clamav/unit_tests/check_regex.c

685 lines
23 KiB
C
Raw Normal View History

/*
* Unit tests for regular expression processing.
*
2020-01-03 15:44:07 -05:00
* Copyright (C) 2013-2020 Cisco Systems, Inc. and/or its affiliates. All rights reserved.
* Copyright (C) 2008-2013 Sourcefire, Inc.
*
* Authors: Török Edvin
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2 as
* published by the Free Software Foundation.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
* MA 02110-1301, USA.
*/
#if HAVE_CONFIG_H
#include "clamav-config.h"
#endif
#include <stdio.h>
#include <stdlib.h>
#include <limits.h>
#include <string.h>
#include <check.h>
2014-02-08 00:31:12 -05:00
#include "../libclamav/clamav.h"
#include "../libclamav/others.h"
#include "../libclamav/mbox.h"
#include "../libclamav/message.h"
#include "../libclamav/htmlnorm.h"
#include "../libclamav/phishcheck.h"
#include "../libclamav/regex_suffix.h"
#include "../libclamav/regex_list.h"
#include "../libclamav/phish_domaincheck_db.h"
#include "../libclamav/phish_whitelist.h"
#include "checks.h"
static size_t cb_called = 0;
static cl_error_t cb_fail(void *cbdata, const char *suffix, size_t len, const struct regex_list *regex)
{
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
UNUSEDPARAM(cbdata);
UNUSEDPARAM(suffix);
UNUSEDPARAM(len);
UNUSEDPARAM(regex);
ck_abort_msg("this pattern is not supposed to have a suffix");
return CL_EMEM;
}
static cl_error_t cb_expect_single(void *cbdata, const char *suffix, size_t len, const struct regex_list *regex)
{
const char *expected = cbdata;
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
UNUSEDPARAM(len);
UNUSEDPARAM(regex);
cb_called++;
ck_assert_msg(suffix && strcmp(suffix, expected) == 0,
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
"suffix mismatch, was: %s, expected: %s\n", suffix, expected);
return CL_SUCCESS;
}
static struct regex_list regex;
START_TEST(empty)
{
const char pattern[] = "";
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
cl_error_t rc;
regex_t *preg;
errmsg_expected();
preg = malloc(sizeof(*regex.preg));
ck_assert_msg(!!preg, "malloc");
rc = cli_regex2suffix(pattern, preg, cb_fail, NULL);
free(preg);
ck_assert_msg(rc == REG_EMPTY, "empty pattern");
ck_assert_msg(cb_called == 0, "callback shouldn't be called");
}
END_TEST
START_TEST(one)
{
char pattern[] = "a";
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
cl_error_t rc;
regex_t *preg;
preg = malloc(sizeof(*regex.preg));
ck_assert_msg(!!preg, "malloc");
rc = cli_regex2suffix(pattern, preg, cb_expect_single, pattern);
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
ck_assert_msg(rc == CL_SUCCESS, "single character pattern");
cli_regfree(preg);
free(preg);
ck_assert_msg(cb_called == 1, "callback should be called once");
}
END_TEST
static const char *ex1[] =
{"com|de", "moc", "ed", NULL};
static const char *ex2[] =
{"xd|(a|e)bc", "dx", "cba", "cbe", NULL};
static const char **tests[] = {
ex1,
ex2};
static cl_error_t cb_expect_multi(void *cbdata, const char *suffix, size_t len, const struct regex_list *r)
{
const char **exp = cbdata;
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
UNUSEDPARAM(r);
ck_assert_msg(!!exp, "expected data");
exp++;
ck_assert_msg(!!*exp, "expected no suffix, got: %s\n", suffix);
ck_assert_msg(!!exp[cb_called], "expected less suffixes, but already got: %d\n", cb_called);
ck_assert_msg(strcmp(exp[cb_called], suffix) == 0,
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
"suffix mismatch, was: %s, expected: %s\n", suffix, exp[cb_called]);
ck_assert_msg(strlen(suffix) == len, "incorrect suffix len, expected: %d, got: %d\n", strlen(suffix), len);
cb_called++;
return CL_SUCCESS;
}
