svn commit: r353419 - head/sys/net

[Oleksandr Tymoshenko]

Gleb Smirnoff (glebius at freebsd.org) wrote:
> On Fri, Aug 28, 2020 at 02:31:30PM -0700, Oleksandr Tymoshenko wrote:
> O> Gleb Smirnoff (glebius at FreeBSD.org) wrote:
> O> > Author: glebius
> O> > Date: Thu Oct 10 23:42:55 2019
> O> > New Revision: 353419
> O> > URL: https://svnweb.freebsd.org/changeset/base/353419
> O> > 
> O> > Log:
> O> >   Provide new KPI for network drivers to access lists of interface
> O> >   addresses.  The KPI doesn't reveal neither how addresses are stored,
> O> >   how the access to them is synchronized, neither reveal struct ifaddr
> O> >   and struct ifmaddr.
> O> >   
> O> >   Reviewed by:	gallatin, erj, hselasky, philip, stevek
> O> >   Differential Revision:	https://reviews.freebsd.org/D21943
> O> 
> O> Hi Gleb,
> O> 
> O> Are there any plans to MFC this change and the subsequent API consumer changes?
> O> Lack of this API in 12 makes MFCing unrelated eth driver fixes hard.
> 
> I don't plan to MFC it, but there is nothing that would blocks such MFC.
> 
> Of course internals of the functions would be different - using mutex instead
> of the epoch to sync access to address lists.
> 
> I can provide patch for you, but you would be responsive for MFC. I don't
> have any 12-based systems to test changes.

I only need it for if_dwc, I don't plan to merge code for any other
drivers. I tested this patch[1] and there were no issues. Does it look
sane? Do I need to run specific steps to trigger codepaths with these
new API calls?

[1] https://people.freebsd.org/~gonzo/patches/if-foreach-mfc.diff
-- 
gonzo