[PATCH] genirq: spurious.c: fix regression that broke irqfixup,irqpoll

[PATCH] genirq: spurious.c: fix regression that broke irqfixup,irqpoll

[edward.donovan]

From: Edward Donovan <edward.donovan@numble.net>

commit  d05c65fff0 ("genirq: spurious: Run only one poller at a time")
introduced a regression, leaving the boot options 'irqfixup' and
'irqpoll' non-functional.  The patch placed tests in each function, to
exit if the function is already running.  The test in 'misrouted_irq'
exited when it should have proceeded, effectively disabling
'misrouted_irq' and 'poll_spurious_irqs'.  Reversing the "==" operator
to "!=" fixes the regression.

Signed-off-by: Edward Donovan <edward.donovan@numble.net>
---
 kernel/irq/spurious.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/kernel/irq/spurious.c b/kernel/irq/spurious.c
index aa57d5d..b5f4742 100644
--- a/kernel/irq/spurious.c
+++ b/kernel/irq/spurious.c
@@ -115,7 +115,7 @@ static int misrouted_irq(int irq)
 	struct irq_desc *desc;
 	int i, ok = 0;
 
-	if (atomic_inc_return(&irq_poll_active) == 1)
+	if (atomic_inc_return(&irq_poll_active) != 1)
 		goto out;
 
 	irq_poll_cpu = smp_processor_id();
-- 
1.7.5.4

Re: [PATCH] genirq: spurious.c: fix regression that broke irqfixup,irqpoll

[Thomas Gleixner]

On Tue, 1 Nov 2011, edward.donovan@numble.net wrote:

> From: Edward Donovan <edward.donovan@numble.net>
> 
> commit  d05c65fff0 ("genirq: spurious: Run only one poller at a time")
> introduced a regression, leaving the boot options 'irqfixup' and
> 'irqpoll' non-functional.  The patch placed tests in each function, to
> exit if the function is already running.  The test in 'misrouted_irq'
> exited when it should have proceeded, effectively disabling
> 'misrouted_irq' and 'poll_spurious_irqs'.  Reversing the "==" operator
> to "!=" fixes the regression.
> 
> Signed-off-by: Edward Donovan <edward.donovan@numble.net>

Good catch! I was staring at it with my conference and flu damaged
brain ....

Thanks,

	tglx

> ---
>  kernel/irq/spurious.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/kernel/irq/spurious.c b/kernel/irq/spurious.c
> index aa57d5d..b5f4742 100644
> --- a/kernel/irq/spurious.c
> +++ b/kernel/irq/spurious.c
> @@ -115,7 +115,7 @@ static int misrouted_irq(int irq)
>  	struct irq_desc *desc;
>  	int i, ok = 0;
>  
> -	if (atomic_inc_return(&irq_poll_active) == 1)
> +	if (atomic_inc_return(&irq_poll_active) != 1)
>  		goto out;
>  
>  	irq_poll_cpu = smp_processor_id();
> -- 
> 1.7.5.4
> 
> 

Re: [PATCH] genirq: spurious.c: fix regression that broke irqfixup,irqpoll

[Edward Donovan]

On Nov 2, 2011 2:02 PM, "Thomas Gleixner" <tglx@linutronix.de> wrote:
>
> Good catch! I was staring at it with my conference and flu damaged
> brain ....
>

Thanks, Thomas.  It had me running around for quite a while before I
caught it. :)

I hoped this had nailed the failure of irqfixup/poll, since the .39
merge window.  But I have one machine where this fix doesn't help.
(With this patch, 2.6.39+ went from never handling that bad irq, to
handling it in a fraction of bootups.  I have to dig into that
more.)

On my Ubuntu bug report I have asked the other reporters for
testing, but haven't heard yet.

It may be back to the bisecting board for me, and I'll post again
if I learn anything.

 Ed