ClamAV is an open source (GPLv2) anti-virus toolkit.
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
clamav/libclamav/pe_icons.c

1723 lines
80 KiB

/*
* Copyright (C) 2013-2023 Cisco Systems, Inc. and/or its affiliates. All rights reserved.
* Copyright (C) 2009-2013 Sourcefire, Inc.
*
* Authors: aCaB <acab@clamav.net>
*
* 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
16 years ago
#include <string.h>
#include <math.h>
#include "clamav.h"
#include "pe_icons.h"
#include "others.h"
14 years ago
#define READ32(x) cli_readint32(&(x))
#define READ16(x) cli_readint16(&(x))
16 years ago
#define USE_FLOATS
#ifdef USE_FLOATS
#define LABDIFF(x) labdiff(x)
16 years ago
#else
#define LABDIFF(x) labdiff2(x)
16 years ago
#endif
/* #define LOGPARSEICONDETAILS */
struct ICON_ENV {
cli_ctx *ctx;
unsigned int gcnt, hcnt; /* gcnt -> number of icon groups parsed, hcnt -> "actual" image count */
uint32_t lastg;
int result;
icon_groupset *set;
PE parsing code improvements, db loading bug fixes Consolidate the PE parsing code into one function. I tried to preserve all existing functionality from the previous, distinct implementations to a large extent (with the exceptions mentioned below). If I noticed potential bugs/improvements, I added a TODO statement about those so that they can be fixed in a smaller commit later. Also, there are more TODOs in places where I'm not entirely sure why certain actions are performed - more research is needed for these. I'm submitting a pull request now so that regression testing can be done, and because merging what I have thus far now will likely have fewer conflicts than if I try to merge later PE parsing code improvements: - PEs without all 16 data directories are parsed more appropriately now - Added lots more debug statements Also: - Allow MAX_BC and MAX_TRACKED_PCRE to be specified via CFLAGS When doing performance testing with the latest CVD, MAX_BC and MAX_TRACKED_PCRE need to be raised to track all the events. Allow these to be specified via CFLAGS by not redefining them if they are already defined - Fix an issue preventing wildcard sizes in .MDB/.MSB rules I'm not sure what the original intent of the check I removed was, but it prevents using wildcard sizes in .MDB/.MSB rules. AFAICT these wildcard sizes should be handled appropriately by the MD5 section hash computation code, so I don't think a check on that is needed. - Fix several issues related to db loading - .imp files will now get loaded if they exist in a directory passed via clamscan's '-d' flag - .pwdb files will now get loaded if they exist in a directory passed via clamscan's '-d' flag even when compiling without yara support - Changes to .imp, .ign, and .ign2 files will now be reflected in calls to cl_statinidir and cl_statchkdir (and also .pwdb files, even when compiling without yara support) - The contents of .sfp files won't be included in some of the signature counts, and the contents of .cud files will be - Any local.gdb files will no longer be loaded twice - For .imp files, you are no longer required to specify a minimum flevel for wildcard rules, since this isn't needed
6 years ago
struct cli_exe_info *peinfo;
uint32_t icnt; /* number of icon entries parsed, declared images */
uint32_t max_icons;
uint32_t err_oof; /* parseicon: offset to icon is out of file */
uint32_t err_bhoof; /* parseicon: bmp header is out of file */
uint32_t err_bhts; /* parseicon: BMP header too small */
uint32_t err_tstl; /* parseicon: Image too small or too big */
uint32_t err_insl; /* parseicon: Image not square enough */
};
int cli_groupiconscan(struct ICON_ENV *icon_env, uint32_t rva);
static int groupicon_scan_cb(void *ptr, uint32_t type, uint32_t name, uint32_t lang, uint32_t rva)
{
struct ICON_ENV *icon_env = ptr;
int ret = CL_CLEAN;
UNUSEDPARAM(type);
UNUSEDPARAM(lang);
cli_dbgmsg("groupicon_cb: scanning group %x\n", name);
if (!icon_env->gcnt || icon_env->lastg == name) {
icon_env->gcnt++;
icon_env->lastg = name;
/* scan icon group */
ret = cli_groupiconscan(icon_env, rva);
if (ret != CL_CLEAN)
return 1;
return 0;
}
return 1;
}
static int parseicon(struct ICON_ENV *icon_env, uint32_t rva);
static int icon_scan_cb(void *ptr, uint32_t type, uint32_t name, uint32_t lang, uint32_t rva)
{
struct ICON_ENV *icon_env = ptr;
UNUSEDPARAM(type);
UNUSEDPARAM(lang);
UNUSEDPARAM(name);
/* scan icon */
icon_env->result = parseicon(icon_env, rva);
icon_env->hcnt++;
if (icon_env->result != CL_CLEAN)
return 1;
return 0;
}
PE parsing code improvements, db loading bug fixes Consolidate the PE parsing code into one function. I tried to preserve all existing functionality from the previous, distinct implementations to a large extent (with the exceptions mentioned below). If I noticed potential bugs/improvements, I added a TODO statement about those so that they can be fixed in a smaller commit later. Also, there are more TODOs in places where I'm not entirely sure why certain actions are performed - more research is needed for these. I'm submitting a pull request now so that regression testing can be done, and because merging what I have thus far now will likely have fewer conflicts than if I try to merge later PE parsing code improvements: - PEs without all 16 data directories are parsed more appropriately now - Added lots more debug statements Also: - Allow MAX_BC and MAX_TRACKED_PCRE to be specified via CFLAGS When doing performance testing with the latest CVD, MAX_BC and MAX_TRACKED_PCRE need to be raised to track all the events. Allow these to be specified via CFLAGS by not redefining them if they are already defined - Fix an issue preventing wildcard sizes in .MDB/.MSB rules I'm not sure what the original intent of the check I removed was, but it prevents using wildcard sizes in .MDB/.MSB rules. AFAICT these wildcard sizes should be handled appropriately by the MD5 section hash computation code, so I don't think a check on that is needed. - Fix several issues related to db loading - .imp files will now get loaded if they exist in a directory passed via clamscan's '-d' flag - .pwdb files will now get loaded if they exist in a directory passed via clamscan's '-d' flag even when compiling without yara support - Changes to .imp, .ign, and .ign2 files will now be reflected in calls to cl_statinidir and cl_statchkdir (and also .pwdb files, even when compiling without yara support) - The contents of .sfp files won't be included in some of the signature counts, and the contents of .cud files will be - Any local.gdb files will no longer be loaded twice - For .imp files, you are no longer required to specify a minimum flevel for wildcard rules, since this isn't needed
6 years ago
int cli_scanicon(icon_groupset *set, cli_ctx *ctx, struct cli_exe_info *peinfo)
{
struct ICON_ENV icon_env;
libclamav: Fix scan recursion tracking Scan recursion is the process of identifying files embedded in other files and then scanning them, recursively. Internally this process is more complex than it may sound because a file may have multiple layers of types before finding a new "file". At present we treat the recursion count in the scanning context as an index into both our fmap list AND our container list. These two lists are conceptually a part of the same thing and should be unified. But what's concerning is that the "recursion level" isn't actually incremented or decremented at the same time that we add a layer to the fmap or container lists but instead is more touchy-feely, increasing when we find a new "file". To account for this shadiness, the size of the fmap and container lists has always been a little longer than our "max scan recursion" limit so we don't accidentally overflow the fmap or container arrays (!). I've implemented a single recursion-stack as an array, similar to before, which includes a pointer to each fmap at each layer, along with the size and type. Push and pop functions add and remove layers whenever a new fmap is added. A boolean argument when pushing indicates if the new layer represents a new buffer or new file (descriptor). A new buffer will reset the "nested fmap level" (described below). This commit also provides a solution for an issue where we detect embedded files more than once during scan recursion. For illustration, imagine a tarball named foo.tar.gz with this structure: | description | type | rec level | nested fmap level | | ------------------------- | ----- | --------- | ----------------- | | foo.tar.gz | GZ | 0 | 0 | | └── foo.tar | TAR | 1 | 0 | | ├── bar.zip | ZIP | 2 | 1 | | │   └── hola.txt | ASCII | 3 | 0 | | └── baz.exe | PE | 2 | 1 | But suppose baz.exe embeds a ZIP archive and a 7Z archive, like this: | description | type | rec level | nested fmap level | | ------------------------- | ----- | --------- | ----------------- | | baz.exe | PE | 0 | 0 | | ├── sfx.zip | ZIP | 1 | 1 | | │   └── hello.txt | ASCII | 2 | 0 | | └── sfx.7z | 7Z | 1 | 1 | |    └── world.txt | ASCII | 2 | 0 | (A) If we scan for embedded files at any layer, we may detect: | description | type | rec level | nested fmap level | | ------------------------- | ----- | --------- | ----------------- | | foo.tar.gz | GZ | 0 | 0 | | ├── foo.tar | TAR | 1 | 0 | | │ ├── bar.zip | ZIP | 2 | 1 | | │ │   └── hola.txt | ASCII | 3 | 0 | | │ ├── baz.exe | PE | 2 | 1 | | │ │ ├── sfx.zip | ZIP | 3 | 1 | | │ │ │   └── hello.txt | ASCII | 4 | 0 | | │ │ └── sfx.7z | 7Z | 3 | 1 | | │ │    └── world.txt | ASCII | 4 | 0 | | │ ├── sfx.zip | ZIP | 2 | 1 | | │ │   └── hello.txt | ASCII | 3 | 0 | | │ └── sfx.7z | 7Z | 2 | 1 | | │   └── world.txt | ASCII | 3 | 0 | | ├── sfx.zip | ZIP | 1 | 1 | | └── sfx.7z | 7Z | 1 | 1 | (A) is bad because it scans content more than once. Note that for the GZ layer, it may detect the ZIP and 7Z if the signature hits on the compressed data, which it might, though extracting the ZIP and 7Z will likely fail. The reason the above doesn't happen now is that we restrict embedded type scans for a bunch of archive formats to include GZ and TAR. (B) If we scan for embedded files at the foo.tar layer, we may detect: | description | type | rec level | nested fmap level | | ------------------------- | ----- | --------- | ----------------- | | foo.tar.gz | GZ | 0 | 0 | | └── foo.tar | TAR | 1 | 0 | | ├── bar.zip | ZIP | 2 | 1 | | │   └── hola.txt | ASCII | 3 | 0 | | ├── baz.exe | PE | 2 | 1 | | ├── sfx.zip | ZIP | 2 | 1 | | │   └── hello.txt | ASCII | 3 | 0 | | └── sfx.7z | 7Z | 2 | 1 | |    └── world.txt | ASCII | 3 | 0 | (B) is almost right. But we can achieve it easily enough only scanning for embedded content in the current fmap when the "nested fmap level" is 0. The upside is that it should safely detect all embedded content, even if it may think the sfz.zip and sfx.7z are in foo.tar instead of in baz.exe. The biggest risk I can think of affects ZIPs. SFXZIP detection is identical to ZIP detection, which is why we don't allow SFXZIP to be detected if insize of a ZIP. If we only allow embedded type scanning at fmap-layer 0 in each buffer, this will fail to detect the embedded ZIP if the bar.exe was not compressed in foo.zip and if non-compressed files extracted from ZIPs aren't extracted as new buffers: | description | type | rec level | nested fmap level | | ------------------------- | ----- | --------- | ----------------- | | foo.zip | ZIP | 0 | 0 | | └── bar.exe | PE | 1 | 1 | | └── sfx.zip | ZIP | 2 | 2 | Provided that we ensure all files extracted from zips are scanned in new buffers, option (B) should be safe. (C) If we scan for embedded files at the baz.exe layer, we may detect: | description | type | rec level | nested fmap level | | ------------------------- | ----- | --------- | ----------------- | | foo.tar.gz | GZ | 0 | 0 | | └── foo.tar | TAR | 1 | 0 | | ├── bar.zip | ZIP | 2 | 1 | | │   └── hola.txt | ASCII | 3 | 0 | | └── baz.exe | PE | 2 | 1 | | ├── sfx.zip | ZIP | 3 | 1 | | │   └── hello.txt | ASCII | 4 | 0 | | └── sfx.7z | 7Z | 3 | 1 | |    └── world.txt | ASCII | 4 | 0 | (C) is right. But it's harder to achieve. For this example we can get it by restricting 7ZSFX and ZIPSFX detection only when scanning an executable. But that may mean losing detection of archives embedded elsewhere. And we'd have to identify allowable container types for each possible embedded type, which would be very difficult. So this commit aims to solve the issue the (B)-way. Note that in all situations, we still have to scan with file typing enabled to determine if we need to reassign the current file type, such as re-identifying a Bzip2 archive as a DMG that happens to be Bzip2- compressed. Detection of DMG and a handful of other types rely on finding data partway through or near the ned of a file before reassigning the entire file as the new type. Other fixes and considerations in this commit: - The utf16 HTML parser has weak error handling, particularly with respect to creating a nested fmap for scanning the ascii decoded file. This commit cleans up the error handling and wraps the nested scan with the recursion-stack push()/pop() for correct recursion tracking. Before this commit, each container layer had a flag to indicate if the container layer is valid. We need something similar so that the cli_recursion_stack_get_*() functions ignore normalized layers. Details... Imagine an LDB signature for HTML content that specifies a ZIP container. If the signature actually alerts on the normalized HTML and you don't ignore normalized layers for the container check, it will appear as though the alert is in an HTML container rather than a ZIP container. This commit accomplishes this with a boolean you set in the scan context before scanning a new layer. Then when the new fmap is created, it will use that flag to set similar flag for the layer. The context flag is reset those that anything after this doesn't have that flag. The flag allows the new recursion_stack_get() function to ignore normalized layers when iterating the stack to return a layer at a requested index, negative or positive. Scanning normalized extracted/normalized javascript and VBA should also use the 'layer is normalized' flag. - This commit also fixes Heuristic.Broken.Executable alert for ELF files to make sure that: A) these only alert if cli_append_virus() returns CL_VIRUS (aka it respects the FP check). B) all broken-executable alerts for ELF only happen if the SCAN_HEURISTIC_BROKEN option is enabled. - This commit also cleans up the error handling in cli_magic_scan_dir(). This was needed so we could correctly apply the layer-is-normalized-flag to all VBA macros extracted to a directory when scanning the directory. - Also fix an issue where exceeding scan maximums wouldn't cause embedded file detection scans to abort. Granted we don't actually want to abort if max filesize or max recursion depth are exceeded... only if max scansize, max files, and max scantime are exceeded. Add 'abort_scan' flag to scan context, to protect against depending on correct error propagation for fatal conditions. Instead, setting this flag in the scan context should guarantee that a fatal condition deep in scan recursion isn't lost which result in more stuff being scanned instead of aborting. This shouldn't be necessary, but some status codes like CL_ETIMEOUT never used to be fatal and it's easier to do this than to verify every parser only returns CL_ETIMEOUT and other "fatal status codes" in fatal conditions. - Remove duplicate is_tar() prototype from filestypes.c and include is_tar.h instead. - Presently we create the fmap hash when creating the fmap. This wastes a bit of CPU if the hash is never needed. Now that we're creating fmap's for all embedded files discovered with file type recognition scans, this is a much more frequent occurence and really slows things down. This commit fixes the issue by only creating fmap hashes as needed. This should not only resolve the perfomance impact of creating fmap's for all embedded files, but also should improve performance in general. - Add allmatch check to the zip parser after the central-header meta match. That way we don't multiple alerts with the same match except in allmatch mode. Clean up error handling in the zip parser a tiny bit. - Fixes to ensure that the scan limits such as scansize, filesize, recursion depth, # of embedded files, and scantime are always reported if AlertExceedsMax (--alert-exceeds-max) is enabled. - Fixed an issue where non-fatal alerts for exceeding scan maximums may mask signature matches later on. I changed it so these alerts use the "possibly unwanted" alert-type and thus only alert if no other alerts were found or if all-match or heuristic-precedence are enabled. - Added the "Heuristics.Limits.Exceeded.*" events to the JSON metadata when the --gen-json feature is enabled. These will show up once under "ParseErrors" the first time a limit is exceeded. In the present implementation, only one limits-exceeded events will be added, so as to prevent a malicious or malformed sample from filling the JSON buffer with millions of events and using a tonne of RAM.
