cascardo/linux.git
7 years agoMerge tag 'renesas-rcar-sysc2-for-v4.8' of git://git.kernel.org/pub/scm/linux/kernel...
Olof Johansson [Thu, 7 Jul 2016 05:12:04 +0000 (22:12 -0700)]
Merge tag 'renesas-rcar-sysc2-for-v4.8' of git://git./linux/kernel/git/horms/renesas into next/drivers

Second Round of Renesas ARM Based SoC R-Car SYSC Updates for v4.8

* Prepare for handling SYSC interrupt configuration purely
  from DT in the rcar-sysc driver for new SoCs, while preserving
  backward compatibility with old DTBs for R-Car H1, H2, and M2-W
* Add R8A7792 support

* tag 'renesas-rcar-sysc2-for-v4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/horms/renesas:
  soc: renesas: rcar-sysc: Improve SYSC interrupt config in legacy wrapper
  soc: renesas: rcar-sysc: Move SYSC interrupt config to rcar-sysc driver
  soc: renesas: rcar-sysc: Make rcar_sysc_init() init the PM domains
  soc: renesas: rcar-sysc: Fix uninitialized error code in rcar_sysc_pd_init()
  soc: renesas: rcar-sysc: add R8A7792 support

Signed-off-by: Olof Johansson <olof@lixom.net>
7 years agovideo: clps711x-fb: Changing the compatibility string to match with the smallest...
Alexander Shiyan [Sat, 4 Jun 2016 07:10:02 +0000 (10:10 +0300)]
video: clps711x-fb: Changing the compatibility string to match with the smallest supported chip

This patch changes the compatibility string to match with the smallest
supported chip (EP7209). Since the DT-support for this CPU is not yet
announced, this change is safe.

Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
7 years agoinput: clps711x-keypad: Changing the compatibility string to match with the smallest...
Alexander Shiyan [Sat, 4 Jun 2016 07:10:01 +0000 (10:10 +0300)]
input: clps711x-keypad: Changing the compatibility string to match with the smallest supported chip

This patch changes the compatibility string to match with the smallest
supported chip (EP7209). Since the DT-support for this CPU is not yet
announced, this change is safe.

Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
7 years agopwm: clps711x: Changing the compatibility string to match with the smallest supported...
Alexander Shiyan [Sat, 4 Jun 2016 07:09:58 +0000 (10:09 +0300)]
pwm: clps711x: Changing the compatibility string to match with the smallest supported chip

This patch changes the compatibility string to match with the smallest
supported chip (EP7209). Since the DT-support for this CPU is not yet
announced, this change is safe.

Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
7 years agoserial: clps711x: Changing the compatibility string to match with the smallest suppor...
Alexander Shiyan [Sat, 4 Jun 2016 07:09:57 +0000 (10:09 +0300)]
serial: clps711x: Changing the compatibility string to match with the smallest supported chip

This patch changes the compatibility string to match with the smallest
supported chip (EP7209). Since the DT-support for this CPU is not yet
announced, this change is safe.

Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
7 years agoirqchip: clps711x: Changing the compatibility string to match with the smallest suppo...
Alexander Shiyan [Sat, 4 Jun 2016 07:09:56 +0000 (10:09 +0300)]
irqchip: clps711x: Changing the compatibility string to match with the smallest supported chip

This patch changes the compatibility string to match with the smallest
supported chip (EP7209). Since the DT-support for this CPU is not yet
announced, this change is safe.

Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
7 years agoclocksource: clps711x: Changing the compatibility string to match with the smallest...
Alexander Shiyan [Sat, 4 Jun 2016 07:09:55 +0000 (10:09 +0300)]
clocksource: clps711x: Changing the compatibility string to match with the smallest supported chip

This patch changes the compatibility string to match with the smallest
supported chip (EP7209). Since the DT-support for this CPU is not yet
announced, this change is safe.

Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
7 years agoclk: clps711x: Changing the compatibility string to match with the smallest supported...
Alexander Shiyan [Sat, 4 Jun 2016 07:09:54 +0000 (10:09 +0300)]
clk: clps711x: Changing the compatibility string to match with the smallest supported chip

This patch changes the compatibility string to match with the smallest
supported chip (EP7209). Since the DT-support for this CPU is not yet
announced, this change is safe.

Signed-off-by: Alexander Shiyan <shc_work@mail.ru>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
7 years agoMerge tag 'samsung-drivers-4.8-2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Olof Johansson [Wed, 6 Jul 2016 05:43:41 +0000 (22:43 -0700)]
Merge tag 'samsung-drivers-4.8-2' of git://git./linux/kernel/git/krzk/linux into next/drivers

Samsung drivers/soc update for v4.8, part 2:
1. Endian-friendly fixes.
2. Make SROMC driver explicitly non-module.

* tag 'samsung-drivers-4.8-2' of git://git.kernel.org/pub/scm/linux/kernel/git/krzk/linux:
  cpufreq: s5pv210: use relaxed IO accesors
  memory: samsung: exynos-srom: make it explicitly non-modular
  memory: samsung: endian fixes for IO

Signed-off-by: Olof Johansson <olof@lixom.net>
7 years agoMerge tag 'qcom-drivers-for-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git...
Olof Johansson [Tue, 5 Jul 2016 05:30:11 +0000 (22:30 -0700)]
Merge tag 'qcom-drivers-for-4.8' of git://git./linux/kernel/git/agross/linux into next/drivers

Qualcomm ARM Based Driver Updates for v4.8

* Rework of SCM driver
* Add file patterns for Qualcomm Maintainers entry
* Add worker for wcnss_ctrl signaling
* Fixes for smp2p
* Update smem_state properties to match documentation
* Add SCM Peripheral Authentication service
* Expose SCM PAS command 10 as a reset controller

* tag 'qcom-drivers-for-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/agross/linux:
  firmware: qcom: scm: Expose PAS command 10 as reset-controller
  firmware: qcom: scm: Peripheral Authentication Service
  soc: qcom: Update properties for smem state referencing
  soc: qcom: smp2p: Drop io-accessors
  soc: qcom: smp2p: Correct addressing of outgoing value
  soc: qcom: wcnss_ctrl: Make wcnss_ctrl parent the other components
  firmware: qcom: scm: Add support for ARM64 SoCs
  firmware: qcom: scm: Convert to streaming DMA APIS
  firmware: qcom: scm: Generalize shared error map
  firmware: qcom: scm: Use atomic SCM for cold boot
  firmware: qcom: scm: Convert SCM to platform driver
  MAINTAINERS: Add file patterns for qcom device tree bindings

Signed-off-by: Olof Johansson <olof@lixom.net>
7 years agoMerge tag 'gpmc-omap-for-v4.8' of https://github.com/rogerq/linux into next/drivers
Olof Johansson [Tue, 5 Jul 2016 05:20:25 +0000 (22:20 -0700)]
Merge tag 'gpmc-omap-for-v4.8' of https://github.com/rogerq/linux into next/drivers

OMAP-GPMC: driver updates for v4.8

* Make GPMC driver non modular as it is going to
be built-in only till arch/arm/mach-omap2 users
are removed.

* Fix build if CONFIG_OF is not defined.

* tag 'gpmc-omap-for-v4.8' of https://github.com/rogerq/linux:
  memory: omap-gpmc: Move gpio functions out of #ifdef CONFIG_OF
  memory: omap-gpmc: make it explicitly non-modular

Signed-off-by: Olof Johansson <olof@lixom.net>
7 years agoMerge tag 'scpi-updates-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/sudeep...
Olof Johansson [Tue, 5 Jul 2016 04:32:07 +0000 (21:32 -0700)]
Merge tag 'scpi-updates-4.8' of git://git./linux/kernel/git/sudeep.holla/linux into next/drivers

SCPI updates and fixes for v4.8

1. Adds support for device power state management using generic power
   domains and runtime PM

2. Other minor/miscellaneous fixes to the driver

* tag 'scpi-updates-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/sudeep.holla/linux:
  firmware: scpi: add device power domain support using genpd
  Documentation: add DT bindings for ARM SCPI power domains
  firmware: arm_scpi: add support for device power state management
  firmware: arm_scpi: make it depend on MAILBOX instead of ARM_MHU
  firmware: arm_scpi: mark scpi_get_sensor_value as static
  firmware: arm_scpi: remove dvfs_get packed structure

Signed-off-by: Olof Johansson <olof@lixom.net>
7 years agoMerge tag 'at91-ab-4.8-drivers2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Olof Johansson [Tue, 5 Jul 2016 03:59:31 +0000 (20:59 -0700)]
Merge tag 'at91-ab-4.8-drivers2' of git://git./linux/kernel/git/abelloni/linux into next/drivers

Drivers for 4.8 #2:
 - Make memory drivers explicitly non-modular

* tag 'at91-ab-4.8-drivers2' of git://git.kernel.org/pub/scm/linux/kernel/git/abelloni/linux:
  memory: atmel-ebi: make it explicitly non-modular
  memory: atmel-sdramc: make it explicitly non-modular

Signed-off-by: Olof Johansson <olof@lixom.net>
7 years agosoc: renesas: rcar-sysc: Improve SYSC interrupt config in legacy wrapper
Geert Uytterhoeven [Tue, 28 Jun 2016 14:10:34 +0000 (16:10 +0200)]
soc: renesas: rcar-sysc: Improve SYSC interrupt config in legacy wrapper

Align SYSC interrupt configuration in the legacy wrapper with the DT
version:
  - Mask SYSC interrupt sources before enabling them (doesn't matter
    much as they're disabled at the GIC level anyway),
  - Make sure not to clear reserved SYSCIMR bits that were set before.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Ulrich Hecht <ulrich.hecht+renesas@gmail.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
7 years agosoc: renesas: rcar-sysc: Move SYSC interrupt config to rcar-sysc driver
Geert Uytterhoeven [Tue, 28 Jun 2016 14:10:33 +0000 (16:10 +0200)]
soc: renesas: rcar-sysc: Move SYSC interrupt config to rcar-sysc driver

On R-Car H1 and Gen2, the SYSC interrupt registers are always configured
using hardcoded values in platform code. For R-Car Gen2, values are
provided for H2 and M2-W only, other SoCs are not yet supported, and
never will be.

