Canonical Ubuntu Linux 19.10

Incomplete cleanup from specific special register read operations in some Intel(R) Processors may allow an authenticated user to potentially enable information disclosure via local access.

GnuTLS 3.6.x before 3.6.14 uses incorrect cryptography for encrypting a session ticket (a loss of confidentiality in TLS 1.2, and an authentication bypass in TLS 1.3). The earliest affected version is 3.6.4 (2018-09-24) because of an error in a 2018-09-18 commit. Until the first key rotation, the TLS server always uses wrong data in place of an encryption key derived from an application.

An issue was discovered in Django 2.2 before 2.2.13 and 3.0 before 3.0.7. In cases where a memcached backend does not perform key validation, passing malformed cache keys could result in a key collision, and potential data leakage.

A vulnerability was found in DPDK versions 18.11 and above. The vhost-crypto library code is missing validations for user-supplied values, potentially allowing an information leak through an out-of-bounds memory read.

A memory corruption issue was found in DPDK versions 17.05 and above. This flaw is caused by an integer truncation on the index of a payload. Under certain circumstances, the index (a UInt) is copied and truncated into a uint16, which can lead to out of bound indexing and possible memory corruption.

A vulnerability was found in DPDK versions 18.05 and above. A missing check for an integer overflow in vhost_user_set_log_base() could result in a smaller memory map than requested, possibly allowing memory corruption.

An Ubuntu-specific modification to Pulseaudio to provide security mediation for Snap-packaged applications was found to have a bypass of intended access restriction for snaps which plugs any of pulseaudio, audio-playback or audio-record via unloading the pulseaudio snap policy module. This issue affects: pulseaudio 1:8.0 versions prior to 1:8.0-0ubuntu3.12; 1:11.1 versions prior to 1:11.1-1ubuntu7.7; 1:13.0 versions prior to 1:13.0-1ubuntu1.2; 1:13.99.1 versions prior to 1:13.99.1-1ubuntu3.2;

libfreerdp/core/update.c in FreeRDP versions > 1.1 through 2.0.0-rc4 has an Out-of-bounds Read.

libfreerdp/cache/bitmap.c in FreeRDP versions > 1.0 through 2.0.0-rc4 has an Out of bounds read.

libfreerdp/codec/interleaved.c in FreeRDP versions > 1.0 through 2.0.0-rc4 has an Out-of-bounds Write.

libfreerdp/gdi/region.c in FreeRDP versions > 1.0 through 2.0.0-rc4 has an Integer Overflow.

libfreerdp/codec/planar.c in FreeRDP version > 1.0 through 2.0.0-rc4 has an Out-of-bounds Write.

Apache Ant 1.1 to 1.9.14 and 1.10.0 to 1.10.7 uses the default temporary directory identified by the Java system property java.io.tmpdir for several tasks and may thus leak sensitive information. The fixcrlf and replaceregexp tasks also copy files from the temporary directory back into the build tree allowing an attacker to inject modified source files into the build process.

In FreeRDP after 1.1 and before 2.0.0, a stream out-of-bounds seek in rdp_read_font_capability_set could lead to a later out-of-bounds read. As a result, a manipulated client or server might force a disconnect due to an invalid data read. This has been fixed in 2.0.0.

In FreeRDP after 1.1 and before 2.0.0, there is an out-of-bound read of client memory that is then passed on to the protocol parser. This has been patched in 2.0.0.

In FreeRDP after 1.0 and before 2.0.0, there is an out-of-bounds read. It only allows to abort a session. No data extraction is possible. This has been fixed in 2.0.0.

In FreeRDP after 1.1 and before 2.0.0, there is an out-of-bounds read in autodetect_recv_bandwidth_measure_results. A malicious server can extract up to 8 bytes of client memory with a manipulated message by providing a short input and reading the measurement result data. This has been patched in 2.0.0.

In the Linux kernel through 5.6.7 on the s390 platform, code execution may occur because of a race condition, as demonstrated by code in enable_sacf_uaccess in arch/s390/lib/uaccess.c that fails to protect against a concurrent page table upgrade, aka CID-3f777e19d171. A crash could also occur.

