Re: [RFC PATCH v1] stop_machine: Add stop_housekeeping_cpuslocked()

[Costa Shulyupin]

On Mon, 6 Jan 2025 at 23:25, Frederic Weisbecker <frederic@xxxxxxxxxx> wrote:
> I must confess I don't understand well your constraints. Why would you
> change the set of isolated CPUs while running realtime tests (btw. did you mean
> "tasks"?).
>
> Do I understand it correctly that your server may run different kinds of
> workloads concurrently, some of them isolated and some of them not, and these
> workloads may be added / removed concurrently _anytime_? And therefore
> a newly added isolated workload (which then adds CPUs to the isolated set)
> mustn't disturb unrelated already running isolated workloads?
Right.

Our client telco companies deploy 5G RAN low-latency realtime
applications on our platform.

Typical System Stack:

Workload: DPDK runs in containers on isolated CPUs.
Orchestration: OpenShift/Kubernetes runs on housekeeping CPUs.
Operating System: Linux with real-time preemption and statically
configured CPU isolation.
Hardware: Intel CPUs with 64 or more hyperthreads.

Current Limitations:

Currently, CPU isolation can only be configured statically. Changing
CPU isolation requires modifying boot arguments and rebooting the
system. The objective is to enable dynamic CPU isolation configuration
without requiring a reboot.

We conduct various realtime latency tests, including RTLA, to ensure
latency remains below 30 µs.

> When you refer to realtime tasks, do you mean isolated?
Realtime tasks require realtime preemption and CPU isolation.

Thank you,
Costa