Lpc2000

Hi,

I'm trying to enable interrupt with timer1 on the LPC2106. I 
configured Timer1 as IRQ source when its value correspond with the 
match register 0... but interrupt routine service doesn't run and is 
never reached. Here is a piece of my code:

  __disable_interrupt(); // IAR intrinsic
//VIC enable registers
  VICProtection=0x0;    

//clears bits in the Interrupt Enable register  
  VICIntEnClear=0xffffffff;   

  VICIntEnable=0x20;    //Timer1
  VICIntSelect=0x00;    //all interrupts are IRQs
  VICVectCntl0=0x25;    
  VICVectAddr0=(unsigned long)&Timer1ISR;

  __enable_intrinsic();

  Then I also initialize Timer1:

  T1_IR=0x00;
  T1_MR0=TIMERMATCH;
  T1_MCR=0x03;  // reset timer and fire interrupt        
  T1_PR=0x0;         
  T1_TC=0x00;   
  T1_TCR=0x01;  // start the timer

__arm __irq void Timer1ISR(void) //Timer1 ISR
{
T1_IR=0x01; //reset interrupt flag
VICVectAddr=0; //reset VIC
return;
}

My questions are:

- what is wrong with this code?
- could it be a problem with the intrinsic functions of IAR    
__enable_interrupt() and __disable_interrupt(); they seem not to 
work; does anybody their asm code??

Any help would be appreciated!

thanx

I'm curious about 2 things.

1. "__enable_intrinsic()"
   Is this really the function you used or do you really mean 
__enable_interrupt().

2. Where is your IRQ Handler? The only time I used "_irq" was to 
define the IRQ handler...not the ISR.

something like...

__arm __irq void IRQ_handler(void)
{
 ISRAddress = VICVectAddr;     //load source address
 (*ISRAddress)()               //function call
 VICVectAddr=0; //reset VIC
 return;
}

void Timer1ISR(void) //Timer1 ISR
{
   .....
}


That way ISRs get called by the handler whenever any enabled 
interrupt channel is activated. 

Some example code should come with your IAR compiler...somewhere in 
the /ARM/src directory. Make sure u look at those. They'll come in 
very handy.

Leighton

--- In lpc2000@yahoogroups.com, "ducatman2000" <545545@b...> wrote:
> Hi,
> 
> I'm trying to enable interrupt with timer1 on the LPC2106. I 
> configured Timer1 as IRQ source when its value correspond with the 
> match register 0... but interrupt routine service doesn't run and 
is 

> never reached. Here is a piece of my code:
> 
>   __disable_interrupt(); // IAR intrinsic
> //VIC enable registers
>   VICProtection=0x0;    
> 
> //clears bits in the Interrupt Enable register  
>   VICIntEnClear=0xffffffff;   
> 
>   VICIntEnable=0x20;    //Timer1
>   VICIntSelect=0x00;    //all interrupts are IRQs
>   VICVectCntl0=0x25;    
>   VICVectAddr0=(unsigned long)&Timer1ISR;
> 
>   __enable_intrinsic();
> 
>   Then I also initialize Timer1:
> 
>   T1_IR=0x00;
>   T1_MR0=TIMERMATCH;
>   T1_MCR=0x03;  // reset timer and fire interrupt        
>   T1_PR=0x0;         
>   T1_TC=0x00;   
>   T1_TCR=0x01;  // start the timer
> 
> __arm __irq void Timer1ISR(void) //Timer1 ISR
> {
> T1_IR=0x01; //reset interrupt flag
> VICVectAddr=0; //reset VIC
> return;
> }
> 
> My questions are:
> 
> - what is wrong with this code?
> - could it be a problem with the intrinsic functions of IAR    
> __enable_interrupt() and __disable_interrupt(); they seem not to 
> work; does anybody their asm code??
> 
> Any help would be appreciated!
> 
> thanx

Can anyone steer me towards some simple tone generators that can easily be
interfaced to the lpc using gpio (as few pins as possible) or i2c?  Nothing
fancy, we need to control frequency and duration (volume will be controlled
manually if at all).

--- In lpc2000@yahoogroups.com, "Curt Powell" <curt.powell@s...> 
wrote:
> Can anyone steer me towards some simple tone generators that can 
easily be
> interfaced to the lpc using gpio (as few pins as possible) or i2c?  
Nothing
> fancy, we need to control frequency and duration (volume will be 
controlled
> manually if at all).

Take a look at using a PWM output.

You are right, I made a mistake writing my previous mail, it was not 
__enable_intrinsic() but of course __enable_interrupt().

Now I have introduced the IRQ Handler exactly as you told me. In that 
way, the Timer1 ISR should be called by the irq_handler whenever an 
interrupt occurs. BUT, it sill not work! In debug mode I can see the 
timer incrementing and when it reaches the match value register, it 
resets and put the IR1 bit to '1' but the ISR is never reached... I 
think there is a problem with the IAR intrinsic __enable_interrupt() 
and __disable_interrupt(). In step by step mode, when the 
__enable_interrupt() is executed, that blocks the programm... Any 
idea of the problem, or do you know the asm code of __enable_interrupt
() ans __disable_interrupt ?

(the code follow at the end of this mail)

Thank you very much !

