Re: [RFA:] The CRIS simulator.

[Andrew Cagney <cagney@gnu.org>]

Hans-Peter Nilsson wrote:
> Assumes ok on all my previous patches.  The simulator calls use the Linux
> ABI, which means that statically linked binaries work on both target and
> simulator.  (It shouldn't be too hard to support dynamically linked
> binaries, but I haven't needed it yet.)  Not all syscalls are implemented
> of course, and only a subset of valid parameters, but a good deal more
> than what's necessary for "hello, world" and the GCC testsuite.  I don't
> use the nltvals machinery for any of the constants, because it has
> mismatches with the Linux definitions.  I also didn't find it useful to
> spread out the ABI; better to keep it in one file.  The manually written
> files are warnings-free with "-W -Wall" using the RH 7.3 derivate of gcc,
> which cannot be said of the automatically generated files nor of
> sim/common.  Ugh.  Testsuite directory later; this patch is already 167k.
> 
> Ok to commit?
> 
> 	* cris: New directory, simulator for Axis Communications CRIS
> 	including CRISv32, CGEN-based.
> 	* configure.in: Add corresponding configury.

To be explicit, the above (once you've updated it to meddle with 
configure.ac), is ok.  You're also free to commit further changes to 
cris code (which includes this cris sim :-).

Please be careful to keep Orjan in the loop on any more complicated changes.

Have fun!

Andrew