5 * LTTng filter expression parser
7 * Copyright 2012 - Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
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.
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.
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
22 * Grammar inspired from http://www.quut.com/c/ANSI-C-grammar-y.html
32 #include "filter-ast.h"
33 #include "filter-parser.h"
35 #include <common/macros.h>
40 int filter_parser_debug = 0;
43 int yyparse(struct filter_parser_ctx *parser_ctx);
45 int yylex(union YYSTYPE *yyval, struct filter_parser_ctx *parser_ctx);
47 int yylex_init_extra(struct filter_parser_ctx *parser_ctx, yyscan_t * ptr_yy_globals);
49 int yylex_destroy(yyscan_t yyparser_ctx);
51 void yyrestart(FILE * in_str, yyscan_t parser_ctx);
54 struct cds_list_head gc;
59 static const char *node_type_to_str[] = {
60 [ NODE_UNKNOWN ] = "NODE_UNKNOWN",
61 [ NODE_ROOT ] = "NODE_ROOT",
62 [ NODE_EXPRESSION ] = "NODE_EXPRESSION",
63 [ NODE_OP ] = "NODE_OP",
64 [ NODE_UNARY_OP ] = "NODE_UNARY_OP",
68 const char *node_type(struct filter_node *node)
70 if (node->type < NR_NODE_TYPES)
71 return node_type_to_str[node->type];
76 static struct gc_string *gc_string_alloc(struct filter_parser_ctx *parser_ctx,
79 struct gc_string *gstr;
82 /* TODO: could be faster with find first bit or glib Gstring */
83 /* sizeof long to account for malloc header (int or long ?) */
84 for (alloclen = 8; alloclen < sizeof(long) + sizeof(*gstr) + len;
87 gstr = malloc(alloclen);
88 cds_list_add(&gstr->gc, &parser_ctx->allocated_strings);
89 gstr->alloclen = alloclen;
94 * note: never use gc_string_append on a string that has external references.
95 * gsrc will be garbage collected immediately, and gstr might be.
96 * Should only be used to append characters to a string literal or constant.
99 struct gc_string *gc_string_append(struct filter_parser_ctx *parser_ctx,
100 struct gc_string *gstr,
101 struct gc_string *gsrc)
103 size_t newlen = strlen(gsrc->s) + strlen(gstr->s) + 1;
106 /* TODO: could be faster with find first bit or glib Gstring */
107 /* sizeof long to account for malloc header (int or long ?) */
108 for (alloclen = 8; alloclen < sizeof(long) + sizeof(*gstr) + newlen;
111 if (alloclen > gstr->alloclen) {
112 struct gc_string *newgstr;
114 newgstr = gc_string_alloc(parser_ctx, newlen);
115 strcpy(newgstr->s, gstr->s);
116 strcat(newgstr->s, gsrc->s);
117 cds_list_del(&gstr->gc);
121 strcat(gstr->s, gsrc->s);
123 cds_list_del(&gsrc->gc);
129 void setstring(struct filter_parser_ctx *parser_ctx, YYSTYPE *lvalp, const char *src)
131 lvalp->gs = gc_string_alloc(parser_ctx, strlen(src) + 1);
132 strcpy(lvalp->gs->s, src);
135 static struct filter_node *make_node(struct filter_parser_ctx *scanner,
138 struct filter_ast *ast = filter_parser_get_ast(scanner);
139 struct filter_node *node;
141 node = malloc(sizeof(*node));
144 memset(node, 0, sizeof(*node));
146 cds_list_add(&node->gc, &ast->allocated_nodes);
150 fprintf(stderr, "[error] %s: trying to create root node\n", __func__);
