The cli_max_malloc, cli_max_calloc, and cli_max_realloc functions
provide a way to protect against allocating too much memory
when the size of the allocation is derived from the untrusted input.
Specifically, we worry about values in the file being scanned being
manipulated to exhaust the RAM and crash the application.
There is no need to check the limits if the size of the allocation
is fixed, or if the size of the allocation is necessary for signature
loading, or the general operation of the applications.
E.g. checking the max-allocation limit for the size of a hash, or
for the size of the scan recursion stack, is a complete waste of
time.
Although we significantly increased the max-allocation limit in
a recent release, it is best not to check an allocation if the
allocation will be safe. It would be a waste of time.
I am also hopeful that if we can reduce the number allocations
that require a limit-check to those that require it for the safe
scan of a file, then eventually we can store the limit in the scan-
context, and make it configurable.
We have some special functions to wrap malloc, calloc, and realloc to
make sure we don't allocate more than some limit, similar to the
max-filesize and max-scansize limits. Our wrappers are really only
needed when allocating memory for scans based on untrusted user input,
where a scan file could have bytes that claim you need to allocate
some ridiculous amount of memory. Right now they're named:
- cli_malloc
- cli_calloc
- cli_realloc
- cli_realloc2
... and these names do not convey their purpose
This commit renames them to:
- cli_max_malloc
- cli_max_calloc
- cli_max_realloc
- cli_max_realloc2
The realloc ones also have an additional feature in that they will not
free your pointer if you try to realloc to 0 bytes. Freeing the memory
is undefined by the C spec, and only done with some realloc
implementations, so this stabilizes on the behavior of not doing that,
which should prevent accidental double-free's.
So for the case where you may want to realloc and do not need to have a
maximum, this commit adds the following functions:
- cli_safer_realloc
- cli_safer_realloc2
These are used for the MPOOL_REALLOC and MPOOL_REALLOC2 macros when
MPOOL is disabled (e.g. because mmap-support is not found), so as to
match the behavior in the mpool_realloc/2 functions that do not make use
of the allocation-limit.
Also fixed a number of conditions where magic_scan() critical errors may
be ignored.
To ensure that the scan truly aborts for signature matches (not in
allmatch mode) and for timeouts, the `ctx->abort` option is now set in
these two conditions, and checked in several spots in magic_scan().
Additionally, I've consolidated some of the "scan must halt" type of
checks (mostly large switch statements) into a function so that we can
use the exact same logic in a few places in magic_scan().
I've also fixed a few minor warnings and code format issues.
Rework the append_virus mechanism to store evidence (strong indicators,
pua indicators, and eventually weak indicators) in vectors. When
appending a "virus", we will return CLEAN when in allmatch-mode, and
simply add the indicator to the appropriate vector.
Later we can check if there were any alerts to return a vector by
summing the lengths of the strong and pua indicator vectors.
This does away with storing the latest "virname" in the scan context.
Instead, we can query for the last indicator in the evidence, giving
priority to strong indicators.
When heuristic-precendence is enabled, add PUA as Strong instead of
as PotentiallyUnwanted. This way, they will be treated equally and
reported in order in allmatch mode.
Also document reason for disabling cache with metadata JSON enabled
The fmap module provides a mechanism for creating a mapping into an
existing map at an offset and length that's used when a file is found
with an uncompressed archive or when embedded files are found with
embedded file type recognition in scanraw(). This is the
"fmap_duplicate()" function. Duplicate fmaps just reference the original
fmap's 'data' or file handle/descriptor while allowing the caller to
treat it like a new map using offsets and lengths that don't account for
the original/actual file dimensions.
fmap's keep track of this with m->nested_offset & m->real_len, which
admittedly have confusing names. I found incorrect uses of these in a
handful of locations. Notably:
- In cli_magic_scan_nested_fmap_type().
The force-to-disk feature would have been checking incorrect sizes and
may have written incorrect offsets for duplicate fmaps.
- In XDP parser.
- A bunch of places from the previous commit when making dupe maps.
This commit fixes those and adds lots of documentation to the fmap.h API
to try to prevent confusion in the future.
nested_offset should never be referenced outside of fmap.c/h.
The fmap_* functions for accessing or reading map data have two
implementations, mem_* or handle_*, depending the data source.
I found issues with some of these so I made a unit test that covers each
of the functions I'm concerned about for both types of data sources and
for both original fmaps and nested/duplicate fmaps.
With the tests, I found and fixed issues in these fmap functions:
- handle_need_offstr(): must account for the nested_offset in dupe maps.
- handle_gets(): must account for nested_offset and use len & real_len
correctly.
- mem_need_offstr(): must account for nested_offset in dupe maps.
- mem_gets(): must account for nested_offset and use len & real_len
correctly.
Moved CDBRANGE() macro out of function definition so for better
legibility.
Fixed a few warnings.
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.
The named "shared" is confusing, especially now that these features are
built as a static library instead of being directly compiled into the
various applications.
On Windows, files open()'ed without the O_BINARY flag will have new-line
LF (aka \n) converted to CRLF (aka \r\n) automatically when read from or
written to. This is undesirable for all scan targets AND temp files
because it affects pattern matching and with hashing.
This commit converts a handful of instances throughout the codebase
where it appears that O_BINARY was mistakenly omitted and could result
in unexpected behavior on Windows.
Git on Windows also converts LF -> CRLF for "text" files, for editing
purposes.
This is problematic for scan files and test files that should match
verbatim.
