X-Git-Url: https://git.exim.org/exim.git/blobdiff_plain/9f6b3bf5187562bac4c96e3ed6a17740d01489fa..44b3172e369435c2c1baa4e9c837252f729d2905:/src/src/regex.c diff --git a/src/src/regex.c b/src/src/regex.c index 5c0f7c4e0..285ab8dbd 100644 --- a/src/src/regex.c +++ b/src/src/regex.c @@ -3,9 +3,10 @@ *************************************************/ /* - * Copyright (c) The Exim Maintainers 2016 - 2022 + * Copyright (c) The Exim Maintainers 2016 - 2023 * Copyright (c) Tom Kistner 2003-2015 * License: GPL + * SPDX-License-Identifier: GPL-2.0-or-later */ /* Code for matching regular expressions against headers and body. @@ -18,41 +19,34 @@ /* Structure to hold a list of Regular expressions */ typedef struct pcre_list { - pcre2_code *re; - uschar *pcre_text; - struct pcre_list *next; + const pcre2_code * re; + uschar * pcre_text; + struct pcre_list * next; } pcre_list; -uschar regex_match_string_buffer[1024]; - extern FILE *mime_stream; extern uschar *mime_current_boundary; + static pcre_list * -compile(const uschar * list) +compile(const uschar * list, BOOL cacheable, int * cntp) { -int sep = 0; -uschar *regex_string; -pcre_list *re_list_head = NULL; -pcre_list *ri; +int sep = 0, cnt = 0; +uschar * regex_string; +pcre_list * re_list_head = NULL, * ri; /* precompile our regexes */ while ((regex_string = string_nextinlist(&list, &sep, NULL, 0))) if (strcmpic(regex_string, US"false") != 0 && Ustrcmp(regex_string, "0") != 0) { - pcre2_code * re; - int err; - PCRE2_SIZE pcre_erroffset; - /* compile our regular expression */ - if (!(re = pcre2_compile( (PCRE2_SPTR) regex_string, PCRE2_ZERO_TERMINATED, - 0, &err, &pcre_erroffset, pcre_cmp_ctx))) + uschar * errstr; + const pcre2_code * re = regex_compile(regex_string, + cacheable ? MCS_CACHEABLE : MCS_NOFLAGS, &errstr, pcre_gen_cmp_ctx); + + if (!re) { - uschar errbuf[128]; - pcre2_get_error_message(err, errbuf, sizeof(errbuf)); - log_write(0, LOG_MAIN, - "regex acl condition warning - error in regex '%s': %s at offset %ld, skipped.", - regex_string, errbuf, (long)pcre_erroffset); + log_write(0, LOG_MAIN, "regex acl condition warning - %s, skipped", errstr); continue; } @@ -61,10 +55,19 @@ while ((regex_string = string_nextinlist(&list, &sep, NULL, 0))) ri->pcre_text = regex_string; ri->next = re_list_head; re_list_head = ri; + cnt++; } +if (cntp) *cntp = cnt; return re_list_head; } + +/* Check list of REs against buffer, returning OK for (first) match, +else FAIL. On match return allocated result strings in regex_vars[]. + +We use the perm-pool for that, so that our caller can release +other allocations. +*/ static int matcher(pcre_list * re_list_head, uschar * linebuffer, int len) { @@ -75,39 +78,50 @@ for (pcre_list * ri = re_list_head; ri; ri = ri->next) int n; /* try matcher on the line */ - if ((n = pcre2_match(ri->re, (PCRE2_SPTR)linebuffer, len, 0, 0, md, pcre_mtc_ctx)) > 0) + if ((n = pcre2_match(ri->re, (PCRE2_SPTR)linebuffer, len, 0, 0, md, pcre_gen_mtc_ctx)) > 0) { - Ustrncpy(regex_match_string_buffer, ri->pcre_text, - sizeof(regex_match_string_buffer)-1); - regex_match_string = regex_match_string_buffer; + int save_pool = store_pool; + store_pool = POOL_PERM; + + regex_match_string = string_copy(ri->pcre_text); for (int nn = 1; nn < n; nn++) { - PCRE2_UCHAR * cstr; - PCRE2_SIZE cslen; - pcre2_substring_get_bynumber(md, nn, &cstr, &cslen); - regex_vars[nn-1] = CUS cstr; + PCRE2_SIZE * ovec = pcre2_get_ovector_pointer(md); + int off = nn * 2; + int len = ovec[off + 1] - ovec[off]; + regex_vars[nn-1] = string_copyn(linebuffer + ovec[off], len); } + store_pool = save_pool; return OK; } } -pcre2_match_data_free(md); +/* pcre2_match_data_free(md); gen ctx needs no free */ return FAIL; } + +/* reset expansion variables */ +void +regex_vars_clear(void) +{ +regex_match_string = NULL; +for (int i = 0; i < REGEX_VARS; i++) regex_vars[i] = NULL; +} + + + int -regex(const uschar **listptr) +regex(const uschar ** listptr, BOOL cacheable) { unsigned long mbox_size; -FILE *mbox_file; -pcre_list *re_list_head; -uschar *linebuffer; +FILE * mbox_file; +pcre_list * re_list_head; long f_pos = 0; -int ret = FAIL; +int ret = FAIL, cnt, lcount = REGEX_LOOPCOUNT_STORE_RESET; -/* reset expansion variable */ -regex_match_string = NULL; +regex_vars_clear(); if (!mime_stream) /* We are in the DATA ACL */ { @@ -129,26 +143,33 @@ else mbox_file = mime_stream; } -/* precompile our regexes */ -if (!(re_list_head = compile(*listptr))) - return FAIL; /* no regexes -> nothing to do */ + /* precompile our regexes */ + if ((re_list_head = compile(*listptr, cacheable, &cnt))) + { + rmark reset_point = store_mark(); -/* match each line against all regexes */ -linebuffer = store_get(32767, GET_TAINTED); -while (fgets(CS linebuffer, 32767, mbox_file)) - { - if ( mime_stream && mime_current_boundary /* check boundary */ - && Ustrncmp(linebuffer, "--", 2) == 0 - && Ustrncmp((linebuffer+2), mime_current_boundary, - Ustrlen(mime_current_boundary)) == 0) - break; /* found boundary */ - - if ((ret = matcher(re_list_head, linebuffer, (int)Ustrlen(linebuffer))) == OK) - goto done; - } -/* no matches ... */ + /* match each line against all regexes */ + while (fgets(CS big_buffer, big_buffer_size, mbox_file)) + { + if ( mime_stream && mime_current_boundary /* check boundary */ + && Ustrncmp(big_buffer, "--", 2) == 0 + && Ustrncmp((big_buffer+2), mime_current_boundary, + Ustrlen(mime_current_boundary)) == 0) + break; /* found boundary */ + + if ((ret = matcher(re_list_head, big_buffer, (int)Ustrlen(big_buffer))) == OK) + break; + + if ((lcount -= cnt) <= 0) + { + store_reset(reset_point); reset_point = store_mark(); + lcount = REGEX_LOOPCOUNT_STORE_RESET; + } + } + + store_reset(reset_point); + } -done: if (!mime_stream) (void)fclose(mbox_file); else @@ -167,20 +188,16 @@ return ret; int -mime_regex(const uschar **listptr) +mime_regex(const uschar **listptr, BOOL cacheable) { -pcre_list *re_list_head = NULL; -FILE *f; -uschar *mime_subject = NULL; +pcre_list * re_list_head = NULL; +FILE * f; +uschar * mime_subject = NULL; int mime_subject_len = 0; -int ret; - -/* reset expansion variable */ -regex_match_string = NULL; +int ret = FAIL; +rmark reset_point; -/* precompile our regexes */ -if (!(re_list_head = compile(*listptr))) - return FAIL; /* no regexes -> nothing to do */ +regex_vars_clear(); /* check if the file is already decoded */ if (!mime_decoded_filename) @@ -204,12 +221,20 @@ if (!(f = fopen(CS mime_decoded_filename, "rb"))) return DEFER; } -/* get 32k memory, tainted */ -mime_subject = store_get(32767, GET_TAINTED); +reset_point = store_mark(); + { + /* precompile our regexes */ + if ((re_list_head = compile(*listptr, cacheable, NULL))) + { + /* get 32k memory, tainted */ + mime_subject = store_get(32767, GET_TAINTED); -mime_subject_len = fread(mime_subject, 1, 32766, f); + mime_subject_len = fread(mime_subject, 1, 32766, f); -ret = matcher(re_list_head, mime_subject, mime_subject_len); + ret = matcher(re_list_head, mime_subject, mime_subject_len); + } + } +store_reset(reset_point); (void)fclose(f); return ret; }