Re: [PATCH 1/2] genirq: Fix IRQ threads affinity VS cpuset isolated partitions

[Thomas Gleixner]

On Thu, Nov 20 2025 at 16:50, Frederic Weisbecker wrote:
> Le Thu, Nov 20, 2025 at 04:00:39PM +0100, Thomas Gleixner a écrit :
>> +	if (!secondary)
>> +		t = kthread_create_on_cpu(irq_thread, new, cpu, "irq/%d-%s", irq, new->name);
>> +	else
>> +		t = kthread_create_on_cpu(irq_thread, new, cpu, "irq/%d-s-%s",
>> -		irq, new->name);
>
> Right I though about something like that, it involved:
>
> kthread_bind_mask(t, cpu_possible_mask);

That's way simpler and also solves the problem with the
kthread_create_on_cpu() name which Marek pointed out.

> Which do you prefer? Also do you prefer such a fixup or should I refactor my
> patches you merged?

Can you split out the wakeup change into a separate patch
(Suggested-by-me) with it's own change log and fold the
kthread_bind_mask() + set(AFFINITY) bit into this one.

I just go and zap the existing commits (they are on top of the branch).

Thanks,

        tglx