[dpdk-dev] [PATCH v2] nfp: report link speed using hardware info

Alejandro Lucero alejandro.lucero at netronome.com
Fri Nov 18 17:50:35 CET 2016


On Fri, Nov 18, 2016 at 4:29 PM, Thomas Monjalon <thomas.monjalon at 6wind.com>
wrote:

> 2016-11-18 16:06, Alejandro Lucero:
> > Previous reported speed was hardcoded.
> >
> > Signed-off-by: Alejandro Lucero <alejandro.lucero at netronome.com>
> > ---
> >  drivers/net/nfp/nfp_net.c      | 28 ++++++++++++++++++++++++++--
> >  drivers/net/nfp/nfp_net_ctrl.h | 13 +++++++++++++
> >  2 files changed, 39 insertions(+), 2 deletions(-)
>
> You should update the doc in the same patch:
>         doc/guides/nics/features/nfp.ini
> It will be the first feature as the file appears to be empty.
> So you will need another patch to fill other existing features.
>

Yes. I'm just working on updating that file properly.
May I delay this doc change for including it with that other one?
It will be a bit weird to just have one feature there.


>
> I have an unrelated question: why nfp is disabled in the default build?
>
>
Because NFP PMD can just work if Netronome BSP is installed in the system.
We do not support PF with the PMD, so it requires a Linux PF driver which
comes with the BSP.

The compilation has no dependencies, but we had our own UIO driver before
(now using igb_uio). So basically, we wanted the people aware of this
dependency and to specifically configure this option.

I know what you are surely going to ask about DPDK in Linux distributions,
and that this being a bad idea. The fact is, we have people using NFP PMD
as part of a product, so installing that product implies to (automatically)
install the BSP and a specific DPDK version with the NFP PMD enabled. But
yes, maybe we should modify this and to add some sort of BSP check inside
the PMD.

So, thanks for the heads up. I will think about this.


More information about the dev mailing list