> -----Original Message-----
> From: Jakub Kicinski <jakub.kicin...@netronome.com>
> Sent: Saturday, July 6, 2019 12:47 AM
> To: Parav Pandit <pa...@mellanox.com>
> Cc: netdev@vger.kernel.org; Jiri Pirko <j...@mellanox.com>; Saeed Mahameed
> <sae...@mellanox.com>
> Subject: Re: [PATCH net-next v2 1/3] devlink: Introduce PCI PF port flavour
> and
> port attribute
>
> On Fri, 5 Jul 2019 02:37:09 -0500, Parav Pandit wrote:
> > @@ -38,14 +38,24 @@ struct devlink {
> > char priv[0] __aligned(NETDEV_ALIGN); };
> >
> > +struct devlink_port_pci_pf_attrs {
> > + u16 pf; /* Associated PCI PF for this port. */
> > +};
> > +
> > struct devlink_port_attrs {
> > u8 set:1,
> > split:1,
> > switch_port:1;
> > enum devlink_port_flavour flavour;
> > - u32 port_number; /* same value as "split group" */
> > + u32 port_number; /* same value as "split group".
> > + * Valid only when a port is physical and visible
> > + * to the user for a given port flavour.
> > + */
>
> port_number can be in the per-flavour union below.
>
Ack.
> > u32 split_subport_number;
>
> As can split_subport_number.
>
Ack.
> > struct netdev_phys_item_id switch_id;
> > + union {
> > + struct devlink_port_pci_pf_attrs pci_pf;
> > + };
> > };
> >
> > struct devlink_port {
>
> > @@ -515,8 +523,14 @@ static int devlink_nl_port_attrs_put(struct sk_buff
> *msg,
> > return 0;
> > if (nla_put_u16(msg, DEVLINK_ATTR_PORT_FLAVOUR, attrs->flavour))
> > return -EMSGSIZE;
> > - if (nla_put_u32(msg, DEVLINK_ATTR_PORT_NUMBER, attrs-
> >port_number))
> > + if (is_devlink_phy_port_num_supported(devlink_port) &&
> > + nla_put_u32(msg, DEVLINK_ATTR_PORT_NUMBER, attrs-
> >port_number))
> > return -EMSGSIZE;
> > + if (devlink_port->attrs.flavour == DEVLINK_PORT_FLAVOUR_PCI_PF) {
> > + if (nla_put_u16(msg, DEVLINK_ATTR_PORT_PCI_PF_NUMBER,
> > + attrs->pci_pf.pf))
> > + return -EMSGSIZE;
> > + }
> > if (!attrs->split)
> > return 0;
> > if (nla_put_u32(msg, DEVLINK_ATTR_PORT_SPLIT_GROUP,
> > attrs->port_number))
>
> Split attributes as well, please:
>
Ack.
> On Tue, 2 Jul 2019 16:42:52 -0700, Jakub Kicinski wrote:
> > port_number, and split attributes should not be exposed for PCI ports.
>
> We have no clear semantics for those, yet, and the phys_port_name
> implementation in this patch doesn't handle split PCI, so let's leave them out
> for now.
Ok. Sending v3.
Thanks for the review.
