After commit 66d228a2bf ("regulator: core: Don't use regulators as
supplies until the parent is bound"), input supplies aren't resolved
if the input supplies parent device has not been bound. This prevent
regulators to hold an invalid reference if its supply parent device
driver probe is deferred.
But this causes issues on some boards where a PMIC's regulator use as
input supply a regulator from another PMIC whose driver is registered
after the driver for the former.
In this case the regulators for the first PMIC will fail to resolve
input supplies on regulators registration (since the other PMIC wasn't
probed yet). And when the core attempts to resolve again latter when
the other PMIC registers its own regulators, it will fail again since
the parent device isn't bound yet.
This will cause some parent supplies to never be resolved and wrongly
be disabled on boot due taking them as unused.
To solve this problem, also attempt to resolve the pending regulators
input supplies before disabling the unused regulators.
Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
It is allowed to call regulator_get with a NULL dev argument
(_regulator_get explicitly checks for it) but this causes an error later
when printing /sys/kernel/debug/regulator_summary.
Fix this by explicitly handling "deviceless" consumers in the debugfs code.
Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Cc: stable@vger.kernel.org
The code in _regulator_get() got a bit confusing over time, with control
flow jumping to a label from couple of places. Let's untangle it a bit by
doing the following:
1. Make handling of missing supplies and substituting them with dummy
regulators more explicit:
- check if we not have full constraints and refuse considering dummy
regulators with appropriate message;
- use "switch (get_type)" to handle different types of request explicitly
as well. "Normal" requests will get dummies, exclusive will not and
will notify user about that; optional will fail silently.
2. Stop jumping to a label in the middle of the function but instead have
proper conditional flow. I believe jumps should be reserved for error
handling, breaking from inner loop, or restarting a loop, but not for
implementing normal conditional flow.
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Instead of returning both regulator_dev structure as return value and
auxiliary error code in 'ret' argument, let's switch to using ERR_PTR
encoded values. This makes it more obvious what is going on at call sites.
Also, let's not unlock the mutex in the middle of a loop, but rather break
out and have single unlock path.
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
There is no need to have two loops there, we can store error for subsequent
reporting.
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Instead of separate "exclusive" and "allow_dummy" arguments, that formed 3
valid combinations (normal, exclusive and optional) and an invalid one,
let's accept explicit "get_type", like we did in devm-managed code.
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
There is no point in assigning value to 'ret' before calling
regulator_dev_lookup() as it will clobber 'ret' anyway.
Also, let's explicitly return -PROBE_DEFER when try_module_get() fails,
instead of relying that earlier initialization of "regulator" carries
correct value.
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
When regulators are successfully registered, we check to see if the
regulator is a supply for any other registered regulator and if so
add the new regulator as the supply for the existing regulator(s).
Some devices, such as Power Management ICs, may register a series of
regulators when probed and there are cases where one of the regulators
may fail to register and defer the probing of the parent device. In this
case any successfully registered regulators would be unregistered so
that they can be re-registered at some time later when the probe is
attempted again. However, if one of the regulators that was registered
was added as a supply to another registered regulator (that did not
belong to the same parent device), then this supply regulator was
unregister again because the parent device is probe deferred, then a
regulator could be holding an invalid reference to a supply regulator
that has been unregistered. This will lead to a system crash if that
regulator is then used.
Although it would be possible to check when unregistering a regulator
if any other regulator in the system is using it as a supply, it still
may not be possible to remove it as a supply if this other regulator is
in use. Therefore, fix this by preventing any regulator from adding
another regulator as a supply if the parent device for the supply
regulator has not been bound and if the parent device for the supply
and the regulator are different. This will allow a parent device that is
registering regulators to be probe deferred and ensure that none of the
regulators it has registered are used as supplies for any other
regulator from another device.
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Every function handling the mode within the regulator core uses an unsigned
int for mode, except for regulator_mode_constrain. This patch changes the
type of mode within regulator_mode_constrain which fixes several instances
where we are passing pointers to unsigned ints then treating them as an int
within this function.
Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Regulator consumers can receive event notifications when
errors are reported to the driver, but currently, there is
no way for a regulator consumer to know when the error is over.
To allow a regulator consumer to poll for error conditions
add a new API: regulator_get_error_flags.
Signed-off-by: Axel Haslam <ahaslam@baylibre.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
commit 73e705bf81 ("regulator: core: Add set_voltage_time op")
introduced a new rdev_warn() if the ramp_delay is 0.
Apparently, on omap3/twl4030 platforms with dynamic voltage
management this results in non-ending spurious messages like
[ 511.143066] VDD1: ramp_delay not set
[ 511.662322] VDD1: ramp_delay not set
[ 513.903625] VDD1: ramp_delay not set
[ 514.222198] VDD1: ramp_delay not set
[ 517.062835] VDD1: ramp_delay not set
[ 517.382568] VDD1: ramp_delay not set
[ 520.142791] VDD1: ramp_delay not set
[ 520.502593] VDD1: ramp_delay not set
[ 523.062896] VDD1: ramp_delay not set
[ 523.362701] VDD1: ramp_delay not set
[ 526.143035] VDD1: ramp_delay not set
I have observed this on GTA04 while it is reported to occur on
N900 as well: https://bugzilla.kernel.org/show_bug.cgi?id=178371
This patch makes the warning appear only in debugging mode.
Signed-off-by: H. Nikolaus Schaller <hns@goldelico.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
drms_uA_update() always returns failure when it cannot find regulator's
input voltage. But if hardware supports load configuration with
ops->set_load() and the input regulator isn't specified with valid reason
such as the input regulator is battery, not finding input voltage is
normal so such case should not return with an error.
Avoid such inadequate error return by checking input/output voltages
only when drms_uA_update() is about to configure load with enum based
ops->set_mode().
Cc: Liam Girdwood <lgirdwood@gmail.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Bjorn Andersson <bjorn.andersson@linaro.org>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Joonwoo Park <joonwoop@codeaurora.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
The new op is analogous to set_voltage_time_sel. It can be used by
regulators which don't have a table of discrete voltages. The function
returns the time for the regulator output voltage to stabilize after
being set to a new value, in microseconds. If the op is not set a
default implementation is used to calculate the delay.
This change also removes the ramp_delay calculation in the PWM
regulator, since the driver now uses the core code for the calculation
of the delay.
Signed-off-by: Matthias Kaehlcke <mka@chromium.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
The current code assumes that only the ramp_delay is used to determine
the time needed for the voltage to stabilize. This may be true for the
calculation done by regulator_set_voltage_time_sel(), however regulators
can implement their own set_voltage_time_sel() op which would be skipped
if no ramp delay is specified. Remove the check in
_regulator_do_set_voltage(), the functions calculating the ramp delay
return 0 anyway when the ramp delay is not configured.
Signed-off-by: Matthias Kaehlcke <mka@chromium.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
If the voltage can not be set jump to the end of the function. This
avoids having to check for an error multiple times and eliminates one
level of nesting in a follow-up change.
Signed-off-by: Matthias Kaehlcke <mka@chromium.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
The patch was based on my missinterpretation of the API and only
accidentally worked for me. Let's clean it out to not confuse others.
This reverts commit 3ff3f518a1.
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
There is little obvious use case for a regualtor driver to know if it is
possible to vary voltages at all by itself. If a consumer needs to
limit what voltages it tries to set based on the system configuration
then it will need to enumerate the possible voltages, and without that
even if it is possible to change voltages that doesn't mean that
constraints or other consumers will allow whatever change the driver is
trying to do at a given time. It doesn't even indicate if _set_voltage()
calls will work as noop _set_voltage() calls return success.
There were no users of this API that weren't abusing it and now they're
all gone so remove the API.
Signed-off-by: Mark Brown <broonie@kernel.org>
The call to set_machine_constraints() in regulator_register(), will
attempt to get the voltage for the regulator. If a regulator is in
bypass will fail to get the voltage (ie. it's bypass voltage) and
hence register the regulator, because the supply for the regulator has
not been resolved yet.
