[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <87o6mgeyn7.fsf@redhat.com>
Date: Tue, 27 Jan 2026 00:34:52 +0100
From: Paolo Valerio <pvalerio@...hat.com>
To: Andrew Lunn <andrew@...n.ch>
Cc: netdev@...r.kernel.org, Nicolas Ferre <nicolas.ferre@...rochip.com>,
Claudiu Beznea <claudiu.beznea@...on.dev>, Andrew Lunn
<andrew+netdev@...n.ch>, "David S. Miller" <davem@...emloft.net>, Eric
Dumazet <edumazet@...gle.com>, Jakub Kicinski <kuba@...nel.org>, Paolo
Abeni <pabeni@...hat.com>, Lorenzo Bianconi <lorenzo@...nel.org>,
Théo Lebrun <theo.lebrun@...tlin.com>
Subject: Re: [PATCH net-next 3/8] cadence: macb: Add page pool support
handle multi-descriptor frame rx
On 26 Jan 2026 at 03:29:46 PM, Andrew Lunn <andrew@...n.ch> wrote:
>> > I was more interested in plain networking, not XDP. Does it perform
>> > better with page pool? You at least need to show it is not worse, you
>> > need to avoid performance regressions.
>> >
>>
>> I retested with iperf3. The target has a single rx queue with iperf3
>> running with no cpu affinity set.
>>
>> | | 64 | 128 |
>> | baseline | 273 | 545 |
>> | pp (page) | 273 | 544 |
>> | pp (2 frags) | 272 | 544 |
>
> So no real difference. That is unusual, it is typically faster, or if
> it is always doing line rate, it uses less CPU time. That might
> suggest the page pool integration is not optimal?
>
Thanks Andrew,
I can't tell, but the same numbers were a little surprising to me as well.
I'll spend some more time on this on my hw.
Powered by blists - more mailing lists