It is necessary to clear MC13892_SWITCHERS0_SWxHI bit when set voltage to the
voltage range from 1100000 to 1375000. Leaving MC13892_SWITCHERS0_SWxHI bit
untouched may result in wrong voltage setting.

For example, currently switch voltage from 1400000 to 1300000 will set the
voltage to 1800000 because the HI bit is still set.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@sirena.org.uk>

It's safe to call regulator_unregister() with NULL, thus remove the NULL test
before regulator_unregister() calls.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@sirena.org.uk>

Patch adds device tree probe support for mc13783-regulator driver.
Signed-off-by: NAlexander Shiyan <shc_work@mail.ru>
Signed-off-by: NMark Brown <broonie@sirena.org.uk>

This patch adds a warning about incorrect regulators instead of
printing the names of non-information message about the wrong amount.
Signed-off-by: NAlexander Shiyan <shc_work@mail.ru>
Signed-off-by: NMark Brown <broonie@sirena.org.uk>

Ensure max77686->opmode always has correct status.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@sirena.org.uk>

Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Ensure to unregister all regulators before return error in probe().

The regulator register order depends on the regulator ID pass to
ab3100_regulator_register() function. Thus we need to scan ab3100_regulator_desc
and find the index of successfully registered regulators, or alternatively just
call ab3100_regulators_remove() to unregister all registered regulators.

Since current code uses a static ab3100_regulators table, explicitly set
reg->rdev = NULL after regulator_unregister() call to ensure calling
ab3100_regulators_remove() in the unwind path always work.

Also move ab3100_regulators_remove() to avoid forward declaration.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

All regulators have ascendant voltage list in this driver.
Use regulator_map_voltage_ascend for them.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

All regulators have ascendant voltage list in this driver.
Use regulator_map_voltage_ascend for them.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

All regulators have ascendant voltage list in this driver.
Use regulator_map_voltage_ascend for them.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Both mc13892_sw1 and mc13892_sw voltage table have ascendant voltage list.
Use regulator_map_voltage_ascend for them.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

All regulators have ascendant voltage list in this driver.
Use regulator_map_voltage_ascend for them.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

tps65020 ldo1 and ldo2 vsel tables are identical, merge them.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

All regulators have ascendant voltage list in this driver.
Use regulator_map_voltage_ascend for them.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

It is necessary to clear MC13892_SWITCHERS0_SWxHI bit when set voltage to the
voltage range from 1100000 to 1375000. Leaving MC13892_SWITCHERS0_SWxHI bit
untouched may result in wrong voltage setting.

For example, currently switch voltage from 1400000 to 1300000 will set the
voltage to 1800000 because the HI bit is still set.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

This implements device tree support for the AB3100 regulators
driver. The initial settings are moved out of platform data
and into the driver for the device tree case, as it appears
that there is no way to supply this as AUXDATA for an I2C
device. The style and bindings are heavily inspired by
Lee Jones' style for AB8500.
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>
Acked-by: NLee Jones <lee.jones@linaro.org>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

This refactors the AB3100 regulator probe to use regulator IDs
and pass this to a separate registration function. This works
much smoother when migrating to device tree, as we can use a
match table with this regulator ID encoded in the .driver_data.
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Current code overrides control1 variable when setting ramp delay bits.
Fix it by just setting ramp_delay bits.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Acked-by: NLaxman Dewangan <ldewangan@nvidia.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Currently we have a special code in tps80031_ldo_set_voltage_sel() to handle
LDO2 track mode for TPS80031 or TPS80032-ES1.0. The purpose is to address below
issues:

Issue description:
- LDO2 traking mode is enabled
- LDO2 tracks SMPS2 voltage.
- LDO2 automatically switch-off when LDO2_CFG_VOLTAGE is changed to some discrete values (non exhaustive list):
            00011001, 00011010, 00011011, 00011100, .
- LDO2 switch-on again when LDO2_CFG_VOLTAGE is changed to other values (non exhaustive list):
        00011000, 00010111, .

LDOs have reserved codes. For these codes, LDO is switch-off.