4 years ago
fmap_t *map = ctx->fmap;
uint32_t err_total = 0;
icon_env.ctx = ctx;
icon_env.gcnt = 0;
icon_env.hcnt = 0;
icon_env.icnt = 0;
icon_env.lastg = 0;
icon_env.result = CL_CLEAN;
6 years ago
icon_env.set = set;
PE parsing code improvements, db loading bug fixes Consolidate the PE parsing code into one function. I tried to preserve all existing functionality from the previous, distinct implementations to a large extent (with the exceptions mentioned below). If I noticed potential bugs/improvements, I added a TODO statement about those so that they can be fixed in a smaller commit later. Also, there are more TODOs in places where I'm not entirely sure why certain actions are performed - more research is needed for these. I'm submitting a pull request now so that regression testing can be done, and because merging what I have thus far now will likely have fewer conflicts than if I try to merge later PE parsing code improvements: - PEs without all 16 data directories are parsed more appropriately now - Added lots more debug statements Also: - Allow MAX_BC and MAX_TRACKED_PCRE to be specified via CFLAGS When doing performance testing with the latest CVD, MAX_BC and MAX_TRACKED_PCRE need to be raised to track all the events. Allow these to be specified via CFLAGS by not redefining them if they are already defined - Fix an issue preventing wildcard sizes in .MDB/.MSB rules I'm not sure what the original intent of the check I removed was, but it prevents using wildcard sizes in .MDB/.MSB rules. AFAICT these wildcard sizes should be handled appropriately by the MD5 section hash computation code, so I don't think a check on that is needed. - Fix several issues related to db loading - .imp files will now get loaded if they exist in a directory passed via clamscan's '-d' flag - .pwdb files will now get loaded if they exist in a directory passed via clamscan's '-d' flag even when compiling without yara support - Changes to .imp, .ign, and .ign2 files will now be reflected in calls to cl_statinidir and cl_statchkdir (and also .pwdb files, even when compiling without yara support) - The contents of .sfp files won't be included in some of the signature counts, and the contents of .cud files will be - Any local.gdb files will no longer be loaded twice - For .imp files, you are no longer required to specify a minimum flevel for wildcard rules, since this isn't needed
6 years ago
icon_env.peinfo = peinfo;
icon_env.max_icons = ctx->engine->maxiconspe;
icon_env.err_oof = 0;
icon_env.err_bhoof = 0;
icon_env.err_bhts = 0;
icon_env.err_tstl = 0;
icon_env.err_insl = 0;
/* icon group scan callback --> groupicon_scan_cb() */
PE parsing code improvements, db loading bug fixes Consolidate the PE parsing code into one function. I tried to preserve all existing functionality from the previous, distinct implementations to a large extent (with the exceptions mentioned below). If I noticed potential bugs/improvements, I added a TODO statement about those so that they can be fixed in a smaller commit later. Also, there are more TODOs in places where I'm not entirely sure why certain actions are performed - more research is needed for these. I'm submitting a pull request now so that regression testing can be done, and because merging what I have thus far now will likely have fewer conflicts than if I try to merge later PE parsing code improvements: - PEs without all 16 data directories are parsed more appropriately now - Added lots more debug statements Also: - Allow MAX_BC and MAX_TRACKED_PCRE to be specified via CFLAGS When doing performance testing with the latest CVD, MAX_BC and MAX_TRACKED_PCRE need to be raised to track all the events. Allow these to be specified via CFLAGS by not redefining them if they are already defined - Fix an issue preventing wildcard sizes in .MDB/.MSB rules I'm not sure what the original intent of the check I removed was, but it prevents using wildcard sizes in .MDB/.MSB rules. AFAICT these wildcard sizes should be handled appropriately by the MD5 section hash computation code, so I don't think a check on that is needed. - Fix several issues related to db loading - .imp files will now get loaded if they exist in a directory passed via clamscan's '-d' flag - .pwdb files will now get loaded if they exist in a directory passed via clamscan's '-d' flag even when compiling without yara support - Changes to .imp, .ign, and .ign2 files will now be reflected in calls to cl_statinidir and cl_statchkdir (and also .pwdb files, even when compiling without yara support) - The contents of .sfp files won't be included in some of the signature counts, and the contents of .cud files will be - Any local.gdb files will no longer be loaded twice - For .imp files, you are no longer required to specify a minimum flevel for wildcard rules, since this isn't needed
6 years ago
findres(14, 0xffffffff, map, peinfo, groupicon_scan_cb, &icon_env);
/* CL_EMAXSIZE is used to track the icon limit */
if (icon_env.result == CL_EMAXSIZE)
cli_dbgmsg("cli_scanicon: max icon count reached\n");
cli_dbgmsg("cli_scanicon: scanned a total of %u[%u actual] icons across %u groups\n", icon_env.icnt, icon_env.hcnt, icon_env.gcnt);
if (icon_env.hcnt < icon_env.icnt)
cli_warnmsg("cli_scanicon: found %u invalid icon entries of %u total\n", icon_env.icnt - icon_env.hcnt, icon_env.icnt);
err_total = icon_env.err_oof + icon_env.err_bhoof + icon_env.err_bhts + icon_env.err_tstl + icon_env.err_insl;
if (err_total > 0) {
cli_dbgmsg("cli_scanicon: detected %u total image parsing issues\n", err_total);
if (icon_env.err_oof > 0)
cli_dbgmsg("cli_scanicon: detected %u cases of 'parseicon: offset to icon is out of file'\n", icon_env.err_oof);
if (icon_env.err_bhoof > 0)
cli_dbgmsg("cli_scanicon: detected %u cases of 'parseicon: bmp header is out of file'\n", icon_env.err_bhoof);
if (icon_env.err_bhts > 0)
cli_dbgmsg("cli_scanicon: detected %u cases of 'parseicon: BMP header too small'\n", icon_env.err_bhts);
if (icon_env.err_tstl > 0)
cli_dbgmsg("cli_scanicon: detected %u cases of 'parseicon: Image too small or too big'\n", icon_env.err_tstl);
if (icon_env.err_insl > 0)
cli_dbgmsg("cli_scanicon: detected %u cases of 'parseicon: Image not square enough'\n", icon_env.err_insl);
}
/* ignore all error returns (previous behavior) */
if (icon_env.result == CL_VIRUS)
return CL_VIRUS;
return CL_CLEAN;
}
int cli_groupiconscan(struct ICON_ENV *icon_env, uint32_t rva)
{
/* import environment */
6 years ago
cli_ctx *ctx = icon_env->ctx;
PE parsing code improvements, db loading bug fixes Consolidate the PE parsing code into one function. I tried to preserve all existing functionality from the previous, distinct implementations to a large extent (with the exceptions mentioned below). If I noticed potential bugs/improvements, I added a TODO statement about those so that they can be fixed in a smaller commit later. Also, there are more TODOs in places where I'm not entirely sure why certain actions are performed - more research is needed for these. I'm submitting a pull request now so that regression testing can be done, and because merging what I have thus far now will likely have fewer conflicts than if I try to merge later PE parsing code improvements: - PEs without all 16 data directories are parsed more appropriately now - Added lots more debug statements Also: - Allow MAX_BC and MAX_TRACKED_PCRE to be specified via CFLAGS When doing performance testing with the latest CVD, MAX_BC and MAX_TRACKED_PCRE need to be raised to track all the events. Allow these to be specified via CFLAGS by not redefining them if they are already defined - Fix an issue preventing wildcard sizes in .MDB/.MSB rules I'm not sure what the original intent of the check I removed was, but it prevents using wildcard sizes in .MDB/.MSB rules. AFAICT these wildcard sizes should be handled appropriately by the MD5 section hash computation code, so I don't think a check on that is needed. - Fix several issues related to db loading - .imp files will now get loaded if they exist in a directory passed via clamscan's '-d' flag - .pwdb files will now get loaded if they exist in a directory passed via clamscan's '-d' flag even when compiling without yara support - Changes to .imp, .ign, and .ign2 files will now be reflected in calls to cl_statinidir and cl_statchkdir (and also .pwdb files, even when compiling without yara support) - The contents of .sfp files won't be included in some of the signature counts, and the contents of .cud files will be - Any local.gdb files will no longer be loaded twice - For .imp files, you are no longer required to specify a minimum flevel for wildcard rules, since this isn't needed
6 years ago
struct cli_exe_info *peinfo = icon_env->peinfo;
int err = 0;
libclamav: Fix scan recursion tracking Scan recursion is the process of identifying files embedded in other files and then scanning them, recursively. Internally this process is more complex than it may sound because a file may have multiple layers of types before finding a new "file". At present we treat the recursion count in the scanning context as an index into both our fmap list AND our container list. These two lists are conceptually a part of the same thing and should be unified. But what's concerning is that the "recursion level" isn't actually incremented or decremented at the same time that we add a layer to the fmap or container lists but instead is more touchy-feely, increasing when we find a new "file". To account for this shadiness, the size of the fmap and container lists has always been a little longer than our "max scan recursion" limit so we don't accidentally overflow the fmap or container arrays (!). I've implemented a single recursion-stack as an array, similar to before, which includes a pointer to each fmap at each layer, along with the size and type. Push and pop functions add and remove layers whenever a new fmap is added. A boolean argument when pushing indicates if the new layer represents a new buffer or new file (descriptor). A new buffer will reset the "nested fmap level" (described below). This commit also provides a solution for an issue where we detect embedded files more than once during scan recursion. For illustration, imagine a tarball named foo.tar.gz with this structure: | description | type | rec level | nested fmap level | | ------------------------- | ----- | --------- | ----------------- | | foo.tar.gz | GZ | 0 | 0 | | └── foo.tar | TAR | 1 | 0 | | ├── bar.zip | ZIP | 2 | 1 | | │   └── hola.txt | ASCII | 3 | 0 | | └── baz.exe | PE | 2 | 1 | But suppose baz.exe embeds a ZIP archive and a 7Z archive, like this: | description | type | rec level | nested fmap level | | ------------------------- | ----- | --------- | ----------------- | | baz.exe | PE | 0 | 0 | | ├── sfx.zip | ZIP | 1 | 1 | | │   └── hello.txt | ASCII | 2 | 0 | | └── sfx.7z | 7Z | 1 | 1 | |    └── world.txt | ASCII | 2 | 0 | (A) If we scan for embedded files at any layer, we may detect: | description | type | rec level | nested fmap level | | ------------------------- | ----- | --------- | ----------------- | | foo.tar.gz | GZ | 0 | 0 | | ├── foo.tar | TAR | 1 | 0 | | │ ├── bar.zip | ZIP | 2 | 1 | | │ │   └── hola.txt | ASCII | 3 | 0 | | │ ├── baz.exe | PE | 2 | 1 | | │ │ ├── sfx.zip | ZIP | 3 | 1 | | │ │ │   └── hello.txt | ASCII | 4 | 0 | | │ │ └── sfx.7z | 7Z | 3 | 1 | | │ │    └── world.txt | ASCII | 4 | 0 | | │ ├── sfx.zip | ZIP | 2 | 1 | | │ │   └── hello.txt | ASCII | 3 | 0 | | │ └── sfx.7z | 7Z | 2 | 1 | | │   └── world.txt | ASCII | 3 | 0 | | ├── sfx.zip | ZIP | 1 | 1 | | └── sfx.7z | 7Z | 1 | 1 | (A) is bad because it scans content more than once. Note that for the GZ layer, it may detect the ZIP and 7Z if the signature hits on the compressed data, which it might, though extracting the ZIP and 7Z will likely fail. The reason the above doesn't happen now is that we restrict embedded type scans for a bunch of archive formats to include GZ and TAR. (B) If we scan for embedded files at the foo.tar layer, we may detect: | description | type | rec level | nested fmap level | | ------------------------- | ----- | --------- | ----------------- | | foo.tar.gz | GZ | 0 | 0 | | └── foo.tar | TAR | 1 | 0 | | ├── bar.zip | ZIP | 2 | 1 | | │   └── hola.txt | ASCII | 3 | 0 | | ├── baz.exe | PE | 2 | 1 | | ├── sfx.zip | ZIP | 2 | 1 | | │   └── hello.txt | ASCII | 3 | 0 | | └── sfx.7z | 7Z | 2 | 1 | |    └── world.txt | ASCII | 3 | 0 | (B) is almost right. But we can achieve it easily enough only scanning for embedded content in the current fmap when the "nested fmap level" is 0. The upside is that it should safely detect all embedded content, even if it may think the sfz.zip and sfx.7z are in foo.tar instead of in baz.exe. The biggest risk I can think of affects ZIPs. SFXZIP detection is identical to ZIP detection, which is why we don't allow SFXZIP to be detected if insize of a ZIP. If we only allow embedded type scanning at fmap-layer 0 in each buffer, this will fail to detect the embedded ZIP if the bar.exe was not compressed in foo.zip and if non-compressed files extracted from ZIPs aren't extracted as new buffers: | description | type | rec level | nested fmap level | | ------------------------- | ----- | --------- | ----------------- | | foo.zip | ZIP | 0 | 0 | | └── bar.exe | PE | 1 | 1 | | └── sfx.zip | ZIP | 2 | 2 | Provided that we ensure all files extracted from zips are scanned in new buffers, option (B) should be safe. (C) If we scan for embedded files at the baz.exe layer, we may detect: | description | type | rec level | nested fmap level | | ------------------------- | ----- | --------- | ----------------- | | foo.tar.gz | GZ | 0 | 0 | | └── foo.tar | TAR | 1 | 0 | | ├── bar.zip | ZIP | 2 | 1 | | │   └── hola.txt | ASCII | 3 | 0 | | └── baz.exe | PE | 2 | 1 | | ├── sfx.zip | ZIP | 3 | 1 | | │   └── hello.txt | ASCII | 4 | 0 | | └── sfx.7z | 7Z | 3 | 1 | |    └── world.txt | ASCII | 4 | 0 | (C) is right. But it's harder to achieve. For this example we can get it by restricting 7ZSFX and ZIPSFX detection only when scanning an executable. But that may mean losing detection of archives embedded elsewhere. And we'd have to identify allowable container types for each possible embedded type, which would be very difficult. So this commit aims to solve the issue the (B)-way. Note that in all situations, we still have to scan with file typing enabled to determine if we need to reassign the current file type, such as re-identifying a Bzip2 archive as a DMG that happens to be Bzip2- compressed. Detection of DMG and a handful of other types rely on finding data partway through or near the ned of a file before reassigning the entire file as the new type. Other fixes and considerations in this commit: - The utf16 HTML parser has weak error handling, particularly with respect to creating a nested fmap for scanning the ascii decoded file. This commit cleans up the error handling and wraps the nested scan with the recursion-stack push()/pop() for correct recursion tracking. Before this commit, each container layer had a flag to indicate if the container layer is valid. We need something similar so that the cli_recursion_stack_get_*() functions ignore normalized layers. Details... Imagine an LDB signature for HTML content that specifies a ZIP container. If the signature actually alerts on the normalized HTML and you don't ignore normalized layers for the container check, it will appear as though the alert is in an HTML container rather than a ZIP container. This commit accomplishes this with a boolean you set in the scan context before scanning a new layer. Then when the new fmap is created, it will use that flag to set similar flag for the layer. The context flag is reset those that anything after this doesn't have that flag. The flag allows the new recursion_stack_get() function to ignore normalized layers when iterating the stack to return a layer at a requested index, negative or positive. Scanning normalized extracted/normalized javascript and VBA should also use the 'layer is normalized' flag. - This commit also fixes Heuristic.Broken.Executable alert for ELF files to make sure that: A) these only alert if cli_append_virus() returns CL_VIRUS (aka it respects the FP check). B) all broken-executable alerts for ELF only happen if the SCAN_HEURISTIC_BROKEN option is enabled. - This commit also cleans up the error handling in cli_magic_scan_dir(). This was needed so we could correctly apply the layer-is-normalized-flag to all VBA macros extracted to a directory when scanning the directory. - Also fix an issue where exceeding scan maximums wouldn't cause embedded file detection scans to abort. Granted we don't actually want to abort if max filesize or max recursion depth are exceeded... only if max scansize, max files, and max scantime are exceeded. Add 'abort_scan' flag to scan context, to protect against depending on correct error propagation for fatal conditions. Instead, setting this flag in the scan context should guarantee that a fatal condition deep in scan recursion isn't lost which result in more stuff being scanned instead of aborting. This shouldn't be necessary, but some status codes like CL_ETIMEOUT never used to be fatal and it's easier to do this than to verify every parser only returns CL_ETIMEOUT and other "fatal status codes" in fatal conditions. - Remove duplicate is_tar() prototype from filestypes.c and include is_tar.h instead. - Presently we create the fmap hash when creating the fmap. This wastes a bit of CPU if the hash is never needed. Now that we're creating fmap's for all embedded files discovered with file type recognition scans, this is a much more frequent occurence and really slows things down. This commit fixes the issue by only creating fmap hashes as needed. This should not only resolve the perfomance impact of creating fmap's for all embedded files, but also should improve performance in general. - Add allmatch check to the zip parser after the central-header meta match. That way we don't multiple alerts with the same match except in allmatch mode. Clean up error handling in the zip parser a tiny bit. - Fixes to ensure that the scan limits such as scansize, filesize, recursion depth, # of embedded files, and scantime are always reported if AlertExceedsMax (--alert-exceeds-max) is enabled. - Fixed an issue where non-fatal alerts for exceeding scan maximums may mask signature matches later on. I changed it so these alerts use the "possibly unwanted" alert-type and thus only alert if no other alerts were found or if all-match or heuristic-precedence are enabled. - Added the "Heuristics.Limits.Exceeded.*" events to the JSON metadata when the --gen-json feature is enabled. These will show up once under "ParseErrors" the first time a limit is exceeded. In the present implementation, only one limits-exceeded events will be added, so as to prevent a malicious or malformed sample from filling the JSON buffer with millions of events and using a tonne of RAM.
