[RESEND] regulator: core: Disable unused regulators after deferred probing is done

Commit Message

Nishanth Menon April 23, 2014, 11:10 p.m. UTC

From: Saravana Kannan <[e-mail]>

regulator_init_complete does a scan of regulators which dont have
always-on or consumers are automatically disabled as being unused.
However, with deferred probing, late_initcall() is too soon to
declare a regulator as unused as the regulator itself might not
have registered due to defferal - Example: A regulator deffered due
to i2bus not available which in turn is deffered due to pinctrl
availability.

Since deferred probing is done in late_initcall(), do the cleanup of
unused regulators by regulator_init_complete in late_initcall_sync
instead of late_initcall.

Cc: Liam Girdwood <[e-mail]>
Cc: Mark Brown <[e-mail]>
Cc: Markus Pargmann <[e-mail]>
Signed-off-by: Saravana Kannan <[e-mail]>
[[e-mail]: minor rewording]
Signed-off-by: Nishanth Menon <[e-mail]>
---
Applies on v3.15-rc2 and on next-20140423
Original post: https://patchwork.kernel.org/patch/2545061/

I rediscovered this patch many times now :(:
Latest Example: https://patchwork.kernel.org/patch/4044811/
http://slexy.org/raw/s21MCHgeJo shows none of the unused regulators are disabled
and with this fix:
http://slexy.org/raw/s2TH36cThR (which auto disables unused regulators).

Ccying Markus as author of 66fda75f47dc (regulator: core: Replace
direct ops->disable usage) he might be interested as well.

Btw, this might open up a lot of broken boards - like we discovered
for DRA7 - so it might be a good idea for next instead of current rc
cyle giving it some time to cook and get platforms fixed.

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

Comments

Mark Brown April 24, 2014, 1:34 p.m. UTC | #1

On Wed, Apr 23, 2014 at 06:10:50PM -0500, Nishanth Menon wrote:
> From: Saravana Kannan <[e-mail]>
> 
> regulator_init_complete does a scan of regulators which dont have
> always-on or consumers are automatically disabled as being unused.
> However, with deferred probing, late_initcall() is too soon to
> declare a regulator as unused as the regulator itself might not
> have registered due to defferal - Example: A regulator deffered due
> to i2bus not available which in turn is deffered due to pinctrl
> availability.
> 
> Since deferred probing is done in late_initcall(), do the cleanup of
> unused regulators by regulator_init_complete in late_initcall_sync
> instead of late_initcall.

I'll apply this however since as we discussed on IRC last night it's
going to make the issues with unconfigured regulators getting powered
off more severe before I do so I'll change things so that regulators
with no configuration at all don't get touched by the core.

Nishanth Menon April 24, 2014, 2:18 p.m. UTC | #2

On 04/24/2014 08:34 AM, Mark Brown wrote:
> On Wed, Apr 23, 2014 at 06:10:50PM -0500, Nishanth Menon wrote:
>> From: Saravana Kannan <[e-mail]>
>>
>> regulator_init_complete does a scan of regulators which dont have
>> always-on or consumers are automatically disabled as being unused.
>> However, with deferred probing, late_initcall() is too soon to
>> declare a regulator as unused as the regulator itself might not
>> have registered due to defferal - Example: A regulator deffered due
>> to i2bus not available which in turn is deffered due to pinctrl
>> availability.
>>
>> Since deferred probing is done in late_initcall(), do the cleanup of
>> unused regulators by regulator_init_complete in late_initcall_sync
>> instead of late_initcall.
> 
> I'll apply this however since as we discussed on IRC last night it's
> going to make the issues with unconfigured regulators getting powered
> off more severe before I do so I'll change things so that regulators
> with no configuration at all don't get touched by the core.
> 
Thanks Mark.

Patch

diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c
index 9a09f3c..5a05de3 100644
--- a/drivers/regulator/core.c
+++ b/drivers/regulator/core.c
@@ -3867,4 +3867,4 @@  unlock:
 
 	return 0;
 }
-late_initcall(regulator_init_complete);
+late_initcall_sync(regulator_init_complete);