gnupic@linuxhacker.org

On Tue, 14 Oct 2003, Alexandre Pereira Nunes wrote:

> That's great news (at least for me)! :-)

Good!

> 
> But this behaviour with the erase command is curious ... Here it (i.e. 
> -ef command) takes a couple of seconds and return ok. In fact, the 
> reply  from the picstart to the erase operation is almost instantaneous, 
> the delay is due to the verify command issued afterwards, which is 
> perceptible on devices with some amount of memory, like the 18f452.


I made a mistake. I essentialy issued an -ep instead of an -ef.


> What firmware version are you using on the ps+?

version 3.11.00

> 
> Thanks for your report, I hope to release a new version of the patch 
> next week, if I find the time to implement some few missing things.

You may wish to consider this:

It appears that the configuration fuses are read from the hex file, but 
only the first word is getting programmed. At least when I attempt to 
program the CP bits there's no effect. I glanced briefly at the code and 
it appears that DoWritePgm only checks the case when an address is equal 
to the start of configuration memory and *not* when it is in the range of 
configuration memory. 

However, the command line option for writing the configuration bits looks
like it'll write more of the configuration bits. OTOH, it still will not
write all of them. picdev.c says that the configuration space for the
18f452 is only 7 bytes. That's not true. There are 34 configuration bits
spanning 13 configuration addresses. Perhaps the '7' refers to the number
of 16-bit words and not 8-bit bytes? In any case, a -wc 0x0 only will 
clear the first 7 or so configuration addresses.

Scott