Replace devm_add_action() with devm_add_action_or_reset(),
and check its return value.
Reviewed-by: NJean Delvare <jdelvare@suse.de>
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>

Coverity reports the following error.

>>>     CID 1364474:  Error handling issues  (CHECKED_RETURN)
>>>     Calling "lm90_read_reg" without checking return value (as is done
>>>     elsewhere 28 out of 29 times).
532             lm90_read_reg(client, LM90_REG_R_REMOTE_LOWH);
533             if (val < 0)
534                     return val;

Fixes: 10bfef47 ("hwmon: (lm90) Read limit registers only once")
Reviewed-by: NJean Delvare <jdelvare@suse.de>
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>

checkpatch rightfully complains that else after return is unnecessary.
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>

Use bool for valid flag and leave it up to the compiler to find
an optimal representation.
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>

Read limit registers only once at startup or after errors to improve
driver performance.
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>

Return both error code and register value as return code from
read functions, and always check for errors.

This reduces code size on x86_64 by more than 1k while at
the same time improving error resiliency.
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>

Since all other cleanup handled with devm_add_action, we can use
devm_hwmon_device_register_with_groups() to register the hwmon
device, and drop the remove function.
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>

Use devm_add_action where possible to simplify error handling and
cleanup on remove.
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>

.alert() is meant to be generic, but there is currently no way
for the device driver to know which protocol generated the alert.
Add a parameter in .alert() to help the device driver to understand
what is given in data.

This patch is required to have the support of SMBus Host Notify protocol
through .alert().
Tested-by: NAndrew Duggan <aduggan@synaptics.com>
For hwmon:
Acked-by: NGuenter Roeck <linux@roeck-us.net>
For IPMI:
Acked-by: NCorey Minyard <cminyard@mvista.com>
Signed-off-by: NBenjamin Tissoires <benjamin.tissoires@redhat.com>
Signed-off-by: NWolfram Sang <wsa@the-dreams.de>

The code handles this variable always as unsigned, so adapt the type.
Signed-off-by: NWolfram Sang <wsa@the-dreams.de>
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>

Simplify code, reduce code size, and attach hwmon attributes to hwmon device.
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Signed-off-by: NJean Delvare <jdelvare@suse.de>

Create all sysfs groups in one call by using sysfs_create_groups
instead of calling sysfs_create_group individually for each group.
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Signed-off-by: NJean Delvare <jdelvare@suse.de>

Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Signed-off-by: NJean Delvare <jdelvare@suse.de>

With the new hwmon API, all attributes have to be created as groups.
Use sysfs_create_group and sysfs_remove_group instead of device_create_file
and device_remove_file to prepare for the new API.

Exception is the 'pec' attribute which will stay with the i2c device.
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Signed-off-by: NJean Delvare <jdelvare@suse.de>

Signed-off-by: NJean Delvare <khali@linux-fr.org>

Commit 109b1283 (hwmon: (lm90) Add support to handle IRQ) introduced
interrupt support. Its error handling code fails to unregister the already
registered hwmon device.

Fixes: 109b1283Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Signed-off-by: NJean Delvare <khali@linux-fr.org>

The device lm90 can be controlled by the vcc rail.
Adding the regulator support to power on/off the vcc rail.
Enable the "vcc" regulator before accessing the device.

[JD: Rename variables to avoid confusion with registers.]
Signed-off-by: NWei Ni <wni@nvidia.com>
Signed-off-by: NJean Delvare <khali@linux-fr.org>

TI TMP451 is mostly compatible with ADT7461, except for
local temperature low byte and max conversion rate.
Add support to the LM90 driver.
Signed-off-by: NWei Ni <wni@nvidia.com>
Reviewed-by: NGuenter Roeck <linux@roeck-us.net>
Signed-off-by: NJean Delvare <khali@linux-fr.org>

Using enums for the indexes and nrs of temp8 and temp11.
This make the code much more readable.
Signed-off-by: NWei Ni <wni@nvidia.com>
Signed-off-by: NJean Delvare <khali@linux-fr.org>

When the temperature exceed the limit range value,
the driver can handle the interrupt.
Signed-off-by: NWei Ni <wni@nvidia.com>
Signed-off-by: NJean Delvare <khali@linux-fr.org>

Add bit defines for the status register. And add a function
lm90_is_tripped() which will read status register and return
tripped or not, then lm90_alert can call it directly, and in the
future the IRQ thread also can use it.

