net/intel-ix-kmod: New port for 10g Intel Ethernet Base Driver

[skozlov]

Summary

New port for updated 10g Intel Ethernet Base Driver, based on
net/intel-ixl-kmod as a master port.

Depends on D17440, must be committed after it.

Test Plan

Compile and sanity tested using poudriere w/portlint:

intel-ix-kmod-3.3.6.log102 KBDownload

[sbruno]

Does this require a LICENSE or some sort?

[skozlov]

In D17443#372074, @sbruno wrote:

Does this require a LICENSE or some sort?

License is BSD2CLAUSE and is set in the master port

[koobs]

Looks good except for -Werror. This all matches the port WIP I have in my tree, except for the additions:

MANPAGENAME=        ixgbe
MANPAGELINKS=        ix if_ix if_ixgbe
​MAX_NETMAP_OSVERSION=        1199999  # Doesn't build w/NETMAP on 12

net/intel-ix-kmod/Makefile

10

The build uses -Werror which is great for upstream, but not good for distribution (the distributed source).

Please add CFLAGS+=-Wno-error

[erj]

net/intel-ix-kmod/pkg-descr

4 ↗

(On Diff #48813)

Shouldn't this be expanded to X520/X540/X550 series, too? Or just 500 series adapters, in general?

[skozlov]

In D17443#372344, @koobs wrote:

Looks good except for -Werror. This all matches the port WIP I have in my tree, except for the additions:

MANPAGENAME=        ixgbe
MANPAGELINKS=        ix if_ix if_ixgbe
​MAX_NETMAP_OSVERSION=        1199999  # Doesn't build w/NETMAP on 12

Will do. Do you have any specific warnings showing up for you?

[koobs]

In D17443#372892, @skozlov wrote:

In D17443#372344, @koobs wrote:

Looks good except for -Werror. This all matches the port WIP I have in my tree, except for the additions:

MANPAGENAME=        ixgbe
MANPAGELINKS=        ix if_ix if_ixgbe
​MAX_NETMAP_OSVERSION=        1199999  # Doesn't build w/NETMAP on 12

Will do. Do you have any specific warnings showing up for you?

None on CURRENT/amd64

[skozlov]

Add X5* controllers to supported list

[skozlov]

I'll add -Wno-error to the master port, but because of the amount of testing it will require - I'd like to make it a separate review.

[skozlov]

net/intel-{ix,em}-kmod: freeze PORTREVISION and PORTEPOCH

PORTREVISION wasn't specified in net/intel-{ix,em}-kmod, which are
slave ports of net/intel-ixl-kmod. Because of that when PORTREVISION
was bumped in intel-ixl-kmod - it automatically appeared in the slave
ports, which shouldn't happen.

To avoid this in the future - explicitly specify PORTREVISION and
PORTEPOCH in the slave ports.

[skozlov]

Original patch committed in r481900
The revision wasn't automatically closed because I wasn't able to use the arcanist properly :(