linux-stable.git/drivers/iio, branch v3.16.78

iio: core: fix a possible circular locking dependency

2019-08-13T11:38:58+00:00

commit 7f75591fc5a123929a29636834d1bcb8b5c9fee3 upstream.

This fixes a possible circular locking dependency detected warning seen
with:
- CONFIG_PROVE_LOCKING=y
- consumer/provider IIO devices (ex: "voltage-divider" consumer of "adc")

When using the IIO consumer interface, e.g. iio_channel_get(), the consumer
device will likely call iio_read_channel_raw() or similar that rely on
'info_exist_lock' mutex.

typically:
...
	mutex_lock(&chan->indio_dev->info_exist_lock);
	if (chan->indio_dev->info == NULL) {
		ret = -ENODEV;
		goto err_unlock;
	}
	ret = do_some_ops()
err_unlock:
	mutex_unlock(&chan->indio_dev->info_exist_lock);
	return ret;
...

Same mutex is also hold in iio_device_unregister().

The following deadlock warning happens when:
- the consumer device has called an API like iio_read_channel_raw()
  at least once.
- the consumer driver is unregistered, removed (unbind from sysfs)

======================================================
WARNING: possible circular locking dependency detected
4.19.24 #577 Not tainted
------------------------------------------------------
sh/372 is trying to acquire lock:
(kn->count#30){++++}, at: kernfs_remove_by_name_ns+0x3c/0x84

but task is already holding lock:
(&dev->info_exist_lock){+.+.}, at: iio_device_unregister+0x18/0x60

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #1 (&dev->info_exist_lock){+.+.}:
       __mutex_lock+0x70/0xa3c
       mutex_lock_nested+0x1c/0x24
       iio_read_channel_raw+0x1c/0x60
       iio_read_channel_info+0xa8/0xb0
       dev_attr_show+0x1c/0x48
       sysfs_kf_seq_show+0x84/0xec
       seq_read+0x154/0x528
       __vfs_read+0x2c/0x15c
       vfs_read+0x8c/0x110
       ksys_read+0x4c/0xac
       ret_fast_syscall+0x0/0x28
       0xbedefb60

-> #0 (kn->count#30){++++}:
       lock_acquire+0xd8/0x268
       __kernfs_remove+0x288/0x374
       kernfs_remove_by_name_ns+0x3c/0x84
       remove_files+0x34/0x78
       sysfs_remove_group+0x40/0x9c
       sysfs_remove_groups+0x24/0x34
       device_remove_attrs+0x38/0x64
       device_del+0x11c/0x360
       cdev_device_del+0x14/0x2c
       iio_device_unregister+0x24/0x60
       release_nodes+0x1bc/0x200
       device_release_driver_internal+0x1a0/0x230
       unbind_store+0x80/0x130
       kernfs_fop_write+0x100/0x1e4
       __vfs_write+0x2c/0x160
       vfs_write+0xa4/0x17c
       ksys_write+0x4c/0xac
       ret_fast_syscall+0x0/0x28
       0xbe906840

other info that might help us debug this:

 Possible unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(&dev->info_exist_lock);
                               lock(kn->count#30);
                               lock(&dev->info_exist_lock);
  lock(kn->count#30);

 *** DEADLOCK ***
...

cdev_device_del() can be called without holding the lock. It should be safe
as info_exist_lock prevents kernelspace consumers to use the exported
routines during/after provider removal. cdev_device_del() is for userspace.

Help to reproduce:
See example: Documentation/devicetree/bindings/iio/afe/voltage-divider.txt
sysv {
	compatible = "voltage-divider";
	io-channels = <&adc 0>;
	output-ohms = <22>;
	full-ohms = <222>;
};

First, go to iio:deviceX for the "voltage-divider", do one read:
$ cd /sys/bus/iio/devices/iio:deviceX
$ cat in_voltage0_raw

Then, unbind the consumer driver. It triggers above deadlock warning.
$ cd /sys/bus/platform/drivers/iio-rescale/
$ echo sysv > unbind

Note I don't actually expect stable will pick this up all the
way back into IIO being in staging, but if's probably valid that
far back.