Move this configuration from SoC-specific platform code to the
rcar_sysc_init() wrapper, so it can be skipped if the SYSC is configured
from DT. This would be the case not only for H1, H2, and M2-W using a
modern DTS, but also for other R-Car Gen2 SoCs not supported by the
platform code, relying purely on DT.

There is no longer a need to return the mapped register block, hence
make the function return void.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Ulrich Hecht <ulrich.hecht+renesas@gmail.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
7 years agosoc: renesas: rcar-sysc: Make rcar_sysc_init() init the PM domains
Geert Uytterhoeven [Tue, 28 Jun 2016 14:10:32 +0000 (16:10 +0200)]
soc: renesas: rcar-sysc: Make rcar_sysc_init() init the PM domains

Let rcar_sysc_init() trigger initialization of the SYSC PM domains from
DT if called before the early_initcall.
On failure, it falls back to mapping the passed register block, as
before.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
7 years agosoc: renesas: rcar-sysc: Fix uninitialized error code in rcar_sysc_pd_init()
Geert Uytterhoeven [Tue, 28 Jun 2016 14:10:31 +0000 (16:10 +0200)]
soc: renesas: rcar-sysc: Fix uninitialized error code in rcar_sysc_pd_init()

On success, rcar_sysc_pd_init() returns an uninitialized error code.
Use the return value of of_genpd_add_provider_onecell() to fix this.

This went unnoticed, as early_initcall() doesn't care about the return
value.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
7 years agofirmware: qcom: scm: Expose PAS command 10 as reset-controller
Bjorn Andersson [Fri, 17 Jun 2016 17:40:43 +0000 (10:40 -0700)]
firmware: qcom: scm: Expose PAS command 10 as reset-controller

PAS command 10 is used to assert and deassert the MSS reset via
TrustZone, expose this as a reset-controller to mimic the direct
access case.

Cc: Stephen Boyd <sboyd@codeaurora.org>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Acked-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Andy Gross <andy.gross@linaro.org>
7 years agofirmware: qcom: scm: Peripheral Authentication Service
Bjorn Andersson [Wed, 23 Sep 2015 19:56:12 +0000 (12:56 -0700)]
firmware: qcom: scm: Peripheral Authentication Service

This adds the Peripheral Authentication Service (PAS) interface to the
Qualcomm SCM interface. The API is used to authenticate and boot a range
of external processors in various Qualcomm platforms.

Signed-off-by: Bjorn Andersson <bjorn.andersson@sonymobile.com>
Signed-off-by: Andy Gross <andy.gross@linaro.org>
7 years agosoc: qcom: Update properties for smem state referencing
Bjorn Andersson [Fri, 10 Jun 2016 04:11:22 +0000 (21:11 -0700)]
soc: qcom: Update properties for smem state referencing

Update the property names to match device tree bindings, the correct
values should be qcom,smem-states and qcom,smem-state-names.

Also update the #qcom,smem-state-cells for consistency, before we merge
any users of these properties.

Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Andy Gross <andy.gross@linaro.org>
7 years agosoc: qcom: smp2p: Drop io-accessors
Bjorn Andersson [Fri, 10 Jun 2016 00:22:57 +0000 (17:22 -0700)]
soc: qcom: smp2p: Drop io-accessors

SMEM is now mapped write-combine and we can use memcpy to access the
name of the entires.

Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Andy Gross <andy.gross@linaro.org>
7 years agosoc: qcom: smp2p: Correct addressing of outgoing value
Bjorn Andersson [Fri, 10 Jun 2016 00:22:56 +0000 (17:22 -0700)]
soc: qcom: smp2p: Correct addressing of outgoing value

The valid_entries index should not be incremented until after we have
acquired the pointer to the value, or we will read and write data one
item off.

Fixes: 50e99641413e ("soc: qcom: smp2p: Qualcomm Shared Memory Point to Point")
Cc: stable@vger.kernel.org
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Andy Gross <andy.gross@linaro.org>
7 years agosoc: qcom: wcnss_ctrl: Make wcnss_ctrl parent the other components
Bjorn Andersson [Mon, 6 Jun 2016 23:58:20 +0000 (16:58 -0700)]
soc: qcom: wcnss_ctrl: Make wcnss_ctrl parent the other components

We need the signal from wcnss_ctrl indicating that the firmware is up
and running before we can communicate with the other components of the
chip. So make these other components children of the wcnss_ctrl device,
so they can be probed in order.

The process seems to take between 1/2-5 seconds, so this is done in a
worker, instead of holding up the probe.

Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Andy Gross <andy.gross@linaro.org>
7 years agofirmware: qcom: scm: Add support for ARM64 SoCs
Kumar Gala [Fri, 3 Jun 2016 23:25:26 +0000 (18:25 -0500)]
firmware: qcom: scm: Add support for ARM64 SoCs

Add an implementation of the SCM interface that works on ARM64 SoCs.  This
is used by things like determine if we have HDCP support or not on the
system.

Signed-off-by: Kumar Gala <galak@codeaurora.org>
Signed-off-by: Andy Gross <andy.gross@linaro.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
7 years agofirmware: qcom: scm: Convert to streaming DMA APIS
Andy Gross [Fri, 3 Jun 2016 23:25:25 +0000 (18:25 -0500)]
firmware: qcom: scm: Convert to streaming DMA APIS

This patch converts the Qualcomm SCM driver to use the streaming DMA APIs
for communication buffers.  This is being done so that the
secure_flush_area call can be removed.  Using the DMA APIs will also make
the SCM32 symmetric to the coming SCM64 code.

Signed-off-by: Andy Gross <andy.gross@linaro.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
7 years agofirmware: qcom: scm: Generalize shared error map
Andy Gross [Fri, 3 Jun 2016 23:25:24 +0000 (18:25 -0500)]
firmware: qcom: scm: Generalize shared error map

This patch moves the qcom_scm_remap_error function to the include file
where can be used by both the 32 and 64 bit versions of the code.

Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
Acked-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Andy Gross <agross@codeaurora.org>
Signed-off-by: Andy Gross <andy.gross@linaro.org>
7 years agofirmware: qcom: scm: Use atomic SCM for cold boot
Andy Gross [Fri, 3 Jun 2016 23:25:23 +0000 (18:25 -0500)]
firmware: qcom: scm: Use atomic SCM for cold boot

This patch changes the cold_set_boot_addr function to use atomic SCM
calls.  cold_set_boot_addr required adding qcom_scm_call_atomic2 to
support the two arguments going to the smc call.  Using atomic removes
the need for memory allocation and instead places all arguments in
registers.

Signed-off-by: Andy Gross <andy.gross@linaro.org>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
Acked-by: Bjorn Andersson <bjorn.andersson@linaro.org>
7 years agofirmware: qcom: scm: Convert SCM to platform driver
Andy Gross [Fri, 3 Jun 2016 23:25:22 +0000 (18:25 -0500)]
firmware: qcom: scm: Convert SCM to platform driver

This patch converts the Qualcomm SCM firmware driver into a platform
driver.  It also adds clock management for firmware calls which require
clocks to be enabled during the duration of their execution.  Rate
setting of the core clock is also in place for higher performance.

Signed-off-by: Andy Gross <andy.gross@linaro.org>
Acked-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Reviewed-by: Stephen Boyd <sboyd@codeaurora.org>
7 years agoMAINTAINERS: Add file patterns for qcom device tree bindings
Geert Uytterhoeven [Sun, 22 May 2016 09:05:41 +0000 (11:05 +0200)]
MAINTAINERS: Add file patterns for qcom device tree bindings

Submitters of device tree binding documentation may forget to CC
the subsystem maintainer if this is missing.