--- In lpc2000@yahoogroups.com, "Leighton Rowe" <leightonsrowe@y...> 
wrote:
> I'm curious about 2 things.
> 
> 1. "__enable_intrinsic()"
>    Is this really the function you used or do you really mean 
> __enable_interrupt().
> 
> 2. Where is your IRQ Handler? The only time I used "_irq" was to 
> define the IRQ handler...not the ISR.
> 
> something like...
> 
> __arm __irq void IRQ_handler(void)
> {
>  ISRAddress = VICVectAddr;     //load source address
>  (*ISRAddress)()               //function call
>  VICVectAddr=0; //reset VIC
>  return;
> }
> 
> void Timer1ISR(void) //Timer1 ISR
> {
>    .....
> }
> 
> 
> That way ISRs get called by the handler whenever any enabled 
> interrupt channel is activated. 
> 
> Some example code should come with your IAR compiler...somewhere in 
> the /ARM/src directory. Make sure u look at those. They'll come in 
> very handy.
> 
> Leighton

Here is a piece of my code:

> > 
> >   __disable_interrupt(); // IAR intrinsic
> >   VICProtection=0x0;    
> > 
> >   VICIntEnClear=0xffffffff;   
> > 
> >   VICIntEnable=0x20;    //Timer1
> >   VICIntSelect=0x00;    //all interrupts are IRQs
> >   VICVectCntl0=0x25;    
> >   VICVectAddr0=(unsigned long)&Timer1ISR;
> >   __enable_interrupt();
> > 
> >   Then I also initialize Timer1:
> > 
> >   T1_IR=0x00;
> >   T1_MR0=TIMERMATCH;
> >   T1_MCR=0x03;  // reset timer and fire interrupt        
> >   T1_PR=0x0;         
> >   T1_TC=0x00;   
> >   T1_TCR=0x01;  // start the timer
> > 
> > __arm __irq void irq_handler(void) //Timer1 ISR
> > {
> >   ISRAddress = VICVectAddr;     //load source address
> >   (*ISRAddress)()               //function call
> >   VICVectAddr=0; //reset VIC
> >   return;
> > }
> >
> > void Timer1ISR()
> > {
> >  IR1 = 0x01;
> > }
> >

--- In lpc2000@yahoogroups.com, "ducatman2000" <545545@b...> wrote:
> You are right, I made a mistake writing my previous mail, it was 
not 
> __enable_intrinsic() but of course __enable_interrupt().
> 
> Now I have introduced the IRQ Handler exactly as you told me. In 
that 
> way, the Timer1 ISR should be called by the irq_handler whenever 
an 
> interrupt occurs. BUT, it sill not work! In debug mode I can see 
the 
> timer incrementing and when it reaches the match value register, 
it 
> resets and put the IR1 bit to '1' but the ISR is never reached... 
I 
> think there is a problem with the IAR intrinsic __enable_interrupt
() and __disable_interrupt().
> In step by step mode, when the 
> __enable_interrupt() is executed, that blocks the programm... 

I remember this happening to me before, but at the time I wasn't 
sure why. I'm still learning the IAR library functions. Do you have 
the vectors in cstartup configured?

org     0x00
pc,=?cstartup
....
org	0x18
ldr	pc,=irq_handler
...

If u did, then check the status of cpsr after both intrinsic 
enable/disable functions. If the I & F bits (global IRQ & FIQ flags) 
don't change correctly, then those 2 library functions may very well 
be the fault.

> do you know the asm code of __enable_interrupt () ans 
__disable_interrupt ?
> 
> (the code follow at the end of this mail)

...that you can do for enable/disable with 3 asm instructions.

read cpsr --> register
modify I F bits 
store register --> cpsr

As long as you configure your processor to the required priviledge 
level (see ?cstartup examples & the ARM7TDMIS Documentation), you 
can easily modify the I & F bits of the cpsr register to 
enable/disable IRQ and FIQ global interrupts. 

Good luck!
Leighton

Has anyone ever received a conclusive reply from Philips concerning the 
reason for the slow GPIO? I just would like to know why this port is so 
slow, I am curious. Initially I thought it had to do with the set/reset 
logic and the goal of  synchronizing the output port. But the port is 
still slow even if I write directly to the IOPIN or IODIR registers.

The sequence below where r6 and r4 are loaded with bit patterns while r0 
holds the IOPIN address runs 7 times slower than the CPU clock with a 
VPBDIV=1.

    str    r6,[r0]
    str    r4,[r0]       
    str    r6,[r0]
    str    r4,[r0]       
    str    r6,[r0]
    str    r4,[r0]       


Something that I miss from the LPC21xx datasheets is a good old logic 
diagram for the GPIO which most micro datasheets normally feature.

--
Peter Jakacki

--- In lpc2000@yahoogroups.com, Peter Jakacki <peterjak@t...> wrote:
> Has anyone ever received a conclusive reply from Philips concerning 
the 
> reason for the slow GPIO? I just would like to know why this port is 
so 
> slow, I am curious.

Peter

This is not conclusive stuff from Philips, but the main reason I can 
see for the problem is that the GPIO is a peripheral hanging off the 
relatively slow peripherla bus. The major time seems to be taken up by 
the sychronisation between the core and this bus.

If the bus had been designed more like an external memory bus then 
we'd probably see full-speed (or at least much faster) GPIOs.

-- CHarles