Re: is there still any need PPC checking for "chosen@0"?

2018-08-08 Thread Michael Ellerman
"Robert P. J. Day"  writes:
>   given that there are no .dts files in the current kernel code base
> that define the node name "/chosen@0" instead of the proper "/chosen",

A good portion of PPC machines get their device tree from firmware, not
from a dts that's in the kernel tree, so that's not a good indication.

> is there any need for arch/powerpc/boot/oflib.c to still make this
> test:
>
> chosen = of_finddevice("/chosen");
> if (chosen == (phandle) -1) {
> chosen = of_finddevice("/chosen@0");   <--- this
> if (chosen == (phandle) -1) {
> printf("no chosen\n");
> return 0;
> }
> }
>
> are there still PPC machines that require the recognition of
> "/chosen@0"?

It was added by Paul in:

66a45dd3620e ("powerpc: Make COFF zImages for old 32-bit powermacs")

So presumably there's an old powermac somewhere that needs it.

Given it's basically 2 extra lines of code I'd be inclined to leave it
alone.

cheers


is there still any need PPC checking for "chosen@0"?

2018-08-06 Thread Robert P. J. Day


  given that there are no .dts files in the current kernel code base
that define the node name "/chosen@0" instead of the proper "/chosen",
is there any need for arch/powerpc/boot/oflib.c to still make this
test:

chosen = of_finddevice("/chosen");
if (chosen == (phandle) -1) {
chosen = of_finddevice("/chosen@0");   <--- this
if (chosen == (phandle) -1) {
printf("no chosen\n");
return 0;
}
}

are there still PPC machines that require the recognition of
"/chosen@0"?

rday

-- 


Robert P. J. Day Ottawa, Ontario, CANADA
  http://crashcourse.ca/dokuwiki

Twitter:   http://twitter.com/rpjday
LinkedIn:   http://ca.linkedin.com/in/rpjday