Convert to devm_gpio_request to save a few error handling code.

This patch properly handle the gpio_request failure, we should return error
when gpio_request fails rather than just show warning.

I think one of the reason we got -EBUSY is because current code does not free
gpios in s5m8767_pmic_remove(). So it got -EBUSY when reload the module.

Yest another reason is in current code if gpio_request() returns error,
the rest of the code still calls gpio_direction_output to config buck_gpios
and buck_ds gpios. This looks wrong to me.
Signed-off-by: NAxel Lin <axel.lin@gmail.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

In current code, .list_voltage and .set_voltage_sel callbacks for BUCK7 and
BUCK8 return -EINVAL.

This patch adds s5m8767_buck78_ops for BUCK7 and BUCK8 which does not set
.list_voltage, .get_voltage_sel and .set_voltage_sel. ( This has the same
effect of returning -EINVAL in the callbacks)

Then for all the users of s5m8767_list_voltage, we don't need to worry about
the case reg_voltage_map[reg_id] is NULL.
So we can convert s5m8767_list_voltage to regulator_list_voltage_linear.
Signed-off-by: NAxel Lin <axel.lin@gmail.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

As s5m8767a is revisioned, ramp_delay register is changed.
5mV/uS, 10mV/uS, 25mV/uS, 50mV/uS, 100mV/uS ramp delay can be selected.
Signed-off-by: NSangbeom Kim <sbkim73@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

The S5M8767A can't know status of ap reset.
So, After AP watchdog reset, AP can't boot normally.

Problem can be happened like below condition.
- AP Bootable lowest voltage(vdd_arm): 0.9v
- AP DVFS voltage table: 0.8v, 0.9v, 1.0v
- During AP works on lowest voltage(0.8V), watchdog reset is asserted
- AP can't boot, because vdd arm is still 0.8v

Solution
- Basic concept:
  After ap watchdog reset, GPIO configuration is changed by default value
- S5M8767A has function of voltage control with gpio (8 levels with 3 gpios)
- Set bootable voltage on level 0 -> can work with default gpio configuration
- In the probing, Change voltage control level from level 0 to level 1
- Execute normal dvfs operation on level 1
- After watchdog reset, ap gpio is set by default value
- PMIC operation mode is changed by ap reset (level1 -> level0)
- Regardless of previous vdd_arm voltage, AP always can be booted.
Signed-off-by: NSangbeom Kim <sbkim73@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

If ramp_delay is 0, delay value can be divided by zero.
This patch can fix the problem.
Signed-off-by: NSangbeom Kim <sbkim73@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

To apply delay time for voltage change,
replace s5m8767_set_voltage with s5m8767_set_voltage_sel.
Signed-off-by: NSangbeom Kim <sbkim73@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

As s5m8767a is revisioned, remove unnecessary buck 2,3,4 initialization routine.
Signed-off-by: NSangbeom Kim <sbkim73@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

All the drivers that need delay for the regulator voltage output voltage to
stabilize after being enabled or after being set to a new value has been
converted to implement enable_time and set_voltage_time_sel callbacks.
Then regulator core will take care of the necessary delay.

For the drivers that don't need the delay, don't need to include linux/delay.h.
This patch removes the unneeded include of linux/delay.h in regulator drivers.
Signed-off-by: NAxel Lin <axel.lin@gmail.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Integer division may truncate the result.
Use DIV_ROUND_UP to ensure new voltage setting falls within specified range.

Also properly handle the case min_vol < desc->min to ensure we don't return
negative value for selector.
Signed-off-by: NAxel Lin <axel.lin@gmail.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Signed-off-by: NAxel Lin <axel.lin@gmail.com>
Acked-by: NSangbeom Kim <sbkim73@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

All the callback function implementation are the same for both LDOs and BUCKs.
Merge them to one s5m8767_ops.
Signed-off-by: NAxel Lin <axel.lin@gmail.com>
Acked-by: NSangbeom Kim <sbkim73@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

s5m8767_set_voltage not only implement set_voltage callbacks for LDOs,
but also for BUCKs when s5m8767->buck_gpioindex is not set.
s5m8767_set_voltage_buck actually only for buck[2|3|4] when
s5m8767->buck_gpioindex is set.
Conditionally calling s5m8767_set_voltage in s5m8767_set_voltage_buck makes
the code hard to read.

