Interrupt handling triggered by GPIO-input on 405EP?

Interrupt handling triggered by GPIO-input on 405EP?

[Martin Egholm Nielsen]

Hello there,

I've done some initial coding with the ibmgpio driver with my PPC405EP
(PPChamelon) board, and inspired by the "gpio_test.c"-application, I've
succeeded in turning on some attached LEDs on some of the ports, and
reading whether some of the others ports are set high or low using the
(manually) attached pushbutton-switches.

However, the way the application is at the moment I need to poll the
input-ports every so often in order to detect a pushbutton-activation.
Hence, I would like to register an interrupt-handler (of some sort) that
informs me when these events occur. Is that somehow possible?

Best regards,
  Martin Egholm


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: Interrupt handling triggered by GPIO-input on 405EP?

[Martin Egholm Nielsen]

[-- Attachment #1: Type: text/plain, Size: 650 bytes --]

Hi Andriy,

>> I've done some initial coding with the ibmgpio driver with my
>> PPC405EP (PPChamelon) board, and inspired by the
>> "gpio_test.c"-application, I've succeeded in turning on some
>> attached LEDs on some of the ports
> can you please share code for manipulating LEDs on PPChameleon? I'm
> using the same board and it would be fine to have such possibility.
I've attached my source-file for turning on LEDs (indirectly) connected
to GPIO-pins 11,13 and 15 on the PPChamelon board.
First of all I created the /dev/gpio device by the command:

mknod /dev/gpio c 10 185

and then you can open and close the pins as you like. :o)

// Martin

[-- Attachment #2: gpio_test.c --]
[-- Type: text/plain, Size: 836 bytes --]

#include <stdio.h>
#include <fcntl.h>
#include <errno.h>
#include <string.h>

#include <linux/types.h>
#include <linux/ioctl.h>

#include <linux/ibm_ocp_gpio.h>

#define GPIO_DEV "/dev/gpio"

int main(int argc, char *argv[])
{
	/* Open the device */
	int fd = open("/dev/gpio", O_RDWR);

	struct ibm_gpio_ioctl_data gpio_ioctl;
	int result;
	int command;

	if(fd==-1) {
		fprintf(stderr,"Unable to open /dev/gpio\n");
		exit(1);
	}

	gpio_ioctl.device=0;

	/* Set IRQ4 pin (GPIO0_21) high - turn on LED */
	gpio_ioctl.mask=0x00000400;
	gpio_ioctl.data=0x00000400;
	ioctl(fd,IBMGPIO_OUT,(void *)&gpio_ioctl);

	/* Sleep 1 second */
	sleep(1);

	/* Set IRQ3-5 pin (GPIO0_20-GPIO0_22) high - turn on LEDs */
	gpio_ioctl.mask=0x00000E00;
	gpio_ioctl.data=0x00000E00;
	ioctl(fd,IBMGPIO_OUT,(void *)&gpio_ioctl);

	close( fd );

	return 0;
}

Re: Interrupt handling triggered by GPIO-input on 405EP?

[Martin Egholm Nielsen]

Hi Mark,

>>However, the way the application is at the moment I need to poll the
>>input-ports every so often in order to detect a pushbutton-activation.
>>Hence, I would like to register an interrupt-handler (of some sort) that
>>informs me when these events occur. Is that somehow possible?
> I don't know this hardware so I don't know if you can bring the GPIO into an
> actual IRQ line.
Well, according to the manual GPIO0_17-23 (IRQ 0-6) is connected to
interrupt 25-31, but I have no clue how to utilise that...

> But if you can't, a pretty efficient way to poll the GPIO is to hook into
> the heartbeat timer (100Hz, 10ms).
=== SNIP ===
> I've only done this from within a device driver, btw - the app opens
> /dev/mydevice and gets a character whenever the gpio (in my case a keypad)
> changes.
So you've created your own device-driver on top of the gpio-dev.-driver?
  And what kind of device is this? Does the driver buffer keypad inputs
even if the device is not open?
Can you please post a source example describing this in further details?

Regards,
  Martin


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Memory management problem with RAMDISK, any help?

[dong in kang]

 Hello,

 My Linux (v.2.4.22) running on a PowerPC 750 system does not seem to know
the memory bound when ramdisk is installed. The system has 128 M memory and
28 Meg of the memory is used as ramdisk. I setup the memory size as 128M
bytes at the kernel parameter and ppc_md.find_end_of_memory() routine
returns constant 128M.

  The kernel dies with " Oops: kernel access of bad area, sig: 11".
The area it accessed is the start of the last 28 Megabytes of the 128 M
memory. It happens when the kernel tries to allocate memory for socket
communication.

 How can I know the kernel knows the end of memory correctly?
Could it be other problem?

 Thanks,

 Dong-In


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/