arm: omap: make wakeupgen_lock raw

Commit Message

Felipe Balbi Dec. 7, 2012, 2:49 p.m. UTC

From: Thomas Gleixner <[e-mail]>

When applying RT patch on top of Linux, spinlocks are
implemented as RT-mutexes, which means they are preemptable.

Current GIC implementation on OMAP is using a spinlock
to protect against preemption. As it turns out, we need
to convert that lock into a raw_spinlock so that OMAP's
interrupt controller works as expected after RT-patch
is applied.

This patch is simply to decrease the amount of changes
RT-team needs to carry out of tree. It doesn't cause any
changes in behavior.

Signed-off-by: Thomas Gleixner <[e-mail]>
Signed-off-by: Felipe Balbi <[e-mail]>
---

Thomas, all I did here was take the patch from v3.6.2-rt21
patchset, apply to current Linus' tree (v3.7-rc8) and compile
test with our omap2plus_defconfig.

If you think the commit log I wrong is lacking, please suggest
changes and I will fix it up. Just thought this is the kind of
patch which can go to mainline today and decrease the amount of
patches out-of-tree regarding RT support.

cheers

 arch/arm/mach-omap2/omap-wakeupgen.c | 14 +++++++-------
 1 file changed, 7 insertions(+), 7 deletions(-)

Comments

Felipe Balbi Jan. 24, 2013, 6:22 p.m. UTC | #1

Hi,

On Fri, Dec 07, 2012 at 04:49:47PM +0200, Felipe Balbi wrote:
> From: Thomas Gleixner <[e-mail]>
> 
> When applying RT patch on top of Linux, spinlocks are
> implemented as RT-mutexes, which means they are preemptable.
> 
> Current GIC implementation on OMAP is using a spinlock
> to protect against preemption. As it turns out, we need
> to convert that lock into a raw_spinlock so that OMAP's
> interrupt controller works as expected after RT-patch
> is applied.
> 
> This patch is simply to decrease the amount of changes
> RT-team needs to carry out of tree. It doesn't cause any
> changes in behavior.
> 
> Signed-off-by: Thomas Gleixner <[e-mail]>
> Signed-off-by: Felipe Balbi <[e-mail]>

ping

Santosh Shilimkar Jan. 25, 2013, 7:51 a.m. UTC | #2

On Thursday 24 January 2013 11:52 PM, Felipe Balbi wrote:
> Hi,
>
> On Fri, Dec 07, 2012 at 04:49:47PM +0200, Felipe Balbi wrote:
>> From: Thomas Gleixner <[e-mail]>
>>
>> When applying RT patch on top of Linux, spinlocks are
>> implemented as RT-mutexes, which means they are preemptable.
>>
>> Current GIC implementation on OMAP is using a spinlock
>> to protect against preemption. As it turns out, we need
>> to convert that lock into a raw_spinlock so that OMAP's
>> interrupt controller works as expected after RT-patch
>> is applied.
>>
>> This patch is simply to decrease the amount of changes
>> RT-team needs to carry out of tree. It doesn't cause any
>> changes in behavior.
>>
>> Signed-off-by: Thomas Gleixner <[e-mail]>
>> Signed-off-by: Felipe Balbi <[e-mail]>
>
Sorry i missed this patch. Tony can you please
pick it up?

For the patch,
Acked-by: Santosh Shilimkar <[e-mail]>

Tony Lindgren Feb. 1, 2013, 10:38 p.m. UTC | #3

* Santosh Shilimkar <[e-mail]> [130124 23:54]:
> On Thursday 24 January 2013 11:52 PM, Felipe Balbi wrote:
> >Hi,
> >
> >On Fri, Dec 07, 2012 at 04:49:47PM +0200, Felipe Balbi wrote:
> >>From: Thomas Gleixner <[e-mail]>
> >>
> >>When applying RT patch on top of Linux, spinlocks are
> >>implemented as RT-mutexes, which means they are preemptable.
> >>
> >>Current GIC implementation on OMAP is using a spinlock
> >>to protect against preemption. As it turns out, we need
> >>to convert that lock into a raw_spinlock so that OMAP's
> >>interrupt controller works as expected after RT-patch
> >>is applied.
> >>
> >>This patch is simply to decrease the amount of changes
> >>RT-team needs to carry out of tree. It doesn't cause any
> >>changes in behavior.
> >>
> >>Signed-off-by: Thomas Gleixner <[e-mail]>
> >>Signed-off-by: Felipe Balbi <[e-mail]>
> >
> Sorry i missed this patch. Tony can you please
> pick it up?
> 
> For the patch,
> Acked-by: Santosh Shilimkar <[e-mail]>
> 

Looks like no comments, so applying this into
omap-for-v3.9/fixes-non-critical.

Regards,

Tony

Patch

diff --git a/arch/arm/mach-omap2/omap-wakeupgen.c b/arch/arm/mach-omap2/omap-wakeupgen.c
index 5d3b4f4..8633a43 100644
--- a/arch/arm/mach-omap2/omap-wakeupgen.c
+++ b/arch/arm/mach-omap2/omap-wakeupgen.c
@@ -46,7 +46,7 @@ 
 
 static void __iomem *wakeupgen_base;
 static void __iomem *sar_base;
-static DEFINE_SPINLOCK(wakeupgen_lock);
+static DEFINE_RAW_SPINLOCK(wakeupgen_lock);
 static unsigned int irq_target_cpu[MAX_IRQS];
 static unsigned int irq_banks = MAX_NR_REG_BANKS;
 static unsigned int max_irqs = MAX_IRQS;
@@ -134,9 +134,9 @@  static void wakeupgen_mask(struct irq_data *d)
 {
 	unsigned long flags;
 
-	spin_lock_irqsave(&wakeupgen_lock, flags);
+	raw_spin_lock_irqsave(&wakeupgen_lock, flags);
 	_wakeupgen_clear(d->irq, irq_target_cpu[d->irq]);
-	spin_unlock_irqrestore(&wakeupgen_lock, flags);
+	raw_spin_unlock_irqrestore(&wakeupgen_lock, flags);
 }
 
 /*
@@ -146,9 +146,9 @@  static void wakeupgen_unmask(struct irq_data *d)
 {
 	unsigned long flags;
 
-	spin_lock_irqsave(&wakeupgen_lock, flags);
+	raw_spin_lock_irqsave(&wakeupgen_lock, flags);
 	_wakeupgen_set(d->irq, irq_target_cpu[d->irq]);
-	spin_unlock_irqrestore(&wakeupgen_lock, flags);
+	raw_spin_unlock_irqrestore(&wakeupgen_lock, flags);
 }
 
 #ifdef CONFIG_HOTPLUG_CPU
@@ -189,7 +189,7 @@  static void wakeupgen_irqmask_all(unsigned int cpu, unsigned int set)
 {
 	unsigned long flags;
 
-	spin_lock_irqsave(&wakeupgen_lock, flags);
+	raw_spin_lock_irqsave(&wakeupgen_lock, flags);
 	if (set) {
 		_wakeupgen_save_masks(cpu);
 		_wakeupgen_set_all(cpu, WKG_MASK_ALL);
@@ -197,7 +197,7 @@  static void wakeupgen_irqmask_all(unsigned int cpu, unsigned int set)
 		_wakeupgen_set_all(cpu, WKG_UNMASK_ALL);
 		_wakeupgen_restore_masks(cpu);
 	}
-	spin_unlock_irqrestore(&wakeupgen_lock, flags);
+	raw_spin_unlock_irqrestore(&wakeupgen_lock, flags);
 }
 #endif