[coreboot] some thoughts on dts vs. Kconfig vs. makefiles vs. C code

Carl-Daniel Hailfinger c-d.hailfinger.devel.2006 at gmx.net
Wed Mar 5 17:33:36 CET 2008


On 05.03.2008 17:10, ron minnich wrote:
> On Wed, Mar 5, 2008 at 8:03 AM, Carl-Daniel Hailfinger
> <c-d.hailfinger.devel.2006 at gmx.net> wrote:
>
>   
>>  No, not the tree mechanism. I was thinking more a long the lines of
>>  another dtc output mode which creates a flat .h file which has no
>>  structure, but contains all specially marked dts properties as #define.
>>     
>
> go ahead and propose a syntax, I'm very slow in the morning.
>   

The syntax you propose below would suit me just fine.

>>  I hope I clarified this. statictree.c in stage 1 would indeed be a bad idea.
>>     
>
> Ah, but would it. Look at the cs5536 startup code -- we don't
> parameterize it yet. We need to.
>
> We could do this: if there is a property in a node called "cpp", it is
> a list of values in that node that should be emitted as #defines.
> The names will be prefixed with the pathname of the dts, so
> if we had
> cpp = "com1enable";
> we get (ugly)
> SOUTHBRiDGE_AMD_CS5536_COM1ENABLE
> or some such.
>   

Nice one! I really like it.

Regards,
Carl-Daniel

-- 
http://www.hailfinger.org/





More information about the coreboot mailing list