pm+acpi-3.11-rc1-more

- Fix for a recent cpufreq regression that caused WARN() to trigger
  overzealously in a couple of places and spam the kernel log with
  useless garbage as a result.  From Viresh Kumar.

- ACPI dock fix removing a discrepancy between the definition of
  acpi_dock_init(), which says that the function returns int, and
  its header in the header file, which says that it is a void
  function.  The function is now defined as void too.

- ACPI PM fix for failures to update device power states as needed,
  for example, during resume from system suspend, because the old
  state was deeper than the new one, but the new one is not D0.

- Fix for two debug messages in the ACPI power resources code that
  don't have a newline at the end and make the kernel log difficult
  to read.  From Mika Westerberg.

- Two ACPI cleanups from Naresh Bhat and Haicheng Li.

- cpupower updates from Thomas Renninger, including Intel Haswell
  support improvements and a new idle-set subcommand among other
  things.

/