START_TEST(test_suffix)
{
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
cl_error_t rc;
regex_t *preg;
const char *pattern = tests[_i][0];
size_t n = 0;
const char **p = tests[_i];
ck_assert_msg(!!pattern, "test pattern");
preg = malloc(sizeof(*regex.preg));
ck_assert_msg(!!preg, "malloc");
rc = cli_regex2suffix(pattern, preg, cb_expect_multi, tests[_i]);
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
ck_assert_msg(rc == CL_SUCCESS, "single character pattern");
cli_regfree(preg);
free(preg);
p++;
while (*p++) n++;
ck_assert_msg(cb_called == n,
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
"suffix number mismatch, expected: %d, was: %d\n", n, cb_called);
}
END_TEST
static void setup(void)
{
cb_called = 0;
}
static void teardown(void)
{
}
static struct regex_matcher matcher;
static void rsetup(void)
{
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
cl_error_t rc;
#ifdef USE_MPOOL
matcher.mempool = mpool_create();
#endif
rc = init_regex_list(&matcher, 1);
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
ck_assert_msg(rc == CL_SUCCESS, "init_regex_list");
}
static void rteardown(void)
{
regex_list_done(&matcher);
#ifdef USE_MPOOL
mpool_destroy(matcher.mempool);
#endif
}
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
typedef enum rtest_result {
RTR_PHISH,
RTR_WHITELISTED,
RTR_CLEAN,
RTR_BLACKLISTED, // if 2nd db is loaded
RTR_INVALID_REGEX
} rtr_t;
static const struct rtest {
const char *pattern; /* NULL if not meant for whitelist testing */
const char *realurl;
const char *displayurl;
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
rtr_t result;
} rtests[] = {
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
{NULL, "http://fake.example.com", "http://foo@key.com/", RTR_PHISH},
{NULL, "http://fake.example.com", "foo.example.com@key.com", RTR_PHISH},
{NULL, "http://fake.example.com", "foo@key.com", RTR_CLEAN},
{NULL, "http://fake.example.com", "&#61;&#61;&#61;&#61;&#61;key.com", RTR_PHISH},
{NULL, "http://key.com", "&#61;&#61;&#61;&#61;&#61;key.com", RTR_CLEAN},
{NULL, " http://key.com", "&#61;&#61;&#61;&#61;&#61;key.com", RTR_CLEAN},
{NULL, "http://key.com@fake.example.com", "key.com", RTR_PHISH},
{NULL, " http://key.com@fake.example.com", "key.com", RTR_PHISH},
{NULL, " http://key.com@fake.example.com ", "key.com", RTR_PHISH},
/* entry taken from .wdb with a / appended */
{".+\\.ebayrtm\\.com([/?].*)?:.+\\.ebay\\.(de|com|co\\.uk)([/?].*)?/",
"http://srx.main.ebayrtm.com",
"pages.ebay.de",
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
RTR_WHITELISTED /* should be whitelisted */},
{".+\\.ebayrtm\\.com([/?].*)?:.+\\.ebay\\.(de|com|co\\.uk)([/?].*)?/",
"http://srx.main.ebayrtm.com.evil.example.com",
"pages.ebay.de",
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
RTR_PHISH},
{".+\\.ebayrtm\\.com([/?].*)?:.+\\.ebay\\.(de|com|co\\.uk)([/?].*)?/",
"www.www.ebayrtm.com?somecgi",
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
"www.ebay.com/something", RTR_WHITELISTED},
{NULL,
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
"http://key.com", "go to key.com", RTR_CLEAN},
{":.+\\.paypal\\.(com|de|fr|it)([/?].*)?:.+\\.ebay\\.(at|be|ca|ch|co\\.uk|de|es|fr|ie|in|it|nl|ph|pl|com(\\.(au|cn|hk|my|sg))?)([/?].*)?/",
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
"http://www.paypal.com", "pics.ebay.com", RTR_WHITELISTED},
{NULL, "http://somefakeurl.example.com", "someotherdomain-key.com", RTR_CLEAN},
{NULL, "http://somefakeurl.example.com", "someotherdomain.key.com", RTR_PHISH},
{NULL, "http://1.test.example.com/something", "test", RTR_BLACKLISTED},
{NULL, "http://1.test.example.com/2", "test", RTR_BLACKLISTED},
{NULL, "http://user@1.test.example.com/2", "test", RTR_BLACKLISTED},
