[RFC PATCH] regulator: core: do not disable regulator if boot_on is set

[RFC PATCH] regulator: core: do not disable regulator if boot_on is set

[Pramod Gurav]

Currently the regulator core disables the regulators which are unused
or whose reference count is zero or if they are configured always_on.
This change adds a check in this logic to see if a regulator is
configured as boot_on and does not disable it if found true.

Signed-off-by: Pramod Gurav <pramod.gurav@smartplayin.com>

---

The issue was found on apq8064 based IFC6410 on which a fixed regulator
configured as regulator-boot-on in DT and was being disabled when not in
use. Tested this change on this board and found working.

 drivers/regulator/core.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c
index cd87c0c..9f7a13f 100644
--- a/drivers/regulator/core.c
+++ b/drivers/regulator/core.c
@@ -4019,7 +4019,7 @@ static int __init regulator_init_complete(void)
 		ops = rdev->desc->ops;
 		c = rdev->constraints;
 
-		if (c && c->always_on)
+		if (c && (c->always_on || c->boot_on))
 			continue;
 
 		if (c && !(c->valid_ops_mask & REGULATOR_CHANGE_STATUS))
-- 
1.7.9.5

Re: [RFC PATCH] regulator: core: do not disable regulator if boot_on is set

[Lucas Stach]

Am Dienstag, den 25.11.2014, 16:23 +0530 schrieb Pramod Gurav:
> Currently the regulator core disables the regulators which are unused
> or whose reference count is zero or if they are configured always_on.
> This change adds a check in this logic to see if a regulator is
> configured as boot_on and does not disable it if found true.
> 
> Signed-off-by: Pramod Gurav <pramod.gurav@smartplayin.com>
> 
> ---
> 
> The issue was found on apq8064 based IFC6410 on which a fixed regulator
> configured as regulator-boot-on in DT and was being disabled when not in
> use. Tested this change on this board and found working.
> 
Um, why would this be the correct fix? regulator-boot-on just tells the
regulator core that the bootloader might have left this regulator
enabled. If you want it to stay on after the kernel finished init you
need to mark it as always-on.

>  drivers/regulator/core.c |    2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c
> index cd87c0c..9f7a13f 100644
> --- a/drivers/regulator/core.c
> +++ b/drivers/regulator/core.c
> @@ -4019,7 +4019,7 @@ static int __init regulator_init_complete(void)
>  		ops = rdev->desc->ops;
>  		c = rdev->constraints;
>  
> -		if (c && c->always_on)
> +		if (c && (c->always_on || c->boot_on))
>  			continue;
>  
>  		if (c && !(c->valid_ops_mask & REGULATOR_CHANGE_STATUS))

-- 
Pengutronix e.K.             | Lucas Stach                 |
Industrial Linux Solutions   | http://www.pengutronix.de/  |

Re: [RFC PATCH] regulator: core: do not disable regulator if boot_on is set

[Pramod Gurav]

On Tuesday 25 November 2014 04:27 PM, Lucas Stach wrote:
> Am Dienstag, den 25.11.2014, 16:23 +0530 schrieb Pramod Gurav:
>> Currently the regulator core disables the regulators which are unused
>> or whose reference count is zero or if they are configured always_on.
>> This change adds a check in this logic to see if a regulator is
>> configured as boot_on and does not disable it if found true.
>>
>> Signed-off-by: Pramod Gurav <pramod.gurav@smartplayin.com>
>>
>> ---
>>
>> The issue was found on apq8064 based IFC6410 on which a fixed regulator
>> configured as regulator-boot-on in DT and was being disabled when not in
>> use. Tested this change on this board and found working.
>>
> Um, why would this be the correct fix? regulator-boot-on just tells the
> regulator core that the bootloader might have left this regulator
> enabled. If you want it to stay on after the kernel finished init you
> need to mark it as always-on.

Thanks Lucas. Shall mark it as alway_on.
> 
>>  drivers/regulator/core.c |    2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c
>> index cd87c0c..9f7a13f 100644
>> --- a/drivers/regulator/core.c
>> +++ b/drivers/regulator/core.c
>> @@ -4019,7 +4019,7 @@ static int __init regulator_init_complete(void)
>>  		ops = rdev->desc->ops;
>>  		c = rdev->constraints;
>>  
>> -		if (c && c->always_on)
>> +		if (c && (c->always_on || c->boot_on))
>>  			continue;
>>  
>>  		if (c && !(c->valid_ops_mask & REGULATOR_CHANGE_STATUS))
> 

Re: [RFC PATCH] regulator: core: do not disable regulator if boot_on is set

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 762 bytes --]

On Tue, Nov 25, 2014 at 04:23:23PM +0530, Pramod Gurav wrote:
> Currently the regulator core disables the regulators which are unused
> or whose reference count is zero or if they are configured always_on.

No, it does *not* disable them if they are configured always_on (as the
code you're modifying shows).

> This change adds a check in this logic to see if a regulator is
> configured as boot_on and does not disable it if found true.

> -		if (c && c->always_on)
> +		if (c && (c->always_on || c->boot_on))
>  			continue;

This isn't what boot_on means.  It just means that the regulator is
expected to be enabled at initial power on, it doesn't mean it needs to
be enabled all the time.  Otherwise there'd be no point in having a
separate always_on flag.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

Re: [RFC PATCH] regulator: core: do not disable regulator if boot_on is set

[Pramod Gurav]

Hi Mark,

On Tuesday 25 November 2014 04:35 PM, Mark Brown wrote:
> On Tue, Nov 25, 2014 at 04:23:23PM +0530, Pramod Gurav wrote:
>> Currently the regulator core disables the regulators which are unused
>> or whose reference count is zero or if they are configured always_on.
> 
> No, it does *not* disable them if they are configured always_on (as the
> code you're modifying shows).

Yes, thats a typo.
> 
>> This change adds a check in this logic to see if a regulator is
>> configured as boot_on and does not disable it if found true.
> 
>> -		if (c && c->always_on)
>> +		if (c && (c->always_on || c->boot_on))
>>  			continue;
> 
> This isn't what boot_on means.  It just means that the regulator is
> expected to be enabled at initial power on, it doesn't mean it needs to
> be enabled all the time.  Otherwise there'd be no point in having a
> separate always_on flag.

Thanks for the comment. :-)
>