We can prevent this issue by marking .ref test files as "binary" in the
.gitattributes file and by always opening scan files and temp files as
binary.
In this commit I've also removed the `ChangeLog merge=cl-merge` line
that was once used to reduce ChangeLog merge conflicts by using the
gnulib git-merge-changlog tool. This project now categorizes changes in
the NEWS.md.
For finer detail, git commit history is fully accessible on github.com.
This patch adds experimental-quality CMake build tooling.
The libmspack build required a modification to use "" instead of <> for
header #includes. This will hopefully be included in the libmspack
upstream project when adding CMake build tooling to libmspack.
Removed use of libltdl when using CMake.
Flex & Bison are now required to build.
If -DMAINTAINER_MODE, then GPERF is also required, though it currently
doesn't actually do anything. TODO!
I found that the autotools build system was generating the lexer output
but not actually compiling it, instead using previously generated (and
manually renamed) lexer c source. As a consequence, changes to the .l
and .y files weren't making it into the build. To resolve this, I
removed generated flex/bison files and fixed the tooling to use the
freshly generated files. Flex and bison are now required build tools.
On Windows, this adds a dependency on the winflexbison package,
which can be obtained using Chocolatey or may be manually installed.
CMake tooling only has partial support for building with external LLVM
library, and no support for the internal LLVM (to be removed in the
future). I.e. The CMake build currently only supports the bytecode
interpreter.
Many files used include paths relative to the top source directory or
relative to the current project, rather than relative to each build
target. Modern CMake support requires including internal dependency
headers the same way you would external dependency headers (albeit
with "" instead of <>). This meant correcting all header includes to
be relative to the build targets and not relative to the workspace.
For example, ...
```c
include "../libclamav/clamav.h"
include "clamd/clamd_others.h"
```
... becomes:
```c
// libclamav
include "clamav.h"
// clamd
include "clamd_others.h"
```
Fixes header name conflicts by renaming a few of the files.
Converted the "shared" code into a static library, which depends on
libclamav. The ironically named "shared" static library provides
features common to the ClamAV apps which are not required in
libclamav itself and are not intended for use by downstream projects.
This change was required for correct modern CMake practices but was
also required to use the automake "subdir-objects" option.
This eliminates warnings when running autoreconf which, in the next
version of autoconf & automake are likely to break the build.
libclamav used to build in multiple stages where an earlier stage is
a static library containing utils required by the "shared" code.
Linking clamdscan and clamdtop with this libclamav utils static lib
allowed these two apps to function without libclamav. While this is
nice in theory, the practical gains are minimal and it complicates
the build system. As such, the autotools and CMake tooling was
simplified for improved maintainability and this feature was thrown
out. clamdtop and clamdscan now require libclamav to function.
Removed the nopthreads version of the autotools
libclamav_internal_utils static library and added pthread linking to
a couple apps that may have issues building on some platforms without
it, with the intention of removing needless complexity from the
source. Kept the regular version of libclamav_internal_utils.la
though it is no longer used anywhere but in libclamav.
Added an experimental doxygen build option which attempts to build
clamav.h and libfreshclam doxygen html docs.
The CMake build tooling also may build the example program(s), which
isn't a feature in the Autotools build system.
Changed C standard to C90+ due to inline linking issues with socket.h
when linking libfreshclam.so on Linux.
Generate common.rc for win32.
Fix tabs/spaces in shared Makefile.am, and remove vestigial ifndef
from misc.c.
Add CMake files to the automake dist, so users can try the new
CMake tooling w/out having to build from a git clone.
clamonacc changes:
- Renamed FANOTIFY macro to HAVE_SYS_FANOTIFY_H to better match other
similar macros.
- Added a new clamav-clamonacc.service systemd unit file, based on
the work of ChadDevOps & Aaron Brighton.
- Added missing clamonacc man page.
Updates to clamdscan man page, add missing options.
Remove vestigial CL_NOLIBCLAMAV definitions (all apps now use
libclamav).
Rename Windows mspack.dll to libmspack.dll so all ClamAV-built
libraries have the lib-prefix with Visual Studio as with CMake.
Add notices to man pages and help strings cautioning against running
bytecode signatures from untrusted sources.
Also adds missing BytecodeUnsigned option to clamd.conf.sample 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.
New features added to freshclam:
- Update signature definitions over HTTPS.
- Support for HTTP protocol v1.1 (formerly v1.0).
- New libfreshclam library with an all new API and versioning separate from libclamav (v2.0.0). This library is now build and installed alongside libclamav as a hard dependency of freshclam.
- The ability to opt-in and opt-out of standard and optional official ClamAV databases (ExtraDatabase, ExcludeDatabase)
- The option to specify the protocol and port number of official and private mirror servers.
- Support for additional types of proxy servers beyond plain HTTP (SOCKS 4, SOCKS 5).
Features removed from freshclam:
- Mirror management (mirrors.dat) file. This feature is no longer needed as official signature databases are distributed using a paid content delivery network (Cloudflare).
This commit also adds the following features for Windows users:
- The clamsubmit tool.
- The json-c library dependency, which will enable the --gen-json option in clamscan.
- Third party libraries under the win32/3rdparty directory have been removed. Developers will need to build the libraries separately from ClamAV and provide the headers and lib/dll library files the same way they do for OpenSSL. This includes libxml2, pthread-win32, bzip2, zlib, pcre2 as well as new dependencies: curl, json-c. Developers are encouraged to use the build tool Mussels to simplify this task.
Val Snyder