In Tensorflow before versions 2.2.1 and 2.3.1, if a user passes a list of strings to `dlpack.to_dlpack` there is a memory leak following an expected validation failure. The issue occurs because the `status` argument during validation failures is not properly checked. Since each of the above methods can return an error status, the `status` value must be checked before continuing. The issue is patched in commit 22e07fb204386768e5bcbea563641ea11f96ceb8 and is released in TensorFlow versions 2.2.1, or 2.3.1.

In tensorflow-lite before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, if a TFLite saved model uses the same tensor as both input and output of an operator, then, depending on the operator, we can observe a segmentation fault or just memory corruption. We have patched the issue in d58c96946b and will release patch releases for all versions between 1.15 and 2.3. We recommend users to upgrade to TensorFlow 1.15.4, 2.0.3, 2.1.2, 2.2.1, or 2.3.1.

In Tensorflow before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, by controlling the `fill` argument of tf.strings.as_string, a malicious attacker is able to trigger a format string vulnerability due to the way the internal format use in a `printf` call is constructed. This may result in segmentation fault. The issue is patched in commit 33be22c65d86256e6826666662e40dbdfe70ee83, and is released in TensorFlow versions 1.15.4, 2.0.3, 2.1.2, 2.2.1, or 2.3.1.

Insufficient data validation in media in Google Chrome prior to 85.0.4183.121 allowed a remote attacker to potentially exploit heap corruption via a crafted HTML page.

In Tensorflow before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, changing the TensorFlow's `SavedModel` protocol buffer and altering the name of required keys results in segfaults and data corruption while loading the model. This can cause a denial of service in products using `tensorflow-serving` or other inference-as-a-service installments. Fixed were added in commits f760f88b4267d981e13f4b302c437ae800445968 and fcfef195637c6e365577829c4d67681695956e7d (both going into TensorFlow 2.2.0 and 2.3.0 but not yet backported to earlier versions). However, this was not enough, as #41097 reports a different failure mode. The issue is patched in commit adf095206f25471e864a8e63a0f1caef53a0e3a6, and is released in TensorFlow versions 1.15.4, 2.0.3, 2.1.2, 2.2.1, or 2.3.1.

In Tensorflow before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, the `tf.raw_ops.Switch` operation takes as input a tensor and a boolean and outputs two tensors. Depending on the boolean value, one of the tensors is exactly the input tensor whereas the other one should be an empty tensor. However, the eager runtime traverses all tensors in the output. Since only one of the tensors is defined, the other one is `nullptr`, hence we are binding a reference to `nullptr`. This is undefined behavior and reported as an error if compiling with `-fsanitize=null`. In this case, this results in a segmentation fault The issue is patched in commit da8558533d925694483d2c136a9220d6d49d843c, and is released in TensorFlow versions 1.15.4, 2.0.3, 2.1.2, 2.2.1, or 2.3.1.

Insufficient data validation in dialogs in Google Chrome on OS X prior to 86.0.4240.75 allowed a remote attacker to obtain potentially sensitive information from disk via a crafted HTML page.

Insufficient data validation in webUI in Google Chrome on ChromeOS prior to 86.0.4240.75 allowed a local attacker to bypass content security policy via a crafted HTML page.

In Tensorflow before versions 2.2.1 and 2.3.1, if a user passes an invalid argument to `dlpack.to_dlpack` the expected validations will cause variables to bind to `nullptr` while setting a `status` variable to the error condition. However, this `status` argument is not properly checked. Hence, code following these methods will bind references to null pointers. This is undefined behavior and reported as an error if compiling with `-fsanitize=null`. The issue is patched in commit 22e07fb204386768e5bcbea563641ea11f96ceb8 and is released in TensorFlow versions 2.2.1, or 2.3.1.

Insufficient data validation in navigation in Google Chrome on Android prior to 86.0.4240.75 allowed a remote attacker who had compromised the renderer process to bypass navigation restrictions via a crafted HTML page.

The (1) git-remote-ext and (2) unspecified other remote helper programs in Git before 2.3.10, 2.4.x before 2.4.10, 2.5.x before 2.5.4, and 2.6.x before 2.6.1 do not properly restrict the allowed protocols, which might allow remote attackers to execute arbitrary code via a URL in a (a) .gitmodules file or (b) unknown other sources in a submodule.

parsetagx.c in w3m before 0.5.3+git20161009 does not properly initialize values, which allows remote attackers to crash the application via a crafted html file, related to a <i> tag.