In tracking, LDO2 ref comes from SMPS2.
However LDO2 enable is still gated by LDO2 VSEL decoding.
As a result, in tracking mode LDO2 will be disabled for following code (SMPS VSEL format):
000000 & 100000 (MSB not decoded)
011001 & 111001 (MSB not decoded)
011010 & 111010 (MSB not decoded)
011100 & 111100 (MSB not decoded)
011101 & 111101 (MSB not decoded)
011110 & 111110 (MSB not decoded)

However, current code has below bugs:
1. It uses regulator_list_voltage_linear, so list_voltage() still shows above
   invalid selectors have supported voltage.
2. Current code may return -EINVAL in tps80031_ldo_set_voltage_sel() for
   supported voltage. This is because when we use regulator_list_voltage_linear
   as list_voltage callback, regulator core will default use
   regulator_map_voltage_linear(). regulator_map_voltage_linear() has an
   assumption that the voltages are linear which is not true for this case.

   For example,
   when request voltage range is: min_uV=950000 uV && max_uV=1200000 uV
   regulator_map_voltage_linear() returns the selector is 29 (0x1D),
   set_voltage_sel() returns -EINVAL.
   (The selector is in invalid range, 0x19 ~ 0x1f).

   In above case, map_voltage() should find the lowest valid voltage within
   specific range (selector = 0x20) and set_voltage_sel() should successfully
   set the voltage to 987500 uV.

This patch fixes these issues by:
1. Add checking valid setting for LDO2 track mode of TPS80031 or TPS80032-ES1.0
   in list_voltage.  So it returns -EINVAL for invalid selectors.
2. Implement tps80031_ldo_map_voltage, use regulator_map_voltage_iterate()
   to find the lowest voltage within specific range for TPS80031 or
   TPS80032-ES1.0. This is required when the voltage map is no longer linear.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

All regulators have ascendant voltage list in this driver.
Thus use regulator_map_voltage_ascend is more efficient than the default
regulator_map_voltage_iterate.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

All regulators have ascendant voltage list in this driver.
Some regulators have more than 200 supported voltages.
e.g.
For TPS65910_REG_VDD1 and TPS65910_REG_VDD2:
n_voltages = VDD1_2_NUM_VOLT_FINE * VDD1_2_NUM_VOLT_COARSE
           = 73 * 3
           = 219

Thus it worth converting to regulator_map_voltage_ascend rather than use
default regulator_map_voltage_iterate.

For consistent, convert all regulators to regulator_map_voltage_ascend.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

All regulators have ascendant voltage list in this driver.
Thus use regulator_map_voltage_ascend is more efficient than the default
regulator_map_voltage_iterate.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Acked-by: NMarek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

All regulators have ascendant voltage list in this driver.
Thus use regulator_map_voltage_ascend is more efficient than the default
regulator_map_voltage_iterate.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Acked-by: NMarek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Simply the [get|set]_voltage_sel implementation by using
regulator_[set|get]_voltage_sel_regmap instead of open coded.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

According to the datasheet[1]

Register 0x06h Description: RmpCtrl (REGISTER ADDRESS: 0x06h Read/Write)
BIT[5..7]:
        RMP[2:0] Output voltage ramp timing
        D7-D5   Slope
        000     32mV/us
        001     16mV/us
        010     8mV/us
        ...
        110     0.5mV/us
        111     0.25mV/us

Thus to get correct ramp_ctrl value, we need to right-shift 5 bits.

[1] http://www.ti.com/lit/ds/symlink/tps62360.pdfSigned-off-by: NAxel Lin <axel.lin@ingics.com>
Acked-by: NLaxman Dewangan <ldewangan@nvidia.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Datasheet says:
Time Step (TSTEP) selection, when changing the output voltage, the new value is
reached through successive voltage steps (if not bypassed). The equivalent
programmable slew rate of the output voltage is:

TSTEP[1:0]: 00 Jump (no slope control)
TSTEP[1:0]: 01 10mV/us
TSTEP[1:0]: 10 5mV/us (default)
TSTEP[1:0]: 11 2.5mV/us
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Acked-by: NLaxman Dewangan <ldewangan@nvidia.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

The ctrl_addr setting for SMPS10 is missed, thus palmas_smps_init() read/write
wrong register for SMPS10 in current code. Fix it.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Acked-by: NLaxman Dewangan <ldewangan@nvidia.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

