Intel 10G SRIOV support

[jfv]

Summary

This revision adds support for PF (host-side) SRIOV support to the ixgbe driver. Thanks belongs to Ryan
Stone of Sandvine who did all the early implementation, and to Marc Snyder of Dell who did some further
polish and debugging.

Test Plan

Due to internal schedules Intel has not yet been able to do more than touch test this feature,
so please understand that this is a Technology Preview, not fully tested functionaliy. We will
be continuing to test and enhance this over this year, and any early adapters are welcome to
submit changes and fixes to the code.

[jfv]

Remove x550 device that is not yet available. Also remove unneeded shared code prototype,
and some added whitespace.

[rstone]

head/sys/dev/ixgbe/if_ix.c

4990

Hm. There's a good chance I am ultimately the person who commented this line out, but I have no idea why. :(

5337

this can go. :)

5337

that was supposed to be on the CTR3 call about

5467

bad indentation here?

[gnn]

I believe there is definitely a bug in the !! case (my first comment).

Also, can you split this commit into two, one for the IOV add and another for the non IOV changes?

head/sys/dev/ixgbe/if_ix.c

3760

Don't you have too many ! there?

head/sys/dev/ixgbe/ix_txrx.c

624–625

Is there a carriage return issue here or is that because of Phabricator?

head/sys/dev/ixgbe/ixgbe.h

450

Remove extra cr/lf

[erj]

head/sys/dev/ixgbe/if_ix.c

3760

It's supposed to be like that because this value gets exposed to the user in the eee.enable sysctl; we want to init the value here to either 1 or 0 instead a function pointer or 0.

[jfv]

head/sys/dev/ixgbe/if_ix.c

3760

Eric did this, its as intended.

4990

Well, are you saying it should be uncommented, or removed?

5337

What do you mean can go exactly?

head/sys/dev/ixgbe/ix_txrx.c

624–625

It was a long line, i divided it to keep it under 80, there's no extra line inbetween
if that's what you mean?

[jfv]

Correct a couple minor points in the last revision.

[jfv]

Integrated in r283883