2023-06-12 10:17:10

by Vladimir Oltean

[permalink] [raw]
Subject: Re: [PATCH net-next 2/2] net: dsa: mv88e6xxx: implement egress tbf qdisc for 6393x family

Hi Sunil,

On Mon, Jun 12, 2023 at 12:04:56PM +0530, Sunil Kovvuri wrote:
> For setting up simple per-port ratelimit, instead of TBF isn't "egress
> matchall" suitable here ?

"matchall" is a filter. What would be the associated action for a
port-level shaper?


2023-06-12 18:27:50

by Sunil Kovvuri

[permalink] [raw]
Subject: Re: [PATCH net-next 2/2] net: dsa: mv88e6xxx: implement egress tbf qdisc for 6393x family

On Mon, Jun 12, 2023 at 3:13 PM Vladimir Oltean <[email protected]> wrote:
>
> Hi Sunil,
>
> On Mon, Jun 12, 2023 at 12:04:56PM +0530, Sunil Kovvuri wrote:
> > For setting up simple per-port ratelimit, instead of TBF isn't "egress
> > matchall" suitable here ?
>
> "matchall" is a filter. What would be the associated action for a
> port-level shaper?

As Alexis mentioned I was referring to "matchall + policer".

Thanks,
Sunil.

2023-06-12 19:06:01

by Vladimir Oltean

[permalink] [raw]
Subject: Re: [PATCH net-next 2/2] net: dsa: mv88e6xxx: implement egress tbf qdisc for 6393x family

On Mon, Jun 12, 2023 at 11:53:06PM +0530, Sunil Kovvuri wrote:
> On Mon, Jun 12, 2023 at 3:13 PM Vladimir Oltean <[email protected]> wrote:
> >
> > Hi Sunil,
> >
> > On Mon, Jun 12, 2023 at 12:04:56PM +0530, Sunil Kovvuri wrote:
> > > For setting up simple per-port ratelimit, instead of TBF isn't "egress
> > > matchall" suitable here ?
> >
> > "matchall" is a filter. What would be the associated action for a
> > port-level shaper?
>
> As Alexis mentioned I was referring to "matchall + policer".

The idea would be to pick a software representation which matches the
hardware behavior. A policer drops excess packets, a shaper queues them.
This hardware supports some sort of egress rate shaping.