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Andy Gross <andy.gross@linaro.org>
Cc: David Brown <david.brown@linaro.org>
Cc: linux-arm-msm@vger.kernel.org
Cc: linux-soc@vger.kernel.org
Signed-off-by: Andy Gross <andy.gross@linaro.org>
7 years agocpufreq: s5pv210: use relaxed IO accesors
Ben Dooks [Wed, 22 Jun 2016 09:36:37 +0000 (10:36 +0100)]
cpufreq: s5pv210: use relaxed IO accesors

The use of __raw IO accesors is not endian safe and should be used
sparingly. The relaxed variants should be as lightweight and also
are endian safe.

Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
7 years agomemory: atmel-ebi: make it explicitly non-modular
Paul Gortmaker [Fri, 17 Jun 2016 00:37:48 +0000 (20:37 -0400)]
memory: atmel-ebi: make it explicitly non-modular

The Kconfig currently controlling compilation of this code is:

drivers/memory/Kconfig:config ATMEL_EBI
drivers/memory/Kconfig: bool "Atmel EBI driver"

...meaning that it currently is not being built as a module by anyone.

Lets remove the few remaining modular references, so that when reading
the driver there is no doubt it is builtin-only.

Since module_platform_driver() uses the same init level priority as
builtin_platform_driver() the init ordering remains unchanged with
this commit.

Also note that MODULE_DEVICE_TABLE is a no-op for non-modular code.

We also delete the MODULE_LICENSE tag etc. since all that information
is already contained at the top of the file in the comments.

Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Acked-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
7 years agomemory: atmel-sdramc: make it explicitly non-modular
Paul Gortmaker [Fri, 17 Jun 2016 00:37:43 +0000 (20:37 -0400)]
memory: atmel-sdramc: make it explicitly non-modular

The Kconfig for this option is currently:

config ATMEL_SDRAMC
        bool "Atmel (Multi-port DDR-)SDRAM Controller"

...meaning that it currently is not being built as a module by anyone.
Lets remove the couple traces of modularity, so that when reading the
driver there is no doubt it is builtin-only.

Since module_init translates to device_initcall in the non-modular
case, the init ordering remains unchanged with this commit.  An
alternate init level might be worth considering at a later date.

Also note that MODULE_DEVICE_TABLE is a no-op for non-modular code.

We also delete the MODULE_LICENSE tag etc. since all that information
was (or is now) contained at the top of the file in the comments.

Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
7 years agofirmware: scpi: add device power domain support using genpd
Sudeep Holla [Thu, 2 Jun 2016 15:34:03 +0000 (16:34 +0100)]
firmware: scpi: add device power domain support using genpd

This patch hooks up the support for device power domain provided by
SCPI using the Linux generic power domain infrastructure.

Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
Tested-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Tested-by: Jon Medhurst <tixy@linaro.org>
Reviewed-by: Jon Medhurst <tixy@linaro.org>
Reviewed-by: Kevin Hilman <khilman@baylibre.com>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
7 years agoDocumentation: add DT bindings for ARM SCPI power domains
Sudeep Holla [Sun, 5 Jun 2016 17:50:50 +0000 (18:50 +0100)]
Documentation: add DT bindings for ARM SCPI power domains

The System Control Processor (SCP) provides peripheral devices with
power domains that can be enabled and disabled viathe System Control
and Power Interface (SCPI) Message Protocol. Add bindings to allow
probing of these device power domians.

Cc: Rob Herring <robh+dt@kernel.org>
Tested-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Tested-by: Jon Medhurst <tixy@linaro.org>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Reviewed-by: Jon Medhurst <tixy@linaro.org>
Reviewed-by: Kevin Hilman <khilman@baylibre.com>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
7 years agofirmware: arm_scpi: add support for device power state management
Sudeep Holla [Wed, 20 Apr 2016 13:05:14 +0000 (14:05 +0100)]
firmware: arm_scpi: add support for device power state management

SCPI protocol supports device power state management. This deals with
power states of various peripheral devices in the system other than the
core compute subsystem.

This patch adds support for the power state management of those
peripheral devices.

Tested-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Tested-by: Jon Medhurst <tixy@linaro.org>
Reviewed-by: Jon Medhurst <tixy@linaro.org>
Reviewed-by: Kevin Hilman <khilman@baylibre.com>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
7 years agomemory: samsung: exynos-srom: make it explicitly non-modular
Paul Gortmaker [Fri, 17 Jun 2016 00:37:47 +0000 (20:37 -0400)]
memory: samsung: exynos-srom: make it explicitly non-modular

The Kconfig currently controlling compilation of this code is:

memory/samsung/Kconfig:config EXYNOS_SROM
memory/samsung/Kconfig: bool "Exynos SROM controller driver" if COMPILE_TEST

...meaning that it currently is not being built as a module by anyone.

Lets remove the modular code that is essentially orphaned, so that
when reading the driver there is no doubt it is builtin-only.

We explicitly disallow a driver unbind, since that doesn't have a
sensible use case anyway, and it allows us to drop the ".remove"
code for non-modular drivers.

Since module_platform_driver() uses the same init level priority as
builtin_platform_driver() the init ordering remains unchanged with
this commit.

Also note that MODULE_DEVICE_TABLE is a no-op for non-modular code.

We also delete the MODULE_LICENSE tag etc. since all that information
is already contained at the top of the file in the comments.

Cc: Pankaj Dubey <pankaj.dubey@samsung.com>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
7 years agomemory: samsung: endian fixes for IO
Ben Dooks [Fri, 17 Jun 2016 17:08:10 +0000 (18:08 +0100)]
memory: samsung: endian fixes for IO

Use the relaxed versions of the IO accessors to avoid any issues
if running in big endian.

Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
7 years agoMerge tag 'arm-soc/for-4.8/drivers' of http://github.com/Broadcom/stblinux into next...
Olof Johansson [Mon, 20 Jun 2016 05:55:16 +0000 (22:55 -0700)]
Merge tag 'arm-soc/for-4.8/drivers' of github.com/Broadcom/stblinux into next/drivers

This pull request contains drivers related changes for Broadcom SoCs:

- Florian moves drivers/soc/brcmstb to drivers/soc/bcm/brcmstb to be consistent with
  how other SoCs are doing it

- Chris provides a reset driver which is common to the BCM21664 and BCM23550 SoCs

- Ben fixes a warning by providing the appropriate include file

* tag 'arm-soc/for-4.8/drivers' of http://github.com/Broadcom/stblinux:
  soc: brcmstb: fix warning from missing include
  power: Introduce Broadcom kona reset driver
  soc: Move brcmstb to bcm/brcmstb

Signed-off-by: Olof Johansson <olof@lixom.net>
7 years agoMerge tag 'samsung-drivers-exynos-mfc-4.8' of git://git.kernel.org/pub/scm/linux...
Olof Johansson [Mon, 20 Jun 2016 05:39:15 +0000 (22:39 -0700)]
Merge tag 'samsung-drivers-exynos-mfc-4.8' of git://git./linux/kernel/git/krzk/linux into next/drivers

Topic branch for Exynos MFC changes for v4.8:
Pull s5p-mfc changes from media tree so the arm/mach-exynos code
could be removed. The bindings are converted to generic reserved memory
bindings.

* tag 'samsung-drivers-exynos-mfc-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/krzk/linux:
  ARM: dts: exynos: Enable MFC device on Exynos4412 Odroid boards
  ARM: dts: exynos: Convert MFC device to generic reserved memory bindings
  ARM: EXYNOS: Remove code for MFC custom reserved memory handling
  media: s5p-mfc: add iommu support
  media: s5p-mfc: replace custom reserved memory handling code with generic one
  media: s5p-mfc: use generic reserved memory bindings
  of: reserved_mem: add support for using more than one region for given device
  media: set proper max seg size for devices on Exynos SoCs
  media: vb2-dma-contig: add helper for setting dma max seg size
  s5p-mfc: Fix race between s5p_mfc_probe() and s5p_mfc_open()
  s5p-mfc: Add release callback for memory region devs
  s5p-mfc: Set device name for reserved memory region devs

Signed-off-by: Olof Johansson <olof@lixom.net>
7 years agomemory: omap-gpmc: Move gpio functions out of #ifdef CONFIG_OF
Roger Quadros [Fri, 17 Jun 2016 07:16:50 +0000 (10:16 +0300)]
memory: omap-gpmc: Move gpio functions out of #ifdef CONFIG_OF

The gpio related functions were mistakenly built only if CONFIG_OF
is defined. They are needed even otherwise and will cause build
failures if CONFIG_OF is not defined. Move the gpio functions
outside #ifdef CONFIG_OF.

Fixes: d2d00862dfbb ("memory: omap-gpmc: Support general purpose input for WAITPINs")

Signed-off-by: Roger Quadros <rogerq@ti.com>
7 years agomemory: omap-gpmc: make it explicitly non-modular
Paul Gortmaker [Fri, 17 Jun 2016 00:37:45 +0000 (20:37 -0400)]
memory: omap-gpmc: make it explicitly non-modular

The Kconfig currently controlling compilation of this code is:

drivers/memory/Kconfig:config OMAP_GPMC
drivers/memory/Kconfig:  bool

...meaning that it currently is not being built as a module by anyone.

