netdev - Re: [PATCH net-next 01/13] net: sched: Pass qdisc reference in struct flow_block

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] [day] [month] [year] [list]

Message-ID: <87pn93fdv4.fsf@mellanox.com>
Date:   Fri, 10 Jul 2020 20:01:51 +0200
From:   Petr Machata <petrm@...lanox.com>
To:     Pablo Neira Ayuso <pablo@...filter.org>
Cc:     Ido Schimmel <idosch@...sch.org>, netdev@...r.kernel.org,
        davem@...emloft.net, kuba@...nel.org, jiri@...lanox.com,
        mlxsw@...lanox.com, michael.chan@...adcom.com, saeedm@...lanox.com,
        leon@...nel.org, kadlec@...filter.org, fw@...len.de,
        jhs@...atatu.com, xiyou.wangcong@...il.com,
        simon.horman@...ronome.com, Ido Schimmel <idosch@...lanox.com>
Subject: Re: [PATCH net-next 01/13] net: sched: Pass qdisc reference in struct flow_block_offload


Pablo Neira Ayuso <pablo@...filter.org> writes:

> On Fri, Jul 10, 2020 at 06:22:03PM +0200, Petr Machata wrote:
>> 
>> Pablo Neira Ayuso <pablo@...filter.org> writes:
>> 
>> > Moreover, the flow_offload infrastructure should also remain
>> > independent from the front-end, either tc/netfilter/ethtool, this is
>> > pulling in tc specific stuff into it, eg.
>> 
>> Hmm, OK, so I should not have assumed there is always a qdisc associated
>> with a block.
>> 
>> I'm not sure how strong your objection to pulling in TC is. Would it be
>> OK, instead of replacing the device with a qdisc in flow_block_indr, to
>> put in both? The qdisc can be NULL for the "normal" binder types,
>> because there the block is uniquely identified just by the type. For the
>> "non-normal" ones it would be obvious how to initialize it.
>
> Adding an extra field to flow_block_indr instead of replacing struct
> net_device should be ok for your new qevent use-case, right? This new
> Qdisc field will be NULL for the existing use-cases, that is what you
> mean, correct?

Yes, that is what I have in mind.

OK, I'll update the patches accordingly. Thanks!

> I still did not have a look at this new qevent infrastructure, so I
> don't have a better proposal right now.