Bit 7 of the status register indicates that the chip is busy
doing a conversion. It does not indicate an alarm status.
Stop reporting it as alarm status bit.

Signed-off-by: Guenter Roeck <linux@roeck-us.net>

Tests with a real chip and a closer look into the datasheet reveals
that the local and remote critical alarm status bits are swapped for
MAX6680/MAX6681.

Signed-off-by: Guenter Roeck <linux@roeck-us.net>

Tests with a real chip and a closer look into the datasheet show that
MAX6654 does not support CRIT/THERM/OVERTEMP limits, so drop support
of the respective attributes for this chip.

Introduce LM90_HAVE_CRIT flag and use it to instantiate critical limit
attributes to solve the problem.

Cc: Josh Lehan <krellan@google.com>
Fixes: 229d495d ("hwmon: (lm90) Add max6654 support to lm90 driver")
Signed-off-by: Guenter Roeck <linux@roeck-us.net>

Commit b50aa496 ("hwmon: (lm90) Prevent integer underflows of
temperature calculations") addressed a number of underflow situations
when writing temperature limits. However, it missed one situation, seen
when an attempt is made to set the hysteresis value to MAX_LONG and the
critical temperature limit is negative.

Use clamp_val() when setting the hysteresis temperature to ensure that
the provided value can never overflow or underflow.

Fixes: b50aa496 ("hwmon: (lm90) Prevent integer underflows of temperature calculations")
Cc: Dmitry Osipenko <digetx@gmail.com>
Reviewed-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>

The detect function had a comment "Make compiler happy" when id did not
read the second configuration register. As it turns out, the code was
checking the contents of this register for manufacturer ID 0xA1 (NXP
Semiconductor/Philips), but never actually read the register. So it
wasn't surprising that the compiler complained, and it indeed had a point.
Fix the code to read the register contents for manufacturer ID 0xa1.

At the same time, the code was reading the register for manufacturer ID
0x41 (Analog Devices), but it was not using the results. In effect it was
just checking if reading the register returned an error. That doesn't
really add much if any value, so stop doing that.

Fixes: f90be42f ("hwmon: (lm90) Refactor reading of config2 register")
Signed-off-by: Guenter Roeck <linux@roeck-us.net>

Before commit 86585c61 ("hwmon: (pwm-fan) stop using legacy
PWM functions and some cleanups") pwm_apply_state() was called
unconditionally in pwm_fan_probe(). In this commit this direct
call was replaced by a call to __set_pwm(ct, MAX_PWM) which
however is a noop if ctx->pwm_value already matches the value to
set.
After probe the fan is supposed to run at full speed, and the
internal driver state suggests it does, but this isn't asserted
and depending on bootloader and pwm low-level driver, the fan
might just be off.
So drop setting pwm_value to MAX_PWM to ensure the check in
__set_pwm doesn't make it exit early and the fan goes on as
intended.

Cc: stable@vger.kernel.org
Fixes: 86585c61 ("hwmon: (pwm-fan) stop using legacy PWM functions and some cleanups")
Signed-off-by: Billy Tsai <billy_tsai@aspeedtech.com>
Reviewed-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Link: https://lore.kernel.org/r/20211130092212.17783-1-billy_tsai@aspeedtech.com


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

Per datasheet, SHT4x may need up to 8.2ms for a "high repeatability"
measurement to complete.  Attempting to read the result too early
triggers a NAK which then causes an EREMOTEIO error.

This behavior has been confirmed with a logic analyzer while running
the I2C bus at only 40kHz.  The low frequency precludes any
signal-integrity issues, which was also confirmed by the absence of
any CRC8 errors.  In this configuration, a NAK occurred on any read
that followed the measurement command within less than 8.2ms.

Signed-off-by: David Mosberger-Tang <davidm@egauge.net>
Link: https://lore.kernel.org/r/20211120212849.2300854-2-davidm@egauge.net


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

The first call to nct6775_asuswmi_read() in nct6775_wmi_read_value()
had been passing the full bank+register number instead of just the
lower 8 bits.  It didn't end up actually causing problems because the
second argument of that function is a u8 anyway, but it seems
preferable to be explicit about it at the call site (and consistent
with the rest of the code).

