From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 13569 invoked by alias); 16 Jun 2005 15:42:07 -0000 Mailing-List: contact sid-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: sid-owner@sources.redhat.com Received: (qmail 13552 invoked by uid 22791); 16 Jun 2005 15:42:00 -0000 Received: from neon-gw-l3.transmeta.com (HELO neon-gw.transmeta.com) (63.209.4.196) by sourceware.org (qpsmtpd/0.30-dev) with ESMTP; Thu, 16 Jun 2005 15:42:00 +0000 Received: from victor.transmeta.com (victor.transmeta.com [10.0.2.120]) by neon-gw.transmeta.com (Postfix) with ESMTP id 67CBA5F8041; Thu, 16 Jun 2005 08:41:57 -0700 (PDT) Received: from localhost (localhost.localdomain [127.0.0.1]) by localhost.transmeta.com (Postfix) with ESMTP id C40744F802C; Thu, 16 Jun 2005 08:41:58 -0700 (PDT) Received: from victor.transmeta.com ([127.0.0.1]) by localhost (victor [127.0.0.1]) (amavisd-new, port 10022) with LMTP id 29033-04-49; Thu, 16 Jun 2005 08:41:58 -0700 (PDT) Received: from casey.transmeta.com (casey.transmeta.com [10.10.25.22]) by victor.transmeta.com (Postfix) with ESMTP id ABC7D4F8003; Thu, 16 Jun 2005 08:41:58 -0700 (PDT) Received: (from dje@localhost) by casey.transmeta.com (8.11.6/8.11.6) id j5GFfwL09900; Thu, 16 Jun 2005 08:41:58 -0700 From: Doug Evans MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <17073.40390.435362.400645@casey.transmeta.com> Date: Thu, 16 Jun 2005 15:42:00 -0000 To: "Frank Ch. Eigler" Cc: cgen@sources.redhat.com, sid@sources.redhat.com Subject: Re: [RFC] New cpu --- Morpho ms1 In-Reply-To: <20050616152212.GC27067@redhat.com> References: <42A752B2.8090407@rogers.com> <20050613145452.GE18930@redhat.com> <42B09F6F.7020302@redhat.com> <17072.41406.147361.75051@casey.transmeta.com> <20050616152212.GC27067@redhat.com> X-SW-Source: 2005-q2/txt/msg00040.txt.bz2 Frank Ch. Eigler writes: > Hi - > > On Wed, Jun 15, 2005 at 02:46:38PM -0700, Doug Evans wrote: > > Imagine if bison or autoconf had encoded in them knowledge of particular > > applications that used them. > > [...] > > Well, the RTL model needed extension to allow for delayed operations. > Since this part of the RTL was only used for instruction semantics, > and thus simulators, adding some more specialized code is not that bad. - nit: it's not necessarily the case that insn semantics would only be used by simulators - e.g. the operand tables derived from insn semantics are used by m32r gas - even if it were, the patch hardwires the test to SID-SIMULATOR - adding specialized code can be "that bad" because without adequate commenting like "do as I say, not as I do", more and more such code creeps in over time - imagine if there's 20 different apps (as in (eq? APPLICATION mumble)) all with their own little tweaks to rtl-c.scm Remove the appearance of the text "SID" from rtl-c.scm and parameterize the feature on something generic and I'd be happier.