d746f78e8d82918c6109170e6fced1f1b70a5dca
[lttng-tools.git] / src / lib / lttng-ctl / filter / filter-parser.y
1 %{
2 /*
3 * filter-parser.y
4 *
5 * LTTng filter expression parser
6 *
7 * Copyright 2012 - Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
8 *
9 * This library is free software; you can redistribute it and/or modify it
10 * under the terms of the GNU Lesser General Public License, version 2.1 only,
11 * as published by the Free Software Foundation.
12 *
13 * This library is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 * Lesser General Public License for more details.
17 *
18 * You should have received a copy of the GNU Lesser General Public License
19 * along with this library; if not, write to the Free Software Foundation,
20 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21 *
22 * Grammar inspired from http://www.quut.com/c/ANSI-C-grammar-y.html
23 */
24
25 #include <stdio.h>
26 #include <unistd.h>
27 #include <string.h>
28 #include <stdlib.h>
29 #include <assert.h>
30 #include <errno.h>
31 #include <inttypes.h>
32 #include "filter-ast.h"
33 #include "filter-parser.h"
34
35 #include <common/macros.h>
36
37 #define WIDTH_u64_SCANF_IS_A_BROKEN_API "20"
38 #define WIDTH_o64_SCANF_IS_A_BROKEN_API "22"
39 #define WIDTH_x64_SCANF_IS_A_BROKEN_API "17"
40 #define WIDTH_lg_SCANF_IS_A_BROKEN_API "4096" /* Hugely optimistic approximation */
41
42 LTTNG_HIDDEN
43 int yydebug;
44 LTTNG_HIDDEN
45 int filter_parser_debug = 0;
46
47 LTTNG_HIDDEN
48 int yyparse(struct filter_parser_ctx *parser_ctx, yyscan_t scanner);
49 LTTNG_HIDDEN
50 int yylex(union YYSTYPE *yyval, yyscan_t scanner);
51 LTTNG_HIDDEN
52 int yylex_init_extra(struct filter_parser_ctx *parser_ctx, yyscan_t * ptr_yy_globals);
53 LTTNG_HIDDEN
54 int yylex_destroy(yyscan_t yyparser_ctx);
55 LTTNG_HIDDEN
56 void yyrestart(FILE * in_str, yyscan_t parser_ctx);
57
58 struct gc_string {
59 struct cds_list_head gc;
60 size_t alloclen;
61 char s[];
62 };
63
64 static const char *node_type_to_str[] = {
65 [ NODE_UNKNOWN ] = "NODE_UNKNOWN",
66 [ NODE_ROOT ] = "NODE_ROOT",
67 [ NODE_EXPRESSION ] = "NODE_EXPRESSION",
68 [ NODE_OP ] = "NODE_OP",
69 [ NODE_UNARY_OP ] = "NODE_UNARY_OP",
70 };
71
72 LTTNG_HIDDEN
73 const char *node_type(struct filter_node *node)
74 {
75 if (node->type < NR_NODE_TYPES)
76 return node_type_to_str[node->type];
77 else
78 return NULL;
79 }
80
81 static struct gc_string *gc_string_alloc(struct filter_parser_ctx *parser_ctx,
82 size_t len)
83 {
84 struct gc_string *gstr;
85 size_t alloclen;
86
87 /* TODO: could be faster with find first bit or glib Gstring */
88 /* sizeof long to account for malloc header (int or long ?) */
89 for (alloclen = 8; alloclen < sizeof(long) + sizeof(*gstr) + len;
90 alloclen *= 2);
91
92 gstr = malloc(alloclen);
93 cds_list_add(&gstr->gc, &parser_ctx->allocated_strings);
94 gstr->alloclen = alloclen;
95 return gstr;
96 }
97
98 /*
99 * note: never use gc_string_append on a string that has external references.
100 * gsrc will be garbage collected immediately, and gstr might be.
101 * Should only be used to append characters to a string literal or constant.
