net/intel-ixl-kmod Updated 40 gigabit FreeBSD Base Driver for Intel(R) Network Connections
ClosedPublic
Actions

Authored by skozlov on Mar 27 2016, 8:03 PM.

Details

Reviewers

jeffrey.e.pieper_intel.com
rpokala
gnn
sbruno

Group Reviewers

Intel Networking

Commits

rP413649: Add new port to provide updated ixl(4) drivers via ports to override

Summary

New port for updated version of Intel(R) ixl driver.

Driver is able to override the one in GENERIC without rebuilding the kernel, please see pkg-message for the details.
Port is made as master to allow easy deployment of other Intel network drivers in the future.
Putting my Intel hat on for this port.

Diff Detail

Repository

rP FreeBSD ports repository

Lint

Lint Not Applicable

Unit

Tests Not Applicable

Event Timeline

skozlov updated this revision to Diff 14655.Mar 27 2016, 8:03 PM

skozlov retitled this revision from to net/intel-ixl-kmod Updated 40 gigabit FreeBSD Base Driver for Intel(R) Network Connections.

skozlov updated this object.

skozlov edited the test plan for this revision. (Show Details)

skozlov added a reviewer: Intel Networking.

skozlov set the repository for this revision to rP FreeBSD ports repository.

Herald added a subscriber: mat. · View Herald TranscriptMar 27 2016, 8:03 PM

sbruno added a reviewer: gnn.Mar 27 2016, 8:50 PM

sbruno added a subscriber: adrian.

To my not-very-ports-savvy eyes, this looks right. The only thing which threw me was SITE_SUBDIR being defined after it was used to define MASTER_SITES; perhaps re-order them?

Why does it need to be called kmod? I'd just name it intel-ixl

In D5761#123295, @gnn wrote:

Why does it need to be called kmod? I'd just name it intel-ixl

This is a ports "thing".

I requested this change to fall inline with the other kernel modules in the ports system.

Carry on then!

This revision is now accepted and ready to land.Mar 29 2016, 4:04 PM

sbruno added inline comments.Mar 29 2016, 4:04 PM

net/intel-ixl-kmod/Makefile
26 ↗	(On Diff #14655)	Change this variable to "MASTER_SITE_SUBDIR" Move it to line 9, under MASTER_SITES for clarity.

What does this warning in the log mean?

WARN: Consider to set DEVELOPER=yes in /etc/make.conf

In D5761#123330, @erj wrote:
What does this warning in the log mean?
WARN: Consider to set DEVELOPER=yes in /etc/make.conf

That is specifically for testing. If you set DEVELOPER=yes in make.conf then the ports MK system will check things more aggressively for conformance to ports "style".

Changed and moved SITE_SUBDIR -> MASTER_SITE_SUBDIR

'portlint -AC' is still happy:

WARN: Makefile: unless this is a master port, PORTNAME has to be set by "=", not by "?=".
WARN: Makefile: unless this is a master port, PORTVERSION has to be set by "=", not by "?=".
WARN: Makefile: unless this is a master port, COMMENT has to be set by "=", not by "?=".
0 fatal errors and 3 warnings found.

This revision now requires review to proceed.Mar 29 2016, 9:39 PM

skozlov marked an inline comment as done.Mar 29 2016, 9:40 PM

In D5761#123383, @sbruno wrote:
In D5761#123330, @erj wrote:
What does this warning in the log mean?
WARN: Consider to set DEVELOPER=yes in /etc/make.conf
That is specifically for testing. If you set DEVELOPER=yes in make.conf then the ports MK system will check things more aggressively for conformance to ports "style".

This warning is inevitable when using poudriere - I have 'DEVELOPER=yes' set in my make.conf. You can see that there's no such warning in my prev comment.

In D5761#123418, @kozlov.sergey.404_gmail.com wrote:

Changed and moved SITE_SUBDIR -> MASTER_SITE_SUBDIR

'portlint -AC' is still happy:

WARN: Makefile: unless this is a master port, PORTNAME has to be set by "=", not by "?=".
WARN: Makefile: unless this is a master port, PORTVERSION has to be set by "=", not by "?=".
WARN: Makefile: unless this is a master port, COMMENT has to be set by "=", not by "?=".
0 fatal errors and 3 warnings found.

Those seem reasonable to change to "=". This is not a slave port.

In D5761#123510, @sbruno wrote:

Those seem reasonable to change to "=". This is not a slave port.

It will be a master port, there are plans to make ports for other drivers. I didn't want to touch this port the time e.g. intel-ix-kmod will be on the review to narrow the scope to one port only.

skozlov mentioned this in D5828: sysutils/intel-nvmupdate: NVM Update Utility for Intel(R) Ethernet Adapters.Apr 4 2016, 4:14 PM

jeffrey.e.pieper_intel.com accepted this revision.Apr 6 2016, 4:34 PM

jeffrey.e.pieper_intel.com added a reviewer: jeffrey.e.pieper_intel.com.

This revision is now accepted and ready to land.Apr 6 2016, 4:34 PM

To my not-very-ports-savvy eyes, this looks fine now.

Need a "LGTM" from portmgr. Then I'll commit it.

Added previously mistakenly excluded files/pkg-message.in

Portlint is still happy:
WARN: Makefile: unless this is a master port, PORTNAME has to be set by "=", not by "?=".
WARN: Makefile: unless this is a master port, PORTVERSION has to be set by "=", not by "?=".
WARN: Makefile: unless this is a master port, COMMENT has to be set by "=", not by "?=".
0 fatal errors and 3 warnings found.

This revision now requires review to proceed.Apr 15 2016, 10:38 AM

bapt removed a reviewer: portmgr.Apr 19 2016, 7:09 AM

Closed by commit rP413649: Add new port to provide updated ixl(4) drivers via ports to override (authored by sbruno). · Explain WhyApr 19 2016, 3:11 PM

This revision was automatically updated to reflect the committed changes.