Update to fix flash and other related bugs

[lalalandrus]

As per issue 10 comments

[p-h-a-i-l]

see #10

[btsimonh]

looks good to me without building and testing.
Qs:
1/ does myFlashSection (RESERVED_FLASH) exist in the flashable image? (we don't really want it to...)
2/ would it be better to, instead of
volatile uint8_t *flash_data = (uint8_t *) 0x08000000 + 0x40000 - 0x800;
use .myFlashSection -
or would this force it to be inside the flashable image?

[lalalandrus]

1. the way the ld is set so that there is that I will exist in the flash image if this is undesirable i think you can remove keep from the mysection in the linker script and include flasharea.c through a define only if you want to format the flash. This way the petitions still exist so there is no chance of conflicr but it won't be in the firmware file either.

2. you can use any region you want. We should have both the linker and flash access use the same defines. If you use a region in the middle I don't see any benefit at the moment as it reserves a chunk in the middle of the flash which could cause future issues. You can grow sections from either side and work towards the middle this way. Tail end for storage and head end for code space.

[btsimonh]

yep, see if you can remove it from the flash image. The flash should initialise if the magic is not valid, but I moved from an area inside the flash image to one outside to try to keep existing flash values through firmware updates. We can scrap 'flasharea.c' completely - I don't see it as relevant any more?

[lalalandrus]

You could change #if def Flash_storage to soemthinglike flash_init and keep the file for formatting the flash. Don't know if anyone sees value in this.

[btsimonh]

you can always init the flash via serial.... (if it does not detect that it needs to init because magic is wrong)?

[lalalandrus]

yes i just confimred it works. remove flasharea.c and remove the keep before the .mysections. in the future you can use someting like flasharea.c to define some initial good settings

[btsimonh]

great - add the commit, and i'll see if i can test tonight.

[btsimonh]

haha... built 'env hoverboard'; did not notice the firmware.bin was 256k, ran my flash.sh, and it failed.
Now I'll have to hold my power button until the Rpizero has booted, and operate all the logging in and flashing with the other hand! serves me right for not being vigilent....

[lalalandrus]

Make sure you change the ld file too. I fried my St link so we both had some setbacks.

[btsimonh]

just flashed back to @p-h-a-i-l 's 'size' stuff and it's back. will try build again.

[btsimonh]

are you building the hoverboard env in pio?

[btsimonh]

ahh... did you not push the remove of keep?

[lalalandrus]

Miscommunication I thought you would add that commit. In fact i dont seem to know how to edit pull requests

[lalalandrus]

oh i think it worked, if not the code is commited to my fork

[btsimonh]

ok, can confirm hoverboard operation with this combined with @p-h-a-i-l 's latest.
Now need to check protocol.
protocol seems ok :).
need to check flash....
flash write confirmed with 'fsl300\n', 'fm1234\n', reset, 'fa\n'.