Tags

Linux 4.9-rc4

phy: for 4.9 -rc

phy fixes:
*) Add a empty function for phy_reset when CONFIG_GENERIC_PHY is not set
*) change the phy lookup table for da8xx-usb to match it with the name
   present in the board configuraion file (used for non-dt boot)
*) Fix incorrect programming sequence in w.r.t deassert of phy_rst
   in phy-rockchip-pcie
*) Fix to avoid NULL pointer dereferencing error in sun4i phy

Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>

ASoC: Fixes for v4.9

The most important fix in here is a change which removes the #error
making the topology API unusable as-is since we have recently discovered
some production uses on Chromebooks so need to acknowledge that what
we've got there now is an ABI.

There's also a very big batch of driver specific fixes here which have
kept on being delayed due to more arriving so the update is another of
these bigger than I would like ones.  There is one especially big one in
there, for the Qualcomm code which fixes simultaneous playback and
capture which was broken during the merge window.  The diff for that is
large because it moves blocks of code to different functions but it's
functionally fairly simple and if it breaks it should have been very
obvious in testing.

Some GPIO fixes for the v4.9 series:

- Fix a nasty file descriptor leak when getting line handles.

- A fix for a cleanup that seemed innocent but created a problem
  for drivers instantiating several gpiochips for one single
  OF node.

- Fix a unpredictable problem using irq_domain_simple() in the
  mvebu driver by converting it to a lineas irqdomain.

drm fixes for amdgpu, radeon, intel, imx and virtio-gpu

media fixes for v4.9-rc4

This is the 4.4.30 stable release

media fixes for v4.9-rc4

regulator: Fix ramp_delay warnings for v4.9

A new warning was introduced for missing information about the time that
regulators take to power on in v4.9.  This is in theory a real issue but
for most practical regulators the communication overhead of talking to
the device is greater than the ramp time so a lot of drivers don't set
it and the warning is far too noisy without identifying practical
issues.  Just remove the warning for now.

This is the 4.4.29 stable release

This is the 4.8.6 stable release

wireless-drivers fixes for 4.9

iwlwifi

* some fixes for suspend/resume with unified FW images
* a fix for a false-positive lockdep report
* a fix for multi-queue that caused an unnecessary 1 second latency
* a fix for an ACPI parsing bug that caused a misleading error message

brcmfmac

* fix a variable uninitialised warning in brcmf_cfg80211_start_ap()

Linux 4.9-rc3

driver core fixes for 4.9-rc3

Here are two small driver core / kernfs fixes for 4.9-rc3.  One makes
the Kconfig entry for DEBUG_TEST_DRIVER_REMOVE a bit more explicit that
this is a crazy thing to enable for a distro kernel (thanks for trying
Fedora!), the other resolves an issue with vim opening kernfs files
(sysfs, configfs, etc.).

Both have been in linux-next with no reported issues.

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

Among various cleanups and improvements, we have the following:
 * client FILS authentication support in mac80211 (Jouni)
 * AP/VLAN multicast improvements (Michael Braun)
 * config/advertising support for differing beacon intervals on
   multiple virtual interfaces (Purushottam Kushwaha, myself)
 * deprecate the old WDS mode for cfg80211-based drivers, the
   mode is hardly usable since it doesn't support any "modern"
   features like WPA encryption (2003), HT (2009) or VHT (2014),
   I'm not even sure WEP (introduced in 1997) could be done.

sound fixes for 4.9-rc3

Here contains the usual stuff -- the fixups and quirks for HD-audio
and USB-audio, in addition to a bad regression fix in ALSA sequencer
timer since 4.8, and a trivial fix for asihpi PCI driver.

This is the 4.4.28 stable release

This is the 4.8.5 stable release