To fix this, add a call to regulator_resolve_supply() before we call
set_machine_constraints(). If the call to regulator_resolve_supply()
fails, rather than returning an error at this point, allow the
registration of the regulator to continue because for some regulators
resolving the supply at this point may not be necessary and it will be
resolved later as more regulators are added. Furthermore, if the supply
is still not resolved for a bypassed regulator, this will be detected
when we attempt to get the voltage for the regulator and an error will
be propagated at this point.
If a bypassed regulator does not have a supply when we attempt to get
the voltage, rather than returing -EINVAL, return -EPROBE_DEFER instead
to allow the registration of the regulator to be deferred and tried
again later.
Please note that regulator_resolve_supply() will call
regulator_dev_lookup() which may acquire the regulator_list_mutex. To
avoid any deadlocks we cannot hold the regulator_list_mutex when calling
regulator_resolve_supply(). Therefore, rather than holding the lock
around a large portion of the registration code, just hold the lock when
aquiring any GPIOs and setting up supplies because these sections may
add entries to the regulator_map_list and regulator_ena_gpio_list,
respectively.
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
To make the code more compat and centralized, this patch add a
unified function - regulator_ops_is_valid. So we can add
some extra checking code easily later.
Signed-off-by: WEN Pingbo <pingbo.wen@linaro.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
The public functions to acquire a regulator, such as regulator_get(),
internally look-up the regulator from the list of regulators that have
been registered with the regulator device class. The registration of
a new regulator with the regulator device class happens before the
regulator has been completely setup. Therefore, it is possible that
the regulator could be acquired before it has been setup successfully.
To avoid this move the device registration of the regulator to the end
of the regulator setup and update the error exit path accordingly.
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
During the resolution of a regulator's supply, we may attempt to enable
the supply if the regulator itself is already enabled. If enabling the
supply fails, then we will call _regulator_put() for the supply.
However, the pointer to the supply has not been cleared for the
regulator and this will cause a crash if we then unregister the
regulator and attempt to call regulator_put() a second time for the
supply. Fix this by clearing the supply pointer if enabling the supply
after fails when resolving the supply for a regulator.
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The function regulator_register_resolve_supply() is called from the
context of class_for_each_dev() (during the regulator registration) to
resolve any supplies added. regulator_register_resolve_supply() will
return an error if a regulator's supply cannot be resolved and this will
terminate the loop in class_for_each_dev(). This means that we will not
attempt to resolve any other supplies after one has failed. Hence, this
may delay the resolution of other regulator supplies until the failing
one itself can be resolved.
Rather than terminating the loop early, don't return an error code and
keep attempting to resolve any other supplies for regulators that have
been registered.
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
There are debugfs entries for voltage and current, but not for
the constraint flags. It's useful for debugging to be able to
see what these flags are so this patch adds a new debugfs file.
We can't use debugfs_create_bool for this because the flags are
bitfields, so as this needs a special read callback they have been
collected into a single file that lists all the flags.
Signed-off-by: Richard Fitzgerald <rf@opensource.wolfsonmicro.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
suspend_prepare can be called during regulator init time also, where
the mutex is not locked yet. This causes a false lockdep warning.
To avoid the problem, remove the lockdep assertion from the function
causing the issue. An alternative would be to lock the mutex during
init, but this would cause other problems (some APIs used during init
will attempt to lock the mutex also, causing deadlock.)
Signed-off-by: Tero Kristo <t-kristo@ti.com>
Reported-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
When we acquire a shareable enable GPIO on probe we do so with the
regulator_list_mutex held. However when we release the GPIOs we do this
immediately after dropping the mutex meaning that the list could become
corrupted. Move the release into the locked region to avoid this.
Signed-off-by: Mark Brown <broonie@kernel.org>
device_register() is calling ->get_voltage() as part of it's sysfs attribute
initialization process, and this functions might need to know the regulator
constraints to return a valid value.
