Page MenuHomeFreeBSD

net/aquantia-atlantic-kmod: Fix CURRENT build post r353868 (iflib changes)
Needs RevisionPublic

Authored by nyan_myuji.xyz on Wed, Jan 13, 2:49 PM.

Details

Reviewers
lwhsu
philip
koobs
Summary
net/aquantia-atlantic-kmod: Fix CURRENT build post r353868 (iflib changes)

This patch enable the port to build working if_atlantic.ko
on 13-CURRENT, broken since r353868, where `if_multi_apply`
and `if_multiaddr_count` were removed.

PR: 252642
Differential_Revision: D28135
MFH: 2021Q1 (blanket: build fix)
Test Plan
  • portlint: ???
  • testport: ??? (poudriere; at least current and 12amd64)
  • hardware: AQC107 (device id 0x07b1) tested

Diff Detail

Repository
rP FreeBSD ports repository
Lint
No Linters Available
Unit
No Unit Test Coverage
Build Status
Buildable 36173
Build 33062: arc lint + arc unit

Event Timeline

lwhsu requested changes to this revision.Wed, Jan 13, 3:44 PM
lwhsu added a subscriber: lwhsu.
lwhsu added inline comments.
net/aquantia-atlantic-kmod/Makefile
4

I believe PORTREVISION is enough here.

This revision now requires changes to proceed.Wed, Jan 13, 3:44 PM
net/aquantia-atlantic-kmod/Makefile
4

The reason why PORTEPOCH is used here is to keep the versioning in-sync with the upstream. or should we change PORTVERSION to something like 0.0.5p1 ?

net/aquantia-atlantic-kmod/Makefile
4

PORTREVISION is exactly what to use to add local patch and keep PORTVSION (or DISTVERSION) sync with upstream. PORTEPOCH should be used carefully because it will stay forever once comes in.

net/aquantia-atlantic-kmod/Makefile
4

my bad, I thought you mean PORTVERSION lmao

Use PORTREVISION instead of PORTEPOCH

This revision is now accepted and ready to land.Wed, Jan 13, 5:37 PM
philip requested changes to this revision.Thu, Jan 14, 12:50 AM

If the iflib changes have been tested and are known to work, they're probably okay.

Is the extra patch to aq_hw.c intentional? Using random() to generate a MAC address isn't great. Not seeding it makes it even worse.

This revision now requires changes to proceed.Thu, Jan 14, 12:50 AM
koobs retitled this revision from Fix net/aquantia-atlantic-kmod post r353868 iflib changes to net/aquantia-atlantic-kmod: Fix CURRENT build post r353868 (iflib changes).Thu, Jan 14, 2:11 AM
koobs edited the summary of this revision. (Show Details)
koobs edited the test plan for this revision. (Show Details)
koobs added a reviewer: koobs.

This needs:

  • Confirmation of QA: portlint/poudriere on at least CURRENT and 12amd64 (fill in TEST PLAN section)
  • PR created upstream
  • Comment added to patch with upstream issue/PR reference and URL

Thank you for the report and patch

koobs added 1 blocking reviewer(s): koobs.
koobs edited the summary of this revision. (Show Details)
net/aquantia-atlantic-kmod/Makefile
5

Is this actually required since the build fails to produce a package?

If the iflib changes have been tested and are known to work, they're probably okay.

Is the extra patch to aq_hw.c intentional? Using random() to generate a MAC address isn't great. Not seeding it makes it even worse.

Yes it is. This is because random has been deprecated and srandom has been removed at https://reviews.freebsd.org/rS356097

net/aquantia-atlantic-kmod/Makefile
5

I think I don't exactly know what you mean..
On my side poudriere testport passed and is able to produce package for both current and 12.2-release...
Portlint also passed.