Lets remove the modular code that is essentially orphaned, so that
when reading the driver there is no doubt it is builtin-only.

Since module_init was not in use by this code, the init ordering
remains unchanged with this commit.

Also note that MODULE_DEVICE_TABLE is a no-op for non-modular code.

We don't replace module.h with init.h since the file already has that.

Cc: Roger Quadros <rogerq@ti.com>
Cc: Tony Lindgren <tony@atomide.com>
Cc: linux-omap@vger.kernel.org
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
7 years agosoc: renesas: rcar-sysc: add R8A7792 support
Sergei Shtylyov [Sun, 12 Jun 2016 20:56:50 +0000 (23:56 +0300)]
soc: renesas: rcar-sysc: add R8A7792 support

Add support for R-Car V2H (R8A7792) SoC power areas to the SYSC driver.

Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
7 years agoMerge tag 'reset-for-4.8-2' of git://git.pengutronix.de/git/pza/linux into next/drivers
Olof Johansson [Tue, 14 Jun 2016 17:30:46 +0000 (10:30 -0700)]
Merge tag 'reset-for-4.8-2' of git://git.pengutronix.de/git/pza/linux into next/drivers

Reset controller changes for v4.8, part 2

- add Amlogic Meson SoC Reset Controller driver

* tag 'reset-for-4.8-2' of git://git.pengutronix.de/git/pza/linux:
  dt-bindings: reset: Add bindings for the Meson SoC Reset Controller
  reset: Add support for the Amlogic Meson SoC Reset Controller

Signed-off-by: Olof Johansson <olof@lixom.net>
7 years agoMerge tag 'at91-ab-4.8-drivers' of git://git.kernel.org/pub/scm/linux/kernel/git...
Olof Johansson [Tue, 14 Jun 2016 17:18:54 +0000 (10:18 -0700)]
Merge tag 'at91-ab-4.8-drivers' of git://git./linux/kernel/git/abelloni/linux into next/drivers

Drivers for 4.8:
 - Add a driver for the EBI

* tag 'at91-ab-4.8-drivers' of git://git.kernel.org/pub/scm/linux/kernel/git/abelloni/linux:
  memory: atmel-ebi: add DT bindings documentation
  memory: add Atmel EBI (External Bus Interface) driver

Signed-off-by: Olof Johansson <olof@lixom.net>
7 years agoMerge tag 'renesas-rcar-sysc-for-v4.8' of git://git.kernel.org/pub/scm/linux/kernel...
Olof Johansson [Mon, 13 Jun 2016 22:39:46 +0000 (15:39 -0700)]
Merge tag 'renesas-rcar-sysc-for-v4.8' of git://git./linux/kernel/git/horms/renesas into next/drivers

Renesas ARM Based SoC R-Car SYSC Updates for v4.8

e0c98b9171ee soc: renesas: rcar-sysc: Add support for R-Car M3-W power areas
74699228b99d soc: renesas: Add r8a7796 SYSC PM Domain Binding Definitions
2ff1bf77e4d4 soc: renesas: rcar-sysc: Document r8a7796 support

* tag 'renesas-rcar-sysc-for-v4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/horms/renesas:
  soc: renesas: rcar-sysc: Add support for R-Car M3-W power areas
  soc: renesas: Add r8a7796 SYSC PM Domain Binding Definitions
  soc: renesas: rcar-sysc: Document r8a7796 support

Signed-off-by: Olof Johansson <olof@lixom.net>
7 years agoMerge tag 'samsung-drivers-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git...
Olof Johansson [Mon, 13 Jun 2016 21:28:37 +0000 (14:28 -0700)]
Merge tag 'samsung-drivers-4.8' of git://git./linux/kernel/git/krzk/linux into next/drivers

Samsung drivers/soc update for v4.8:
1. Move the power domain driver from arm/mach-exynos and prepare
   for supporting ARMv8.
2. Add COMPILE_TEST.

* tag 'samsung-drivers-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/krzk/linux:
  soc: samsung: pm_domains: Enable COMPILE_TEST for build coverage
  soc: samsung: pm_domains: Prepare for supporting ARMv8 Exynos
  ARM: EXYNOS: Move pm_domains driver to drivers/soc/samsung

Signed-off-by: Olof Johansson <olof@lixom.net>
7 years agofirmware: arm_scpi: make it depend on MAILBOX instead of ARM_MHU
Sudeep Holla [Sun, 5 Jun 2016 17:23:21 +0000 (18:23 +0100)]
firmware: arm_scpi: make it depend on MAILBOX instead of ARM_MHU

ARM_SCPI_PROTOCOL can be used with any mailbox and not just ARM MHU
mailbox controller. This patch drops it's dependency on ARM_MHU and
make it depend on just mailbox framework.

Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
7 years agofirmware: arm_scpi: mark scpi_get_sensor_value as static
Sudeep Holla [Tue, 23 Feb 2016 16:21:16 +0000 (16:21 +0000)]
firmware: arm_scpi: mark scpi_get_sensor_value as static

scpi_get_sensor_value like other scpi operations needs to be static.
This patch marks it as static to be consistent with others.

Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
7 years agofirmware: arm_scpi: remove dvfs_get packed structure
Sudeep Holla [Wed, 20 Apr 2016 12:25:01 +0000 (13:25 +0100)]
firmware: arm_scpi: remove dvfs_get packed structure

dvfs_get packed structure is used to read the DVFS/OPP index from the
firmware. It just contains a single byte that needs no packing making
the whole structure defination unnecessary.

This patch replaces the unnecessary dvfs_get packed structure with an
unsigned byte.

Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
7 years agosoc: brcmstb: fix warning from missing include
Ben Dooks [Tue, 7 Jun 2016 17:54:54 +0000 (18:54 +0100)]
soc: brcmstb: fix warning from missing include

The brcmstb_biuctrl_init() is defined in the soc specific header
file, but wasn't included in the driver file. Fix the following
warning by including <linux/soc/brcmstb/brcmstb.h> in the driver:

drivers/soc/brcmstb/biuctrl.c:101:13: warning: symbol 'brcmstb_biuctrl_init' was not declared. Should it be static?

Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
7 years agopower: Introduce Broadcom kona reset driver
Chris Brand [Wed, 11 May 2016 21:36:17 +0000 (14:36 -0700)]
power: Introduce Broadcom kona reset driver

This driver supports reset on both BCM21664 and BCM23550.
Code is being moved from arch/arm/mach-bcm/board_bcm21664.c

Signed-off-by: Chris Brand <chris.brand@broadcom.com>
Acked-By: Sebastian Reichel <sre@kernel.org>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
7 years agoARM: dts: exynos: Enable MFC device on Exynos4412 Odroid boards
Marek Szyprowski [Tue, 24 May 2016 13:31:30 +0000 (15:31 +0200)]
ARM: dts: exynos: Enable MFC device on Exynos4412 Odroid boards

Enable support for Multimedia Codec (MFC) device for all Exynos4412-based
Odroid boards.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Javier Martinez Canillas <javier@osg.samsung.com>
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
7 years agoARM: dts: exynos: Convert MFC device to generic reserved memory bindings
Marek Szyprowski [Tue, 24 May 2016 13:31:29 +0000 (15:31 +0200)]
ARM: dts: exynos: Convert MFC device to generic reserved memory bindings

This patch replaces custom properties for defining reserved memory
regions with generic reserved memory bindings for MFC video codec
device.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Reviewed-by: Javier Martinez Canillas <javier@osg.samsung.com>
Tested-by: Javier Martinez Canillas <javier@osg.samsung.com>
[k.kozlowski: Add Samsung copyrights]
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
7 years agoARM: EXYNOS: Remove code for MFC custom reserved memory handling
Marek Szyprowski [Tue, 24 May 2016 13:31:28 +0000 (15:31 +0200)]
ARM: EXYNOS: Remove code for MFC custom reserved memory handling

Once MFC driver has been converted to generic reserved memory bindings,
there is no need for custom memory reservation code.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Javier Martinez Canillas <javier@osg.samsung.com>
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
7 years agoMerge branch 'for-v4.8/media/exynos-mfc' of git://linuxtv.org/snawrocki/samsung into...
Krzysztof Kozlowski [Mon, 6 Jun 2016 07:21:09 +0000 (09:21 +0200)]
Merge branch 'for-v4.8/media/exynos-mfc' of git://linuxtv.org/snawrocki/samsung into for-v4.8/exynos-mfc

The Exynos MFC driver was converted to use a generic reserved memory
regions and IOMMU. Pull these changes before removal of mach-specific
code from mach-exynos and before changing DT bindings.