I think merging s5m8767_set_voltage_buck and s5m8767_set_voltage actually
simplifies things. It's easy to use buck234_vol pointer to differentiate if
we need to control voltage for buck[2|3|4] by DVS GPIOs.

This patch reworks s5m8767_set_voltage to support both LDOx and BUCKx in all
cases.
Signed-off-by: NAxel Lin <axel.lin@gmail.com>
Acked-by: NSangbeom Kim <sbkim73@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Rather than adding new arguments to regulator_register() every time we
want to add a new bit of dynamic information at runtime change the function
to take these via a struct. By doing this we avoid needing to do further
changes like the recent addition of device tree support which required each
regulator driver to be updated to take an additional parameter.

The regulator_desc which should (mostly) be static data is still passed
separately as most drivers are able to configure this statically at build
time.
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

S5M8767A has 4 operation mode for BUCK/LDOs.
This patch can handle opmode for s5m8767a.
Signed-off-by: NSangbeom Kim <sbkim73@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Current code does not really update the register with new value, fix it.
I rename the variable i to sel for better readability.
Signed-off-by: NAxel Lin <axel.lin@gmail.com>
Acked-by: NSangbeom Kim <sbkim73@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Check s5m8767->buck[2|3|4]_gpiodvs status in s5m8767_get_voltage_register
and return correct register accordingly.
Signed-off-by: NAxel Lin <axel.lin@gmail.com>
Acked-by: NSangbeom Kim <sbkim73@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

If we need to ensure only one of the buck[2|3|4]_gpiodvs can be specificed,
check them earlier.
Signed-off-by: NAxel Lin <axel.lin@gmail.com>
Acked-by: NSangbeom Kim <sbkim73@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

This function finds the smallest voltage that falls within the specified range,
and then returns the selector. This rename makes the intention more clear.

Also remove unneeded local variables min_vol and max_vol in s5m8767_set_voltage
and s5m8767_set_voltage_buck.
Signed-off-by: NAxel Lin <axel.lin@gmail.com>
Acked-by: NSangbeom Kim <sbkim73@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

In the s5m8767_set_voltage_time_sel function, divisor unit is wrong.
ramp_delay is usec unit. So 1000 should be multiplied.
Signed-off-by: NSangbeom Kim <sbkim73@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Signed-off-by: NAxel Lin <axel.lin@gmail.com>
Acked-by: NSangbeom Kim <sbkim73@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

This patch includes below fixes:
1. The mask variable is not used at all here, remove it.
2. We already have the new_sel and old_sel, simply returns the delay by:
  DIV_ROUND_UP(desc->step * (new_sel - old_sel), s5m8767->ramp_delay);
Signed-off-by: NAxel Lin <axel.lin@gmail.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: NSangbeom Kim <sbkim73@samsung.com>

Calling s5m8767_get_reg_id() is exactly the same as calling rdev_get_id().
It is pointless to add s5m8767_get_reg_id() function.
Use rdev_get_id() directly and remove s5m8767_get_reg_id() function.
Signed-off-by: NAxel Lin <axel.lin@gmail.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

The MODULE_DEVICE_TABLE will setup the modalias, thus adding
a MODULE_ALIAS for an entry already in s5m8767_pmic_id is redundant.
Signed-off-by: NAxel Lin <axel.lin@gmail.com>
Acked-by: NSangbeom Kim <sbkim73@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

Signed-off-by: NAxel Lin <axel.lin@gmail.com>
Acked-by: NSangbeom Kim <sbkim73@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>

S5M8767A is a cost-effective PMIC which is designed for mobile applications.
It includes high efficient 9 Buck converters, 28 LDOs.
Especially, S5M8767A is optimized for Multi-core SOCs.
And during DVFS operation, S5M8767A output stable voltage.
This patch implement regulator driver for S5M8767A.
Signed-off-by: NSangbeom Kim <sbkim73@samsung.com>
Signed-off-by: NMark Brown <broonie@opensource.wolfsonmicro.com>