git: 68a46f11eada - main - e1000: Restore VF interface random MAC

[Kevin Bowling]

I would like to discuss this further, it crossed my mind, but none of
the intel drivers currently use it.  If this is the right KPI I think
a review should be prepared that updates e1000, ixgbe (ixv), ixl
(iavf) in one go.

Regards,
Kevin

On Thu, Apr 15, 2021 at 12:36 PM Kyle Evans <kevans at freebsd.org> wrote:
>
> On Thu, Apr 15, 2021 at 2:32 PM Ronald Klop <ronald-lists at klop.ws> wrote:
> >
> > On 4/15/21 8:48 PM, Kevin Bowling wrote:
> > > The branch main has been updated by kbowling (ports committer):
> > >
> > > URL: https://cgit.FreeBSD.org/src/commit/?id=68a46f11eadab48a1da9e3d3900569a6a1ce142e
> > >
> > > commit 68a46f11eadab48a1da9e3d3900569a6a1ce142e
> > > Author:     Kevin Bowling <kbowling at FreeBSD.org>
> > > AuthorDate: 2021-04-15 18:45:02 +0000
> > > Commit:     Kevin Bowling <kbowling at FreeBSD.org>
> > > CommitDate: 2021-04-15 18:45:02 +0000
> > >
> > >      e1000: Restore VF interface random MAC
> > >
> > >      Restore 525e07418c77 after the iflib conversion of igb(4). This
> > >      reenables random MAC address generation when attaching to a VF with a
> > >      zeroed MAC.
> > >
> > >      PR:             253535
> > >      Reported by:    Balaev PA <mail at void.so>
> > >      Reviewed by:    markj
> > >      MFC after:      2 weeks
> > >      Differential Revision:  https://reviews.freebsd.org/D29785
> > > ---
> > >   sys/dev/e1000/if_em.c | 21 ++++++++++++++++++---
> > >   1 file changed, 18 insertions(+), 3 deletions(-)
> > >
> > > diff --git a/sys/dev/e1000/if_em.c b/sys/dev/e1000/if_em.c
> > > index 6f022c80c01c..79a617b3342f 100644
> > > --- a/sys/dev/e1000/if_em.c
> > > +++ b/sys/dev/e1000/if_em.c
> > > @@ -1061,9 +1061,17 @@ em_if_attach_pre(if_ctx_t ctx)
> > >       }
> > >
> > >       if (!em_is_valid_ether_addr(hw->mac.addr)) {
> > > -             device_printf(dev, "Invalid MAC address\n");
> > > -             error = EIO;
> > > -             goto err_late;
> > > +             if (adapter->vf_ifp) {
> > > +                     u8 addr[ETHER_ADDR_LEN];
> > > +                     arc4rand(&addr, sizeof(addr), 0);
> > > +                     addr[0] &= 0xFE;
> > > +                     addr[0] |= 0x02;
> > > +                     bcopy(addr, hw->mac.addr, sizeof(addr));
> > > +             } else {
> > > +                     device_printf(dev, "Invalid MAC address\n");
> > > +                     error = EIO;
> > > +                     goto err_late;
> > > +             }
> >
> >
> > Just curious. Would ether_gen_addr() be useful here?
> > It is implemented in net/if_ethersubr.c.
> >
>
> I had asked myself the same question, but I suspect these machines may
> not necessarily have a hostuuid preloaded to avoid collisions from the
> generated addresses.
>
> Thanks,
>
> Kyle Evans