[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20181208170729.GB25144@lunn.ch>
Date: Sat, 8 Dec 2018 18:07:29 +0100
From: Andrew Lunn <andrew@...n.ch>
To: Ilias Apalodimas <ilias.apalodimas@...aro.org>
Cc: Eric Dumazet <eric.dumazet@...il.com>,
Florian Westphal <fw@...len.de>,
Jesper Dangaard Brouer <brouer@...hat.com>,
netdev@...r.kernel.org, "David S. Miller" <davem@...emloft.net>,
Toke Høiland-Jørgensen <toke@...e.dk>,
ard.biesheuvel@...aro.org, Jason Wang <jasowang@...hat.com>,
BjörnTöpel <bjorn.topel@...el.com>,
w@....eu, Saeed Mahameed <saeedm@...lanox.com>,
mykyta.iziumtsev@...il.com,
Daniel Borkmann <borkmann@...earbox.net>,
Alexei Starovoitov <alexei.starovoitov@...il.com>,
Tariq Toukan <tariqt@...lanox.com>
Subject: Re: [net-next, RFC, 4/8] net: core: add recycle capabilities on skbs
via page_pool API
> I got other concerns on the patchset though. Like how much memory is
> it 'ok' to keep mapped keeping in mind we are using the streaming
> DMA API. Are we going to affect anyone else negatively by doing so ?
For mvneta, you can expect the target to have between 512Mbyte to
3G. You can take a look at the DT files to get a better feel for this.
All of it should be low enough that it is DMA'able.
These SoCs are often used for WiFi access points, and NAS boxes. So i
guess the big memory usage on these boxes is the block cache for a NAS
device. So if you want to see the impact of the driver using more
memory, see if disk performance is affected.
Andrew
Powered by blists - more mailing lists