[JD: Adjusted to include all the new MAX6696 status flags.]
Signed-off-by: NWei Ni <wni@nvidia.com>
Signed-off-by: NJean Delvare <khali@linux-fr.org>

Bit 2 of status register 2 on MAX6696 (external diode 2 open)
sets ALERT; the bit thus has to be listed in alert_alarms.
Also display a message in the alert handler if the condition
is encountered.

Even though not all overtemperature conditions cause ALERT
to be set, we should not ignore them in the alert handler.
Display messages for all out-of-range conditions.
Reported-by: NJean Delvare <khali@linux-fr.org>
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Cc: stable@vger.kernel.org
Signed-off-by: NJean Delvare <khali@linux-fr.org>

time_after (as opposed to time_after_equal) already ensures that the
cache lifetime is at least as much as requested. There is no point in
manually adding another jiffy to that value, and this can confuse the
reader into wrong interpretation.
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Cc: Imre Deak <imre.deak@intel.com>

SENSORS_LIMIT and the generic clamp_val have the same functionality,
and clamp_val is more efficient.

This patch reduces text size by 9052 bytes and bss size by 11624 bytes
for x86_64 builds.
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Acked-by: NGeorge Joseph <george.joseph@fairview5.com>
Acked-by: NJean Delvare <khali@linux-fr.org>

Convert to use devm_ functions to reduce code size and simplify the code.

Cc: Jean Delvare <khali@linux-fr.org>
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Acked-by: NJean Delvare <khali@linux-fr.org>

The lm90 driver restores the original chip configuration in its exit function.
However, the chip configuration is not restored if the probe function fails.
Restore it there as well.
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Signed-off-by: NJean Delvare <khali@linux-fr.org>

GMT G781 is a ADM1032-compatible temperature sensor chip.
Add support to the LM90 driver.

Cc: Mike Gorchak <lestat@i.com.ua>
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Signed-off-by: NJean Delvare <khali@linux-fr.org>

Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Signed-off-by: NJean Delvare <khali@linux-fr.org>

This patch converts the drivers in drivers/hwmon/* to use the
module_i2c_driver() macro which makes the code smaller and a bit simpler.
Signed-off-by: NAxel Lin <axel.lin@gmail.com>
Cc: Corentin Labbe <corentin.labbe@geomatys.fr>
Cc: Dirk Eibach <eibach@gdsys.de>
Cc: "Mark M. Hoffman" <mhoffman@lightlink.com>
Cc: Steve Glendinning <steve.glendinning@smsc.com>
Cc: Riku Voipio <riku.voipio@iki.fi>
Cc: Guillaume Ligneul <guillaume.ligneul@gmail.com>
Cc: David George <david.george@ska.ac.za>
Cc: "Hans J. Koch" <hjk@hansjkoch.de>
Cc: Marc Hulsman <m.hulsman@tudelft.nl>
Cc: Rudolf Marek <r.marek@assembler.cz>
Signed-off-by: NGuenter Roeck <guenter.roeck@ericsson.com>

When writing the update_interval attribute, the parameter value was
not range checked, which could cause an integer overflow and result
in an arbitrary update interval. Fix by limiting the value range to
<0, 100000>.
Reported-by: NJean Delvare <khali@linux-fr.org>
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Signed-off-by: NJean Delvare <khali@linux-fr.org>

replaced strict_strtol with kstrtol and
replaced strict_strtuol with kstrtuol

This satisfies checkpatch -f
Compile tested only: no warnings or errors given
Signed-off-by: NFrans Meulenbroeks <fransmeulenbroeks@gmail.com>
Signed-off-by: NGuenter Roeck <guenter.roeck@ericsson.com>

Clean up the code to make it more readable:
* Remove reg_ and new_ prefixes from variable names, they made the
  names longer, causing extra line breaks, while not adding much
  value.
* Introduce struct device dev* = &client->dev in two functions, to
  avoid repeating client->dev everywhere in these functions.
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Acked-by: NGuenter Roeck <guenter.roeck@ericsson.com>

With some configuration option combinations, we get the following
warnings:

drivers/hwmon/lm90.c: In function 'lm90_detect':
drivers/hwmon/lm90.c:1114: warning: 'chip_id' may be used uninitialized
in this function
drivers/hwmon/lm90.c:1114: warning: 'reg_config1' may be used
uninitialized in this function
drivers/hwmon/lm90.c:1114: warning: 'reg_convrate' may be used
uninitialized in this function
drivers/hwmon/lm90.c:1187: warning: 'reg_emerg2' may be used
uninitialized in this function
drivers/hwmon/lm90.c:1187: warning: 'reg_status2' may be used
uninitialized in this function

We can solve these easily by reading the register values first and
checking for errors later. These errors should be very rare, even in
the case of failed detection, so this change has no impact on
performance. And this makes checkpatch.pl happier.
Reported-by: NGuenter Roeck <guenter.roeck@ericsson.com>
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Acked-by: NGuenter Roeck <guenter.roeck@ericsson.com>

Several vendors implement a second configuration register, which we
check during device detection. Refactor the code to avoid duplication.
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Cc: Stijn Devriendt <sdevrien@cisco.com>
Cc: Guenter Roeck <guenter.roeck@ericsson.com>
Signed-off-by: NGuenter Roeck <guenter.roeck@ericsson.com>

With a device ID register value of 0, the SA56004 detection is rather
weak. Check several other register too to confirm the detection, as we
do for other supported devices.
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Cc: Stijn Devriendt <sdevrien@cisco.com>
Cc: Guenter Roeck <guenter.roeck@ericsson.com>
Acked-by: NStijn Devriendt <sdevrien@cisco.com>
Signed-off-by: NGuenter Roeck <guenter.roeck@ericsson.com>

The optional extended local temperature register can never have
address 0, as this address is already used by another register. Thus
we can get rid of flag LM90_HAVE_LOCAL_EXT and simply rely on
reg_local_ext being non-zero to determine if a given chip has this
extension or not. This makes the code more simple.
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Cc: Stijn Devriendt <sdevrien@cisco.com>
Cc: Guenter Roeck <guenter.roeck@ericsson.com>
Signed-off-by: NGuenter Roeck <guenter.roeck@ericsson.com>

Add support for Philips SA56004, an LM86 compatible temperature sensor.
Signed-off-by: NStijn Devriendt <sdevrien@cisco.com>
Signed-off-by: NGuenter Roeck <guenter.roeck@ericsson.com>

This patch adds support for ADT7461A and NCT1008 to the lm90 driver.
Both chips have identical functionality and report the same manufacturing ID
and device ID values.
Signed-off-by: NGuenter Roeck <guenter.roeck@ericsson.com>
Signed-off-by: NJean Delvare <khali@linux-fr.org>

Fixes generated by 'codespell' and manually reviewed.
Signed-off-by: NLucas De Marchi <lucas.demarchi@profusion.mobi>

I was wondering if that chip ever existed publicly... Apparently yes,
so add support for it.
Signed-off-by: NJean Delvare <khali@linux-fr.org>
Tested-by: NAlexander Stein <alexander.stein@informatik.tu-chemnitz.de>
Acked-by: NGuenter Roeck <guenter.roeck@ericsson.com>