Hi all, Gentle ping on this series - it has been quiet for a couple of months since v4 went out, so checking whether it is stuck in a review queue or just hasn't been picked up yet.
For context, v4 rebuilds the series around the CPU-by-CPU hotplug cycling approach that Thomas suggested during v3 review, replacing the RCU-swap-mask/callback mechanism from earlier versions. Patches 1-9 cover sched/isolation, tick/nohz, context_tracking, watchdog and the cpuset driver logic; 10-11 are docs and selftests. One thing worth flagging: v4 deliberately does not touch RCU/nocb offload state. That piece is still being worked out in a separate, ongoing discussion on Waiman's nohz_full/managed_irq series [1], where the open question is how to change the nocb mask without paying a stop_machine latency spike. We scoped it out of v4 so review of this series would not need to wait on that RCU decision - happy to fold in RCU/nocb support as a follow-up once that discussion converges. Any review, even partial (a single patch or a design-level concern), would be very helpful at this point. [1] https://lore.kernel.org/r/[email protected] Thanks, Jing Wu
