Probing of regulators can be a slow operation and can contribute to
slower boot times. This is especially true if a regulator is turned on
at probe time (with regulator-boot-on or regulator-always-on) and the
regulator requires delays (off-on-time, ramp time, etc).
While the overall kernel is not ready to switch to async probe by
default, as per the discussion on the mailing lists [1] it is believed
that the regulator subsystem is in good shape and we can move
regulator drivers over wholesale. There is no way to just magically
opt in all regulators (regulators are just normal drivers like
platform_driver), so we set PROBE_PREFER_ASYNCHRONOUS for all
regulators found in 'drivers/regulator' individually.
Given the number of drivers touched and the impossibility to test this
ahead of time, it wouldn't be shocking at all if this caused a
regression for someone. If there is a regression caused by this patch,
it's likely to be one of the cases talked about in [1]. As a "quick
fix", drivers involved in the regression could be fixed by changing
them to PROBE_FORCE_SYNCHRONOUS. That being said, the correct fix
would be to directly fix the problem that caused the issue with async
probe.
The approach here follows a similar approach that was used for the mmc
subsystem several years ago [2]. In fact, I ran nearly the same python
script to auto-generate the changes. The only thing I changed was to
search for "i2c_driver", "spmi_driver", and "spi_driver" in addition
to "platform_driver".
[1] https://lore.kernel.org/r/06db017f-e985-4434-8d1d-02ca2100cca0@sirena.org.uk
[2] https://lore.kernel.org/r/20200903232441.2694866-1-dianders@chromium.org/
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Link: https://lore.kernel.org/r/20230316125351.1.I2a4677392a38db5758dee0788b2cea5872562a82@changeid
Signed-off-by: Mark Brown <broonie@kernel.org>
Based on 1 normalized pattern(s):
this program is free software you can redistribute it and or modify
it under the terms of the gnu general public license as published by
the free software foundation either version 2 of the license or at
your option any later version
extracted by the scancode license scanner the SPDX license identifier
GPL-2.0-or-later
has been chosen to replace the boilerplate/reference in 3029 file(s).
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Allison Randal <allison@lohutok.net>
Cc: linux-spdx@vger.kernel.org
Link: https://lkml.kernel.org/r/20190527070032.746973796@linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Declare regulator_ops structure as const as it is only stored in the ops
field of a regulator_desc structure. This field is of type const, so
regulator_ops structures having this property can be made const too.
File size before: drivers/regulator/pcf50633-regulator.o
text data bss dec hex filename
2763 464 0 3227 c9b regulator/pcf50633-regulator.o
File size after: drivers/regulator/pcf50633-regulator.o
text data bss dec hex filename
3019 192 0 3211 c8b regulator/pcf50633-regulator.o
Signed-off-by: Bhumika Goyal <bhumirks@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Change the device name of the regulator function to the one chosen for
MODULE_ALIAS. This fixes kernel auto-module loading for the regulator function.
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Acked-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Mark Brown <broonie@linaro.org>
Use devm_regulator_register() to make cleanup paths simpler,
and remove unnecessary remove().
Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
Use the wrapper function for retrieving the platform data instead of
accessing dev->platform_data directly.
Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Mark Brown <broonie@linaro.org>
Since 0998d06310 "device-core: Ensure drvdata = NULL when no driver is bound",
this is done by driver core after device_release or on probe failure.
Thus we can remove all platform_set_drvdata(pdev, NULL) in drivers.
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
This driver can be converted to use linear_min_sel and
regulator_[map|list]_voltage_linear.
Below shows the equation (from Datasheet) for each LDOs.
For AUTOOUT:
VO(prog) = 0.625 + auto_out x 0.025 V; e.g.
(00000000 to 00101110: reserved)
00101111: 1.8 V (min)
01010011: 2.7 V
01101010: 3.275 V
01101011: 3.300 V
01101100: 3.325 V
01111111 : 3.800 V (max)
The linear mapping start from 0x2f selector.
Thus we convert this equation to:
VO(prog) = 1.8 + (selector - linear_min_sel) x 0.025 V
(min_uV = 1800000, uV_step = 25000, linear_min_sel = 0x2f)
For DOWNxOUT:
VO(prog) = 0.625 + downx_out x 0.025 V; e.g.
00000000 : 0.625 V (min)
00010111 : 1.200 V
00101111 : 1.800 V
01011111 : 3.000 V (max)
For xLDOOUT:
VO(prog) = 0.9 + xldo_out x 0.1 V; e.g.
00000: 0.9 V
00001: 1.0 V
11000 : 3.3 V
11011 : 3.6 V
Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
CONFIG_HOTPLUG is going away as an option so __devexit is no
longer needed.
Signed-off-by: Bill Pemberton <wfp5p@virginia.edu>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
CONFIG_HOTPLUG is going away as an option so __devinit is no longer
needed.
Signed-off-by: Bill Pemberton <wfp5p@virginia.edu>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
CONFIG_HOTPLUG is going away as an option so __devexit_p is no longer
needed.
Signed-off-by: Bill Pemberton <wfp5p@virginia.edu>
Signed-off-by: Mark 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: Mark Brown <broonie@opensource.wolfsonmicro.com>
Convert pcf50633 to get_voltage_sel and then we can remove
pcf50633_regulator_voltage_value function and move its implementation
to pcf50633_regulator_list_voltage.
Signed-off-by: Axel Lin <axel.lin@gmail.com>
Acked-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
The datasheet says 00000000 to 00101110 are reserved, and the min value of the
voltage setting is 1.8 V.
Thus don't write 0 to AUTO output voltage select register (address 1Ah).