{NULL, "http://user@1.test.example.com/2/test", "test", RTR_BLACKLISTED},
{NULL, "http://user@1.test.example.com/", "test", RTR_BLACKLISTED},
{NULL, "http://x.exe", "http:///x.exe", RTR_CLEAN},
{".+\\.ebayrtm\\.com([/?].*)?:[^.]+\\.ebay\\.(de|com|co\\.uk)/",
"http://srx.main.ebayrtm.com",
"pages.ebay.de",
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
RTR_WHITELISTED /* should be whitelisted */},
{".+\\.ebayrtm\\.com([/?].*)?:.+[r-t]\\.ebay\\.(de|com|co\\.uk)/",
"http://srx.main.ebayrtm.com",
"pages.ebay.de",
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
RTR_WHITELISTED /* should be whitelisted */},
{".+\\.ebayrtm\\.com([/?].*)?:.+[r-t]\\.ebay\\.(de|com|co\\.uk)/",
"http://srx.main.ebayrtm.com",
"pages.ebay.de",
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
RTR_WHITELISTED /* should be whitelisted */},
{"[t-", "", "", RTR_INVALID_REGEX},
{NULL, "http://co.uk", "http:// co.uk", RTR_CLEAN},
{NULL, "http://co.uk", " ", RTR_CLEAN},
{NULL, "127.0.0.1", "pages.ebay.de", RTR_CLEAN},
{".+\\.ebayrtm\\.com([/?].*)?:.+\\.ebay\\.(de|com|co\\.uk)([/?].*)?/",
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
"http://pages.ebay.de@fake.example.com", "pages.ebay.de", RTR_PHISH},
{NULL, "http://key.com", "https://key.com", RTR_PHISH},
{NULL, "http://key.com%00fake.example.com", "https://key.com", RTR_PHISH},
{NULL, "http://key.com.example.com", "key.com.invalid", RTR_PHISH}};
START_TEST(regex_list_match_test)
{
const char *info;
const struct rtest *rtest = &rtests[_i];
char *pattern, *realurl;
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
cl_error_t rc;
if (!rtest->pattern) {
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
ck_assert_msg(rtest->result != RTR_WHITELISTED,
"whitelist test must have pattern set");
/* this test entry is not meant for whitelist testing */
return;
}
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
ck_assert_msg(rtest->result == RTR_PHISH || rtest->result == RTR_WHITELISTED || rtest->result == RTR_INVALID_REGEX,
"whitelist test result must be either RTR_PHISH or RTR_WHITELISTED or RTR_INVALID_REGEX");
pattern = cli_strdup(rtest->pattern);
ck_assert_msg(!!pattern, "cli_strdup");
rc = regex_list_add_pattern(&matcher, pattern);
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
if (rtest->result == RTR_INVALID_REGEX) {
ck_assert_msg(rc, "regex_list_add_pattern should return error");
free(pattern);
return;
} else
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
ck_assert_msg(rc == CL_SUCCESS, "regex_list_add_pattern");
free(pattern);
matcher.list_loaded = 1;
rc = cli_build_regex_list(&matcher);
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
ck_assert_msg(rc == CL_SUCCESS, "cli_build_regex_list");
ck_assert_msg(is_regex_ok(&matcher), "is_regex_ok");
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
realurl = cli_strdup(rtest->realurl);
rc = regex_list_match(&matcher, realurl, rtest->displayurl, NULL, 1, &info, 1);
ck_assert_msg(rc == rtest->result, "regex_list_match");
/* regex_list_match is not supposed to modify realurl in this case */
ck_assert_msg(!strcmp(realurl, rtest->realurl), "realurl altered");
free(realurl);
}
END_TEST
static struct cl_engine *engine;
static int loaded_2 = 0;
static void psetup_impl(int load2)
{
FILE *f;
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
cl_error_t rc;
unsigned signo = 0;
engine = cl_engine_new();
ck_assert_msg(!!engine, "cl_engine_new");
phishing_init(engine);
ck_assert_msg(!!engine->phishcheck, "phishing_init");
rc = init_domainlist(engine);
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
ck_assert_msg(rc == CL_SUCCESS, "init_domainlist");
