER	(MOD_CHAR | MOD_SHORT | MOD_LONG_ALL | MOD_SIGNEDNESS)
250 #define MOD_SIZE	(MOD_CHAR | MOD_SHORT | MOD_LONG_ALL)
251 #define MOD_IGNORE	(MOD_STORAGE | MOD_ACCESS | MOD_USERTYPE | MOD_EXPLICITLY_SIGNED | MOD_EXT_VISIBLE)
252 #define	MOD_QUALIFIER	(MOD_CONST | MOD_VOLATILE | MOD_RESTRICT | MOD_ATOMIC)
253 #define MOD_PTRINHERIT	(MOD_QUALIFIER | MOD_NODEREF | MOD_NORETURN | MOD_NOCAST)
254 /* modifiers preserved by typeof() operator */
255 #define MOD_TYPEOF	(MOD_QUALIFIER | MOD_NOCAST | MOD_SPECIFIER)
256 
257 
258 /* Current parsing/evaluation function */
259 extern struct symbol *current_fn;
260 
261 /* Abstract types */
262 extern struct symbol	int_type,
263 			fp_type;
264 
265 /* C types */
266 extern struct symbol	bool_ctype, void_ctype, type_ctype,
267 			char_ctype, schar_ctype, uchar_ctype,
268 			short_ctype, sshort_ctype, ushort_ctype,
269 			int_ctype, sint_ctype, uint_ctype,
270 			long_ctype, slong_ctype, ulong_ctype,
271 			llong_ctype, sllong_ctype, ullong_ctype,
272 			lllong_ctype, slllong_ctype, ulllong_ctype,
273 			float_ctype, double_ctype, ldouble_ctype,
274 			string_ctype, ptr_ctype, lazy_ptr_ctype,
275 			incomplete_ctype, label_ctype, bad_ctype,
276 			null_ctype;
277 extern struct symbol	int_ptr_ctype, uint_ptr_ctype;
278 extern struct symbol	long_ptr_ctype, ulong_ptr_ctype;
279 extern struct symbol	llong_ptr_ctype, ullong_ptr_ctype;
280 extern struct symbol	float32_ctype, float32x_ctype;
281 extern struct symbol	float64_ctype, float64x_ctype;
282 extern struct symbol	float128_ctype;
283 extern struct symbol	const_void_ctype, const_char_ctype;
284 extern struct symbol	const_ptr_ctype, const_string_ctype;
285 
286 #define	uintptr_ctype	 size_t_ctype
287 #define	 intptr_ctype	ssize_t_ctype
288 
289 /* Special internal symbols */
290 extern struct symbol	zero_int;
291 
292 #define __IDENT(n,str,res) \
293 	extern struct ident n
294 #include "ident-list.h"
295 
296 
297 extern struct symbol_list *translation_unit_used_list;
298 
299 extern void access_symbol(struct symbol *);
300 
301 extern const char * type_difference(struct ctype *c1, struct ctype *c2,
302 	unsigned long mod1, unsigned long mod2);
303 
304 extern struct symbol *lookup_symbol(struct ident *, enum namespace);
305 extern struct symbol *create_symbol(int stream, const char *name, int type, int namespace);
306 extern void init_symbols(void);
307 extern void init_builtins(int stream);
308 extern void declare_builtins(void);
309 extern void init_ctype(void);
310 extern void init_target(void);
311 extern struct symbol *alloc_symbol(struct position, int type);
312 extern void show_type(struct symbol *);
313 extern const char *modifier_string(unsigned long mod);
314 extern void show_symbol(struct symbol *);
315 extern int show_symbol_expr_init(struct symbol *sym);
316 extern void show_type_list(struct symbol *);
317 extern void show_symbol_list(struct symbol_list *, const char *);
318 extern void add_symbol(struct symbol_list **, struct symbol *);
319 extern void bind_symbol(struct symbol *, struct ident *, enum namespace);
320 
321 extern struct symbol *examine_symbol_type(struct symbol *);
322 extern struct symbol *examine_pointer_target(struct symbol *);
323 extern const char *show_as(struct ident *as);
324 extern const char *show_typename(struct symbol *sym);
325 extern const char *builtin_typename(struct symbol *sym);
326 extern const char *builtin_type_suffix(struct symbol *sym);
327 extern const char *builtin_ctypename(struct ctype *ctype);
328 extern const char* get_type_name(enum type type);
329 
330 extern void debug_symbol(struct symbol *);
331 extern void merge_type(struct symbol *sym, struct symbol *base_type);
332 extern void check_declaration(struct symbol *sym);
333 extern void check_duplicates(struct symbol *sym);
334 
valid_type(const struct symbol * ctype)335 static inline int valid_type(const struct symbol *ctype)
336 {
337 	return ctype && ctype != &bad_ctype;
338 }
339 
get_base_type(const struct symbol * sym)340 static inline struct symbol *get_base_type(const struct symbol *sym)
341 {
342 	return examine_symbol_type(sym->ctype.base_type);
343 }
344 
345 ///
346 // test if type is an integer type
347 //
348 // @return: ``1`` for plain integer type, enums & bitfields
349 //	but ``0`` for bitwise types!