4 years ago
fmap_t *map = ctx->fmap;
PE parsing code improvements, db loading bug fixes Consolidate the PE parsing code into one function. I tried to preserve all existing functionality from the previous, distinct implementations to a large extent (with the exceptions mentioned below). If I noticed potential bugs/improvements, I added a TODO statement about those so that they can be fixed in a smaller commit later. Also, there are more TODOs in places where I'm not entirely sure why certain actions are performed - more research is needed for these. I'm submitting a pull request now so that regression testing can be done, and because merging what I have thus far now will likely have fewer conflicts than if I try to merge later PE parsing code improvements: - PEs without all 16 data directories are parsed more appropriately now - Added lots more debug statements Also: - Allow MAX_BC and MAX_TRACKED_PCRE to be specified via CFLAGS When doing performance testing with the latest CVD, MAX_BC and MAX_TRACKED_PCRE need to be raised to track all the events. Allow these to be specified via CFLAGS by not redefining them if they are already defined - Fix an issue preventing wildcard sizes in .MDB/.MSB rules I'm not sure what the original intent of the check I removed was, but it prevents using wildcard sizes in .MDB/.MSB rules. AFAICT these wildcard sizes should be handled appropriately by the MD5 section hash computation code, so I don't think a check on that is needed. - Fix several issues related to db loading - .imp files will now get loaded if they exist in a directory passed via clamscan's '-d' flag - .pwdb files will now get loaded if they exist in a directory passed via clamscan's '-d' flag even when compiling without yara support - Changes to .imp, .ign, and .ign2 files will now be reflected in calls to cl_statinidir and cl_statchkdir (and also .pwdb files, even when compiling without yara support) - The contents of .sfp files won't be included in some of the signature counts, and the contents of .cud files will be - Any local.gdb files will no longer be loaded twice - For .imp files, you are no longer required to specify a minimum flevel for wildcard rules, since this isn't needed
6 years ago
const uint8_t *grp = fmap_need_off_once(map, cli_rawaddr(rva, peinfo->sections, peinfo->nsections, (unsigned int *)(&err), map->len, peinfo->hdr_size), 16);
if (grp && !err) {
uint32_t gsz = cli_readint32(grp + 4);
if (gsz > 6) {
uint32_t icnt, raddr;
unsigned int piconcnt;
struct icondir {
uint8_t w;
uint8_t h;
uint8_t palcnt;
uint8_t rsvd;
uint16_t planes;
uint16_t depth;
uint32_t sz;
uint16_t id;
} * dir;
PE parsing code improvements, db loading bug fixes Consolidate the PE parsing code into one function. I tried to preserve all existing functionality from the previous, distinct implementations to a large extent (with the exceptions mentioned below). If I noticed potential bugs/improvements, I added a TODO statement about those so that they can be fixed in a smaller commit later. Also, there are more TODOs in places where I'm not entirely sure why certain actions are performed - more research is needed for these. I'm submitting a pull request now so that regression testing can be done, and because merging what I have thus far now will likely have fewer conflicts than if I try to merge later PE parsing code improvements: - PEs without all 16 data directories are parsed more appropriately now - Added lots more debug statements Also: - Allow MAX_BC and MAX_TRACKED_PCRE to be specified via CFLAGS When doing performance testing with the latest CVD, MAX_BC and MAX_TRACKED_PCRE need to be raised to track all the events. Allow these to be specified via CFLAGS by not redefining them if they are already defined - Fix an issue preventing wildcard sizes in .MDB/.MSB rules I'm not sure what the original intent of the check I removed was, but it prevents using wildcard sizes in .MDB/.MSB rules. AFAICT these wildcard sizes should be handled appropriately by the MD5 section hash computation code, so I don't think a check on that is needed. - Fix several issues related to db loading - .imp files will now get loaded if they exist in a directory passed via clamscan's '-d' flag - .pwdb files will now get loaded if they exist in a directory passed via clamscan's '-d' flag even when compiling without yara support - Changes to .imp, .ign, and .ign2 files will now be reflected in calls to cl_statinidir and cl_statchkdir (and also .pwdb files, even when compiling without yara support) - The contents of .sfp files won't be included in some of the signature counts, and the contents of .cud files will be - Any local.gdb files will no longer be loaded twice - For .imp files, you are no longer required to specify a minimum flevel for wildcard rules, since this isn't needed
6 years ago
raddr = cli_rawaddr(cli_readint32(grp), peinfo->sections, peinfo->nsections, (unsigned int *)(&err), map->len, peinfo->hdr_size);
cli_dbgmsg("cli_scanicon: icon group @%x\n", raddr);
grp = fmap_need_off_once(map, raddr, gsz);
if (grp && !err) {
icnt = cli_readint32(grp + 2) >> 16;
grp += 6;
gsz -= 6;
while (icnt && gsz >= 14 /* && (remaining amount of icons) */) {
piconcnt = icon_env->hcnt;
dir = (struct icondir *)grp;
cli_dbgmsg("cli_scanicon: Icongrp @%x - %ux%ux%u - (id=%x, rsvd=%u, planes=%u, palcnt=%u, sz=%x)\n", rva, dir->w, dir->h, cli_readint16(&dir->depth), cli_readint16(&dir->id), cli_readint16(&dir->planes), dir->palcnt, dir->rsvd, cli_readint32(&dir->sz));
/* icon scan callback --> icon_scan_cb() */
PE parsing code improvements, db loading bug fixes Consolidate the PE parsing code into one function. I tried to preserve all existing functionality from the previous, distinct implementations to a large extent (with the exceptions mentioned below). If I noticed potential bugs/improvements, I added a TODO statement about those so that they can be fixed in a smaller commit later. Also, there are more TODOs in places where I'm not entirely sure why certain actions are performed - more research is needed for these. I'm submitting a pull request now so that regression testing can be done, and because merging what I have thus far now will likely have fewer conflicts than if I try to merge later PE parsing code improvements: - PEs without all 16 data directories are parsed more appropriately now - Added lots more debug statements Also: - Allow MAX_BC and MAX_TRACKED_PCRE to be specified via CFLAGS When doing performance testing with the latest CVD, MAX_BC and MAX_TRACKED_PCRE need to be raised to track all the events. Allow these to be specified via CFLAGS by not redefining them if they are already defined - Fix an issue preventing wildcard sizes in .MDB/.MSB rules I'm not sure what the original intent of the check I removed was, but it prevents using wildcard sizes in .MDB/.MSB rules. AFAICT these wildcard sizes should be handled appropriately by the MD5 section hash computation code, so I don't think a check on that is needed. - Fix several issues related to db loading - .imp files will now get loaded if they exist in a directory passed via clamscan's '-d' flag - .pwdb files will now get loaded if they exist in a directory passed via clamscan's '-d' flag even when compiling without yara support - Changes to .imp, .ign, and .ign2 files will now be reflected in calls to cl_statinidir and cl_statchkdir (and also .pwdb files, even when compiling without yara support) - The contents of .sfp files won't be included in some of the signature counts, and the contents of .cud files will be - Any local.gdb files will no longer be loaded twice - For .imp files, you are no longer required to specify a minimum flevel for wildcard rules, since this isn't needed
6 years ago
findres(3, cli_readint16(&dir->id), map, peinfo, icon_scan_cb, icon_env);
if (icon_env->result != CL_CLEAN)
return icon_env->result;
if (piconcnt == icon_env->hcnt)
cli_dbgmsg("cli_scanicon: invalid icon entry %u in group @%x\n", dir->id, rva);
icon_env->icnt++;
icnt--;
if (icon_env->icnt >= icon_env->max_icons) {
icon_env->result = CL_EMAXSIZE;
return icon_env->result;
}
grp += 14;
gsz -= 14;
}
if (icnt != 0)
cli_dbgmsg("cli_scanicon: could not find %u icons\n", icnt);
if (gsz != 0)
cli_dbgmsg("cli_scanicon: could not parse %u bytes of icon entries\n", gsz);
}
}
}
return icon_env->result;
16 years ago
}
/* static const int gaussk[]={1,10,45,120,210,252,210,120,45,10,1}; */
static const int gaussk[] = {1, 2, 1};
static const int gkernsz = (sizeof(gaussk) / sizeof(gaussk[0]));
16 years ago
#ifndef USE_FLOATS
// clang-format off
static const uint32_t rtable[256][3] = {
{0x00000000,0x00000000,0x00000000}, {0x00033475,0x0001a70c,0x00002675},
{0x000668e9,0x00034e18,0x00004ceb}, {0x00099d5e,0x0004f525,0x00007360},
{0x000cd1d3,0x00069c31,0x000099d6}, {0x00100648,0x0008433d,0x0000c04b},
{0x00133abc,0x0009ea49,0x0000e6c1}, {0x00166f31,0x000b9156,0x00010d36},
{0x0019a3a6,0x000d3862,0x000133ac}, {0x001cd81b,0x000edf6e,0x00015a21},
{0x00200c8f,0x0010867a,0x00018097}, {0x002355ef,0x00123850,0x0001a807},
{0x0026d1df,0x00140438,0x0001d1d7}, {0x002a7f1c,0x0015e98b,0x0001fdf5},
{0x002e5e65,0x0017e8ad,0x00022c6d}, {0x00327076,0x001a01fd,0x00025d46},
{0x0036b606,0x001c35dc,0x00029088}, {0x003b2fca,0x001e84a5,0x0002c63e},
{0x003fde72,0x0020eeb3,0x0002fe6d}, {0x0044c2aa,0x00237461,0x00033920},
{0x0049dd1d,0x00261604,0x0003765d}, {0x004f2e71,0x0028d3f3,0x0003b62d},
{0x0054b749,0x002bae83,0x0003f898}, {0x005a7848,0x002ea606,0x00043da3},
{0x0060720a,0x0031bace,0x00048559}, {0x0066a52c,0x0034ed2c,0x0004cfbe},
{0x006d1247,0x00383d6e,0x00051cdb}, {0x0073b9f3,0x003babe2,0x00056cb7},
{0x007a9cc3,0x003f38d6,0x0005bf59}, {0x0081bb4a,0x0042e494,0x000614c8},
{0x0089161a,0x0046af67,0x00066d09}, {0x0090adbf,0x004a9998,0x0006c825},
{0x009882c8,0x004ea371,0x00072622}, {0x00a095be,0x0052cd38,0x00078705},
{0x00a8e72b,0x00571734,0x0007ead6}, {0x00b17796,0x005b81ab,0x0008519b},
{0x00ba4783,0x00600ce2,0x0008bb5a}, {0x00c35778,0x0064b91c,0x0009281a},
{0x00cca7f6,0x0069869d,0x000997e0}, {0x00d6397e,0x006e75a7,0x000a0ab2},
{0x00e00c90,0x0073867c,0x000a8097}, {0x00ea21a8,0x0078b95d,0x000af994},
{0x00f47945,0x007e0e8a,0x000b75af}, {0x00ff13e0,0x00838642,0x000bf4ef},
{0x0109f1f4,0x008920c5,0x000c7758}, {0x011513f9,0x008ede4f,0x000cfcf0},
{0x01207a66,0x0094bf20,0x000d85bd}, {0x012c25b2,0x009ac373,0x000e11c5},
{0x01381652,0x00a0eb85,0x000ea10c}, {0x01444cb8,0x00a73792,0x000f3399},
{0x0150c959,0x00ada7d5,0x000fc970}, {0x015d8ca4,0x00b43c89,0x00106298},
{0x016a970c,0x00baf5e6,0x0010ff15}, {0x0177e8ff,0x00c1d428,0x00119eec},
{0x018582ed,0x00c8d786,0x00124223}, {0x01936541,0x00d0003a,0x0012e8bf},
{0x01a19069,0x00d74e7b,0x001392c5}, {0x01b004d1,0x00dec280,0x0014403a},
{0x01bec2e3,0x00e65c82,0x0014f123}, {0x01cdcb08,0x00ee1cb5,0x0015a585},
{0x01dd1dab,0x00f60351,0x00165d64}, {0x01ecbb32,0x00fe108b,0x001718c7},
{0x01fca405,0x01064498,0x0017d7b1}, {0x020cd88a,0x010e9fad,0x00189a27},
{0x021d5927,0x011721fe,0x0019602e}, {0x022e2641,0x011fcbc0,0x001a29cc},
{0x023f403c,0x01289d25,0x001af703}, {0x0250a77a,0x01319661,0x001bc7da},
{0x02625c5f,0x013ab7a8,0x001c9c55}, {0x02745f4c,0x0144012a,0x001d7478},
{0x0286b0a2,0x014d731b,0x001e5048}, {0x029950c2,0x01570dab,0x001f2fca},
{0x02ac400b,0x0160d10d,0x00201301}, {0x02bf7edc,0x016abd71,0x0020f9f3},
{0x02d30d94,0x0174d308,0x0021e4a4}, {0x02e6ec90,0x017f1203,0x0022d318},
{0x02fb1c2e,0x01897a90,0x0023c553}, {0x030f9cc9,0x01940ce0,0x0024bb5a},
{0x03246ebe,0x019ec923,0x0025b532}, {0x03399268,0x01a9af87,0x0026b2de},
{0x034f0822,0x01b4c03b,0x0027b462}, {0x0364d045,0x01bffb6d,0x0028b9c4},
{0x037aeb2a,0x01cb614c,0x0029c307}, {0x0391592c,0x01d6f205,0x002ad02f},
{0x03a81aa2,0x01e2adc6,0x002be141}, {0x03bf2fe4,0x01ee94bc,0x002cf640},
{0x03d6994a,0x01faa715,0x002e0f30}, {0x03ee5729,0x0206e4fc,0x002f2c17},
{0x040669d9,0x02134e9f,0x00304cf7}, {0x041ed1ae,0x021fe429,0x003171d5},
{0x04378eff,0x022ca5c7,0x00329ab5}, {0x0450a220,0x023993a5,0x0033c79b},
{0x046a0b65,0x0246aded,0x0034f88a}, {0x0483cb22,0x0253f4ca,0x00362d87},
{0x049de1aa,0x02616869,0x00376695}, {0x04b84f50,0x026f08f3,0x0038a3b9},
{0x04d31467,0x027cd692,0x0039e4f6}, {0x04ee3140,0x028ad173,0x003b2a50},
{0x0509a62c,0x0298f9bd,0x003c73cb}, {0x0525737d,0x02a74f9b,0x003dc16b},
{0x05419984,0x02b5d337,0x003f1334}, {0x055e1890,0x02c484b9,0x00406928},
{0x057af0f1,0x02d3644b,0x0041c34d}, {0x059822f6,0x02e27217,0x004321a5},
{0x05b5aef0,0x02f1ae43,0x00448435}, {0x05d3952b,0x030118fa,0x0045eaff},
{0x05f1d5f6,0x0310b263,0x00475609}, {0x0610719f,0x03207aa7,0x0048c555},
{0x062f6873,0x033071ec,0x004a38e7}, {0x064ebabf,0x0340985c,0x004bb0c3},
{0x066e68d0,0x0350ee1d,0x004d2ceb}, {0x068e72f1,0x03617357,0x004ead65},
{0x06aed96f,0x03722830,0x00503233}, {0x06cf9c96,0x03830cd0,0x0051bb59},
{0x06f0bcaf,0x0394215e,0x005348da}, {0x07123a07,0x03a565ff,0x0054daba},
{0x073414e7,0x03b6dadb,0x005670fd}, {0x07564d99,0x03c88018,0x00580ba5},
{0x0778e468,0x03da55da,0x0059aab7}, {0x079bd99c,0x03ec5c4a,0x005b4e35},
{0x07bf2d7f,0x03fe938b,0x005cf624}, {0x07e2e059,0x0410fbc4,0x005ea286},
{0x0806f273,0x0423951a,0x0060535f}, {0x082b6414,0x04365fb1,0x006208b3},
{0x08503586,0x04495bb0,0x0063c284}, {0x0875670e,0x045c893b,0x006580d7},
{0x089af8f4,0x046fe876,0x006743ae}, {0x08c0eb80,0x04837986,0x00690b0c},
{0x08e73ef6,0x04973c90,0x006ad6f6}, {0x090df39f,0x04ab31b7,0x006ca76e},
{0x093509bf,0x04bf5920,0x006e7c77}, {0x095c819c,0x04d3b2ef,0x00705616},
{0x09845b7d,0x04e83f47,0x0072344c}, {0x09ac97a4,0x04fcfe4c,0x0074171e},
{0x09d53659,0x0511f021,0x0075fe8f}, {0x09fe37de,0x052714ea,0x0077eaa1},
{0x0a279c78,0x053c6cca,0x0079db58}, {0x0a51646c,0x0551f7e4,0x007bd0b8},
{0x0a7b8ffc,0x0567b65b,0x007dcac2}, {0x0aa61f6d,0x057da852,0x007fc97c},
