INL-retro-progdump/firmware
Paul XPS e4debe5a13 Big update for firmware version v2.3.1
this is the verion getting flashed on all v2.0N NESmaker kits
v2.3.0 worked for basic functions, but was never shipped

Majority of effort revolved around testing mapper30 boards with the
smaller v2.0N INLretro with the NES connector alone for NESmaker kits.

added linear feedback shift register for test stream data generated
locally on the device.  I'm not 100% sure if this is any faster than
pushing the actual data via USB though.. :/  It's plenty fast on the
stm32 nearly instantaneous for 32KByte.  But the AVR takes a couple
sec..

Created "stuff" dictionary for things like that were I just want to add
small things and don't want to bother with a whole new dictionary.

Added file verification to the host with files.lua

Have some nes flash algos return post-written data so calling function
can decide if want to retry, fail, etc.

Changed host dictionary calls to assert instead of error because it
really shouldn't continue.  I didn't see an error when sending opcode to
wrong dict and caused head banging..

fwupdate permits bytes to be skipped, or force the update.  Found that
the fwupdater got assigned different addresses of ram depending on what
all other ram gets allocated to the main application

Some clean up of inlretro.lua

TODO:
host learn and keep track of the connected device.
Needed for ciccom right now, or knowing whether ciccom connection
is even present..
In the end maybe ciccom is better placed in firmware, but for small
transfers of only a few bytes it kinda makes sense to keep on the host.
Pinport gets quite messy with these made up pin names when really all I
want to do is toggle a specific pin on the NES connector.  So maybe some
double mappings would actually be okay, need to rethink that..

create different flash modes that either keep going, retry, or error
depending on the goal of the flash operation.  Fanout the return value
from flash algos to all of them.

have fwupdate assigned a specific area of ram so the ram pointer doesn't
change between builds.  Okay to ignore for now.

Realized can have STM32F070C6 devices execute bootloader by erasing all
the flash or perhaps even just the first word of flash according at
AN2606.  This wouldn't work for RB devices though.  This could be done
through the bootloader dict
2018-12-06 01:07:13 -06:00
..
DFU_release Big update for firmware version v2.3.1 2018-12-06 01:07:13 -06:00
build_avr Big update for firmware version v2.3.1 2018-12-06 01:07:13 -06:00
build_stm6 Big update for firmware version v2.3.1 2018-12-06 01:07:13 -06:00
build_stmn Big update for firmware version v2.3.1 2018-12-06 01:07:13 -06:00
include_stm Big release commit for firmware v2.3.0 2018-12-01 02:18:01 -06:00
release Big update from past weeks' work. Most of the work involves converting 2018-11-19 18:00:03 -06:00
source Big update for firmware version v2.3.1 2018-12-06 01:07:13 -06:00
source_stm_only Big release commit for firmware v2.3.0 2018-12-01 02:18:01 -06:00
usbdrv_Vusb Final commit for firmware version 2.01 first public release 2018-09-07 09:13:39 -05:00
Make_avr Big release commit for firmware v2.3.0 2018-12-01 02:18:01 -06:00
Make_stm_adapter Firmware updates are complete for switchless firmware updates via USB! 2018-11-29 22:31:26 -06:00
Make_stm_inl6 Big update for firmware version v2.3.1 2018-12-06 01:07:13 -06:00
Make_stm_inl6p Firmware updates are complete for switchless firmware updates via USB! 2018-11-29 22:31:26 -06:00
Make_stm_nes Big update for firmware version v2.3.1 2018-12-06 01:07:13 -06:00
Makefile Big old first major commit since publicly releasing.. 2018-11-09 21:52:33 -06:00