[PATCH 0/9] New back end ia16: 16-bit Intel x86

[Andrew Jenner <andrew at codesourcery dot com>]

About 10 years ago, Rask Ingemann Lambertsen sent a patch series to add 
a 16-bit x86 (i.e. 8088, 8086, 80186 and 80286 CPUs) back end. This work 
was never committed. Recently I've been doing some work on this back 
end, and today we released a Sourcery CodeBench Lite distribution based 
on it (see 
https://blogs.mentor.com/embedded/blog/2017/04/01/announcing-sourcery-codebench-lite-for-ia16/ 
). Given this, I thought it would be a good idea to send these patches 
upstream today (though obviously if I commit these it won't be until 
we're in Stage 1 and more testing has been done on the middle-end changes).

I have not been able to contact Rask Ingemann Lambertsen, but he does 
have a copyright assignment for GCC on file (dated 2006-11-16) so I do 
not believe there are any copyright assignment issues that would block this.

On top of the original patch, here is what I have done:
* Ported to GCC 6.3.
* Switched from reload to LRA.
* Added C++ support.
* Added startup code, BSP and string functions to Newlib.
* Improved prologue/epilogue generation.
* Implemented stack usage and sibling calls.
* Fixed many bugs.
* Improved the testsuite to have better support for 16-bit ports.

My intention is to continue to develop this back end. I'm happy to be 
maintainer of it as an in-tree port or (if the steering committee does 
not want this port in-tree) as an out-of-tree port. I plan to add 
further code generation improvements and features including:
* Far pointers.
* Support for OMF .obj files and DOS .exe files in binutils.
* x87 floating point.
* Debugging support.
* More C library functions (to improve compatibility with other DOS 
compilers).

Contents of the patch set:
1: Top-level config.sub.
2: The port itself.
3: Documentation.
4: libstdc++ changes.
5: libstdc++ testsuite changes.
6: gcc testsuite changes.
7: ira-color.c fix.
8: subreg_get_info fix.
9: c-family/c-cppbuiltin.c fix.

Test results are at
https://gcc.gnu.org/ml/gcc-testresults/2017-04/msg00088.html


To answer some questions that came up last time the port was submitted:

* This port is only useful for targetting CPUs prior to the 80386. If 
you want to generate code for a 32-bit x86 CPU that runs in 16-bit mode 
(e.g. for a bootloader) the i386 target has the -m16 option for this case.

* I'm not interested in changing the name from ia16. I'd have preferred 
i086 or i86 (by analogy with i386) but I don't feel strongly enough to 
argue about it or change it. If someone else cares strongly enough about 
it to write the patches, I'll consider it.

* I'm also not interested in merging this port with the i386 back end. 
The targets are too different, and it would create an excessive 
maintenance burden for ia16 and (worse) for i386.

Andrew