153 case NODE_EXPRESSION:
162 fprintf(stderr, "[error] %s: unknown node type %d\n", __func__,
170 static struct filter_node *make_op_node(struct filter_parser_ctx *scanner,
172 struct filter_node *lchild,
173 struct filter_node *rchild)
175 struct filter_ast *ast = filter_parser_get_ast(scanner);
176 struct filter_node *node;
178 node = malloc(sizeof(*node));
181 memset(node, 0, sizeof(*node));
182 node->type = NODE_OP;
183 cds_list_add(&node->gc, &ast->allocated_nodes);
184 node->u.op.type = type;
185 node->u.op.lchild = lchild;
186 node->u.op.rchild = rchild;
191 void yyerror(struct filter_parser_ctx *parser_ctx, const char *str)
193 fprintf(stderr, "error %s\n", str);
202 #define parse_error(parser_ctx, str) \
204 yyerror(parser_ctx, YY_("parse error: " str "\n")); \
208 static void free_strings(struct cds_list_head *list)
210 struct gc_string *gstr, *tmp;
212 cds_list_for_each_entry_safe(gstr, tmp, list, gc)
216 static struct filter_ast *filter_ast_alloc(void)
218 struct filter_ast *ast;
220 ast = malloc(sizeof(*ast));
223 memset(ast, 0, sizeof(*ast));
224 CDS_INIT_LIST_HEAD(&ast->allocated_nodes);
225 ast->root.type = NODE_ROOT;
229 static void filter_ast_free(struct filter_ast *ast)
231 struct filter_node *node, *tmp;
233 cds_list_for_each_entry_safe(node, tmp, &ast->allocated_nodes, gc)
239 int filter_parser_ctx_append_ast(struct filter_parser_ctx *parser_ctx)
241 return yyparse(parser_ctx);
245 struct filter_parser_ctx *filter_parser_ctx_alloc(FILE *input)
247 struct filter_parser_ctx *parser_ctx;
250 yydebug = filter_parser_debug;
252 parser_ctx = malloc(sizeof(*parser_ctx));
255 memset(parser_ctx, 0, sizeof(*parser_ctx));
257 ret = yylex_init_extra(parser_ctx, &parser_ctx->scanner);
259 fprintf(stderr, "yylex_init error\n");
260 goto cleanup_parser_ctx;
262 /* Start processing new stream */
263 yyrestart(input, parser_ctx->scanner);
265 parser_ctx->ast = filter_ast_alloc();
266 if (!parser_ctx->ast)
268 CDS_INIT_LIST_HEAD(&parser_ctx->allocated_strings);
271 fprintf(stdout, "parser_ctx input is a%s.\n",
272 isatty(fileno(input)) ? "n interactive tty" :
273 " noninteractive file");
278 ret = yylex_destroy(parser_ctx->scanner);
280 fprintf(stderr, "yylex_destroy error\n");
287 void filter_parser_ctx_free(struct filter_parser_ctx *parser_ctx)
291 free_strings(&parser_ctx->allocated_strings);
292 filter_ast_free(parser_ctx->ast);
293 ret = yylex_destroy(parser_ctx->scanner);
295 fprintf(stderr, "yylex_destroy error\n");
303 %parse-param {struct filter_parser_ctx *parser_ctx}
304 %lex-param {struct filter_parser_ctx *parser_ctx}
305 %start translation_unit
306 %token CHARACTER_CONSTANT_START SQUOTE STRING_LITERAL_START DQUOTE
307 %token ESCSEQ CHAR_STRING_TOKEN
308 %token DECIMAL_CONSTANT OCTAL_CONSTANT HEXADECIMAL_CONSTANT FLOAT_CONSTANT
309 %token LSBRAC RSBRAC LPAREN RPAREN LBRAC RBRAC RARROW
310 %token STAR PLUS MINUS
311 %token MOD_OP DIV_OP RIGHT_OP LEFT_OP
312 %token EQ_OP NE_OP LE_OP GE_OP LT_OP GT_OP AND_OP OR_OP NOT_OP
313 %token ASSIGN COLON SEMICOLON DOTDOTDOT DOT EQUAL COMMA