Apport reads and writes information on a crashed process to /proc/pid with elevated privileges. Apport then determines which user the crashed process belongs to by reading /proc/pid through get_pid_info() in data/apport. An unprivileged user could exploit this to read information about a privileged running process by exploiting PID recycling. This information could then be used to obtain ASLR offsets for a process with an existing memory corruption vulnerability. The initial fix introduced regressions in the Python Apport library due to a missing argument in Report.add_proc_environ in apport/report.py. It also caused an autopkgtest failure when reading /proc/pid and with Python 2 compatibility by reading /proc maps. The initial and subsequent regression fixes are in 2.20.11-0ubuntu16, 2.20.11-0ubuntu8.6, 2.20.9-0ubuntu7.12, 2.20.1-0ubuntu2.22 and 2.14.1-0ubuntu3.29+esm3.

Overlayfs in the Linux kernel and shiftfs, a non-upstream patch to the Linux kernel included in the Ubuntu 5.0 and 5.3 kernel series, both replace vma->vm_file in their mmap handlers. On error the original value is not restored, and the reference is put for the file to which vm_file points. On upstream kernels this is not an issue, as no callers dereference vm_file following after call_mmap() returns an error. However, the aufs patchs change mmap_region() to replace the fput() using a local variable with vma_fput(), which will fput() vm_file, leading to a refcount underflow.

Time-of-check Time-of-use Race Condition vulnerability on crash report ownership change in Apport allows for a possible privilege escalation opportunity. If fs.protected_symlinks is disabled, this can be exploited between the os.open and os.chown calls when the Apport cron script clears out crash files of size 0. A symlink with the same name as the deleted file can then be created upon which chown will be called, changing the file owner to root. Fixed in versions 2.20.1-0ubuntu2.23, 2.20.9-0ubuntu7.14, 2.20.11-0ubuntu8.8 and 2.20.11-0ubuntu22.

Apport creates a world writable lock file with root ownership in the world writable /var/lock/apport directory. If the apport/ directory does not exist (this is not uncommon as /var/lock is a tmpfs), it will create the directory, otherwise it will simply continue execution using the existing directory. This allows for a symlink attack if an attacker were to create a symlink at /var/lock/apport, changing apport's lock file location. This file could then be used to escalate privileges, for example. Fixed in versions 2.20.1-0ubuntu2.23, 2.20.9-0ubuntu7.14, 2.20.11-0ubuntu8.8 and 2.20.11-0ubuntu22.

re2c 1.3 has a heap-based buffer overflow in Scanner::fill in parse/scanner.cc via a long lexeme.

Affected versions of Git have a vulnerability whereby Git can be tricked into sending private credentials to a host controlled by an attacker. This bug is similar to CVE-2020-5260(GHSA-qm7j-c969-7j4q). The fix for that bug still left the door open for an exploit where _some_ credential is leaked (but the attacker cannot control which one). Git uses external "credential helper" programs to store and retrieve passwords or other credentials from secure storage provided by the operating system. Specially-crafted URLs that are considered illegal as of the recently published Git versions can cause Git to send a "blank" pattern to helpers, missing hostname and protocol fields. Many helpers will interpret this as matching _any_ URL, and will return some unspecified stored password, leaking the password to an attacker's server. The vulnerability can be triggered by feeding a malicious URL to `git clone`. However, the affected URLs look rather suspicious; the likely vector would be through systems which automatically clone URLs not visible to the user, such as Git submodules, or package systems built around Git. The root of the problem is in Git itself, which should not be feeding blank input to helpers. However, the ability to exploit the vulnerability in practice depends on which helpers are in use. Credential helpers which are known to trigger the vulnerability: - Git's "store" helper - Git's "cache" helper - the "osxkeychain" helper that ships in Git's "contrib" directory Credential helpers which are known to be safe even with vulnerable versions of Git: - Git Credential Manager for Windows Any helper not in this list should be assumed to trigger the vulnerability.

A use-after-free issue exists in WebKitGTK before 2.28.1 and WPE WebKit before 2.28.1 via crafted web content that allows remote attackers to execute arbitrary code or cause a denial of service (memory corruption and application crash).