Merge tag 'iio-for-3.15b' of git://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio into staging-next
Jonathan writes:
Second round of IIO new driver, functionality and cleanups for the 3.15 series.
There are a few fixes in here that might, earlier in a cycle, have gone
to Greg as fixes. Given they are either minor or have never actually
been observed as causing trouble (the locking bug in the event code) and
are invasive, I have included them in this pull request, targeting the
3.15 merge window instead.
The rest are pretty uncontroversial new drivers, a handy little tool for
the example code in our documentation and little cleanups.
New drivers
* Freescale Vybrid and i.MX6SLX ADC driver.
* HID Sensor hub proximity sensors.
* HID Sensor hub pressure sensors.
* LPS25H Pressure sensors added to the ST micro pressure sensor driver.
New functionality
* lsiio tool. This is added to the staging tree as we haven't yet moved
the example code it sits with out. Moving this code out is now a reasonably
high priority but holding up this tool in the meantime did not seem
worthwhile.
* mag3110 - add missing scale factor for temperature output to userspace.
Cleanups
* Fix a bug in the event reporting in which a spin lock might be held over
when a sleep occured. A similar bug was found by Lars in the buffer code.
It has not to our knowledge been observed as actually occuring and is
a little too invasive to push out as a fix.
* Drop the IIO_ST macro after clearing out all users. This macro was a very
bad idea leading to a number of bugs after it stopped covering all elements
of the structure being assigned and people started making assumptions about
what it did cover. Glad to see it go!
* Avoid applying extended name to shared attributes as it makes no sense.
No in tree drivers were using the combination, hence not pushed out as
a fix.
* ad799x - move to devm_request_threaded_irq to reduce boilerplate clean up.
* bma180 - make the low_pass_filter_3db_frequency info element shared rather
than per attribute. The old approach was valid but not as clean as it might
be and was setting a bad example. Hence the cleanup.
* mxs-lradc - propogate the error code form a platform_get_irq call rather than
eating it up by returning -EINVAL on all errors.
* ad799x - typo fix in the copyright message. Either that or Michael was
asserting a copyright that moved backwards in time by about a thousand years.
* ad799x - use a regulator for vref rather than platform data. The driver
dates from just as the regulator framework was coming into common use so
provides an alternative way of specifying the reference voltage. We no
longer need that approach so drop it in favour of a regulator only approach.
* max1363 - some internal vref values were out by a small amount. The effect
would have been tiny and no one noticed hence not pushing this through as
a fix.
* core - replace some pointless goto error_ret (with no clean up) lines with
direct returns. This is my bad coding style so I'm glad to see it cleaned
up.
* core - avoid a kasprintf that just directly prints a string with no
formatting elements. This has always been there but Lars just noticed it.
Oops.
The event code currently holds a spinlock with IRQs disabled while calling
kfifo_to_user(). kfifo_to_user() can generate a page fault though, which means
we have to be able to sleep, which is not possible if the interrupts are
disabled. The good thing is that kfifo handles concurrent read and write access
just fine as long as there is only one reader and one writer, so we do not any
locking to protect against concurrent access from the read and writer thread. It
is possible though that userspace is trying to read from the event FIFO from
multiple concurrent threads, so we need to add locking to protect against this.
This is done using a mutex. The mutex will only protect the kfifo_to_user()
call, it will not protect the waitqueue. This means that multiple threads can be
waiting for new data and once a new event is added to the FIFO all waiting
threads will be woken up. If one of those threads is unable to read any data
(because another thread already read all the data) it will go back to sleep. The
only remaining issue is that now that the clearing of the BUSY flag and the
emptying of the FIFO does no longer happen in one atomic step it is possible
that a event is added to the FIFO after it has been emptied and this sample will
be visible the next time a new event file descriptor is created. To avoid this
rather move the emptying of the FIFO from iio_event_chrdev_release to
iio_event_getfd().
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de> Signed-off-by: Jonathan Cameron <jic23@kernel.org>
Archana Patni [Thu, 20 Feb 2014 06:30:00 +0000 (06:30 +0000)]
iio: hid-sensors: Added Pressure Sensor driver
Added usage id processing for Pressure Sensor. This uses IIO
interfaces for triggered buffer to present data to user
mode. This uses HID sensor framework for registering callback
events from the sensor hub.
Archana Patni [Thu, 20 Feb 2014 06:29:00 +0000 (06:29 +0000)]
iio: hid-sensors: Added Proximity Sensor Driver
Added usage id processing for Proximity (Human Presence).
This uses IIO interfaces for triggered buffer to present data
to user mode. This uses HID sensor framework for registering
callback events from the sensor hub.
This function probes a number of the boards registers during the
(*attach) to verify that it is actually a PCL-816 compatible board.
