On Thu, 2017-10-26 at 18:24 -0700, Cong Wang wrote:
> ...
> On the other hand, this makes tcf_block_put() ugly and
> harder to understand. Since David and Eric strongly dislike
> adding synchronize_rcu(), this is probably the only
> solution that could make everyone happy.
...
> +static void tcf_block_put_deferred(struct work_struct *work)
> +{
> + struct tcf_block *block = container_of(work, struct tcf_block, work);
> + struct tcf_chain *chain;
>
> + rtnl_lock();
> /* Hold a refcnt for all chains, except 0, in case they are gone. */
> list_for_each_entry(chain, &block->chain_list, list)
> if (chain->index)
> @@ -292,13 +308,27 @@ void tcf_block_put(struct tcf_block *block)
> list_for_each_entry(chain, &block->chain_list, list)
> tcf_chain_flush(chain);
>
> - /* Wait for RCU callbacks to release the reference count. */
> + INIT_WORK(&block->work, tcf_block_put_final);
> + /* Wait for RCU callbacks to release the reference count and make
> + * sure their works have been queued before this.
> + */
> rcu_barrier();
> + tcf_queue_work(&block->work);
> + rtnl_unlock();
> +}
On a loaded server, rcu_barrier() typically takes 4 ms.
Way better than synchronize_rcu() (about 90 ms) but still an issue when
holding RTNL.
We have thousands of filters, and management daemon restarts and rebuild
TC hierarchy from scratch.
Simply getting rid of 1000 old filters might block RTNL for a while, or
maybe I misunderstood your patches.
Thanks.
