[PATCH v9 07/12] net/nfp: add flower ctrl VNIC related logics

[Chaoyong He]

> 21/09/2022 04:02, Chaoyong He:
> > > On 9/15/2022 11:44 AM, Chaoyong He wrote:
> > > > Adds the setup/start logic for the ctrl vNIC. This vNIC is used by
> > > > the PMD and flower firmware application as a communication channel
> > > between
> > > > driver and firmware. In the case of OVS it is also used to
> > > > communicate flow statistics from hardware to the driver.
> > > >
> > > > A rte_eth device is not exposed to DPDK for this vNIC as it is
> > > > strictly used internally by flower logic.
> > > >
> > >
> > > Hi Chaoyong,
> > >
> > > Similar comment with previous versions, interface is created using
> > > regular 'rte_eth_dev_allocate()' API, I think interface will be
> > > visible to application, I can't understand the need of creating an interface
> for control.
> 
> You didn't reply to this.
> Why the control port should be exposed to the application?
> We recommend not using ethdev for this.
> 

Actually, in the v1--v5 of this patch series, we did create a control port which is not
in the rte_eth_devices[] array, so it won't exposed to the application.

> 
> > > What is the communication method between driver and FW?
> > > Since one of the following patches (09/12) introduces Rx/Tx for ctrl
> > > interface, is device interface is control packets (similar to network data
> packets)?
> > >
> >
> > Basically, the 'control message' is exist in the form of normal data packets.
> >
> > When we use the flower firmware application, there exist two types of
> > packets for now, and they are identified only from the prepend meta-data.
> >
> > Bit    3 3 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0
> > -----\ 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0
> > Word  +---------------+---------------+---------------+---------------+
> >    0  |    type       |     type      |     type      |     type      |
> >
> > +---------------+---------------+---------------+---------------+
> > The 'control message' packets are processed by the ctrl vNIC.
> > The 'normal' packets are processed by the pf vNIC.
> >
> > The communication method between driver and firmware is decided by
> the
> > designment of hardware and firmware.
> >
> > The kernel driver also has the same ctrl vNIC and pf vNIC ethdev and the
> usage is same.
> >
> > > > Because of the add of ctrl vNIC, a new PCItoCPPBar is needed.
> > > > Modify the related logics.
> > > >
> > > > Signed-off-by: Chaoyong He <chaoyong.he at corigine.com>
> > > > Reviewed-by: Niklas Söderlund <niklas.soderlund at corigine.com>
> 
>