Re: gEDA-user: C question

[Hans Nieuwenhuis <vzzbx@xxxxxxxxx>]

On Sun, 4 Mar 2007 13:58:45 -0700
Andy Peters <devel@xxxxxxxxx> wrote:

> On Mar 3, 2007, at 4:57 PM, carzrgr8@xxxxxxxxxxxxx wrote:
> 
> > Thanks.  I threw the question to the group because I know there's  
> > plenty of knowledgable people reading willing to answer it.
> >
> > Anyway, I was trying to get the compiler to generate 2 versions of  
> > an initialized array: ROM, and RAM.  The processor I am using has  
> > more ROM than RAM, so
> > placing the code into ROM makes more sense, yet I was having  
> > trouble getting access it correctly due to a routine expecting a  
> > const ptr.  So, I was trying it in
> > RAM, but not sure if the code was correct.
> >
> > Thanks again for the help :)
> 
> You didn't say which compiler you're using, nor which architecture.
> 
> But Keil's 8051 compiler has the "code" keyword, which tells the  
> compiler that the variable should be put into CODE space, which is by  
> definition ROM.
> 

For gcc ports this keyword would be: __attribute__((section
("YOURSECTION"))) where YOURSECTION must be replaced with .text for code
in ROM. 
GCC ports are not consistent though. If I develop for ARM and declare
avariable as const it will end up in RAM also unless I add the
__attribute__... keyword. However the MSP430 port places the variable in
ROM if I only declare it as const...
Maybe the reason for this behaviour is the difference in complexity in
ARM and MSP430 micros...

HTH,

Hans



> -a
> 
> 
> _______________________________________________
> geda-user mailing list
> geda-user@xxxxxxxxxxxxxx
> http://www.seul.org/cgi-bin/mailman/listinfo/geda-user
> 


-- 

$ cat .sig /dev/null


_______________________________________________
geda-user mailing list
geda-user@xxxxxxxxxxxxxx
http://www.seul.org/cgi-bin/mailman/listinfo/geda-user