From mboxrd@z Thu Jan  1 00:00:00 1970
From: grant.likely@secretlab.ca (Grant Likely)
Date: Wed, 21 Nov 2012 14:53:30 +0000
Subject: IRQ code inconsistencies across platforms
In-Reply-To: <53814.210.54.1.170.1349825180.squirrel@server.prisktech.co.nz>
References: <53814.210.54.1.170.1349825180.squirrel@server.prisktech.co.nz>
Message-ID: <20121121145330.3CCF53E0A47@localhost>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

On Wed, 10 Oct 2012 12:26:20 +1300 (NZDT), linux at prisktech.co.nz wrote:
> In arch/arm/common.gic.c:
> 
> static int gic_irq_domain_map(struct irq_domain *d, unsigned int irq,
> urq_hw_number_t hw)
> {
> ...
>   irq_set_chip_and_handler(irq, &gic_chip, handle_fasteoi_irq);
>   set_irq_flags(irq, IRQF_VALID | IRQF_PROBE);
> }
> 
> but in arch/arm/mach-pxa/irq.c:
> 
> static int pxa_irq_map(struct irq_domain *d, unsigned int virq, urq_hw_number_t hw)
> {
> ...
>   irq_set_chip_and_handler(hw, &pxa_internal_irq_chip, handle_level_irq);
>   set_irq_flags(hw, IRQF_VALID);
> }
> 
> Is it correct to use irq/virq or hw for irq_set_chip_and_handler/set_irq_flags?
> 
> Both platforms are adding legacy domains, so while they both presumably work,
> one is slightly incorrect.

It should be irq. The mach-pxa code looks incorrect.

g.