Signed-off-by: Zev Weiss <zev@bewilderbeest.net>
Fixes: 3fbbfc27 ("hwmon: (nct6775) Support access via Asus WMI")
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Link: https://lore.kernel.org/r/20211111025339.27520-1-zev@bewilderbeest.net


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

The removal function is called regardless of whether
/proc/i8k was created successfully or not, the later
causing a WARN() on module removal.
Fix that by only registering the removal function
if /proc/i8k was created successfully.

Tested on a Inspiron 3505.

Fixes: 039ae585 ("hwmon: Allow to compile dell-smm-hwmon driver without /proc/i8k")
Signed-off-by: Armin Wolf <W_Armin@gmx.de>
Acked-by: Pali Rohár <pali@kernel.org>
Link: https://lore.kernel.org/r/20211112171440.59006-1-W_Armin@gmx.de


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

sparse warnings: (new ones prefixed by >>)
>> drivers/hwmon/corsair-psu.c:536:82: sparse: sparse: Using plain
   integer as NULL pointer

Fixes: d115b51e ("hwmon: add Corsair PSU HID controller driver")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Wilken Gottwalt <wilken.gottwalt@posteo.net>
Link: https://lore.kernel.org/r/YY9hAL8MZEQYLYPf@monster.localdomain


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

Not all files in the kernel should include mm.h.  Migrating callers from
kmalloc to kvmalloc is easier if the kvmalloc functions are in slab.h.

[akpm@linux-foundation.org: move the new kvrealloc() also]
[akpm@linux-foundation.org: drivers/hwmon/occ/p9_sbe.c needs slab.h]

Link: https://lkml.kernel.org/r/20210622215757.3525604-1-willy@infradead.org


Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Acked-by: Pekka Enberg <penberg@kernel.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Now that we have all the shared memory region information populated in
the pcc_mbox_chan, let us propagate the pointer to the same as the
return value to pcc_mbox_request channel.

This eliminates the need for the individual users of PCC mailbox to
parse the PCCT subspace entries and fetch the shmem information. This
also eliminates the need for PCC mailbox controller to set con_priv to
PCCT subspace entries. This is required as con_priv is private to the
controller driver to attach private data associated with the channel and
not meant to be used by the mailbox client/users.

Let us convert all the users of pcc_mbox_{request,free}_channel to use
new interface.

Cc: Jean Delvare <jdelvare@suse.com>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
Acked-by: Wolfram Sang <wsa@kernel.org>
Acked-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Jassi Brar <jaswinder.singh@linaro.org>

Fix following coccicheck warning:
./drivers/hwmon/nct7802.c:1152:2-24: WARNING: Function
for_each_child_of_node should have of_node_put() before return.

Early exits from for_each_child_of_node should decrement the
node reference counter.

Signed-off-by: Wan Jiabing <wanjiabing@vivo.com>
Link: https://lore.kernel.org/r/20211029024918.5161-1-wanjiabing@vivo.com


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

TMP461 is almost identical to TMP451, which is already supported by the
lm90 driver. At the same time, unlike other sensors from the TMP401
compatible series, it only supports 8-bit temperature read operations,
and it supports negative temperatures when configured for its default
temperature range, and it supports a temperature offset register.
Supporting this chip in the tmp401 driver adds unnecessary complexity.
Remove its support from this driver and support the chip with the lm90
driver instead.

Fixes: 24333ac2 ("hwmon: (tmp401) use smb word operations instead of 2 smb byte operations")
Reported-by: David T. Wilson <david.wilson@nasa.gov>
Cc: David T. Wilson <david.wilson@nasa.gov>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>

TMP461 is almost identical to TMP451 and was actually detected as TMP451
with the existing lm90 driver if its I2C address is 0x4c. Add support
for it to the lm90 driver. At the same time, improve the chip detection
function to at least try to distinguish between TMP451 and TMP461.

As a side effect, this fixes commit 24333ac2 ("hwmon: (tmp401) use
smb word operations instead of 2 smb byte operations"). TMP461 does not
support word operations on temperature registers, which causes bad
temperature readings with the tmp401 driver. The lm90 driver does not
perform word operations on temperature registers and thus does not have
this problem.

