if_wm: support for INTEL 82571EB_SERDES

[Matthias Drochner]

Hi -
I got a board with such a chip. There is practically
no PHY, no autonegoiation etc. Data are transferred
on a backplane.

The appended patch makes it work. It is obviously quite
destructive -- just to show where changes are needed.
I could turn that into its own set of
mediainit/change/check functions. Just don't want to
step onto any maintainer's toes.
So does anyone care?

Another question is what if_media.h mediatype to use.
Technically it is 1000base-bx (sometimes called -kx).
There is already a -cx, the only difference is the
impedance (which is 100ohm). Using cx is wrong, adding
a new one doesn't make sense because the driver can't
tell the difference anyway... It's about the same as
with singlemode/multimode fiber, there is just too much
nonsense in that header.

best regards
Matthias



------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------
Forschungszentrum Juelich GmbH
52425 Juelich
Sitz der Gesellschaft: Juelich
Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498
Vorsitzende des Aufsichtsrats: MinDir'in Baerbel Brumme-Bothe
Geschaeftsfuehrung: Prof. Dr. Achim Bachem (Vorsitzender),
Dr. Ulrich Krafft (stellv. Vorsitzender), Prof. Dr.-Ing. Harald Bolt,
Prof. Dr. Sebastian M. Schmidt
------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------

[Hisashi T Fujinaka]

Hi, Matthias.

Post by Matthias Drochner
Hi -
I got a board with such a chip. There is practically
no PHY, no autonegoiation etc. Data are transferred
on a backplane.
The appended patch makes it work. It is obviously quite
destructive -- just to show where changes are needed.
I could turn that into its own set of
mediainit/change/check functions. Just don't want to
step onto any maintainer's toes.
So does anyone care?

I will merge your code and send the diff to you before commit because
I have no serdes devices. Is it OK?

Post by Matthias Drochner
Another question is what if_media.h mediatype to use.
Technically it is 1000base-bx (sometimes called -kx).
There is already a -cx, the only difference is the
impedance (which is 100ohm). Using cx is wrong, adding
a new one doesn't make sense because the driver can't
tell the difference anyway... It's about the same as
with singlemode/multimode fiber, there is just too much
nonsense in that header.
best regards
Matthias

I read the IEEE 802.3 document and it said "1000BASE-BX10", so I should
call it as -BX10.

Tell me what you want me to test and I'll try to get the hardware
together here in my "day job".

[Matthias Drochner]

Hi Masanobu -

I will merge your code and send the diff to you before commit because
I have no serdes devices. Is it OK?

Yes that would be great.
Btw, ther is certainly a way to query the link state whether
it is up or not. (which would mean that there are idle characters
coming in from the other side) I just can't test it because my
board is plugged into a backplane, and I can't control the
other side, and for obvious reasons I can't pull out the module.
So as long as the module is there the link is up and I can't
do anything about it.

I read the IEEE 802.3 document and it said "1000BASE-BX10", so I should
call it as -BX10.

But this is not in if_media.h. And the driver will never be able to tell
whether the lanes are connected to a backplane or a cable, so
it doesn't make sense to have distinct media types.
For me, _AUTO or _MANUAL would do it as well because there is
no choice anyway.

best regards
Matthias



------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------
Forschungszentrum Juelich GmbH
52425 Juelich
Sitz der Gesellschaft: Juelich
Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498
Vorsitzende des Aufsichtsrats: MinDir'in Baerbel Brumme-Bothe
Geschaeftsfuehrung: Prof. Dr. Achim Bachem (Vorsitzender),
Dr. Ulrich Krafft (stellv. Vorsitzender), Prof. Dr.-Ing. Harald Bolt,
Prof. Dr. Sebastian M. Schmidt
------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------

--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-***@muc.de

[Jason Thorpe]

Post by Matthias Drochner
But this is not in if_media.h. And the driver will never be able to tell
whether the lanes are connected to a backplane or a cable, so
it doesn't make sense to have distinct media types.
For me, _AUTO or _MANUAL would do it as well because there is
no choice anyway.

I think _MANUAL is fine. I did a similar thing with the tlp driver on
the DEC Multia.

Post by Matthias Drochner
best regards
Matthias
------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------
Forschungszentrum Juelich GmbH
52425 Juelich
Sitz der Gesellschaft: Juelich
Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498
Vorsitzende des Aufsichtsrats: MinDir'in Baerbel Brumme-Bothe
Geschaeftsfuehrung: Prof. Dr. Achim Bachem (Vorsitzender),
Dr. Ulrich Krafft (stellv. Vorsitzender), Prof. Dr.-Ing. Harald Bolt,
Prof. Dr. Sebastian M. Schmidt
------------------------------------------------------------------------------------------------
------------------------------------------------------------------------------------------------

-- thorpej


--
Posted automagically by a mail2news gateway at muc.de e.V.
Please direct questions, flames, donations, etc. to news-***@muc.de