Show Your Support: This site is primarily supported by advertisements. Ads are what have allowed this site to be maintained on a daily basis for the past 18+ years. We do our best to ensure only clean, relevant ads are shown, when any nasty ads are detected, we work to remove them ASAP. If you would like to view the site without ads while still supporting our work, please consider our ad-free Phoronix Premium.
Linux 5.15 To Bring More Scalable + Reliable Open vSwitch
Red Hat's work on per-CPU upcall dispatch handling for the Open vSwitch kernel code has made it into net-next this week, thereby effectively making it among the material now waiting around for the Linux 5.15 cycle.
The kernel code rework moves to an upcall mode using per-CPU dispatch instead of the existing per-vPort dispatch. Long story short, the performance within the Open vSwitch kernel module is more scalable with the number of Netlink sockets scaling to the number of CPUs available rather than the number of vPorts.
Besides greater scaling with today's large servers, the improved Open vSwitch code in net-next also addresses a packet reordering issue currently plaguing the code as well as a thundering herd issue.
The updated OVS code for the kernel is now in net-next. There is also updated OVS user-space patches for distributing upcalls to user-space threads.