GitHub/LineageOS/android_kernel_motorola_exynos9610.git
10 years agopinctrl: baytrail: Add pull type, strength and open drain to debugfs output
Mika Westerberg [Fri, 16 May 2014 09:18:29 +0000 (12:18 +0300)]
pinctrl: baytrail: Add pull type, strength and open drain to debugfs output

In case of resolving power management or similar issues it might be useful
to have these properties included in the debugfs output.

Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Acked-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: baytrail: Register GPIO chip after chip->to_irq is set
Jin Yao [Thu, 15 May 2014 15:28:47 +0000 (18:28 +0300)]
pinctrl: baytrail: Register GPIO chip after chip->to_irq is set

If chip->to_irq is NULL ACPI GPIO helpers don't register GPIO event
handlers thus preventing any ACPI GPIO triggered events. Solve this by
calling gpiochip_add() after we have set up drivers chip->to_irq hook.

Signed-off-by: Jin Yao <yao.jin@intel.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: baytrail: Add back Baytrail-T ACPI ID
Jin Yao [Thu, 15 May 2014 15:28:46 +0000 (18:28 +0300)]
pinctrl: baytrail: Add back Baytrail-T ACPI ID

Now that the x86 dynamic IRQ allocation problem has been resolved with
commmit 62a08ae2a576 (genirq: x86: Ensure that dynamic irq allocation does
not conflict), we can add back Baytrail-T ACPI ID to the pinctrl driver.

This makes the driver to work on Asus T100 where it is needed for several
things like ACPI GPIO events and SD card detection.

References: https://bugzilla.kernel.org/show_bug.cgi?id=68291
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Jin Yao <yao.jin@intel.com>
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: sh-pfc: r8a73a4: Allow Multiplatform Build
Magnus Damm [Tue, 13 May 2014 11:37:49 +0000 (13:37 +0200)]
pinctrl: sh-pfc: r8a73a4: Allow Multiplatform Build

Add #ifdefs to allow r8a73a4 Multiplatform build. Needed
to enable r8a73a4 Multiplatform support.

Signed-off-by: Magnus Damm <damm+renesas@opensource.se>
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: sh-pfc: sh73a0: Allow Multiplatform Build
Magnus Damm [Tue, 13 May 2014 11:37:48 +0000 (13:37 +0200)]
pinctrl: sh-pfc: sh73a0: Allow Multiplatform Build

Add #ifdefs to allow sh73a0 Multiplatform build. Needed
to enable sh73a0 Multiplatform support.

Signed-off-by: Magnus Damm <damm+renesas@opensource.se>
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: sh-pfc: r8a7740: Allow Multiplatform Build
Magnus Damm [Tue, 13 May 2014 11:37:47 +0000 (13:37 +0200)]
pinctrl: sh-pfc: r8a7740: Allow Multiplatform Build

Add #ifdefs to allow r8a7740 Multiplatform build. Needed
to enable r8a7740 Multiplatform support.

Signed-off-by: Magnus Damm <damm+renesas@opensource.se>
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: sh-pfc: Don't set the pinmux_irq irq field for multiplatform
Laurent Pinchart [Tue, 13 May 2014 11:37:46 +0000 (13:37 +0200)]
pinctrl: sh-pfc: Don't set the pinmux_irq irq field for multiplatform

In the multiplatform kernel case the IRQs associated with the PFC GPIOs
are specified through DT. The pinmux_irq irq field is thus ignored by
the code, and doesn't need to be set.

This will allow removing the mach/irq.h include from pfc-*.c files that
was required for the irq_pin() macro used to initialize the irq field.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: sunxi: Enable the pinctrl Kconfig options by default
Maxime Ripard [Tue, 13 May 2014 15:23:17 +0000 (17:23 +0200)]
pinctrl: sunxi: Enable the pinctrl Kconfig options by default

Enable the freshly introduced Kconfig options whenever their matching
architecture is enabled.

Since the Kconfig symbols for these machines are going through a different
tree, keep PINCTRL_SUNXI around for the moment to avoid breaking the defconfig.
It should be removed eventually.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agodt-bindings: adapt rockchip-pinctrl doc to changed bindings
Heiko Stübner [Mon, 5 May 2014 12:00:11 +0000 (14:00 +0200)]
dt-bindings: adapt rockchip-pinctrl doc to changed bindings

Introduce the syscons for grf and pmu and deprecate the previous register
areas.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Tested-by: Max Schwarz <max.schwarz@online.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: rockchip: base regmap supplied by a syscon
Heiko Stübner [Mon, 5 May 2014 11:59:51 +0000 (13:59 +0200)]
pinctrl: rockchip: base regmap supplied by a syscon

This allows the basic registers of the general register files to be supplied
by a syscon instead of being mapped locally.

