[Driver][M68k] (Patch 8/8) Add driver support for M68k

[myhsu]

Summary

Add new toolchain and driver options for the M68k target

[myhsu]

Fix formatting issues

[myhsu]

• Use the canonical CPU name (i.e. names started with upper case 'M')
• Tell the driver to use integrated assembler (i.e. MC) by default

[myhsu]

[NFC] Rename M680x0 to M68k

[RKSimon]

Some very minor nits - the 68060 omission is the biggest one (apple might not have used it but commodore did!)

clang/include/clang/Driver/Options.td

168

(sorting) put this before mips?

clang/lib/Driver/ToolChains/Arch/M68k.cpp

45

Why no 68060 ?

65

(style) remove unnecessary braces?

73

(clang-format) indentation?

95

remove braces

clang/lib/Driver/ToolChains/Arch/M68k.h

36

(clang-format) indentation?

39

(clang-tidy) end namespace tools

clang/lib/Driver/ToolChains/Clang.cpp

383

(sorting) move before msp430 ?

clang/lib/Driver/ToolChains/CommonArgs.cpp

371

(sorting) move down to before mips ?

[rengolin]

clang/include/clang/Driver/Options.td

3234

Same question as @RKSimon had below: Shouldn't this cover all models the back-end recognises?

clang/lib/Driver/ToolChains/Gnu.cpp

2103

The front-end supports FreeBSD, too.

[RKSimon]

clang/lib/Driver/ToolChains/Arch/M68k.cpp

52

Can't we just use StringSwitch here?

[bruno]

clang/include/clang/Driver/Options.td

3234

Unless you are planning to add 100 or more target variations I'd prefer to see these explicitly defined instead of a foreach. If I'm grepping for a specific CPU variation in the code base it's nice to get that information easily.

clang/lib/Driver/ToolChains/Arch/M68k.cpp

38

No need for this else here.

52

+1

65

and also the unnecessary elses.

[myhsu]

• Rebased to latest changes
• Addressed some of the feedbacks

clang/include/clang/Driver/Options.td

3234

@rengolin I think the backend currently doesn't support M68060 either

[myhsu]

• Add support for M68060 sub-target

[myhsu]

clang/include/clang/Driver/Options.td

3234

Update: I've just put the sub-target (skeleton) for M68060. So now you can specific M68060 :-)

clang/lib/Driver/ToolChains/Arch/M68k.cpp

45

@RKSimon Just added 060's support :-)
Currently I don't see any major ISA difference between 060 and 040 (correct me if I'm wrong). So I just put a sub-target skeleton and clang support for 060, and fill in the details in the future

[bruno]

clang/include/clang/Driver/Options.td

164

Looks like this is missing clang-format?

[craig.topper]

clang/include/clang/Driver/Options.td

164

I don't think clang-format really understands tablegen files.

[rengolin]

clang/include/clang/Driver/Options.td

164

It does not. Don't clang-format table-gen files, the end result is horrendous. :)

[myhsu]

• [NFC] Fix minor formatting

[bruno]

LGTM, it would be great if someone else stamps this too, in case I missed something.

clang/include/clang/Driver/Options.td

164

Thanks for pointing out, thinko for "please format"

[rengolin]

LGTM too, thanks!

[MaskRay]

clang/lib/Driver/ToolChains/Arch/M68k.h

38

mismatching end and no end?

clang-format seems to drop end. I think that style can be picked if convenient.

clang/lib/Driver/ToolChains/Gnu.cpp

2101

The second const is redundant

[jrtc27]

clang/lib/Driver/ToolChains/Gnu.cpp

2103

Aren't these arrays only used on multiarch systems so irrelevant for BSDs? There aren't FreeBSD triples listed for X86/X86_64 for example.

[rengolin]

clang/lib/Driver/ToolChains/Gnu.cpp

2103

Makes sense.