Re: [ping 3] New port: Toshiba Media Processor (mep-elf)

[DJ Delorie <dj at redhat dot com>]

Thanks for reviewing my port.  It will take me some time to digest
your suggestions, as MeP is an old port internally and I'll have to
sort through the history to determine the best way to migrate things
forward.

> My only concern is about validate_replace_rtx_subexp, and whether
> there is a newer way of doing what it used to do.

I'm open to suggestions, but I looked and didn't find any.  We use it
in one point - we scan for superfluous register moves and delete them,
so we need a way of replacing rtx found in part of an insn, while
leaving the same rtx in other places in the insn alone, to propogate
the change.

> Similarly most posts are now using constraints.md and predicates.md
> for the constraints and predicates.

I have to also deal with the fact that many of the
coprocessor-specific constraints (when coprocessors are configured in)
are being generated by cgen.  I'll see what I can come up with.

> I don't know if you want farcall_list to be in garbage collected
> memory with GTY(()) markers instead of using xmalloc.  It also may
> be useful to use a hashtab for this instead of a linear list.

It's a relatively short build-once list, I don't think it matters much
how it's implemented.  But again, I'll see what I can do with it.