question on MAC passthrough and multiple devices in r8152

question on MAC passthrough and multiple devices in r8152

[Oliver Neukum]

Hi,

I am looking at the logic the r8152 driver uses to
assign a pass through MAC and I do not understand how
you make sure that it is passed to one device only.
As far as I can see the MAC comes from an ACPI object,
hence the driver will happily read it out and assign
it multiple times. Am I overlooking something?

	Regards
		Oliver

Re: question on MAC passthrough and multiple devices in r8152

[Jakub Kicinski]

On Wed, 27 Jul 2022 12:51:44 +0200 Oliver Neukum wrote:
> I am looking at the logic the r8152 driver uses to
> assign a pass through MAC and I do not understand how
> you make sure that it is passed to one device only.
> As far as I can see the MAC comes from an ACPI object,
> hence the driver will happily read it out and assign
> it multiple times. Am I overlooking something?

Just to be sure - have you seen the giant threads on the topic?
IIRC lenovo was sending patches to extend the mechanism, broke
some people's setups and to much teeth grinding we accepted the
somewhat improved patches and moved on. The entire mechanism is
quite poorly designed, we couldn't come up with a better way :(

RE: question on MAC passthrough and multiple devices in r8152

[Hayes Wang]

Oliver Neukum <oneukum@suse.com>
> Sent: Wednesday, July 27, 2022 6:52 PM
> To: Hayes Wang <hayeswang@realtek.com>
> Cc: netdev@vger.kernel.org
> Subject: question on MAC passthrough and multiple devices in r8152
> 
> Hi,
> 
> I am looking at the logic the r8152 driver uses to
> assign a pass through MAC and I do not understand how
> you make sure that it is passed to one device only.
> As far as I can see the MAC comes from an ACPI object,
> hence the driver will happily read it out and assign
> it multiple times. Am I overlooking something?

You are right. And, the patches are not submitted by me.
As far as I know, our customers don't care the situation.
They would provide the guide to describe it. Besides, the
feature could be disabled through BIOS.

In addition, the Dell limits the specific devices to
use the feature of MAC passthrough. It could reduce such
situation.

Best Regards,
Hayes

Re: question on MAC passthrough and multiple devices in r8152

[Oliver Neukum]

On 28.07.22 09:59, Hayes Wang wrote:

> In addition, the Dell limits the specific devices to
> use the feature of MAC passthrough. It could reduce such
> situation.

Understood. Nevertheless, giving out the same MAC twice
at a time is a bad idea.

	Regards
		Oliver