• Greg Kroah-Hartman's avatar
    Merge tag 'iio-fixes-for-4.1a-take2' of... · ec94efcd
    Greg Kroah-Hartman authored
    Merge tag 'iio-fixes-for-4.1a-take2' of git://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio into staging-linus
    
    Jonathan writes:
    
    The usual mixed bag of fixes for IIO in the 4.1 cycle.
    
    Second version of this pull request as a small fix to a fix turned
    up before Greg pulled it for a cc10001 patch near the top of the tree.
    
    One core fix
    
    * Set updated for a iio kfifo was incorrectly set to false during a failed
      update, resulting in atttempts to repeat the failed operation appearing
      to succeed.
    
    This time I've decided to list the driver fixes in alphabetical order rather
    than 'randomly'.
    
    * axp288_adc - a recent change added a check for valid info masks when
      reading channels from consumer drivers.
    * bmp280 - temperature compensation was failing to read the tfine value, hence
      causing a temperature of 0 to always be returned and incorrect presure
      measurements.
    * cc10001 - Fix channel number mapping when some channels are reserved for
      remote CPUs. Fix an issue with the use of the power-up/power-down register
      (basically wrong polarity). Fix an issue due to the missinterpretting the
      return value from regulator_get_voltage. Add a delay before the start bit
      as recommended for the hardware to avoid data corruption.
    * hid pressure - fix channel spec of modfiied, but no modifier (which makes no
      sense!)
    * hid proximity - fix channel spec of modified, but no modifier (which makes
      no sense!). Fix a memory leak in the probe function.
    * mcp320x - occasional incorrect readings on dma using spi busses due to
      cacheline corruption. Fixed by forcing ___cacheline_aligned for the buffers.
    * mma9551 - buffer overrun fix (miss specified maximum length of buffers)
    * mma9553 - endian fix on status message. Add an enable element for activity
      channel. Input checking for activity period to avoid rather unpredictable
      results.
    * spmi-vadc - fix an overflow in the output value normalization seen on some
      boards.
    * st-snesors - oops due to use of a mutex that is not yet initialized during
      probe.
    * xilinx adc - Some wrong register addresses, a wrong address for vccaux
      channel, incorrect scale on VREFP and incorrect sign on VREFN.
    ec94efcd
hid-sensor-press.c 10.5 KB