On 12/01/16 19:13, Tuan Phan wrote:

On Tue, Jan 12, 2016 at 10:59 AM, Marc Zyngier [off-list ref] wrote:

quoted

On 07/11/15 00:51, Loc Ho wrote:

quoted

The function acpi_gsi_to_irq should returns 0 on success as upper function
caller expect an 0 for sucesss.

I have to ask: why do you feel the need to change an API that behaves
like the rest of the IRQ allocation functions (at least when it comes to
its return value)?

0 is defined as an invalid interrupt, and I wish it stayed that way.

The upper caller expects 0 for success.
In drivers/acpi/apei/ghes.c:
        case ACPI_HEST_NOTIFY_EXTERNAL:
                /* External interrupt vector is GSI */
                rc = acpi_gsi_to_irq(generic->notify.vector, &ghes->irq);
                if (rc) {
                        pr_err(GHES_PFX "Failed to map GSI to IRQ for
generic hardware error source: %d\n",
                               generic->header.source_id);
                        goto err_edac_unreg;
                }

Also the implementation of acpi_gsi_to_irq for X86 in
arch/x86/kernel/acpi/boot.c also return 0 for success and -1 for
failure.
int acpi_gsi_to_irq(u32 gsi, unsigned int *irqp)
{
        int rc, irq, trigger, polarity;

        if (acpi_irq_model == ACPI_IRQ_MODEL_PIC) {
                *irqp = gsi;
                return 0;
        }

        rc = acpi_get_override_irq(gsi, &trigger, &polarity);
        if (rc == 0) {
                trigger = trigger ? ACPI_LEVEL_SENSITIVE : ACPI_EDGE_SENSITIVE;
                polarity = polarity ? ACPI_ACTIVE_LOW : ACPI_ACTIVE_HIGH;
                irq = acpi_register_gsi(NULL, gsi, trigger, polarity);
                if (irq >= 0) {
                        *irqp = irq;
                        return 0;
                }
        }

        return -1;
}

Right. In which case please document this properly in the commit log.

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...