SMPS3 and SMPS7 do not have tstep_addr setting, so current code actually
writes 0 to smps12_ctl (offset is 0) register when set_ramp_delay callback
is called for SMPS3 and SMPS7.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Acked-by: NLaxman Dewangan <ldewangan@nvidia.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

When SMPS45 is set to off and LDO8 tracking is enabled, the output voltage can
be set from 0.45 to 1.65 V. Thus set min_uV to be 450000.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Acked-by: NLaxman Dewangan <ldewangan@nvidia.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

The voltage_shift can be calculated from voltage_mask.
Let's remove voltage_shift fied from struct ab8500_regulator_info, this change
can prevent missing voltage_shift setting issue.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Acked-by: NBengt Jonsson <bengt.g.jonsson@stericsson.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

The voltage_shift field of struct expand_register is not used now, remove it.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Acked-by: NBengt Jonsson <bengt.g.jonsson@stericsson.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

The Palma device like TPS65913 have the mode mask which is also
used for enable/disable the rails. The mode bits are defined as
	00: OFF
	01: AUTO
	10: ECO
	11: Forced PWM

and modes are set accordingly as
	REGULATOR_MODE_NORMAL: AUTO
	REGULATOR_MODE_IDLE: ECO
	REGULATOR_MODE_FAST: PWM

Two issue observed:
1. If client calls following sequence:
	regulator_enable(),
	regulator_set_mode(FAST),
	regulator_disable()

	and again the regulator_enable() then the mode is reset
	to NORMAL inplace of keeping the mode as FAST.

	Fixing this by storing the current mode configured by client
	and restoring modes when enable() is called after disable().

2. In following sequence, the regulator get enabled:
	regulator_disable()
	regulator_set_mode(FAST),

	Fixing this by updating new mode in register only if it is
	enabled.
Signed-off-by: NLaxman Dewangan <ldewangan@nvidia.com>
Acked-by: NGraeme Gregory <gg@slimlogic.co.uk>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Currently Palma regulator driver support the ramp delay
through rail specific platform data.

As regulator framework support the configuration of ramp
delay through regulator constraint, using the framework
method and removing the platform specific data approach.
Signed-off-by: NLaxman Dewangan <ldewangan@nvidia.com>
Acked-by: NGraeme Gregory <gg@slimlogic.co.uk>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

A lot of regulator hardware has ascendant voltage list.
This patch adds regulator_map_voltage_ascend() and export it.

Drivers that have ascendant voltage list can use this as their map_voltage()
operation, this is more efficient than default regulator_map_voltage_iterate()
function.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

When setting voltage for AB8540_LDO_AUX3, current code only updates one of
info->voltage_reg and info->expand_register registers which is wrong.
To ensure we set to correct voltage, it always needs to clear or set
expand_register.voltage_mask bit of expand_register.

The function of the expand register bit is the following (from the user manual):
0: VAUX3 output voltage is determined by Vaux3Sel bit settings in register
   VldoCVaux3Sel
1: VAUX3 output voltage is set to 3.05V regardless of Vaux3Sel settings in
   register VldoCVaux3Sel (VldoCVaux3Sel is the register at 0x0421)

So when going to 3.05V, set the expand register bit.
When leaving 3.05V for another voltage, set the target voltage before clearing
the expand register bit.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Acked-by: NBengt Jonsson <bengt.g.jonsson@stericsson.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Both enable() and disable() functions have only one caller, thus remove them.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Acked-by: NBengt Jonsson <bengt.g.jonsson@stericsson.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

This ensures info->update_val status is still correct if set_mode() call fails.
Otherwise, get_mode() may return wrong status if a set_mode() call fails.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Acked-by: NBengt Jonsson <bengt.g.jonsson@stericsson.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Currently, set invalid mode setting for shared mode regulators may change
sm->lp_mode_req flag. This patch ensures we don't set lp_mode_req flag to wrong
status if set_mode() fails.

This patch includes some clean up, and these changes makes this patch looks like
code refactor. The clean up is mainly to avoid adding ugly code to handle
failure paths.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Acked-by: NBengt Jonsson <bengt.g.jonsson@stericsson.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

This patch ensures info->update_val is consistent with current register value.
Signed-off-by: NAxel Lin <axel.lin@ingics.com>
Acked-by: NBengt Jonsson <bengt.g.jonsson@stericsson.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>