{0x0ad11301,0x0593cdeb,0x0081cce7}, {0x0afc6afb,0x05aa2748,0x0083d507},
{0x0b28279e,0x05c0b48d,0x0085e1de}, {0x0b54492c,0x05d775db,0x0087f371},
{0x0b80cfe8,0x05ee6b54,0x008a09c2}, {0x0badbc13,0x0605951a,0x008c24d4},
{0x0bdb0dee,0x061cf350,0x008e44aa}, {0x0c08c5bc,0x06348617,0x00906948},
{0x0c36e3bd,0x064c4d90,0x009292b0}, {0x0c656832,0x066449dc,0x0094c0e5},
{0x0c94535c,0x067c7b1f,0x0096f3ec}, {0x0cc3a57b,0x0694e177,0x00992bc5},
{0x0cf35ecf,0x06ad7d07,0x009b6875}, {0x0d237f99,0x06c64df0,0x009da9ff},
{0x0d540818,0x06df5451,0x009ff064}, {0x0d84f88a,0x06f8904d,0x00a23baa},
{0x0db65131,0x07120204,0x00a48bd2}, {0x0de8124a,0x072ba995,0x00a6e0df},
{0x0e1a3c15,0x07458722,0x00a93ad5}, {0x0e4cced0,0x075f9acb,0x00ab99b5},
{0x0e7fcab9,0x0779e4b0,0x00adfd84}, {0x0eb3300f,0x079464f1,0x00b06644},
{0x0ee6ff0f,0x07af1bad,0x00b2d3f8}, {0x0f1b37f7,0x07ca0906,0x00b546a3},
{0x0f4fdb05,0x07e52d19,0x00b7be48}, {0x0f84e876,0x08008808,0x00ba3ae9},
{0x0fba6087,0x081c19f2,0x00bcbc8a}, {0x0ff04375,0x0837e2f5,0x00bf432e},
{0x1026917d,0x0853e331,0x00c1ced6}, {0x105d4ada,0x08701ac6,0x00c45f86},
{0x10946fca,0x088c89d3,0x00c6f542}, {0x10cc0089,0x08a93076,0x00c9900b},
{0x1103fd52,0x08c60ece,0x00cc2fe4}, {0x113c6661,0x08e324fa,0x00ced4d1},
{0x11753bf2,0x0900731a,0x00d17ed4}, {0x11ae7e40,0x091df94a,0x00d42def},
{0x11e82d85,0x093bb7ab,0x00d6e227}, {0x122249fe,0x0959ae5a,0x00d99b7d},
{0x125cd3e4,0x0977dd75,0x00dc59f3}, {0x1297cb73,0x0996451b,0x00df1d8e},
{0x12d330e4,0x09b4e56a,0x00e1e64f}, {0x130f0472,0x09d3be80,0x00e4b43a},
{0x134b4657,0x09f2d07b,0x00e78751}, {0x1387f6cd,0x0a121b78,0x00ea5f97},
{0x13c5160d,0x0a319f96,0x00ed3d0e}, {0x1402a451,0x0a515cf2,0x00f01fb9},
{0x1440a1d2,0x0a7153a9,0x00f3079b}, {0x147f0eca,0x0a9183da,0x00f5f4b7},
{0x14bdeb71,0x0ab1eda0,0x00f8e70f}, {0x14fd3800,0x0ad2911b,0x00fbdea5},
{0x153cf4b0,0x0af36e66,0x00fedb7e}, {0x157d21ba,0x0b1485a0,0x0101dd9a},
{0x15bdbf54,0x0b35d6e4,0x0104e4fd}, {0x15fecdb9,0x0b576251,0x0107f1aa},
{0x16404d1f,0x0b792802,0x010b03a3}, {0x16823dbf,0x0b9b2815,0x010e1aeb},
{0x16c49fd0,0x0bbd62a7,0x01113784}, {0x1707738a,0x0bdfd7d3,0x01145970},
{0x174ab923,0x0c0287b7,0x011780b4}, {0x178e70d4,0x0c25726f,0x011aad50},
{0x17d29ad3,0x0c489817,0x011ddf48}, {0x18173757,0x0c6bf8cc,0x0121169e},
{0x185c4697,0x0c8f94aa,0x01245355}, {0x18a1c8c9,0x0cb36bcc,0x01279570},
{0x18e7be24,0x0cd77e50,0x012adcf0}, {0x192e26dd,0x0cfbcc51,0x012e29d9},
{0x1975032d,0x0d2055ea,0x01317c2d}, {0x19bc5347,0x0d451b38,0x0134d3ee},
{0x1a041762,0x0d6a1c57,0x0138311f}, {0x1a4c4fb3,0x0d8f5962,0x013b93c3},
{0x1a94fc71,0x0db4d275,0x013efbdc}, {0x1ade1dd0,0x0dda87aa,0x0142696d},
{0x1b27b406,0x0e00791f,0x0145dc77}, {0x1b71bf48,0x0e26a6ee,0x014954fe},
{0x1bbc3fca,0x0e4d1132,0x014cd305}, {0x1c0735c3,0x0e73b807,0x0150568c},
{0x1c52a165,0x0e9a9b87,0x0153df98}, {0x1c9e82e6,0x0ec1bbcf,0x01576e2a},
{0x1ceada7b,0x0ee918f8,0x015b0245}, {0x1d37a857,0x0f10b31e,0x015e9beb},
{0x1d84ecae,0x0f388a5d,0x01623b20}, {0x1dd2a7b6,0x0f609ecd,0x0165dfe4},
{0x1e20d9a0,0x0f88f08b,0x01698a3b}, {0x1e6f82a2,0x0fb17fb1,0x016d3a27},
{0x1ebea2ef,0x0fda4c59,0x0170efab}, {0x1f0e3aba,0x1003569f,0x0174aac9},
{0x1f5e4a37,0x102c9e9c,0x01786b83}, {0x1faed199,0x1056246b,0x017c31db},
{0x1fffd112,0x107fe827,0x017ffdd5}, {0x205148d7,0x10a9e9e9,0x0183cf72},
{0x20a33919,0x10d429cc,0x0187a6b5}, {0x20f5a20b,0x10fea7ea,0x018b83a1},
{0x214883e1,0x1129645d,0x018f6637}, {0x219bdecc,0x11545f3f,0x01934e7a},
{0x21efb2ff,0x117f98aa,0x01973c6d}, {0x224400ac,0x11ab10b9,0x019b3011},
{0x2298c805,0x11d6c783,0x019f2969}, {0x22ee093c,0x1202bd25,0x01a32878},
{0x2343c484,0x122ef1b6,0x01a72d3f}, {0x2399fa0c,0x125b6552,0x01ab37c2},
{0x23f0aa09,0x12881811,0x01af4802}, {0x2447d4aa,0x12b50a0d,0x01b35e01},
{0x249f7a21,0x12e23b5f,0x01b779c3}, {0x24f79a9f,0x130fac21,0x01bb9b49},
{0x25503656,0x133d5c6d,0x01bfc296}, {0x25a94d77,0x136b4c5b,0x01c3efab},
{0x2602e032,0x13997c04,0x01c8228c}, {0x265ceeb9,0x13c7eb83,0x01cc5b3a},
{0x26b7793c,0x13f69aef,0x01d099b9}, {0x27127feb,0x14258a63,0x01d4de09},
{0x276e02f8,0x1454b9f6,0x01d9282e}, {0x27ca0292,0x148429c2,0x01dd7829},
{0x28267ee9,0x14b3d9e1,0x01e1cdfd}, {0x2883782f,0x14e3ca69,0x01e629ac},
{0x28e0ee92,0x1513fb76,0x01ea8b39}, {0x293ee243,0x15446d1e,0x01eef2a6},
};
static const uint32_t gtable[256][3] = {
{0x00000000,0x00000000,0x00000000}, {0x0002c74a,0x00058e94,0x0000ed19},
{0x00058e94,0x000b1d27,0x0001da31}, {0x000855dd,0x0010abbb,0x0002c74a},
{0x000b1d27,0x00163a4f,0x0003b462}, {0x000de471,0x001bc8e2,0x0004a17b},
{0x0010abbb,0x00215776,0x00058e94}, {0x00137305,0x0026e60a,0x00067bac},
{0x00163a4f,0x002c749d,0x000768c5}, {0x00190198,0x00320331,0x000855dd},
{0x001bc8e2,0x003791c5,0x000942f6}, {0x001ea24f,0x003d449e,0x000a361a},
{0x0021a791,0x00434f22,0x000b37db}, {0x0024d791,0x0049af21,0x000c47db},
{0x002832f4,0x005065e8,0x000d6651}, {0x002bba5d,0x005774ba,0x000e9374},
{0x002f6e6c,0x005edcd8,0x000fcf79}, {0x00334fbc,0x00669f78,0x00111a94},
{0x00375ee6,0x006ebdcd,0x001274f7}, {0x003b9c81,0x00773902,0x0013ded5},
{0x0040091f,0x0080123d,0x0015585f}, {0x0044a550,0x00894aa0,0x0016e1c5},
{0x004971a3,0x0092e346,0x00187b36}, {0x004e6ea3,0x009cdd46,0x001a24e1},
{0x00539cda,0x00a739b4,0x001bdef3}, {0x0058fcce,0x00b1f99c,0x001da999},
{0x005e8f05,0x00bd1e09,0x001f8501}, {0x00645400,0x00c8a801,0x00217155},
{0x006a4c43,0x00d49885,0x00236ec0}, {0x0070784a,0x00e0f094,0x00257d6d},
{0x0076d894,0x00edb128,0x00279d86}, {0x007d6d9c,0x00fadb38,0x0029cf33},
{0x008437dc,0x01086fb7,0x002c129e}, {0x008b37cc,0x01166f97,0x002e67ee},
{0x00926de3,0x0124dbc5,0x0030cf4b}, {0x0099da96,0x0133b52b,0x003348dc},
{0x00a17e58,0x0142fcb1,0x0035d4c7}, {0x00a9599d,0x0152b33b,0x00387333},
{0x00b16cd5,0x0162d9aa,0x003b2446}, {0x00b9b870,0x017370df,0x003de824},
{0x00c23cdb,0x018479b7,0x0040bef3}, {0x00cafa85,0x0195f50a,0x0043a8d6},
{0x00d3f1d9,0x01a7e3b2,0x0046a5f2}, {0x00dd2341,0x01ba4683,0x0049b66a},
{0x00e68f28,0x01cd1e50,0x004cda61}, {0x00f035f6,0x01e06bec,0x005011fb},
{0x00fa1812,0x01f43024,0x00535d5a}, {0x010435e3,0x02086bc6,0x0056bc9f},
{0x010e8fce,0x021d1f9b,0x005a2fee}, {0x01192637,0x02324c6d,0x005db766},
{0x0123f982,0x0247f303,0x00615329}, {0x012f0a11,0x025e1421,0x00650359},
{0x013a5846,0x0274b08b,0x0068c815}, {0x0145e481,0x028bc902,0x006ca17e},
{0x0151af22,0x02a35e45,0x00708fb4}, {0x015db889,0x02bb7112,0x007492d6},
{0x016a0112,0x02d40225,0x0078ab04}, {0x0176891d,0x02ed1239,0x007cd85d},
{0x01835103,0x0306a207,0x00811aff}, {0x01905923,0x0320b246,0x00857309},
{0x019da1d6,0x033b43ac,0x0089e09a}, {0x01ab2b77,0x035656ee,0x008e63d0},
{0x01b8f65f,0x0371ecbd,0x0092fcc8}, {0x01c702e6,0x038e05cd,0x0097ab9f},
{0x01d55166,0x03aaa2cc,0x009c7074}, {0x01e3e235,0x03c7c46a,0x00a14b64},
{0x01f2b5aa,0x03e56b54,0x00a63c8b}, {0x0201cc1b,0x04039836,0x00ab4406},
{0x021125de,0x04224bbb,0x00b061f1}, {0x0220c346,0x0441868c,0x00b59669},
{0x0230a4a9,0x04614952,0x00bae18a}, {0x0240ca5a,0x048194b4,0x00c04370},
{0x025134ac,0x04a26957,0x00c5bc36}, {0x0261e3f0,0x04c3c7e1,0x00cb4bf7},
{0x0272d87a,0x04e5b0f5,0x00d0f2d0}, {0x0284129a,0x05082535,0x00d6b0da},
{0x029592a1,0x052b2543,0x00dc8632}, {0x02a758df,0x054eb1bf,0x00e272f1},
{0x02b965a4,0x0572cb48,0x00e87732}, {0x02cbb93e,0x0597727c,0x00ee9310},
{0x02de53fd,0x05bca7fa,0x00f4c6a5}, {0x02f1362e,0x05e26c5b,0x00fb120b},
{0x0304601f,0x0608c03d,0x0101755b}, {0x0317d21c,0x062fa439,0x0107f0af},
{0x032b8c74,0x065718e7,0x010e8422}, {0x033f8f71,0x067f1ee2,0x01152fcb},
{0x0353db5f,0x06a7b6be,0x011bf3c5}, {0x0368708a,0x06d0e114,0x0122d029},
{0x037d4f3c,0x06fa9e79,0x0129c50f}, {0x039277c0,0x0724ef81,0x0130d290},
{0x03a7ea60,0x074fd4c0,0x0137f8c5}, {0x03bda764,0x077b4ec8,0x013f37c6},
{0x03d3af16,0x07a75e2d,0x01468fac}, {0x03ea01bf,0x07d4037f,0x014e008f},
{0x04009fa7,0x08013f4e,0x01558a87}, {0x04178915,0x082f122a,0x015d2dab},
{0x042ebe51,0x085d7ca3,0x0164ea15}, {0x04463fa2,0x088c7f45,0x016cbfda},
{0x045e0d4f,0x08bc1a9e,0x0174af14}, {0x0476279e,0x08ec4f3b,0x017cb7d8},
{0x048e8ed4,0x091d1da8,0x0184da3f}, {0x04a74337,0x094e866e,0x018d1660},
{0x04c0450d,0x09808a1a,0x01956c52}, {0x04d9949a,0x09b32933,0x019ddc2c},
{0x04f33222,0x09e66444,0x01a66604}, {0x050d1de9,0x0a1a3bd3,0x01af09f1},
{0x05275834,0x0a4eb069,0x01b7c809}, {0x0541e146,0x0a83c28c,0x01c0a064},
{0x055cb961,0x0ab972c3,0x01c99318}, {0x0577e0c9,0x0aefc192,0x01d2a03a},
{0x059357bf,0x0b26af7f,0x01dbc7e2}, {0x05af1e87,0x0b5e3d0d,0x01e50a24},
{0x05cb3561,0x0b966ac1,0x01ee6717}, {0x05e79c8e,0x0bcf391d,0x01f7ded1},
{0x06045451,0x0c08a8a3,0x02017167}, {0x06215cea,0x0c42b9d4,0x020b1eef},
{0x063eb699,0x0c7d6d33,0x0214e77f}, {0x065c619f,0x0cb8c33f,0x021ecb2b},
{0x067a5e3c,0x0cf4bc78,0x0228ca0a}, {0x0698acae,0x0d31595d,0x0232e430},
{0x06b74d37,0x0d6e9a6d,0x023d19b2}, {0x06d64013,0x0dac8026,0x02476aa6},
{0x06f58583,0x0deb0b06,0x0251d721}, {0x07151dc5,0x0e2a3b89,0x025c5f37},
{0x07350916,0x0e6a122d,0x026702fc}, {0x075547b6,0x0eaa8f6c,0x0271c287},
{0x0775d9e1,0x0eebb3c3,0x027c9dea}, {0x0796bfd5,0x0f2d7fab,0x0287953b},
{0x07b7f9d0,0x0f6ff3a0,0x0292a88f}, {0x07d9880d,0x0fb3101a,0x029dd7f8},
{0x07fb6aca,0x0ff6d595,0x02a9238c}, {0x081da243,0x103b4487,0x02b48b5f},
{0x08402eb5,0x10805d69,0x02c00f85}, {0x0863105a,0x10c620b4,0x02cbb011},
{0x0886476f,0x110c8edd,0x02d76d18}, {0x08a9d42f,0x1153a85d,0x02e346ad},
{0x08cdb6d5,0x119b6da9,0x02ef3ce4}, {0x08f1ef9c,0x11e3df37,0x02fb4fd1},
{0x09167ebf,0x122cfd7d,0x03077f87}, {0x093b6478,0x1276c8ef,0x0313cc19},
{0x0960a101,0x12c14202,0x0320359c}, {0x09863495,0x130c692a,0x032cbc23},
{0x09ac1f6d,0x13583eda,0x03395fc0}, {0x09d261c3,0x13a4c385,0x03462087},
{0x09f8fbcf,0x13f1f79e,0x0352fe8b}, {0x0a1fedcc,0x143fdb98,0x035ff9df},
{0x0a4737f1,0x148e6fe3,0x036d1296}, {0x0a6eda79,0x14ddb4f1,0x037a48c3},
{0x0a96d59a,0x152dab34,0x03879c78}, {0x0abf298d,0x157e531b,0x03950dc9},
{0x0ae7d68b,0x15cfad17,0x03a29cc8}, {0x0b10dccb,0x1621b997,0x03b04988},
{0x0b3a3c85,0x1674790a,0x03be141b}, {0x0b63f5f0,0x16c7ebe0,0x03cbfc94},
{0x0b8e0943,0x171c1287,0x03da0304}, {0x0bb876b6,0x1770ed6c,0x03e82780},
{0x0be33e7f,0x17c67cff,0x03f66a18}, {0x0c0e60d5,0x181cc1aa,0x0404cadf},
{0x0c39ddef,0x1873bbdd,0x041349e7}, {0x0c65b601,0x18cb6c03,0x0421e742},
{0x0c91e944,0x1923d288,0x0430a303}, {0x0cbe77ec,0x197cefd8,0x043f7d3b},
{0x0ceb622f,0x19d6c45e,0x044e75fc}, {0x0d18a843,0x1a315086,0x045d8d57},
{0x0d464a5d,0x1a8c94ba,0x046cc360}, {0x0d7448b2,0x1ae89164,0x047c1826},
{0x0da2a377,0x1b4546ef,0x048b8bbd}, {0x0dd15ae2,0x1ba2b5c3,0x049b1e36},
{0x0e006f25,0x1c00de4b,0x04aacfa1}, {0x0e2fe077,0x1c5fc0ee,0x04baa011},
{0x0e5faf0b,0x1cbf5e16,0x04ca8f98}, {0x0e8fdb15,0x1d1fb62a,0x04da9e46},
{0x0ec064c9,0x1d80c993,0x04eacc2c}, {0x0ef14c5c,0x1de298b8,0x04fb195d},
{0x0f229200,0x1e452400,0x050b85e8}, {0x0f5435e9,0x1ea86bd1,0x051c11e0},
{0x0f86384a,0x1f0c7094,0x052cbd56}, {0x0fb89957,0x1f7132ad,0x053d885a},
{0x0feb5941,0x1fd6b283,0x054e72fd}, {0x101e783d,0x203cf07b,0x055f7d51},
{0x1051f67d,0x20a3ecfa,0x0570a765}, {0x1085d433,0x210ba866,0x0581f14c},
{0x10ba1191,0x21742322,0x05935b16}, {0x10eeaeca,0x21dd5d94,0x05a4e4d3},
{0x1123ac0f,0x2247581f,0x05b68e95}, {0x11590993,0x22b21326,0x05c8586b},
{0x118ec787,0x231d8f0e,0x05da4267}, {0x11c4e61d,0x2389cc39,0x05ec4c98},
{0x11fb6585,0x23f6cb0b,0x05fe7710}, {0x123245f2,0x24648be5,0x0610c1df},
{0x12698795,0x24d30f2a,0x06232d15}, {0x12a12a9e,0x2542553b,0x0635b8c2},
{0x12d92f3e,0x25b25e7b,0x064864f7}, {0x131195a6,0x26232b4b,0x065b31c4},
{0x134a5e06,0x2694bc0c,0x066e1f39}, {0x1383888f,0x2707111e,0x06812d66},
{0x13bd1571,0x277a2ae2,0x06945c5c}, {0x13f704dc,0x27ee09b8,0x06a7ac2a},
{0x14315700,0x2862ae01,0x06bb1ce1}, {0x146c0c0e,0x28d8181b,0x06ceae8f},
{0x14a72433,0x294e4867,0x06e26146}, {0x14e29fa2,0x29c53f43,0x06f63515},
{0x151e7e87,0x2a3cfd0f,0x070a2a0c}, {0x155ac114,0x2ab58228,0x071e403b},
{0x15976777,0x2b2eceee,0x073277b1}, {0x15d471df,0x2ba8e3be,0x0746d07d},
{0x1611e07b,0x2c23c0f6,0x075b4ab1}, {0x164fb37a,0x2c9f66f4,0x076fe65b},
{0x168deb0a,0x2d1bd615,0x0784a38b}, {0x16cc875b,0x2d990eb6,0x07998250},
{0x170b889a,0x2e171134,0x07ae82ba}, {0x174aeef6,0x2e95ddeb,0x07c3a4d8},
{0x178aba9c,0x2f157539,0x07d8e8ba}, {0x17caebbc,0x2f95d778,0x07ee4e6f},
{0x180b8282,0x30170504,0x0803d606}, {0x184c7f1d,0x3098fe3a,0x08197f8e},
{0x188de1bb,0x311bc375,0x082f4b18}, {0x18cfaa88,0x319f550f,0x084538b1},
{0x1911d9b2,0x3223b364,0x085b4869}, {0x19546f67,0x32a8dece,0x08717a50},
{0x19976bd4,0x332ed7a8,0x0887ce74}, {0x19dacf26,0x33b59e4b,0x089e44e4},
{0x1a1e9989,0x343d3312,0x08b4ddb0}, {0x1a62cb2b,0x34c59656,0x08cb98e5},
{0x1aa76439,0x354ec872,0x08e27694}, {0x1aec64de,0x35d8c9bd,0x08f976cb},
{0x1b31cd49,0x36639a91,0x09109998}, {0x1b779da4,0x36ef3b48,0x0927df0c},
{0x1bbdd61c,0x377bac38,0x093f4733}, {0x1c0476de,0x3808edbc,0x0956d21e},
{0x1c4b8015,0x3897002b,0x096e7fdb}, {0x1c92f1ee,0x3925e3dc,0x09865078},
{0x1cdacc94,0x39b59928,0x099e4404}, {0x1d231033,0x3a462066,0x09b65a8e},
{0x1d6bbcf7,0x3ad779ee,0x09ce9424}, {0x1db4d30b,0x3b69a616,0x09e6f0d5},
{0x1dfe529b,0x3bfca535,0x09ff70af}, {0x1e483bd1,0x3c9077a2,0x0a1813c1},
{0x1e928eda,0x3d251db4,0x0a30da19}, {0x1edd4be0,0x3dba97c0,0x0a49c3c5},
{0x1f28730e,0x3e50e61d,0x0a62d0d4}, {0x1f740490,0x3ee80920,0x0a7c0154},
{0x1fc00090,0x3f80011f,0x0a955554}, {0x200c6738,0x4018ce70,0x0aaecce1},
{0x205938b4,0x40b27167,0x0ac86809}, {0x20a6752d,0x414cea5b,0x0ae226dc},
{0x20f41ccf,0x41e8399f,0x0afc0967}, {0x21422fc4,0x42845f87,0x0b160fb8},
{0x2190ae35,0x43215c6a,0x0b3039dd}, {0x21df984d,0x43bf309b,0x0b4a87e5},
{0x222eee36,0x445ddc6d,0x0b64f9dd}, {0x227eb01a,0x44fd6035,0x0b7f8fd3},
{0x22cede23,0x459dbc46,0x0b9a49d5}, {0x231f787a,0x463ef0f5,0x0bb527f2},
{0x23707f4a,0x46e0fe93,0x0bd02a36}, {0x23c1f2bb,0x4783e575,0x0beb50b1},
};
static const uint32_t btable[256][3] = {
{0x00000000,0x00000000,0x00000000}, {0x000166ed,0x00008f92,0x00076257},
{0x0002cdda,0x00011f24,0x000ec4ae}, {0x000434c7,0x0001aeb6,0x00162705},
