Path: utzoo!yunexus!torsqnt!dptcdc!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!bloom-beacon!apple!oliveb!sun!pepper!cmcmanis
From: cmcmanis%pepper@Sun.COM (Chuck McManis)
Newsgroups: comp.sys.amiga.tech
Subject: Re: Introduction to comp.sys.amiga.tech
Message-ID: <92970@sun.uucp>
Date: 8 Mar 89 17:14:07 GMT
Article-I.D.: sun.92970
References: <1739@psu-cs.UUCP>
Sender: news@sun.uucp
Reply-To: cmcmanis@sun.UUCP (Chuck McManis)
Organization: Sun Microsystems, Mountain View
Lines: 23

In article <1739@psu-cs.UUCP> bartonr@psu-cs.UUCP (Robert Barton) writes:
>  This is a new one to me.  When the OpenWindow() routine allocates the
>memory for the Window structure it asks for MEMF_CLEAR | MEMF_PUBLIC.  Same
>with OpenScreen() when allocating a Screen or RasInfo structure.  Since a
>Screen contains a ViewPort, a RastPort, a BitMap, and a LayerInfo this
>implies that none of these structures need to be in chip memory either.

You are correct, the custom chips (blitter, dma channels) never "see" 
the various structures, they see the data. That is why you can allocate
a RastPort from anywhere but the actual data areas that the PlanePtrs
of the BitMap structure point to must be in chip memory. Same is true
with Audio device or Trackdisk device requests, the IORequest struct
can be in any kind of memory but the data buffers must be in chip memory.

The only place this gets kinda weird are things like the Image data in 
an Intuition Image structure. This my not appear at first as data for
a custom chip, however Intuition does indeed point the blitter at it
and blit so it too has to be in chip memory. 


--Chuck McManis
uucp: {anywhere}!sun!cmcmanis   BIX: cmcmanis  ARPAnet: cmcmanis@sun.com
These opinions are my own and no one elses, but you knew that didn't you.