Tags

This is the 4.9.121 stable release

This is the 4.4.149 stable release

This is the 3.18.119 stable release

"LA.UM.6.8.r3-01400-SDM710.0"

"LE.UM.2.3.2-02700-SDX24"

media updates for v4.19-rc1

This is the 4.9.120 stable release

This is the 4.14.63 stable release

This is the 4.17.15 stable release

This is the 4.4.148 stable release

This is the 4.18.1 stable release

This is the bulk of GPIO changes for the v4.19 kernel cycle:

Core changes:

- Add a new API for explicitly naming GPIO consumers, when needed.

- Don't let userspace set values on input lines. While we do not
  think anyone would do this crazy thing we better plug the hole
  before someone uses it and think it's a nifty feature.

- Avoid calling chip->request() for unused GPIOs.

New drivers/subdrivers:

- The Mediatek MT7621 is supported which is a big win for OpenWRT
  and similar router distributions using this chip, as it seems
  every major router manufacturer on the planet has made products
  using this chip:
  https://wikidevi.com/wiki/MediaTek_MT7621

- The Tegra 194 is now supported.

- The IT87 driver now supports IT8786E and IT8718F super-IO
  chips.

- Add support for Rockchip RK3328 in the syscon GPIO driver.

Driver changes:

- Handle the get/set_multiple() properly on MMIO chips with
  inverted direction registers. We didn't have this problem
  until a new chip appear that has get/set registers AND
  inverted direction bits, OK now we handle it.

- A patch series making more error codes percolate upward
  properly for different errors on gpiochip_lock_as_irq().

- Get/set multiple for the OMAP driver, accelerating these
  multiple line operations if possible.

- A coprocessor interface for the Aspeed driver. Sometimes a few
  GPIO lines need to be grabbed by a co-processor for doing
  automated tasks, sometimes they are available as GPIO lines.
  By adding an explicit API in this driver we make it possible
  for the two line consumers to coexist. (This work was
  made available on the ib-aspeed branch, which may be appearing
  in other pull requests.)

- Implemented .get_direction() and open drain in the SCH311x
  driver.

- Continuing cleanup of included headers in GPIO drivers.

"LA.UM.6.3.r4-05000-sdm845.0"

Devicetree updates for 4.19:

- Remove an obsolete hack for PPC32 longtrail systems

- Make of_io_request_and_map() "name" arg optional

- Add vendor prefixes for bitmain, Asus, and Y Soft

- Remove 'interrupt-parent' from bindings as it is implicit

- New properties for wm8994 audio codec

- Add 'clocks' property support to SRAM binding

- Add binding for ASPEED coprocessor interrupt controller

- Various binding spelling and link fixes

regulator: Changes for v4.19

The biggest set of changes in here is the addition of the Qualcomm RPMH
driver.  As well as the regualtor driver itself being quite large due to
the usual involved Qualcomm regulator stuff there's also some code
shared with the arm-soc tree, a bus driver required to communicate with
the hardware that actually winds up being much larger than the regulator
driver itself and a LLCC driver that was part of the same signed tag
used with the arm-soc tree.

Other than that it's a fairly standard and quiet release, highlights
include:

 - Addition of device links from regulator consumers to their
   regulators, helping the core avoid dependency issues during suspend.
 - Support for the entertainingly innovative suspend implementation in
   the BD9571MWV.
 - Support for switch regulators on the PFUZE100, this required two goes
   due to backwards compatibility issues with old DTs that were
   discovered.
 - Support for Freescale PFUZE3001 and SocioNext UniPhier.
 - The aforementioned Qualcomm RPMH driver together with the driver
   changes required to support it.

sound updates for 4.19

It's been busy summer weeks and hence lots of changes, partly for a
few new drivers and partly for a wide range of fixes.
Here are highlights:

ALSA Core:
 - Fix rawmidi buffer management, code cleanup / refactoring
 - Fix the SG-buffer page handling with incorrect fallback size
 - Fix the stall at virmidi trigger callback with a large buffer;
   also offloading and code-refactoring along with it
 - Various ALSA sequencer code cleanups

ASoC:
 - Deploy the standard snd_pcm_stop_xrun() helper in several drivers
 - Support for providing name prefixes to generic component nodes
 - Quite a few fixes for DPCM as it gains a bit wider use and more
   robust testing
 - Generalization of the DIO2125 support to a simple amplifier driver
 - Accessory detection support for the audio graph card
 - DT support for PXA AC'97 devices
 - Quirks for a number of new x86 systems
 - Support for AM Logic Meson, Everest ES7154, Intel systems with
   RT5682, Qualcomm QDSP6 and WCD9335, Realtek RT5682 and TI TAS5707

HD-audio:
 - Code refactoring in HD-audio ext codec codes to drop own classes;
   preliminary works for the upcoming legacy codec support
 - Generalized DRM audio component for the upcoming radeon / amdgpu
   support
 - Unification of mic mute-LED and GPIO support for various codecs
 - Further improvement of CA0132 codec support including Recon3D
 - Proper vga_switcheroo handling for AMD i-GPU
 - Update of model list in documentation
 - Fixups for another HP Spectre x360, Conexant codecs, power-save
   blacklist update

USB-audio:
 - Fix the invalid sample rate setup with external clock
 - Support of UAC3 selector units and processing units
 - Basic UAC3 power-domain support
 - Support for Encore mDSD and Thesycon-based DSD devices
 - Preparation for future complete callback changes

Firewire:
 - Add support for MOTU Traveler

Misc:
 - The endianess notation fixes in various drivers
 - Add fall-through comment in lots of drivers
 - Various sparse warning fixes, e.g. about PCM format types

"LA.UM.6.8.2.r1-00500-SDM710.0"

Linux 4.18

Branch point for 4.14-p-release

Branch point for 4.4-p-release