7 years agosoc: renesas: rcar-sysc: Add support for R-Car M3-W power areas
Geert Uytterhoeven [Mon, 30 May 2016 17:05:11 +0000 (19:05 +0200)]
soc: renesas: rcar-sysc: Add support for R-Car M3-W power areas

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
7 years agosoc: renesas: Add r8a7796 SYSC PM Domain Binding Definitions
Geert Uytterhoeven [Mon, 30 May 2016 17:05:10 +0000 (19:05 +0200)]
soc: renesas: Add r8a7796 SYSC PM Domain Binding Definitions

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
7 years agosoc: renesas: rcar-sysc: Document r8a7796 support
Geert Uytterhoeven [Mon, 30 May 2016 17:05:09 +0000 (19:05 +0200)]
soc: renesas: rcar-sysc: Document r8a7796 support

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
7 years agoMerge tag 'reset-for-4.8' of git://git.pengutronix.de/git/pza/linux into next/drivers
Olof Johansson [Fri, 3 Jun 2016 19:18:58 +0000 (12:18 -0700)]
Merge tag 'reset-for-4.8' of git://git.pengutronix.de/git/pza/linux into next/drivers

Reset controller changes for v4.8

- add devm_reset_controller_register API to get rid of a good
  amount of boilerplate across reset drivers.
- move reset controller Kconfig options into a sub-menu
- fix inconsistent return values, let all reset_control_get
  stubs return -ENOTSUPP

* tag 'reset-for-4.8' of git://git.pengutronix.de/git/pza/linux:
  reset: Return -ENOTSUPP when not configured
  reset: oxnas: Use devm register API and get rid of platform remove
  reset: fix Kconfig menu to include reset drivers in sub-menu
  reset: zynq: use devm_reset_controller_register()
  reset: socfpga: use devm_reset_controller_register()
  reset: sunxi: use devm_reset_controller_register()
  reset: pistachio: use devm_reset_controller_register()
  reset: ath79: use devm_reset_controller_register()
  reset: add devm_reset_controller_register API

Signed-off-by: Olof Johansson <olof@lixom.net>
7 years agomedia: s5p-mfc: add iommu support
Marek Szyprowski [Tue, 24 May 2016 13:31:27 +0000 (15:31 +0200)]
media: s5p-mfc: add iommu support

This patch adds support for IOMMU to s5p-mfc device driver. MFC firmware
is limited and it cannot use the default configuration. If IOMMU is
available, the patch disables the default DMA address space
configuration and creates a new address space of size limited to 256M
and base address set to 0x20000000.

For now the same address space is shared by both 'left' and 'right'
memory channels, because the DMA/IOMMU frameworks do not support
configuring them separately. This is not optimal, but besides limiting
total address space available has no other drawbacks (MFC firmware
supports 256M of address space per each channel).

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Javier Martinez Canillas <javier@osg.samsung.com>
Tested-by: Javier Martinez Canillas <javier@osg.samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
7 years agomedia: s5p-mfc: replace custom reserved memory handling code with generic one
Marek Szyprowski [Tue, 24 May 2016 13:31:26 +0000 (15:31 +0200)]
media: s5p-mfc: replace custom reserved memory handling code with generic one

This patch removes custom code for initialization and handling of
reserved memory regions in s5p-mfc driver and replaces it with generic
reserved memory regions api.

s5p-mfc driver now handles two reserved memory regions defined by
generic reserved memory bindings. Support for non-dt platform has been
removed, because all supported platforms have been already converted to
device tree.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Javier Martinez Canillas <javier@osg.samsung.com>
Tested-by: Javier Martinez Canillas <javier@osg.samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
7 years agomedia: s5p-mfc: use generic reserved memory bindings
Marek Szyprowski [Tue, 24 May 2016 13:31:25 +0000 (15:31 +0200)]
media: s5p-mfc: use generic reserved memory bindings

Use generic reserved memory bindings and mark old, custom properties
as obsoleted.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Javier Martinez Canillas <javier@osg.samsung.com>
Acked-by: Rob Herring <robh@kernel.org>
Acked-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
7 years agoof: reserved_mem: add support for using more than one region for given device
Marek Szyprowski [Tue, 24 May 2016 13:31:24 +0000 (15:31 +0200)]
of: reserved_mem: add support for using more than one region for given device

This patch allows device drivers to initialize more than one reserved
memory region assigned to given device. When driver needs to use more
than one reserved memory region, it should allocate child devices and
initialize regions by index for each of its child devices.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
7 years agomedia: set proper max seg size for devices on Exynos SoCs
Marek Szyprowski [Tue, 24 May 2016 07:16:07 +0000 (09:16 +0200)]
media: set proper max seg size for devices on Exynos SoCs

All multimedia devices found on Exynos SoCs support only contiguous
buffers, so set DMA max segment size to DMA_BIT_MASK(32) to let memory
allocator to correctly create contiguous memory mappings.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
7 years agomedia: vb2-dma-contig: add helper for setting dma max seg size
Marek Szyprowski [Tue, 24 May 2016 07:16:06 +0000 (09:16 +0200)]
media: vb2-dma-contig: add helper for setting dma max seg size

Add a helper function for device drivers to set DMA's max_seg_size.
Setting it to largest possible value lets DMA-mapping API always create
contiguous mappings in DMA address space. This is essential for all
devices, which use dma-contig videobuf2 memory allocator and shared
buffers.

Till now, the only case when vb2-dma-contig really 'worked' was a case
where userspace provided USERPTR buffer, which was in fact mmaped
contiguous buffer from the other v4l2/drm device. Also DMABUF made of
contiguous buffer worked only when its exporter did not split it into
several chunks in the scatter-list. Any other buffer failed, regardless
of the arch/platform used and the presence of the IOMMU of the device bus.

This patch provides interface to fix this issue.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
7 years agos5p-mfc: Fix race between s5p_mfc_probe() and s5p_mfc_open()
Javier Martinez Canillas [Tue, 3 May 2016 20:27:18 +0000 (16:27 -0400)]
s5p-mfc: Fix race between s5p_mfc_probe() and s5p_mfc_open()

The s5p_mfc_probe() function registers the video devices before all the
resources needed by s5p_mfc_open() are correctly initalized.

So if s5p_mfc_open() function is called before s5p_mfc_probe() finishes
(since the video dev is already registered), a NULL pointer dereference
will happen due s5p_mfc_open() accessing uninitialized vars such as the
struct s5p_mfc_dev .watchdog_timer and .mfc_ops fields.

An example is following BUG caused by add_timer() getting a NULL pointer:

[   45.765374] kernel BUG at kernel/time/timer.c:790!
[   45.765381] Internal error: Oops - BUG: 0 [#1] PREEMPT SMP ARM
...
[   45.766149] [<c016fdf4>] (mod_timer) from [<bf181d18>] (s5p_mfc_open+0x274/0x4d4 [s5p_mfc])
[   45.766416] [<bf181d18>] (s5p_mfc_open [s5p_mfc]) from [<bf0214a0>] (v4l2_open+0x9c/0x100 [videodev])
[   45.766547] [<bf0214a0>] (v4l2_open [videodev]) from [<c01e355c>] (chrdev_open+0x9c/0x178)
[   45.766575] [<c01e355c>] (chrdev_open) from [<c01dceb4>] (do_dentry_open+0x1e0/0x300)
[   45.766595] [<c01dceb4>] (do_dentry_open) from [<c01ec2f0>] (path_openat+0x800/0x10d4)
[   45.766610] [<c01ec2f0>] (path_openat) from [<c01ed8b8>] (do_filp_open+0x5c/0xc0)
[   45.766624] [<c01ed8b8>] (do_filp_open) from [<c01de218>] (do_sys_open+0x10c/0x1bc)
[   45.766642] [<c01de218>] (do_sys_open) from [<c01078c0>] (ret_fast_syscall+0x0/0x3c)
[   45.766655] Code: eaffffe3 e3a00001 e28dd008 e8bd81f0 (e7f001f2)

Fix it by registering the video devs as the last step in s5p_mfc_probe().

Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
7 years agos5p-mfc: Add release callback for memory region devs
Javier Martinez Canillas [Tue, 3 May 2016 20:27:17 +0000 (16:27 -0400)]
s5p-mfc: Add release callback for memory region devs

When s5p_mfc_remove() calls put_device() for the reserved memory region
devs, the driver core warns that the dev doesn't have a release callback:

WARNING: CPU: 0 PID: 591 at drivers/base/core.c:251 device_release+0x8c/0x90
Device 's5p-mfc-l' does not have a release() function, it is broken and must be fixed.

Also, the declared DMA memory using dma_declare_coherent_memory() isn't
relased so add a dev .release that calls dma_release_declared_memory().

Cc: <stable@vger.kernel.org>
Fixes: 6e83e6e25eb4 ("[media] s5p-mfc: Fix kernel warning on memory init")
Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
7 years agos5p-mfc: Set device name for reserved memory region devs
Javier Martinez Canillas [Tue, 3 May 2016 20:27:16 +0000 (16:27 -0400)]
s5p-mfc: Set device name for reserved memory region devs

The devices don't have a name set, so makes dev_name() returns NULL which
makes harder to identify the devices that are causing issues, for example:

WARNING: CPU: 2 PID: 616 at drivers/base/core.c:251 device_release+0x8c/0x90
Device '(null)' does not have a release() function, it is broken and must be fixed.

