On Tue, Apr 27, 2021 at 06:10:09PM -0400, Joe Korty wrote:
> Balance local_irq_{disable,enable} usage in irq_forced_thread_fn
>
> Re: 0152-genirq-Allow-disabling-of-softirq-processing-in-irq-.patch
>
> In 4.9.263-rt177, irq_forced_thread_fn has potentially unbalanced calls to
> local_irq_disable ... local_irq_enable. This is probably not intentional.
>
> I am not absolutely sure what the proper fix is. Attached is an example
> of what that might look like.
>
> [ Issue detected via compiler warning, using a sufficiently advanced gcc ]
>
> Signed-off-by: Joe Korty <joe.korty@xxxxxxxxxxxxxxxxx>
Joe, I fell a bit behind on v4.9-rt and intend to get up to date this
weekend. I am now satisfied with my current backport of the futex changes
from v4.9.264 and my 96h of pi_stress testing.
Your patch makes sense, taking v4.19-rt as reference, but I will wait
for comments from the wise people you listed on the Cc: before taking
action.
Best regards,
Luis
>
> Index: b/kernel/irq/manage.c
> ===================================================================
> --- a/kernel/irq/manage.c
> +++ b/kernel/irq/manage.c
> @@ -1035,6 +1035,9 @@ irq_forced_thread_fn(struct irq_desc *de
> atomic_inc(&desc->threads_handled);
>
> irq_finalize_oneshot(desc, action);
> +
> + if (!IS_ENABLED(CONFIG_PREEMPT_RT_BASE))
> + local_irq_enable();
> /*
> * Interrupts which have real time requirements can be set up
> * to avoid softirq processing in the thread handler. This is
> @@ -1043,8 +1046,6 @@ irq_forced_thread_fn(struct irq_desc *de
> if (irq_settings_no_softirq_call(desc))
> _local_bh_enable();
> else
> - if (!IS_ENABLED(CONFIG_PREEMPT_RT_BASE))
> - local_irq_enable();
> local_bh_enable();
> return ret;
> }
---end quoted text---
