[PATCH 5/6] Unified AVR32/AT91 MCI Driver Support for a VCC Pin

[PATCH 5/6] Unified AVR32/AT91 MCI Driver Support for a VCC Pin

[Rob Emanuele]

Add support to the atmel-mci driver for a VCC pin in order to conserve
power if the MMC interface is powered down.

Please read the whole set, try it out, and comment.

Thank you,

Rob Emanuele

---
 drivers/mmc/host/atmel-mci.c |   33 ++++++++++++++++++++-------------
 include/linux/atmel-mci.h    |    2 ++
 2 files changed, 22 insertions(+), 13 deletions(-)

diff --git a/drivers/mmc/host/atmel-mci.c b/drivers/mmc/host/atmel-mci.c
index 1c1a4b3..a639b93 100644
--- a/drivers/mmc/host/atmel-mci.c
+++ b/drivers/mmc/host/atmel-mci.c
@@ -214,6 +214,8 @@ struct atmel_mci_slot {

 	int			detect_pin;
 	int			wp_pin;
+	int			vcc_pin;
+	bool			vcc_pin_act_low;

 	struct timer_list	detect_timer;
 };
@@ -926,23 +928,19 @@ static void atmci_set_ios(struct mmc_host *mmc,
struct mmc_ios *ios)
 	}

 	switch (ios->power_mode) {
+	case MMC_POWER_OFF:
+		if (gpio_is_valid(slot->vcc_pin))
+			gpio_set_value(slot->vcc_pin, slot->vcc_pin_act_low);
+		break;
 	case MMC_POWER_UP:
+		if (gpio_is_valid(slot->vcc_pin))
+			gpio_set_value(slot->vcc_pin, !slot->vcc_pin_act_low);
 		set_bit(ATMCI_CARD_NEED_INIT, &slot->flags);
 		break;
-	default:
-		/*
-		 * TODO: None of the currently available AVR32-based
-		 * boards allow MMC power to be turned off. Implement
-		 * power control when this can be tested properly.
-		 *
-		 * We also need to hook this into the clock management
-		 * somehow so that newly inserted cards aren't
-		 * subjected to a fast clock before we have a chance
-		 * to figure out what the maximum rate is. Currently,
-		 * there's no way to avoid this, and there never will
-		 * be for boards that don't support power control.
-		 */
+	case MMC_POWER_ON:
 		break;
+	default:
+		WARN_ON(1);
 	}
 }

@@ -1498,6 +1496,8 @@ static int __init atmci_init_slot(struct atmel_mci *host,
 	slot->host = host;
 	slot->detect_pin = slot_data->detect_pin;
 	slot->wp_pin = slot_data->wp_pin;
+	slot->vcc_pin = slot_data->vcc_pin;
+	slot->vcc_pin_act_low = slot_data->vcc_pin_act_low;
 	slot->sdc_reg = sdc_reg;

 	mmc->ops = &atmci_ops;
@@ -1534,6 +1534,13 @@ static int __init atmci_init_slot(struct atmel_mci *host,
 		}
 	}

+	if (gpio_is_valid(slot->vcc_pin)) {
+		if (gpio_request(slot->vcc_pin, "mmc_pow")) {
+			dev_dbg(&mmc->class_dev, "no power pin available\n");
+			slot->vcc_pin = -EBUSY;
+		}
+	}
+
 	host->slot[id] = slot;
 	mmc_add_host(mmc);

diff --git a/include/linux/atmel-mci.h b/include/linux/atmel-mci.h
index 2f1f957..3b4239a 100644
--- a/include/linux/atmel-mci.h
+++ b/include/linux/atmel-mci.h
@@ -24,6 +24,8 @@ struct mci_slot_pdata {
 	unsigned int		bus_width;
 	int			detect_pin;
 	int			wp_pin;
+	int			vcc_pin;
+	unsigned		vcc_pin_act_low:1;
 };

 /**
-- 
1.6.0.4

Re: [PATCH 5/6] Unified AVR32/AT91 MCI Driver Support for a VCC Pin

[Nicolas Ferre]

Rob Emanuele :
> Add support to the atmel-mci driver for a VCC pin in order to conserve
> power if the MMC interface is powered down.

Can you elaborate a bit more. In particular, I do not understand the
concept of "vcc_pin_act_low".

[..]

Thanks. Bye,
-- 
Nicolas Ferre

Re: [PATCH 5/6] Unified AVR32/AT91 MCI Driver Support for a VCC Pin

[Rob Emanuele]

Nicolas,

On Mon, Jun 15, 2009 at 9:20 AM, Nicolas Ferre<nicolas.ferre@atmel.com> wrote:
> Rob Emanuele :
>> Add support to the atmel-mci driver for a VCC pin in order to conserve
>> power if the MMC interface is powered down.
>
> Can you elaborate a bit more. In particular, I do not understand the
> concept of "vcc_pin_act_low".
>

Like the original at91_mci driver, this patch adds the ability to use
a pin to toggle 3.3V power going to the SD Card.  In my case the pin
is tied to a International Rectifier 6420 P-type MOSFET since a port
pin alone cannot drive the SD Card.  When I pull the vcc_pin low, 3.3V
can reach the SD Card to power it.  The "vcc_pin_act_low" (pin active
low) determines which way the pin should be pulled to enable power to
the SD Card.  If vcc_pin_act_low is true, when the card needs to get
powered up, it sets the vcc_pin low.  Other board designs may need a
pin driven high to power the SD Card.

Thanks,

Rob

Re: [PATCH 5/6] Unified AVR32/AT91 MCI Driver Support for a VCC Pin

[Michael Roth]

> diff --git a/drivers/mmc/host/atmel-mci.c b/drivers/mmc/host/atmel-mci.c
> index 1c1a4b3..a639b93 100644
> --- a/drivers/mmc/host/atmel-mci.c
> +++ b/drivers/mmc/host/atmel-mci.c
> @@ -214,6 +214,8 @@ struct atmel_mci_slot {
>
>  	int			detect_pin;
>  	int			wp_pin;
> +	int			vcc_pin;
> +	bool			vcc_pin_act_low;

Maybe, just 'vcc_active_low' is more clear?

Most other kernel code use 'active_low' or 'xyz_active_low' as far as I can 
see.

Re: [PATCH 5/6] Unified AVR32/AT91 MCI Driver Support for a VCC Pin

[Rob Emanuele]

That would be fine with me.

On Mon, Jun 15, 2009 at 2:35 PM, Michael Roth<mroth@nessie.de> wrote:
>
>> diff --git a/drivers/mmc/host/atmel-mci.c b/drivers/mmc/host/atmel-mci.c
>> index 1c1a4b3..a639b93 100644
>> --- a/drivers/mmc/host/atmel-mci.c
>> +++ b/drivers/mmc/host/atmel-mci.c
>> @@ -214,6 +214,8 @@ struct atmel_mci_slot {
>>
>>       int                     detect_pin;
>>       int                     wp_pin;
>> +     int                     vcc_pin;
>> +     bool                    vcc_pin_act_low;
>
> Maybe, just 'vcc_active_low' is more clear?
>
> Most other kernel code use 'active_low' or 'xyz_active_low' as far as I can
> see.
>
>