And after setting the device name:

WARNING: CPU: 0 PID: 591 at drivers/base/core.c:251 device_release+0x8c/0x90
Device 's5p-mfc-l' does not have a release() function, it is broken and must be fixed.

Cc: <stable@vger.kernel.org>
Fixes: 6e83e6e25eb4 ("[media] s5p-mfc: Fix kernel warning on memory init")
Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
7 years agomemory: atmel-ebi: add DT bindings documentation
Boris Brezillon [Mon, 23 May 2016 07:44:55 +0000 (09:44 +0200)]
memory: atmel-ebi: add DT bindings documentation

The EBI (External Bus Interface) is used to access external peripherals
(NOR, SRAM, NAND, and other specific devices like ethernet controllers).
Each device is assigned a CS line and an address range and can have its
own configuration (timings, access mode, bus width, ...).
This driver provides a generic DT binding to configure a device according
to its requirements.
For specific device controllers (like the NAND one) the SMC timings
should be configured by the controller driver through the matrix and smc
syscon regmaps.

Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
7 years agomemory: add Atmel EBI (External Bus Interface) driver
Boris Brezillon [Mon, 23 May 2016 07:44:54 +0000 (09:44 +0200)]
memory: add Atmel EBI (External Bus Interface) driver

The EBI (External Bus Interface) is used to access external peripherals
(NOR, SRAM, NAND, and other specific devices like ethernet controllers).
Each device is assigned a CS line and an address range and can have its
own configuration (timings, access mode, bus width, ...).
This driver provides a generic DT binding to configure a device according
to its requirements.
For specific device controllers (like the NAND one) the SMC timings
should be configured by the controller driver through the matrix and
smc syscon regmaps.

Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
7 years agodt-bindings: reset: Add bindings for the Meson SoC Reset Controller
Neil Armstrong [Mon, 30 May 2016 13:27:16 +0000 (15:27 +0200)]
dt-bindings: reset: Add bindings for the Meson SoC Reset Controller

Add DT bindings for the Meson SoC Reset Controller documentation and the
associated include file.

Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
7 years agoreset: Add support for the Amlogic Meson SoC Reset Controller
Neil Armstrong [Mon, 30 May 2016 13:27:15 +0000 (15:27 +0200)]
reset: Add support for the Amlogic Meson SoC Reset Controller

This patch adds the platform driver for the Amlogic Meson SoC Reset
Controller.

The Meson8b and GXBB SoCs are supported.

Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
7 years agoreset: Return -ENOTSUPP when not configured
John Youn [Tue, 31 May 2016 23:55:01 +0000 (16:55 -0700)]
reset: Return -ENOTSUPP when not configured

Prior to commit 6c96f05c8bb8 ("reset: Make [of_]reset_control_get[_foo]
functions wrappers"), the "optional" functions returned -ENOTSUPP when
CONFIG_RESET_CONTROLLER was not set.

Revert back to the old behavior by changing the new
__devm_reset_control_get() and __of_reset_control_get() functions to
return ERR_PTR(-ENOTSUPP) when compiled without CONFIG_RESET_CONTROLLER.

Otherwise they will return -EINVAL causing users to think that an error
occurred when CONFIG_RESET_CONTROLLER is not set.

Fixes: 6c96f05c8bb8 ("reset: Make [of_]reset_control_get[_foo] functions wrappers")
Signed-off-by: John Youn <johnyoun@synopsys.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
7 years agosoc: Move brcmstb to bcm/brcmstb
Florian Fainelli [Fri, 20 May 2016 18:35:57 +0000 (11:35 -0700)]
soc: Move brcmstb to bcm/brcmstb

Unify the different Broadcom SoCs directory and have everybody live
under drivers/soc/bcm/*.

Acked-by: Scott Branden <scott.branden@broadcom.com>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
7 years agoreset: oxnas: Use devm register API and get rid of platform remove
Neil Armstrong [Fri, 20 May 2016 12:21:37 +0000 (14:21 +0200)]
reset: oxnas: Use devm register API and get rid of platform remove

Use the brand new devm_reset_controller_register() API to get rid of
the platform driver remove callback.

Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
7 years agoreset: fix Kconfig menu to include reset drivers in sub-menu
Masahiro Yamada [Tue, 3 May 2016 06:29:52 +0000 (15:29 +0900)]
reset: fix Kconfig menu to include reset drivers in sub-menu

In "make menuconfig", reset drivers are currently lined up together
with the reset sub-system menu, like this:

    -*- Reset Controller Support  ----
    < > Hi6220 Reset Driver

(It also means, the menu "Reset Controller Support" is always empty.)

"Hi6220 Reset Driver" should go into the sub-menu of the
"Reset Controller Support".

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
7 years agoreset: zynq: use devm_reset_controller_register()
Masahiro Yamada [Sun, 1 May 2016 10:37:03 +0000 (19:37 +0900)]
reset: zynq: use devm_reset_controller_register()

Use devm_reset_controller_register() for the reset controller
registration and drop the .remove callback.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
7 years agoreset: socfpga: use devm_reset_controller_register()
Masahiro Yamada [Sun, 1 May 2016 10:37:02 +0000 (19:37 +0900)]
reset: socfpga: use devm_reset_controller_register()

Use devm_reset_controller_register() for the reset controller
registration and drop the .remove callback.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
7 years agoreset: sunxi: use devm_reset_controller_register()
Masahiro Yamada [Sun, 1 May 2016 10:37:01 +0000 (19:37 +0900)]
reset: sunxi: use devm_reset_controller_register()

Use devm_reset_controller_register() for the reset controller
registration and drop the .remove callback.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
7 years agoreset: pistachio: use devm_reset_controller_register()
Masahiro Yamada [Sun, 1 May 2016 10:37:00 +0000 (19:37 +0900)]
reset: pistachio: use devm_reset_controller_register()

Use devm_reset_controller_register() for the reset controller
registration and drop the .remove callback.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
7 years agoreset: ath79: use devm_reset_controller_register()
Masahiro Yamada [Sun, 1 May 2016 10:36:58 +0000 (19:36 +0900)]
reset: ath79: use devm_reset_controller_register()

Use devm_reset_controller_register() for the reset controller
registration and remove the unregister call from the .remove callback.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
7 years agoreset: add devm_reset_controller_register API
Masahiro Yamada [Sun, 1 May 2016 10:36:57 +0000 (19:36 +0900)]
reset: add devm_reset_controller_register API

Add a device managed API for reset_controller_register().

This helps in reducing code in .remove callbacks and sometimes
dropping .remove callbacks entirely.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Acked-by: Laxman Dewangan <ldewangan@nvidia.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
7 years agosoc: samsung: pm_domains: Enable COMPILE_TEST for build coverage
Krzysztof Kozlowski [Tue, 10 May 2016 14:31:26 +0000 (16:31 +0200)]
soc: samsung: pm_domains: Enable COMPILE_TEST for build coverage

Introduce a platform selectable symbol EXYNOS_PM_DOMAINS which can be
also toggled on by COMPILE_TEST for some build coverage.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
7 years agosoc: samsung: pm_domains: Prepare for supporting ARMv8 Exynos
Krzysztof Kozlowski [Tue, 10 May 2016 13:52:21 +0000 (15:52 +0200)]
soc: samsung: pm_domains: Prepare for supporting ARMv8 Exynos

The ARMv8 Exynos family (Exynos5433 and Exynos7420) uses different value
(0xf instead of 0x7) for controlling the power domain on/off registers
(both for control and for status).

Choose the value depending on the compatible. This prepares the driver
for supporting ARMv8 SoCs.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Reviewed-by: Javier Martinez Canillas <javier@osg.samsung.com>
7 years agoARM: EXYNOS: Move pm_domains driver to drivers/soc/samsung
Krzysztof Kozlowski [Tue, 10 May 2016 13:52:20 +0000 (15:52 +0200)]
ARM: EXYNOS: Move pm_domains driver to drivers/soc/samsung

Exynos PM domains driver does not have mach-specific dependencies so it
can be safely moved out of arm/mach-exynos to drivers/soc. This in
future will allow re-using it on ARM64 boards.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Reviewed-by: Javier Martinez Canillas <javier@osg.samsung.com>
7 years agoLinux 4.7-rc1 v4.7-rc1
Linus Torvalds [Sun, 29 May 2016 16:29:24 +0000 (09:29 -0700)]
Linux 4.7-rc1

7 years agohash_string: Fix zero-length case for !DCACHE_WORD_ACCESS
George Spelvin [Sun, 29 May 2016 12:05:56 +0000 (08:05 -0400)]
hash_string: Fix zero-length case for !DCACHE_WORD_ACCESS

The self-test was updated to cover zero-length strings; the function
needs to be updated, too.

Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: George Spelvin <linux@sciencehorizons.net>
Fixes: fcfd2fbf22d2 ("fs/namei.c: Add hashlen_string() function")
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years agoRename other copy of hash_string to hashlen_string
George Spelvin [Sun, 29 May 2016 05:26:41 +0000 (01:26 -0400)]
Rename other copy of hash_string to hashlen_string

The original name was simply hash_string(), but that conflicted with a
function with that name in drivers/base/power/trace.c, and I decided
that calling it "hashlen_" was better anyway.

But you have to do it in two places.

[ This caused build errors for architectures that don't define
  CONFIG_DCACHE_WORD_ACCESS   - Linus ]

Signed-off-by: George Spelvin <linux@sciencehorizons.net>
Reported-by: Guenter Roeck <linux@roeck-us.net>
Fixes: fcfd2fbf22d2 ("fs/namei.c: Add hashlen_string() function")
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years agohpfs: implement the show_options method
Mikulas Patocka [Tue, 24 May 2016 20:49:18 +0000 (22:49 +0200)]
hpfs: implement the show_options method

The HPFS filesystem used generic_show_options to produce string that is
displayed in /proc/mounts.  However, there is a problem that the options
may disappear after remount.  If we mount the filesystem with option1
and then remount it with option2, /proc/mounts should show both option1
and option2, however it only shows option2 because the whole option
string is replaced with replace_mount_options in hpfs_remount_fs.

To fix this bug, implement the hpfs_show_options function that prints
options that are currently selected.

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years agoaffs: fix remount failure when there are no options changed
Mikulas Patocka [Tue, 24 May 2016 20:48:33 +0000 (22:48 +0200)]
affs: fix remount failure when there are no options changed

Commit c8f33d0bec99 ("affs: kstrdup() memory handling") checks if the
kstrdup function returns NULL due to out-of-memory condition.

However, if we are remounting a filesystem with no change to
filesystem-specific options, the parameter data is NULL.  In this case,
kstrdup returns NULL (because it was passed NULL parameter), although no
out of memory condition exists.  The mount syscall then fails with
ENOMEM.

This patch fixes the bug.  We fail with ENOMEM only if data is non-NULL.

The patch also changes the call to replace_mount_options - if we didn't
pass any filesystem-specific options, we don't call
replace_mount_options (thus we don't erase existing reported options).

Fixes: c8f33d0bec99 ("affs: kstrdup() memory handling")
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Cc: stable@vger.kernel.org # v4.1+
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years agohpfs: fix remount failure when there are no options changed
Mikulas Patocka [Tue, 24 May 2016 20:47:00 +0000 (22:47 +0200)]
hpfs: fix remount failure when there are no options changed

Commit ce657611baf9 ("hpfs: kstrdup() out of memory handling") checks if
the kstrdup function returns NULL due to out-of-memory condition.

However, if we are remounting a filesystem with no change to
filesystem-specific options, the parameter data is NULL.  In this case,
kstrdup returns NULL (because it was passed NULL parameter), although no
out of memory condition exists.  The mount syscall then fails with
ENOMEM.

This patch fixes the bug.  We fail with ENOMEM only if data is non-NULL.

The patch also changes the call to replace_mount_options - if we didn't
pass any filesystem-specific options, we don't call
replace_mount_options (thus we don't erase existing reported options).

Fixes: ce657611baf9 ("hpfs: kstrdup() out of memory handling")
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years agoMerge branch 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus
Linus Torvalds [Sat, 28 May 2016 23:41:39 +0000 (16:41 -0700)]
Merge branch 'upstream' of git://git.linux-mips.org/ralf/upstream-linus

Pull more MIPS updates from Ralf Baechle:
 "This is the secondnd batch of MIPS patches for 4.7. Summary:

  CPS:
   - Copy EVA configuration when starting secondary VPs.

  EIC:
   - Clear Status IPL.

  Lasat:
   - Fix a few off by one bugs.

  lib:
   - Mark intrinsics notrace.  Not only are the intrinsics
     uninteresting, it would cause infinite recursion.

  MAINTAINERS:
   - Add file patterns for MIPS BRCM device tree bindings.
   - Add file patterns for mips device tree bindings.

  MT7628:
   - Fix MT7628 pinmux typos.
   - wled_an pinmux gpio.
   - EPHY LEDs pinmux support.

  Pistachio:
   - Enable KASLR

  VDSO:
   - Build microMIPS VDSO for microMIPS kernels.
   - Fix aliasing warning by building with `-fno-strict-aliasing' for
     debugging but also tracing them might result in recursion.

  Misc:
   - Add missing FROZEN hotplug notifier transitions.
   - Fix clk binding example for varioius PIC32 devices.
   - Fix cpu interrupt controller node-names in the DT files.
   - Fix XPA CPU feature separation.
   - Fix write_gc0_* macros when writing zero.
   - Add inline asm encoding helpers.
   - Add missing VZ accessor microMIPS encodings.
   - Fix little endian microMIPS MSA encodings.
   - Add 64-bit HTW fields and fix its configuration.
   - Fix sigreturn via VDSO on microMIPS kernel.
   - Lots of typo fixes.
   - Add definitions of SegCtl registers and use them"

* 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus: (49 commits)
  MIPS: Add missing FROZEN hotplug notifier transitions
  MIPS: Build microMIPS VDSO for microMIPS kernels
  MIPS: Fix sigreturn via VDSO on microMIPS kernel
  MIPS: devicetree: fix cpu interrupt controller node-names
  MIPS: VDSO: Build with `-fno-strict-aliasing'
  MIPS: Pistachio: Enable KASLR
  MIPS: lib: Mark intrinsics notrace
  MIPS: Fix 64-bit HTW configuration
  MIPS: Add 64-bit HTW fields
  MAINTAINERS: Add file patterns for mips device tree bindings
  MAINTAINERS: Add file patterns for mips brcm device tree bindings
  MIPS: Simplify DSP instruction encoding macros
  MIPS: Add missing tlbinvf/XPA microMIPS encodings
  MIPS: Fix little endian microMIPS MSA encodings
  MIPS: Add missing VZ accessor microMIPS encodings
  MIPS: Add inline asm encoding helpers
  MIPS: Spelling fix lets -> let's
  MIPS: VR41xx: Fix typo
  MIPS: oprofile: Fix typo
  MIPS: math-emu: Fix typo
  ...

7 years agofs: fix binfmt_aout.c build error
Guenter Roeck [Sat, 28 May 2016 22:26:02 +0000 (15:26 -0700)]
fs: fix binfmt_aout.c build error

Various builds (such as i386:allmodconfig) fail with

  fs/binfmt_aout.c:133:2: error: expected identifier or '(' before 'return'
  fs/binfmt_aout.c:134:1: error: expected identifier or '(' before '}' token

[ Oops. My bad, I had stupidly thought that "allmodconfig" covered this
  on x86-64 too, but it obviously doesn't.  Egg on my face.  - Linus ]

Fixes: 5d22fc25d4fc ("mm: remove more IS_ERR_VALUE abuses")
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
7 years agoMerge branch 'hash' of git://ftp.sciencehorizons.net/linux
Linus Torvalds [Sat, 28 May 2016 23:15:25 +0000 (16:15 -0700)]
Merge branch 'hash' of git://ftp.sciencehorizons.net/linux

Pull string hash improvements from George Spelvin:
 "This series does several related things:

   - Makes the dcache hash (fs/namei.c) useful for general kernel use.

     (Thanks to Bruce for noticing the zero-length corner case)

   - Converts the string hashes in <linux/sunrpc/svcauth.h> to use the
     above.

   - Avoids 64-bit multiplies in hash_64() on 32-bit platforms.  Two
     32-bit multiplies will do well enough.

   - Rids the world of the bad hash multipliers in hash_32.

     This finishes the job started in commit 689de1d6ca95 ("Minimal
     fix-up of bad hashing behavior of hash_64()")

     The vast majority of Linux architectures have hardware support for
     32x32-bit multiply and so derive no benefit from "simplified"
     multipliers.

     The few processors that do not (68000, h8/300 and some models of
     Microblaze) have arch-specific implementations added.  Those
     patches are last in the series.

   - Overhauls the dcache hash mixing.

     The patch in commit 0fed3ac866ea ("namei: Improve hash mixing if
     CONFIG_DCACHE_WORD_ACCESS") was an off-the-cuff suggestion.
     Replaced with a much more careful design that's simultaneously
     faster and better.  (My own invention, as there was noting suitable
     in the literature I could find.  Comments welcome!)

   - Modify the hash_name() loop to skip the initial HASH_MIX().  This
     would let us salt the hash if we ever wanted to.

   - Sort out partial_name_hash().

     The hash function is declared as using a long state, even though
     it's truncated to 32 bits at the end and the extra internal state
     contributes nothing to the result.  And some callers do odd things:

      - fs/hfs/string.c only allocates 32 bits of state
      - fs/hfsplus/unicode.c uses it to hash 16-bit unicode symbols not bytes

   - Modify bytemask_from_count to handle inputs of 1..sizeof(long)
     rather than 0..sizeof(long)-1.  This would simplify users other
     than full_name_hash"

  Special thanks to Bruce Fields for testing and finding bugs in v1.  (I
  learned some humbling lessons about "obviously correct" code.)

