[PATCH 1/6] examples/l3fwd: fix lcore ID restriction

[Morten Brørup]

> From: David Marchand [mailto:david.marchand at redhat.com]
> Sent: Thursday, 7 March 2024 10.22
restriction
> 
> On Thu, Mar 7, 2024 at 10:16 AM Morten Brørup <mb at smartsharesystems.com>
> wrote:
> >
> > > From: David Marchand [mailto:david.marchand at redhat.com]
> > > Sent: Thursday, 7 March 2024 09.34
> > >
> > > On Mon, Dec 18, 2023 at 8:49 AM Sivaprasad Tummala
> > > <sivaprasad.tummala at amd.com> wrote:
> > > >
> > > > Currently the config option allows lcore IDs up to 255,
> > > > irrespective of RTE_MAX_LCORES and needs to be fixed.
> > >
> > > "needs to be fixed" ?
> > > I disagree on the principle.
> > > The examples were written with limitations, this is not a bug.
> >
> > Unfortunately, l3fwd is not only an example; it is also used for
> benchmarking. It really belongs in some other directory.
> >
> > With that in mind, I would consider it a bug that the benchmarking
> application cannot handle the amount of cores available in modern CPUs.
> 
> This is not a bug.

DPDK 23.11 LTS supposedly supports Zen 4, which has 512 cores [1].
If l3fwd does not support it, it is a bug in l3fwd.

[1]: https://elixir.bootlin.com/dpdk/v23.11/source/config/x86/meson.build#L88

> 
> And with careful configuration (using --lcores), you can already start
> l3fwd with 254 datapath threads, right?

Correct, but you cannot start l3fwd with all 512 cores.

There should be some test scenario to set up l3fwd to use all 512 cores.

If there are practical limitations that effectively prevents l3fwd from using all 512 cores, e.g. if no NIC offers more than 256 queues, then I'll let go and not consider it a bug.

> 
> 
> --
> David Marchand