[Icc-mot] Re: Icc-mot Digest, Vol 35, Issue 5

[Wade A Smith]

On Tue, 24 Jul 2007 12:12:44 -0700 (PDT) icc-mot-request at imagecraft.com
writes:
> From: "Edward Karpicz" <ekarpicz at freemail.lt>
> Subject: Re: [Icc-mot] function prototype for specific address

<snip> 

> Do you mean *flash* EEPROM programming? Yes, interrupts should be 
> disabled 
> for flash program/erase time. That's because flash is not readable 
> while 
> programming/erasing; and since it's not readable -  if interrupt 
> would occur 
> while programming, then CPU couldn't fetch valid vector from flash 
> and 
> execute valid interrupt service code.
> Word program time is <75us. You could disable interrupts for word 
> program 
> time and still have quite high rate SCI interrupts working. But 
> setor erase 
> time is indeed really long. If really necessary then RAM could be 
> remapped 
> to vector area, vectors and ISR's should be moved to RAM. Do you 
> really need 
> interrupts always enabled?
> 
> Edward

While we do a sector erase, we have messages coming in (constantly at
4800BAUD, every 2mS) so while doing a sector erase, we could miss the
first few characters of a message (the first indicates Start-of-message,
second indicates bytes to follow and third indicates if it is for us or
not).  So, yes we need to catch every character coming in.

I dread the idea of relocating RAM to 0xF000 because then we would lose
access to our monitor and we would have to include some code to handle
some things that we now pass off to the monitor.

Would you have sample code?

wade a smith
Florida