Re: kvm+nouveau induced lockdep gripe

[Sebastian Andrzej Siewior]

On 2020-10-26 20:15:23 [+0100], Mike Galbraith wrote:
> --- a/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c
> +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/pci/base.c
> @@ -116,7 +116,17 @@ nvkm_pci_oneinit(struct nvkm_subdev *sub
>  			return ret;
>  	}
> 
> +	/*
> +	 * Scheduler code taking cpuset_rwsem during irq thread initialization sets
> +	 * up a cpuset_rwsem vs mm->mmap_lock circular dependency gripe upon later
> +	 * cpuset usage. It's harmless, tell lockdep there's nothing to see here.
> +	 */
> +	if (force_irqthreads)
> +		lockdep_off();
>  	ret = request_irq(pdev->irq, nvkm_pci_intr, IRQF_SHARED, "nvkm", pci);
> +	if (force_irqthreads)
> +		lockdep_on();
> +
>  	if (ret)
>  		return ret;
> 
Could you try this, please?

--- a/kernel/irq/manage.c
+++ b/kernel/irq/manage.c
@@ -1155,6 +1155,8 @@ static int irq_thread(void *data)
 	irqreturn_t (*handler_fn)(struct irq_desc *desc,
 			struct irqaction *action);
 
+	sched_set_fifo(current);
+
 	if (force_irqthreads && test_bit(IRQTF_FORCED_THREAD,
 					&action->thread_flags))
 		handler_fn = irq_forced_thread_fn;
@@ -1320,8 +1322,6 @@ setup_irq_thread(struct irqaction *new,
 	if (IS_ERR(t))
 		return PTR_ERR(t);
 
-	sched_set_fifo(t);
-
 	/*
 	 * We keep the reference to the task struct even if
 	 * the thread dies to avoid that the interrupt code


Sebastian