102 */
103 LTTNG_HIDDEN
104 struct gc_string *gc_string_append(struct filter_parser_ctx *parser_ctx,
105 struct gc_string *gstr,
106 struct gc_string *gsrc)
107 {
108 size_t newlen = strlen(gsrc->s) + strlen(gstr->s) + 1;
109 size_t alloclen;
110
111 /* TODO: could be faster with find first bit or glib Gstring */
112 /* sizeof long to account for malloc header (int or long ?) */
113 for (alloclen = 8; alloclen < sizeof(long) + sizeof(*gstr) + newlen;
114 alloclen *= 2);
115
116 if (alloclen > gstr->alloclen) {
117 struct gc_string *newgstr;
118
119 newgstr = gc_string_alloc(parser_ctx, newlen);
120 strcpy(newgstr->s, gstr->s);
121 strcat(newgstr->s, gsrc->s);
122 cds_list_del(&gstr->gc);
123 free(gstr);
124 gstr = newgstr;
125 } else {
126 strcat(gstr->s, gsrc->s);
127 }
128 cds_list_del(&gsrc->gc);
129 free(gsrc);
130 return gstr;
131 }
132
133 LTTNG_HIDDEN
134 void setstring(struct filter_parser_ctx *parser_ctx, YYSTYPE *lvalp, const char *src)
135 {
136 lvalp->gs = gc_string_alloc(parser_ctx, strlen(src) + 1);
137 strcpy(lvalp->gs->s, src);
138 }
139
140 static struct filter_node *make_node(struct filter_parser_ctx *scanner,
141 enum node_type type)
142 {
143 struct filter_ast *ast = filter_parser_get_ast(scanner);
144 struct filter_node *node;
145
146 node = malloc(sizeof(*node));
147 if (!node)
148 return NULL;
149 memset(node, 0, sizeof(*node));
150 node->type = type;
151 cds_list_add(&node->gc, &ast->allocated_nodes);
152
153 switch (type) {
154 case NODE_ROOT:
155 fprintf(stderr, "[error] %s: trying to create root node\n", __func__);
156 break;
157
158 case NODE_EXPRESSION:
159 break;
160 case NODE_OP:
161 break;
162 case NODE_UNARY_OP:
163 break;
164
165 case NODE_UNKNOWN:
166 default:
167 fprintf(stderr, "[error] %s: unknown node type %d\n", __func__,
168 (int) type);
169 break;
170 }
171
172 return node;
173 }
174
175 static struct filter_node *make_op_node(struct filter_parser_ctx *scanner,
176 enum op_type type,
177 struct filter_node *lchild,
178 struct filter_node *rchild)
179 {
180 struct filter_ast *ast = filter_parser_get_ast(scanner);
181 struct filter_node *node;
182
183 node = malloc(sizeof(*node));
184 if (!node)
185 return NULL;
186 memset(node, 0, sizeof(*node));
187 node->type = NODE_OP;
188 cds_list_add(&node->gc, &ast->allocated_nodes);
189 node->u.op.type = type;
190 node->u.op.lchild = lchild;
191 node->u.op.rchild = rchild;
192 return node;
193 }
194
195 LTTNG_HIDDEN
196 void yyerror(struct filter_parser_ctx *parser_ctx, yyscan_t scanner, const char *str)
197 {
198 fprintf(stderr, "error %s\n", str);
199 }
200
201 LTTNG_HIDDEN
202 int yywrap(void)
203 {
204 return 1;
205 }
206
207 #define parse_error(parser_ctx, str) \
208 do { \
209 yyerror(parser_ctx, parser_ctx->scanner, YY_("parse error: " str "\n")); \
210 YYERROR; \
211 } while (0)
212
