6747 GPIO's

I'm setting it like this in Main() , Am I doing it wrong?

    PINMUX8 = PINMUX8 | 0x8000;//bit 12-15 = 0x08
    PINMUX13 = PINMUX13 | 0x8000000;//bit 24-27=0x08

I have the numbers backwards, it should be

    PINMUX8 = PINMUX8 | 0x8000000;//
    PINMUX13 = PINMUX13 |  0x8000 ;//

but it still does not work.

This comes from the Ch 2.9 of the GPIO Guide:

The following steps are required to configure the GPIO module after a hardware reset:

1. Perform the necessary device pin multiplexing setup (see your device-specific data manual).
2. Program the Power and Sleep Controller (PSC) to enable the GPIO module. For details on the PSC, see see your device-specific System Reference Guide.
3. Program the direction, data, and interrupt control registers to set the configuration of the desired GPIO pins (described in this document).

The GPIO module is now ready to perform data transactions.

No I have not configured that. I assumed it was already configured since GPIOS 12-15 work. I'm reading sprufk4c now trying to figure out how to configure it but it is very confusing.

Is it bits 0-2 of register MDCTL3 I have to set? If so is the register at 0xA12?

In the Gel File there is this:

etup_Psc_All_On( )
{
    int i;
    GEL_TextOut( "Setup Power Modules (All on)... " );

    // PSC0
    PSC0_lPSC_enable(0, 0);
    PSC0_lPSC_enable(0, 1);
    PSC0_lPSC_enable(0, 2);
    PSC0_lPSC_enable(0, 3);  // EMIFA
    PSC0_lPSC_enable(0, 4);
    PSC0_lPSC_enable(0, 5);
    PSC0_lPSC_enable(0, 6);
    PSC0_lPSC_enable(0, 8);
    PSC0_lPSC_enable(0, 9);
    PSC0_lPSC_enable(0, 10);
    PSC0_lPSC_enable(0, 11);
    PSC0_lPSC_enable(0, 12);
    PSC0_lPSC_enable(0, 13);

    // PSC1
    PSC1_lPSC_enable(0, 1);
    PSC1_lPSC_enable(0, 2);
    PSC1_lPSC_enable(0, 3);
    PSC1_lPSC_enable(0, 4);
    PSC1_lPSC_enable(0, 5);
    PSC1_lPSC_enable(0, 6);  // EMIFB
    PSC1_lPSC_enable(0, 7);
    PSC1_lPSC_enable(0, 8);
    PSC1_lPSC_enable(0, 9);
    PSC1_lPSC_enable(0, 10);
    PSC1_lPSC_enable(0, 11);
    PSC1_lPSC_enable(0, 12);
    PSC1_lPSC_enable(0, 13);
    PSC1_lPSC_enable(0, 16);
    PSC1_lPSC_enable(0, 17);
    PSC1_lPSC_enable(0, 20);
    PSC1_lPSC_enable(0, 21);
    PSC1_lPSC_enable(0, 24);
    PSC1_lPSC_enable(0, 25);
    PSC1_lPSC_enable(0, 26);
    PSC1_lPSC_enable(0, 31);

    GEL_TextOut( "[Done]\n" );
}

Is the line

    PSC1_lPSC_enable(0, 3);

what I should be looking at? IF so is setting it to 0 correct? I would have thought I would set it to 3 from the SPRU if I was right about how to set it, but Since GPIO 12-15 work I must be missing something.

Ringo

I wrote a quick little code snippet using #define statements for the necessary registers to remove the dependency of a library.

#define PINMUX8 (*(int *)0x01C14140)
#define PINMUX13 (*(int *)0x01C14154)

#define GPDIR45 (*(int *)0x01E26060)
#define GPSET45 (*(int *)0x01E26068)
#define GPCLR45 (*(int *)0x01E2606C)

void main(){
 PINMUX8 |= 0x8000000;
 PINMUX13 |= 0x8000 ;

GPDIR45 &= ~((1 << 7) | (1 << 27));

 while(1){
  GPSET45 = (1 << 7) | (1 << 27);
  GPCLR45 = (1 << 7) | (1 << 27);
 }
}

I tried the code above and I do see out_data45 toggle from 0x00 to 0x08000080 on every step. However the pin never changes.

Ringo

GPIO72 is low when I power up the board, GPIO 92 is high. When I run the code they never change.

I changed the code to

// PINMUX8 |= 0x8000000;
// PINMUX13 |= 0x8000 ;
 PINMUX8 &= 0x0ffffff;
 PINMUX13 &= 0xf000 ;

so the pins should be tri-stated but it still stays 1 low, 1 high.

What else could be messing with the pins?

Ringo

I think we were messing up our bit-level logic. I got out a scope and changed my PINMUX logic to this

 PINMUX8 &= ~(0xF000000);
 PINMUX8 |= 0x8000000;
 PINMUX13 &= ~(0xF000);
 PINMUX13 |= 0x8000;

I have LEDS on the pins and I have removed the leds and am watching the pins on an O'scope and single stepping through the code. I tried the code you just posted and still get the same thing. What Gel file are you using?I'm using the evmc6747_dsp.gel, could there be something in there conflicting?

