Introducing a nanoMIPS port for Binutils, GDB and GOLD

[Cary Coutant]

> Patches used from the binutils mailing list:
>
> - Extended support of INPUT and GROUP commands to linker scripts added with -T
>   (https://sourceware.org/ml/binutils/2016-07/msg00036.html)

Sorry, I've sat on this patch way too long; I should have reviewed it
long ago. On first look, I didn't see any problems, but I kind of
remember feeling that there may have been a better approach, and meant
to take a closer look. I'll do that.

> - Fixed the assignment of LMA for subsequent sections which do not have their
>   own explicit address set
>   (https://sourceware.org/ml/binutils/2016-10/msg00161.html)

This one looked OK, too. Apologies, again.

> - Fixed NOLOAD behavior
>   (https://sourceware.org/ml/binutils/2015-09/msg00216.html)

There were issues with this patch. See the discussion starting with
this message:

   https://sourceware.org/ml/binutils/2015-09/msg00179.html

The problem is that this makes NOLOAD mean something different from
what's documented, and we had known use cases for what it currently
does (those cases may not exist anymore, but I can't be certain).

There really is no need to declare a section NOBITS, which is what the
patch sought to turn "NOLOAD" into. An output section whose contents
are all NOBITS, and which falls at the end of a segment, should be
turned into a NOBITS section automatically. Gold didn't do that
correctly originally, but that should have been fixed with the
following patch:

   https://sourceware.org/ml/binutils/2016-12/msg00184.html

Looking forward to your patches! Did you implement this as a new
target, or as part of the current mips target?

(Also, please be aware that I'll be away with little or no internet
access from May 10-27, so I won't be able to start reviewing patches
until the end of the month.)

-cary