[PATCH 1/2] opcodes/mips: use .word/.short for undefined instructions

[Maciej W. Rozycki]

Andrew,

> I've updated the patch.  Let me know what you think.

 I'd say it's OK, except that I put your change through my MIPS regression 
tester and that revealed failures from your new case for numerous targets, 
e.g.:

mips-elf  +FAIL: microMIPS source file contains reserved encoding (o32)
mips-img-elf  +FAIL: microMIPS source file contains reserved encoding (o32)
mips-img-elf  +FAIL: microMIPS source file contains reserved encoding (n32)

etc.  The usual suspect is section padding owing to different alignments 
used with individual MIPS targets, e.g.:

extra regexps in 
.../binutils/testsuite/binutils-all/mips/micromips-reserved-enc-o32.d starting with "^
\.\.\.$"
EOF from tmpdir/dump.out
FAIL: microMIPS source file contains reserved encoding (o32)

See e.g. binutils/testsuite/binutils-all/mips/micromips-branch-alias.s for 
how to add suitable padding at the end.

 And then:

mipsisa32r6-elf  +FAIL: microMIPS source file contains reserved encoding (o32)
mipsisa32r6-linux  +FAIL: microMIPS source file contains reserved encoding (o32)
mipsisa32r6el-elf  +FAIL: microMIPS source file contains reserved encoding (o32)

etc., due to:

.../binutils/testsuite/binutils-all/mips/micromips-reserved-enc.s: Assembler messages:
.../binutils/testsuite/binutils-all/mips/micromips-reserved-enc.s:3: Fatal error: `micromips' cannot be used with `mips32r6'

We don't care about different ISA levels here, so let's set a reasonable 
one, as in binutils/testsuite/binutils-all/mips/micromips-branch-alias.s 
again:

	.module	mips64r3

(it could be `.set' too, but let's be consistent, and it has to be a 
64-bit one for the n32/n64 ABIs).

 OK with these updates, thank you for your contribution.

  Maciej