Ringo

I am also using the evmc6747_dsp.gel file. I am not sure how often they update the 'version' number but mine is listed as v1.02 at the top in the comments. I do not think that this is important though.

I am not sure what else to suggest at this point other than to make certain you are looking at the correct pins. I can attach the source code I used to test this on the EVM but considering I have already copied all my code throughout this thread I don't think it would help. I'll post it all at once just in case it does...

void main(){
 // GP4[7] and GP5[11] configured as GPIO pins
 PINMUX8 &= ~(0xF000000);
 PINMUX8 |= 0x8000000;
 PINMUX13 &= ~(0xF000);
 PINMUX13 |= 0x8000;

 // PSC configuration handled in GEL
 
 // GP4[7] and GP5[11] configured as outputs
GPDIR45 &= ~((1 << 7) | (1 << 27));

 // Toggle output on both pins
 while(1){
  if((GPSET45 & (0x80)) != 0x80){
   GPSET45 = (1 << 7) | (1 << 27);
  }
  else{
   GPCLR45 = (1 << 7) | (1 << 27);
  }
 // Some delay loop
 }
}

I pasted your code into my Main and still no luck. I have GP1[8] through 10 on terst points so I also tried this code

#define GPDIR01     (*(int *)0x01E26010)
#define GPSET01     (*(int *)0x01E26018)
#define GPCLR01     (*(int *)0x01E2601C)

 PINMUX16 &= ~(0xF000000);
 PINMUX16 |= 0x88000000;
 PINMUX17 &= ~(0xF000000);
 PINMUX17 |= 0x00000008;

  GPDIR01 |= (1 << 24) | (1 << 25) | (1 << 26);

 while(1)
 {
  GPSET01 = (1 << 24) | (1 << 25) | (1 << 26);
  GPCLR01 = (1 << 24) | (1 << 25) | (1 << 26);
 }

but it does not work either.

The original code to blink GPIO 12-15 does appear to work though, THose pins are multiplexed with HPI pins that I happen to have resistors on so I can see them toggle with the scope. The chips I have are the TMX, not TMS, would this make a difference?

Ringo

I had an error in the DIR code I just posted. It is fixed now but still none of them work. Here is my Main

void main( void )
{

#define PINMUX8 (*(int *)0x01C14140)
#define PINMUX13 (*(int *)0x01C14154)
#define GPDIR01     (*(int *)0x01E26010)
#define GPSET01     (*(int *)0x01E26018)
#define GPCLR01     (*(int *)0x01E2601C)
#define OUT_DATA01     (*(int *)0x01E26014)

#define GPDIR45     (*(int *)0x01E26060)
#define GPSET45     (*(int *)0x01E26068)
#define GPCLR45     (*(int *)0x01E2606C)
#define OUT_DATA45     (*(int *)0x01E26064)

 PINMUX8 &= ~(0xF000000);
 PINMUX8 |= 0x8000000;
 PINMUX13 &= ~(0xF000);
 PINMUX13 |= 0x8000;

 PINMUX16 &= ~(0xF000000);
 PINMUX16 |= 0x88000000;
 PINMUX17 &= ~(0xF000000);
 PINMUX17 |= 0x00000008;

 // PSC configuration handled in GEL
 
 // GP4[7] and GP5[11] configured as outputs
GPDIR45 &= ~((1 << 7) | (1 << 27));
GPDIR01 &= ~((1 << 24) | (1 << 25) | (1 << 26));

 // Toggle output on both pins
 while(1)
 {
  GPSET45 = (1 << 7) | (1 << 27);
  GPCLR45 = (1 << 7) | (1 << 27);
  GPSET01 = (1 << 24) | (1 << 25) | (1 << 26);
  GPCLR01 = (1 << 24) | (1 << 25) | (1 << 26);
 }

}

I'm not sure what to do next. I have 3 DSP's per board and I have tried different boards and they all act the same. I can read and write to external DRAM so I know the chip is working ok, just can't use any of these GPIO's

Ringo

lester davis said:

 PINMUX16 &= ~(0xFF000000);
 PINMUX16 |= 0x88000000;
 PINMUX17 &= ~(0xF);
 PINMUX17 |= 0x00000008;

PINMUX8 = 0xn8nnnnnn
PINMUX13 = 0xnnnn8nnn
PINMUX16 = 0x88nnnnnn
PINMUX17 = 0xnnnnnnn8

I finally got it working. Apparently it was a Code composer issue.

I was working in the Lab using my board and the 510USB Emulator. I dug out my EVM board and tried the code on it and it did not work. So then I tried the original LED project that came with the EVM (since I know it works) and it did not work either. I took the EVM back to my Cubicle and connected it via USB (no emulator) and tried the LED project there and it worked. So I thought maybe it had something to do with the emulator. I went back to the lab with the USB cable and attached it there. I had to reinstall the Emulator driver to get it to work. When I tested the board it worked! So now I switched from the cable to the emulator and it worked as well. I put the EVM up and went back to my board and now it works. Something in reinstalling the driver must have fixed something. It works on every board now.

Ringo

Goodness, talk about an odd issue. I would have never suspected this... Well, in any case I am glad to hear you got this resolved!