[PATCH 1/2] regulator: add regulator_is_dummy function

[PATCH 1/2] regulator: add regulator_is_dummy function

[Kevin Liu]

Introduce a regulator_is_dummy function to check whether the
regulator is dummy.

Signed-off-by: Kevin Liu <kliu5@marvell.com>
---
 drivers/regulator/core.c           |   17 +++++++++++++++++
 include/linux/regulator/consumer.h |    6 ++++++
 2 files changed, 23 insertions(+)

diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c
index e872c8b..d2a8bdb 100644
--- a/drivers/regulator/core.c
+++ b/drivers/regulator/core.c
@@ -1826,6 +1826,23 @@ int regulator_disable_regmap(struct regulator_dev *rdev)
 }
 EXPORT_SYMBOL_GPL(regulator_disable_regmap);
 
+/**
+ * regulator_is_dummy - check if regulator is dummy
+ * @regulator: regulator source
+ *
+ * Returns positive if the regulator is dummy, zero if it's not.
+ */
+int regulator_is_dummy(struct regulator *regulator)
+{
+	struct regulator_dev	*rdev = regulator->rdev;
+
+	if (!strcmp(rdev_get_name(rdev), "regulator-dummy"))
+		return 1;
+	else
+		return 0;
+}
+EXPORT_SYMBOL_GPL(regulator_is_dummy);
+
 static int _regulator_is_enabled(struct regulator_dev *rdev)
 {
 	/* A GPIO control always takes precedence */
diff --git a/include/linux/regulator/consumer.h b/include/linux/regulator/consumer.h
index c43cd35..fe549d0 100644
--- a/include/linux/regulator/consumer.h
+++ b/include/linux/regulator/consumer.h
@@ -164,6 +164,7 @@ int regulator_count_voltages(struct regulator *regulator);
 int regulator_list_voltage(struct regulator *regulator, unsigned selector);
 int regulator_is_supported_voltage(struct regulator *regulator,
 				   int min_uV, int max_uV);
+int regulator_is_dummy(struct regulator *regulator);
 int regulator_set_voltage(struct regulator *regulator, int min_uV, int max_uV);
 int regulator_set_voltage_time(struct regulator *regulator,
 			       int old_uV, int new_uV);
@@ -249,6 +250,11 @@ static inline int regulator_is_enabled(struct regulator *regulator)
 	return 1;
 }
 
+static inline int regulator_is_dummy(struct regulator *regulator)
+{
+	return 0;
+}
+
 static inline int regulator_bulk_get(struct device *dev,
 				     int num_consumers,
 				     struct regulator_bulk_data *consumers)
-- 
1.7.9.5

[PATCH 2/2] mmc: sdhci: fix dummy regulator issue

[Kevin Liu]

Should consider the case that dummy regulator is used for host->vmmc
and host->vqmmc and skip the regulator voltage check.

Signed-off-by: Kevin Liu <kliu5@marvell.com>
---
 drivers/mmc/host/sdhci.c |    6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c
index c13415c..561d126 100644
--- a/drivers/mmc/host/sdhci.c
+++ b/drivers/mmc/host/sdhci.c
@@ -2897,8 +2897,8 @@ int sdhci_add_host(struct sdhci_host *host)
 		}
 	} else {
 		regulator_enable(host->vqmmc);
-		if (!regulator_is_supported_voltage(host->vqmmc, 1700000,
-			1950000))
+		if ((regulator_is_supported_voltage(host->vqmmc, 1700000,
+			1950000) <= 0) && !regulator_is_dummy(host->vqmmc))
 			caps[1] &= ~(SDHCI_SUPPORT_SDR104 |
 					SDHCI_SUPPORT_SDR50 |
 					SDHCI_SUPPORT_DDR50);
@@ -2965,7 +2965,7 @@ int sdhci_add_host(struct sdhci_host *host)
 	}
 
 #ifdef CONFIG_REGULATOR
-	if (host->vmmc) {
+	if (host->vmmc && !regulator_is_dummy(host->vmmc)) {
 		ret = regulator_is_supported_voltage(host->vmmc, 2700000,
 			3600000);
 		if ((ret <= 0) || (!(caps[0] & SDHCI_CAN_VDD_330)))
-- 
1.7.9.5

Re: [PATCH 1/2] regulator: add regulator_is_dummy function

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 298 bytes --]

On Mon, Jan 07, 2013 at 01:38:50PM +0800, Kevin Liu wrote:
> Introduce a regulator_is_dummy function to check whether the
> regulator is dummy.

No, we've been through this repeatedly.  Whatever problem you're trying
to bodge around is going to be a problem with some real physical
regulators too.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [PATCH 1/2] regulator: add regulator_is_dummy function

[Kevin Liu]

2013/1/7 Mark Brown <broonie@opensource.wolfsonmicro.com>:
> On Mon, Jan 07, 2013 at 01:38:50PM +0800, Kevin Liu wrote:
>> Introduce a regulator_is_dummy function to check whether the
>> regulator is dummy.
>
> No, we've been through this repeatedly.  Whatever problem you're trying
> to bodge around is going to be a problem with some real physical
> regulators too.

Then the only way is MUST define the regulator rather than let dummy
regulator handle it, right?

Re: [PATCH 1/2] regulator: add regulator_is_dummy function

[Dr. Rainer Kaluscha]

Am 07.01.2013 12:25, schrieb Mark Brown:
> On Mon, Jan 07, 2013 at 01:38:50PM +0800, Kevin Liu wrote:
>> Introduce a regulator_is_dummy function to check whether the
>> regulator is dummy.
>
> No, we've been through this repeatedly.  Whatever problem you're trying
> to bodge around is going to be a problem with some real physical
> regulators too.

Though I don't know the previous discussion, this seems a different case to me.

The problem is that *only* a dummy regulator is available which refuses to do any regulation at 
all.
So sdhci/mmc should better behave as no regulator was available at all.
This may happen if e.g. in a precompiled kernel of some linux distro regulator support 
(including dummy) is enabled but the machine running that kernel doesn't have a known hardware 
regulator. This results in sdhci/mmc failing to enable the SD card reader though it would work 
well without regulator support in the kernel.

Rainer

Re: [PATCH 1/2] regulator: add regulator_is_dummy function

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 576 bytes --]

On Mon, Jan 07, 2013 at 06:36:36AM -0500, Kevin Liu wrote:
> 2013/1/7 Mark Brown <broonie@opensource.wolfsonmicro.com>:

> > No, we've been through this repeatedly.  Whatever problem you're trying
> > to bodge around is going to be a problem with some real physical
> > regulators too.

> Then the only way is MUST define the regulator rather than let dummy
> regulator handle it, right?

To repeat once again, dummy regulators should never be used in
production.  They are purely a crutch to help get boot going - it is
expected that things like this and cpufreq might fail.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [PATCH 1/2] regulator: add regulator_is_dummy function

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 751 bytes --]

On Mon, Jan 07, 2013 at 01:30:54PM +0100, Dr. Rainer Kaluscha wrote:

> The problem is that *only* a dummy regulator is available which
> refuses to do any regulation at all.
> So sdhci/mmc should better behave as no regulator was available at all.
> This may happen if e.g. in a precompiled kernel of some linux distro
> regulator support (including dummy) is enabled but the machine
> running that kernel doesn't have a known hardware regulator. This
> results in sdhci/mmc failing to enable the SD card reader though it
> would work well without regulator support in the kernel.

Production systems should not enable the dummy regulators, it's that
simple.  They're a crutch to get things booting but they're going to
cause problems in production.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]