{0x00059bb3,0x00023e48,0x001d895c}, {0x000702a0,0x0002cdda,0x0024ebb3},
{0x0008698d,0x00035d6c,0x002c4e0a}, {0x0009d07a,0x0003ecfe,0x0033b061},
{0x000b3767,0x00047c90,0x003b12b8}, {0x000c9e54,0x00050c22,0x0042750f},
{0x000e0541,0x00059bb3,0x0049d765}, {0x000f7554,0x00062eef,0x005169ef},
{0x0010fb87,0x0006cb03,0x005970f8}, {0x0012974a,0x00076fb7,0x0061e996},
{0x001448f2,0x00081d2e,0x006ad585}, {0x001610d2,0x0008d387,0x00743673},
{0x0017ef39,0x000992e4,0x007e0e09}, {0x0019e476,0x000a5b62,0x00885de4},
{0x001bf0d6,0x000b2d23,0x0093279b}, {0x001e14a5,0x000c0842,0x009e6cbc},
{0x0020502e,0x000cecdf,0x00aa2ece}, {0x0022a3b8,0x000ddb16,0x00b66f52},
{0x00250f8c,0x000ed305,0x00c32fbf}, {0x002793f0,0x000fd4c7,0x00d0718a},
{0x002a312a,0x0010e077,0x00de361f}, {0x002ce77d,0x0011f632,0x00ec7ee5},
{0x002fb72c,0x00131612,0x00fb4d3d}, {0x0032a07a,0x00144031,0x010aa283},
{0x0035a3a8,0x001574aa,0x011a800d}, {0x0038c0f6,0x0016b395,0x012ae72e},
{0x003bf8a2,0x0017fd0e,0x013bd932}, {0x003f4aec,0x0019512b,0x014d5761},
{0x0042b811,0x001ab007,0x015f6300}, {0x0046404d,0x001c19b8,0x0171fd4e},
{0x0049e3dc,0x001d8e58,0x01852786}, {0x004da2fa,0x001f0dfe,0x0198e2e0},
{0x00517de1,0x002098c0,0x01ad308f}, {0x005574cb,0x00222eb7,0x01c211c3},
{0x005987f0,0x0023cff9,0x01d787a8}, {0x005db789,0x00257c9d,0x01ed9368},
{0x006203cd,0x002734b9,0x02043626}, {0x00666cf5,0x0028f862,0x021b7106},
{0x006af335,0x002ac7af,0x02334526}, {0x006f96c4,0x002ca2b5,0x024bb3a2},
{0x007457d8,0x002e898a,0x0264bd92}, {0x007936a5,0x00307c42,0x027e640c},
{0x007e335f,0x00327af3,0x0298a823}, {0x00834e39,0x003485b1,0x02b38ae6},
{0x00888768,0x00369c90,0x02cf0d64}, {0x008ddf1d,0x0038bfa5,0x02eb30a7},
{0x0093558b,0x003aef04,0x0307f5b8}, {0x0098eae4,0x003d2ac2,0x03255d9b},
{0x009e9f58,0x003f72f0,0x03436954}, {0x00a47319,0x0041c7a3,0x036219e4},
{0x00aa6656,0x004428ef,0x03817049}, {0x00b07940,0x004696e6,0x03a16d80},
{0x00b6ac06,0x0049119c,0x03c21283}, {0x00bcfed8,0x004b9923,0x03e36049},
{0x00c371e3,0x004e2d8e,0x040557c8}, {0x00ca0556,0x0050ceef,0x0427f9f4},
{0x00d0b960,0x00537d59,0x044b47bf}, {0x00d78e2d,0x005638df,0x046f4218},
{0x00de83ea,0x00590191,0x0493e9ee}, {0x00e59ac5,0x005bd782,0x04b9402a},
{0x00ecd2ea,0x005ebac4,0x04df45b9}, {0x00f42c85,0x0061ab68,0x0505fb82},
{0x00fba7c1,0x0064a981,0x052d626b}, {0x010344cb,0x0067b51e,0x05557b5a},
{0x010b03cd,0x006ace52,0x057e4730}, {0x0112e4f2,0x006df52d,0x05a7c6d0},
{0x011ae864,0x007129c2,0x05d1fb18}, {0x01230e4e,0x00746c1f,0x05fce4e8},
{0x012b56d9,0x0077bc57,0x0628851b}, {0x0133c22f,0x007b1a79,0x0654dc8c},
{0x013c507a,0x007e8697,0x0681ec15}, {0x014501e2,0x008200c1,0x06afb48d},
{0x014dd690,0x00858906,0x06de36cb}, {0x0156ceac,0x00891f78,0x070d73a5},
{0x015fea5f,0x008cc426,0x073d6bed}, {0x016929d1,0x00907720,0x076e2075},
{0x01728d28,0x00943876,0x079f920f}, {0x017c148d,0x00980839,0x07d1c18a},
{0x0185c027,0x009be676,0x0804afb3}, {0x018f901c,0x009fd33f,0x08385d59},
{0x01998494,0x00a3cea2,0x086ccb46}, {0x01a39db4,0x00a7d8af,0x08a1fa45},
{0x01addba3,0x00abf175,0x08d7eb1f}, {0x01b83e87,0x00b01903,0x090e9e9b},
{0x01c2c685,0x00b44f69,0x09461581}, {0x01cd73c4,0x00b894b5,0x097e5095},
{0x01d84667,0x00bce8f6,0x09b7509d}, {0x01e33e95,0x00c14c3c,0x09f1165b},
{0x01ee5c72,0x00c5be94,0x0a2ba292}, {0x01f9a023,0x00ca400e,0x0a66f602},
{0x020509cc,0x00ced0b8,0x0aa3116b}, {0x02109992,0x00d370a1,0x0adff58c},
{0x021c4f98,0x00d81fd6,0x0b1da323}, {0x02282c02,0x00dcde67,0x0b5c1aec},
{0x02342ef4,0x00e1ac62,0x0b9b5da4}, {0x02405892,0x00e689d4,0x0bdb6c05},
{0x024ca8ff,0x00eb76cc,0x0c1c46c8}, {0x0259205d,0x00f07358,0x0c5deea6},
{0x0265becf,0x00f57f86,0x0ca06457}, {0x02728479,0x00fa9b64,0x0ce3a892},
{0x027f717d,0x00ffc6ff,0x0d27bc0c}, {0x028c85fd,0x01050265,0x0d6c9f7b},
{0x0299c21c,0x010a4da5,0x0db25392}, {0x02a725fa,0x010fa8ca,0x0df8d904},
{0x02b4b1bb,0x011513e4,0x0e403084}, {0x02c26580,0x011a8f00,0x0e885ac3},
{0x02d0416a,0x01201a2a,0x0ed15871}, {0x02de459b,0x0125b571,0x0f1b2a3e},
{0x02ec7233,0x012b60e1,0x0f65d0d9}, {0x02fac754,0x01311c88,0x0fb14cef},
{0x03094520,0x0136e873,0x0ffd9f2d}, {0x0317ebb5,0x013cc4af,0x104ac840},
{0x0326bb36,0x0142b149,0x1098c8d4}, {0x0335b3c1,0x0148ae4d,0x10e7a192},
{0x0344d579,0x014ebbca,0x11375325}, {0x0354207c,0x0154d9cb,0x1187de35},
{0x036394eb,0x015b085e,0x11d9436c}, {0x037332e5,0x0161478f,0x122b8370},
{0x0382fa8b,0x0167976b,0x127e9ee8}, {0x0392ebfb,0x016df7fe,0x12d2967b},
{0x03a30755,0x01746955,0x13276ace}, {0x03b34cb9,0x017aeb7d,0x137d1c85},
{0x03c3bc45,0x01817e82,0x13d3ac44}, {0x03d45619,0x01882270,0x142b1aaf},
{0x03e51a53,0x018ed754,0x14836867}, {0x03f60911,0x01959d3a,0x14dc9610},
{0x04072274,0x019c742e,0x1536a449}, {0x04186698,0x01a35c3d,0x159193b4},
{0x0429d59d,0x01aa5572,0x15ed64f0}, {0x043b6fa1,0x01b15fda,0x164a189c},
{0x044d34c1,0x01b87b80,0x16a7af56}, {0x045f251c,0x01bfa872,0x170629bd},
{0x047140d0,0x01c6e6b9,0x1765886e}, {0x048387f9,0x01ce3664,0x17c5cc05},
{0x0495fab7,0x01d5977c,0x1826f51f}, {0x04a89926,0x01dd0a0f,0x18890455},
{0x04bb6364,0x01e48e28,0x18ebfa44}, {0x04ce598d,0x01ec23d2,0x194fd785},
{0x04e17bc0,0x01f3cb1a,0x19b49cb2}, {0x04f4ca19,0x01fb840a,0x1a1a4a64},
{0x050844b5,0x02034eaf,0x1a80e132}, {0x051bebb0,0x020b2b13,0x1ae861b5},
{0x052fbf29,0x02131944,0x1b50cc84}, {0x0543bf3a,0x021b194b,0x1bba2235},
{0x0557ec02,0x02232b34,0x1c24635f}, {0x056c459b,0x022b4f0b,0x1c8f9097},
{0x0580cc22,0x023384db,0x1cfbaa71}, {0x05957fb5,0x023bccaf,0x1d68b183},
{0x05aa606d,0x02442692,0x1dd6a660}, {0x05bf6e68,0x024c9290,0x1e45899b},
{0x05d4a9c2,0x025510b4,0x1eb55bc7}, {0x05ea1295,0x025da109,0x1f261d76},
{0x05ffa8ff,0x02664399,0x1f97cf3a}, {0x06156d1a,0x026ef871,0x200a71a5},
{0x062b5f01,0x0277bf9a,0x207e0546}, {0x06417ed1,0x02809920,0x20f28aaf},
{0x0657cca4,0x0289850f,0x2168026e}, {0x066e4896,0x0292836f,0x21de6d13},
{0x0684f2c2,0x029b944e,0x2255cb2c}, {0x069bcb43,0x02a4b7b4,0x22ce1d48},
{0x06b2d233,0x02adedae,0x234763f5}, {0x06ca07ae,0x02b73646,0x23c19fc0},
{0x06e16bce,0x02c09186,0x243cd134}, {0x06f8feae,0x02c9ff79,0x24b8f8e0},
{0x0710c068,0x02d3802a,0x2536174e}, {0x0728b117,0x02dd13a3,0x25b42d0a},
{0x0740d0d6,0x02e6b9ef,0x26333a9f}, {0x07591fbe,0x02f07319,0x26b34097},
{0x07719de9,0x02fa3f2a,0x27343f7c}, {0x078a4b73,0x03041e2e,0x27b637d8},
{0x07a32874,0x030e102e,0x28392a34}, {0x07bc3507,0x03181536,0x28bd1718},
{0x07d57146,0x03222d4f,0x2941ff0e}, {0x07eedd4a,0x032c5884,0x29c7e29b},
{0x0808792e,0x033696df,0x2a4ec249}, {0x0822450a,0x0340e86b,0x2ad69e9d},
{0x083c40f9,0x034b4d30,0x2b5f781f}, {0x08566d13,0x0355c53b,0x2be94f54},
{0x0870c973,0x03605094,0x2c7424c3}, {0x088b5631,0x036aef47,0x2cfff8f0},
{0x08a61367,0x0375a15c,0x2d8ccc60}, {0x08c1012e,0x038066df,0x2e1a9f98},
{0x08dc1f9e,0x038b3fd9,0x2ea9731c}, {0x08f76ed2,0x03962c54,0x2f39476f},
{0x0912eee1,0x03a12c5a,0x2fca1d16}, {0x092e9fe6,0x03ac3ff5,0x305bf491},
{0x094a81f7,0x03b76730,0x30eece65}, {0x0966952f,0x03c2a213,0x3182ab14},
{0x0982d9a6,0x03cdf0a9,0x32178b1e}, {0x099f4f74,0x03d952fb,0x32ad6f06},
{0x09bbf6b2,0x03e4c914,0x3344574c}, {0x09d8cf78,0x03f052fd,0x33dc4471},
{0x09f5d9df,0x03fbf0c0,0x347536f5}, {0x0a1315ff,0x0407a266,0x350f2f58},
{0x0a3083f0,0x041367fa,0x35aa2e19}, {0x0a4e23cb,0x041f4184,0x364633b9},
{0x0a6bf5a6,0x042b2f0f,0x36e340b4}, {0x0a89f99b,0x043730a5,0x3781558b},
{0x0aa82fc2,0x0443464d,0x382072ba}, {0x0ac69831,0x044f7014,0x38c098c0},
{0x0ae53302,0x045bae01,0x3961c81a}, {0x0b04004b,0x0468001e,0x3a040145},
{0x0b230024,0x04746675,0x3aa744be}, {0x0b4232a6,0x0480e10f,0x3b4b9301},
{0x0b6197e7,0x048d6ff6,0x3bf0ec8a}, {0x0b812fff,0x049a1333,0x3c9751d4},
{0x0ba0fb05,0x04a6cacf,0x3d3ec35b}, {0x0bc0f911,0x04b396d4,0x3de7419b},
{0x0be12a3b,0x04c0774b,0x3e90cd0d}, {0x0c018e98,0x04cd6c3d,0x3f3b662c},
{0x0c222641,0x04da75b4,0x3fe70d72}, {0x0c42f14d,0x04e793b8,0x4093c359},
{0x0c63efd2,0x04f4c654,0x4141885a}, {0x0c8521e8,0x05020d90,0x41f05ced},
{0x0ca687a5,0x050f6975,0x42a0418d}, {0x0cc82121,0x051cda0d,0x435136b1},
{0x0ce9ee71,0x052a5f61,0x44033cd2}, {0x0d0befae,0x0537f979,0x44b65467},
{0x0d2e24ee,0x0545a85f,0x456a7de7}, {0x0d508e46,0x05536c1c,0x461fb9ca},
{0x0d732bcf,0x056144b9,0x46d60888}, {0x0d95fd9e,0x056f323f,0x478d6a95},
{0x0db903c9,0x057d34b7,0x4845e069}, {0x0ddc3e68,0x058b4c2a,0x48ff6a7a},
{0x0dffad91,0x059978a0,0x49ba093d}, {0x0e235159,0x05a7ba24,0x4a75bd29},
{0x0e4729d8,0x05b610bd,0x4b3286b1}, {0x0e6b3722,0x05c47c74,0x4bf0664a},
{0x0e8f794f,0x05d2fd53,0x4caf5c6a}, {0x0eb3f075,0x05e19362,0x4d6f6985},
{0x0ed89ca9,0x05f03eaa,0x4e308e0e}, {0x0efd7e02,0x05feff34,0x4ef2ca79},
{0x0f229495,0x060dd508,0x4fb61f3a}, {0x0f47e078,0x061cc030,0x507a8cc4},
{0x0f6d61c1,0x062bc0b4,0x51401389}, {0x0f931886,0x063ad69c,0x5206b3fd},
{0x0fb904dd,0x064a01f2,0x52ce6e91}, {0x0fdf26db,0x065942be,0x539743b7},
{0x10057e95,0x06689909,0x546133e2}, {0x102c0c22,0x067804da,0x552c3f83},
{0x1052cf97,0x0687863c,0x55f8670b}, {0x1079c909,0x06971d37,0x56c5aaeb},
{0x10a0f88e,0x06a6c9d2,0x57940b94}, {0x10c85e3a,0x06b68c17,0x58638976},
{0x10effa24,0x06c6640f,0x59342501}, {0x1117cc61,0x06d651c0,0x5a05dea6},
{0x113fd505,0x06e65535,0x5ad8b6d4}, {0x11681427,0x06f66e76,0x5bacadfa},
{0x119089da,0x07069d8a,0x5c81c488}, {0x11b93635,0x0716e27b,0x5d57faec},
{0x11e2194b,0x07273d51,0x5e2f5196}, {0x120b3333,0x0737ae14,0x5f07c8f3},
};
// clang-format on
#endif
16 years ago
#ifdef USE_FLOATS
static void lab(double r, double g, double b, double *L, double *A, double *B)
{
double x, y, z;
r /= 255.0f;
g /= 255.0f;
b /= 255.0f;
if (r > 0.04045f)
r = pow(((r + 0.055f) / 1.055f), 2.4f);
else
r /= 12.92f;
if (g > 0.04045f)
g = pow(((g + 0.055f) / 1.055f), 2.4f);
else
g /= 12.92f;
if (b > 0.04045f)
b = pow(((b + 0.055f) / 1.055f), 2.4f);
else
b /= 12.92f;
r *= 100.0f;
g *= 100.0f;
b *= 100.0f;
x = r * 0.4124f + g * 0.3576f + b * 0.1805f;
y = r * 0.2126f + g * 0.7152f + b * 0.0722f;
z = r * 0.0193f + g * 0.1192f + b * 0.9505f;
x /= 95.047f;
y /= 100.000f;
z /= 108.883f;
if (x > 0.008856f)
x = pow(x, 1.0f / 3.0f);
else
x = (7.787f * x) + (16.0f / 116.0f);
if (y > 0.008856f)
y = pow(y, (1.0f / 3.0f));
else
y = (7.787f * y) + (16.0f / 116.0f);
if (z > 0.008856f)
z = pow(z, (1.0f / 3.0f));
else
z = (7.787f * z) + (16.0f / 116.0f);
*L = (116.0f * y) - 16.0f;
*A = 500.0f * (x - y);
*B = 200.0f * (y - z);
}
static double labdiff(unsigned int rgb)
{
unsigned int r, g, b;
const double L1 = 53.192777691077211f, A1 = 0.0031420942181448197f, B1 = -0.0062075877844014471f;
double L2, A2, B2;
r = (rgb >> 16) & 0xff;
g = (rgb >> 8) & 0xff;
b = rgb & 0xff;
lab(r, g, b, &L2, &A2, &B2);
return sqrt(pow(L1 - L2, 2.0f) + pow(A1 - A2, 2.0f) + pow(B1 - B2, 2.0f));
}
16 years ago
#else
static void lab2(uint32_t r, uint32_t g, uint32_t b, int32_t *L, int32_t *A, int32_t *B)
{
uint32_t xx, yy, zz;
xx = rtable[r][0] + gtable[g][0] + btable[b][0];
yy = rtable[r][1] + gtable[g][1] + btable[b][1];
zz = rtable[r][2] + gtable[g][2] + btable[b][2];
if (xx > 148587) {
xx = (1 << 24) * pow(xx / (95.047 * (1 << 24)), 1.0 / 3.0);
} else {
xx = xx * 24389 / 3132 + 2314099;
}
if (yy > 148587) {
yy = (1 << 24) * pow(yy / (100.0 * (1 << 24)), 1.0 / 3.0);
} else {
yy = yy * 24389 / 3132 + 2314099;
}
if (zz > 148587) {
zz = (1 << 24) * pow(zz / (108.883 * (1 << 24)), 1.0 / 3.0);
} else {
zz = zz * 24389 / 3132 + 2314099;
}
*L = (116 * yy - 116 * 2314099);
*A = 500 / 4 * (xx - yy);
*B = 200 / 4 * (yy - zz); /* /4 to avoid overflow */
}
static uint32_t labdiff2(unsigned int b)
{
unsigned int r2, g2, b2;
int32_t L1, A1, B1, L2, A2, B2;
int64_t ld, ad, bd;
r2 = (b >> 16) & 0xff;
g2 = (b >> 8) & 0xff;
b2 = b & 0xff;
/* ref = 0x7f7f7f -> L*a*b ~ (53,0,0) */
L1 = 53 * (1 << 24);
A1 = 0;
B1 = 0;
lab2(r2, g2, b2, &L2, &A2, &B2);
ld = L1 - L2;
ld *= ld;
ad = A1 - A2;
ad *= ad;
bd = B1 - B2;
bd *= bd;
ld += ad + bd;
return ((uint32_t)(sqrt(ld / 1024.0))) >> 17;
}
16 years ago
#endif
static void makebmp(const char *step, const char *tempd, int w, int h, void *data)
{
unsigned int tmp1 = 0, tmp2 = 0, tmp3 = 0, tmp4 = 0, y;
char *fname;
FILE *f;
if (!tempd)
return;
Improve tmp sub-directory names At present many parsers create tmp subdirectories to store extracted files. For parsers like the vba parser, this is required as the directory is later scanned. For other parsers, these subdirectories are probably not helpful now that we provide recursive sub-dirs when --leave-temps is enabled. It's not quite as simple as removing the extra subdirectories, however. Certain parsers, like autoit, don't create very unique filenames and would result in file name collisions when --leave-temps is not enabled. The best thing to do would be to make sure each parser uses unique filenames and doesn't rely on cli_magic_scan_dir() to scan extracted content before removing the extra subdirectory. In the meantime, this commit gives the extra subdirectories meaningful names to improve readability. This commit also: - Provides the 'bmp' prefix for extracted PE icons. - Removes empty tmp subdirs when extracting rtf files, to eliminate clutter. - The PDF parser sometimes creates tmp files when decompressing streams before it knows if there is actually any content to decompress. This resulted in a large number of empty files. While it would be best to avoid creating empty files in the first place, that's not quite as as it sounds. This commit does the next best thing and deletes the tmp files if nothing was actually extracted, even if --leave-temps is enabled. - Removes the "scantemp" prefix for unnamed fmaps scanned with cli_magic_scan(). The 5-character hashes given to tmp files with prefixes resulted in occasional file name collisions when extracting certain file types with thousands of embedded files. - The VBA and TAR parsers mistakenly used NAME_MAX instead of PATH_MAX, resulting in truncated file paths and failed extraction when --leave-temps is enabled and a lot of recursion is in play. This commit switches them from NAME_MAX to PATH_MAX.
