[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <7a2d23b2-5a7d-d68e-1ae4-13f114c5a380@redhat.com>
Date: Wed, 23 Feb 2022 19:10:15 +0100
From: Jesper Dangaard Brouer <jbrouer@...hat.com>
To: Joe Damato <jdamato@...tly.com>, Jakub Kicinski <kuba@...nel.org>
Cc: brouer@...hat.com, Jesper Dangaard Brouer <jbrouer@...hat.com>,
netdev@...r.kernel.org, ilias.apalodimas@...aro.org,
davem@...emloft.net, hawk@...nel.org, saeed@...nel.org,
ttoukan.linux@...il.com
Subject: Re: [net-next v6 1/2] page_pool: Add page_pool stats
On 23/02/2022 18.45, Joe Damato wrote:
> On Wed, Feb 23, 2022 at 9:40 AM Jakub Kicinski <kuba@...nel.org> wrote:
>>
>> On Wed, 23 Feb 2022 09:05:06 -0800 Joe Damato wrote:
>>> Are the cache-line placement and per-cpu designations the only
>>> remaining issues with this change?
>>
>> page_pool_get_stats() has no callers as is, I'm not sure how we can
>> merge it in current form.
>>
>> Maybe I'm missing bigger picture or some former discussion.
>
> I wrote the mlx5 code to call this function and export the data via
> ethtool. I had assumed the mlx5 changes should be in a separate
> patchset. I can include that code as part of this change, if needed.
I agree with Jakub we need to see how this is used by drivers.
--Jesper
Powered by blists - more mailing lists