On 06/13/2019 09:06 AM, Vlad Buslov wrote:
> On Wed 12 Jun 2019 at 19:33, Daniel Borkmann <dan...@iogearbox.net> wrote:
>> On 06/12/2019 09:14 AM, Vlad Buslov wrote:
>>> To remove rtnl lock dependency in tc filter update API when using ingress
>>> Qdisc, set QDISC_CLASS_OPS_DOIT_UNLOCKED flag in ingress Qdisc_class_ops.
>>>
>>> Ingress Qdisc ops don't require any modifications to be used without rtnl
>>> lock on tc filter update path. Ingress implementation never changes its
>>> q->block and only releases it when Qdisc is being destroyed. This means it
>>> is enough for RTM_{NEWTFILTER|DELTFILTER|GETTFILTER} message handlers to
>>> hold ingress Qdisc reference while using it without relying on rtnl lock
>>> protection. Unlocked Qdisc ops support is already implemented in filter
>>> update path by unlocked cls API patch set.
>>>
>>> Signed-off-by: Vlad Buslov <vla...@mellanox.com>
>>> ---
>>> net/sched/sch_ingress.c | 1 +
>>> 1 file changed, 1 insertion(+)
>>>
>>> diff --git a/net/sched/sch_ingress.c b/net/sched/sch_ingress.c
>>> index 0f65f617756b..d5382554e281 100644
>>> --- a/net/sched/sch_ingress.c
>>> +++ b/net/sched/sch_ingress.c
>>> @@ -114,6 +114,7 @@ static int ingress_dump(struct Qdisc *sch, struct
>>> sk_buff *skb)
>>> }
>>>
>>> static const struct Qdisc_class_ops ingress_class_ops = {
>>> + .flags = QDISC_CLASS_OPS_DOIT_UNLOCKED,
>>> .leaf = ingress_leaf,
>>> .find = ingress_find,
>>> .walk = ingress_walk,
>>>
>>
>> Vlad, why is clsact_class_ops not updated here? Please elaborate!
>
> Daniel, no particular reason to not enable unlocked execution for
> clsact. I set the unlocked flag for ingress because that was the Qdisc
> that I tested all my parallel tc changes on. However, ingress and clsact
> implementations are quite similar, so I can send a followup patch that
> updates clsact_class_ops, if you want.
Yes, please do, they are pretty much the same except that clsact supersedes
ingress in that it also offers an egress hook.
