[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <f0c50993-f8aa-528e-1128-218cc0f01d57@gmail.com>
Date: Thu, 5 Sep 2024 11:50:37 +0100
From: Edward Cree <ecree.xilinx@...il.com>
To: Jakub Kicinski <kuba@...nel.org>, edward.cree@....com
Cc: linux-net-drivers@....com, davem@...emloft.net, edumazet@...gle.com,
pabeni@...hat.com, netdev@...r.kernel.org
Subject: Re: [PATCH net-next 3/6] sfc: add n_rx_overlength to ethtool stats
On 29/08/2024 01:47, Jakub Kicinski wrote:
> On Wed, 28 Aug 2024 14:45:12 +0100 edward.cree@....com wrote:
>> This counter is the main difference between the old and new locations
>> of the rx_packets increment (the other is scatter errors which
>> produce a WARN_ON). It previously was not reported anywhere; add it
>> to ethtool -S output to ensure users still have this information.
>
> What is it tho? Not IEEE 802.3 30.3.1.1.25 aFrameTooLongErrors ?
No, it doesn't appear to be.
If I'm understanding the code correctly, it counts "RX packets which
SG placed in a single RX buffer but whose length (from the RX event)
is too big to fit in that RX buffer". Which doesn't sound like a
thing that should ever happen (and when it does we netif_err() under
ratelimit, see efx_rx_packet__check_len()).
I'll put this into the commit message.
On 28/08/2024 23:22, Jacob Keller wrote:
> The description makes sense in context with the whole series but doesn't
> quite work for me if I think about viewing it without context. Perhaps a
> little more clarification about the rx_packets behavioral change?
Sure, will do.
Powered by blists - more mailing lists