When using more than one trigger consumer it can happen that multiple threads
perform a read-modify-update cycle on 'use_count' concurrently. This can cause
updates to be lost and use_count can get stuck at non-zero value, in which case
the IIO core assumes that at least one thread is still running and will wait for
it to finish before running any trigger handlers again. This effectively renders
the trigger disabled and a reboot is necessary before it can be used again. To
fix this make use_count an atomic variable. Also set it to the number of
consumers before starting the first consumer, otherwise it might happen that
use_count drops to 0 even though not all consumers have been run yet.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Tested-by: NDenis Ciocca <denis.ciocca@st.com>
Signed-off-by: NJonathan Cameron <jic23@kernel.org>

iio_trigger unregistration and freeing has been separated in this
code for some time, but it looks like the calls to the device
handling were not appropriately updated.
Signed-off-by: NJonathan Cameron <jic23@kernel.org>
Reported-by: NOtavio Salvador <otavio@ossystems.com.br>
Tested-by: NOtavio Salvador <otavio@ossystems.com.br>
Reviewed-by: NLars-Peter Clausen <lars@metafoo.de>

Signed-off-by: NPeter Meerwald <pmeerw@pmeerw.net>
Signed-off-by: NJonathan Cameron <jic23@kernel.org>

The name sysfs attribute is the same for all triggers, so there is no need to
register them dynamically at runtime. Create a attribute group for it and set it
up for the bus attribute group. This also avoids a possible race condition
where the uevent for the device is sent before the name sysfs attribute has been
added.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Signed-off-by: NJonathan Cameron <jic23@kernel.org>

Signed-off-by: NPeter Meerwald <pmeerw@pmeerw.net>
Signed-off-by: NJonathan Cameron <jic23@kernel.org>

Use to_iio_trigger(dev) instead of dev_get_drvdata(dev). Both will return the
trigger which belongs to the device, but the the first on is a bit more
lightweight.

Since this is the last location where we used dev_get_drvdata() for retrieving
the trigger there is no need anymore to assign the the trigger to the devices
drvdata, so we can remove that as well.
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Acked-by: NJonathan Cameron <jic23@kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Replace open-coded instances of getting a iio_dev struct from a device struct
with dev_to_iio_dev().
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Acked-by: NJonathan Cameron <jic23@kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Currently we use two different naming schemes in the IIO API, iio_verb_object
and iio_object_verb. E.g iio_device_register and iio_allocate_device. This
patches renames instances of the later to the former. The patch also renames allocate to
alloc as this seems to be the preferred form throughout the kernel.

In particular the following renames are performed by the patch:
	iio_put_device -> iio_device_put
	iio_allocate_device -> iio_device_alloc
	iio_free_device -> iio_device_free
	iio_get_trigger -> iio_trigger_get
	iio_put_trigger -> iio_trigger_put
	iio_allocate_trigger -> iio_trigger_alloc
	iio_free_trigger -> iio_trigger_free

The conversion was done with the following coccinelle patch with manual fixes to
comments and documentation.

<smpl>
@@
@@
-iio_put_device
+iio_device_put
@@
@@
-iio_allocate_device
+iio_device_alloc
@@
@@
-iio_free_device
+iio_device_free
@@
@@
-iio_get_trigger
+iio_trigger_get
@@
@@
-iio_put_trigger
+iio_trigger_put
@@
@@
-iio_allocate_trigger
+iio_trigger_alloc
@@
@@
-iio_free_trigger
+iio_trigger_free
</smpl>
Signed-off-by: NLars-Peter Clausen <lars@metafoo.de>
Acked-by: NJonathan Cameron <jic23@kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Take the core support + the kfifo buffer implentation out of
staging.  Whilst we are far from done in improving this subsystem
it is now at a stage where the userspae interfaces (provided by
the core) can be considered stable.

Drivers will follow over a longer time scale.
Signed-off-by: NJonathan Cameron <jic23@kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Step 1 in moving the IIO core out of staging.
Signed-off-by: NJonathan Cameron <jic23@kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The below patch fixes some typos in various parts of the kernel, as well as fixes some comments.
Please let me know if I missed anything, and I will try to get it changed and resent.
Signed-off-by: NJustin P. Mattock <justinmattock@gmail.com>
Acked-by: NRandy Dunlap <rdunlap@xenotime.net>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

Both of these are decidedly silly bugs show up whilst testing
completely different code paths.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Postenable and predisable are called via buffer->ops so  don't
need to check if buffer exists.

