blitter, CPU cache, NOVA, TOS

From: Petr Stehlik <>
Date: 7 Feb 99 11:28:42 +0000


I am still working on the NOVA routines integrated into TOS. I have just
written a (very) simple replacement for Bconout that does not use
blitter and so it works OK on the NOVA.

Now I would need to fix the last thing - XBoot. I donīt know how XBoot
does its output - it doesnīt use AES but a well written AES replacement
(XBoot is cool anyway since it contains something like N.AES in itself
:-) but as for the actual output on the screen it might use VDI (I heard
it was possible to initialize VDI from AUTO folder).

In the XBoot the worst is the text output (icons and bitmaps are nearly
OK, the mouse pointer is 100% OK). I wonder how VDI does its text output
(I wouldnīt like to disassemble VDI so thatīs why I am asking here).
Disabling Textblt routine of LineA didnīt help - what else the VDI can

Also, Dougīs Toolkit disables all 68030 cache instructions. At the end
of all Blitterīs routines thereīs an instruction that clears the data
cache of 68030. As I said this cache clearing is disabled in patched TOS
and I wonder if this could cause that all output on the NOVA is messed
and all odd (or even?) columns are not blitted correctly.

What is the 68040 analogy for setting the CD bit of CACR register of
68030 CPU? Cache flushing, or disabling or what? Sorry I have a very
poor knowledge of 68040 :-( A pointer to a good RTFM would be

E-mail:                        PARCP developer
   WWW:              MiNTOS/Linux user
mirror:                 Atari 800XL emulation
mirror:              Atari Falcon040
Received on sÃļ. feb. 07 1999 - 12:28:42 CET

This archive was generated by hypermail 2.3.0 : ti. nov. 03 2015 - 20:07:54 CET