Commit Graph

3 Commits

Author SHA1 Message Date
Paul XPS 7584bbeb70 Big update from past weeks' work. Most of the work involves converting
NES scripts to use new dump/flash methodology that MMC3 started.
Includes BNROM, UNROM, MMC1, and new scripts for FME7 & MMC4 (SOP flash).
Adding more general support to SNES with v2proto_hirom that script is
actually becoming more of a master script supporting both LoROM and
HiROM including flash, dump, and save backups.
SNES Rd/Wr now designate the state of /ROMSEL so have to manually
determine if access should be in /ROMSEL space of the SNES memory map or
not.  (ie all SNES cart memories are /ROMSEL space except HiROM SRAM).
2018-11-19 18:00:03 -06:00
Paul XPS 86e8d3d215 Big old first major commit since publicly releasing..
I made a commit earlier this week but messed things up by not pulling
from the master and things weren't updated...  Here's the notes from
that commit:

 Author: Paul XPS <paul@infiniteneslives.com>
 Date:   Tue Nov 6 22:45:52 2018 -0600

 Large commit biggest feature add is NES MMC3 support including Save RAM
 (aka WRAM/PRGRAM) support including dumping and writing save files.

 The MMC3 script & method of dumping/flashing is the most forward
 thinking script/firmware so far.  Finally starting to form a clear
 vision of how I want to handle flashing/dumping variations with mappers.
 Biggest thing is having the host handle the mapper init & banking
 control.  In the view of the firmware, it's only responsible for knowing
 how to flash a bank.  And dumping is even more generic with the host
 just telling what address range to read.  Things will get more complex
 with support of mappers with bus conflicts.  But ready to start
 converting these old hacked methods to be more like the MMC3 means.

 Have some early support for dumping gameboy using the snes script as the
 pinouts are nearly identical.  Along with testing of toggling between 3v
 GBA and 5v DMG.

 Have some early support for INLretro NES only version which uses a
 smaller mcu because it doesn't need to support large 16bit carts.  Still
 have to get this completed.

 Added support for CNROM, but I'm not sure if it's actually working.  Going
 to restart with NROM and start updating the currently supported
 mappers to be more like MMC3.

But this also includes some new updates from the second half of the week:

Started updating existing NES scripts to use new MMC3 methodology.  Got
NROM, CNROM, BNROM, & Color Dreams working.  On the host side only
needed to add nes.c functions for specific flash algos.  Able to delete
significant amounts of mapper specific code from flash.c

Got some basic SNES script support with new methodology for Catskull
elect 5v PLCC SNES LoROM board.  And INL SNES HI/LO-ROM 3v board as
well.  These don't yet use buffer writes, just single byte writes.

Also having issues with Mirroring test/sensing again.  Driving me crazy,
but don't really care about it at the moment and not sure what's wrong..
So just committing that broken for now.  Looking to remove this
functionality from the firmware side as the host should be controlling
most of this.

Looking to add SNES RAM & buffered writes.  Also need to test some of
the HIROM code as I just added it in there while I did the LOROM stuff..
2018-11-09 21:52:33 -06:00
Paul Black ASUS win7 9c57f1bdb3 Massively overdue commit of several months worth of random work.
Mostly adding support for mappers as I needed it for my own hardware
builds:
-MMC1
-mapper 30
-easy NSF (still need to update for mapper verilog fix)
-action53 (still need to update for mapper verilog fix)
-dual port board flashing
-colordreams, not sure if I actually got this working
-color ninja, just a special CPLD version of colordreams for ninja boards

Just started working on SNES code.  slowly getting things up and working
outside of main inlretro.lua script similar to how NES has been handling
everything with it's own script.  Able to flash v3 boards fine.  v1 boards
flash without errors, but still having some mapping problems where it
verifies but won't boot.  v2 prototype flashes most bytes but not all,
seems v2 boards are much slower to output valid data..  But that may just
be the manufacturer ID codes..?

TODO next:
-bootloader dictionary that jumps to bootloader so don't have to manually
close jumper on the board.
-turn on the watchdog timer for stm32
-create some sort of host timeout so reset button on programmer isn't as
useful
-allow firmware programing algos to be uploaded and executed from SRAM for
faster code that also doesn't require specific firmware builds to support
new mapers.
-Finish JTAG to simplify programing NES & SNES CPLDs
-Sort out swim issue with stm8s001 CICs
-add SWIM support for avr
2018-07-08 20:23:44 -05:00