The GRF registers contain a lot more than pinctrl functions like dma, usb-phy
and general soc control and status registers, intermixed with the iomux, pull
and drive-strength registers.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Tested-by: Max Schwarz <max.schwarz@online.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: rockchip: only map bank0-pull-region when pmu regmap missing
Heiko Stübner [Mon, 5 May 2014 11:59:30 +0000 (13:59 +0200)]
pinctrl: rockchip: only map bank0-pull-region when pmu regmap missing

When the pmu registers are supplied through a syscon regmap we do not need
to map the registers ourself.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Tested-by: Max Schwarz <max.schwarz@online.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: rockchip: let pmu registers be supplied by a syscon
Heiko Stübner [Mon, 5 May 2014 11:59:09 +0000 (13:59 +0200)]
pinctrl: rockchip: let pmu registers be supplied by a syscon

Currently the pmu registers containing pin pull settings on the rk3188 are mapped
locally when bank0 is instantiated. Add an alternative that can resolve the pmu
from a syscon phandle.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Tested-by: Max Schwarz <max.schwarz@online.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: rockchip: rockchip_pinctrl in rockchip_get_bank_data
Heiko Stübner [Mon, 5 May 2014 11:58:46 +0000 (13:58 +0200)]
pinctrl: rockchip: rockchip_pinctrl in rockchip_get_bank_data

Convert rockchip_get_bank_data to use the struct rockchip_pinctrl because
later on we need to check a value from it when registering the gpio banks.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Tested-by: Max Schwarz <max.schwarz@online.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: rockchip: use regmaps instead of raw mappings
Heiko Stübner [Mon, 5 May 2014 11:58:20 +0000 (13:58 +0200)]
pinctrl: rockchip: use regmaps instead of raw mappings

This allows us to use syscons in the future.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Tested-by: Max Schwarz <max.schwarz@online.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: rockchip: do not require 2nd register area
Heiko Stübner [Mon, 5 May 2014 11:58:00 +0000 (13:58 +0200)]
pinctrl: rockchip: do not require 2nd register area

Deprecate secondary register area for rk3188 pulls. Instead use big enough
initial mapping of grf registers to catch all.

The now deprecated register is still supported though.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Tested-by: Max Schwarz <max.schwarz@online.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agoMerge tag 'sunxi-pinctrl-for-3.16' of https://github.com/mripard/linux into devel
Linus Walleij [Fri, 9 May 2014 06:47:16 +0000 (08:47 +0200)]
Merge tag 'sunxi-pinctrl-for-3.16' of https://github.com/mripard/linux into devel

Pinctrl cleanup and reworks for 3.16

This serie of patch:
  - Moves the Allwinner pinctrl driver to a folder of its own
  - removes the sunxi-pinctrl-pins header, and split the driver into a core
    one, with all the logic, and smaller drivers, one for each SoC, that
    declare the pins, and will provide to the core the set of pins.
  - And does a few cleanups here and there.

10 years agopinctrl: sunxi: Move the reset handling functions out of the core
Maxime Ripard [Sat, 26 Apr 2014 20:28:54 +0000 (22:28 +0200)]
pinctrl: sunxi: Move the reset handling functions out of the core

The way that reset is handled right now is that it is made optional for every
pinctrl driver, while actually, it isn't used at all for the main pin
controllers so far, and while it's mandatory for the A31's secondary pin
controller.

Move the reset functions out of the core and in the driver, where they can be
made mandatory.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
10 years agopinctrl: sunxi: Introduce per-driver Kconfig options
Maxime Ripard [Sat, 26 Apr 2014 15:06:57 +0000 (17:06 +0200)]
pinctrl: sunxi: Introduce per-driver Kconfig options

Add one Kconfig option for each driver. This will allow to better control which
driver is enabled, instead of having either all or nothing.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
10 years agopinctrl: sunxi: Move Allwinner A20 pinctrl driver to a driver of its own
Maxime Ripard [Fri, 18 Apr 2014 18:12:50 +0000 (20:12 +0200)]
pinctrl: sunxi: Move Allwinner A20 pinctrl driver to a driver of its own

Move the pin description to a driver specific to be.

This is the final step toward retiring pinctrl-sunxi-pins.h that used to define
all the pins for all the Allwinner SoCs in a single header, that would have in
turn result in having these structures in the final binary as many times as the
header was included.