  On the arch-specific front, the m68k assembly has been tested in a
  standalone test harness, I've been in contact with the Microblaze
  maintainers who mostly don't care, as the hardware multiplier is never
  omitted in real-world applications, and I haven't heard anything from
  the H8/300 world"

* 'hash' of git://ftp.sciencehorizons.net/linux:
  h8300: Add <asm/hash.h>
  microblaze: Add <asm/hash.h>
  m68k: Add <asm/hash.h>
  <linux/hash.h>: Add support for architecture-specific functions
  fs/namei.c: Improve dcache hash function
  Eliminate bad hash multipliers from hash_32() and  hash_64()
  Change hash_64() return value to 32 bits
  <linux/sunrpc/svcauth.h>: Define hash_str() in terms of hashlen_string()
  fs/namei.c: Add hashlen_string() function
  Pull out string hash to <linux/stringhash.h>

7 years agoh8300: Add <asm/hash.h>
George Spelvin [Wed, 25 May 2016 18:19:49 +0000 (14:19 -0400)]
h8300: Add <asm/hash.h>

This will improve the performance of hash_32() and hash_64(), but due
to complete lack of multi-bit shift instructions on H8, performance will
still be bad in surrounding code.

Designing H8-specific hash algorithms to work around that is a separate
project.  (But if the maintainers would like to get in touch...)

Signed-off-by: George Spelvin <linux@sciencehorizons.net>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: uclinux-h8-devel@lists.sourceforge.jp
7 years agomicroblaze: Add <asm/hash.h>
George Spelvin [Wed, 25 May 2016 15:06:09 +0000 (11:06 -0400)]
microblaze: Add <asm/hash.h>

Microblaze is an FPGA soft core that can be configured various ways.

If it is configured without a multiplier, the standard __hash_32()
will require a call to __mulsi3, which is a slow software loop.

Instead, use a shift-and-add sequence for the constant multiply.
GCC knows how to do this, but it's not as clever as some.

Signed-off-by: George Spelvin <linux@sciencehorizons.net>
Cc: Alistair Francis <alistair.francis@xilinx.com>
Cc: Michal Simek <michal.simek@xilinx.com>
7 years agom68k: Add <asm/hash.h>
George Spelvin [Thu, 26 May 2016 15:36:19 +0000 (11:36 -0400)]
m68k: Add <asm/hash.h>

This provides a multiply by constant GOLDEN_RATIO_32 = 0x61C88647
for the original mc68000, which lacks a 32x32-bit multiply instruction.

Yes, the amount of optimization effort put in is excessive. :-)

Shift-add chain found by Yevgen Voronenko's Hcub algorithm at
http://spiral.ece.cmu.edu/mcm/gen.html

Signed-off-by: George Spelvin <linux@sciencehorizons.net>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Greg Ungerer <gerg@linux-m68k.org>
Cc: Andreas Schwab <schwab@linux-m68k.org>
Cc: Philippe De Muyter <phdm@macq.eu>
Cc: linux-m68k@lists.linux-m68k.org
7 years ago<linux/hash.h>: Add support for architecture-specific functions
George Spelvin [Fri, 27 May 2016 02:11:51 +0000 (22:11 -0400)]
<linux/hash.h>: Add support for architecture-specific functions

This is just the infrastructure; there are no users yet.

This is modelled on CONFIG_ARCH_RANDOM; a CONFIG_ symbol declares
the existence of <asm/hash.h>.

That file may define its own versions of various functions, and define
HAVE_* symbols (no CONFIG_ prefix!) to suppress the generic ones.

Included is a self-test (in lib/test_hash.c) that verifies the basics.
It is NOT in general required that the arch-specific functions compute
the same thing as the generic, but if a HAVE_* symbol is defined with
the value 1, then equality is tested.

Signed-off-by: George Spelvin <linux@sciencehorizons.net>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Greg Ungerer <gerg@linux-m68k.org>
Cc: Andreas Schwab <schwab@linux-m68k.org>
Cc: Philippe De Muyter <phdm@macq.eu>
Cc: linux-m68k@lists.linux-m68k.org
Cc: Alistair Francis <alistai@xilinx.com>
Cc: Michal Simek <michal.simek@xilinx.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: uclinux-h8-devel@lists.sourceforge.jp
7 years agofs/namei.c: Improve dcache hash function
George Spelvin [Mon, 23 May 2016 11:43:58 +0000 (07:43 -0400)]
fs/namei.c: Improve dcache hash function

Patch 0fed3ac866 improved the hash mixing, but the function is slower
than necessary; there's a 7-instruction dependency chain (10 on x86)
each loop iteration.

Word-at-a-time access is a very tight loop (which is good, because
link_path_walk() is one of the hottest code paths in the entire kernel),
and the hash mixing function must not have a longer latency to avoid
slowing it down.

There do not appear to be any published fast hash functions that:
1) Operate on the input a word at a time, and
2) Don't need to know the length of the input beforehand, and
3) Have a single iterated mixing function, not needing conditional
   branches or unrolling to distinguish different loop iterations.

One of the algorithms which comes closest is Yann Collet's xxHash, but
that's two dependent multiplies per word, which is too much.

The key insights in this design are:

1) Barring expensive ops like multiplies, to diffuse one input bit
   across 64 bits of hash state takes at least log2(64) = 6 sequentially
   dependent instructions.  That is more cycles than we'd like.
2) An operation like "hash ^= hash << 13" requires a second temporary
   register anyway, and on a 2-operand machine like x86, it's three
   instructions.
3) A better use of a second register is to hold a two-word hash state.
   With careful design, no temporaries are needed at all, so it doesn't
   increase register pressure.  And this gets rid of register copying
   on 2-operand machines, so the code is smaller and faster.
4) Using two words of state weakens the requirement for one-round mixing;
   we now have two rounds of mixing before cancellation is possible.
5) A two-word hash state also allows operations on both halves to be
   done in parallel, so on a superscalar processor we get more mixing
   in fewer cycles.

I ended up using a mixing function inspired by the ChaCha and Speck
round functions.  It is 6 simple instructions and 3 cycles per iteration
(assuming multiply by 9 can be done by an "lea" instruction):

x ^= *input++;
y ^= x; x = ROL(x, K1);
x += y; y = ROL(y, K2);
y *= 9;

Not only is this reversible, two consecutive rounds are reversible:
if you are given the initial and final states, but not the intermediate
state, it is possible to compute both input words.  This means that at
least 3 words of input are required to create a collision.

(It also has the property, used by hash_name() to avoid a branch, that
it hashes all-zero to all-zero.)

The rotate constants K1 and K2 were found by experiment.  The search took
a sample of random initial states (I used 1023) and considered the effect
of flipping each of the 64 input bits on each of the 128 output bits two
rounds later.  Each of the 8192 pairs can be considered a biased coin, and
adding up the Shannon entropy of all of them produces a score.

The best-scoring shifts also did well in other tests (flipping bits in y,
trying 3 or 4 rounds of mixing, flipping all 64*63/2 pairs of input bits),
so the choice was made with the additional constraint that the sum of the
shifts is odd and not too close to the word size.

The final state is then folded into a 32-bit hash value by a less carefully
optimized multiply-based scheme.  This also has to be fast, as pathname
components tend to be short (the most common case is one iteration!), but
there's some room for latency, as there is a fair bit of intervening logic
before the hash value is used for anything.

(Performance verified with "bonnie++ -s 0 -n 1536:-2" on tmpfs.  I need
a better benchmark; the numbers seem to show a slight dip in performance
between 4.6.0 and this patch, but they're too noisy to quote.)

Special thanks to Bruce fields for diligent testing which uncovered a
nasty fencepost error in an earlier version of this patch.

[checkpatch.pl formatting complaints noted and respectfully disagreed with.]

Signed-off-by: George Spelvin <linux@sciencehorizons.net>
Tested-by: J. Bruce Fields <bfields@redhat.com>
7 years agoEliminate bad hash multipliers from hash_32() and hash_64()
George Spelvin [Fri, 27 May 2016 03:00:23 +0000 (23:00 -0400)]
Eliminate bad hash multipliers from hash_32() and  hash_64()

The "simplified" prime multipliers made very bad hash functions, so get rid
of them.  This completes the work of 689de1d6ca.

To avoid the inefficiency which was the motivation for the "simplified"
multipliers, hash_64() on 32-bit systems is changed to use a different
algorithm.  It makes two calls to hash_32() instead.

drivers/media/usb/dvb-usb-v2/af9015.c uses the old GOLDEN_RATIO_PRIME_32
for some horrible reason, so it inherits a copy of the old definition.

Signed-off-by: George Spelvin <linux@sciencehorizons.net>
Cc: Antti Palosaari <crope@iki.fi>
Cc: Mauro Carvalho Chehab <m.chehab@samsung.com>