On Fri, Nov 8, 2019 at 12:53 PM Eric Dumazet <edumazet@xxxxxxxxxx> wrote:
>
> per cpu SNMP counters mostly, with no IRQ safety requirements.
>
> Note that this could be implemented using local{64}_add() on arches like x86_64,
> while others might have to fallback to WRITE_ONCE(variable, variable + add)
raw_cpu_add()?
We already use those for vm_counters where we intentionally accept races.
Linus