Table 50. AUTOOUT - AUTO output voltage select register (address 1Ah) bit description[1]
Bit Symbol Access Description
7:0 auto_out R/W VO(prog) = 0.625 + auto_out × 0.025 V
eg. 00000000 to 00101110: reserved
00101111: 1.8 V (min)
01010011: 2.7 V
01101010: 3.275 V
01101011: 3.300 V
01101100: 3.325 V
01111111 : 3.800 V (max)
..... .....
11111110 : 3.800 V
11111111 : 3.800 V
This patch also fixes a bug in pcf50633_regulator_list_voltage:
In regulator core _regulator_do_set_voltage function:
if (rdev->desc->ops->set_voltage) {
ret = rdev->desc->ops->set_voltage(rdev, min_uV, max_uV,
&selector);
if (rdev->desc->ops->list_voltage)
selector = rdev->desc->ops->list_voltage(rdev,
selector);
else
selector = -1;
The list_voltage call here takes the selector got from set_voltage callback.
Thus adding 0x2f to the index in pcf50633_regulator_list_voltage looks wrong to me.
e.g.
If min_uV < 1.8V, pcf50633_regulator_set_voltage sets 0 to selector.
For this case, adding 0x2f to the index in pcf50633_regulator_list_voltage is correct.
However, if min_uV == 1.8V, pcf50633_regulator_set_voltage sets 0x2f to selector.
Adding 0x2f to the index in pcf50633_regulator_list_voltage in this case is wrong.
What this patch does is:
The minimal voltage setting for AUTOOUT is 0x2f.
Thus for the case min_uV < 1.8, set the voltage setting to 1.8V by writting
0x2f to AUTOOUT register and set selector = 0x2f.
So we don't write the rserved range to AUTOOUT register.
Which means the possible range of AUTOOUT register value is 0x2f ~ 0xff.
We have no problem in regulator_get_voltage.
Since we won't write 0~0x2e to AUTOOUT register, we have no problem converting
the bits we read to voltage. The equation in auto_voltage_value works fine.
For list_voltage, we need to take into account the case selector is 0 ~ 0x2e
because the regulator core assumes the selector is starting from 0.
This patch returns 0 for the cases selector is 0 ~ 0x2e, which means
"this selector code can't be used on this system".
The regulator core iterates from 0 to n_voltages to find the small voltage
in the specific range. The n_voltages settings for AUTOOUT should be 128 now,
including the reserved range of AUTOOUT.
Signed-off-by: Axel Lin <axel.lin@gmail.com>
Acked-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Current code has off-by-one n_voltage settings for AUTO/DOWN*/LDO* regulators.
Take ldo1 as example:
n_voltage should be (3.6 - 0.9) / 0.1 + 1 = 28
Table 76. LDO1OUT - LDO1 output voltage select register (address 2Dh) bit description[1]
Bit Symbol Access Description
4:0 ldo1_out R/W VO(prog) = 0.9 + ldo1_out × 0.1 V (max 3.6V); e.g.
00000 : 0.9 V
00001 : 1.0 V
11000 : 3.3 V
11011 : 3.6 V
11111 : 3.6 V
The n_voltage settings for HCLDO and MEMLDO are also wrong.
n_voltage for HCLDO and MEMLDO should be (3.6 - 0.9) / 0.1 + 1 = 28
Table 88. HCLDOOUT - HCLDO output voltage select register (addr. 39h) bit description[1]
Bit Symbol Access Description
4:0 hcldo_out R/W VO(prog) = 0.9 + hcldo_out × 0.1 V (max 3.6 V); e.g.
00000 : 0.9 V
00001 : 1.0 V
11011 : 3.6 V
11111 : 3.6 V
Table 62. MEMLDOOUT - MEMLDO o/p voltage select reg. (address 26h) bit description[1]
Bit Symbol Access Description
4:0 memldo_out R/W VO(prog) = 0.9 + memldo_out × 0.1 V; e.g.
00000: 0.9 V
00001: 1.0 V
11000 : 3.3 V
11011 : 3.6 V
11111 : 3.6 V
Signed-off-by: Axel Lin <axel.lin@gmail.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
With device tree support for regulators, its needed that the
regulator_dev->dev device has the right of_node attached.
To be able to do this add an additional parameter to the
regulator_register() api, wherein the dt-adapted driver can
then pass this additional info onto the regulator core.
Signed-off-by: Rajendra Nayak <rnayak@ti.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Change the interface used by set_voltage() to report the selected value
to the regulator core in terms of a selector used by list_voltage().
This allows the regulator core to know the voltage that was chosen
without having to do an explict get_voltage(), which would be much more
expensive as it will generally access hardware.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
Currently the pcf50633-regulator driver data is set to the pcf50633 core
structure, but the pcf50633-regulator remove handler assumes that it is set to
the regulator device. This patch fixes the issue by accessing the pcf506533
core structure through its parent device and setting the driver data to the
regulator device.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
This patch implements list_voltage for the pcf50644 regulator driver.
As the voltages are linearly scaled the code to convert register values to
voltages can be reused and most of the code can be shared with get_voltage.
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
Regulators need to be available early in init in order to allow them
to be available for consumers when requested. This is generally done
by registering them at subsys_initcall() time but not all regulator
drivers have done that. Convert these drivers to do so in order to
mimimise future support.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: Mike Rapoport <mike@compulab.co.il>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
Rather than having the regulator init data read from the platform_data
member of the struct device that is registered for the regulator make
the init data an explict argument passed in when registering. This
allows drivers to use the platform data for their own purposes if they
wish.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>
Changes from V1:
- Removed support for suspend_enable & suspend_disable functions.
Signed-off-by: Balaji Rao <balajirrao@openmoko.org>
Cc: Andy Green <andy@openmoko.com>
Cc: Liam Girdwood <lrg@slimlogic.co.uk>
Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Samuel Ortiz <sameo@openedhand.com>