f = fdopen(open_testfile("input/daily.pdb"), "r");
ck_assert_msg(!!f, "fopen daily.pdb");
rc = load_regex_matcher(engine, engine->domainlist_matcher, f, &signo, 0, 0, NULL, 1);
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
ck_assert_msg(rc == CL_SUCCESS, "load_regex_matcher");
fclose(f);
ck_assert_msg(signo == 201, "Incorrect number of signatures: %u, expected %u", signo, 201);
if (load2) {
f = fdopen(open_testfile("input/daily.gdb"), "r");
ck_assert_msg(!!f, "fopen daily.gdb");
signo = 0;
rc = load_regex_matcher(engine, engine->domainlist_matcher, f, &signo, 0, 0, NULL, 1);
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
ck_assert_msg(rc == CL_SUCCESS, "load_regex_matcher");
fclose(f);
ck_assert_msg(signo == 4, "Incorrect number of signatures: %u, expected %u", signo, 4);
}
loaded_2 = load2;
rc = init_whitelist(engine);
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
ck_assert_msg(rc == CL_SUCCESS, "init_whitelist");
f = fdopen(open_testfile("input/daily.wdb"), "r");
signo = 0;
rc = load_regex_matcher(engine, engine->whitelist_matcher, f, &signo, 0, 1, NULL, 1);
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
ck_assert_msg(rc == CL_SUCCESS, "load_regex_matcher");
fclose(f);
ck_assert_msg(signo == 31, "Incorrect number of signatures: %u, expected %u", signo, 31);
rc = cli_build_regex_list(engine->whitelist_matcher);
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
ck_assert_msg(rc == CL_SUCCESS, "cli_build_regex_list");
rc = cli_build_regex_list(engine->domainlist_matcher);
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
ck_assert_msg(rc == CL_SUCCESS, "cli_build_regex_list");
ck_assert_msg(is_regex_ok(engine->whitelist_matcher), "is_regex_ok");
ck_assert_msg(is_regex_ok(engine->domainlist_matcher), "is_regex_ok");
}
static void psetup(void)
{
psetup_impl(0);
}
static void psetup2(void)
{
psetup_impl(1);
}
static void pteardown(void)
{
if (engine) {
cl_engine_free(engine);
}
engine = NULL;
}
static void do_phishing_test(const struct rtest *rtest)
{
char *realurl;
cli_ctx ctx;
struct cl_scan_options options;
const char *virname = NULL;
tag_arguments_t hrefs;
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
cl_error_t rc;
memset(&ctx, 0, sizeof(ctx));
memset(&options, 0, sizeof(struct cl_scan_options));
ctx.options = &options;
realurl = cli_strdup(rtest->realurl);
ck_assert_msg(!!realurl, "cli_strdup");
hrefs.count = 1;
hrefs.value = cli_malloc(sizeof(*hrefs.value));
ck_assert_msg(!!hrefs.value, "cli_malloc");
hrefs.value[0] = (unsigned char *)realurl;
hrefs.contents = cli_malloc(sizeof(*hrefs.contents));
ck_assert_msg(!!hrefs.contents, "cli_malloc");
hrefs.tag = cli_malloc(sizeof(*hrefs.tag));
ck_assert_msg(!!hrefs.tag, "cli_malloc");
hrefs.tag[0] = (unsigned char *)cli_strdup("href");
hrefs.contents[0] = (unsigned char *)cli_strdup(rtest->displayurl);
ctx.engine = engine;
ctx.virname = &virname;
rc = phishingScan(&ctx, &hrefs);
html_tag_arg_free(&hrefs);
ck_assert_msg(rc == CL_CLEAN, "phishingScan");
switch (rtest->result) {
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
case RTR_PHISH:
ck_assert_msg(ctx.found_possibly_unwanted,
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
"this should be phishing, realURL: %s, displayURL: %s",
rtest->realurl, rtest->displayurl);
break;
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
case RTR_WHITELISTED:
ck_assert_msg(!ctx.found_possibly_unwanted,
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
"this should be whitelisted, realURL: %s, displayURL: %s",
rtest->realurl, rtest->displayurl);
break;
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
case RTR_CLEAN:
ck_assert_msg(!ctx.found_possibly_unwanted,