For aesthetics, move the function closer to the (*attach).
To better match the pcl818 driver, allocate the private data before
calling pcl816_check().
Refactor the function to return an errno if fails. Change the errno
from -EIO to -ENODEV and remove the unnecessary dev_err() noise.
Make sure the CONTROL register is reset to a known state after the
check. The 0x18 value actually defines an invalid interrupt selection
and sets an undefined bit.
Add a couple comments to clarify the magic values.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: pcl816: convert private data flags to bit-fields
The 'irq_was_now_closed' member is actually a flag, devpriv->int816_mode
will always be > 0 when it's used to set irq_was_now_closed in the cancel
function.
Convert the flags in the private data to bit-fields to save a bit of
space.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: pcl816: don't calc the timer divisors twice
The timer divisors are calculated in the (*do_cmdtest) before the (*do_cmd)
is called by the comedi core. The extra sanity checks in the (*do_cmd) are
not necessary, the values returned from i8253_cascade_ns_to_timer() will be
greater than 1. Save the values in the private data so they don't need to be
recalced.
Refactor pcl816_start_pacer() to use the values from the private data.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: pcl816: clarify irq request in pcl816_attach()
All the board types can use IRQ 2-7 for async command support. Remove
the 'IRQbits', which is a mask of the valid IRQs, from the boardinfo
and refactor pcl816_attach().
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Chase Southwood [Wed, 19 Feb 2014 07:37:54 +0000 (01:37 -0600)]
Staging: comedi: addi-data: replace printk() with dev_err() in hwdrv_apci1564.c
There were a small handful of printk() calls in hwdrv_apci1564.c. It is
generally better to use dev_err() for error messages instead, so I
switched all the printk() calls out, as well as cleaned up the error
strings.
Chase Southwood [Wed, 19 Feb 2014 07:37:42 +0000 (01:37 -0600)]
Staging: comedi: addi-data: cleanup brace usage in hwdrv_apci1564.c
hwdrv_apci1564.c had many single statments wrapped in braces, so we can
delete these. Also, some else statements were improperly placed, fix
these too.
Chase Southwood [Wed, 19 Feb 2014 07:36:46 +0000 (01:36 -0600)]
Staging: comedi: addi-data: cleanup comments in hwdrv_apci1564.c
hwdrv_apci1564.c had a lot of commented out conditional statements that
were often identical to other un-commented out statements nearby, so it
should be safe to just delete all of these commented out lines. This
patch also converts the remaining comments to the preferred kernel style
for comments.
Signed-off-by: Chase Southwood <chase.southwood@yahoo.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Alan [Mon, 17 Feb 2014 14:13:08 +0000 (14:13 +0000)]
staging: et131x: fix allocation failures
We should check the ring allocations don't fail.
If we get a fail we need to clean up properly. The allocator assumes the
deallocator will be used on failure, but it isn't. Make sure the
right deallocator is always called and add a missing check against
fbr allocation failure.
[v2]: Correct check logic
Signed-off-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This function probes a number of the boards registers during the
(*attach) to verify that it is actually a PCL-818 compatible board.
For aesthetics, move the function closer to the (*attach).
Refactor the function to return an errno if fails. Change the errno
from -EIO to -ENODEV and remove the unnecessary comedi_error() noise.
Make sure the CONTROL register is reset to a known state after the
check. The 0x18 value actually defines an invalid interrupt selection
and sets an undefined bit.
Add a couple comments to clarify the magic values.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: pcl818: don't calc the timer divisors twice
The timer divisors are calculated in the (*do_cmdtest) before the (*do_cmd)
is called by the comedi core. The extra sanity checks in the (*do_cmd) are
not necessary, the values returned from i8253_cascade_ns_to_timer() will be
greater than 1. Save the values in the private data so they don't need to be
recalced.
Refactor pcl818_start_pacer() to use the values from the private data.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: pcl812: don't calc the timer divisors twice
The timer divisors are calculated in the (*do_cmdtest) before the (*do_cmd)
is called by the comedi core. Save the values in the private data so they
don't need to be recalced.
Refactor pcl812_start_pacer() to use the values from the private data.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: pcl818: tidy up dma buffer allocation
This driver uses 2 buffers for DMA. Refactor the buffer allocation to
use a for loop to remove code duplication. Remove the dev_err() messages
when __get_dma_pages() fails and change the errno returned from -EBUSY
to -ENOMEM.
Both buffers are the same size so replace the 'dmapages' and 'hwdmasize'
arrays in the private data with variables to save a bit of space.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: pcl816: tidy up dma buffer allocation
This driver uses 2 buffers for DMA. Refactor the buffer allocation to
use a for loop to remove code duplication. Remove the dev_err() messages
when __get_dma_pages() fails and change the errno returned from -EBUSY
to -ENOMEM.
