[PATCH] irqchip/mips-gic: allow forced affinity for current cpu during hotplug

[PATCH] irqchip/mips-gic: allow forced affinity for current cpu during hotplug

[Markus Stockhausen]

Devices of the Realtek MIPS Otto platform use the official rtl-otto-timer
as clock event generator and cpu clocksource. It is registered for cpu
startup via cpuhp_setup_state() and forces the affinity of the clockevent
interrupts to the appropriate cpu via irq_force_affinity().

On the "smaller" devices with a vendor specific interrupt controller
(supported by irq-realtek-rtl) the registration works fine. The "larger"
RTL931x series is based on a MIPS interAptiv dual core with a MIPS GIC
controller. Interrupt routing setup is cancelled because gic_set_affinity()
does not accept the current (not yet online) cpu as a target.

Relax the checks by evaluating the force parameter that is provided for
exactly this purpose. With this patch affinity can be set as follows:

- force = false: works like before
- force = true: allow to set affinity to the current not yet online cpu

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
Signed-off-by: Sebastian Gottschall <s.gottschall@dd-wrt.com>
---
 drivers/irqchip/irq-mips-gic.c | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

diff --git a/drivers/irqchip/irq-mips-gic.c b/drivers/irqchip/irq-mips-gic.c
index bca8053864b2..0a88f785e114 100644
--- a/drivers/irqchip/irq-mips-gic.c
+++ b/drivers/irqchip/irq-mips-gic.c
@@ -378,9 +378,18 @@ static int gic_set_affinity(struct irq_data *d, const struct cpumask *cpumask,
 	 * the first online CPU in the mask.
 	 */
 	cpu = cpumask_first_and(cpumask, cpu_online_mask);
-	if (cpu >= NR_CPUS)
+
+	if ((cpu >= NR_CPUS) && !force)
+		/* In normal mode allow only online CPUs. */
 		return -EINVAL;
 
+	if (cpu >= NR_CPUS) {
+		/* In force mode allow current not yet online CPU for hotplug handlers. */
+		cpu = cpumask_first(cpumask);
+		if (cpu != get_cpu())
+			return -EINVAL;
+	}
+
 	old_cpu = cpumask_first(irq_data_get_effective_affinity_mask(d));
 	old_cl = cpu_cluster(&cpu_data[old_cpu]);
 	cl = cpu_cluster(&cpu_data[cpu]);
-- 
2.47.0

Re: [PATCH] irqchip/mips-gic: allow forced affinity for current cpu during hotplug

[Thomas Gleixner]

On Fri, May 23 2025 at 11:15, Markus Stockhausen wrote:
> +
> +	if ((cpu >= NR_CPUS) && !force)
> +		/* In normal mode allow only online CPUs. */
>  		return -EINVAL;
>  
> +	if (cpu >= NR_CPUS) {
> +		/* In force mode allow current not yet online CPU for hotplug handlers. */
> +		cpu = cpumask_first(cpumask);
> +		if (cpu != get_cpu())
> +			return -EINVAL;
> +	}

This logic really makes my brain hurt. Why not doing the obvious:

	if (cpu >= NR_CPUS) {
        	/* Sensible comment */	
        	if (!force)
                	return -EINVAL;
                ...
        }

Hmm?

Thanks

        tglx

AW: [PATCH] irqchip/mips-gic: allow forced affinity for current cpu during hotplug

[markus.stockhausen]

> Von: Thomas Gleixner <tglx@linutronix.de> 
>
> On Fri, May 23 2025 at 11:15, Markus Stockhausen wrote:
> > +
> > +	if ((cpu >= NR_CPUS) && !force)
> > +		/* In normal mode allow only online CPUs. */
> >  		return -EINVAL;
> >  
> > +	if (cpu >= NR_CPUS) {
> > +		/* In force mode allow current not yet online CPU for
hotplug handlers. */
> > +		cpu = cpumask_first(cpumask);
> > +		if (cpu != get_cpu())
> > +			return -EINVAL;
> > +	}
>
> This logic really makes my brain hurt. Why not doing the obvious:
>
>	if (cpu >= NR_CPUS) {
>         	/* Sensible comment */	
>         	if (!force)
>                 	return -EINVAL;
>                 ...
>         }

Then what about an even more relaxed and cleaner version like in other
drivers?

If (force)
  cpu = cpumask_first(cpumask);
else
  cpu = cpumask_first_and(cpumask, cpu_online_mask);

Markus

Re: AW: [PATCH] irqchip/mips-gic: allow forced affinity for current cpu during hotplug

[Thomas Gleixner]

On Sun, May 25 2025 at 11:43, markus stockhausen wrote:
>> Von: Thomas Gleixner <tglx@linutronix.de> 
>>
>> On Fri, May 23 2025 at 11:15, Markus Stockhausen wrote:
>> > +
>> > +	if ((cpu >= NR_CPUS) && !force)
>> > +		/* In normal mode allow only online CPUs. */
>> >  		return -EINVAL;
>> >  
>> > +	if (cpu >= NR_CPUS) {
>> > +		/* In force mode allow current not yet online CPU for
> hotplug handlers. */
>> > +		cpu = cpumask_first(cpumask);
>> > +		if (cpu != get_cpu())
>> > +			return -EINVAL;
>> > +	}
>>
>> This logic really makes my brain hurt. Why not doing the obvious:
>>
>>	if (cpu >= NR_CPUS) {
>>         	/* Sensible comment */	
>>         	if (!force)
>>                 	return -EINVAL;
>>                 ...
>>         }
>
> Then what about an even more relaxed and cleaner version like in other
> drivers?
>
> If (force)
>   cpu = cpumask_first(cpumask);
> else
>   cpu = cpumask_first_and(cpumask, cpu_online_mask);

Fine with me.