Signed-off-by: Fabrice Gasnier 
Fixes: ac917a81117c ("staging:iio:core set the iio_dev.info pointer to null on unregister")
Signed-off-by: Jonathan Cameron 
[bwh: Backported to 3.16: adjust context]
Signed-off-by: Ben Hutchings

iio: ad_sigma_delta: select channel when reading register

2019-08-13T11:38:53+00:00

commit fccfb9ce70ed4ea7a145f77b86de62e38178517f upstream.

The desired channel has to be selected in order to correctly fill the
buffer with the corresponding data.
The `ad_sd_write_reg()` already does this, but for the
`ad_sd_read_reg_raw()` this was omitted.

Fixes: af3008485ea03 ("iio:adc: Add common code for ADI Sigma Delta devices")
Signed-off-by: Dragos Bogdan 
Signed-off-by: Alexandru Ardelean 
Signed-off-by: Jonathan Cameron 
Signed-off-by: Ben Hutchings

iio: dac: mcp4725: add missing powerdown bits in store eeprom

2019-08-13T11:38:42+00:00

commit 06003531502d06bc89d32528f6ec96bf978790f9 upstream.

When issuing the write DAC register and write eeprom command, the two
powerdown bits (PD0 and PD1) are assumed by the chip to be present in
the bytes sent. Leaving them at 0 implies "powerdown disabled" which is
a different state that the current one. By adding the current state of
the powerdown in the i2c write, the chip will correctly power-on exactly
like as it is at the moment of store_eeprom call.

This is documented in MCP4725's datasheet, FIGURE 6-2: "Write Commands
for DAC Input Register and EEPROM" and MCP4726's datasheet, FIGURE 6-3:
"Write All Memory Command".

Signed-off-by: Jean-Francois Dagenais 
Acked-by: Peter Meerwald-Stadler 
Signed-off-by: Jonathan Cameron 
[bwh: Backported to 3.16: adjust context]
Signed-off-by: Ben Hutchings

iio: adc: at91: disable adc channel interrupt in timeout case

2019-08-13T11:38:40+00:00

commit 09c6bdee51183a575bf7546890c8c137a75a2b44 upstream.

Having a brief look at at91_adc_read_raw() it is obvious that in the case
of a timeout the setting of AT91_ADC_CHDR and AT91_ADC_IDR registers is
omitted. If 2 different channels are queried we can end up with a
situation where two interrupts are enabled, but only one interrupt is
cleared in the interrupt handler. Resulting in a interrupt loop and a
system hang.

Signed-off-by: Georg Ottinger 
Acked-by: Ludovic Desroches 
Signed-off-by: Jonathan Cameron 
Signed-off-by: Ben Hutchings

iio: Fix scan mask selection

2019-08-13T11:38:40+00:00

commit 20ea39ef9f2f911bd01c69519e7d69cfec79fde3 upstream.

The trialmask is expected to have all bits set to 0 after allocation.
Currently kmalloc_array() is used which does not zero the memory and so
random bits are set. This results in random channels being enabled when
they shouldn't. Replace kmalloc_array() with kcalloc() which has the same
interface but zeros the memory.

Note the fix is actually required earlier than the below fixes tag, but
will require a manual backport due to move from kmalloc to kmalloc_array.

Signed-off-by: Lars-Peter Clausen 
Signed-off-by: Alexandru Ardelean 
Fixes commit 057ac1acdfc4 ("iio: Use kmalloc_array() in iio_scan_mask_set()").
Signed-off-by: Jonathan Cameron 
Signed-off-by: Ben Hutchings

iio: Use kmalloc_array() in iio_scan_mask_set()

2019-08-13T11:38:39+00:00

commit 057ac1acdfc4743f066fcefe359385cad00549eb upstream.

A multiplication for the size determination of a memory allocation
indicated that an array data structure should be processed.
Thus use the corresponding function "kmalloc_array".

