IPv6 routing headers issue

The IPv6 protocol allows remote attackers to cause a denial of service via crafted IPv6 type 0 route headers (IPV6_RTHDR_TYPE_0) that create network amplification between two routers.

The display driver allocattr functions in NetBSD 3.0 through 4.0_BETA2, and NetBSD-current before 20070728, allow local users to cause a denial of service (panic) via a (1) negative or (2) large value in an ioctl call, as demonstrated by the vga_allocattr function.

A certain pseudo-random number generator (PRNG) algorithm that uses XOR and 2-bit random hops (aka "Algorithm X2"), as used in OpenBSD 2.6 through 3.4, Mac OS X 10 through 10.5.1, FreeBSD 4.4 through 7.0, and DragonFlyBSD 1.0 through 1.10.1, allows remote attackers to guess sensitive values such as IP fragmentation IDs by observing a sequence of previously generated values. NOTE: this issue can be leveraged for attacks such as injection into TCP packets and OS fingerprinting.

A certain pseudo-random number generator (PRNG) algorithm that uses ADD with 0 random hops (aka "Algorithm A0"), as used in OpenBSD 3.5 through 4.2 and NetBSD 1.6.2 through 4.0, allows remote attackers to guess sensitive values such as (1) DNS transaction IDs or (2) IP fragmentation IDs by observing a sequence of previously generated values. NOTE: this issue can be leveraged for attacks such as DNS cache poisoning, injection into TCP packets, and OS fingerprinting.

The ipsec4_get_ulp function in the kernel in NetBSD 2.0 through 3.1 and NetBSD-current before 20071028, when the fast_ipsec subsystem is enabled, allows remote attackers to bypass the IPsec policy by sending packets from a source machine with a different endianness than the destination machine, a different vulnerability than CVE-2006-0905.

The BSD make program allows local users to modify files via a symlink attack when the -j option is being used.

Operating systems with shared memory implementations based on BSD 4.4 code allow a user to conduct a denial of service and bypass memory limits (e.g., as specified with rlimits) using mmap or shmget to allocate memory and cause page faults.

A race condition between the select() and accept() calls in NetBSD TCP servers allows remote attackers to cause a denial of service.

The at program in IRIX 6.2 and NetBSD 1.3.2 and earlier allows local users to read portions of arbitrary files by submitting the file to at with the -f argument, which generates error messages that at sends to the user via e-mail.

rpc.mountd on Linux, Ultrix, and possibly other operating systems, allows remote attackers to determine the existence of a file on the server by attempting to mount that file, which generates different error messages depending on whether the file exists or not.

NetBSD netstat command allows local users to access kernel memory.

Buffer overflow of rlogin program using TERM environmental variable.

The (1) dump and (2) dump_lfs commands in NetBSD 1.4.x through 1.5.1 do not properly drop privileges, which could allow local users to gain privileges via the RCMD_CMD environment variable.

sendmsg function in NetBSD 1.3 through 1.5 allows local users to cause a denial of service (kernel trap or panic) via a msghdr structure with a large msg_controllen length.

fts routines in FreeBSD 4.3 and earlier, NetBSD before 1.5.2, and OpenBSD 2.9 and earlier can be forced to change (chdir) into a different directory than intended when the directory above the current directory is moved, which could cause scripts to perform dangerous actions on the wrong directories.

Multiple TCP implementations could allow remote attackers to cause a denial of service (bandwidth and CPU exhaustion) by setting the maximum segment size (MSS) to a very small number and requesting large amounts of data, which generates more packets with less TCP-level data that amplify network traffic and consume more server CPU to process.

security flaw

Heap corruption vulnerability in the "at" program allows local users to execute arbitrary code via a malformed execution time, which causes at to free the same memory twice.

The TCP implementation in various BSD operating systems (tcp_input.c) does not properly block connections to broadcast addresses, which could allow remote attackers to bypass intended filters via packets with a unicast link layer address and an IP broadcast address.

KAME-derived implementations of IPsec on NetBSD 1.5.2, FreeBSD 4.5, and other operating systems, does not properly consult the Security Policy Database (SPD), which could cause a Security Gateway (SG) that does not use Encapsulating Security Payload (ESP) to forward forged IPv4 packets.

The asynchronous I/O facility in 4.4 BSD kernel does not check user credentials when setting the recipient of I/O notification, which allows local users to cause a denial of service by using certain ioctl and fcntl calls to cause the signal to be sent to an arbitrary process ID.