In the Linux kernel, the following vulnerability has been resolved: netfilter: nf_conncount: increase the connection clean up limit to 64 After the optimization to only perform one GC per jiffy, a new problem was introduced. If more than 8 new connections are tracked per jiffy the list won't be cleaned up fast enough possibly reaching the limit wrongly. In order to prevent this issue, only skip the GC if it was already triggered during the same jiffy and the increment is lower than the clean up limit. In addition, increase the clean up limit to 64 connections to avoid triggering GC too often and do more effective GCs. This has been tested using a HTTP server and several performance tools while having nft_connlimit/xt_connlimit or OVS limit configured. Output of slowhttptest + OVS limit at 52000 connections: slow HTTP test status on 340th second: initializing: 0 pending: 432 connected: 51998 error: 0 closed: 0 service available: YES
https://git.kernel.org/stable/c/fa85432d58c8e74b39333edbf8d28df2985dfc79
https://git.kernel.org/stable/c/a5c9e14e0e8923218ae881d5e78c990c07694966
https://git.kernel.org/stable/c/6e5fa7add3e76da068a478d905be64be8fa4e80a
https://git.kernel.org/stable/c/3d0994ed0aa1fc0a2c5e620b765e8defdd021bff
https://git.kernel.org/stable/c/21d033e472735ecec677f1ae46d6740b5e47a4f3
https://git.kernel.org/stable/c/13eede458fdf231f1bf96a398feea4ad1553f14c
https://git.kernel.org/stable/c/0af0812baf2d363176c9b76fc07e33f13aede8db
https://git.kernel.org/stable/c/0792ad077d776c2dcf20f0484e2461ded1b77a24