5 years ago
if (!(fname = cli_gentemp_with_prefix(tempd, "bmp")))
return;
if (!(f = fopen(fname, "wb"))) {
cli_unlink(fname);
cli_dbgmsg("makebmp: failed to create file %s\n", fname);
free(fname);
return;
}
cli_writeint32(&tmp1, 0x28 + 0xe + w * h * 4);
cli_writeint32(&tmp2, (32 << 16) | 1);
tmp3 = 0;
cli_writeint32(&tmp4, w * h * 4);
if (!fwrite("BM", 2, 1, f) ||
!fwrite(&tmp1, 4, 1, f) ||
!fwrite("aCaB\x36\x00\x00\x00\x28\x00\x00\x00", 12, 1, f) ||
!fwrite(&w, 4, 1, f) ||
!fwrite(&h, 4, 1, f) ||
!fwrite(&tmp2, 4, 1, f) ||
!fwrite(&tmp3, 4, 1, f) ||
!fwrite(&tmp4, 4, 1, f) ||
!fwrite("\1\0\0\0\1\0\0\0\0\0\0\0\0\0\0\0", 16, 1, f)) {
fclose(f);
cli_unlink(fname);
free(fname);
cli_dbgmsg("makebmp: failed to write output\n");
return;
}
16 years ago
for (y = h - 1; y < (unsigned int)h; y--)
#if WORDS_BIGENDIAN == 0
if (!fwrite(&((unsigned int *)data)[y * w], w * 4, 1, f))
break;
#else
{
int x;
for (x = 0; x < w; x++) {
cli_writeint32(&tmp1, ((unsigned int *)data)[y * w]);
if (!fwrite(&tmp1, 4, 1, f))
break;
}
if (x != w)
break;
16 years ago
}
#endif
fclose(f);
if (y < (unsigned int)h)
cli_unlink(fname);
else
cli_dbgmsg("makebmp: Image %s dumped to %s\n", step, fname);
free(fname);
}
16 years ago
static unsigned int matchpoint(unsigned int side, unsigned int *x1, unsigned int *y1, unsigned int *avg1, const unsigned int *x2, const unsigned int *y2, const unsigned int *avg2, unsigned int max)
{
unsigned int i, j, best, match = 0, ksize = side / 4;
for (i = 0; i < 3; i++) {
best = 0;
for (j = 0; j < 3; j++) {
/* approximately measure the distance from the best matching reference - avoid N*N total war */
int diffx = (int)x1[i] - (int)x2[j];
int diffy = ((int)y1[i] - (int)y2[j]);
unsigned int diff = sqrt(diffx * diffx + diffy * diffy);
if (diff > ksize * 3 / 4 || (unsigned int)abs((int)avg1[i] - (int)avg2[j]) > max / 5)
continue;
diff = 100 - diff * 60 / (ksize * 3 / 4);
if (diff > best)
best = diff;
}
match += best;
16 years ago
}
return match / 3;
}
16 years ago
static unsigned int matchbwpoint(unsigned int side, unsigned int *x1a, unsigned int *y1a, unsigned int *avg1a, unsigned int *x1b, unsigned int *y1b, unsigned int *avg1b, const unsigned int *x2a, const unsigned int *y2a, const unsigned int *avg2a, const unsigned int *x2b, const unsigned int *y2b, const unsigned int *avg2b)
{
unsigned int i, j, best, match = 0, ksize = side / 4;
unsigned int x1[6], y1[6], avg1[6], x2[6], y2[6], avg2[6];
for (i = 0; i < 3; i++) {
x1[i] = x1a[i];
y1[i] = y1a[i];
avg1[i] = avg1a[i];
x2[i] = x2a[i];
y2[i] = y2a[i];
avg2[i] = avg2a[i];
x1[i + 3] = x1b[i];
y1[i + 3] = y1b[i];
avg1[i + 3] = avg1b[i];
x2[i + 3] = x2b[i];
y2[i + 3] = y2b[i];
avg2[i + 3] = avg2b[i];
}
16 years ago
for (i = 0; i < 6; i++) {
best = 0;
for (j = 0; j < 6; j++) {
/* approximately measure the distance from the best matching reference - avoid N*N total war */
int diffx = (int)x1[i] - (int)x2[j];
int diffy = ((int)y1[i] - (int)y2[j]);
unsigned int diff = sqrt(diffx * diffx + diffy * diffy);
if (diff > ksize * 3 / 4 || (unsigned int)abs((int)avg1[i] - (int)avg2[j]) > 255 / 5)
continue;
diff = 100 - diff * 60 / (ksize * 3 / 4);
if (diff > best)
best = diff;
}
match += best;
}
return match / 6;
}
16 years ago
static void hsv(unsigned int c, unsigned int *r, unsigned int *g, unsigned int *b, unsigned int *s, unsigned int *v, unsigned int *delta)
{
unsigned int min, max;
*r = (c >> 16) & 0xff;
*g = (c >> 8) & 0xff;
*b = c & 0xff;
min = MIN(*r, MIN(*g, *b));
max = MAX(*r, MAX(*g, *b));
*v = max;
*delta = max - min;
if (!*delta)
*s = 0;
else
*s = 255 * (*delta) / max;
}
static int getmetrics(unsigned int side, unsigned int *imagedata, struct icomtr *res, const char *tempd)
{
unsigned int x, y, xk, yk, i, j, *tmp;
unsigned int ksize = side / 4, bwonly = 0;
unsigned int edge_avg[6], edge_x[6] = {0, 0, 0, 0, 0, 0}, edge_y[6] = {0, 0, 0, 0, 0, 0}, noedge_avg[6], noedge_x[6] = {0, 0, 0, 0, 0, 0}, noedge_y[6] = {0, 0, 0, 0, 0, 0};
double *sobel;
if (!(tmp = cli_malloc(side * side * 4 * 2))) {
cli_errmsg("getmetrics: Unable to allocate memory for tmp %u\n", (side * side * 4 * 2));
return CL_EMEM;
16 years ago
}
memset(res, 0, sizeof(*res));
16 years ago
/* compute colored, gray, bright and dark areas, color presence */
for (y = 0; y <= side - ksize; y++) {
for (x = 0; x <= side - ksize; x++) {
unsigned int colsum = 0, lightsum = 0;
unsigned int r, g, b, s, v, delta;
if (x == 0 && y == 0) {
/* Here we handle the 1st window which is fully calculated */
for (yk = 0; yk < ksize; yk++) {
for (xk = 0; xk < ksize; xk++) {
hsv(imagedata[yk * side + xk], &r, &g, &b, &s, &v, &delta);
colsum += (unsigned int)sqrt(s * s * v);
lightsum += v;
/* count colors (full square) */
if (s > 85 && v > 85) {
res->ccount++;
res->rsum += 100 - 100 * abs((int)g - (int)b) / delta;
res->gsum += 100 - 100 * abs((int)r - (int)b) / delta;
res->bsum += 100 - 100 * abs((int)r - (int)g) / delta;
}
}
}
} else if (x) { /* Here we incrementally calculate rows and columns
code is split as gcc produces faster code this way */
colsum = tmp[y * side + x - 1];
lightsum = tmp[side * side + y * side + x - 1];
for (yk = 0; yk < ksize; yk++) {
/* remove previous column */
hsv(imagedata[(y + yk) * side + x - 1], &r, &g, &b, &s, &v, &delta);
colsum -= (unsigned int)sqrt(s * s * v);
lightsum -= v;
/* add next column */
hsv(imagedata[(y + yk) * side + x + ksize - 1], &r, &g, &b, &s, &v, &delta);
colsum += (unsigned int)sqrt(s * s * v);
lightsum += v;
/* count colors (full column or only the last px) */
if ((y == 0 || yk == ksize - 1) && s > 85 && v > 85) {
res->ccount++;
res->rsum += 100 - 100 * abs((int)g - (int)b) / delta;
res->gsum += 100 - 100 * abs((int)r - (int)b) / delta;
res->bsum += 100 - 100 * abs((int)r - (int)g) / delta;
}
}
} else {
colsum = tmp[(y - 1) * side];
lightsum = tmp[side * side + (y - 1) * side];
for (xk = 0; xk < ksize; xk++) {
/* remove previous row */
hsv(imagedata[(y - 1) * side + xk], &r, &g, &b, &s, &v, &delta);
colsum -= (unsigned int)sqrt(s * s * v);
lightsum -= v;
/* add next row */
hsv(imagedata[(y + ksize - 1) * side + xk], &r, &g, &b, &s, &v, &delta);
colsum += (unsigned int)sqrt(s * s * v);
lightsum += v;
/* count colors (full row) */
if (s > 85 && v > 85) {
res->ccount++;
res->rsum += 100 - 100 * abs((int)g - (int)b) / delta;
res->gsum += 100 - 100 * abs((int)r - (int)b) / delta;
res->bsum += 100 - 100 * abs((int)r - (int)g) / delta;
}
}
}
tmp[y * side + x] = colsum;
tmp[side * side + y * side + x] = lightsum;
}
}
16 years ago
/* extract top 3 non overlapping areas for: colored, gray, bright and dark areas, color presence */
for (i = 0; i < 3; i++) {
res->gray_avg[i] = 0xffffffff;
res->dark_avg[i] = 0xffffffff;
for (y = 0; y < side - ksize; y++) {
for (x = 0; x < side - 1 - ksize; x++) {
unsigned int colsum = tmp[y * side + x], lightsum = tmp[side * side + y * side + x];
if (colsum > res->color_avg[i]) {
for (j = 0; j < i; j++) {
if (x + ksize > res->color_x[j] && x < res->color_x[j] + ksize &&
y + ksize > res->color_y[j] && y < res->color_y[j] + ksize)
break;
}
if (j == i) {
res->color_avg[i] = colsum;
res->color_x[i] = x;
res->color_y[i] = y;
}
}
if (colsum < res->gray_avg[i]) {
for (j = 0; j < i; j++) {
if (x + ksize > res->gray_x[j] && x < res->gray_x[j] + ksize &&
y + ksize > res->gray_y[j] && y < res->gray_y[j] + ksize)
break;
}
if (j == i) {
res->gray_avg[i] = colsum;
res->gray_x[i] = x;
res->gray_y[i] = y;
}
}
if (lightsum > res->bright_avg[i]) {
for (j = 0; j < i; j++) {
if (x + ksize > res->bright_x[j] && x < res->bright_x[j] + ksize &&
y + ksize > res->bright_y[j] && y < res->bright_y[j] + ksize)
break;
}
if (j == i) {
res->bright_avg[i] = lightsum;
res->bright_x[i] = x;
res->bright_y[i] = y;
}
}
if (lightsum < res->dark_avg[i]) {
for (j = 0; j < i; j++) {
if (x + ksize > res->dark_x[j] && x < res->dark_x[j] + ksize &&
y + ksize > res->dark_y[j] && y < res->dark_y[j] + ksize)
break;
}
if (j == i) {
res->dark_avg[i] = lightsum;
res->dark_x[i] = x;
res->dark_y[i] = y;
}
}
}
}
}
16 years ago
/* abs->avg */
for (i = 0; i < 3; i++) {
res->color_avg[i] /= ksize * ksize;
res->gray_avg[i] /= ksize * ksize;
res->bright_avg[i] /= ksize * ksize;
res->dark_avg[i] /= ksize * ksize;
}
16 years ago
if (res->ccount * 100 / side / side > 5) {
res->rsum /= res->ccount;
res->gsum /= res->ccount;
res->bsum /= res->ccount;
res->ccount = res->ccount * 100 / side / side;
} else {
res->ccount = 0;
res->rsum = 0;
res->gsum = 0;
res->bsum = 0;
bwonly = 1;
}
/* Edge detection - Sobel */
16 years ago
/* Sobel 1 - gradients */
i = 0;
16 years ago
#ifdef USE_FLOATS
sobel = cli_malloc(side * side * sizeof(double));
if (!sobel) {
cli_errmsg("getmetrics: Unable to allocate memory for edge detection %llu\n", (long long unsigned)(side * side * sizeof(double)));
free(tmp);
return CL_EMEM;
}
16 years ago
#else
#define sobel imagedata
16 years ago
#endif
for (y = 0; y < side; y++) {
for (x = 0; x < side; x++) {
sobel[y * side + x] = LABDIFF(imagedata[y * side + x]);
}
}
for (y = 1; y < side - 1; y++) {
for (x = 1; x < side - 1; x++) {
unsigned int sob;
16 years ago
#ifdef USE_FLOATS
double gx, gy;
16 years ago
#else
unsigned int gx, gy;
16 years ago
#endif
/* X matrix */
gx = sobel[(y - 1) * side + (x - 1)];
gx += sobel[(y + 0) * side + (x - 1)] * 2;
gx += sobel[(y + 1) * side + (x - 1)];
gx -= sobel[(y - 1) * side + (x + 1)];
gx -= sobel[(y + 0) * side + (x + 1)] * 2;
gx -= sobel[(y + 1) * side + (x + 1)];
/* Y matrix */
gy = sobel[(y - 1) * side + (x - 1)];
gy += sobel[(y - 1) * side + (x + 0)] * 2;
gy += sobel[(y - 1) * side + (x + 1)];
gy -= sobel[(y + 1) * side + (x - 1)];
gy -= sobel[(y + 1) * side + (x + 0)] * 2;
gy -= sobel[(y + 1) * side + (x + 1)];
sob = (int)sqrt(gx * gx + gy * gy);
tmp[y * side + x] = sob;
if (sob > i)
i = sob;
}
}
16 years ago
#ifdef USE_FLOATS
free(sobel);
16 years ago
#endif
16 years ago
/* Sobel 2 - norm to max */
if (i) {
for (y = 1; y < side - 1; y++) {
for (x = 1; x < side - 1; x++) {
unsigned int c = tmp[y * side + x];
c = c * 255 / i;
imagedata[y * side + x] = 0xff000000 | c | (c << 8) | (c << 16);
}
}
}
16 years ago
/* black borders */
for (x = 0; x < side; x++) {
imagedata[x] = 0xff000000;
imagedata[(side - 1) * side + x] = 0xff000000;
}
for (y = 0; y < side; y++) {
imagedata[y * side] = 0xff000000;
imagedata[y * side + side - 1] = 0xff000000;
}
makebmp("3-edge", tempd, side, side, imagedata);
/* gaussian blur */
for (y = 1; y < side - 1; y++) {
for (x = 1; x < side - 1; x++) {
unsigned int sum = 0, tot = 0;
int disp;
for (disp = -MIN((int)x, gkernsz / 2); disp <= MIN((int)(side - 1 - x), gkernsz / 2); disp++) {
unsigned int c = imagedata[y * side + x + disp] & 0xff;
sum += c * gaussk[disp + gkernsz / 2];
tot += gaussk[disp + gkernsz / 2];
}
sum /= tot;
imagedata[y * side + x] &= 0xff;
imagedata[y * side + x] |= sum << 8;
}
}
i = 0;
for (y = 1; y < side - 1; y++) {
for (x = 1; x < side - 1; x++) {
unsigned int sum = 0, tot = 0;
int disp;
for (disp = -MIN((int)y, gkernsz / 2); disp <= MIN((int)(side - 1 - y), gkernsz / 2); disp++) {
unsigned int c = (imagedata[(y + disp) * side + x] >> 8) & 0xff;
sum += c * gaussk[disp + gkernsz / 2];
tot += gaussk[disp + gkernsz / 2];
}
sum /= tot;
if (sum > i)
i = sum;
imagedata[y * side + x] = 0xff000000 | sum | (sum << 8) | (sum << 16);
}
}
makebmp("4-gauss", tempd, side, side, imagedata);
/* calculate edges */
for (y = 0; y <= side - ksize; y++) {
for (x = 0; x <= side - 1 - ksize; x++) {
unsigned int sum = 0;
if (x == 0 && y == 0) { /* 1st windows */
for (yk = 0; yk < ksize; yk++) {
for (xk = 0; xk < ksize; xk++)
sum += imagedata[(y + yk) * side + x + xk] & 0xff;
}
} else if (x) { /* next column */
sum = tmp[y * side + x - 1];
for (yk = 0; yk < ksize; yk++) {
sum -= imagedata[(y + yk) * side + x - 1] & 0xff;
sum += imagedata[(y + yk) * side + x + ksize - 1] & 0xff;
}
} else { /* next row */
sum = tmp[(y - 1) * side];
for (xk = 0; xk < ksize; xk++) {
sum -= imagedata[(y - 1) * side + xk] & 0xff;
sum += imagedata[(y + ksize - 1) * side + xk] & 0xff;
}
}
tmp[y * side + x] = sum;
}
}
16 years ago
/* calculate best and worst 3 (or 6) edged areas */
for (i = 0; i < 3 * (bwonly + 1); i++) {
edge_avg[i] = 0;
noedge_avg[i] = 0xffffffff;
for (y = 0; y < side - ksize; y++) {
for (x = 0; x < side - 1 - ksize; x++) {
unsigned int sum = tmp[y * side + x];
if (sum > edge_avg[i]) {
for (j = 0; j < i; j++) {
if (x + ksize > edge_x[j] && x < edge_x[j] + ksize &&
y + ksize > edge_y[j] && y < edge_y[j] + ksize)
break;
}
if (j == i) {
edge_avg[i] = sum;
edge_x[i] = x;
edge_y[i] = y;
}
}
if (sum < noedge_avg[i]) {
for (j = 0; j < i; j++) {
if (x + ksize > noedge_x[j] && x < noedge_x[j] + ksize &&
y + ksize > noedge_y[j] && y < noedge_y[j] + ksize)
break;
}
if (j == i) {
noedge_avg[i] = sum;
noedge_x[i] = x;
noedge_y[i] = y;
}
}
}
}
}
16 years ago
free(tmp);
/* abs->avg */
for (i = 0; i < 3; i++) {
res->edge_avg[i] = edge_avg[i] / ksize / ksize;
res->edge_x[i] = edge_x[i];
res->edge_y[i] = edge_y[i];
res->noedge_avg[i] = noedge_avg[i] / ksize / ksize;
res->noedge_x[i] = noedge_x[i];
res->noedge_y[i] = noedge_y[i];
}
if (bwonly) {
for (i = 0; i < 3; i++) {
res->color_avg[i] = edge_avg[i + 3] / ksize / ksize;
res->color_x[i] = edge_x[i + 3];
res->color_y[i] = edge_y[i + 3];
res->gray_avg[i] = noedge_avg[i + 3] / ksize / ksize;
res->gray_x[i] = edge_x[i + 3];
res->gray_y[i] = edge_y[i + 3];
}
}
16 years ago
cli_dbgmsg("edge areas: %u@(%u,%u) %u@(%u,%u) %u@(%u,%u)\n", res->edge_avg[0], res->edge_x[0], res->edge_y[0], res->edge_avg[1], res->edge_x[1], res->edge_y[1], res->edge_avg[2], res->edge_x[2], res->edge_y[2]);
cli_dbgmsg("noedge areas: %u@(%u,%u) %u@(%u,%u) %u@(%u,%u)\n", res->noedge_avg[0], res->noedge_x[0], res->noedge_y[0], res->noedge_avg[1], res->noedge_x[1], res->noedge_y[1], res->noedge_avg[2], res->noedge_x[2], res->noedge_y[2]);
cli_dbgmsg("%s areas: %u@(%u,%u) %u@(%u,%u) %u@(%u,%u)\n", bwonly ? "edge(2nd)" : "color", res->color_avg[0], res->color_x[0], res->color_y[0], res->color_avg[1], res->color_x[1], res->color_y[1], res->color_avg[2], res->color_x[2], res->color_y[2]);
cli_dbgmsg("%s areas: %u@(%u,%u) %u@(%u,%u) %u@(%u,%u)\n", bwonly ? "noedge(2nd)" : "gray", res->gray_avg[0], res->gray_x[0], res->gray_y[0], res->gray_avg[1], res->gray_x[1], res->gray_y[1], res->gray_avg[2], res->gray_x[2], res->gray_y[2]);
cli_dbgmsg("bright areas: %u@(%u,%u) %u@(%u,%u) %u@(%u,%u)\n", res->bright_avg[0], res->bright_x[0], res->bright_y[0], res->bright_avg[1], res->bright_x[1], res->bright_y[1], res->bright_avg[2], res->bright_x[2], res->bright_y[2]);
cli_dbgmsg("dark areas: %u@(%u,%u) %u@(%u,%u) %u@(%u,%u)\n", res->dark_avg[0], res->dark_x[0], res->dark_y[0], res->dark_avg[1], res->dark_x[1], res->dark_y[1], res->dark_avg[2], res->dark_x[2], res->dark_y[2]);
if (!bwonly)
cli_dbgmsg("color spread: %u,%u,%u %u%%\n", res->rsum, res->gsum, res->bsum, res->ccount);
if (cli_debug_flag) {
#define ICOSIGSZ (2 + (3 + 2 + 2) * 3 * 2 + (2 + 2 + 2) * 3 * 4 + 2 + 2 + 2 + 2)
char mstr[ICOSIGSZ + 1], *ptr = mstr;
sprintf(ptr, "%02x", side);
ptr += 2;
for (i = 0; i < 3; i++) {
sprintf(ptr, "%03x", res->color_avg[i]);
ptr += 3;
sprintf(ptr, "%02x", res->color_x[i]);
ptr += 2;
sprintf(ptr, "%02x", res->color_y[i]);
ptr += 2;
}
for (i = 0; i < 3; i++) {
sprintf(ptr, "%03x", res->gray_avg[i]);
ptr += 3;
sprintf(ptr, "%02x", res->gray_x[i]);
ptr += 2;
sprintf(ptr, "%02x", res->gray_y[i]);
ptr += 2;
}
for (i = 0; i < 3; i++) {
sprintf(ptr, "%02x", res->bright_avg[i]);
ptr += 2;
sprintf(ptr, "%02x", res->bright_x[i]);
ptr += 2;
sprintf(ptr, "%02x", res->bright_y[i]);
ptr += 2;
}
for (i = 0; i < 3; i++) {
sprintf(ptr, "%02x", res->dark_avg[i]);
ptr += 2;
sprintf(ptr, "%02x", res->dark_x[i]);
ptr += 2;
sprintf(ptr, "%02x", res->dark_y[i]);
ptr += 2;
}
for (i = 0; i < 3; i++) {
sprintf(ptr, "%02x", res->edge_avg[i]);
ptr += 2;
sprintf(ptr, "%02x", res->edge_x[i]);
ptr += 2;
sprintf(ptr, "%02x", res->edge_y[i]);
ptr += 2;
}
for (i = 0; i < 3; i++) {
sprintf(ptr, "%02x", res->noedge_avg[i]);
ptr += 2;
sprintf(ptr, "%02x", res->noedge_x[i]);
ptr += 2;
