Re: [RFC PATCH] pci_request_irq (was [-mm patch] aic7xxx: check irq validity)

[Matthew Wilcox <matthew@wil.cx>]

On Mon, Oct 02, 2006 at 08:00:48PM +0000, Frederik Deweerdt wrote:
>  /**
> + * pci_request_irq - Reserve an IRQ for a PCI device
> + * @pdev: The PCI device whose irq is to be reserved
> + * handler: The interrupt handler function,

> + * pci_get_drvdata(pdev) shall be passed as an argument to that function

I don't think you can (or should) do this.  Move it to the body of the
comment below.

> + * @flags: The flags to be passed to request_irq()
> + * @name: The name of the device to be associated with the irq
> + *
> + * Returns 0 on success, or a negative value on error.  A warning
> + * message is also printed on failure.
> + */
> +int pci_request_irq(struct pci_dev *pdev,
> +		    irqreturn_t (*handler)(int, void *, struct pt_regs *),
> +		    unsigned long flags, const char *name)
> +{
> +	int rc;
> +	const char *actual_name = name;
> +
> +	rc = is_irq_valid(pdev->irq);
> +	if (!rc) {
> +		dev_printk(KERN_ERR, &pdev->dev, "invalid irq #%d\n", pdev->irq);
> +		return -EINVAL;
> +	}

Why is that more readable than

	if (!is_irq_valid(pdev->irq)) {
		dev_err(&pdev->dev, "invalid irq #%d\n", pdev->irq);
		return -EINVAL;
	}

> +	if (!actual_name)
> +		actual_name = pci_name(pdev);
> +
> +	return request_irq(pdev->irq, handler, flags | IRQF_SHARED,
> +			   actual_name, pci_get_drvdata(pdev));

The driver name is a far more common usage than the pci_name.

	return request_irq(pdev->irq, handler, flags | IRQF_SHARED,
			name ? name : pdev->driver->name,
			pci_get_drvdata(pdev));