213 static void free_strings(struct cds_list_head *list)
214 {
215 struct gc_string *gstr, *tmp;
216
217 cds_list_for_each_entry_safe(gstr, tmp, list, gc)
218 free(gstr);
219 }
220
221 static struct filter_ast *filter_ast_alloc(void)
222 {
223 struct filter_ast *ast;
224
225 ast = malloc(sizeof(*ast));
226 if (!ast)
227 return NULL;
228 memset(ast, 0, sizeof(*ast));
229 CDS_INIT_LIST_HEAD(&ast->allocated_nodes);
230 ast->root.type = NODE_ROOT;
231 return ast;
232 }
233
234 static void filter_ast_free(struct filter_ast *ast)
235 {
236 struct filter_node *node, *tmp;
237
238 cds_list_for_each_entry_safe(node, tmp, &ast->allocated_nodes, gc)
239 free(node);
240 free(ast);
241 }
242
243 LTTNG_HIDDEN
244 int filter_parser_ctx_append_ast(struct filter_parser_ctx *parser_ctx)
245 {
246 return yyparse(parser_ctx, parser_ctx->scanner);
247 }
248
249 LTTNG_HIDDEN
250 struct filter_parser_ctx *filter_parser_ctx_alloc(FILE *input)
251 {
252 struct filter_parser_ctx *parser_ctx;
253 int ret;
254
255 yydebug = filter_parser_debug;
256
257 parser_ctx = malloc(sizeof(*parser_ctx));
258 if (!parser_ctx)
259 return NULL;
260 memset(parser_ctx, 0, sizeof(*parser_ctx));
261
262 ret = yylex_init_extra(parser_ctx, &parser_ctx->scanner);
263 if (ret) {
264 fprintf(stderr, "yylex_init error\n");
265 goto cleanup_parser_ctx;
266 }
267 /* Start processing new stream */
268 yyrestart(input, parser_ctx->scanner);
269
270 parser_ctx->ast = filter_ast_alloc();
271 if (!parser_ctx->ast)
272 goto cleanup_lexer;
273 CDS_INIT_LIST_HEAD(&parser_ctx->allocated_strings);
274
275 if (yydebug)
276 fprintf(stdout, "parser_ctx input is a%s.\n",
277 isatty(fileno(input)) ? "n interactive tty" :
278 " noninteractive file");
279
280 return parser_ctx;
281
282 cleanup_lexer:
283 ret = yylex_destroy(parser_ctx->scanner);
284 if (!ret)
285 fprintf(stderr, "yylex_destroy error\n");
286 cleanup_parser_ctx:
287 free(parser_ctx);
288 return NULL;
289 }
290
291 LTTNG_HIDDEN
292 void filter_parser_ctx_free(struct filter_parser_ctx *parser_ctx)
293 {
294 int ret;
295
296 free_strings(&parser_ctx->allocated_strings);
297 filter_ast_free(parser_ctx->ast);
298 ret = yylex_destroy(parser_ctx->scanner);
299 if (ret)
300 fprintf(stderr, "yylex_destroy error\n");
301 free(parser_ctx);
302 }
303
304 %}
305
306 %define api.pure
307 /* %locations */
308 %parse-param {struct filter_parser_ctx *parser_ctx}
309 %parse-param {yyscan_t scanner}
310 %lex-param {yyscan_t scanner}
311 %start translation_unit
312 %token CHARACTER_CONSTANT_START SQUOTE STRING_LITERAL_START DQUOTE
313 %token ESCSEQ CHAR_STRING_TOKEN
314 %token DECIMAL_CONSTANT OCTAL_CONSTANT HEXADECIMAL_CONSTANT FLOAT_CONSTANT
315 %token LSBRAC RSBRAC LPAREN RPAREN LBRAC RBRAC RARROW
316 %token STAR PLUS MINUS
317 %token MOD_OP DIV_OP RIGHT_OP LEFT_OP
318 %token EQ_OP NE_OP LE_OP GE_OP LT_OP GT_OP AND_OP OR_OP NOT_OP
