Re: [PATCH v2 09/12] cgroup/cpuset: Introduce CPUSet-driven dynamic housekeeping (DHM)

[Frederic Weisbecker <frederic@kernel.org>]

Le Mon, Apr 13, 2026 at 03:43:15PM +0800, Qiliang Yuan a écrit :
> Currently, subsystem housekeeping masks are generally static and can
> only be configured via boot-time parameters (e.g., isolcpus, nohz_full).
> This inflexible approach forces a system reboot whenever an orchestrator
> needs to change workload isolation boundaries.
> 
> This patch introduces CPUSet-driven Dynamic Housekeeping Management (DHM)
> by exposing the `cpuset.housekeeping.cpus` control file on the root cgroup.
> Writing a new cpumask to this file dynamically updates the housekeeping
> masks of all registered subsystems (scheduler, RCU, timers, tick, workqueues,
> and managed IRQs) simultaneously, without restarting the node.

There is already the "isolated" partition type which does scheduler, timers
and workqueues isolation. Shouldn't we extend that to dynamically apply nohz_full
instead of adding a new unrelated file?

I don't know which form that should take. Perhaps reuse the "isolated" partition
but add some sort of parameter to define if we want only domain isolation or
also full isolation (that is nohz_full). Waiman should have a better idea for an
interface here.

> 
> At the cpuset and isolation core level, this change implements:
> 1. `housekeeping_update_all_types(const struct cpumask *new_mask)` API inside
>    `isolation.c` to safely allocate, update, and replace all enabled hk_type
> masks.

HK_TYPE_DOMAIN is handled by "isolated" partitions. What remains to handle
is HK_TYPE_KERNEL_NOISE.

As for managed IRQs this will require more thinking but we should include that
into "full isolation" in the future.

> +int housekeeping_update_all_types(const struct cpumask *new_mask)

Please reuse housekeeping_update().

Thanks.

-- 
Frederic Weisbecker
SUSE Labs