is_int_type(const struct symbol * type)350 static inline int is_int_type(const struct symbol *type)
351 {
352 	if (type->type == SYM_NODE)
353 		type = type->ctype.base_type;
354 	if (type->type == SYM_ENUM)
355 		type = type->ctype.base_type;
356 	return type->type == SYM_BITFIELD ||
357 	       type->ctype.base_type == &int_type;
358 }
359 
is_enum_type(const struct symbol * type)360 static inline int is_enum_type(const struct symbol *type)
361 {
362 	if (type->type == SYM_NODE)
363 		type = type->ctype.base_type;
364 	return (type->type == SYM_ENUM);
365 }
366 
is_signed_type(struct symbol * sym)367 static inline int is_signed_type(struct symbol *sym)
368 {
369 	if (sym->type == SYM_NODE)
370 		sym = sym->ctype.base_type;
371 	if (sym->type == SYM_PTR)
372 		return 0;
373 	return !(sym->ctype.modifiers & MOD_UNSIGNED);
374 }
375 
is_type_type(struct symbol * type)376 static inline int is_type_type(struct symbol *type)
377 {
378 	return (type->ctype.modifiers & MOD_TYPE) != 0;
379 }
380 
is_ptr_type(struct symbol * type)381 static inline int is_ptr_type(struct symbol *type)
382 {
383 	if (!type)
384 		return 0;
385 	if (type->type == SYM_NODE)
386 		type = type->ctype.base_type;
387 	return type->type == SYM_PTR || type->type == SYM_ARRAY || type->type == SYM_FN;
388 }
389 
is_func_type(struct symbol * type)390 static inline int is_func_type(struct symbol *type)
391 {
392 	if (type->type == SYM_NODE)
393 		type = type->ctype.base_type;
394 	return type->type == SYM_FN;
395 }
396 
is_array_type(struct symbol * type)397 static inline int is_array_type(struct symbol *type)
398 {
399 	if (type->type == SYM_NODE)
400 		type = type->ctype.base_type;
401 	return type->type == SYM_ARRAY;
402 }
403 
is_float_type(struct symbol * type)404 static inline int is_float_type(struct symbol *type)
405 {
406 	if (type->type == SYM_NODE)
407 		type = type->ctype.base_type;
408 	return type->ctype.base_type == &fp_type;
409 }
410 
is_byte_type(struct symbol * type)411 static inline int is_byte_type(struct symbol *type)
412 {
413 	return type->bit_size == bits_in_char && type->type != SYM_BITFIELD;
414 }
415 
is_void_type(struct symbol * type)416 static inline int is_void_type(struct symbol *type)
417 {
418 	if (type->type == SYM_NODE)
419 		type = type->ctype.base_type;
420 	return type == &void_ctype;
421 }
422 
is_bool_type(struct symbol * type)423 static inline int is_bool_type(struct symbol *type)
424 {
425 	if (type->type == SYM_NODE)
426 		type = type->ctype.base_type;
427 	return type == &bool_ctype;
428 }
429 
is_scalar_type(struct symbol * type)430 static inline int is_scalar_type(struct symbol *type)
431 {
432 	if (type->type == SYM_NODE)
433 		type = type->ctype.base_type;
434 	switch (type->type) {
435 	case SYM_ENUM:
436 	case SYM_BITFIELD:
437 	case SYM_PTR:
438 	case SYM_RESTRICT:	// OK, always integer types
