netdev - Re: [net-next 03/14] net/mlx5e: Expose rx_oversize_pkts

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 19 Jul 2022 14:13:39 +0300
From:   Gal Pressman <gal@...dia.com>
To:     Jakub Kicinski <kuba@...nel.org>, Saeed Mahameed <saeed@...nel.org>
Cc:     "David S. Miller" <davem@...emloft.net>,
        Paolo Abeni <pabeni@...hat.com>,
        Eric Dumazet <edumazet@...gle.com>,
        Saeed Mahameed <saeedm@...dia.com>, netdev@...r.kernel.org,
        Tariq Toukan <tariqt@...dia.com>
Subject: Re: [net-next 03/14] net/mlx5e: Expose rx_oversize_pkts_buffer
 counter

On 19/07/2022 06:25, Jakub Kicinski wrote:
> On Sun, 17 Jul 2022 14:33:41 -0700 Saeed Mahameed wrote:
>> From: Gal Pressman <gal@...dia.com>
>>
>> Add the rx_oversize_pkts_buffer counter to ethtool statistics.
>> This counter exposes the number of dropped received packets due to
>> length which arrived to RQ and exceed software buffer size allocated by
>> the device for incoming traffic. It might imply that the device MTU is
>> larger than the software buffers size.
> Is it counted towards any of the existing stats as well? It needs 
> to end up in struct rtnl_link_stats64::rx_length_errors somehow.

Probably makes sense to count it in rx_over_errors:
 *   The recommended interpretation for high speed interfaces is -
 *   number of packets dropped because they did not fit into buffers
 *   provided by the host, e.g. packets larger than MTU or next buffer
 *   in the ring was not available for a scatter transfer.

It doesn't fit the rx_length_errors (802.3) as these packets are not
dropped on the MAC.
Will change.

> On ethtool side - are you not counting this towards FrameTooLongErrors
> because it's not dropped in the MAC? Can we count it as RMON's
> oversize_pkts?

 etherStatsOversizePkts OBJECT-TYPE
     SYNTAX     Counter32
     UNITS      "Packets"
     MAX-ACCESS read-only
     STATUS     current
     DESCRIPTION
         "The total number of packets received that were
         longer than 1518 octets (excluding framing bits,
         but including FCS octets) and were otherwise
         well formed."
     ::= { etherStatsEntry 10 }

This counter isn't necessarily tied to 1518 bytes.

Thanks for the review, Jakub.