This issue was detected by using the Coccinelle software.

Signed-off-by: Markus Elfring 
Signed-off-by: Jonathan Cameron 
Signed-off-by: Ben Hutchings

iio/hid-sensors: Fix IIO_CHAN_INFO_RAW returning wrong values for signed numbers

2019-02-11T17:54:05+00:00

commit 0145b50566e7de5637e80ecba96c7f0e6fff1aad upstream.

Before this commit sensor_hub_input_attr_get_raw_value() failed to take
the signedness of 16 and 8 bit values into account, returning e.g.
65436 instead of -100 for the z-axis reading of an accelerometer.

This commit adds a new is_signed parameter to the function and makes all
callers pass the appropriate value for this.

While at it, this commit also fixes up some neighboring lines where
statements were needlessly split over 2 lines to improve readability.

Signed-off-by: Hans de Goede 
Acked-by: Srinivas Pandruvada 
Acked-by: Benjamin Tissoires 
Signed-off-by: Jonathan Cameron 
[bwh: Backported to 3.16:
 - sensor_hub_input_attr_get_raw_value() doesn't take a sync/async flag
   parameter
 - In sensor_hub_input_attr_get_raw_value() keep using data->pending instead of
   hsdev->pending
 - In magn_3d_read_raw() keep using chan->scan_index intstead of chan->address
 - Drop changes in hid-sensor-{custom,humidity,temperature}
 - Adjust context]
Signed-off-by: Ben Hutchings

iio: ad5064: Fix regulator handling

2019-02-11T17:53:24+00:00

commit 8911a43bc198877fad9f4b0246a866b26bb547ab upstream.

The correct way to handle errors returned by regualtor_get() and friends is
to propagate the error since that means that an regulator was specified,
but something went wrong when requesting it.

For handling optional regulators, e.g. when the device has an internal
vref, regulator_get_optional() should be used to avoid getting the dummy
regulator that the regulator core otherwise provides.

Signed-off-by: Lars-Peter Clausen 
Signed-off-by: Jonathan Cameron 
[bwh: Backported to 3.16: Keep using ad5064_write() instead of
 ad5064_set_config().]
Signed-off-by: Ben Hutchings

iio: adc: at91: fix wrong channel number in triggered buffer mode

2019-02-11T17:53:24+00:00

commit aea835f2dc8a682942b859179c49ad1841a6c8b9 upstream.

When channels are registered, the hardware channel number is not the
actual iio channel number.
This is because the driver is probed with a certain number of accessible
channels. Some pins are routed and some not, depending on the description of
the board in the DT.
Because of that, channels 0,1,2,3 can correspond to hardware channels
2,3,4,5 for example.
In the buffered triggered case, we need to do the translation accordingly.
Fixed the channel number to stop reading the wrong channel.

Fixes: 0e589d5fb ("ARM: AT91: IIO: Add AT91 ADC driver.")
Cc: Maxime Ripard 
Signed-off-by: Eugen Hristev 
Acked-by: Ludovic Desroches 
Signed-off-by: Jonathan Cameron 
Signed-off-by: Ben Hutchings

iio: adc: at91: fix acking DRDY irq on simple conversions

2019-02-11T17:53:23+00:00

commit bc1b45326223e7e890053cf6266357adfa61942d upstream.

When doing simple conversions, the driver did not acknowledge the DRDY irq.
If this irq status is not acked, it will be left pending, and as soon as a
trigger is enabled, the irq handler will be called, it doesn't know why
this status has occurred because no channel is pending, and then it will go
int a irq loop and board will hang.
To avoid this situation, read the LCDR after a raw conversion is done.

Fixes: 0e589d5fb ("ARM: AT91: IIO: Add AT91 ADC driver.")
Cc: Maxime Ripard 
Signed-off-by: Eugen Hristev 
Acked-by: Ludovic Desroches 
Signed-off-by: Jonathan Cameron 
Signed-off-by: Ben Hutchings