home *** CD-ROM | disk | FTP | other *** search
- Path: shellx.best.com!usenet
- From: Remi Lenoir <remi@dvsystems.com>
- Newsgroups: comp.sys.amiga.emulations
- Subject: Re: Bridgeboard to AmiTCP through Zorro Bus
- Date: Fri, 09 Feb 1996 12:50:26 -0800
- Organization: Digital Video Systems
- Message-ID: <311BB392.4F45@dvsystems.com>
- References: <4eqma4$9oo@kocrsv08.delcoelect.com> <311270CE.7F80@dvsystems.com> <4ffolh$59s@vixen.cso.uiuc.edu>
- NNTP-Posting-Host: pepper.vip.best.com
- Mime-Version: 1.0
- Content-Type: text/plain; charset=us-ascii
- Content-Transfer-Encoding: 7bit
- X-Mailer: Mozilla 2.0b6a (Win95; I)
-
- Brent Busby wrote:
-
- <snip>
-
- > (Part of the reason I want this is I'd like to find the answer to the
- > age old question, by now certainly in the FAQ for comp.sys.amiga.emulations
- > if there is one, does the fact that Enforcer is needed for Janus really
- > mean that there is something illegal happening at the machine code level
- > whenever the bridgeboard side and the native side talk to eachother
- > through the shared memory communications port, or is it all completely
- > safe to do things like AREADs and AWRITEs as often as you want, even
- > requiring Enforcer as it does? I'd like to know...)
-
- The reason enforcer is needed is because the ZorroII expansion space is
- considered as cacheable by default and has no cache control signals. What
- enforcer does is inhibit the cache for the memory area used by the
- BridgeBoard.
-
- This memory area can not be cached because it is a dual-port memory. The
- 680x0 (its cache controller to be precise) has no way to know that the memory
- has been modified by the PC side and thus could -and quickly would- hold
- obsolete data into its cache leading the software to be "confused".
-
- - Remi.
-