Support is listed as basic because TMP461 supports a sensor resolution
of 0.0625 degrees C, while the lm90 driver assumes a resolution of 0.125
degrees C. Also, the TMP461 supports negative temperatures with its
default temperature range, which is not the case for similar chips
supported by the lm90 and the tmp401 drivers. Those limitations will be
addressed with follow-up patches.

Fixes: 24333ac2 ("hwmon: (tmp401) use smb word operations instead of 2 smb byte operations")
Reported-by: David T. Wilson <david.wilson@nasa.gov>
Cc: David T. Wilson <david.wilson@nasa.gov>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>

A flag indicating extended temperature support makes it easier
to add support for additional chips with this functionality.

Cc: David T. Wilson <david.wilson@nasa.gov>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>

ASUS ProArt X570-CREATOR WIFI board has got an nct6775 chip, but by default
there's no use of it because of resource conflict with WMI method.

This commit adds ProArt X570-CREATOR WIFI to the list of boards that can be
monitored using ASUS WMI.

BugLink: https://bugzilla.kernel.org/show_bug.cgi?id=204807


Signed-off-by: Denis Pauk <pauk.denis@gmail.com>
Tested-by: Joel Wirāmu Pauling <jwp@redhat.com>
Link: https://lore.kernel.org/r/20211025194748.6784-1-pauk.denis@gmail.com


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

This change allows LTD and RTD inputs to be configured via
device tree bindings. If the DT bindings are not present or
invalid, the input configuration is not modified and left at
HW defaults.

Signed-off-by: Oskar Senft <osk@google.com>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Link: https://lore.kernel.org/r/20211020164213.174597-2-osk@google.com


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

Eddie James authored 3 years ago and Joel Stanley committed 3 years ago

Save any FFDC provided by the OCC driver, and provide it to userspace
through a binary sysfs entry. Notify userspace pollers when there is an
error too.

Signed-off-by: Eddie James <eajames@linux.ibm.com>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Link: https://lore.kernel.org/r/20211019205307.36946-5-eajames@linux.ibm.com


Signed-off-by: Joel Stanley <joel@jms.id.au>

When setting the fan speed, i8k_set_fan() calls i8k_get_fan_status(),
causing an unnecessary SMM call since from the two users of this
function, only i8k_ioctl_unlocked() needs to know the new fan status
while dell_smm_write() ignores the new fan status.
Since SMM calls can be very slow while also making error reporting
difficult for dell_smm_write(), remove the function call from
i8k_set_fan() and call it separately in i8k_ioctl_unlocked().

Tested on a Dell Inspiron 3505.

Signed-off-by: Armin Wolf <W_Armin@gmx.de>
Reviewed-by: Pali Rohár <pali@kernel.org>
Link: https://lore.kernel.org/r/20211021190531.17379-6-W_Armin@gmx.de


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

i8k_config_data[] should only be used for applying device specific
quirks in case autoconfig does not work properly on certain
devices.

Signed-off-by: Armin Wolf <W_Armin@gmx.de>
Reviewed-by: Pali Rohár <pali@kernel.org>
Link: https://lore.kernel.org/r/20211021190531.17379-5-W_Armin@gmx.de


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

Returning -ENOIOCTLCMD gives the callers a better
hint of what went wrong and is the recommended
behavior.

Signed-off-by: Armin Wolf <W_Armin@gmx.de>
Acked-by: Pali Rohár <pali@kernel.org>
Link: https://lore.kernel.org/r/20211021190531.17379-4-W_Armin@gmx.de


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

Using strscpy_pad() allows for fewer memory accesses
since memset() will not unconditionally zero-out
the whole buffer.

Signed-off-by: Armin Wolf <W_Armin@gmx.de>
Acked-by: Pali Rohár <pali@kernel.org>
Link: https://lore.kernel.org/r/20211021190531.17379-3-W_Armin@gmx.de


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

Sort includes for better overview.

