[developer] Review: 5250 update oce from source provided by Emulex

[Hans Rosenfeld via illumos-developer]

Hi,

a while back we received an updated oce driver from Emulex to support
their newer 10G NICs. This is apparently the same driver that they ship
for Solaris.

Dan McDonald made a few adjustments to make it compile against our
subset of GLDv3 when he was still at Nexenta. I made a few more changes
to make it actually work.

This is a rather big change in a rather big driver that we don't have
hardware documentation for. I don't expect to get a real thorough review
for this, but it would be nice if we at least made sure theres no
obviously questionable code in there. Also, I'd like someone who really
knows GLDv3 to look at the changes I made in oce_ring.c around line 135.

Here's the webrev of Dan's and my changes on top of the Emulex source
drop:
http://ma.nexenta.com/~woodstock/illumos-oce-1/

And the webrev for all changes from Emulex and us put together:
http://ma.nexenta.com/~woodstock/illumos-oce/


Hans
--
%SYSTEM-F-ANARCHISM, The operating system has been overthrown


-------------------------------------------
illumos-developer
Archives: https://www.listbox.com/member/archive/182179/=now
RSS Feed: https://www.listbox.com/member/archive/rss/182179/21175072-86d49504
Modify Your Subscription: https://www.listbox.com/member/?member_id=21175072&id_secret=21175072-abdf7b7e
Powered by Listbox: http://www.listbox.com

[Dan McDonald via illumos-developer]

Post by Hans Rosenfeld via illumos-developer
Here's the webrev of Dan's and my changes on top of the Emulex source
http://ma.nexenta.com/~woodstock/illumos-oce-1/

You've a comment on lines 136-140 of oce_ring.c. I'd like to make sure Sebastien or Robert looks at that. They know Nemo/MAC better than I, and might be able to give you an answer and drive away the ugly XXX. :)

Dan



-------------------------------------------
illumos-developer
Archives: https://www.listbox.com/member/archive/182179/=now
RSS Feed: https://www.listbox.com/member/archive/rss/182179/21175072-86d49504
Modify Your Subscription: https://www.listbox.com/member/?member_id=21175072&id_secret=21175072-abdf7b7e
Powered by Listbox: http://www.listbox.com

[Garrett D'Amore via illumos-developer]

I have this sneaking suspicion that multiple transmit groups expressed by hardware were not implemented.

Anyway, my read of the code is that if mr_gnum is zero, the system will dynamically create one group per actual ring. This is probably actually what you want anyway.

I’d probably set mr_gnum = 0 explicitly, with a comment that says /* Let system dynamically create TX groups. */

I have not read any of the rest of the driver code, ENOTIME.

- Garrett

Post by Dan McDonald via illumos-developer

Post by Hans Rosenfeld via illumos-developer
Here's the webrev of Dan's and my changes on top of the Emulex source
http://ma.nexenta.com/~woodstock/illumos-oce-1/

You've a comment on lines 136-140 of oce_ring.c. I'd like to make sure Sebastien or Robert looks at that. They know Nemo/MAC better than I, and might be able to give you an answer and drive away the ugly XXX. :)
Dan
-------------------------------------------
illumos-developer
Archives: https://www.listbox.com/member/archive/182179/=now
RSS Feed: https://www.listbox.com/member/archive/rss/182179/21239177-3604570e
Modify Your Subscription: https://www.listbox.com/member/?&
Powered by Listbox: http://www.listbox.com

-------------------------------------------
illumos-developer
Archives: https://www.listbox.com/member/archive/182179/=now
RSS Feed: https://www.listbox.com/member/archive/rss/182179/21175072-86d49504
Modify Your Subscription: https://www.listbox.com/member/?member_id=21175072&id_secret=21175072-abdf7b7e
Powered by Listbox: http://www.listbox.com