Mozilla Firefox before 43.0 mishandles the # (number sign) character in a data: URI, which allows remote attackers to spoof web sites via unspecified vectors.

zypp-refresh-patches in zypper in SUSE openSUSE 10.2, 10.3, and 11.0 does not ask the user before accepting repository keys, which allows remote repositories to cause a denial of service (package data corruption) via a spoofed key.

os_unix.c in SQLite before 3.13.0 improperly implements the temporary directory search algorithm, which might allow local users to obtain sensitive information, cause a denial of service (application crash), or have unspecified other impact by leveraging use of the current working directory for temporary files.

The gdImageCropThreshold function in gd_crop.c in the GD Graphics Library (aka libgd) before 2.2.3, as used in PHP before 7.0.9, allows remote attackers to cause a denial of service (application crash) via an invalid color index.

The parse_chunk_header function in libtorrent before 1.1.1 allows remote attackers to cause a denial of service (crash) via a crafted (1) HTTP response or possibly a (2) UPnP broadcast.

The mkdumprd script called "dracut" in the current working directory "." allows local users to trick the administrator into executing code as root.

The bcpowmod function in ext/bcmath/bcmath.c in PHP before 5.5.35, 5.6.x before 5.6.21, and 7.x before 7.0.6 accepts a negative integer for the scale argument, which allows remote attackers to cause a denial of service or possibly have unspecified other impact via a crafted call.

Libksba before 1.3.4 allows remote attackers to cause a denial of service (out-of-bounds read and crash) via unspecified vectors, related to the "returned length of the object from _ksba_ber_parse_tl."

The bcpowmod function in ext/bcmath/bcmath.c in PHP before 5.5.35, 5.6.x before 5.6.21, and 7.x before 7.0.6 modifies certain data structures without considering whether they are copies of the _zero_, _one_, or _two_ global variable, which allows remote attackers to cause a denial of service or possibly have unspecified other impact via a crafted call.

Multiple unspecified vulnerabilities in Google Chrome before 53.0.2785.143 allow remote attackers to cause a denial of service or possibly have other impact via unknown vectors.

The _rsvg_css_normalize_font_size function in librsvg 2.40.2 allows context-dependent attackers to cause a denial of service (stack consumption and application crash) via circular definitions in an SVG document.

Mercurial before 3.7.3 allows remote attackers to execute arbitrary code via a crafted git ext:: URL when cloning a subrepository.

In Tensorflow before versions 1.15.4, 2.0.3, 2.1.2, 2.2.1 and 2.3.1, the `SparseFillEmptyRowsGrad` implementation has incomplete validation of the shapes of its arguments. Although `reverse_index_map_t` and `grad_values_t` are accessed in a similar pattern, only `reverse_index_map_t` is validated to be of proper shape. Hence, malicious users can pass a bad `grad_values_t` to trigger an assertion failure in `vec`, causing denial of service in serving installations. The issue is patched in commit 390611e0d45c5793c7066110af37c8514e6a6c54, and is released in TensorFlow versions 1.15.4, 2.0.3, 2.1.2, 2.2.1, or 2.3.1."

The (1) xmlParserEntityCheck and (2) xmlParseAttValueComplex functions in parser.c in libxml2 2.9.3 do not properly keep track of the recursion depth, which allows context-dependent attackers to cause a denial of service (stack consumption and application crash) via a crafted XML document containing a large number of nested entity references.

Stack-based buffer overflow in the getaddrinfo function in sysdeps/posix/getaddrinfo.c in the GNU C Library (aka glibc or libc6) allows remote attackers to cause a denial of service (crash) via vectors involving hostent conversion. NOTE: this vulnerability exists because of an incomplete fix for CVE-2013-4458.

The bgp_dump_routes_func function in bgpd/bgp_dump.c in Quagga does not perform size checks when dumping data, which might allow remote attackers to cause a denial of service (assertion failure and daemon crash) via a large BGP packet.

The (1) EPHEMERAL, (2) HTTPS, (3) MVG, (4) MSL, (5) TEXT, (6) SHOW, (7) WIN, and (8) PLT coders in ImageMagick before 6.9.3-10 and 7.x before 7.0.1-1 allow remote attackers to execute arbitrary code via shell metacharacters in a crafted image, aka "ImageTragick."

The JSGenericLowering class in compiler/js-generic-lowering.cc in Google V8, as used in Google Chrome before 50.0.2661.94, mishandles comparison operators, which allows remote attackers to obtain sensitive information via crafted JavaScript code.

Mozilla Firefox before 44.0 allows user-assisted remote attackers to spoof a trailing substring in the address bar by leveraging a user's paste of a (1) wyciwyg: URI or (2) resource: URI.

Perl might allow context-dependent attackers to bypass the taint protection mechanism in a child process via duplicate environment variables in envp.

