[U-Boot-Users] MPC8272 Family

[U-Boot-Users] MPC8272 Family

[Paul Bilke]

I have a dual 8247 platform that I ported u-boot 1.1.4 to awhile back. 
I have been asked to go ahead to move the changes to the current
version.  Right now the include/mcp8260.h has only minimal knowledge of
the 8272 series, only changing the CPI_ID_STR as appropriate.  On these
processors the SIUMCR register and the bits in the HRCW are different
than they are on 8260 (and 8270, 8280 etc).  In 1.1.4 I just had #ifdef
's to support the differences, but I know people don't like the headers
littered with them (actually 4).  This platform will not have wide
distribution so pushing all the changes upstream is not necessary, but I
would like to push these base infrastructure changes up so others can
use them as they build with these chips.  So I am looking for an opinion
as to which strategy to use:
Put the #ifdefs in
Put the new options in without #ifdefs (a mpc8260 define changes the
same bit as a mpc8272 one does)
Duplicate the mpc8260.h into a mpc8272.h file
or split out the processor specific sections into a separate include and
have a single ifdef in the mpc8260.h that includes the right one.
Or maybe no one is interested in these changes, I am open to suggestions.

Also this platform has two processors (core disabled on the second one)
and was wondering if anyone was interested in supporting more than one
iop_conf_tab in the system.  Right now the functions that set up the CPM
assume there is only one in a system (one immr) so the changes are not
pretty.  This could be fixed to support generic configuration of
multiple CPM's by passing the immr to be configured.  If there is
interest I can also clean up these changes and send them up.  Or did
slave processors go out of vogue with the 68360.

Paul

[U-Boot-Users] MPC8272 Family

[Wolfgang Denk]

In message <45F8784B.901@conspiracy.net> you wrote:
>
> use them as they build with these chips.  So I am looking for an opinion
> as to which strategy to use:
> Put the #ifdefs in
> Put the new options in without #ifdefs (a mpc8260 define changes the
> same bit as a mpc8272 one does)
> Duplicate the mpc8260.h into a mpc8272.h file
> or split out the processor specific sections into a separate include and
> have a single ifdef in the mpc8260.h that includes the right one.
> Or maybe no one is interested in these changes, I am open to suggestions.

We already support other MPC8272 based systems.  They  did  not  need
such changes. What am I missing?

> Also this platform has two processors (core disabled on the second one)
> and was wondering if anyone was interested in supporting more than one
> iop_conf_tab in the system.  Right now the functions that set up the CPM

Sure - if it's done in a clean way.

> interest I can also clean up these changes and send them up.  Or did
> slave processors go out of vogue with the 68360.

I know only a single customer who did this (1 x MPC860 + 3 x  68360),
but this predates U-Boot and Embedded Linux by a couple of years.

Best regards,

Wolfgang Denk

-- 
DENX Software Engineering GmbH, HRB 165235 Munich, CEO: Wolfgang Denk
Office:  Kirchenstr. 5,       D-82194 Groebenzell,            Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
He'd been wrong, there _was_ a light at the end of the tunnel, and it
was a flamethrower.                         - Terry Pratchett, _Mort_