Tags

Final power management fixes for 3.15

 - Taking non-idle time into account when calculating core busy
   time was a mistake and led to a performance regression.  Since
   the problem it was supposed to address is now taken care of in
   a different way, we don't need to do it any more, so drop the
   non-idle time tracking from intel_pstate.  Dirk Brandewie.

 - Changing to fixed point math throughout the busy calculation
   introduced rounding errors that adversely affect the accuracy
   of intel_pstate's computations.  Fix from Dirk Brandewie.

 - The PID controller algorithm used by intel_pstate assumes that
   the time interval between two adjacent samples will always be the
   same which is not the case for deferable timers (used by
   intel_pstate) when the system is idle. This leads to inaccurate
   predictions and artificially increases convergence times for
   the minimum P-state.  Fix from Dirk Brandewie.

 - intel_pstate carries out computations using 32-bit variables
   that may overflow for large enough values of APERF/MPERF.  Switch
   to using 64-bit variables for computations, from Doug Smythies.

/

regulator: Updates for v3.16

The bulk of the changes for this release are a few new drivers however
there are a couple of noticable core changes and the usual stream of
cleanups and fixes:

 - Move disable of unused regulators later in init so it comes after
   deferred probe has iterated making startup smoother.
 - Fixes reference counting of the DT nodes for constraints from Charles
   Keepax.  This has little practical impact since all real users of
   the regulator bindings use FDT which doesn't need the reference
   counting.
 - Lots of cleanups, especially to the Samsung drivers.
 - Support for Linear Technologies LTC3589, Texas Instruments TPS658640
   and X-Powers AXP20x.

ASoC: Final updates for v3.16

A few more updates from the last week of development, nothing too
exciting.  Highlights include:

- GPIO descriptor support for jacks
- More updates and fixes to the Freescale SSI, Intel and rsnd drivers.
- New drivers for Analog Devices ADAU1361, ADAU1381, ADAU1761 and
  ADAU1781, and Realtek RT5677.

sound fixes for 3.15-final

A few addition of HD-audio fixups for ALC260 and AD1986A codecs.
All marked as stable fixes.

The fixes are pretty local and they are old machines, so quite safe
to apply.

This is the bulk of pin control changes for the v3.16
development cycle:

- Antoine Tenart made the get_group_pins() vtable entry
  optional.

- Antoine also provides an entirely new driver for the
  Marvell Berlin SoC. This is unrelated to the existing
  MVEBU hardware driver and warrants its own separate
  driver.

- Reflected from the GPIO subsystem there is a number of
  refactorings to make pin control drivers with gpiochips
  use the new gpiolib irqchip helpers. The following
  drivers were converted to use the new infrastructure:

  - ST Microelectronics STiH416 and friends

  - The Atmel AT91

  - The CSR SiRF (Prima2)

  - The Qualcomm MSM series

- Massive improvements in the Qualcomm MSM driver from
  Bjorn Andersson, Andy Gross and Kumar Gala. Among those
  new support for the IPQ8064 and MSM8x74 SoC variants.

- Support for the Freescale i.MX6 SoloX SoC variant.

- Massive improvements in the Allwinner sunxi driver from
  Boris Brezillon, Maxime Ripard and Chen-Yu Tsai.

- Renesas PFC updates from Laurent Pinchart, Kuninori
  Morimoto, Wolfram Sang and Magnus Damm.

- Cleanups and refactorings of the nVidia Tegra driver from
  Stepgen Warren.

- The Exynos driver now supports the Exynos3250 SoC.

- Intel BayTrail updates from Jin Yao, Mika Westerberg.

- The MVEBU driver now supports the Orion5x SoC
  variants, which is part of the effort of getting rid of
  the old Marvell kludges in arch/arm/mach-orion5x

- Rockchip driver updates from Heiko Stuebner.

- A ton of cleanups and janitorial patches from Axel Lin.

- Some minor fixes and improvements here and there.

Staging driver patches for 3.16-rc1

Here is the big staging driver pull request for 3.16-rc1.

Lots of stuff here, tons of cleanup patches, a few new drivers, and some
removed as well, but I think we are still adding a few thousand more
lines than we remove, due to the new drivers being bigger than the ones
deleted.

One notible bit of work did stand out, Jes Sorensen has gone on a tear,
fixing up a wireless driver to be "more sane" than it originally was
from the vendor, with over 500 patches merged here.  Good stuff, and a
number of users laptops are better off for it.

All of this has been in linux-next for a while.

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging driver fixes for 3.15-rc8

Here are some staging driver fixes for 3.15.  3 are for the speakup
drivers (one fix a regression caused in 3.15-rc, and the other 2 resolve
a tty issue found by Ben Hutchings)  The comedi and r8192e_pci driver
fixes also resolve reported issues.

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ARM: SoC updates for 3.16 (part 1)

A quite large set of SoC updates this cycle. In no particular order:

- Multi-cluster power management for Samsung Exynos, adding support for
  big.LITTLE CPU switching on EXYNOS5
- SMP support for Marvell Armada 375 and 38x
- SMP rework on Allwinner A31
- Xilinx Zynq support for SOC_BUS, big endian
- Marvell orion5x platform cleanup, modernizing the implementation and
  moving to DT.
- _Finally_ moving Samsung Exynos over to support MULTIPLATFORM, so
  that their platform can be enabled in the same kernel binary as most
  of the other v7 platforms in the tree. \o/ The work isn't quite complete,
  there's some driver fixes still needed, but the basics now work.

New SoC support added:
- Freescale i.MX6SX
- LSI Axxia AXM55xx SoCs
- Samsung EXYNOS 3250, 5260, 5410, 5420 and 5800
- STi STIH407

Plus a large set of various smaller updates for different platforms. I'm
probably missing some important one here.

