[U-Boot-Users] PHY library?

[U-Boot-Users] PHY library?

[David Saada]

Hi,
I have a new PHY that needs to be handled in U-boot. Now this PHY is used in
two differenet ports - TSEC & UCC (it's an MPC8568 board). The problem is
that now I need to duplicate the handling of this PHY in both the TSEC and
UCC drivers. Wouldn't it be better if U-boot had an indepedent PHY library
(as in Linux) used by the different Ethernet drivers?
Regards,
David.
-- 
View this message in context: http://www.nabble.com/PHY-library--tf4849909.html#a13876488
Sent from the Uboot - Users mailing list archive at Nabble.com.

[U-Boot-Users] PHY library?

[Haavard Skinnemoen]

On Wed, 21 Nov 2007 04:51:19 -0800 (PST)
David Saada <David.Saada@ecitele.com> wrote:

> Wouldn't it be better if U-boot had an indepedent PHY library
> (as in Linux) used by the different Ethernet drivers?

Absolutely. At the very least, the multiple MII header files and
duplicate MII register definitions at various places under include/
should be consolidated. But I think it makes sense to take it one step
further and do a MAC-independent PHY library as well.

H?vard

[U-Boot-Users] PHY library?

[Ben Warren]

David Saada wrote:
> Hi,
> I have a new PHY that needs to be handled in U-boot. Now this PHY is used in
> two differenet ports - TSEC & UCC (it's an MPC8568 board). The problem is
> that now I need to duplicate the handling of this PHY in both the TSEC and
> UCC drivers. Wouldn't it be better if U-boot had an indepedent PHY library
> (as in Linux) used by the different Ethernet drivers?
> Regards,
> David.
>   
Absolutely. In fact, I'm working on one right now... Access methods to 
PHYs are currently common (miiphybb.c and miiphyutil.c) but there's no 
reason why the TSEC, UCC and other drivers should have hard wiring for 
individual PHY devices. Any design suggestions you or others have are 
welcome.

regards,
Ben

[U-Boot-Users] PHY library?

[Stefan Roese]

On Wednesday 21 November 2007, Ben Warren wrote:
> > I have a new PHY that needs to be handled in U-boot. Now this PHY is used
> > in two differenet ports - TSEC & UCC (it's an MPC8568 board). The problem
> > is that now I need to duplicate the handling of this PHY in both the TSEC
> > and UCC drivers. Wouldn't it be better if U-boot had an indepedent PHY
> > library (as in Linux) used by the different Ethernet drivers?
> > Regards,
> > David.
>
> Absolutely. In fact, I'm working on one right now...

That's great news. Thanks.

> Access methods to 
> PHYs are currently common (miiphybb.c and miiphyutil.c) but there's no
> reason why the TSEC, UCC and other drivers should have hard wiring for
> individual PHY devices. Any design suggestions you or others have are
> welcome.

Not really. Just "Go Ben, go go go ...!" ;)

Best regards,
Stefan

=====================================================================
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-0 Fax: +49-8142-66989-80  Email: office at denx.de
=====================================================================

[U-Boot-Users] PHY library?

[Jon Loeliger]

On Wed, 2007-11-21 at 09:01, Ben Warren wrote:

> Absolutely. In fact, I'm working on one right now... Access methods to 
> PHYs are currently common (miiphybb.c and miiphyutil.c) but there's no 
> reason why the TSEC, UCC and other drivers should have hard wiring for 
> individual PHY devices. Any design suggestions you or others have are 
> welcome.


Well, there _is_ the long standing tradition of, er, borrowing
code from Linux...where there just happens to be a PHY lib
already in place, written by somebody named Andy Fleming.
We might consider leveraging that some...?

jdl

[U-Boot-Users] PHY library?

[Ben Warren]

Jon Loeliger wrote:
> On Wed, 2007-11-21 at 09:01, Ben Warren wrote:
>
>   
>> Absolutely. In fact, I'm working on one right now... Access methods to 
>> PHYs are currently common (miiphybb.c and miiphyutil.c) but there's no 
>> reason why the TSEC, UCC and other drivers should have hard wiring for 
>> individual PHY devices. Any design suggestions you or others have are 
>> welcome.
>>     
>
>
> Well, there _is_ the long standing tradition of, er, borrowing
> code from Linux...where there just happens to be a PHY lib
> already in place, written by somebody named Andy Fleming.
> We might consider leveraging that some...?
>
> jdl
>
>
>
>   
There will be much blatant ripping-off of Andy's fine library.

Ben

[U-Boot-Users] PHY library?

[Andy Fleming]

On Nov 21, 2007 10:49 AM, Ben Warren <bwarren@qstreams.com> wrote:
> Jon Loeliger wrote:
> > On Wed, 2007-11-21 at 09:01, Ben Warren wrote:
> >
> >
> >> Absolutely. In fact, I'm working on one right now... Access methods to
> >> PHYs are currently common (miiphybb.c and miiphyutil.c) but there's no
> >> reason why the TSEC, UCC and other drivers should have hard wiring for
> >> individual PHY devices. Any design suggestions you or others have are
> >> welcome.
> >>
> >
 [...]
> There will be much blatant ripping-off of Andy's fine library.

Excellent.  It was on my personal to-do list, but I'm still bogged
down in other tasks.  So I second the "Go Ben Go!" sentiment!

If I had one potentially non-obvious design request it's that the PHYs
be accessible by board code, too.  There are a number of boards that
have specific demands on the PHYs.  I suspect you've already got that,
though, since it's currently possible. :)

Andy