Hi Rich,

On Thu, Aug 4, 2016 at 6:30 AM, Rich Felker [off-list ref] wrote:

+int __init aic_irq_of_init(struct device_node *node, struct device_node *parent)
+{
+       unsigned min_irq = JCORE_AIC2_MIN_HWIRQ;
+       unsigned dom_sz = JCORE_AIC_MAX_HWIRQ+1;
+       struct irq_domain *domain;
+
+       pr_info("Initializing J-Core AIC\n");
+
+       /* AIC1 needs priority initialization to receive interrupts. */
+       if (of_device_is_compatible(node, "jcore,aic1")) {
+               unsigned cpu;
+
+               for_each_present_cpu(cpu) {
+                       void __iomem *base = of_iomap(node, cpu);

Just double checking, these regions are per-cpu hardware registers,
and not related to other functionality at all?

I.e. when booting on an SMP-capable system a kernel compiled with
CONFIG_SMP=n, or using the kernel command line option maxcpusto reduce the number of CPUs, no ill effects happen by not mapping the
region and not writing to the register below?

+
+                       if (!base) {
+                               pr_err("Unable to map AIC for cpu %u\n", cpu);
+                               return -ENOMEM;
+                       }
+                       __raw_writel(0xffffffff, base + JCORE_AIC1_INTPRI_REG);
+                       iounmap(base);
+               }
+               min_irq = JCORE_AIC1_MIN_HWIRQ;
+       }

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds