On 2020-10-15 10:11:31 [+0200], Eelco Chaudron wrote:
> Thanks for your reply! Yes I had it replaced with local_bh_disable() in my
> v2, as I noticed the hard IRQ to softirq part earlier.
Okay. Resend the complete thing once you ready and I take a look.
> Thanks,
>
> Eelco
Sebastian
On 15 Oct 2020, at 9:55, Sebastian Andrzej Siewior wrote:
On 2020-10-14 12:44:23 [+0200], Eelco Chaudron wrote:
Let me know your thoughts.
better. If your seccount is per-CPU then you get away without explicit
writer locking if you rely on global per-CPU locking. You can't do
preempt_disab
On 2020-10-14 12:44:23 [+0200], Eelco Chaudron wrote:
> Let me know your thoughts.
better. If your seccount is per-CPU then you get away without explicit
writer locking if you rely on global per-CPU locking. You can't do
preempt_disable() because this section can be interrupt by softirq. You
need
On 13 Oct 2020, at 14:53, Sebastian Andrzej Siewior wrote:
On 2020-10-13 14:44:19 [+0200], Eelco Chaudron wrote:
The flow_lookup() function uses per CPU variables, which must not be
preempted. However, this is fine in the general napi use case where
the local BH is disabled. But, it's also c
On 2020-10-13 14:44:19 [+0200], Eelco Chaudron wrote:
> The flow_lookup() function uses per CPU variables, which must not be
> preempted. However, this is fine in the general napi use case where
> the local BH is disabled. But, it's also called in the netlink
> context, which is preemptible. The be
The flow_lookup() function uses per CPU variables, which must not be
preempted. However, this is fine in the general napi use case where
the local BH is disabled. But, it's also called in the netlink
context, which is preemptible. The below patch makes sure that even
in the netlink path, preemption