Signed-off-by: Armin Wolf <W_Armin@gmx.de>
Acked-by: Pali Rohár <pali@kernel.org>
Link: https://lore.kernel.org/r/20211021190531.17379-2-W_Armin@gmx.de


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

Fix following coccicheck warning:
./drivers/hwmon/tmp421.c:416:1-23: WARNING: Function
for_each_child_of_node should have of_node_put() before return.

Early exits from for_each_child_of_node should decrement the
node reference counter.

Signed-off-by: Wan Jiabing <wanjiabing@vivo.com>
Link: https://lore.kernel.org/r/20211018121538.16482-1-wanjiabing@vivo.com


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

When an spi driver's remove function returns a non-zero error code
nothing happens apart from emitting a generic error message. Make this
error message more device specific and return zero instead.

Acked-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Link: https://lore.kernel.org/r/20211012153945.2651412-5-u.kleine-koenig@pengutronix.de


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

In read_domain_devices(), acpi_bus_get_device() is called to obtain
the ACPI device object attached to the given ACPI handle and
subsequently that object is passed to get_device() for reference
counting, but there is a window between the acpi_bus_get_device()
and get_device() calls in which the ACPI device object in question
may go away.

To address this issue, make read_domain_devices() use
acpi_bus_get_acpi_device() to reference count and return the given
ACPI device object in one go and export that function to modules.

While at it, also make read_domain_devices() and
remove_domain_devices() use acpi_dev_put() instead of calling
put_device() directly on the ACPI device objects returned by
acpi_bus_get_acpi_device().

Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/11871063.O9o76ZdvQC@kreacher


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

The nominal speed of each fan can be obtained with
i8k_get_fan_nominal_speed(), however the result is not available
from userspace.
Change that by adding fanX_min, fanX_max and fanX_target attributes.
All are RO since fan control happens over pwm.

Tested on a Dell Inspiron 3505 and a Dell Latitude C600.

Signed-off-by: Armin Wolf <W_Armin@gmx.de>
Reviewed-by: Pali Rohár <pali@kernel.org>
Link: https://lore.kernel.org/r/20210926221044.14327-2-W_Armin@gmx.de


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

In case the DT contains some nodes not describing the input channels,
ignore them instead of exiting with error.

Signed-off-by: Krzysztof Adamski <krzysztof.adamski@nokia.com>
Link: https://lore.kernel.org/r/8e9e332b18dc2cf545f8e8255157e408d356f916.1634206677.git.krzysztof.adamski@nokia.com


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

Since the recent patches added possibility of disabling sensor channels
via DT, it only make sense to allow controlling that from userspace via
HWMON_T_ENABLE mechanism. This patches adds support for that.

Signed-off-by: Krzysztof Adamski <krzysztof.adamski@nokia.com>
Link: https://lore.kernel.org/r/a64c22e7323bd5a083f37aaaca91a745ac1beef3.1634206677.git.krzysztof.adamski@nokia.com


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

Recent patch added possibility to disable selected channels. That would
only make sure that the ENODATA is returned for those channels but would
not configure the actual hardware.

With this patch, the config register is written to make sure the
channels are disabled also at hardware level.

Signed-off-by: Krzysztof Adamski <krzysztof.adamski@nokia.com>
Link: https://lore.kernel.org/r/d451cacdf21bf8eff38a47c055aad8c0c6e8755a.1634206677.git.krzysztof.adamski@nokia.com


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

Previous patches added a way to specify some channel specific parameters
in DT and n-factor is definitely one of them. This calibration mechanism
is board specific as its value depends on the diodes/transistors being
connected to the sensor and thus the DT seems like a right fit for that
information. It is very similar to the value of shunt resistor that some
drivers allows specifying in DT.

This patch adds a possibility to set n-factor for each channel via
"n-factor" DT property in each channel subnode.

Signed-off-by: Krzysztof Adamski <krzysztof.adamski@nokia.com>
Link: https://lore.kernel.org/r/69d0bfcc5ba27c67f21d3eabfb100656a14c75b9.1634206677.git.krzysztof.adamski@nokia.com


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

The previous patch introduced per channel subnodes in DT that let us
specify some channel specific properties. This built a ground for easily
disabling individual channels of the sensor that may not be connected to
any external diode and thus are not returning any meaningful data.

