[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <d805bea3-cb2f-4e2c-a07a-27b8b4c5f294@intel.com>
Date: Wed, 31 Jul 2024 09:41:16 -0700
From: Jacob Keller <jacob.e.keller@...el.com>
To: Jakub Kicinski <kuba@...nel.org>, "Song, Yoong Siang"
<yoong.siang.song@...el.com>
CC: "Neftin, Sasha" <sasha.neftin@...el.com>, Brett Creeley
<brett.creeley@....com>, "linux-doc@...r.kernel.org"
<linux-doc@...r.kernel.org>, Alexei Starovoitov <ast@...nel.org>, "Eric
Dumazet" <edumazet@...gle.com>, "Nguyen, Anthony L"
<anthony.l.nguyen@...el.com>, "Blanco Alcaine, Hector"
<hector.blanco.alcaine@...el.com>, Daniel Borkmann <daniel@...earbox.net>,
Jonathan Corbet <corbet@....net>, "Gomes, Vinicius"
<vinicius.gomes@...el.com>, "Kitszel, Przemyslaw"
<przemyslaw.kitszel@...el.com>, John Fastabend <john.fastabend@...il.com>,
Shinas Rasheed <srasheed@...vell.com>, "intel-wired-lan@...ts.osuosl.org"
<intel-wired-lan@...ts.osuosl.org>, Paolo Abeni <pabeni@...hat.com>, "Tian,
Kevin" <kevin.tian@...el.com>, Jesper Dangaard Brouer <hawk@...nel.org>,
Richard Cochran <richardcochran@...il.com>, "netdev@...r.kernel.org"
<netdev@...r.kernel.org>, "Hay, Joshua A" <joshua.a.hay@...el.com>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"bpf@...r.kernel.org" <bpf@...r.kernel.org>, "David S . Miller"
<davem@...emloft.net>
Subject: Re: [Intel-wired-lan] [PATCH iwl-next, v1 0/3] Add Default Rx Queue
Setting for igc driver
On 7/31/2024 7:43 AM, Jakub Kicinski wrote:
> On Wed, 31 Jul 2024 07:40:11 +0000 Song, Yoong Siang wrote:
>> Regarding your suggestion of implementing a "wildcard rule,"
>> are you suggesting the use of an ethtool command similar to the following?
>>
>> ethtool -U <iface name> flow-type ether action <default queue>
>>
>> I have a concern that users might be not aware that this nfc rule is having lowest priority,
>> as the default queue would only take effect when no other filtering rules match.
>
> I believe that ethtool n-tuple rules are expected to be executed in
> order. User can use the 'loc' argument to place the rule at the end
> of the table.
>
Yes. Some drivers lack support for ordered rules, but instead enforce
that no rule can be added if it would cause such a violation.
In this case, (I haven't dug into the actual patches or code), I suspect
the driver will need to validate the location values when adding rules
to ensure that all rules which don't use the default queue have higher
priority than the wild card rule. The request to add a filter should
reject the rule in the case where a default queue rule was added with a
higher priority location.
>> Do you see this as a potential issue? If not, I am willing to proceed with
>> exploring the ethtool options you've mentioned.
Powered by blists - more mailing lists