sprintf(ptr, "%02x", res->noedge_y[i]);
ptr += 2;
}
sprintf(ptr, "%02x", res->rsum);
ptr += 2;
sprintf(ptr, "%02x", res->gsum);
ptr += 2;
sprintf(ptr, "%02x", res->bsum);
ptr += 2;
sprintf(ptr, "%02x", res->ccount);
cli_dbgmsg("IDB SIGNATURE: ICON_NAME:GROUP1:GROUP2:%s\n", mstr);
}
16 years ago
return CL_CLEAN;
}
static int parseicon(struct ICON_ENV *icon_env, uint32_t rva)
{
6 years ago
icon_groupset *set = icon_env->set;
cli_ctx *ctx = icon_env->ctx;
PE parsing code improvements, db loading bug fixes Consolidate the PE parsing code into one function. I tried to preserve all existing functionality from the previous, distinct implementations to a large extent (with the exceptions mentioned below). If I noticed potential bugs/improvements, I added a TODO statement about those so that they can be fixed in a smaller commit later. Also, there are more TODOs in places where I'm not entirely sure why certain actions are performed - more research is needed for these. I'm submitting a pull request now so that regression testing can be done, and because merging what I have thus far now will likely have fewer conflicts than if I try to merge later PE parsing code improvements: - PEs without all 16 data directories are parsed more appropriately now - Added lots more debug statements Also: - Allow MAX_BC and MAX_TRACKED_PCRE to be specified via CFLAGS When doing performance testing with the latest CVD, MAX_BC and MAX_TRACKED_PCRE need to be raised to track all the events. Allow these to be specified via CFLAGS by not redefining them if they are already defined - Fix an issue preventing wildcard sizes in .MDB/.MSB rules I'm not sure what the original intent of the check I removed was, but it prevents using wildcard sizes in .MDB/.MSB rules. AFAICT these wildcard sizes should be handled appropriately by the MD5 section hash computation code, so I don't think a check on that is needed. - Fix several issues related to db loading - .imp files will now get loaded if they exist in a directory passed via clamscan's '-d' flag - .pwdb files will now get loaded if they exist in a directory passed via clamscan's '-d' flag even when compiling without yara support - Changes to .imp, .ign, and .ign2 files will now be reflected in calls to cl_statinidir and cl_statchkdir (and also .pwdb files, even when compiling without yara support) - The contents of .sfp files won't be included in some of the signature counts, and the contents of .cud files will be - Any local.gdb files will no longer be loaded twice - For .imp files, you are no longer required to specify a minimum flevel for wildcard rules, since this isn't needed
6 years ago
struct cli_exe_info *peinfo = icon_env->peinfo;
struct
{
unsigned int sz;
unsigned int w;
unsigned int h;
unsigned short planes;
unsigned short depth;
unsigned int comp;
unsigned int imagesize;
unsigned int dpix;
unsigned int dpiy;
unsigned int used;
unsigned int important;
} bmphdr;
struct icomtr metrics;
const unsigned char *rawimage;
const char *tempd;
const uint32_t *palette = NULL;
uint32_t *imagedata;
unsigned int scanlinesz, andlinesz;
unsigned int width, height, depth, x, y;
unsigned int err, scalemode = 2, enginesize;
fmap_t *map;
uint32_t icoff;
struct icon_matcher *matcher;
unsigned int special_32_is_32 = 0;
if (!ctx || !ctx->engine || !(matcher = ctx->engine->iconcheck))
return CL_SUCCESS;
libclamav: Fix scan recursion tracking Scan recursion is the process of identifying files embedded in other files and then scanning them, recursively. Internally this process is more complex than it may sound because a file may have multiple layers of types before finding a new "file". At present we treat the recursion count in the scanning context as an index into both our fmap list AND our container list. These two lists are conceptually a part of the same thing and should be unified. But what's concerning is that the "recursion level" isn't actually incremented or decremented at the same time that we add a layer to the fmap or container lists but instead is more touchy-feely, increasing when we find a new "file". To account for this shadiness, the size of the fmap and container lists has always been a little longer than our "max scan recursion" limit so we don't accidentally overflow the fmap or container arrays (!). I've implemented a single recursion-stack as an array, similar to before, which includes a pointer to each fmap at each layer, along with the size and type. Push and pop functions add and remove layers whenever a new fmap is added. A boolean argument when pushing indicates if the new layer represents a new buffer or new file (descriptor). A new buffer will reset the "nested fmap level" (described below). This commit also provides a solution for an issue where we detect embedded files more than once during scan recursion. For illustration, imagine a tarball named foo.tar.gz with this structure: | description | type | rec level | nested fmap level | | ------------------------- | ----- | --------- | ----------------- | | foo.tar.gz | GZ | 0 | 0 | | └── foo.tar | TAR | 1 | 0 | | ├── bar.zip | ZIP | 2 | 1 | | │   └── hola.txt | ASCII | 3 | 0 | | └── baz.exe | PE | 2 | 1 | But suppose baz.exe embeds a ZIP archive and a 7Z archive, like this: | description | type | rec level | nested fmap level | | ------------------------- | ----- | --------- | ----------------- | | baz.exe | PE | 0 | 0 | | ├── sfx.zip | ZIP | 1 | 1 | | │   └── hello.txt | ASCII | 2 | 0 | | └── sfx.7z | 7Z | 1 | 1 | |    └── world.txt | ASCII | 2 | 0 | (A) If we scan for embedded files at any layer, we may detect: | description | type | rec level | nested fmap level | | ------------------------- | ----- | --------- | ----------------- | | foo.tar.gz | GZ | 0 | 0 | | ├── foo.tar | TAR | 1 | 0 | | │ ├── bar.zip | ZIP | 2 | 1 | | │ │   └── hola.txt | ASCII | 3 | 0 | | │ ├── baz.exe | PE | 2 | 1 | | │ │ ├── sfx.zip | ZIP | 3 | 1 | | │ │ │   └── hello.txt | ASCII | 4 | 0 | | │ │ └── sfx.7z | 7Z | 3 | 1 | | │ │    └── world.txt | ASCII | 4 | 0 | | │ ├── sfx.zip | ZIP | 2 | 1 | | │ │   └── hello.txt | ASCII | 3 | 0 | | │ └── sfx.7z | 7Z | 2 | 1 | | │   └── world.txt | ASCII | 3 | 0 | | ├── sfx.zip | ZIP | 1 | 1 | | └── sfx.7z | 7Z | 1 | 1 | (A) is bad because it scans content more than once. Note that for the GZ layer, it may detect the ZIP and 7Z if the signature hits on the compressed data, which it might, though extracting the ZIP and 7Z will likely fail. The reason the above doesn't happen now is that we restrict embedded type scans for a bunch of archive formats to include GZ and TAR. (B) If we scan for embedded files at the foo.tar layer, we may detect: | description | type | rec level | nested fmap level | | ------------------------- | ----- | --------- | ----------------- | | foo.tar.gz | GZ | 0 | 0 | | └── foo.tar | TAR | 1 | 0 | | ├── bar.zip | ZIP | 2 | 1 | | │   └── hola.txt | ASCII | 3 | 0 | | ├── baz.exe | PE | 2 | 1 | | ├── sfx.zip | ZIP | 2 | 1 | | │   └── hello.txt | ASCII | 3 | 0 | | └── sfx.7z | 7Z | 2 | 1 | |    └── world.txt | ASCII | 3 | 0 | (B) is almost right. But we can achieve it easily enough only scanning for embedded content in the current fmap when the "nested fmap level" is 0. The upside is that it should safely detect all embedded content, even if it may think the sfz.zip and sfx.7z are in foo.tar instead of in baz.exe. The biggest risk I can think of affects ZIPs. SFXZIP detection is identical to ZIP detection, which is why we don't allow SFXZIP to be detected if insize of a ZIP. If we only allow embedded type scanning at fmap-layer 0 in each buffer, this will fail to detect the embedded ZIP if the bar.exe was not compressed in foo.zip and if non-compressed files extracted from ZIPs aren't extracted as new buffers: | description | type | rec level | nested fmap level | | ------------------------- | ----- | --------- | ----------------- | | foo.zip | ZIP | 0 | 0 | | └── bar.exe | PE | 1 | 1 | | └── sfx.zip | ZIP | 2 | 2 | Provided that we ensure all files extracted from zips are scanned in new buffers, option (B) should be safe. (C) If we scan for embedded files at the baz.exe layer, we may detect: | description | type | rec level | nested fmap level | | ------------------------- | ----- | --------- | ----------------- | | foo.tar.gz | GZ | 0 | 0 | | └── foo.tar | TAR | 1 | 0 | | ├── bar.zip | ZIP | 2 | 1 | | │   └── hola.txt | ASCII | 3 | 0 | | └── baz.exe | PE | 2 | 1 | | ├── sfx.zip | ZIP | 3 | 1 | | │   └── hello.txt | ASCII | 4 | 0 | | └── sfx.7z | 7Z | 3 | 1 | |    └── world.txt | ASCII | 4 | 0 | (C) is right. But it's harder to achieve. For this example we can get it by restricting 7ZSFX and ZIPSFX detection only when scanning an executable. But that may mean losing detection of archives embedded elsewhere. And we'd have to identify allowable container types for each possible embedded type, which would be very difficult. So this commit aims to solve the issue the (B)-way. Note that in all situations, we still have to scan with file typing enabled to determine if we need to reassign the current file type, such as re-identifying a Bzip2 archive as a DMG that happens to be Bzip2- compressed. Detection of DMG and a handful of other types rely on finding data partway through or near the ned of a file before reassigning the entire file as the new type. Other fixes and considerations in this commit: - The utf16 HTML parser has weak error handling, particularly with respect to creating a nested fmap for scanning the ascii decoded file. This commit cleans up the error handling and wraps the nested scan with the recursion-stack push()/pop() for correct recursion tracking. Before this commit, each container layer had a flag to indicate if the container layer is valid. We need something similar so that the cli_recursion_stack_get_*() functions ignore normalized layers. Details... Imagine an LDB signature for HTML content that specifies a ZIP container. If the signature actually alerts on the normalized HTML and you don't ignore normalized layers for the container check, it will appear as though the alert is in an HTML container rather than a ZIP container. This commit accomplishes this with a boolean you set in the scan context before scanning a new layer. Then when the new fmap is created, it will use that flag to set similar flag for the layer. The context flag is reset those that anything after this doesn't have that flag. The flag allows the new recursion_stack_get() function to ignore normalized layers when iterating the stack to return a layer at a requested index, negative or positive. Scanning normalized extracted/normalized javascript and VBA should also use the 'layer is normalized' flag. - This commit also fixes Heuristic.Broken.Executable alert for ELF files to make sure that: A) these only alert if cli_append_virus() returns CL_VIRUS (aka it respects the FP check). B) all broken-executable alerts for ELF only happen if the SCAN_HEURISTIC_BROKEN option is enabled. - This commit also cleans up the error handling in cli_magic_scan_dir(). This was needed so we could correctly apply the layer-is-normalized-flag to all VBA macros extracted to a directory when scanning the directory. - Also fix an issue where exceeding scan maximums wouldn't cause embedded file detection scans to abort. Granted we don't actually want to abort if max filesize or max recursion depth are exceeded... only if max scansize, max files, and max scantime are exceeded. Add 'abort_scan' flag to scan context, to protect against depending on correct error propagation for fatal conditions. Instead, setting this flag in the scan context should guarantee that a fatal condition deep in scan recursion isn't lost which result in more stuff being scanned instead of aborting. This shouldn't be necessary, but some status codes like CL_ETIMEOUT never used to be fatal and it's easier to do this than to verify every parser only returns CL_ETIMEOUT and other "fatal status codes" in fatal conditions. - Remove duplicate is_tar() prototype from filestypes.c and include is_tar.h instead. - Presently we create the fmap hash when creating the fmap. This wastes a bit of CPU if the hash is never needed. Now that we're creating fmap's for all embedded files discovered with file type recognition scans, this is a much more frequent occurence and really slows things down. This commit fixes the issue by only creating fmap hashes as needed. This should not only resolve the perfomance impact of creating fmap's for all embedded files, but also should improve performance in general. - Add allmatch check to the zip parser after the central-header meta match. That way we don't multiple alerts with the same match except in allmatch mode. Clean up error handling in the zip parser a tiny bit. - Fixes to ensure that the scan limits such as scansize, filesize, recursion depth, # of embedded files, and scantime are always reported if AlertExceedsMax (--alert-exceeds-max) is enabled. - Fixed an issue where non-fatal alerts for exceeding scan maximums may mask signature matches later on. I changed it so these alerts use the "possibly unwanted" alert-type and thus only alert if no other alerts were found or if all-match or heuristic-precedence are enabled. - Added the "Heuristics.Limits.Exceeded.*" events to the JSON metadata when the --gen-json feature is enabled. These will show up once under "ParseErrors" the first time a limit is exceeded. In the present implementation, only one limits-exceeded events will be added, so as to prevent a malicious or malformed sample from filling the JSON buffer with millions of events and using a tonne of RAM.
4 years ago
map = ctx->fmap;
Record names of extracted files A way is needed to record scanned file names for two purposes: 1. File names (and extensions) must be stored in the json metadata properties recorded when using the --gen-json clamscan option. Future work may use this to compare file extensions with detected file types. 2. File names are useful when interpretting tmp directory output when using the --leave-temps option. This commit enables file name retention for later use by storing file names in the fmap header structure, if a file name exists. To store the names in fmaps, an optional name argument has been added to any internal scan API's that create fmaps and every call to these APIs has been modified to pass a file name or NULL if a file name is not required. The zip and gpt parsers required some modification to record file names. The NSIS and XAR parsers fail to collect file names at all and will require future work to support file name extraction. Also: - Added recursive extraction to the tmp directory when the --leave-temps option is enabled. When not enabled, the tmp directory structure remains flat so as to prevent the likelihood of exceeding MAX_PATH. The current tmp directory is stored in the scan context. - Made the cli_scanfile() internal API non-static and added it to scanners.h so it would be accessible outside of scanners.c in order to remove code duplication within libmspack.c. - Added function comments to scanners.h and matcher.h - Converted a TDB-type macros and LSIG-type macros to enums for improved type safey. - Converted more return status variables from `int` to `cl_error_t` for improved type safety, and corrected ooxml file typing functions so they use `cli_file_t` exclusively rather than mixing types with `cl_error_t`. - Restructured the magic_scandesc() function to use goto's for error handling and removed the early_ret_from_magicscan() macro and magic_scandesc_cleanup() function. This makes the code easier to read and made it easier to add the recursive tmp directory cleanup to magic_scandesc(). - Corrected zip, egg, rar filename extraction issues. - Removed use of extra sub-directory layer for zip, egg, and rar file extraction. For Zip, this also involved changing the extracted filenames to be randomly generated rather than using the "zip.###" file name scheme.