314 %token XOR_BIN AND_BIN OR_BIN NOT_BIN
316 %token <gs> IDENTIFIER GLOBAL_IDENTIFIER
322 struct gc_string *gs;
323 struct filter_node *n;
326 %type <gs> s_char s_char_sequence c_char c_char_sequence
328 %type <n> primary_expression
329 %type <n> postfix_expression
330 %type <n> unary_expression
331 %type <n> unary_operator
332 %type <n> multiplicative_expression
333 %type <n> additive_expression
334 %type <n> shift_expression
335 %type <n> relational_expression
336 %type <n> equality_expression
337 %type <n> and_expression
338 %type <n> exclusive_or_expression
339 %type <n> inclusive_or_expression
340 %type <n> logical_and_expression
341 %type <n> logical_or_expression
352 | c_char_sequence c_char
353 { $$ = gc_string_append(parser_ctx, $1, $2); }
361 parse_error(parser_ctx, "escape sequences not supported yet");
365 /* 1.6 String literals */
370 | s_char_sequence s_char
371 { $$ = gc_string_append(parser_ctx, $1, $2); }
379 parse_error(parser_ctx, "escape sequences not supported yet");
386 $$ = make_node(parser_ctx, NODE_EXPRESSION);
387 $$->u.expression.type = AST_EXP_IDENTIFIER;
388 $$->u.expression.u.identifier = yylval.gs->s;
392 $$ = make_node(parser_ctx, NODE_EXPRESSION);
393 $$->u.expression.type = AST_EXP_GLOBAL_IDENTIFIER;
394 $$->u.expression.u.identifier = yylval.gs->s;
399 $$ = make_node(parser_ctx, NODE_EXPRESSION);
400 $$->u.expression.type = AST_EXP_CONSTANT;
401 sscanf(yylval.gs->s, "%" PRIu64,
402 &$$->u.expression.u.constant);
406 $$ = make_node(parser_ctx, NODE_EXPRESSION);
407 $$->u.expression.type = AST_EXP_CONSTANT;
408 sscanf(yylval.gs->s, "0%" PRIo64,
409 &$$->u.expression.u.constant);
411 | HEXADECIMAL_CONSTANT
413 $$ = make_node(parser_ctx, NODE_EXPRESSION);
414 $$->u.expression.type = AST_EXP_CONSTANT;
415 sscanf(yylval.gs->s, "0x%" PRIx64,
416 &$$->u.expression.u.constant);
420 $$ = make_node(parser_ctx, NODE_EXPRESSION);
421 $$->u.expression.type = AST_EXP_FLOAT_CONSTANT;
422 sscanf(yylval.gs->s, "%lg",
423 &$$->u.expression.u.float_constant);
425 | STRING_LITERAL_START DQUOTE
427 $$ = make_node(parser_ctx, NODE_EXPRESSION);
428 $$->u.expression.type = AST_EXP_STRING;
429 $$->u.expression.u.string = "";
431 | STRING_LITERAL_START s_char_sequence DQUOTE
433 $$ = make_node(parser_ctx, NODE_EXPRESSION);
434 $$->u.expression.type = AST_EXP_STRING;
435 $$->u.expression.u.string = $2->s;
437 | CHARACTER_CONSTANT_START c_char_sequence SQUOTE
439 $$ = make_node(parser_ctx, NODE_EXPRESSION);
440 $$->u.expression.type = AST_EXP_STRING;
441 $$->u.expression.u.string = $2->s;
443 | LPAREN expression RPAREN
445 $$ = make_node(parser_ctx, NODE_EXPRESSION);
446 $$->u.expression.type = AST_EXP_NESTED;
447 $$->u.expression.u.child = $2;
454 | postfix_expression DOT IDENTIFIER
456 $$ = make_node(parser_ctx, NODE_EXPRESSION);
457 $$->u.expression.type = AST_EXP_IDENTIFIER;
458 $$->u.expression.post_op = AST_LINK_DOT;
459 $$->u.expression.u.identifier = $3->s;
460 $$->u.expression.prev = $1;
462 | postfix_expression RARROW IDENTIFIER
464 $$ = make_node(parser_ctx, NODE_EXPRESSION);
