RE: Interrupts on GPIO

Message: 7
Date: Mon, 30 Mar 2009 10:32:17 -0600
From: Grant Likely <redacted>
Subject: > To: Yann Pelletier [off-list ref]
Cc: "linuxppc-dev@ozlabs.org" <redacted>
Message-ID:
	[off-list ref]
Content-Type: text/plain; charset=3DISO-8859-1
=20
On Mon, Mar 30, 2009 at 9:57 AM, Yann Pelletier
[off-list ref] wrote:

quoted

Hi,
I'd like to add supports for interrupts handling from GPIO controller.

?Some of these interrupts are from I2C devices and some others are from
SPI devices.

quoted

What is the best approach to enable interrupts handling for those

devices through GPIO.
=20
Modify the GPIO driver to also be a cascaded IRQ driver (see
http://patchwork.ozlabs.org/patch/21914/ for an example).
=20

I've looked at the patch but I wonder if this would be applicable to the MP=
C8313.  If my understanding is good, in the MPC52xx you have 8 gpio only in=
terrupts and 8 gpio/gpt.  With the MPC8313, each GPIO can act as an interru=
pt but they are all muxed to only 1 interrupt in the IPIC.  So I'm concern =
about making the GPIO controller act as an interrupt controller.

quoted

Also,

# cat /proc/interrupts
? ? ? ? ? CPU0
?16: ? ? ? ? 60 ? IPIC ? Level ? ? serial
?17: ? ? ? ? ?0 ? IPIC ? Level ? ? fpga
?18: ? ? ? ? ?4 ? IPIC ? Level ? ? mpc83xx_spi
?21: ? ? ? ? ?0 ? IPIC ? Level ? ? i2c-mpc
?22: ? ? ? ? ?7 ? IPIC ? Level ? ? i2c-mpc

According to my platform device tree, serial irq should be on int9, SPI

on int16 and i2c on int14,15. ?This causes me some problems as I need
int18 to map to external IRQ2.
=20
These numbers are *Linux internal* virtual irq numbers.  They have no
relation to the physical IRQ number for each interrupt controller.
Just make sure your Device Tree maps the IRQs correctly and you'll be
okay.
=20

This is good now, thanks

g.
=20
--
Grant Likely, B.Sc., P.Eng.
Secret Lab Technologies Ltd.
=20
=20

On Tue, Mar 31, 2009 at 2:59 PM, Yann Pelletier
[off-list ref] wrote:

quoted

Message: 7
Date: Mon, 30 Mar 2009 10:32:17 -0600
From: Grant Likely <redacted>
Subject: > To: Yann Pelletier [off-list ref]
Cc: "linuxppc-dev@ozlabs.org" <redacted>
Message-ID:
=A0 =A0 =A0 [off-list ref]
Content-Type: text/plain; charset=3DISO-8859-1

On Mon, Mar 30, 2009 at 9:57 AM, Yann Pelletier
[off-list ref] wrote:

quoted

Hi,
I'd like to add supports for interrupts handling from GPIO controller.

?Some of these interrupts are from I2C devices and some others are from
SPI devices.

quoted

What is the best approach to enable interrupts handling for those

devices through GPIO.

Modify the GPIO driver to also be a cascaded IRQ driver (see
http://patchwork.ozlabs.org/patch/21914/ for an example).

I've looked at the patch but I wonder if this would be applicable to the =

MPC8313. =A0If my understanding is good, in the MPC52xx you have 8 gpio onl=
y interrupts and 8 gpio/gpt. =A0With the MPC8313, each GPIO can act as an i=
nterrupt but they are all muxed to only 1 interrupt in the IPIC. =A0So I'm =
concern about making the GPIO controller act as an interrupt controller.

Why?  The situation really isn't any different.  Using the GPIO
controller hardware as a Linux interrupt controller means that other
drivers can just use it without having to do anything special.  It's
just and IRQ.

g.

--=20
Grant Likely, B.Sc., P.Eng.
Secret Lab Technologies Ltd.