[v3] irqchip: omap-intc: add support for spurious irq handling
diff mbox

Message ID 9c78a6db02ac55f7af7371b417b6e414d2c3095b.1450188128.git.nsekhar@ti.com
State New
Headers show

Commit Message

Sekhar Nori Dec. 15, 2015, 2:26 p.m. UTC
Under some conditions, irq sorting procedure used
by INTC can go wrong resulting in a spurious irq
getting reported.

If this condition is not handled, it results in
endless stream of:

    unexpected IRQ trap at vector 00

messages from ack_bad_irq()

Handle the spurious interrupt condition in omap-intc
driver to prevent this.

Measurements using kernel function profiler on AM335x
EVM running at 720MHz show that after this patch
omap_intc_handle_irq() takes about 37.4us against
34us before this patch.

Signed-off-by: Sekhar Nori <nsekhar@ti.com>
---
v3: add unlikely(), add profiling information
    to commit message.

v2: increment error irq counter, use pr_err_once,
    add a comment on tips to debug spurious irq
    condition.

This patch results in a checkpatch warning about
extern definition of irq_err_count, but looks like
thats the prevalent method of accessing that counter.

 drivers/irqchip/irq-omap-intc.c | 27 ++++++++++++++++++++++++++-
 1 file changed, 26 insertions(+), 1 deletion(-)

Comments

Tony Lindgren Dec. 15, 2015, 3:28 p.m. UTC | #1
* Sekhar Nori <nsekhar@ti.com> [151215 06:26]:
> Under some conditions, irq sorting procedure used
> by INTC can go wrong resulting in a spurious irq
> getting reported.
> 
> If this condition is not handled, it results in
> endless stream of:
> 
>     unexpected IRQ trap at vector 00
> 
> messages from ack_bad_irq()
> 
> Handle the spurious interrupt condition in omap-intc
> driver to prevent this.
> 
> Measurements using kernel function profiler on AM335x
> EVM running at 720MHz show that after this patch
> omap_intc_handle_irq() takes about 37.4us against
> 34us before this patch.
> 
> Signed-off-by: Sekhar Nori <nsekhar@ti.com>

Looks good to me, probably should get tagged Cc stable when
committing:

Acked-by: Tony Lindgren <tony@atomide.com>


> ---
> v3: add unlikely(), add profiling information
>     to commit message.
> 
> v2: increment error irq counter, use pr_err_once,
>     add a comment on tips to debug spurious irq
>     condition.
> 
> This patch results in a checkpatch warning about
> extern definition of irq_err_count, but looks like
> thats the prevalent method of accessing that counter.
> 
>  drivers/irqchip/irq-omap-intc.c | 27 ++++++++++++++++++++++++++-
>  1 file changed, 26 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/irqchip/irq-omap-intc.c b/drivers/irqchip/irq-omap-intc.c
> index 8587d0f8d8c0..f6cb1b8bb981 100644
> --- a/drivers/irqchip/irq-omap-intc.c
> +++ b/drivers/irqchip/irq-omap-intc.c
> @@ -47,6 +47,7 @@
>  #define INTC_ILR0		0x0100
>  
>  #define ACTIVEIRQ_MASK		0x7f	/* omap2/3 active interrupt bits */
> +#define SPURIOUSIRQ_MASK	(0x1ffffff << 7)
>  #define INTCPS_NR_ILR_REGS	128
>  #define INTCPS_NR_MIR_REGS	4
>  
> @@ -330,11 +331,35 @@ static int __init omap_init_irq(u32 base, struct device_node *node)
>  static asmlinkage void __exception_irq_entry
>  omap_intc_handle_irq(struct pt_regs *regs)
>  {
> +	extern unsigned long irq_err_count;
>  	u32 irqnr;
>  
>  	irqnr = intc_readl(INTC_SIR);
> +
> +	/*
> +	 * A spurious IRQ can result if interrupt that triggered the
> +	 * sorting is no longer active during the sorting (10 INTC
> +	 * functional clock cycles after interrupt assertion). Or a
> +	 * change in interrupt mask affected the result during sorting
> +	 * time. There is no special handling required except ignoring
> +	 * the SIR register value just read and retrying.
> +	 * See section 6.2.5 of AM335x TRM Literature Number: SPRUH73K
> +	 *
> +	 * Many a times, a spurious interrupt situation has been fixed
> +	 * by adding a flush for the posted write acking the IRQ in
> +	 * the device driver. Typically, this is going be the device
> +	 * driver whose interrupt was handled just before the spurious
> +	 * IRQ occurred. Pay attention to those device drivers if you
> +	 * run into hitting the spurious IRQ condition below.
> +	 */
> +	if (unlikely((irqnr & SPURIOUSIRQ_MASK) == SPURIOUSIRQ_MASK)) {
> +		pr_err_once("%s: spurious irq!\n", __func__);
> +		irq_err_count++;
> +		omap_ack_irq(NULL);
> +		return;
> +	}
> +
>  	irqnr &= ACTIVEIRQ_MASK;
> -	WARN_ONCE(!irqnr, "Spurious IRQ ?\n");
>  	handle_domain_irq(domain, irqnr, regs);
>  }
>  
> -- 
> 2.6.3
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Sekhar Nori Jan. 4, 2016, 8:35 a.m. UTC | #2
Hi Thomas,

On Tuesday 15 December 2015 08:58 PM, Tony Lindgren wrote:
> * Sekhar Nori <nsekhar@ti.com> [151215 06:26]:
>> Under some conditions, irq sorting procedure used
>> by INTC can go wrong resulting in a spurious irq
>> getting reported.
>>
>> If this condition is not handled, it results in
>> endless stream of:
>>
>>     unexpected IRQ trap at vector 00
>>
>> messages from ack_bad_irq()
>>
>> Handle the spurious interrupt condition in omap-intc
>> driver to prevent this.
>>
>> Measurements using kernel function profiler on AM335x
>> EVM running at 720MHz show that after this patch
>> omap_intc_handle_irq() takes about 37.4us against
>> 34us before this patch.
>>
>> Signed-off-by: Sekhar Nori <nsekhar@ti.com>
> 
> Looks good to me, probably should get tagged Cc stable when
> committing:
> 
> Acked-by: Tony Lindgren <tony@atomide.com>

Can you please apply this if it looks good?

Thanks,
Sekhar
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch
diff mbox

diff --git a/drivers/irqchip/irq-omap-intc.c b/drivers/irqchip/irq-omap-intc.c
index 8587d0f8d8c0..f6cb1b8bb981 100644
--- a/drivers/irqchip/irq-omap-intc.c
+++ b/drivers/irqchip/irq-omap-intc.c
@@ -47,6 +47,7 @@ 
 #define INTC_ILR0		0x0100
 
 #define ACTIVEIRQ_MASK		0x7f	/* omap2/3 active interrupt bits */
+#define SPURIOUSIRQ_MASK	(0x1ffffff << 7)
 #define INTCPS_NR_ILR_REGS	128
 #define INTCPS_NR_MIR_REGS	4
 
@@ -330,11 +331,35 @@  static int __init omap_init_irq(u32 base, struct device_node *node)
 static asmlinkage void __exception_irq_entry
 omap_intc_handle_irq(struct pt_regs *regs)
 {
+	extern unsigned long irq_err_count;
 	u32 irqnr;
 
 	irqnr = intc_readl(INTC_SIR);
+
+	/*
+	 * A spurious IRQ can result if interrupt that triggered the
+	 * sorting is no longer active during the sorting (10 INTC
+	 * functional clock cycles after interrupt assertion). Or a
+	 * change in interrupt mask affected the result during sorting
+	 * time. There is no special handling required except ignoring
+	 * the SIR register value just read and retrying.
+	 * See section 6.2.5 of AM335x TRM Literature Number: SPRUH73K
+	 *
+	 * Many a times, a spurious interrupt situation has been fixed
+	 * by adding a flush for the posted write acking the IRQ in
+	 * the device driver. Typically, this is going be the device
+	 * driver whose interrupt was handled just before the spurious
+	 * IRQ occurred. Pay attention to those device drivers if you
+	 * run into hitting the spurious IRQ condition below.
+	 */
+	if (unlikely((irqnr & SPURIOUSIRQ_MASK) == SPURIOUSIRQ_MASK)) {
+		pr_err_once("%s: spurious irq!\n", __func__);
+		irq_err_count++;
+		omap_ack_irq(NULL);
+		return;
+	}
+
 	irqnr &= ACTIVEIRQ_MASK;
-	WARN_ONCE(!irqnr, "Spurious IRQ ?\n");
 	handle_domain_irq(domain, irqnr, regs);
 }