No, Most sectors are 4 pages (64KB), except for $7C-$7D, which is one, and $7E (certificate page) is two 8KB sectors. Boot code is its own. They're laid out at the beginning of the data sheet, btw
(Note: translate pages for 84+ and 83+, the non-SE's)
The extra $4000 is just to get it onto the flash page. I could have added $5000, $6000, or $7000 if I wanted. (most significant hex digit doesn't count)

Wait, how many bytes is a sector? If I wanted to erase page 7E (god forbid), how?

You have to erase both sectors of it. first addresses starting with $4000, the other set with $6000. My code would have to be modified.

So an ErasePage routine would have to be adaptable for different pages?

Nah, only the certificate page is special everything else works normally since the other sectors are at least a page in size.

At least a page? So are some sectors multiple pages in length?

Most sectors are 4 pages. One sector, $7C-$7D, is two, and $7F is its own sector.

Very nice, thanks for clarifying that, although I could have saved myself the stupid question if I had read the other Flash topic first.

Ameh. All the raw flash control code you could ever need is here (except the full-chip erase). By the way, the documented autoselect commands don't work. This program helpfully provides code that proves this.

Very impressive! Is that your own work, DrDnar?

Yeah. You'll recognize the style, especially the menu routine, from other software I've released on ticalc.org. I never uploaded because I don't figure it's really a useful program. Not that that stops other people.

I'd say that not only is it somewhat useful on its own, the routines it contains are extremely valuable, so I think you should at least release it in the Routines section! You obviously put plenty of hard work into this, and you deserve more recognition (not even your name in the program?) for it.

Hm... that code and your username look familiar...
Oh, that's right, I've seen you on UTI.
I can't believe I didn't remember this earlier; I could have saved myself some trouble by just linking to your posts

Hehe, and we return to the reinventing-the-wheel lessons that I was helping Sonlen to learn. Still, I'm glad there's a bunch of information in this topic now from different people that all agrees.

I agree with KermMartian. This is hard stuff, and not well documented, until this thread.

As Brandon likes to point out, there's no significant advantage to writing your own flash manipulation routines, perhaps aside from being able to run other code while the flash chip is busy. But you wouldn't be able to do much, because you can't read from flash while a write/erase operation is in progress.

Or perhaps it could be required if OP1 can't be corrupted...

I'm most interested in independence from the TI-OS.

Okay, so the ErasePage routine isn't working, ideas?

Code:

Did you try comparing your routine with Dr. Dnar's routine?