pq2fads support...

Hello Scott and all,

In your patch for the pq2fads support, the device tree seems to be missing  
a memory node, for which dtc complains loudly. Is this board tested to  
boot with these patches (because I'm now using this as a template for my  
board - with your patchset applied to Paul's git)

thanks a lot

Alex

PS. What's about this new cpm-serial driver? is this supposed to be a  
replacement for cpm_uart?

On Fri, Jul 27, 2007 at 03:24:08PM +0300, Alexandros Kostopoulos wrote:

In your patch for the pq2fads support, the device tree seems to be missing  
a memory node, for which dtc complains loudly. Is this board tested to  
boot with these patches (because I'm now using this as a template for my  
board - with your patchset applied to Paul's git)

Sorry about that -- dtc complains about a lot of silly things, falsely
assuming that the result is intended to be a complete tree ready for
passing to Linux (which is rarely true).  I'd been running with a patch
to remove some of those checks, but it got rejected, and I forgot to
stick the templates back into the dts.

The next patchset will have a dts that works with current dtc; in the
meantime, you can add the -f flag to dtc in arch/powerpc/boot/wrapper, or
add the missing nodes based on the mpc8272ads device tree (the actual
information will be filled in by the bootwrapper).

PS. What's about this new cpm-serial driver? is this supposed to be a  
replacement for cpm_uart?

No, it's a minimal driver for the bootwrapper.

-Scott

Thanks for your reply, Scott,

one question though: I am NOT using the bootwrapper, I'm instead feeding  
dtb and uImage directly to u-boot. This probably means that I should fill  
in lots of stuff manually into dts, right? But, (and please forgive my  
ignorance) who does actually fill in stuff (like e.g. clock-frequency or  
bus-frequency), except, of course, from the trivial modifications u-boot  
does to dtb? The wrapper script seems to do nothing more that simply  
running dtc on dts.

Thanks again

Alex

On Fri, 27 Jul 2007 19:35:29 +0300, Scott Wood [off-list ref]  
wrote:

On Fri, Jul 27, 2007 at 03:24:08PM +0300, Alexandros Kostopoulos wrote:

quoted

In your patch for the pq2fads support, the device tree seems to be  
missing
a memory node, for which dtc complains loudly. Is this board tested to
boot with these patches (because I'm now using this as a template for my
board - with your patchset applied to Paul's git)

Sorry about that -- dtc complains about a lot of silly things, falsely
assuming that the result is intended to be a complete tree ready for
passing to Linux (which is rarely true).  I'd been running with a patch
to remove some of those checks, but it got rejected, and I forgot to
stick the templates back into the dts.

The next patchset will have a dts that works with current dtc; in the
meantime, you can add the -f flag to dtc in arch/powerpc/boot/wrapper, or
add the missing nodes based on the mpc8272ads device tree (the actual
information will be filled in by the bootwrapper).

quoted

PS. What's about this new cpm-serial driver? is this supposed to be a
replacement for cpm_uart?

No, it's a minimal driver for the bootwrapper.

-Scott

Alexandros Kostopoulos wrote:

Thanks for your reply, Scott,

one question though: I am NOT using the bootwrapper, I'm instead 
feeding  dtb and uImage directly to u-boot. This probably means that
I should fill  in lots of stuff manually into dts, right?

No, u-boot should fill in the same stuff.  It may require the empty
fields to be there, though (it definitely requires it if you're using
the old, non-libfdt code -- if you are, you should also remove the 
/chosen node, or else you'll get two of them).

But, (and please forgive my  ignorance) who does actually fill in
stuff (like e.g. clock-frequency or  bus-frequency), except, of
course, from the trivial modifications u-boot  does to dtb? The
wrapper script seems to do nothing more that simply  running dtc on
dts.

In the bootwrapper case, it's the code in cuboot-pq2.c and devtree.c. 
In the no-bootwrapper case, it's u-boot.

-Scott