netdev - Re: [RFC PATCH net-next 0/2] Configuring PFC stall prevention via ethtool

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]

Message-ID: <20171116010722.GD2130@lunn.ch>
Date:   Thu, 16 Nov 2017 02:07:22 +0100
From:   Andrew Lunn <andrew@...n.ch>
To:     Eran Ben Elisha <eranbe@...lanox.com>
Cc:     netdev@...r.kernel.org, "David S. Miller" <davem@...emloft.net>,
        "John W. Linville" <linville@...driver.com>,
        Saeed Mahameed <saeedm@...lanox.com>,
        Gal Pressman <galp@...lanox.com>,
        Ariel Almog <ariela@...lanox.com>,
        Inbar Karmy <inbark@...lanox.com>
Subject: Re: [RFC PATCH net-next 0/2] Configuring PFC stall prevention via
 ethtool

On Wed, Nov 15, 2017 at 09:00:09PM +0200, Eran Ben Elisha wrote:
> From: Inbar Karmy <inbark@...lanox.com>
> 
> This RFC adds support for configuring PFC stall prevention through ethtool.
> 
> In the event where the device unexpectedly becomes unresponsive for a long
> period of time, flow control mechanism may propagate pause frames which will
> cause congestion spreading to the entire network.
> 
> To prevent this scenario, the device may implement a protection mechanism for
> monitoring and resolving such state.  The following patches allow the user to
> control the stall prevention functionality.
> 
> PFC stall prevention configuration is done via ethtool -a (pause).
> Two modes are introduced:
> Default - current behavior per driver.
> Auto - protection mechanism controlled automatically by the driver.

Why Auto?

Down in the driver you seem to translate this to a time. And it looks
like your hardware is flexible on that time, it can probably do at
least 8s to 100ms.

Why not specify a time?

What do other vendors support? Time? Number of pause frames sent?

     Andrew