5 years ago
tempd = (cli_debug_flag && ctx->engine->keeptmp) ? (ctx->sub_tmpdir ? ctx->sub_tmpdir : cli_gettmpdir()) : NULL;
PE parsing code improvements, db loading bug fixes Consolidate the PE parsing code into one function. I tried to preserve all existing functionality from the previous, distinct implementations to a large extent (with the exceptions mentioned below). If I noticed potential bugs/improvements, I added a TODO statement about those so that they can be fixed in a smaller commit later. Also, there are more TODOs in places where I'm not entirely sure why certain actions are performed - more research is needed for these. I'm submitting a pull request now so that regression testing can be done, and because merging what I have thus far now will likely have fewer conflicts than if I try to merge later PE parsing code improvements: - PEs without all 16 data directories are parsed more appropriately now - Added lots more debug statements Also: - Allow MAX_BC and MAX_TRACKED_PCRE to be specified via CFLAGS When doing performance testing with the latest CVD, MAX_BC and MAX_TRACKED_PCRE need to be raised to track all the events. Allow these to be specified via CFLAGS by not redefining them if they are already defined - Fix an issue preventing wildcard sizes in .MDB/.MSB rules I'm not sure what the original intent of the check I removed was, but it prevents using wildcard sizes in .MDB/.MSB rules. AFAICT these wildcard sizes should be handled appropriately by the MD5 section hash computation code, so I don't think a check on that is needed. - Fix several issues related to db loading - .imp files will now get loaded if they exist in a directory passed via clamscan's '-d' flag - .pwdb files will now get loaded if they exist in a directory passed via clamscan's '-d' flag even when compiling without yara support - Changes to .imp, .ign, and .ign2 files will now be reflected in calls to cl_statinidir and cl_statchkdir (and also .pwdb files, even when compiling without yara support) - The contents of .sfp files won't be included in some of the signature counts, and the contents of .cud files will be - Any local.gdb files will no longer be loaded twice - For .imp files, you are no longer required to specify a minimum flevel for wildcard rules, since this isn't needed
6 years ago
icoff = cli_rawaddr(rva, peinfo->sections, peinfo->nsections, &err, map->len, peinfo->hdr_size);
/* read the bitmap header */
if (err || !(rawimage = fmap_need_off_once(map, icoff, 4))) {
icon_env->err_oof++;
// cli_dbgmsg("parseicon: offset to icon is out of file\n");
return CL_SUCCESS;
}
16 years ago
rva = cli_readint32(rawimage);
PE parsing code improvements, db loading bug fixes Consolidate the PE parsing code into one function. I tried to preserve all existing functionality from the previous, distinct implementations to a large extent (with the exceptions mentioned below). If I noticed potential bugs/improvements, I added a TODO statement about those so that they can be fixed in a smaller commit later. Also, there are more TODOs in places where I'm not entirely sure why certain actions are performed - more research is needed for these. I'm submitting a pull request now so that regression testing can be done, and because merging what I have thus far now will likely have fewer conflicts than if I try to merge later PE parsing code improvements: - PEs without all 16 data directories are parsed more appropriately now - Added lots more debug statements Also: - Allow MAX_BC and MAX_TRACKED_PCRE to be specified via CFLAGS When doing performance testing with the latest CVD, MAX_BC and MAX_TRACKED_PCRE need to be raised to track all the events. Allow these to be specified via CFLAGS by not redefining them if they are already defined - Fix an issue preventing wildcard sizes in .MDB/.MSB rules I'm not sure what the original intent of the check I removed was, but it prevents using wildcard sizes in .MDB/.MSB rules. AFAICT these wildcard sizes should be handled appropriately by the MD5 section hash computation code, so I don't think a check on that is needed. - Fix several issues related to db loading - .imp files will now get loaded if they exist in a directory passed via clamscan's '-d' flag - .pwdb files will now get loaded if they exist in a directory passed via clamscan's '-d' flag even when compiling without yara support - Changes to .imp, .ign, and .ign2 files will now be reflected in calls to cl_statinidir and cl_statchkdir (and also .pwdb files, even when compiling without yara support) - The contents of .sfp files won't be included in some of the signature counts, and the contents of .cud files will be - Any local.gdb files will no longer be loaded twice - For .imp files, you are no longer required to specify a minimum flevel for wildcard rules, since this isn't needed
6 years ago
icoff = cli_rawaddr(rva, peinfo->sections, peinfo->nsections, &err, map->len, peinfo->hdr_size);
if (err || fmap_readn(map, &bmphdr, icoff, sizeof(bmphdr)) != sizeof(bmphdr)) {
icon_env->err_bhoof++;
// cli_dbgmsg("parseicon: bmp header is out of file\n");
return CL_SUCCESS;
}
16 years ago
if ((size_t)READ32(bmphdr.sz) < sizeof(bmphdr)) {
icon_env->err_bhts++;
// cli_dbgmsg("parseicon: BMP header too small\n");
return CL_SUCCESS;
}
16 years ago
/* seek to the end of v4/v5 header */
icoff += READ32(bmphdr.sz);
width = READ32(bmphdr.w);
height = READ32(bmphdr.h) / 2;
depth = READ16(bmphdr.depth);
if (width > 256 || height > 256 || width < 16 || height < 16) {
icon_env->err_tstl++;
// cli_dbgmsg("parseicon: Image too small or too big (%ux%u)\n", width, height);
return CL_SUCCESS;
}
if (width < height * 3 / 4 || height < width * 3 / 4) {
icon_env->err_insl++;
// cli_dbgmsg("parseicon: Image not square enough (%ux%u)\n", width, height);
return CL_SUCCESS;
}
16 years ago
/* scaling logic */
if (width == height) {
if (width == 16 || width == 24 || width == 32)
scalemode = 0;
else if (!(width % 32) || !(width % 24))
scalemode = 1;
else
scalemode = 2;
}
15 years ago
cli_dbgmsg("parseicon: Bitmap - %ux%ux%u\n", width, height, depth);
/* check color depth and load palette */
switch (depth) {
default:
case 0:
/* PNG OR JPEG */
cli_dbgmsg("parseicon: PNG icons are not yet sported\n");
return CL_SUCCESS;
case 1:
case 4:
case 8:
/* HAVE PALETTE */
if (!(palette = fmap_need_off(map, icoff, (1 << depth) * sizeof(int))))
return CL_SUCCESS;
icoff += (1 << depth) * sizeof(int);
/* for(j=0; j<pcolcnt; j++) */
/* cli_dbgmsg("Palette[%u] = %08x\n", j, palette[j]); */
break;
case 16:
case 24:
case 32:
/* NO PALETTE */
break;
}
16 years ago
/* compute line sizes */
scanlinesz = 4 * (width * depth / 32) + 4 * (width * depth % 32 != 0);
andlinesz = ((depth & 0x1f) != 0) * (4 * (width / 32) + 4 * (width % 32 != 0));
16 years ago
/* read the raw image */
16 years ago
if (!(rawimage = fmap_need_off_once(map, icoff, height * (scanlinesz + andlinesz)))) {
if (palette)
fmap_unneed_ptr(map, palette, (1 << depth) * sizeof(int));
return CL_SUCCESS;
}
if (!(imagedata = cli_malloc(width * height * sizeof(*imagedata)))) {
if (palette)
fmap_unneed_ptr(map, palette, (1 << depth) * sizeof(int));
return CL_SUCCESS;
}
16 years ago
/* decode the image to an RGBA array */
for (y = 0; y < height; y++) {
unsigned int x_off = y * scanlinesz;
switch (depth) {
case 1:
case 4:
case 8: {
unsigned int have = 0;
unsigned char c = 0; // will be set in first loop
for (x = 0; x < width; x++) {
if (!have) {
c = rawimage[x_off++];
have = 8;
}
have -= depth;
imagedata[(height - 1 - y) * width + x] = READ32(palette[(c >> have) & ((1 << depth) - 1)]);
}
break;
}
case 16: {
for (x = 0; x < width; x++) {
unsigned int b = (rawimage[x_off] & 0x1f);
unsigned int g = ((rawimage[x_off] >> 5) | ((rawimage[x_off + 1] & 0x3) << 3));
unsigned int r = (rawimage[x_off + 1] & 0xfc);
b = (b << 3) | (b >> 2);
g = ((g << 3) | (g >> 2)) << 11;
r = ((r << 3) | (r >> 2)) << 17;
imagedata[(height - 1 - y) * width + x] = r | g | b;
x_off += 2;
}
break;
}
case 24:
for (x = 0; x < width; x++) {
unsigned int c = rawimage[x_off] | (rawimage[x_off + 1] << 8) | (rawimage[x_off + 2] << 16);
imagedata[(height - 1 - y) * width + x] = c;
x_off += 3;
}
break;
case 32:
for (x = 0; x < width; x++) {
unsigned int a = rawimage[x_off + 3] << 24;
imagedata[(height - 1 - y) * width + x] = rawimage[x_off] | (rawimage[x_off + 1] << 8) | (rawimage[x_off + 2] << 16) | a;
special_32_is_32 |= a;
x_off += 4;
}
break;
}
}
16 years ago
if (palette)
fmap_unneed_ptr(map, palette, (1 << depth) * sizeof(int));
makebmp("0-noalpha", tempd, width, height, imagedata);
if (depth == 32 && !special_32_is_32) { /* Sometimes it really is 24. Exploited live - see sample 0013839101 */
andlinesz = 4 * (width / 32) + 4 * (width % 32 != 0);
if (!(rawimage = fmap_need_off_once(map, icoff + height * scanlinesz, height * andlinesz))) {
/* Likely a broken sample - 32bit icon with 24bit data and a broken mask:
i could really break out here but i've got the full image, so i'm just forcing full alpha
Found in samples: 0008777448, 0009116157, 0009116157 */
for (y = 0; y < height; y++)
for (x = 0; x < width; x++)
imagedata[y * width + x] |= 0xff000000;
special_32_is_32 = 1;
cli_dbgmsg("parseicon: found a broken and stupid icon\n");
} else
cli_dbgmsg("parseicon: found a stupid icon\n");
} else
rawimage += height * scanlinesz;
/* Set alpha on or off based on the mask */
if ((depth & 0x1f) || !special_32_is_32) {
for (y = 0; y < height; y++) {
unsigned int x_off = y * andlinesz;
unsigned int have = 0;
unsigned char c = 0; // will be set in first loop
for (x = 0; x < width; x++) {
if (!have) {
c = rawimage[x_off++];
have = 8;
}
have--;
imagedata[(height - 1 - y) * width + x] |= (!((c >> have) & 1)) * 0xff000000;
}
}
}
makebmp("1-alpha-mask", tempd, width, height, imagedata);
/* Blend alpha */
for (y = 0; y < height; y++) {
for (x = 0; x < width; x++) {
unsigned int r, g, b, a;
unsigned int c = imagedata[y * width + x];
a = c >> 24;
r = (c >> 16) & 0xff;
g = (c >> 8) & 0xff;
b = c & 0xff;
r = 0xff - a + a * r / 0xff;
g = 0xff - a + a * g / 0xff;
b = 0xff - a + a * b / 0xff;
imagedata[y * width + x] = 0xff000000 | (r << 16) | (g << 8) | b;
}
}
16 years ago
switch (scalemode) {
case 0:
break;
case 1:
/* Fast 50% scaler with linear gamma */
while (width > 32) {
for (y = 0; y < height; y += 2) {
for (x = 0; x < width; x += 2) {
unsigned int c1 = imagedata[y * width + x], c2 = imagedata[y * width + x + 1], c3 = imagedata[(y + 1) * width + x], c4 = imagedata[(y + 1) * width + x + 1];
c1 = (((c1 ^ c2) & 0xfefefefe) >> 1) + (c1 & c2);
c2 = (((c3 ^ c4) & 0xfefefefe) >> 1) + (c3 & c4);
imagedata[y / 2 * width / 2 + x / 2] = (((c1 ^ c2) & 0xfefefefe) >> 1) + (c1 & c2);
}
}
width /= 2;
height /= 2;
cli_dbgmsg("parseicon: Fast scaling to %ux%u\n", width, height);
}
break;
case 2:
/* Slow up/down scale */
{
double scalex, scaley;
unsigned int newsize;
uint32_t *newdata;
if (abs((int)width - 32) + abs((int)height - 32) < abs((int)width - 24) + abs((int)height - 24))
newsize = 32;
else if (abs((int)width - 24) + abs((int)height - 24) < abs((int)width - 16) + abs((int)height - 16))
newsize = 24;
else
newsize = 16;
scalex = (double)width / newsize;
scaley = (double)height / newsize;
if (!(newdata = cli_malloc(newsize * newsize * sizeof(*newdata)))) {
cli_errmsg("parseicon: Unable to allocate memory for scaling image\n");
return CL_EMEM;
}
cli_dbgmsg("parseicon: Slow scaling to %ux%u (%f, %f)\n", newsize, newsize, scalex, scaley);
for (y = 0; y < newsize; y++) {
unsigned int oldy = (unsigned int)(y * scaley) * width;
for (x = 0; x < newsize; x++)
newdata[y * newsize + x] = imagedata[oldy + (unsigned int)(x * scalex + 0.5f)];
}
free(imagedata);
height = newsize;
width = newsize;
imagedata = newdata;
}
}
makebmp("2-alpha-blend", tempd, width, height, imagedata);
16 years ago
getmetrics(width, imagedata, &metrics, tempd);
free(imagedata);
enginesize = (width >> 3) - 2;
for (x = 0; x < matcher->icon_counts[enginesize]; x++) {
unsigned int color = 0, gray = 0, bright, dark, edge, noedge, reds, greens, blues, ccount;
unsigned int colors, confidence, bwmatch = 0, positivematch = 64 + 4 * (2 - enginesize);
unsigned int i, j;
i = matcher->icons[enginesize][x].group[0];
j = i % 64;
i /= 64;
if (!(set->v[0][i] & ((uint64_t)1 << j)))
continue;
i = matcher->icons[enginesize][x].group[1];
j = i % 64;
i /= 64;
if (!(set->v[1][i] & ((uint64_t)1 << j)))
continue;
if (!metrics.ccount && !matcher->icons[enginesize][x].ccount) {
/* BW matching */
edge = matchbwpoint(width, metrics.edge_x, metrics.edge_y, metrics.edge_avg, metrics.color_x, metrics.color_y, metrics.color_avg, matcher->icons[enginesize][x].edge_x, matcher->icons[enginesize][x].edge_y, matcher->icons[enginesize][x].edge_avg, matcher->icons[enginesize][x].color_x, matcher->icons[enginesize][x].color_y, matcher->icons[enginesize][x].color_avg);
noedge = matchbwpoint(width, metrics.noedge_x, metrics.noedge_y, metrics.noedge_avg, metrics.gray_x, metrics.gray_y, metrics.gray_avg, matcher->icons[enginesize][x].noedge_x, matcher->icons[enginesize][x].noedge_y, matcher->icons[enginesize][x].noedge_avg, matcher->icons[enginesize][x].gray_x, matcher->icons[enginesize][x].gray_y, matcher->icons[enginesize][x].gray_avg);
bwmatch = 1;
} else {
edge = matchpoint(width, metrics.edge_x, metrics.edge_y, metrics.edge_avg, matcher->icons[enginesize][x].edge_x, matcher->icons[enginesize][x].edge_y, matcher->icons[enginesize][x].edge_avg, 255);
noedge = matchpoint(width, metrics.noedge_x, metrics.noedge_y, metrics.noedge_avg, matcher->icons[enginesize][x].noedge_x, matcher->icons[enginesize][x].noedge_y, matcher->icons[enginesize][x].noedge_avg, 255);
if (metrics.ccount && matcher->icons[enginesize][x].ccount) {
/* color matching */
color = matchpoint(width, metrics.color_x, metrics.color_y, metrics.color_avg, matcher->icons[enginesize][x].color_x, matcher->icons[enginesize][x].color_y, matcher->icons[enginesize][x].color_avg, 4072);
gray = matchpoint(width, metrics.gray_x, metrics.gray_y, metrics.gray_avg, matcher->icons[enginesize][x].gray_x, matcher->icons[enginesize][x].gray_y, matcher->icons[enginesize][x].gray_avg, 4072);
}
}
bright = matchpoint(width, metrics.bright_x, metrics.bright_y, metrics.bright_avg, matcher->icons[enginesize][x].bright_x, matcher->icons[enginesize][x].bright_y, matcher->icons[enginesize][x].bright_avg, 255);
dark = matchpoint(width, metrics.dark_x, metrics.dark_y, metrics.dark_avg, matcher->icons[enginesize][x].dark_x, matcher->icons[enginesize][x].dark_y, matcher->icons[enginesize][x].dark_avg, 255);
reds = abs((int)metrics.rsum - (int)matcher->icons[enginesize][x].rsum) * 10;
reds = (reds < 100) * (100 - reds);
greens = abs((int)metrics.gsum - (int)matcher->icons[enginesize][x].gsum) * 10;
greens = (greens < 100) * (100 - greens);
blues = abs((int)metrics.bsum - (int)matcher->icons[enginesize][x].bsum) * 10;
blues = (blues < 100) * (100 - blues);
ccount = abs((int)metrics.ccount - (int)matcher->icons[enginesize][x].ccount) * 10;
ccount = (ccount < 100) * (100 - ccount);
colors = (reds + greens + blues + ccount) / 4;
if (bwmatch) {
confidence = (bright + dark + edge * 2 + noedge) / 6;
positivematch = 70;
} else
confidence = (color + (gray + bright + noedge) * 2 / 3 + dark + edge + colors) / 6;
#ifdef LOGPARSEICONDETAILS
cli_dbgmsg("parseicon: edge confidence: %u%%\n", edge);
cli_dbgmsg("parseicon: noedge confidence: %u%%\n", noedge);
if (!bwmatch) {
cli_dbgmsg("parseicon: color confidence: %u%%\n", color);
cli_dbgmsg("parseicon: gray confidence: %u%%\n", gray);
}
cli_dbgmsg("parseicon: bright confidence: %u%%\n", bright);
cli_dbgmsg("parseicon: dark confidence: %u%%\n", dark);
if (!bwmatch)
cli_dbgmsg("parseicon: spread confidence: red %u%%, green %u%%, blue %u%% - colors %u%%\n", reds, greens, blues, ccount);
#endif
if (confidence >= positivematch) {
cli_dbgmsg("confidence: %u\n", confidence);
return CL_VIRUS;
}
}
return CL_SUCCESS;
}
void cli_icongroupset_add(const char *groupname, icon_groupset *set, unsigned int type, cli_ctx *ctx)
{
struct icon_matcher *matcher;
unsigned int i, j;
if (type > 1 || !ctx || !ctx->engine || !(matcher = ctx->engine->iconcheck) || !matcher->group_counts[type])
return;
j = matcher->group_counts[type];
if (groupname[0] == '*' && !groupname[1]) {
set->v[type][0] = set->v[type][1] = set->v[type][2] = set->v[type][3] = ~0;
return;
}
for (i = 0; i < j; i++) {
if (!strcmp(groupname, matcher->group_names[type][i]))
break;
}
if (i == j)
cli_dbgmsg("cli_icongroupset_add: failed to locate icon group%u %s\n", type, groupname);
else {
j = i % 64;
i /= 64;
set->v[type][i] |= (uint64_t)1 << j;
}
}