439 		return 1;
440 	default:
441 		break;
442 	}
443 	if (type->ctype.base_type == &int_type)
444 		return 1;
445 	if (type->ctype.base_type == &fp_type)
446 		return 1;
447 	return 0;
448 }
449 
450 /// return true for integer & pointer types
is_integral_type(struct symbol * type)451 static inline bool is_integral_type(struct symbol *type)
452 {
453 	if (type->type == SYM_NODE)
454 		type = type->ctype.base_type;
455 	switch (type->type) {
456 	case SYM_ENUM:
457 	case SYM_PTR:
458 	case SYM_RESTRICT:	// OK, always integer types
459 		return 1;
460 	default:
461 		break;
462 	}
463 	if (type->ctype.base_type == &int_type)
464 		return 1;
465 	return 0;
466 }
467 
is_function(struct symbol * type)468 static inline int is_function(struct symbol *type)
469 {
470 	return type && type->type == SYM_FN;
471 }
472 
is_extern_inline(struct symbol * sym)473 static inline int is_extern_inline(struct symbol *sym)
474 {
475 	return (sym->ctype.modifiers & MOD_EXTERN) &&
476 		(sym->ctype.modifiers & MOD_INLINE) &&
477 		is_function(sym->ctype.base_type);
478 }
479 
get_sym_type(struct symbol * type)480 static inline int get_sym_type(struct symbol *type)
481 {
482 	if (type->type == SYM_NODE)
483 		type = type->ctype.base_type;
484 	if (type->type == SYM_ENUM)
485 		type = type->ctype.base_type;
486 	return type->type;
487 }
488 
extend_value(long long val,struct symbol * ctype)489 static inline long long extend_value(long long val, struct symbol *ctype)
490 {
491 	int is_signed = !(ctype->ctype.modifiers & MOD_UNSIGNED);
492 	unsigned size = ctype->bit_size;
493 
494 	return bits_extend(val, size, is_signed);
495 }
496 
lookup_keyword(struct ident * ident,enum namespace ns)497 static inline struct symbol *lookup_keyword(struct ident *ident, enum namespace ns)
498 {
499 	if (!ident->keyword)
500 		return NULL;
501 	return lookup_symbol(ident, ns);
502 }
503 
504 #define is_restricted_type(type) (get_sym_type(type) == SYM_RESTRICT)
505 #define is_fouled_type(type) (get_sym_type(type) == SYM_FOULED)
506 #define is_bitfield_type(type)   (get_sym_type(type) == SYM_BITFIELD)
507 
508 void create_fouled(struct symbol *type);
509 struct symbol *befoul(struct symbol *type);
510 
511 
512 extern struct ident bad_address_space;
513 
valid_as(struct ident * as)514 static inline bool valid_as(struct ident *as)
515 {
516 	return as && as != &bad_address_space;
517 }
518 
combine_address_space(struct position pos,struct ident ** tas,struct ident * sas)519 static inline void combine_address_space(struct position pos,
520 	struct ident **tas, struct ident *sas)
521 {
522 	struct ident *as;
523 	if (!sas)
524 		return;
525 	as = *tas;
526 	if (!as)
527 		*tas = sas;
528 	else if (as != sas) {
529 		*tas = &bad_address_space;
530 		sparse_error(pos, "multiple address spaces given");
531 	}
532 }
533 
534 #endif /* SYMBOL_H */
535