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
"this should be clean, realURL: %s, displayURL: %s",
rtest->realurl, rtest->displayurl);
break;
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
case RTR_BLACKLISTED:
if (!loaded_2)
ck_assert_msg(!ctx.found_possibly_unwanted,
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
"this should be clean, realURL: %s, displayURL: %s",
rtest->realurl, rtest->displayurl);
else {
ck_assert_msg(ctx.found_possibly_unwanted,
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
"this should be blacklisted, realURL: %s, displayURL: %s",
rtest->realurl, rtest->displayurl);
if (*ctx.virname) {
char *phishingFound = strstr((const char *)*ctx.virname, "Heuristics.Safebrowsing.Suspected-malware_safebrowsing.clamav.net");
ck_assert_msg(phishingFound != NULL, "\n\t should be: Heuristics.Safebrowsing.Suspected-malware_safebrowsing.clamav.net,\n\t but is: %s\n", *ctx.virname);
}
}
break;
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
case RTR_INVALID_REGEX:
/* don't worry about it, this was tested in regex_list_match_test() */
break;
}
}
static void do_phishing_test_allscan(const struct rtest *rtest)
{
char *realurl;
cli_ctx ctx;
const char *virname = NULL;
tag_arguments_t hrefs;
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
cl_error_t rc;
struct cl_scan_options options;
memset(&ctx, 0, sizeof(ctx));
memset(&options, 0, sizeof(struct cl_scan_options));
ctx.options = &options;
realurl = cli_strdup(rtest->realurl);
ck_assert_msg(!!realurl, "cli_strdup");
hrefs.count = 1;
hrefs.value = cli_malloc(sizeof(*hrefs.value));
ck_assert_msg(!!hrefs.value, "cli_malloc");
hrefs.value[0] = (unsigned char *)realurl;
hrefs.contents = cli_malloc(sizeof(*hrefs.contents));
ck_assert_msg(!!hrefs.contents, "cli_malloc");
hrefs.tag = cli_malloc(sizeof(*hrefs.tag));
ck_assert_msg(!!hrefs.tag, "cli_malloc");
hrefs.tag[0] = (unsigned char *)cli_strdup("href");
hrefs.contents[0] = (unsigned char *)cli_strdup(rtest->displayurl);
ctx.engine = engine;
ctx.virname = &virname;
ctx.options->general |= CL_SCAN_GENERAL_ALLMATCHES;
rc = phishingScan(&ctx, &hrefs);
html_tag_arg_free(&hrefs);
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
if (rtest->result == RTR_PHISH || (loaded_2 != 0 && rtest->result == RTR_BLACKLISTED)) {
ck_assert_msg(rc == CL_VIRUS, "phishingScan returned \"%s\", expected \"%s\". \n\trealURL: %s \n\tdisplayURL: %s",
cl_strerror(rc),
cl_strerror(CL_VIRUS),
rtest->realurl, rtest->displayurl);
} else {
ck_assert_msg(rc == CL_CLEAN, "phishingScan returned \"%s\", expected \"%s\". \n\trealURL: %s \n\tdisplayURL: %s",
cl_strerror(rc),
cl_strerror(CL_CLEAN),
rtest->realurl, rtest->displayurl);
}
switch (rtest->result) {
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
case RTR_PHISH:
ck_assert_msg(ctx.num_viruses,
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
"this should be phishing, realURL: %s, displayURL: %s",
rtest->realurl, rtest->displayurl);
break;
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
case RTR_WHITELISTED:
ck_assert_msg(!ctx.num_viruses,
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
"this should be whitelisted, realURL: %s, displayURL: %s",
rtest->realurl, rtest->displayurl);
break;
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
case RTR_CLEAN:
ck_assert_msg(!ctx.num_viruses,
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
"this should be clean, realURL: %s, displayURL: %s",
rtest->realurl, rtest->displayurl);
break;
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
case RTR_BLACKLISTED:
if (!loaded_2)
ck_assert_msg(!ctx.num_viruses,
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