ARM: SoC driver changes

SoC-near driver changes that we're merging through our tree. Mostly
because they depend on other changes we have staged, but in some cases
because the driver maintainers preferred that we did it this way.

This contains a largeish cleanup series of the omap_l3_noc bus driver,
cpuidle rework for Exynos, some reset driver conversions and a long
branch of TI EDMA fixes and cleanups, with more to come next release.

The TI EDMA cleanups is a shared branch with the dmaengine tree, with
a handful of Davinci-specific fixes on top.

After discussion at last year's KS (and some more on the mailing lists),
we are here adding a drivers/soc directory. The purpose of this is
to keep per-vendor shared code that's needed by different drivers but
that doesn't fit into the MFD (nor drivers/platform) model. We expect
to keep merging contents for this hierarchy through arm-soc so we can
keep an eye on what the vendors keep adding here and not making it a
free-for-all to shove in crazy stuff.

ARM: SoC low-priority fixes for 3.16

A small selection of fixes coming in late during the release cycle and
not being critical enough for 3.15 inclusion.

ARM: SoC devicetree updates for 3.16

As with previous release, this continues to be among the largest branches
we merge, with lots of new contents.

New things for this release are among other things:

- DTSI contents for the new SoCs supported in 3.16 (see SoC pull request)
- Qualcomm APQ8064 and APQ8084 SoCs and eval boards
- Nvidia Jetson TK1 development board (Tegra T124-based)

Two new SoCs that didn't need enough new platform code to stand out
enough for me to notice when writing the SoC tag, but that adds new DT
contents are:

- TI DRA72
- Marvell Berlin 2Q

ARM: SoC defconfig updates for 3.16

A number of defconfig updates for 3.16. We have quite a few of these as
new SoCs and drivers get supported, and we've started collecting them
in one shared branch.

We started halfway through the merge window so there'll be some small
conflicts with other branches on nearby changes, but nothing major. Next
cycle should be smoother.

ARM: SoC cleanups for 3.16

Cleanups for 3.16. Among these are:

- A bunch of misc cleanups for Broadcom platforms, mostly housekeeping
- Enabling Common Clock Framework on the older s3c24xx Samsung chipsets
- Cleanup of the Versatile Express system controller code, moving it to syscon
- Power management cleanups for OMAP platforms

+ a handful of other cleanups across the place

This is the bulk of GPIO changes for the v3.16 series:

- We are finalizing and fixing up the gpiochip irqchip helpers
  bringing a helpful irqchip implementation into the gpiolib
  core and avoiding duplicate code and, more importantly,
  duplicate bug fixes:

  - Support for using the helpers with threaded interrupt
    handlers as used on sleeping GPIO-irqchips

  - Do not set up hardware triggers for edges or levels if
    the default IRQ type is IRQ_TYPE_NONE - some drivers
    would exploit the fact that you could get default
    initialization of the IRQ type from the core at probe()
    but if no default type is set up from the helper, we
    should not call the driver to configure anything. Wait
    until a consumer requests the interrupt instead.

  - Make the irqchip helpers put the GPIO irqs into their
    own lock class. The GPIO irqchips can often emit
    (harmless, but annoying) lockdep warnings about recursions
    when they are in fact just cascaded IRQs. By putting
    them into their own lock class we help the lockdep core
    to keep track of things.

  - Switch the tc3589x GPIO expanders to use the irqchip
    helpers

  - Switch the OMAP GPIO driver to use the irqchip helpers

  - Add some documentation for the irqchip helpers

  - select IRQ_DOMAIN when using the helpers since some
    platforms may not be using this by default and it's a
    strict dependency.

- Continued GPIO descriptor refactoring:

  - Remove the one instance of gpio_to_desc() from the
    device tree code, making the OF GPIO code use GPIO
    descriptors only.

  - Introduce gpiod_get_optional() and
    gpiod_get_optional_index() akin to the similar
    regulator functions for cases where the use of GPIO
    is optional and not strictly required.

  - Make of_get_named_gpiod_flags() private - we do not
    want to unnecessarily expose APIs to drivers that
    make the gpiolib harder than necessary to maintain
    and refactor. Privatize this function.

- Support "-gpio" suffix for the OF GPIO retrieveal path.
  We used to look for "foo-gpios" or just "gpios" in device
  tree nodes, but it turns out that some drivers with a
  single GPIO line will just state "foo-gpio" (singularis).
  Sigh. Support this with a fallback looking for it, as
  this simplifies driver code and handles it in core code.

- Switch the ACPI GPIO core to fetch GPIOs with the
  *_cansleep function variants as the GPIO operation
  region handler can sleep, and shall be able to handle
  gpiochips that sleep.

- Tons of cleanups and janitorial work from Jingoo Han,
  Axel Lin, Javier Martinez Canillas and Abdoulaye Berthe.
  Notably Jingoo cut off a ton of pointless OOM messages.

- Incremental development and fixes for various drivers,
  nothing really special here.

This is the 3.12.21 stable release

Linux 3.15-rc8

This is the 3.14.5 stable release

This is the 3.10.41 stable release

ACPI and power management fixes for 3.15-rc8

 - A workqueue is destroyed too early during the ACPI thermal driver
   module unload which leads to a NULL pointer dereference in the
   driver's remove callback.  Fix from Aaron Lu.

 - A wrong argument is passed to devm_regulator_get_optional() in
   the probe routine of the cpu0 cpufreq driver which leads to
   resource leaks if the driver is unbound from the cpufreq
   platform device.  Fix from Lucas Stach.

 - A lock is missing in cpufreq_governor_dbs() which leads to
   memory corruption and NULL pointer dereferences during
   system suspend/resume, for example.  Fix from Bibek Basu.

/