I'll try to get some work done here once 13 is out. This will be a new feature :-)

Looks good.

Looks good. Don't forget to MFC.

I might have some time later this month to help prepare a patch the way I think is more correct to implement this.

Looks good.

Like said:
Channel prepare must use the configure message callback API, else this control request might end up out of order!

wulf@ Can you please have a look?

Drew, can you look at this patch.

Only if RSS is enabled it binds interrupts. Let's add Drew for comments.

Both locks are typically locked from the same CPU. Can you show some benefits of this, or is it just theoretical?

Added examples of new API usage - requested by kib@ .

Probably should do the same for aarch64 .

@kib: Let me start from scratch:

man blockcount
No manual entry for blockcount

Address comments from @kp

Looks good. Don't forget to MFC to 11,12,13.

Looks good.

Targetting a review for Monday.

I will test and review this patch on Monday.

I'll have a look at this tomorrow.

@rrs: I'm fine with your proposal to commit everything except the mlx5en(4) parts, which we would like split out in an own commit.

Test OK!

Let me test your latest patch tomorrow. Thank you!

Can if_next_snd_tag() be committed first?

The following additional diff make things work. Can you integrate and test?

It is still not working. Trying to investigate why.

It looks like the Mellanox bits can be committed separately and before the other patches.

Still doesn't work like it should:

No more comments from me.

Can you please test the patch? I get bogus results:

Run ifconfig ue0 in a loop on a USB interface which has got a mii interface attached to it.

Just one minor comment and I will accept.