In Silicon Labs uC/TCP-IP 3.6.0, TCP ISNs are improperly random.

In Oryx CycloneTCP 1.9.6, TCP ISNs are improperly random.

An issue was discovered in picoTCP through 1.7.0. The DNS domain name record decompression functionality in picodnsdecompressname() in picodnscommon.c does not validate the compression pointer offset values with respect to the actual data present in a DNS response packet, causing out-of-bounds writes that lead to Denial-of-Service and Remote Code Execution.

An issue was discovered in the IPv6 stack in Contiki through 3.0. There is an insufficient check for the IPv6 header length. This leads to Denial-of-Service and potential Remote Code Execution via a crafted ICMPv6 echo packet.

An issue was discovered in the DNS implementation in Ethernut in Nut/OS 5.1. There is no check on whether a domain name has '\0' termination. This may lead to successful Denial-of-Service, and possibly Remote Code Execution.

An issue was discovered in the IPv6 stack in Contiki through 3.0. There are inconsistent checks for IPv6 header extension lengths. This leads to Denial-of-Service and potential Remote Code Execution via a crafted ICMPv6 echo packet.

An issue was discovered in the DNS implementation in Ethernut in Nut/OS 5.1. The number of DNS queries/responses (set in a DNS header) is not checked against the data present. This may lead to successful Denial-of-Service, and possibly Remote Code Execution.

An issue was discovered in the DNS implementation in Ethernut in Nut/OS 5.1. The length byte of a domain name in a DNS query/response is not checked, and is used for internal memory operations. This may lead to successful Denial-of-Service, and possibly Remote Code Execution.

An issue was discovered in the DNS implementation in Ethernut in Nut/OS 5.1. The DNS response data length is not checked (it can be set to an arbitrary value from a packet). This may lead to successful Denial-of-Service, and possibly Remote Code Execution.

An issue was discovered in Contiki through 3.0 and Contiki-NG through 4.5. The code for parsing Type A domain name answers in ip64-dns64.c doesn't verify whether the address in the answer's length is sane. Therefore, when copying an address of an arbitrary length, a buffer overflow can occur. This bug can be exploited whenever NAT64 is enabled.

An issue was discovered in uIP 1.0, as used in Contiki 3.0 and other products. The code that reassembles fragmented packets fails to properly validate the total length of an incoming packet specified in its IP header, as well as the fragmentation offset value specified in the IP header. By crafting a packet with specific values of the IP header length and the fragmentation offset, attackers can write into the .bss section of the program (past the statically allocated buffer that is used for storing the fragmented data) and cause a denial of service in uipreass() in uip.c, or possibly execute arbitrary code on some target architectures.

In Microchip MPLAB Net 3.6.1, TCP ISNs are improperly random.

In Contiki 4.5, TCP ISNs are improperly random.

In FNET 4.6.3, TCP ISNs are improperly random.

In PicoTCP 1.7.0, TCP ISNs are improperly random.

An issue was discovered in FNET through 4.6.4. The code for processing the hostname from an LLMNR request doesn't check for '\0' termination. Therefore, the deduced length of the hostname doesn't reflect the correct length of the actual data. This may lead to Information Disclosure in fnetllmnrpoll in fnetllmnr.c during a response to a malicious request of the DNS class IN.

An issue was discovered in FNET through 4.6.4. The code for processing resource records in mDNS queries doesn't check for proper '\0' termination of the resource record name string, leading to an out-of-bounds read, and potentially causing information leak or Denial-or-Service.

An issue was discovered in picoTCP and picoTCP-NG through 1.7.0. The TCP input data processing function in picotcp.c does not validate the length of incoming TCP packets, which leads to an out-of-bounds read when assembling received packets into a data segment, eventually causing Denial-of-Service or an information leak.

An issue was discovered in picoTCP 1.7.0. The code for processing the IPv6 headers does not validate whether the IPv6 payload length field is equal to the actual size of the payload, which leads to an Out-of-Bounds read during the ICMPv6 checksum calculation, resulting in either Denial-of-Service or Information Disclosure. This affects picoipv6extensionheaders and picochecksumadder (in picoipv6.c and picoframe.c).