Both buffers are the same size so replace the 'dmapages' and 'hwdmasize'
arrays in the private data with variables to save a bit of space.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: pcl812: tidy up dma buffer allocation
This driver uses 2 buffers for DMA. Refactor the buffer allocation to
use a for loop to remove code duplication. Remove the dev_err() messages
when __get_dma_pages() fails and change the errno returned from -EBUSY
to -ENOMEM.
Both buffers are the same size so replace the 'dmapages' and 'hwdmasize'
arrays in the private data with variables to save a bit of space.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This (*insn_read) function was used to read 16-bit analog input data
from the boardACL8216 boardtypes. The 12/16-bit differences are now
handled by the pcl812_ai_eoc() and pcl812_ai_get_sample() helpers.
Remove this function and use pcl812_ai_insn_read() for all boardtypes.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Introduce a helper function to read the 12-bit analog input data
sample and optionally return the channel that the sample was for.
The channel is only used in the interrupt routine to check for
dropped samples.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: pcl812: remove 'ai_maxdata' from boardinfo
Most of the board types supported by this driver have 12-bit analog
inputs. Two of them, the acl8216 and a826pg, have 16-bit analog inputs.
Remove the 'ai_maxdata' member from the boardinfo and replace it with
a bit-field flag 'has_16bit_ai'. Refactor pcl812_attach() to use this
new boardinfo.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: pcl818: clarify irq request in pcl818_attach()
All the board types can use IRQ 2-7 for async command support. Remove
the 'IRQbits', which is a mask of the valid IRQs, from the boardinfo
and refactor pcl818_attach().
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: pcl812: tidy up differential ai user option
Some of the boards supported by this driver can do differential analog
input when configureg with jumpers on the board. When diff. ai is used
the number of input channels is half the single-ended number of channels.
The user specifies the analog input mode for these boards when attaching
to the driver by using a configuration option.
Remove the unnecessary 'n_aichan_diff' member from the boardinfo. Add a
comment for the boards that can do differential ai.
Refactor pcl812_attach() to parse the user option before setting up the
subdevices. We can then use the 'use_diff' flag to determine if the ai
is single-ended or differential and set the subdev_flags and n_chan.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: pcl818: factor analog input range selection out of (*attach)
The analog input subdevice range is setup in this driver based on a config
option passed by the user. Factor the code that sets the range_table out
of pcl818_attach() to clarify the (*attach).
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: pcl812: factor analog input range selection out of (*attach)
The analog input subdevice range is setup in this driver based on a config
option passed by the user. Factor the code that sets the range_table out
of pcl812_attach() to clarify the (*attach).
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: pcl812: tidy up digital subdevice boardinfo
For the board types that have digital inputs and outputs there are always
16 input channels and 16 output channels.
Remove the 'n_dichan' and 'n_dochan' members in the boardinfo and replace
them with a bit-field flag 'has_dio'. Refactor pcl812_attach() to use the
new boardinfo.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: pcl818: clarify dma channel request in pcl818_attach()
All the board types that can do DMA can use DMA channels 3 or 1. Remove
the 'DMAbits', which is a mask of the valid channels, from the boardinfo
and replace it with a bit-field flag 'has_dma'.
Refactor pcl818_attach() to use the new flag and remove the need for the
goto.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: pcl812: clarify dma channel request in pcl812_attach()
All the board types that can do DMA can use DMA channels 3 or 1. Remove
the 'DMAbits', which is a mask of the valid channels, from the boardinfo
and replace it with a bit-field flag 'has_dma'.
Refactor pcl812_attach() to use the new flag and remove the need for the
goto.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Dan Carpenter [Mon, 17 Feb 2014 19:56:06 +0000 (22:56 +0300)]
staging/bcm: two information leaks in ioctl
There are a couple paths where we don't check how much data we copy back
to the user.
Cc: Dave Jones <davej@redhat.com> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Zhao, Gang [Tue, 18 Feb 2014 15:03:27 +0000 (23:03 +0800)]
staging: wlan-ng: replace function ieee80211_dsss_chan_to_freq()
Replace ieee80211_dsss_chan_to_freq() with more generic
ieee80211_channel_to_frequency(), and add a variable to deal with 80
characters problem.
File cfg80211.c is included by p80211netdev.c, p80211netdev.c includes
<net/cfg80211.h>, both ieee80211_channel_to_frequency() and
IEEE80211_BAND_2GHZ is defined / declared in <net/cfg80211.h>. So this
change is safe.
This change is a preparation for the removal of function
ieee80211_{dsss_chan_to_freq, freq_to_dsss_chan}.
Signed-off-by: Zhao, Gang <gamerh2o@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
projects / mirror_ubuntu-kernels.git / log