Change in coreboot[master]: Makefile.inc: Update the master header pointer in a separate target

[Arthur Heymans (Code Review)]

Attention is currently required from: Arthur Heymans, Martin Roth, Julius Werner, Yu-Ping Wu.
Arthur Heymans has posted comments on this change. ( https://review.coreboot.org/c/coreboot/+/59132 )

Change subject: Makefile.inc: Update the master header pointer in a separate target
......................................................................

Patch Set 4:

(3 comments)

Commit Message:

https://review.coreboot.org/c/coreboot/+/59132/comment/92eee83c_068d892c 
PS4, Line 9: The makefiles don't like cbfs file names with spaces
> I'm good with how it's done now, but honestly, this seems odd to me.  We should be able to escape it and make it work.  maybe '\ ' so that after the 1st evaluation we get '\ ' presented to the shell?

Thanks, I'll try that.

Patchset:

PS4:
> It might be good to give some background on the cbfs master header in the commit message just so that people have a better understanding of what's going on without having to go search gerrit.  (This is not meant as a criticism.)
>
> Maybe add some of the comments between you and Julius in CB:56120
>
> It probably would have been better in the commit message of that revert, but that's too late now.  It's just a little confusing going through commits and seeing the comments:
> "# the cbfs master header is a deprecated feature only used on x86" and "If FMAP is used anyway there is no need for a cbfs master header.", but still seeing us using it on all platforms.
>
> Again, not a big deal, just some documentation of what's going on with the master header might be useful.

Got it. btw I still plan on removing it in the end, but Libpayload needs to support FMAP for that.

File Makefile.inc:

https://review.coreboot.org/c/coreboot/+/59132/comment/8f207308_ab80de66 
PS4, Line 1150: 32
> Nit: Maybe put some of these numbers into variables?  Some comments as to what's going on here might be good too.
>
> # Add a pointer to the cbfs master header in the last 4 bytes of CBFS

Now that I think of it. I can probably just create a c struct and directly add something useful here. That would allow to drop the whole add-master-header functionality in cbfstool.