I am writing data to the internal flash of a STM32F746 and I'm wondering what the difference between pages and sectors is? And what is the purpose of using sectors instead ? Notice that I'm very new and the reason to why I'm asking is because I have previously written to flash knowing the page size and amount of pages. But now I have 8 sectors instead.
Unless I have missed it somewhere I can't find any reference to pages or page size in any any document to the STM32F746
The concept is the same, between page and sectors. STM32F7 documentation uses the sector term because they are much bigger than the typical pages in external flash memory. But it doesn't matter.
So, you can't find reference to pages because what you called pages are here called sectors.
The constraint is: when you want to erase, you have to erase the whole sector. As you have seen, on STM32F7, there are 8 sectors, with different sizes (4 sectors of 32 Kbytes, 1 sector of 128 Kbytes, and 3 sectors of 256 Kbytes). The fact they have different sizes allow you to have more flexibility in the way you organize your flash. For example, you could have a sector reserved for configuration data. This way, when you want to reprogram this part, you don't have to erase the whole firmware code. But you still have to erase the whole configuration sector, even if you need to change just a single byte of it. And it is true these sectors are huge, so you can't use this flash as easily as typical external serial flashes on which you could organize things in a more flexible way (or even use a filesystem). But this is not the purpose of these embedded flash memories.
If you need a more flexible way to store persistent data that can change over time, use an external flash, or external EEPROM, depending on you constraints. There is, however, a trick you could use on the main MCU flash: the erase is required only when you switch bits from 0 to 1. To switch bits from 1 to 0, you just need to program the flash bytes (see the note in chapter 3.3.7 in the ref manual). Programming the flash bytes can be made as many times as you want on the sector, and can be made on individual bytes. So you could take advantage of this by dividing the huge "configuration" sector in small chunks that you program individually, when your data changes. Then, you need an index in which you "fuse" the bits from 1 to 0 to know in which chunk the latest data can be found. Finally, when all the chunks have been used, you erase the whole sector and start over. This wastes a lot of space, but may give you the flexibility you need without compromising too much on flash endurance.