We can finally remove that header, and remove all the driver part of the
pinctrl-sunxi core.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
10 years agopinctrl: sunxi: Move Allwinner A31 special pins driver to a driver of its own
Maxime Ripard [Fri, 18 Apr 2014 18:12:50 +0000 (20:12 +0200)]
pinctrl: sunxi: Move Allwinner A31 special pins driver to a driver of its own

Move the pin description to a driver specific to be. This is one more step
toward retiring pinctrl-sunxi-pins.h that used to define all the pins for all
the Allwinner SoCs in a single header, that would have in turn result in having
these structures in the final binary as many times as the header was included.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
10 years agopinctrl: sunxi: Move Allwinner A31 pinctrl driver to a driver of its own
Maxime Ripard [Fri, 18 Apr 2014 18:12:50 +0000 (20:12 +0200)]
pinctrl: sunxi: Move Allwinner A31 pinctrl driver to a driver of its own

Move the pin description to a driver specific to be. This is one more step
toward retiring pinctrl-sunxi-pins.h that used to define all the pins for all
the Allwinner SoCs in a single header, that would have in turn result in having
these structures in the final binary as many times as the header was included.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
10 years agopinctrl: sunxi: Move Allwinner A13 pinctrl driver to a driver of its own
Maxime Ripard [Fri, 18 Apr 2014 18:12:50 +0000 (20:12 +0200)]
pinctrl: sunxi: Move Allwinner A13 pinctrl driver to a driver of its own

Move the pin description to a driver specific to be. This is one more step
toward retiring pinctrl-sunxi-pins.h that used to define all the pins for all
the Allwinner SoCs in a single header, that would have in turn result in having
these structures in the final binary as many times as the header was included.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
10 years agopinctrl: sunxi: Move Allwinner A10s pinctrl driver to a driver of its own
Maxime Ripard [Fri, 18 Apr 2014 18:12:50 +0000 (20:12 +0200)]
pinctrl: sunxi: Move Allwinner A10s pinctrl driver to a driver of its own

Move the pin description to a driver specific to be. This is one more step
toward retiring pinctrl-sunxi-pins.h that used to define all the pins for all
the Allwinner SoCs in a single header, that would have in turn result in having
these structures in the final binary as many times as the header was included.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
10 years agopinctrl: sunxi: Move Allwinner A10 pinctrl driver to a driver of its own
Maxime Ripard [Fri, 18 Apr 2014 18:12:50 +0000 (20:12 +0200)]
pinctrl: sunxi: Move Allwinner A10 pinctrl driver to a driver of its own

Move the pin description to a driver specific to be. This is one more step
toward retiring pinctrl-sunxi-pins.h that used to define all the pins for all
the Allwinner SoCs in a single header, that would have in turn result in having
these structures in the final binary as many times as the header was included.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
10 years agopinctrl: sunxi: Libraryse the driver
Maxime Ripard [Fri, 18 Apr 2014 18:10:41 +0000 (20:10 +0200)]
pinctrl: sunxi: Libraryse the driver

This will allow to have multiple drivers using the same core code, and
eventually, retire pinctrl-sunxi-pins.h

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
10 years agopinctrl: sunxi: Switch to devm_ioremap_resource
Maxime Ripard [Sat, 26 Apr 2014 19:59:50 +0000 (21:59 +0200)]
pinctrl: sunxi: Switch to devm_ioremap_resource

The previous code was calling of_iomap, which doesn't do any resource
management, and doesn't call request_mem_region either. Use
devm_ioremap_resource that do both.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
10 years agopinctrl: sunxi: Replace hardcoded pin defines by a macro
Maxime Ripard [Thu, 24 Apr 2014 14:06:52 +0000 (16:06 +0200)]
pinctrl: sunxi: Replace hardcoded pin defines by a macro

We previously had an evergrowing (and exhaustive) list of the pins that could
be used on any Allwinner SoCs. These defines were then used by each pinctrl
driver to declare the list of functions for this pin. Since it's pretty much
all boilerplate, we can remove it just by a single macro.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
10 years agopinctrl: sunxi: Move the Allwinner pinctrl driver to its own directory
Maxime Ripard [Fri, 18 Apr 2014 16:53:02 +0000 (18:53 +0200)]
pinctrl: sunxi: Move the Allwinner pinctrl driver to its own directory

This will allow to create numerous files without crippling the main pinctrl
directory.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
10 years agopinctrl: sunxi: Drop unused structure members
Maxime Ripard [Sat, 26 Apr 2014 18:11:47 +0000 (20:11 +0200)]
pinctrl: sunxi: Drop unused structure members

The ranges and nranges were never used. Remove them.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
10 years agopinctrl: sunxi: Add const qualifier to the pin descriptor
Maxime Ripard [Fri, 18 Apr 2014 17:34:07 +0000 (19:34 +0200)]
pinctrl: sunxi: Add const qualifier to the pin descriptor

The pins description structure were declared as const, but the of_device_id
data magic was losing it silently.

Make sure we have it on both sides.

And now that we're using const, we can also remove the useless cast in probe.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
10 years agopinctrl: qcom: Correct name for pin 0
Andy Gross [Fri, 25 Apr 2014 20:41:55 +0000 (15:41 -0500)]
pinctrl: qcom: Correct name for pin 0

Fix copy/paste error in pinctrl_pin_desc for pin 0.