This patch adds support for parsing the "status" property of channels DT
subnodes and makes sure the -ENODATA is returned when disabled channels
value is read.

Signed-off-by: Krzysztof Adamski <krzysztof.adamski@nokia.com>
Link: https://lore.kernel.org/r/a85d623f0792b862870933a875bdf802f4c017f1.1634206677.git.krzysztof.adamski@nokia.com


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

tmp42x is a multichannel temperature sensor with several external
channels. Since those channels can be used to connect diodes placed
anywhere in the system, their meaning will vary depending on the
project. For this case, the hwmon framework has an idea of labels which
allows us to assign the meaning to each channel.

The similar concept is already implemented in ina3221 - the label for
each channel can be defined via device tree. See commit a9e9dd9c
("hwmon: (ina3221) Read channel input source info from DT")

This patch adds support for similar feature to tmp421.

Signed-off-by: Krzysztof Adamski <krzysztof.adamski@nokia.com>
Link: https://lore.kernel.org/r/dab0fda6ac0e8d7f163c3762a7fb1e595a4d8159.1634206677.git.krzysztof.adamski@nokia.com


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

Eddie James authored 3 years ago and Joel Stanley committed 3 years ago

Checksumming of the request and sequence numbering is now done in the
OCC interface driver in order to keep unique sequence numbers. So
remove those in the hwmon driver. Also, add the command length to the
send_cmd function pointer, since the checksum must be placed in the
last two bytes of the command. The submit interface must receive the
exact size of the command - previously it could be rounded to the
nearest 8 bytes with no consequence.

Signed-off-by: Eddie James <eajames@linux.ibm.com>
Acked-by: Guenter Roeck <linux@roeck-us.net>
Link: https://lore.kernel.org/r/20210721190231.117185-3-eajames@linux.ibm.com


Signed-off-by: Joel Stanley <joel@jms.id.au>

This is a preparatory change. Upcoming patches will introduce more
per-channel parameters so it's worth organizing them into a struct.

Signed-off-by: Krzysztof Adamski <krzysztof.adamski@nokia.com>
Link: https://lore.kernel.org/r/baf34d95983a6b58a3e39e4c098e5979e541572e.1634029538.git.krzysztof.adamski@nokia.com


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

Up to now adt7x10_remove() returns zero unconditionally. Make it return
void instead which makes it easier to see in the callers that there is
no error to handle.

Also the return value of i2c and spi remove callbacks is ignored anyway.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Link: https://lore.kernel.org/r/20211011132754.2479853-3-u.kleine-koenig@pengutronix.de


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

sched.h and io.h are not used anywhere in dell-smm-hwmon.c.

Signed-off-by: Armin Wolf <W_Armin@gmx.de>
Link: https://lore.kernel.org/r/20210926221044.14327-3-W_Armin@gmx.de


Signed-off-by: Guenter Roeck <linux@roeck-us.net>

We have bool so use it consistently in all the drivers.

The following Coccinelle script was used:

@@
identifier T;
type t = { char, int };
@@
struct T {
...
-	t valid;
+	bool valid;
...
}

@@
identifier v;
@@
(
- v->valid = 0
+ v->valid = false
|
- v->valid = 1
+ v->valid = true
)

followed by sed to fixup the comments:
sed '/bool valid;/{s/!=0/true/;s/zero/false/}'

Few whitespace changes were fixed manually. All modified drivers were
compile-tested.

Signed-off-by: Paul Fertser <fercerpav@gmail.com>
Link: https://lore.kernel.org/r/20210924195202.27917-1-fercerpav@gmail.com


[groeck: Fixed up 'u8 valid' to 'boool valid' in atxp1.c]
Signed-off-by: Guenter Roeck <linux@roeck-us.net>

Use devm_hwmon_device_register_with_info() which will make thermal
framework work.

Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Link: https://lore.kernel.org/r/20211007125301.3030-1-o.rempel@pengutronix.de


Signed-off-by: Guenter Roeck <linux@roeck-us.net>