Re: [PATCH net v2] usbnet: fix smp_processor_id() use in preemptible context

On Mon,  7 Sep 2026 21:52:20 +0300
Ömer Mete Kaya [off-list ref] wrote:

Hi Oliver,

Thank you for the review. You are right.

I used local_irq_save()/restore() thinking that softirq writers could
race on tstats. After checking the tstats update paths, I agreed that
preemption protection is sufficient for these accesses.


I will send v3 using preempt_disable()/preempt_enable() as suggested.

Shouldn't you use get_cpu_ptr() and put_cpu_ptr() ?

David

Best regards,

Ömer Mete Kaya