[patch] collie: enable frontlight

[patch] collie: enable frontlight

[Pavel Machek]

Enable frontlight on collie, so that display can be actually read
indoors.

Signed-off-by: Pavel Machek <pavel@suse.cz>

---
commit dc507abbd6658927bdef7a521c4aa16c511ec62f
tree 666babf57b70cdf7fbac66c96defc9be4eb2518a
parent 8d04f63101774a56e80ef76b4d40ab782541dd94
author <pavel@amd.(none)> Wed, 02 Nov 2005 00:20:26 +0100
committer <pavel@amd.(none)> Wed, 02 Nov 2005 00:20:26 +0100

 arch/arm/common/locomo.c |   22 ++++++++++++++++++++++
 1 files changed, 22 insertions(+), 0 deletions(-)

diff --git a/arch/arm/common/locomo.c b/arch/arm/common/locomo.c
--- a/arch/arm/common/locomo.c
+++ b/arch/arm/common/locomo.c
@@ -631,6 +631,23 @@ static int locomo_resume(struct device *
 }
 #endif
 
+static spinlock_t fl_lock = SPIN_LOCK_UNLOCKED;
+
+#define LCM_ALC_EN	0x8000
+
+void frontlight_set(struct locomo *lchip, int duty, int vr, int bpwf)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&fl_lock, flags);
+	locomo_writel(bpwf, lchip->base + LOCOMO_FRONTLIGHT + LOCOMO_ALS);
+	udelay(100);
+	locomo_writel(duty, lchip->base + LOCOMO_FRONTLIGHT + LOCOMO_ALD);
+	locomo_writel(bpwf | LCM_ALC_EN, lchip->base + LOCOMO_FRONTLIGHT + LOCOMO_ALS);
+	spin_unlock_irqrestore(&fl_lock, flags);
+}
+
+
 /**
  *	locomo_probe - probe for a single LoCoMo chip.
  *	@phys_addr: physical address of device.
@@ -690,6 +707,11 @@ __locomo_probe(struct device *me, struct
 	/* FrontLight */
 	locomo_writel(0, lchip->base + LOCOMO_FRONTLIGHT + LOCOMO_ALS);
 	locomo_writel(0, lchip->base + LOCOMO_FRONTLIGHT + LOCOMO_ALD);
+
+	/* Same constants can be used for collie and poodle
+	   (depending on CONFIG options in original sharp code)? */
+	frontlight_set(lchip, 163, 0, 148);
+
 	/* Longtime timer */
 	locomo_writel(0, lchip->base + LOCOMO_LTINT);
 	/* SPI */

-- 
Thanks, Sharp!

Re: [patch] collie: enable frontlight

[Andrew Morton]

Pavel Machek <pavel@ucw.cz> wrote:
>
> +static spinlock_t fl_lock = SPIN_LOCK_UNLOCKED;
> +
> +#define LCM_ALC_EN	0x8000
> +
> +void frontlight_set(struct locomo *lchip, int duty, int vr, int bpwf)
> +{
> +	unsigned long flags;
> +
> +	spin_lock_irqsave(&fl_lock, flags);
> +	locomo_writel(bpwf, lchip->base + LOCOMO_FRONTLIGHT + LOCOMO_ALS);
> +	udelay(100);
> +	locomo_writel(duty, lchip->base + LOCOMO_FRONTLIGHT + LOCOMO_ALD);
> +	locomo_writel(bpwf | LCM_ALC_EN, lchip->base + LOCOMO_FRONTLIGHT + LOCOMO_ALS);
> +	spin_unlock_irqrestore(&fl_lock, flags);
> +}

ick, a 100 microsecond glitch, quite unnecessary.  Why not use a semaphore
here, if any locking is actually needed?  (It's called from device probe -
there's higher-level serialisation, no?)

Re: [patch] collie: enable frontlight

[Pavel Machek]

Hi!

> > +static spinlock_t fl_lock = SPIN_LOCK_UNLOCKED;
> > +
> > +#define LCM_ALC_EN	0x8000
> > +
> > +void frontlight_set(struct locomo *lchip, int duty, int vr, int bpwf)
> > +{
> > +	unsigned long flags;
> > +
> > +	spin_lock_irqsave(&fl_lock, flags);
> > +	locomo_writel(bpwf, lchip->base + LOCOMO_FRONTLIGHT + LOCOMO_ALS);
> > +	udelay(100);
> > +	locomo_writel(duty, lchip->base + LOCOMO_FRONTLIGHT + LOCOMO_ALD);
> > +	locomo_writel(bpwf | LCM_ALC_EN, lchip->base + LOCOMO_FRONTLIGHT + LOCOMO_ALS);
> > +	spin_unlock_irqrestore(&fl_lock, flags);
> > +}
> 
> ick, a 100 microsecond glitch, quite unnecessary.  Why not use a semaphore
> here, if any locking is actually needed?  (It's called from device probe -
> there's higher-level serialisation, no?)

Suspend (locomo_suspend) touches frontlight registers, too, and it
does it under spinlock => converting to semaphore would not too
simple. As it is only one glitch during probe, it should not hurt that
much...
								Pavel
-- 
Thanks, Sharp!

Re: [patch] collie: enable frontlight

[Pavel Machek]

Hi!

> Pavel Machek <pavel@ucw.cz> wrote:
> >
> > +static spinlock_t fl_lock = SPIN_LOCK_UNLOCKED;
> > +
> > +#define LCM_ALC_EN	0x8000
> > +
> > +void frontlight_set(struct locomo *lchip, int duty, int vr, int bpwf)
> > +{
> > +	unsigned long flags;
> > +
> > +	spin_lock_irqsave(&fl_lock, flags);
> > +	locomo_writel(bpwf, lchip->base + LOCOMO_FRONTLIGHT + LOCOMO_ALS);
> > +	udelay(100);
> > +	locomo_writel(duty, lchip->base + LOCOMO_FRONTLIGHT + LOCOMO_ALD);
> > +	locomo_writel(bpwf | LCM_ALC_EN, lchip->base + LOCOMO_FRONTLIGHT + LOCOMO_ALS);
> > +	spin_unlock_irqrestore(&fl_lock, flags);
> > +}
> 
> ick, a 100 microsecond glitch, quite unnecessary.  Why not use a semaphore
> here, if any locking is actually needed?  (It's called from device probe -
> there's higher-level serialisation, no?)

Oops, I see; I invented my own lock when I should have used
lchip->lock. Fixed locally, I'll push it up eventually.

							Pavel
-- 
Thanks, Sharp!