This is at least true for the pwm regulator driver (when operating in
continuous mode) which needs to know the minimum and maximum voltage values
to calculate the current voltage:
min_uV + (((max_uV - min_uV) * dutycycle) / 100);
Move device_register() after set_machine_constraints() to make sure those
constraints are correctly initialized when ->get_voltage() is called.
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Reported-by: Stephen Barber <smbarber@chromium.org>
Tested-by: Caesar Wang <wxt@rock-chips.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
When a regulator is in bypass mode it is functioning as a switch
returning the voltage set in the regulator will not give the voltage
being output by the regulator as it's just passing through its supply.
This means that when we are getting the voltage from a regulator we
should check to see if it is in bypass mode and if it is we should
report the voltage from the supply rather than that which is set on the
regulator.
Reported-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
[treding@nvidia.com: return early for bypass mode]
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Commit 5e3ca2b349 ("regulator: Try to resolve regulators supplies on
registration") added a call to regulator_resolve_supply() within
regulator_register() where the regulator_list_mutex is held. This causes
a deadlock to occur on the Tegra114 Dalmore board when the palmas PMIC
is registered because regulator_register_resolve_supply() calls
regulator_dev_lookup() which may try to acquire the regulator_list_mutex
again.
Fix this by releasing the mutex before calling
regulator_register_resolve_supply() and update the error exit path to
ensure the mutex is released on an error.
[Made commit message more legible -- broonie]
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
This aids in debugging problems triggered by the regulator core applying
its constraints, we could potentially crash immediately after updating
the voltage if the constraints are buggy.
Signed-off-by: Mark Brown <broonie@kernel.org>
Commit 6261b06de5 ("regulator: Defer lookup of supply to regulator_get")
moved the regulator supplies lookup logic from the regulators registration
to the regulators get time.
Unfortunately, that changed the behavior of the regulator core since now a
parent supply with a child regulator marked as always-on, won't be enabled
unless a client driver attempts to get the child regulator during boot.
This patch tries to resolve the parent supply for the already registered
regulators each time that a new regulator is registered. So the regulators
that have child regulators marked as always on will be enabled regardless
if a driver gets the child regulator or not.
That was the behavior before the mentioned commit, since parent supplies
were looked up at regulator registration time instead of during child get.
Since regulator_resolve_supply() checks for rdev->supply, most of the times
it will be a no-op. Errors aren't checked to keep the possible out of order
dependencies which was the motivation for the mentioned commit.
Also, the supply being available will be enforced on regulator get anyways
in case the resolve fails on regulators registration.
Fixes: 6261b06de5 ("regulator: Defer lookup of supply to regulator_get")
Suggested-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Cc: <stable@vger.kernel.org> # 4.1+
Currently we only attempt to set the voltage during constraints
application if an exact voltage is specified. Extend this so that if
the currently set voltage for the regulator is outside the bounds set in
constraints we will move the voltage to the nearest constraint, raising
to the minimum or lowering to the maximum as needed. This ensures that
drivers can probe without the hardware being driven out of spec.
Reported-by: Ivaylo Dimitrov <ivo.g.dimitrov.75@gmail.com>
Tested-by: Ivaylo Dimitrov <ivo.g.dimitrov.75@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Apparently due to a wrongly resolved merge conflict between two
branches, which contained the same commit, the commit contents
partially was added two times in a row.
This change reverts the latter wrong inclusion of commit 909f7ee0b5
("regulator: core: Add support for active-discharge configuration").
The first applied commit 670666b9e0 ("regulator: core: Add support
for active-discharge configuration") is not touched.
Signed-off-by: Vladimir Zapolskiy <vladimir_zapolskiy@mentor.com>
Cc: Laxman Dewangan <ldewangan@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
The regulator_resolve_supply() function checks if a supply has been
associated with a regulator to avoid enabling it if that is not the
case.
But the supply was already looked up with regulator_resolve_supply()
and set with set_supply() before the check and both return on error.
So the fact that this statement has been reached means that neither
of them failed and a supply must be associated with the regulator.
Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Mark Brown