"this should be clean, realURL: %s, displayURL: %s",
rtest->realurl, rtest->displayurl);
else {
ck_assert_msg(ctx.num_viruses,
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
"this should be blacklisted, realURL: %s, displayURL: %s",
rtest->realurl, rtest->displayurl);
if (*ctx.virname) {
char *phishingFound = strstr((const char *)*ctx.virname, "Heuristics.Safebrowsing.Suspected-malware_safebrowsing.clamav.net");
ck_assert_msg(phishingFound != NULL, "\n\t should be: Heuristics.Safebrowsing.Suspected-malware_safebrowsing.clamav.net,\n\t but is: %s\n", *ctx.virname);
}
}
break;
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
case RTR_INVALID_REGEX:
/* don't worry about it, this was tested in regex_list_match_test() */
break;
}
}
START_TEST(phishingScan_test)
{
do_phishing_test(&rtests[_i]);
}
END_TEST
START_TEST(phishingScan_test_allscan)
{
do_phishing_test_allscan(&rtests[_i]);
}
END_TEST
static struct uc {
const char *in;
const char *host;
const char *path;
} uc[] = {
{":example/%25%32%35", "example/", "%25"},
{":example/%25%32%35%25%32%35", "example/", "%25%25"},
{":example/abc%25%32%35asd", "example/", "abc%25asd"},
{":www.example.com/", "www.example.com/", ""},
{":%31%32%37%2e%30%2e%30%2e%31/%2E%73%65%63%75%72%65/%77%77%77%2e%65%78%61%6d%70%6c%65%2e%63%6f%6d/",
"127.0.0.1/", ".secure/www.example.com/"},
{":127.0.0.1/uploads/%20%20%20%20/.verify/.blah=abcd-ef=gh/",
"127.0.0.1/", "uploads/%20%20%20%20/.verify/.blah=abcd-ef=gh/"},
{"http://example%23.com/%61%40%62%252B",
"example%23.com/", "a@b+"},
{"http://example.com/blah/..", "example.com/", ""},
{"http://example.com/blah/../x", "example.com/", "x"},
{"http://example.com/./a", "example.com/", "a"}};
START_TEST(test_url_canon)
{
char urlbuff[1024 + 3];
char *host = NULL;
2009-02-19 08:50:04 +00:00
const char *path = NULL;
size_t host_len, path_len;
struct uc *u = &uc[_i];
cli_url_canon(u->in, strlen(u->in), urlbuff, sizeof(urlbuff), &host, &host_len, &path, &path_len);
ck_assert_msg(!!host && !!path, "null results\n");
ck_assert_msg(!strcmp(u->host, host), "host incorrect: %s\n", host);
ck_assert_msg(!strcmp(u->path, path), "path incorrect: %s\n", path);
}
END_TEST
static struct regex_test {
const char *regex;
const char *text;
int match;
} rg[] = {
{"\\.exe$", "test.exe", 1},
{"\\.exe$", "test.eXe", 0},
{"(?i)\\.exe$", "test.exe", 1},
{"(?i)\\.exe$", "test.eXe", 1}};
START_TEST(test_regexes)
{
regex_t reg;
struct regex_test *tst = &rg[_i];
int match;
ck_assert_msg(cli_regcomp(&reg, tst->regex, REG_EXTENDED | REG_NOSUB) == 0, "cli_regcomp");
match = (cli_regexec(&reg, tst->text, 0, NULL, 0) == REG_NOMATCH) ? 0 : 1;
ck_assert_msg(match == tst->match, "cli_regexec failed for %s and %s\n", tst->regex, tst->text);
cli_regfree(&reg);
}
END_TEST
START_TEST(phishing_fake_test)
{
char buf[4096];
FILE *f = fdopen(open_testfile("input/daily.pdb"), "r");
ck_assert_msg(!!f, "fopen daily.pdb");
while (fgets(buf, sizeof(buf), f)) {
struct rtest rtest;
const char *pdb = strchr(buf, ':');
ck_assert_msg(!!pdb, "missing : in pdb");
rtest.realurl = pdb;
rtest.displayurl = pdb;
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
rtest.result = RTR_CLEAN;
do_phishing_test(&rtest);
rtest.realurl = "http://fake.example.com";
rtest.result = 0;
do_phishing_test(&rtest);
}
fclose(f);
}
END_TEST
START_TEST(phishing_fake_test_allscan)
{
char buf[4096];
FILE *f = fdopen(open_testfile("input/daily.pdb"), "r");
ck_assert_msg(!!f, "fopen daily.pdb");
while (fgets(buf, sizeof(buf), f)) {
struct rtest rtest;
const char *pdb = strchr(buf, ':');
ck_assert_msg(!!pdb, "missing : in pdb");
rtest.realurl = pdb;
rtest.displayurl = pdb;