319 %token ASSIGN COLON SEMICOLON DOTDOTDOT DOT EQUAL COMMA
320 %token XOR_BIN AND_BIN OR_BIN NOT_BIN
321
322 %token <gs> IDENTIFIER GLOBAL_IDENTIFIER
323 %token ERROR
324 %union
325 {
326 long long ll;
327 char c;
328 struct gc_string *gs;
329 struct filter_node *n;
330 }
331
332 %type <gs> s_char s_char_sequence c_char c_char_sequence
333
334 %type <n> primary_expression
335 %type <n> postfix_expression
336 %type <n> unary_expression
337 %type <n> unary_operator
338 %type <n> multiplicative_expression
339 %type <n> additive_expression
340 %type <n> shift_expression
341 %type <n> relational_expression
342 %type <n> equality_expression
343 %type <n> and_expression
344 %type <n> exclusive_or_expression
345 %type <n> inclusive_or_expression
346 %type <n> logical_and_expression
347 %type <n> logical_or_expression
348 %type <n> expression
349
350 %%
351
352
353 /* 1.5 Constants */
354
355 c_char_sequence:
356 c_char
357 { $$ = $1; }
358 | c_char_sequence c_char
359 { $$ = gc_string_append(parser_ctx, $1, $2); }
360 ;
361
362 c_char:
363 CHAR_STRING_TOKEN
364 { $$ = yylval.gs; }
365 | ESCSEQ
366 {
367 parse_error(parser_ctx, "escape sequences not supported yet");
368 }
369 ;
370
371 /* 1.6 String literals */
372
373 s_char_sequence:
374 s_char
375 { $$ = $1; }
376 | s_char_sequence s_char
377 { $$ = gc_string_append(parser_ctx, $1, $2); }
378 ;
379
380 s_char:
381 CHAR_STRING_TOKEN
382 { $$ = yylval.gs; }
383 | ESCSEQ
384 {
385 parse_error(parser_ctx, "escape sequences not supported yet");
386 }
387 ;
388
389 primary_expression
390 : IDENTIFIER
391 {
392 $$ = make_node(parser_ctx, NODE_EXPRESSION);
393 $$->u.expression.type = AST_EXP_IDENTIFIER;
394 $$->u.expression.u.identifier = yylval.gs->s;
395 }
396 | GLOBAL_IDENTIFIER
397 {
398 $$ = make_node(parser_ctx, NODE_EXPRESSION);
399 $$->u.expression.type = AST_EXP_GLOBAL_IDENTIFIER;
400 $$->u.expression.u.identifier = yylval.gs->s;
401 }
402
403 | DECIMAL_CONSTANT
404 {
405 $$ = make_node(parser_ctx, NODE_EXPRESSION);
406 $$->u.expression.type = AST_EXP_CONSTANT;
407 if (sscanf(yylval.gs->s, "%" WIDTH_u64_SCANF_IS_A_BROKEN_API SCNu64,
408 &$$->u.expression.u.constant) != 1) {
409 parse_error(parser_ctx, "cannot scanf decimal constant");
410 }
411 }
412 | OCTAL_CONSTANT
413 {
414 $$ = make_node(parser_ctx, NODE_EXPRESSION);
415 $$->u.expression.type = AST_EXP_CONSTANT;
416 if (!strcmp(yylval.gs->s, "0")) {
417 $$->u.expression.u.constant = 0;
418 } else if (sscanf(yylval.gs->s, "0%" WIDTH_o64_SCANF_IS_A_BROKEN_API SCNo64,
419 &$$->u.expression.u.constant) != 1) {
420 parse_error(parser_ctx, "cannot scanf octal constant");
421 }
422 }
423 | HEXADECIMAL_CONSTANT
424 {
425 $$ = make_node(parser_ctx, NODE_EXPRESSION);
426 $$->u.expression.type = AST_EXP_CONSTANT;
427 if (sscanf(yylval.gs->s, "0x%" WIDTH_x64_SCANF_IS_A_BROKEN_API SCNx64,
