Le 21/11/2018 à 17:03, Mario.Limonciello@xxxxxxxx a écrit :
Sent from Workspace ONE Boxer
On Nov 21, 2018 09:51, Frédéric Parrenin
<frederic.parrenin@xxxxxxxxxxxxxxxxxxxxxx> wrote:
>
>
> [EXTERNAL EMAIL]
>
>
> Le 21/11/2018 à 15:19, Frédéric Parrenin a écrit :
> >
> > Le 21/11/2018 à 15:01, Mario.Limonciello@xxxxxxxx a écrit :
> >> On Nov 21, 2018 02:09, Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> >> >
> >> >
> >> > [EXTERNAL EMAIL]
> >> >
> >> > On Tue, Nov 20, 2018 at 09:10:22PM +0100, Bjørn Mork wrote:
> >> > > Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> writes:
> >> > >
> >> > > > I do not see any USB networking device here at all.
> >> > >
> >> > > No, It wasn't easy to see. But it's there both with and
without the
> >> > > feature enabled:
> >> > >
> >> > > /: Bus 02.Port 1: Dev 1, Class=root_hub, Driver=xhci_hcd/6p,
5000M
> >> > > | Port 4: Dev 10, If 0, Class=Hub, Driver=hub/7p, 5000M
> >> > > |__ Port 2: Dev 11, If 0, Class=Vendor Specific Class,
> >> Driver=r8152, 5000M
> >> >
> >> > Ah, sorry, I missed that.
> >> >
> >> > > This feature is only for convenience and should never prevent the
> >> > > network from working. Unless the other end is confused by seeing
> >> the
> >> > > same mac address on two ports... Better take down or
disconnect the
> >> > > undocked interface after docking.
> >> >
> >> > Ok, so it seems that this driver does not support the pass-through
> >> > option for some reason.
> >> >
> >> > Mario, any ideas what Frédéric can do here to help debug the
> >> issue? He
> >> > has some Dell hardware that does not seem to be working
properly with
> >> > the pass-through option.
> >> >
> >> > I place odds that the driver needs some kind of update for newer
> >> device
> >> > signatures, which is what some of us worried about a long time ago
> >> when
> >> > this patch first went in...
> >> >
> >>
> >> I don't have the whole context for this thread but I think I'm
> >> following that the question is why MAC address pass thru isn't
> >> activating on DS1000?
> >>
> >> The criteria that we previously put in for this patch is still the
> >> correct criteria; R8153-AD with a particular eFuse blown.
> >>
> >> I never heard DS1000 is supposed to meet that criteria. I thought
> >> official documentation referred to TB16 and WD15 only for Mac
> >> passthrough.
> >>
> >> FYI I'm on holiday right now so it might take me some time to get
> >> back after this response.
> >>
> >>
> > Some online shops selling the DS1000 mention it supports Mac Address
> > pass-through, for example:
> >
> > https://shop.bechtle.de/medias/VXav7KboWIcl8FlvCBYV58-30.pdf
> >
https://www.amazon.com/Dell-USB-Type-Monitor-Release-VT96R/dp/B01G6VL3I4/ref=sr_1_1?ie=UTF8&qid=1542809915&sr=8-1&keywords=DS1000
> >
> >
> > But actually, this might be a wrong information.
> > Sorry for the noise if this is the case, and too bad the DS1000 does
> > not support pass-through.
> >
> I asked the question on amazon.com, and the guy said:
>
> > Dell Mac address pass-through is enabled in the BIOS of the Dell
laptop
> which over rides the Mac on the dock. So, the answer is, "it's not an
> issue if the dock supports it, as the laptop BIOS is what is
determining
> if it is supported".
>
> So what is the truth?
>
> Unfortunately, I don't have Windows on that machine to test.
>
They work in tandem. A Mac address for the system needs to be reserved
at manufacturing time. It's stored in NVRAM for the system BIOS to
access and populate into an ACPI table.
If your system is too old or no Mac address is reserved for that model
(it's currently typically Latitude, Precision, XPS only feature) then
no ACPI entry is created.
The driver will parse that information when using appropriate
components like I described in previous email.
Yes, I have a Latitude 5285 which does support Mac Address pass through.
But when I activate that option in the bios and connect to the DS1000,
the network does not work and "ip a" does not display the laptop's Mac
address but the dock Mac address.
The question I have if whether this is the DS1000's fault (it does not
support that feature) or linux's fault.