bb12506: Fix phishing/heuristic alert verbosity Some detections, like phishing, are considered heuristic alerts because they match based on behavior more than on content. A subset of these are considered "potentially unwanted" (low-severity). These low-severity alerts include: - phishing - PDFs with obfuscated object names - bytecode signature alerts that start with "BC.Heuristics" The concept is that unless you enable "heuristic precedence" (a method of lowing the threshold to immediateley alert on low-severity detections), the scan should continue after a match in case a higher severity match is found. Only at the end will it print the low-severity match if nothing else was found. The current implementation is buggy though. Scanning of archives does not correctly bail out for the entire archive if one email contains a phishing link. Instead, it sets the "heuristic found" flag then and alerts for every subsequent file in the archive because it doesn't know if the heuristic was found in an embedded file or the target file. Because it's just a heuristic and the status is "clean", it keeps scanning. This patch corrects the behavior by checking if a low-severity alerts were found at the end of scanning the target file, instead of at the end of each embedded file. Additionally, this patch fixes an in issue with phishing alerts wherein heuristic precedence mode did not cause a scan to stop after the first alert. The above changes required restructuring to create an fmap inside of cl_scandesc_callback() so that scan_common() could be modified to require an fmap and set up so that the current *ctx->fmap pointer is never NULL when scan_common() evaluates match results. Also fixed a couple minor bugs in the phishing unit tests and cleaned up the test code for improved legitibility and type safety.
2020-04-18 10:46:57 -04:00
rtest.result = RTR_CLEAN;
do_phishing_test_allscan(&rtest);
rtest.realurl = "http://fake.example.com";
rtest.result = 0;
do_phishing_test_allscan(&rtest);
}
fclose(f);
}
END_TEST
Suite *test_regex_suite(void)
{
Suite *s = suite_create("regex");
TCase *tc_api, *tc_matching, *tc_phish, *tc_phish2, *tc_regex;
tc_api = tcase_create("cli_regex2suffix");
suite_add_tcase(s, tc_api);
tcase_add_checked_fixture(tc_api, setup, teardown);
tcase_add_test(tc_api, empty);
tcase_add_test(tc_api, one);
tcase_add_loop_test(tc_api, test_suffix, 0, sizeof(tests) / sizeof(tests[0]));
tc_matching = tcase_create("regex_list");
suite_add_tcase(s, tc_matching);
tcase_add_checked_fixture(tc_matching, rsetup, rteardown);
tcase_add_loop_test(tc_matching, regex_list_match_test, 0, sizeof(rtests) / sizeof(rtests[0]));
tc_phish = tcase_create("phishingScan");
suite_add_tcase(s, tc_phish);
tcase_add_unchecked_fixture(tc_phish, psetup, pteardown);
tcase_add_loop_test(tc_phish, phishingScan_test, 0, sizeof(rtests) / sizeof(rtests[0]));
tcase_add_loop_test(tc_phish, phishingScan_test_allscan, 0, sizeof(rtests) / sizeof(rtests[0]));
tcase_add_test(tc_phish, phishing_fake_test);
tcase_add_test(tc_phish, phishing_fake_test_allscan);
tc_phish2 = tcase_create("phishingScan with 2 dbs");
suite_add_tcase(s, tc_phish2);
tcase_add_unchecked_fixture(tc_phish2, psetup2, pteardown);
tcase_add_loop_test(tc_phish2, phishingScan_test, 0, sizeof(rtests) / sizeof(rtests[0]));
tcase_add_loop_test(tc_phish2, phishingScan_test_allscan, 0, sizeof(rtests) / sizeof(rtests[0]));
tcase_add_test(tc_phish2, phishing_fake_test);
tcase_add_test(tc_phish2, phishing_fake_test_allscan);
tcase_add_loop_test(tc_phish, test_url_canon, 0, sizeof(uc) / sizeof(uc[0]));
tc_regex = tcase_create("cli_regcomp/execute");
suite_add_tcase(s, tc_regex);
tcase_add_loop_test(tc_regex, test_regexes, 0, sizeof(rg) / sizeof(rg[0]));
return s;
}