[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Yx7b5Jg051jFhLea@DEN-LT-70577>
Date: Mon, 12 Sep 2022 07:03:24 +0000
From: <Daniel.Machon@...rochip.com>
To: <vladimir.oltean@....com>
CC: <netdev@...r.kernel.org>, <Allan.Nielsen@...rochip.com>,
<UNGLinuxDriver@...rochip.com>, <maxime.chevallier@...tlin.com>,
<petrm@...dia.com>, <kuba@...nel.org>, <vinicius.gomes@...el.com>,
<thomas.petazzoni@...tlin.com>
Subject: Re: [RFC PATCH net-next 2/2] net: dcb: add new apptrust attribute
Den Fri, Sep 09, 2022 at 12:29:50PM +0000 skrev Vladimir Oltean:
> EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
>
> Hi Daniel,
>
> On Thu, Sep 08, 2022 at 02:04:42PM +0200, Daniel Machon wrote:
> > Add a new apptrust extension attribute to the 8021Qaz APP managed
> > object.
> >
> > The new attribute is meant to allow drivers, whose hw supports the
> > notion of trust, to be able to set whether a particular app selector is
> > to be trusted - and also the order of precedence of selectors.
> >
> > A new structure ieee_apptrust has been created, which contains an array
> > of selectors, where lower indexes has higher precedence.
> >
> > Signed-off-by: Daniel Machon <daniel.machon@...rochip.com>
> > ---
>
> Let's say I have a switch which only looks at VLAN PCP/DEI if the bridge
> vlan_filtering setting is enabled (otherwise, the switch is completely
> VLAN unaware, including for QoS purposes).
>
> Would it be ok to report through ieee_getapptrust() that the PCP
> selector is trusted when under a vlan_filtering bridge, not trusted when
> not under a vlan_filtering bridge, and deny changes to ieee_setapptrust()
> for the PCP selector? I see the return value is not cached anywhere
> within the kernel, just passed to the user.
There *might* be a distinction between enabled and trusted, disabled and not-trusted.
For instance, sparx5 switch has this distinction (at least for dscp) - but really that is
hw dependent. Therefore, in your particular case, with the vlan_filtering on/off,
yes that would be OK IMO. Any concerns?
This patch merely provides the means for drivers to implement a user-specified trust
order and report it back to the user, just like with many of the other dcb attributes
(maxrate, buffer etc.).
Powered by blists - more mailing lists