The SSH transport protocol with certain OpenSSH extensions, found in OpenSSH before 9.6 and other products, allows remote attackers to bypass integrity checks such that some packets are omitted (from the extension negotiation message), and a client and server may consequently end up with a connection for which some security features have been downgraded or disabled, aka a Terrapin attack. This occurs because the SSH Binary Packet Protocol (BPP), implemented by these extensions, mishandles the handshake phase and mishandles use of sequence numbers. For example, there is an effective attack against SSH's use of ChaCha20-Poly1305 (and CBC with Encrypt-then-MAC). The bypass occurs in chacha20-poly1305@openssh.com and (if CBC is used) the -etm@openssh.com MAC algorithms. This also affects Maverick Synergy Java SSH API before 3.1.0-SNAPSHOT, Dropbear through 2022.83, Ssh before 5.1.1 in Erlang/OTP, PuTTY before 0.80, AsyncSSH before 2.14.2, golang.org/x/crypto before 0.17.0, libssh before 0.10.6, libssh2 through 1.11.0, Thorn Tech SFTP Gateway before 3.4.6, Tera Term before 5.1, Paramiko before 3.4.0, jsch before 0.2.15, SFTPGo before 2.5.6, Netgate pfSense Plus through 23.09.1, Netgate pfSense CE through 2.7.2, HPN-SSH through 18.2.0, ProFTPD before 1.3.8b (and before 1.3.9rc2), ORYX CycloneSSH before 2.3.4, NetSarang XShell 7 before Build 0144, CrushFTP before 10.6.0, ConnectBot SSH library before 2.2.22, Apache MINA sshd through 2.11.0, sshj through 0.37.0, TinySSH through 20230101, trilead-ssh2 6401, LANCOM LCOS and LANconfig, FileZilla before 3.66.4, Nova before 11.8, PKIX-SSH before 14.4, SecureCRT before 9.4.3, Transmit5 before 5.10.4, Win32-OpenSSH before 9.5.0.0p1-Beta, WinSCP before 6.2.2, Bitvise SSH Server before 9.32, Bitvise SSH Client before 9.33, KiTTY through 0.76.1.13, the net-ssh gem 7.2.0 for Ruby, the mscdex ssh2 module before 1.15.0 for Node.js, the thrussh library before 0.35.1 for Rust, and the Russh crate before 0.40.2 for Rust.

In versions of FreeBSD 12.4-RELEASE prior to 12.4-RELEASE-p7 and FreeBSD 13.2-RELEASE prior to 13.2-RELEASE-p5 the __sflush() stdio function in libc does not correctly update FILE objects' write space members for write-buffered streams when the write(2) system call returns an error.  Depending on the nature of an application that calls libc's stdio functions and the presence of errors returned from the write(2) system call (or an overridden stdio write routine) a heap buffer overflow may occur. Such overflows may lead to data corruption or the execution of arbitrary code at the privilege level of the calling program.

On an msdosfs filesystem, the 'truncate' or 'ftruncate' system calls under certain circumstances populate the additional space in the file with unallocated data from the underlying disk device, rather than zero bytes. This may permit a user with write access to files on a msdosfs filesystem to read unintended data (e.g. from a previously deleted file).

In pf packet processing with a 'scrub fragment reassemble' rule, a packet containing multiple IPv6 fragment headers would be reassembled, and then immediately processed. That is, a packet with multiple fragment extension headers would not be recognized as the correct ultimate payload. Instead a packet with multiple IPv6 fragment headers would unexpectedly be interpreted as a fragmented packet, rather than as whatever the real payload is. As a result, IPv6 fragments may bypass pf firewall rules written on the assumption all fragments have been reassembled and, as a result, be forwarded or processed by the host.

pam_krb5 authenticates a user by essentially running kinit with the password, getting a ticket-granting ticket (tgt) from the Kerberos KDC (Key Distribution Center) over the network, as a way to verify the password. However, if a keytab is not provisioned on the system, pam_krb5 has no way to validate the response from the KDC, and essentially trusts the tgt provided over the network as being valid. In a non-default FreeBSD installation that leverages pam_krb5 for authentication and does not have a keytab provisioned, an attacker that is able to control both the password and the KDC responses can return a valid tgt, allowing authentication to occur for any user on the system.

In FreeBSD 13.0-STABLE before n247428-9352de39c3dc, 12.2-STABLE before r370674, 13.0-RELEASE before p6, and 12.2-RELEASE before p12, certain conditions involving use of the highlight buffer while text is scrolling on the console, console data may overwrite data structures associated with the system console or other kernel memory.

In FreeBSD 13.0-STABLE before n246938-0729ba2f49c9, 12.2-STABLE before r370383, 11.4-STABLE before r370381, 13.0-RELEASE before p4, 12.2-RELEASE before p10, and 11.4-RELEASE before p13, the ggatec daemon does not validate the size of a response before writing it to a fixed-sized buffer allowing a malicious attacker in a privileged network position to overwrite the stack of ggatec and potentially execute arbitrary code.

In FreeBSD 13.0-STABLE before n246941-20f96f215562, 12.2-STABLE before r370400, 11.4-STABLE before r370399, 13.0-RELEASE before p4, 12.2-RELEASE before p10, and 11.4-RELEASE before p13, certain VirtIO-based device models in bhyve failed to handle errors when fetching I/O descriptors. A malicious guest may cause the device model to operate on uninitialized I/O vectors leading to memory corruption, crashing of the bhyve process, and possibly arbitrary code execution in the bhyve process.

