Re: [PATCH v3] uio/pdrv_genirq: Add OF support

["Hans J. Koch" <hjk@hansjkoch.de>]

On Tue, Apr 19, 2011 at 12:08:16AM -0600, Grant Likely wrote:
> On Mon, Apr 18, 2011 at 10:50:54AM +0200, Michal Simek wrote:
> > Adding OF binding to genirq.
> > Version string is setup to the "devicetree".
> > 
> > Compatible string is not setup for now but you can add your
> > custom compatible string to uio_of_genirq_match structure.
> > 
> > For example with "uio" compatible string:
> > static const struct of_device_id __devinitconst uio_of_genirq_match[] = {
> > 	{ .compatible = "uio", },
> > 	{ /* empty for now */ },
> > };
> > 
> > Signed-off-by: Michal Simek <monstr@monstr.eu>
> > 
> > ---
> > 
> > v3: Fix irq binding
> >     Use "devicetree" as version string
> > 
> > v2: Remove additional resource binding
> >     Setup correct version string
> >     Clear compatible string
> > ---
> >  drivers/uio/uio_pdrv_genirq.c |   42 +++++++++++++++++++++++++++++++++++++++++
> >  1 files changed, 42 insertions(+), 0 deletions(-)
> > 
> > diff --git a/drivers/uio/uio_pdrv_genirq.c b/drivers/uio/uio_pdrv_genirq.c
> > index 7174d51..c8d565a 100644
> > --- a/drivers/uio/uio_pdrv_genirq.c
> > +++ b/drivers/uio/uio_pdrv_genirq.c
> > @@ -23,6 +23,10 @@
> >  #include <linux/pm_runtime.h>
> >  #include <linux/slab.h>
> >  
> > +#include <linux/of.h>
> > +#include <linux/of_platform.h>
> > +#include <linux/of_address.h>
> > +
> >  #define DRIVER_NAME "uio_pdrv_genirq"
> >  
> >  struct uio_pdrv_genirq_platdata {
> > @@ -97,6 +101,30 @@ static int uio_pdrv_genirq_probe(struct platform_device *pdev)
> >  	int ret = -EINVAL;
> >  	int i;
> >  
> > +	if (!uioinfo) {
> > +		int irq;
> > +
> > +		/* alloc uioinfo for one device */
> > +		uioinfo = kzalloc(sizeof(*uioinfo), GFP_KERNEL);
> > +		if (!uioinfo) {
> > +			ret = -ENOMEM;
> > +			dev_err(&pdev->dev, "unable to kmalloc\n");
> > +			goto bad2;
> > +		}
> > +		uioinfo->name = pdev->dev.of_node->name;
> > +		uioinfo->version = "devicetree";
> > +
> > +		/* Multiple IRQs are not supported */
> > +		irq = platform_get_irq(pdev, 0);
> > +		if (irq < 0) {
> 
> PowerPC and x86 will return 0 for an unassigned IRQ, as will most platforms.

That might be right for these architectures. On ARM SoCs, IRQ0 is often a
normal irq like any other (e.g. "Audio DMA Controller 1" on Telechips TCC8000).

We've had this discussion before. In my original uio_driver.h, UIO_IRQ_NONE
was #defined as -1. There was heavy opposition against that, and I finally
accepted a patch changing UIO_IRQ_NONE to 0. It's a tradeoff between 0 being
a valid interrupt number on most archs and the convenience of the compiler
automatically inserting a 0 into an unassigned variable.

The above line

if (irq < 0) {

is probably a copy-paste leftover of some older code. Today, irq < 0 should
never be true in UIO, so the line is wrong.

Thanks for pointing to it,
Hans