Blink, as used in Google Chrome before 50.0.2661.94, mishandles assertions in the WTF::BitArray and WTF::double_conversion::Vector classes, which allows remote attackers to cause a denial of service (out-of-bounds write) or possibly have unspecified other impact via a crafted web site.

The media subsystem in Google Chrome before 50.0.2661.75 does not initialize an unspecified data structure, which allows remote attackers to cause a denial of service (invalid read operation) via unknown vectors.

The verify function in the RSA package for Python (Python-RSA) before 3.3 allows attackers to spoof signatures with a small public exponent via crafted signature padding, aka a BERserk attack.

print-wb.c in tcpdump before 4.7.4 allows remote attackers to cause a denial of service (segmentation fault and process crash).

The MSL interpreter in ImageMagick before 6.9.6-4 allows remote attackers to cause a denial of service (segmentation fault and application crash) via a crafted XML file.

Stack consumption vulnerability in Zend/zend_exceptions.c in PHP before 5.4.44, 5.5.x before 5.5.28, and 5.6.x before 5.6.12 allows remote attackers to cause a denial of service (segmentation fault) via recursive method calls.

The gdk-pixbuf configuration in Mozilla Firefox before 43.0 on Linux GNOME platforms incorrectly enables the JasPer decoder, which allows remote attackers to cause a denial of service or possibly have unspecified other impact via a crafted JPEG 2000 image.

The API in the WebKit Plug-ins component in Apple Safari before 9 does not provide notification of an HTTP Redirection (aka 3xx) status code to a plugin, which allows remote attackers to bypass intended request restrictions via a crafted web site.

IcedTea-Web before 1.5.3 and 1.6.x before 1.6.1 does not properly determine the origin of unsigned applets, which allows remote attackers to bypass the approval process or trick users into approving applet execution via a crafted web page.

An issue was discovered in Rsyslog v8.1908.0. contrib/pmcisconames/pmcisconames.c has a heap overflow in the parser for Cisco log messages. The parser tries to locate a log message delimiter (in this case, a space or a colon), but fails to account for strings that do not satisfy this constraint. If the string does not match, then the variable lenMsg will reach the value zero and will skip the sanity check that detects invalid log messages. The message will then be considered valid, and the parser will eat up the nonexistent colon delimiter. In doing so, it will decrement lenMsg, a signed integer, whose value was zero and now becomes minus one. The following step in the parser is to shift left the contents of the message. To do this, it will call memmove with the right pointers to the target and destination strings, but the lenMsg will now be interpreted as a huge value, causing a heap overflow.

IcedTea-Web before 1.5.3 and 1.6.x before 1.6.1 does not properly sanitize applet URLs, which allows remote attackers to inject applets into the .appletTrustSettings configuration file and bypass user approval to execute the applet via a crafted web page, possibly related to line breaks.

jcc.c in Privoxy before 3.0.23 allows remote attackers to cause a denial of service (abort) via a crafted chunk-encoded body.

The dissect_atn_cpdlc_heur function in asn1/atn-cpdlc/packet-atn-cpdlc-template.c in the ATN-CPDLC dissector in Wireshark 1.12.x before 1.12.4 does not properly follow the TRY/ENDTRY code requirements, which allows remote attackers to cause a denial of service (stack memory corruption and application crash) via a crafted packet.

The LocalFrame::isURLAllowed function in core/frame/LocalFrame.cpp in Blink, as used in Google Chrome before 44.0.2403.89, does not properly check for a page's maximum number of frames, which allows remote attackers to cause a denial of service (invalid count value and use-after-free) or possibly have unspecified other impact via crafted JavaScript code that makes many createElement calls for IFRAME elements.

asn1/lpp/lpp.cnf in the LPP dissector in Wireshark 1.10.x before 1.10.12 and 1.12.x before 1.12.3 does not validate a certain index value, which allows remote attackers to cause a denial of service (out-of-bounds memory access and application crash) via a crafted packet.

The MagickRealloc function in memory.c in Graphicsmagick 1.3.25 allows remote attackers to cause a denial of service (crash) via large dimensions in a jpeg image.

parsers.c in Privoxy before 3.0.23 allows remote attackers to cause a denial of service (invalid read and crash) via vectors related to an HTTP time header.

A vulnerability in the Object Linking & Embedding (OLE2) file scanning functionality of Clam AntiVirus (ClamAV) Software versions 0.101.1 and prior could allow an unauthenticated, remote attacker to cause a denial of service condition on an affected device. The vulnerability is due to a lack of proper input and validation checking mechanisms for OLE2 files sent an affected device. An attacker could exploit this vulnerability by sending malformed OLE2 files to the device running an affected version ClamAV Software. An exploit could allow the attacker to cause an out-of-bounds write condition, resulting in a crash that could result in a denial of service condition on an affected device.