In FreeBSD 13.0-STABLE before n245765-bec0d2c9c841, 12.2-STABLE before r369859, 11.4-STABLE before r369866, 13.0-RELEASE before p1, 12.2-RELEASE before p7, and 11.4-RELEASE before p10, missing message validation in libradius(3) could allow malicious clients or servers to trigger denial of service in vulnerable servers or clients respectively.

In FreeBSD 13.0-STABLE before n245764-876ffe28796c, 12.2-STABLE before r369857, 13.0-RELEASE before p1, and 12.2-RELEASE before p7, a system call triggering a fault could cause SMAP protections to be disabled for the duration of the system call. This weakness could be combined with other kernel bugs to craft an exploit.

In FreeBSD 13.0-STABLE before n245050, 12.2-STABLE before r369525, 13.0-RC4 before p0, and 12.2-RELEASE before p6, listening socket accept filters implementing the accf_create callback incorrectly freed a process supplied argument string. Additional operations on the socket can lead to a double free or use after free.

In FreeBSD 13.0-STABLE before n245117, 12.2-STABLE before r369551, 11.4-STABLE before r369559, 13.0-RC5 before p1, 12.2-RELEASE before p6, and 11.4-RELEASE before p9, copy-on-write logic failed to invalidate shared memory page mappings between multiple processes allowing an unprivileged process to maintain a mapping after it is freed, allowing the process to read private data belonging to other processes or the kernel.

In FreeBSD 13.0-STABLE before n245118, 12.2-STABLE before r369552, 11.4-STABLE before r369560, 13.0-RC5 before p1, 12.2-RELEASE before p6, and 11.4-RELEASE before p9, a superuser inside a FreeBSD jail configured with the non-default allow.mount permission could cause a race condition between the lookup of ".." and remounting a filesystem, allowing access to filesystem hierarchy outside of the jail.

In FreeBSD 12.2-STABLE before r369334, 11.4-STABLE before r369335, 12.2-RELEASE before p4 and 11.4-RELEASE before p8 when a process, such as jexec(8) or killall(1), calls jail_attach(2) to enter a jail, the jailed root can attach to it using ptrace(2) before the current working directory is changed.

In FreeBSD 12.2-STABLE before r369312, 11.4-STABLE before r369313, 12.2-RELEASE before p4 and 11.4-RELEASE before p8 due to a race condition in the jail_remove(2) implementation, it may fail to kill some of the processes.

In FreeBSD 12.2-STABLE before r369346, 11.4-STABLE before r369345, 12.2-RELEASE before p4 and 11.4-RELEASE before p8 a regression in the login.access(5) rule processor has the effect of causing rules to fail to match even when they should not. This means that rules denying access may be ignored.

In FreeBSD 12.2-STABLE before r368969, 11.4-STABLE before r369047, 12.2-RELEASE before p3, 12.1-RELEASE before p13 and 11.4-RELEASE before p7 msdosfs(5) was failing to zero-fill a pair of padding fields in the dirent structure, resulting in a leak of three uninitialized bytes.

In FreeBSD 12.2-STABLE before r368969, 11.4-STABLE before r369047, 12.2-RELEASE before p3, 12.1-RELEASE before p13 and 11.4-RELEASE before p7 several file systems were not properly initializing the d_off field of the dirent structures returned by VOP_READDIR. In particular, tmpfs(5), smbfs(5), autofs(5) and mqueuefs(5) were failing to do so. As a result, eight uninitialized kernel stack bytes may be leaked to userspace by these file systems.

The X509_V_FLAG_X509_STRICT flag enables additional security checks of the certificates present in a certificate chain. It is not set by default. Starting from OpenSSL version 1.1.1h a check to disallow certificates in the chain that have explicitly encoded elliptic curve parameters was added as an additional strict check. An error in the implementation of this check meant that the result of a previous check to confirm that certificates in the chain are valid CA certificates was overwritten. This effectively bypasses the check that non-CA certificates must not be able to issue other certificates. If a "purpose" has been configured then there is a subsequent opportunity for checks that the certificate is a valid CA. All of the named "purpose" values implemented in libcrypto perform this check. Therefore, where a purpose is set the certificate chain will still be rejected even when the strict flag has been used. A purpose is set by default in libssl client and server certificate verification routines, but it can be overridden or removed by an application. In order to be affected, an application must explicitly set the X509_V_FLAG_X509_STRICT verification flag and either not set a purpose for the certificate verification or, in the case of TLS client or server applications, override the default purpose. OpenSSL versions 1.1.1h and newer are affected by this issue. Users of these versions should upgrade to OpenSSL 1.1.1k. OpenSSL 1.0.2 is not impacted by this issue. Fixed in OpenSSL 1.1.1k (Affected 1.1.1h-1.1.1j).

An OpenSSL TLS server may crash if sent a maliciously crafted renegotiation ClientHello message from a client. If a TLSv1.2 renegotiation ClientHello omits the signature_algorithms extension (where it was present in the initial ClientHello), but includes a signature_algorithms_cert extension then a NULL pointer dereference will result, leading to a crash and a denial of service attack. A server is only vulnerable if it has TLSv1.2 and renegotiation enabled (which is the default configuration). OpenSSL TLS clients are not impacted by this issue. All OpenSSL 1.1.1 versions are affected by this issue. Users of these versions should upgrade to OpenSSL 1.1.1k. OpenSSL 1.0.2 is not impacted by this issue. Fixed in OpenSSL 1.1.1k (Affected 1.1.1-1.1.1j).