netdev - Re: [RFC PATCH net-next 1/2] net: dcb: add new pcp selector to app object

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: <YyAmZaQegw2zMNLb@DEN-LT-70577>
Date:   Tue, 13 Sep 2022 06:33:30 +0000
From:   <Daniel.Machon@...rochip.com>
To:     <petrm@...dia.com>
CC:     <netdev@...r.kernel.org>, <Allan.Nielsen@...rochip.com>,
        <UNGLinuxDriver@...rochip.com>, <maxime.chevallier@...tlin.com>,
        <vladimir.oltean@....com>, <kuba@...nel.org>,
        <vinicius.gomes@...el.com>, <thomas.petazzoni@...tlin.com>
Subject: Re: [RFC PATCH net-next 1/2] net: dcb: add new pcp selector to app
 object

> > The purpose of adding the PCP selector, is to be able to offload
> > PCP-based queue classification to the 8021Q Priority Code Point table,
> > see 6.9.3 of IEEE Std 802.1Q-2018.
> >
> > PCP and DEI is encoded in the protocol field as 8*dei+pcp, so that a
> > mapping of PCP 2 and DEI 1 to priority 3 is encoded as {255, 10, 3}.
> >
> > While PCP is not a standard 8021Qaz selector, it seems very convenient
> > to add it to the APP object, as this is where similar priority mapping
> > is handled, and it perfectly fits the {selector, protocol, priority}
> > triplet.
> >
> > Signed-off-by: Daniel Machon <daniel.machon@...rochip.com>
> > ---
> >  include/uapi/linux/dcbnl.h | 1 +
> >  1 file changed, 1 insertion(+)
> >
> > diff --git a/include/uapi/linux/dcbnl.h b/include/uapi/linux/dcbnl.h
> > index a791a94013a6..8eab16e5bc13 100644
> > --- a/include/uapi/linux/dcbnl.h
> > +++ b/include/uapi/linux/dcbnl.h
> > @@ -217,6 +217,7 @@ struct cee_pfc {
> >  #define IEEE_8021QAZ_APP_SEL_DGRAM   3
> >  #define IEEE_8021QAZ_APP_SEL_ANY     4
> >  #define IEEE_8021QAZ_APP_SEL_DSCP       5
> > +#define IEEE_8021QAZ_APP_SEL_PCP     255
> >
> >  /* This structure contains the IEEE 802.1Qaz APP managed object. This
> >   * object is also used for the CEE std as well.
> 
> I'm thinking how to further isolate this from the IEEE standard values.
> I think it would be better to pass the non-standard APP contributions in
> a different attribute. IIUIC, this is how the APP table is passed:
> 
> DCB_ATTR_IEEE_APP_TABLE {
>     DCB_ATTR_IEEE_APP {
>         struct dcb_app { ... };
>     }
>     DCB_ATTR_IEEE_APP {
>         struct dcb_app { ... };
>     }
> }
> 
> Well, instead, the non-standard stuff could be passed in a different
> attribute:
> 
> DCB_ATTR_IEEE_APP_TABLE {
>     DCB_ATTR_IEEE_APP {
>         struct dcb_app { ... }; // standard contribution to APP table
>     }
>     DCB_ATTR_DCB_APP {
>         struct dcb_app { ... }; // non-standard contribution
>     }
> }
> 
> The new selector could still stay as 255. This will allow us to keep the
> internal bookkeeping simple for the likely case that 255 never becomes a
> valid IEEE selector. But if it ever does, the UAPI can stay the same,
> just the internals will need to be updated.

I get your sentiment, but it seems a little far-fetched to me. The 
trade-off will be extra code, in trade for something that IMO very likely 
will not happen. Like you said earlier - how many selectors could one
possibly prioritize on?