I think it was happening due to compiling the interrupt routines into a
library object and then linking  the library with main.  Ill test more when
I get back to the office.

On Mar 23, 2010 4:21 AM, "JMGross" <[email protected]> wrote:


That's an interesting observation. I never encountered this as I never had
an interrupt function on its own in a separate compilation unit.
Yes, there are no references, even the entry for the vector table (which is
referenceing the interrupt routine, of course) is local and therefore does
not force the linker to link the file.
Perhaps if one would generate weak references to the dummy ISR for all
interrupt vectors, the linker would detect that there is something
overriding these weak references an dlink. But only if the vecto table with
these
weak references is in a separate file other than the dummy ISR, or lese the
references woul dbe resolved locally and again no external reference would
be created.

Anyway, I usually put the ISRs together with the related function. The timer
ISR together with the timer init function, the UART ISRs together with the
code stuffing data into the ringbuffer etc. This way, if none of the
UART functions is used, none of the code is linked, but if, all of the code
gets linked.


JMGross

----- Ursprüngliche Nachricht -----
Von: Anthony Asterisk
An: [email protected]...
Gesendet am: 17 Mrz 2010 21:08:26
Betreff: Re: [Mspgcc-users] mspgcc4 usci.h error in definition of UCBxCTL0/1


Will these patches being going into GCC cvs?

BTW, other than these minor header file issues I have...

------------------------------------------------------------------------------
Download Intel&#174; ...

Reply via email to