The return value of iio_device_register_trigger_consumer is
always zero and it isn't checked anyway so get rid of it.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

We had a random missmatch of these two.  Lets pick the most common
and get rid of the other.  This patch covers the core.  Others
will clean up the drivers.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

also, IIO_RING_HARDWARE_BUFFER -> IIO_BUFFER_HARDWARE

These aren't always rings so the naming should not imply that.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Not always a ring so naming is missleading.
Also, kfifo_buf is probably first buffer to take out of staging and it
definitely isn't a ring.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This set also includes quite a number of bug fixes of particularly
remove functions.

Necessary due to issue pointed out in Bart Van Assche's patch:
docs/driver-model: Document device.groups

V2: Rebase due to patch reordering.
V3: Pull various error fixes and cleanups out into their own patches.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Acked-by: NMichael Hennerich <Michael.Hennerich@analog.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

All iio dependent modules statically depend on industrialio
so these aren't needed. Not sure why they originally existed,
but they aren't now.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Use new simple ida allocation functions to remove some boilerplate
code.

Also, now we only have one chdev per device we don't need to maintain
a separate ida for minor numbers. Just use the devices id.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Stop exporting functions only intended for core usage.
Return void from function which can never return anything other than 0.
Trivial stype cleanups.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Need this out the way to create a trigger_consumer.h that actually is
for trigger consumers.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This is always true, so lets make it explicit.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This is to do with dynamically allocated triggers and the need to ensure
the underlying structures don't go away when a consumer is using them.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Also get rid of a few function defs where they are only now in one core file
anyway.
Whilst here add mask = 0 to get rid of warning.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Right now this results in increased code, but I still think it is worth doing to avoid
replication across instances of drivers etc and move as much stuff as possible to constant.
Ops structure is optional for the occasional driver that uses none of it (currently
only the ad7793).
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Allow devices to reject triggers and vice versa.

Changes since V1:
Added kernel-doc
Moved callback into iio_info
Changed function naming
Revised return value passing
Add forward declaration to avoid warnings
Signed-off-by: NMichael Hennerich <michael.hennerich@analog.com>
Acked-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Failure to set iio_poll_func private_data, causes zero pointer access
violations in all consumer trigger handlers.
Signed-off-by: NMichael Hennerich <michael.hennerich@analog.com>
Acked-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Add a function to neatly deal with allocation of poll functions.
Ultimately this allows us to more easily change the implementation.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

staging:iio:trigger handle name attr in core, remove old alloc and register any control_attrs via struct device

As the majority of triggers don't actually have any other control_attrs lets use the fact
that struct device has a groups element when we do need to have these attributes registered.
A vargs function is used to cut down on lots of building strings in every single driver
just in order to pass them into the allocate.

Also iio_allocate_trigger_named -> iio_allocate_trigger as there is no
unamed version any more, so that is now just confusing.

Blackfin tested and fixed by Michael Hennerich.

V2: Elements from Michael Hennerich's patches for the ade7758
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Awaiting comments on using the nested_irq_trick so that may change.
Moves away from platform device to sysfs controlled creation and
removal of these triggers.

Fix double free of name on trigger allocation failure thanks
to Michael Hennerich.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Reviewed-by: NMichael Hennerich <michael.hennerich@analog.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This case is extremely common, so let us only have the one
copy.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

V3: Get rid of separate interrupt pool. This is well handled
    by irq_get_descs and irq_free_descs.  Two functions I simply
    wasn't aware of previously.  Thus the allocation for a given
    trigger is now handled by core code rather than us reinventing
    the wheel.

V2: Stop silly name duplication.
    Move pool handling to industrialio-trigger as that is the only user.
    Changed over to using irq_modify_status rather than the arm
    specific set_irq_flags as per Thomas Gleixner's suggestion.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Trivial reorganization.
Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: NJonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>