{"id":123598,"date":"2025-04-18T15:05:14","date_gmt":"2025-04-18T13:05:14","guid":{"rendered":"https:\/\/prohoster.info\/blog\/novosti-interneta\/v-yadre-linux-6-16-budet-prekrashhena-podderzhka-protokola-dccp"},"modified":"2025-04-18T15:05:14","modified_gmt":"2025-04-18T13:05:14","slug":"v-yadre-linux-6-16-budet-prekrashhena-podderzhka-protokola-dccp","status":"publish","type":"post","link":"https:\/\/prohoster.info\/en\/blog\/news\/v-yadre-linux-6-16-budet-prekrashhena-podderzhka-protokola-dccp","title":{"rendered":"Support for the DCCP Protocol Will Be Discontinued in Linux Kernel 6.16","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"
The net-next branch, where changes are accumulated for the upcoming release of the Linux kernel 6.16, has accepted a change that removes support for the DCCP (Datagram Congestion Control Protocol). Support for netfilter modules for filtering DCCP packets has been retained. The DCCP protocol was standardized and accepted into the kernel in 2006. It was intended that DCCP would improve the efficiency of streaming and VoIP in congested networks, but in practice, the protocol proved to be unpopular and did not gain traction. <\/p>\n
The DCCP code in the kernel has been unmaintained for five years, and changes in recent years have been limited to fixes due to API refactoring. A few years ago, as part of the Multipath DCCP project, an attempt was made to further develop the DCCP protocol in the kernel, but part of the code from this project remains proprietary. The developer of Multipath DCCP intended to open the proprietary code, transfer the accumulated changes to the main part of the kernel, and take on the responsibilities of maintaining DCCP in the Linux kernel, but he has not been in touch for several years. <\/p>\n
Removing DCCP from the Linux kernel will eliminate barriers that hinder the redesign of the inet_connection_sock data structure to improve the efficiency of the TCP stack. Currently, the inet_connection_sock structure is shared between TCP and DCCP, which prevents some optimizations for TCP without reworking the DCCP code. In particular, redesigning the inet_connection_sock structure will enable more efficient use of the CPU cache for accelerated packet processing for an already established TCP connection (TCP fastpath). \n Source: opennet.ru<\/a> <\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"