Re: [PATCH] irqchip/gic: Return IRQ_SET_MASK_OK_DONE in the set_affinity method

2016-02-19 Thread Jason Cooper
On Fri, Feb 19, 2016 at 03:00:29PM +, Marc Zyngier wrote: > Moving an SPI around doesn't require any extra work from the rest > of the stack, and specially not for MSI-generated SPIs. > > It is then worth returning IRQ_SET_MASK_OK_DONE instead of > IRQ_SET_MASK_OK, and simplify the other

Re: [PATCH] irqchip/gic: Return IRQ_SET_MASK_OK_DONE in the set_affinity method

2016-02-19 Thread Jason Cooper
On Fri, Feb 19, 2016 at 03:00:29PM +, Marc Zyngier wrote: > Moving an SPI around doesn't require any extra work from the rest > of the stack, and specially not for MSI-generated SPIs. > > It is then worth returning IRQ_SET_MASK_OK_DONE instead of > IRQ_SET_MASK_OK, and simplify the other

Re: [PATCH] irqchip/gic: Return IRQ_SET_MASK_OK_DONE in the set_affinity method

2016-02-19 Thread Thomas Petazzoni
Marc, On Fri, 19 Feb 2016 15:00:29 +, Marc Zyngier wrote: > Moving an SPI around doesn't require any extra work from the rest > of the stack, and specially not for MSI-generated SPIs. > > It is then worth returning IRQ_SET_MASK_OK_DONE instead of > IRQ_SET_MASK_OK, and simplify the other

Re: [PATCH] irqchip/gic: Return IRQ_SET_MASK_OK_DONE in the set_affinity method

2016-02-19 Thread Thomas Petazzoni
Marc, On Fri, 19 Feb 2016 15:00:29 +, Marc Zyngier wrote: > Moving an SPI around doesn't require any extra work from the rest > of the stack, and specially not for MSI-generated SPIs. > > It is then worth returning IRQ_SET_MASK_OK_DONE instead of > IRQ_SET_MASK_OK, and simplify the other

[PATCH] irqchip/gic: Return IRQ_SET_MASK_OK_DONE in the set_affinity method

2016-02-19 Thread Marc Zyngier
Moving an SPI around doesn't require any extra work from the rest of the stack, and specially not for MSI-generated SPIs. It is then worth returning IRQ_SET_MASK_OK_DONE instead of IRQ_SET_MASK_OK, and simplify the other irqchips that rely on this behaviour (GICv2m and Marvell's ODMI controller).

[PATCH] irqchip/gic: Return IRQ_SET_MASK_OK_DONE in the set_affinity method

2016-02-19 Thread Marc Zyngier
Moving an SPI around doesn't require any extra work from the rest of the stack, and specially not for MSI-generated SPIs. It is then worth returning IRQ_SET_MASK_OK_DONE instead of IRQ_SET_MASK_OK, and simplify the other irqchips that rely on this behaviour (GICv2m and Marvell's ODMI controller).