An issue was discovered in uIP 1.0, as used in Contiki 3.0 and other products. The code that parses incoming DNS packets does not validate that the incoming DNS replies match outgoing DNS queries in newdata() in resolv.c. Also, arbitrary DNS replies are parsed if there was any outgoing DNS query with a transaction ID that matches the transaction ID of an incoming reply. Provided that the default DNS cache is quite small (only four records) and that the transaction ID has a very limited set of values that is quite easy to guess, this can lead to DNS cache poisoning.

The code that processes DNS responses in uIP through 1.0, as used in Contiki and Contiki-NG, does not check whether the number of responses specified in the DNS packet header corresponds to the response data available in the DNS packet, leading to an out-of-bounds read and Denial-of-Service in resolv.c.

An issue was discovered in picoTCP and picoTCP-NG through 1.7.0. The code that processes DNS responses in picomdnshandledataasanswersgeneric() in picomdns.c does not check whether the number of answers/responses specified in a DNS packet header corresponds to the response data available in the packet, leading to an out-of-bounds read, invalid pointer dereference, and Denial-of-Service.

An issue was discovered in picoTCP 1.7.0. The code for creating an ICMPv6 echo replies doesn't check whether the ICMPv6 echo request packet's size is shorter than 8 bytes. If the size of the incoming ICMPv6 request packet is shorter than this, the operation that calculates the size of the ICMPv6 echo replies has an integer wrap around, leading to memory corruption and, eventually, Denial-of-Service in picoicmp6sendechoreplynotfrag in picoicmp6.c.

An issue was discovered in uIP 1.0, as used in Contiki 3.0 and other products. When the Urgent flag is set in a TCP packet, and the stack is configured to ignore the urgent data, the stack attempts to use the value of the Urgent pointer bytes to separate the Urgent data from the normal data, by calculating the offset at which the normal data should be present in the global buffer. However, the length of this offset is not checked; therefore, for large values of the Urgent pointer bytes, the data pointer can point to memory that is way beyond the data buffer in uipprocess in uip.c.

An issue was discovered in Contiki through 3.0. An Out-of-Bounds Read vulnerability exists in the uIP TCP/IP Stack component when calculating the checksums for IP packets in upperlayerchksum in net/ipv4/uip.c.

In SIMATIC MV400 family versions prior to v7.0.6, the ISN generator is initialized with a constant value and has constant increments. An attacker could predict and hijack TCP sessions.

An issue was discovered in Ethernut Nut/OS 5.1. The code that generates Initial Sequence Numbers (ISNs) for TCP connections derives the ISN from an insufficiently random source. As a result, an attacker may be able to determine the ISN of current and future TCP connections and either hijack existing ones or spoof future ones. While the ISN generator seems to adhere to RFC 793 (where a global 32-bit counter is incremented roughly every 4 microseconds), proper ISN generation should aim to follow at least the specifications outlined in RFC 6528.

An issue was discovered in FNET through 4.6.4. The code for IPv6 fragment reassembly tries to access a previous fragment starting from a network incoming fragment that still doesn't have a reference to the previous one (which supposedly resides in the reassembly list). When faced with an incoming fragment that belongs to a non-empty fragment list, IPv6 reassembly must check that there are no empty holes between the fragments: this leads to an uninitialized pointer dereference in fnetip6reassembly in fnetip6.c, and causes Denial-of-Service.

An issue was discovered in picoTCP and picoTCP-NG through 1.7.0. When an unsupported TCP option with zero length is provided in an incoming TCP packet, it is possible to cause a Denial-of-Service by achieving an infinite loop in the code that parses TCP options, aka tcpparseoptions() in picotcp.c.

An issue was discovered in picoTCP 1.7.0. The code for parsing the hop-by-hop IPv6 extension headers does not validate the bounds of the extension header length value, which may result in Integer Wraparound. Therefore, a crafted extension header length value may cause Denial-of-Service because it affects the loop in which the extension headers are parsed in picoipv6processhopbyhop() in picoipv6.c.