This is the mail archive of the
cgen@sourceware.org
mailing list for the CGEN project.
Re: better use of disassembly hash table by mep port
- From: DJ Delorie <dj at redhat dot com>
- To: Doug Evans <dje at sebabeach dot org>
- Cc: fche at redhat dot com, cgen at sourceware dot org
- Date: Mon, 21 Dec 2009 14:25:46 -0500
- Subject: Re: better use of disassembly hash table by mep port
- References: <4B2DE87B.2030806@sebabeach.org>
> I think I understand the MEP port enough to have it starting using the
> disassembly hash table better (rather than hashing every insn to the
> same slot, blech).
MeP does that too? I mentioned m32c had that problem, and it was a
problem even with one specific ISA (the chip is weird as far as opcode
encoding goes). m32c is compounded by the explosion of opcodes caused
by the way cgen does variable-length insns.
My hash thoughts for m32c was to take each opcode, figure out which
bits are decodable for *that* opcode, and use those to put the opcode
into one or more hash entries. That allows us to use a wider set of
bits, thus fewer entries per bucket, while sidestepping the m32c's
decodable bits problem.
OTOH I've also started working on a non-cgen port of m32c.