netdev - Re: [RFC v2 1/1] net: Add support to configure SR-IOV VF minimum and maximum Tx rate through ip tool.

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 for Android: free password hash cracker in your pocket

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

Date:	Fri, 28 Mar 2014 00:26:05 +0000
From:	Ben Hutchings <ben@...adent.org.uk>
To:	Michal Kubecek <mkubecek@...e.cz>
Cc:	Sucheta Chakraborty <sucheta.chakraborty@...gic.com>,
	netdev@...r.kernel.org, Dept-HSGLinuxNICDev@...gic.com,
	gregory.v.rose@...el.com, linux-net-drivers@...arflare.com,
	Ariel.Elior@...gic.com, amirv@...lanox.com
Subject: Re: [RFC v2 1/1] net: Add support to configure SR-IOV VF minimum
 and maximum Tx rate through ip tool.

On Wed, 2014-03-26 at 08:37 +0100, Michal Kubecek wrote:
> On Monday 24 of March 2014 18:55:00 Ben Hutchings wrote:
> > 
> > Also, this special-casing of -1 isn't documented anywhere.  Is it even
> > necessary?  If userland needs to set just one limit, it can read the
> > existing limits and set both.
> 
> Wouldn't this open a window for a race if one process wanted to change 
> one limit and another process wanted to change the other at the same 
> time? Such scenario doesn't sound very realistic but our customers 
> taught me that things I don't find very realistic tend to be used quite 
> frequently by them.

If there are two processes trying to change the minimum and maximum
bandwidth of a VF at the same time, something has gone wrong already.
This is a problem for userland to solve.

Ben.

> On the other hand, if changing only one limit is going to be common, it 
> might be more appropriate to add IFLA_VF_TX_MIN_RATE instead and always 
> pass the minimum and maximum rate separately (and pass only one if only 
> one is going to be changed).

-- 
Ben Hutchings
73.46% of all statistics are made up.

Download attachment "signature.asc" of type "application/pgp-signature" (812 bytes)