465 $$->u.expression.type = AST_EXP_IDENTIFIER;
466 $$->u.expression.post_op = AST_LINK_RARROW;
467 $$->u.expression.u.identifier = $3->s;
468 $$->u.expression.prev = $1;
475 | unary_operator unary_expression
478 $$->u.unary_op.child = $2;
485 $$ = make_node(parser_ctx, NODE_UNARY_OP);
486 $$->u.unary_op.type = AST_UNARY_PLUS;
490 $$ = make_node(parser_ctx, NODE_UNARY_OP);
491 $$->u.unary_op.type = AST_UNARY_MINUS;
495 $$ = make_node(parser_ctx, NODE_UNARY_OP);
496 $$->u.unary_op.type = AST_UNARY_NOT;
500 $$ = make_node(parser_ctx, NODE_UNARY_OP);
501 $$->u.unary_op.type = AST_UNARY_BIN_NOT;
505 multiplicative_expression
508 | multiplicative_expression STAR unary_expression
510 $$ = make_op_node(parser_ctx, AST_OP_MUL, $1, $3);
512 | multiplicative_expression DIV_OP unary_expression
514 $$ = make_op_node(parser_ctx, AST_OP_DIV, $1, $3);
516 | multiplicative_expression MOD_OP unary_expression
518 $$ = make_op_node(parser_ctx, AST_OP_MOD, $1, $3);
523 : multiplicative_expression
525 | additive_expression PLUS multiplicative_expression
527 $$ = make_op_node(parser_ctx, AST_OP_PLUS, $1, $3);
529 | additive_expression MINUS multiplicative_expression
531 $$ = make_op_node(parser_ctx, AST_OP_MINUS, $1, $3);
536 : additive_expression
538 | shift_expression LEFT_OP additive_expression
540 $$ = make_op_node(parser_ctx, AST_OP_LSHIFT, $1, $3);
542 | shift_expression RIGHT_OP additive_expression
544 $$ = make_op_node(parser_ctx, AST_OP_RSHIFT, $1, $3);
548 relational_expression
551 | relational_expression LT_OP shift_expression
553 $$ = make_op_node(parser_ctx, AST_OP_LT, $1, $3);
555 | relational_expression GT_OP shift_expression
557 $$ = make_op_node(parser_ctx, AST_OP_GT, $1, $3);
559 | relational_expression LE_OP shift_expression
561 $$ = make_op_node(parser_ctx, AST_OP_LE, $1, $3);
563 | relational_expression GE_OP shift_expression
565 $$ = make_op_node(parser_ctx, AST_OP_GE, $1, $3);
570 : relational_expression
572 | equality_expression EQ_OP relational_expression
574 $$ = make_op_node(parser_ctx, AST_OP_EQ, $1, $3);
576 | equality_expression NE_OP relational_expression
578 $$ = make_op_node(parser_ctx, AST_OP_NE, $1, $3);
583 : equality_expression
585 | and_expression AND_BIN equality_expression
587 $$ = make_op_node(parser_ctx, AST_OP_BIN_AND, $1, $3);
591 exclusive_or_expression
594 | exclusive_or_expression XOR_BIN and_expression
596 $$ = make_op_node(parser_ctx, AST_OP_BIN_XOR, $1, $3);
600 inclusive_or_expression
601 : exclusive_or_expression
603 | inclusive_or_expression OR_BIN exclusive_or_expression
605 $$ = make_op_node(parser_ctx, AST_OP_BIN_OR, $1, $3);
609 logical_and_expression
610 : inclusive_or_expression
612 | logical_and_expression AND_OP inclusive_or_expression
614 $$ = make_op_node(parser_ctx, AST_OP_AND, $1, $3);
618 logical_or_expression
619 : logical_and_expression
621 | logical_or_expression OR_OP logical_and_expression
623 $$ = make_op_node(parser_ctx, AST_OP_OR, $1, $3);
628 : logical_or_expression
635 parser_ctx->ast->root.u.root.child = $1;