netdev - Re: [PATCH net-next 00/13] Control action percpu counters allocation by netlink flag

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Sun, 27 Oct 2019 08:31:32 +0000
From:   Vlad Buslov <vladbu@...lanox.com>
To:     Jamal Hadi Salim <jhs@...atatu.com>
CC:     Vlad Buslov <vladbu@...lanox.com>, Roman Mashak <mrv@...atatu.com>,
        Jiri Pirko <jiri@...nulli.us>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        "xiyou.wangcong@...il.com" <xiyou.wangcong@...il.com>,
        "davem@...emloft.net" <davem@...emloft.net>,
        "mleitner@...hat.com" <mleitner@...hat.com>,
        "dcaratti@...hat.com" <dcaratti@...hat.com>,
        Eric Dumazet <edumazet@...gle.com>
Subject: Re: [PATCH net-next 00/13] Control action percpu counters allocation
 by netlink flag


On Sat 26 Oct 2019 at 21:38, Jamal Hadi Salim <jhs@...atatu.com> wrote:
> On 2019-10-26 12:42 p.m., Vlad Buslov wrote:
>>
>> On Sat 26 Oct 2019 at 19:06, Jamal Hadi Salim <jhs@...atatu.com> wrote:
>
>> Hmm, yes, this looks quite redundant.
>
> It's legit.
> Two different code paths for two different objects
> that can be configured independently or together.
> Only other thing that does this is FIB/NH path.
>
>> At the same time having basically
>> same flags in two different spaces is ugly. Maybe correct approach would
>> be not to add act API at all and only extend tcf_action_init_1() to be
>> used from cls API? I don't see a use for act API at the moment because
>> the only flag value (skip percpu allocation) is hardware offloads
>> specific and such clients generally create action through cls new filter
>> API. WDYT?
>>
>
> I am not sure if there is a gain. Your code path is
> tcf_exts_validate()->tcf_action_init()->tcf_action_init_1()
>
> Do you want to replicate the tcf_action_init() in
> cls?
>
> cheers,
> jamal

No, I meant that we don't need to change iproute2 to always send new
TCA_ACT_ROOT_FLAGS. They can be set conditionally (only when user set
the flag) or only when creating filters with actions attached (cls API).
Such approach wouldn't introduce any additional overhead to netlink
packets when batch creating actions.

Regards,
Vlad