[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[M16C-dev] Hello

From: Dr. Manuel Kessler
Subject: [M16C-dev] Hello
Date: Mon, 7 Apr 2003 13:01:48 +0200 (CEST)

Hello and welcome

to all members interested in the development of a port of the GNU
toolchain to the Mitsubishi M16C/M32C microcontrollers.

This list has been silent for some time now, mostly because we (the
initial developers, Sven and myself) were not aware of the first
subscribers and thus started using private communication. Sorry to those
who considered themselves as forgotten. As interest is growing, this list
is definitely more appropriate and we hope to see some traffic here soon.

As you may already know, we have started on the first parts of GNU
binutils, that is, the assembler. Of course, this requires some help from
BFD, and is much more amenable to testing if the disassembler,
incorporated for example in objdump, works as well. We decided to simplify
this slightly by using CGEN, which generates the better part of the
assembler/disassembler using a common description of the cpu and its
variants. However, CGEN is more directed towards RISC instruction sets, so
we have to extend this tool a little bit as well, but we still expect a
faster development as if we had written these parts from scratch by hand.

Basically, the current state of affairs (loadable from CVS) is as this:
binutils does compile more or less without hassle, and the assembler can
generate some code. Implemented are the no-operand single opcodes,
starting with nop, as well as the single operand read-modify-write
instructions abs,neg,not,pop,push,rolc,rorc (pop and push are a little bit
different, but share the same instruction format). Others will certainly
follow shortly.

The assembler is able to generate some relocations as well, that the
linker even resolves, but not in the correct way. Sven will take a look at
these issues, but maybe could do with some help.

Before adding more instructions, I will try to consolidate the existing
ones with their many operand combinations using the anyof-operand feature
of CGEN. However, it seems to be unusable right now, so I will contact the
CGEN mailing list for some help. CGEN experts could point me in the
correct direction.

Anyway, that's it for now, I am looking forward to helpful discussions,
close cooperation and fast success.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]