On Fri, 11 Sep 2009 21:34:17 -0500, you wrote:

>Does the volatile keyword for a variable (using C30 PIC24) guarantee an 
>atomic operation when accessed regarding interrupts?  i.e. if I declare 
>a volatile DWORD elapsed_time and increment it in the timer interrupt, 
>do I need to worry about disabling interrupts to access it in the main 
>loop?  I always do, just wondering if I could live without it, and can't 
>think of a simple test to prove it.

No - Volatile simply tells the compiler that a variable's value can change outside of the normal
program flow, and so should not, for example, be optimised into a register ( not really applicable
to PICs in practice). Atomicity  is a completely seperate issue as far as the compiler is concerned

To guarantee an atomic read in the foreground you do need to either disable  the interrupt, set a
flag to tell the int code not to touch it, synchronise the read with the int process, or re-read
values to check for changes.

e.g. read lsbyte
read msbyte(s)
read lsbyte again and compare to first read - if different read msbyte(s) again, 

The latter method makes the (usually reasonable)  assumptoin that only a single int can occur during
the read process. 



-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist