[PATCH 0/1] ARM: at91: fix drivers broken because of sparse irq support

[PATCH 0/1] ARM: at91: fix drivers broken because of sparse irq support

[ludovic.desroches at atmel.com]

From: Ludovic Desroches <ludovic.desroches@atmel.com>

Hello,

This patch fixes the rm9200 breakage (introduced by sparse irq support) reported
by Joachim Eastwood.

Regards

Ludovic

Ludovic Desroches (1):
  ARM: at91: fix drivers broken because of sparse irq support

 arch/arm/mach-at91/at91rm9200_time.c |  2 +-
 drivers/rtc/rtc-at91sam9.c           | 11 ++++++-----
 2 files changed, 7 insertions(+), 6 deletions(-)

-- 
1.7.11.3

[PATCH] ARM: at91: fix drivers broken because of sparse irq support

[ludovic.desroches at atmel.com]

From: Ludovic Desroches <ludovic.desroches@atmel.com>

Some drivers directly use AT91_ID_SYS as virq so they are broken because of
the NR_IRQS_LEGACY virq offset introduced with sparse irq support.

Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com>
Cc: stable at vger.kernel.org
---
 arch/arm/mach-at91/at91rm9200_time.c |  2 +-
 drivers/rtc/rtc-at91sam9.c           | 11 ++++++-----
 2 files changed, 7 insertions(+), 6 deletions(-)

diff --git a/arch/arm/mach-at91/at91rm9200_time.c b/arch/arm/mach-at91/at91rm9200_time.c
index 104ca40..aaa443b 100644
--- a/arch/arm/mach-at91/at91rm9200_time.c
+++ b/arch/arm/mach-at91/at91rm9200_time.c
@@ -197,7 +197,7 @@ void __init at91rm9200_timer_init(void)
 	at91_st_read(AT91_ST_SR);
 
 	/* Make IRQs happen for the system timer */
-	setup_irq(AT91_ID_SYS, &at91rm9200_timer_irq);
+	setup_irq(NR_IRQS_LEGACY + AT91_ID_SYS, &at91rm9200_timer_irq);
 
 	/* The 32KiHz "Slow Clock" (tick every 30517.58 nanoseconds) is used
 	 * directly for the clocksource and all clockevents, after adjusting
diff --git a/drivers/rtc/rtc-at91sam9.c b/drivers/rtc/rtc-at91sam9.c
index 8318689..2969047 100644
--- a/drivers/rtc/rtc-at91sam9.c
+++ b/drivers/rtc/rtc-at91sam9.c
@@ -345,11 +345,12 @@ static int __devinit at91_rtc_probe(struct platform_device *pdev)
 	}
 
 	/* register irq handler after we know what name we'll use */
-	ret = request_irq(AT91_ID_SYS, at91_rtc_interrupt,
+	ret = request_irq(NR_IRQS_LEGACY + AT91_ID_SYS, at91_rtc_interrupt,
 				IRQF_SHARED,
 				dev_name(&rtc->rtcdev->dev), rtc);
 	if (ret) {
-		dev_dbg(&pdev->dev, "can't share IRQ %d?\n", AT91_ID_SYS);
+		dev_dbg(&pdev->dev, "can't share IRQ %d?\n",
+			NR_IRQS_LEGACY + AT91_ID_SYS);
 		rtc_device_unregister(rtc->rtcdev);
 		goto fail_register;
 	}
@@ -386,7 +387,7 @@ static int __devexit at91_rtc_remove(struct platform_device *pdev)
 
 	/* disable all interrupts */
 	rtt_writel(rtc, MR, mr & ~(AT91_RTT_ALMIEN | AT91_RTT_RTTINCIEN));
-	free_irq(AT91_ID_SYS, rtc);
+	free_irq(NR_IRQS_LEGACY + AT91_ID_SYS, rtc);
 
 	rtc_device_unregister(rtc->rtcdev);
 
@@ -423,7 +424,7 @@ static int at91_rtc_suspend(struct platform_device *pdev,
 	rtc->imr = mr & (AT91_RTT_ALMIEN | AT91_RTT_RTTINCIEN);
 	if (rtc->imr) {
 		if (device_may_wakeup(&pdev->dev) && (mr & AT91_RTT_ALMIEN)) {
-			enable_irq_wake(AT91_ID_SYS);
+			enable_irq_wake(NR_IRQS_LEGACY + AT91_ID_SYS);
 			/* don't let RTTINC cause wakeups */
 			if (mr & AT91_RTT_RTTINCIEN)
 				rtt_writel(rtc, MR, mr & ~AT91_RTT_RTTINCIEN);
@@ -441,7 +442,7 @@ static int at91_rtc_resume(struct platform_device *pdev)
 
 	if (rtc->imr) {
 		if (device_may_wakeup(&pdev->dev))
-			disable_irq_wake(AT91_ID_SYS);
+			disable_irq_wake(NR_IRQS_LEGACY + AT91_ID_SYS);
 		mr = rtt_readl(rtc, MR);
 		rtt_writel(rtc, MR, mr | rtc->imr);
 	}
-- 
1.7.11.3

[PATCH] ARM: at91: fix drivers broken because of sparse irq support

[Joachim Eastwood]

Hi Ludovic,

On Mon, Aug 13, 2012 at 5:38 PM,  <ludovic.desroches@atmel.com> wrote:
> From: Ludovic Desroches <ludovic.desroches@atmel.com>
>
> Some drivers directly use AT91_ID_SYS as virq so they are broken because of
> the NR_IRQS_LEGACY virq offset introduced with sparse irq support.

This patch makes my board boot again. Thanks.

Feel free to add;
Tested-by: Joachim Eastwood <joachim.eastwood@jotron.com>

> Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com>
> Cc: stable at vger.kernel.org

Why are you sending this to stable?
This is a regression only in 3.6 and should go upstream ASAP.

regards
Joachim Eastwood

[PATCH] ARM: at91: fix drivers broken because of sparse irq support

[Jean-Christophe PLAGNIOL-VILLARD]

On 17:38 Mon 13 Aug     , ludovic.desroches at atmel.com wrote:
> From: Ludovic Desroches <ludovic.desroches@atmel.com>
> 
> Some drivers directly use AT91_ID_SYS as virq so they are broken because of
> the NR_IRQS_LEGACY virq offset introduced with sparse irq support.
> 
> Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com>
> Cc: stable at vger.kernel.org
> ---
>  arch/arm/mach-at91/at91rm9200_time.c |  2 +-
>  drivers/rtc/rtc-at91sam9.c           | 11 ++++++-----
>  2 files changed, 7 insertions(+), 6 deletions(-)
> 
> diff --git a/arch/arm/mach-at91/at91rm9200_time.c b/arch/arm/mach-at91/at91rm9200_time.c
> index 104ca40..aaa443b 100644
> --- a/arch/arm/mach-at91/at91rm9200_time.c
> +++ b/arch/arm/mach-at91/at91rm9200_time.c
> @@ -197,7 +197,7 @@ void __init at91rm9200_timer_init(void)
>  	at91_st_read(AT91_ST_SR);
>  
>  	/* Make IRQs happen for the system timer */
> -	setup_irq(AT91_ID_SYS, &at91rm9200_timer_irq);
> +	setup_irq(NR_IRQS_LEGACY + AT91_ID_SYS, &at91rm9200_timer_irq);
>  
>  	/* The 32KiHz "Slow Clock" (tick every 30517.58 nanoseconds) is used
>  	 * directly for the clocksource and all clockevents, after adjusting
> diff --git a/drivers/rtc/rtc-at91sam9.c b/drivers/rtc/rtc-at91sam9.c
> index 8318689..2969047 100644
> --- a/drivers/rtc/rtc-at91sam9.c
> +++ b/drivers/rtc/rtc-at91sam9.c
> @@ -345,11 +345,12 @@ static int __devinit at91_rtc_probe(struct platform_device *pdev)
>  	}
>  
>  	/* register irq handler after we know what name we'll use */
> -	ret = request_irq(AT91_ID_SYS, at91_rtc_interrupt,
> +	ret = request_irq(NR_IRQS_LEGACY + AT91_ID_SYS, at91_rtc_interrupt,
the irq need to be pass by resources and not hardcoded

Best Regrds,
J.