Re: [PATCH 5/7] MIPS: SGI-IP27: rework HUB interrupts

[Christoph Hellwig]

>  struct slice_data {
>  	unsigned long irq_enable_mask[2];
> -	int level_to_irq[LEVELS_PER_SLICE];
>  };

Any reaason to keep struct slice_data around at all?

> +	HUB_S(hd->irq_mask_addr[0], si->irq_enable_mask[0]);
> +	HUB_S(hd->irq_mask_addr[1], si->irq_enable_mask[1]);

I know the HUB_S name is pre-existing, but maybe you can throw in
a patch to give it a more descriptive name?  Or maybe just kill
it off entirely at least for new code and use __raw_readq
directly.

>  #endif
>  	{
> -		/* "map" swlevel to irq */
> -		struct slice_data *si = cpu_data[cpu].data;
> -
> -		irq = si->level_to_irq[swlevel];
> -		do_IRQ(irq);
> +		do_IRQ(swlevel + IP27_HUB_IRQ_BASE);
>  	}

Looks like we can just kill the { } and additional indentation here.