svn commit: r271601 - in head/sys/arm: arm include

Andrew Turner andrew at fubar.geek.nz
Mon Sep 15 07:48:51 UTC 2014


On Sun, 14 Sep 2014 21:21:04 +0000 (UTC)
Ian Lepore <ian at FreeBSD.org> wrote:

...

> Modified: head/sys/arm/arm/gic.c
> ==============================================================================
> --- head/sys/arm/arm/gic.c	Sun Sep 14 20:13:07 2014
> (r271600) +++ head/sys/arm/arm/gic.c	Sun Sep 14 21:21:03
> 2014	(r271601) @@ -185,6 +185,51 @@ gic_init_secondary(void)
>  	gic_d_write_4(GICD_ISENABLER(30 >> 5), (1UL << (30 & 0x1F)));
>  }
>  
> +int
> +gic_decode_fdt(uint32_t iparent, uint32_t *intr, int *interrupt,
> +    int *trig, int *pol)
> +{
> +	static u_int num_intr_cells;
> +
> +	if (num_intr_cells == 0) {
> +		if (OF_searchencprop(OF_node_from_xref(iparent), 
> +		    "#interrupt-cells", &num_intr_cells, 
> +		    sizeof(num_intr_cells)) == -1) {
> +			num_intr_cells = 1;
> +		}
> +	}
> +
> +	if (num_intr_cells == 1) {
> +		*interrupt = fdt32_to_cpu(intr[0]);
> +		*trig = INTR_TRIGGER_CONFORM;
> +		*pol = INTR_POLARITY_CONFORM;
> +	} else {
> +		if (intr[0] == 0)
> +			*interrupt = fdt32_to_cpu(intr[1]) + 32;
> +		else
> +			*interrupt = fdt32_to_cpu(intr[1]);

Shouldn't the PPI case have a +16? The PPI interrupts start from 16
with 0-15 being used for software-generated interrupts. My copy of the
binding document state the values for the PPI interrupts should be in
the range 0 to 15 meaning they will be assigned an SGI value here.

Andrew


More information about the svn-src-all mailing list