2026-04-07, 15:07:47 +0000, Cosmin Ratiu wrote:
> On Thu, 2026-04-02 at 16:48 +0200, Sabrina Dubroca wrote:
> > If this happens on a real device, the VLAN filters will be broken.
> > I'm
> > not sure what the right behavior would be:
> > 
> > 1. reject the request to enable offload
> > 2. switch to promiscuous mode
> 
> I implemented and tested option 1. In the unlikely scenario adding VLAN
> filters prevents offloading

How unlikely is it? Resource allocation, talking to HW, device limits,
anything else that could fail?

> , it's better for the driver to be explicit
> and let the user turn on promisc mode themselves. Keeping track of
> whether VLAN filters failed and promisc was used as a fallback adds
> some extra complexity.

If it's indeed (very) unlikely, sure. There's still a bit of a
regression/change of behavior for users compared to before the
IFF_UNICAST_FLT patch, but I think we can wait until someone
complains, and then add the tracking if that happens.

> What would be the point of IFF_UNICAST_FLT then?

The point is that this would just be a fallback.

> Please let me know if you agree with this approach, so I can send v8
> with it.

If you can confirm it's on the "very unlikely" side, yes, this
approach sounds ok. Thanks.

> > OTOH maybe we don't need to care, since __netdev_update_features also
> > (kind of) ignores those errors:
> > 
[...]
> 
> Well, in this case we have the chance to do something nicer (even
> proper error message back to the user via extack) for a small

(That reminds me I have a branch of "macsec: add extack to X" patches
I still need to polish and submit. I don't think I'll get to it before
the merge window opens, I'll rebase on top of your changes.)

> complexity cost. Perhaps the VLAN filter handling could be improved
> separately.

I'm not sure if this would be an "improvement to VLAN filter handling"
or a "(breaking) change of user-visible behavior". Probably
improvement. Or maybe it's "unlikely enough" that nobody has ever
cared.

-- 
Sabrina

Reply via email to