428 &$$->u.expression.u.constant) != 1) {
429 parse_error(parser_ctx, "cannot scanf hexadecimal constant");
430 }
431 }
432 | FLOAT_CONSTANT
433 {
434 $$ = make_node(parser_ctx, NODE_EXPRESSION);
435 $$->u.expression.type = AST_EXP_FLOAT_CONSTANT;
436 if (sscanf(yylval.gs->s, "%" WIDTH_lg_SCANF_IS_A_BROKEN_API "lg",
437 &$$->u.expression.u.float_constant) != 1) {
438 parse_error(parser_ctx, "cannot scanf float constant");
439 }
440 }
441 | STRING_LITERAL_START DQUOTE
442 {
443 $$ = make_node(parser_ctx, NODE_EXPRESSION);
444 $$->u.expression.type = AST_EXP_STRING;
445 $$->u.expression.u.string = "";
446 }
447 | STRING_LITERAL_START s_char_sequence DQUOTE
448 {
449 $$ = make_node(parser_ctx, NODE_EXPRESSION);
450 $$->u.expression.type = AST_EXP_STRING;
451 $$->u.expression.u.string = $2->s;
452 }
453 | CHARACTER_CONSTANT_START c_char_sequence SQUOTE
454 {
455 $$ = make_node(parser_ctx, NODE_EXPRESSION);
456 $$->u.expression.type = AST_EXP_STRING;
457 $$->u.expression.u.string = $2->s;
458 }
459 | LPAREN expression RPAREN
460 {
461 $$ = make_node(parser_ctx, NODE_EXPRESSION);
462 $$->u.expression.type = AST_EXP_NESTED;
463 $$->u.expression.u.child = $2;
464 }
465 ;
466
467 postfix_expression
468 : primary_expression
469 { $$ = $1; }
470 | postfix_expression DOT IDENTIFIER
471 {
472 $$ = make_node(parser_ctx, NODE_EXPRESSION);
473 $$->u.expression.type = AST_EXP_IDENTIFIER;
474 $$->u.expression.post_op = AST_LINK_DOT;
475 $$->u.expression.u.identifier = $3->s;
476 $$->u.expression.prev = $1;
477 }
478 | postfix_expression RARROW IDENTIFIER
479 {
480 $$ = make_node(parser_ctx, NODE_EXPRESSION);
481 $$->u.expression.type = AST_EXP_IDENTIFIER;
482 $$->u.expression.post_op = AST_LINK_RARROW;
483 $$->u.expression.u.identifier = $3->s;
484 $$->u.expression.prev = $1;
485 }
486 ;
487
488 unary_expression
489 : postfix_expression
490 { $$ = $1; }
491 | unary_operator unary_expression
492 {
493 $$ = $1;
494 $$->u.unary_op.child = $2;
495 }
496 ;
497
498 unary_operator
499 : PLUS
500 {
501 $$ = make_node(parser_ctx, NODE_UNARY_OP);
502 $$->u.unary_op.type = AST_UNARY_PLUS;
503 }
504 | MINUS
505 {
506 $$ = make_node(parser_ctx, NODE_UNARY_OP);
507 $$->u.unary_op.type = AST_UNARY_MINUS;
508 }
509 | NOT_OP
510 {
511 $$ = make_node(parser_ctx, NODE_UNARY_OP);
512 $$->u.unary_op.type = AST_UNARY_NOT;
513 }
514 | NOT_BIN
515 {
516 $$ = make_node(parser_ctx, NODE_UNARY_OP);
517 $$->u.unary_op.type = AST_UNARY_BIN_NOT;
518 }
519 ;
520
521 multiplicative_expression
522 : unary_expression
523 { $$ = $1; }
524 | multiplicative_expression STAR unary_expression
525 {
526 $$ = make_op_node(parser_ctx, AST_OP_MUL, $1, $3);
527 }
528 | multiplicative_expression DIV_OP unary_expression
529 {
530 $$ = make_op_node(parser_ctx, AST_OP_DIV, $1, $3);
531 }
532 | multiplicative_expression MOD_OP unary_expression
533 {
534 $$ = make_op_node(parser_ctx, AST_OP_MOD, $1, $3);
535 }
536 ;
537
538 additive_expression
539 : multiplicative_expression
540 { $$ = $1; }
541 | additive_expression PLUS multiplicative_expression
542 {
543 $$ = make_op_node(parser_ctx, AST_OP_PLUS, $1, $3);
544 }
545 | additive_expression MINUS multiplicative_expression
546 {
547 $$ = make_op_node(parser_ctx, AST_OP_MINUS, $1, $3);
548 }
549 ;
550
551 shift_expression
552 : additive_expression
553 { $$ = $1; }
554 | shift_expression LEFT_OP additive_expression
555 {
556 $$ = make_op_node(parser_ctx, AST_OP_LSHIFT, $1, $3);
557 }
558 | shift_expression RIGHT_OP additive_expression
559 {
560 $$ = make_op_node(parser_ctx, AST_OP_RSHIFT, $1, $3);
561 }
562 ;
563
564 relational_expression
565 : shift_expression
566 { $$ = $1; }
567 | relational_expression LT_OP shift_expression
568 {
569 $$ = make_op_node(parser_ctx, AST_OP_LT, $1, $3);
570 }
571 | relational_expression GT_OP shift_expression
572 {
573 $$ = make_op_node(parser_ctx, AST_OP_GT, $1, $3);
574 }
575 | relational_expression LE_OP shift_expression
576 {
577 $$ = make_op_node(parser_ctx, AST_OP_LE, $1, $3);
578 }
579 | relational_expression GE_OP shift_expression
580 {
581 $$ = make_op_node(parser_ctx, AST_OP_GE, $1, $3);
582 }
583 ;
584
585 equality_expression
586 : relational_expression
587 { $$ = $1; }
588 | equality_expression EQ_OP relational_expression
589 {
590 $$ = make_op_node(parser_ctx, AST_OP_EQ, $1, $3);
591 }
592 | equality_expression NE_OP relational_expression
593 {
594 $$ = make_op_node(parser_ctx, AST_OP_NE, $1, $3);
595 }
596 ;
597
598 and_expression
599 : equality_expression
600 { $$ = $1; }
601 | and_expression AND_BIN equality_expression
602 {
603 $$ = make_op_node(parser_ctx, AST_OP_BIN_AND, $1, $3);
604 }
605 ;
606
607 exclusive_or_expression
608 : and_expression
609 { $$ = $1; }
610 | exclusive_or_expression XOR_BIN and_expression
611 {
612 $$ = make_op_node(parser_ctx, AST_OP_BIN_XOR, $1, $3);
613 }
614 ;
615
616 inclusive_or_expression
617 : exclusive_or_expression
618 { $$ = $1; }
619 | inclusive_or_expression OR_BIN exclusive_or_expression
620 {
621 $$ = make_op_node(parser_ctx, AST_OP_BIN_OR, $1, $3);
622 }
623 ;
624
625 logical_and_expression
626 : inclusive_or_expression
627 { $$ = $1; }
628 | logical_and_expression AND_OP inclusive_or_expression
629 {
630 $$ = make_op_node(parser_ctx, AST_OP_AND, $1, $3);
631 }
632 ;
633
634 logical_or_expression
635 : logical_and_expression
636 { $$ = $1; }
637 | logical_or_expression OR_OP logical_and_expression
638 {
639 $$ = make_op_node(parser_ctx, AST_OP_OR, $1, $3);
640 }
641 ;
642
643 expression
644 : logical_or_expression
645 { $$ = $1; }
646 ;
647
648 translation_unit
649 : expression
650 {
651 parser_ctx->ast->root.u.root.child = $1;
652 }
653 ;
This page took 0.042181 seconds and 4 git commands to generate.