Signed-off-by: Andy Gross <agross@codeaurora.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: mvebu: new driver for Orion platforms
Thomas Petazzoni [Tue, 22 Apr 2014 21:26:07 +0000 (23:26 +0200)]
pinctrl: mvebu: new driver for Orion platforms

This commit extends the pinctrl mvebu logic with a new driver to cover
Orion5x SoC. It supports the definitions for the 5181l, 5182 and 5281
variants of Orion5x, which are the three ones supported by the old
style MPP code in arch/arm/mach-orion5x/.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: rockchip: implement PIN_CONFIG_OUTPUT handling
Heiko Stübner [Wed, 23 Apr 2014 12:28:59 +0000 (14:28 +0200)]
pinctrl: rockchip: implement PIN_CONFIG_OUTPUT handling

In some cases it is nice to be able to simply control a gpio output
via the PIN_CONFIG_OUTPUT option without having a driver control it.
Thus add support for it to the rockchip pinctrl driver.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: rockchip: return a complete config in pinconf_get
Heiko Stübner [Wed, 23 Apr 2014 12:27:51 +0000 (14:27 +0200)]
pinctrl: rockchip: return a complete config in pinconf_get

Till now pinconf_get only set the argument value into the config parameter
effectively removing the actual config param value. As other pinctrl drivers
do, it might be nicer to keep the config param intact.
Therefore construct a real pinconfig value from param and arg in pinconf_get

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: sirf: switch driver to use gpiolib irqchip helpers
Linus Walleij [Tue, 15 Apr 2014 06:43:47 +0000 (14:43 +0800)]
pinctrl: sirf: switch driver to use gpiolib irqchip helpers

This switches the SiRF pinctrl driver over to using the gpiolib
irqchip helpers simplifying some of the code.

Signed-off-by: Barry Song <Baohua.Song@csr.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: sirf: wrap all gpio banks into one gpio_chip
Barry Song [Tue, 15 Apr 2014 06:43:46 +0000 (14:43 +0800)]
pinctrl: sirf: wrap all gpio banks into one gpio_chip

all gpio banks are in one chip, that makes software clean in mapping
irq and gpio.

Signed-off-by: Barry Song <Baohua.Song@csr.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: sirf: rename inlined accessor
Linus Walleij [Tue, 8 Apr 2014 08:59:29 +0000 (10:59 +0200)]
pinctrl: sirf: rename inlined accessor

The sirfsoc_irqchip_to_bank() is obviously misnamed, as it is
not converting an irqchip to a bank but converts a gpiochip
to a bank so rename it sirfsoc_gpiochip_to_bank().

Acked-by: Barry Song <Baohua.Song@csr.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agodt: Document Qualcomm IPQ8064 pinctrl binding
Andy Gross [Tue, 15 Apr 2014 03:10:36 +0000 (22:10 -0500)]
dt: Document Qualcomm IPQ8064 pinctrl binding

Define a new binding for the Qualcomm TLMMv2 based pin controller inside the
IPQ8064.

Signed-off-by: Andy Gross <agross@codeaurora.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@sonymobile.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: qcom: Add definitions for IPQ8064
Andy Gross [Tue, 15 Apr 2014 03:10:35 +0000 (22:10 -0500)]
pinctrl: qcom: Add definitions for IPQ8064

This adds pinctrl definitions for the GPIO pins of the TLMM v2 block in the
Qualcomm IPQ8064 platform.

Signed-off-by: Andy Gross <agross@codeaurora.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@sonymobile.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: sunxi: list all pinctrl compatible strings
Boris BREZILLON [Tue, 22 Apr 2014 13:38:06 +0000 (15:38 +0200)]
pinctrl: sunxi: list all pinctrl compatible strings

List all sunxi pinctrl compatible strings in order to be able to grep for
those values.

Signed-off-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: exynos: Add driver data for Exynos3250
Tomasz Figa [Mon, 14 Apr 2014 01:45:47 +0000 (10:45 +0900)]
pinctrl: exynos: Add driver data for Exynos3250

This patch adds driver data (bank list and EINT layout) for Exynos3250
to pinctrl-exynos driver. Exynos3250 includes 158 multi-functional input/output
ports. There are 23 general port groups.

Changes from v1:
- Add signed-off of sender
- Post only separated patch for pinctrl from following patchset(v1)
  : https://lkml.org/lkml/2014/4/10/286

Cc: Thomas Abraham <thomas.abraham@linaro.org>
Cc: Linus Walleij <linus.walleij@linaro.org>
Cc: Kukjin Kim <kgene.kim@samsung.com>
Signed-off-by: Tomasz Figa <t.figa@samsung.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl/at91: Fix mask creation in at91_gpio_dbg_show
Alexander Stein [Mon, 14 Apr 2014 18:53:08 +0000 (20:53 +0200)]
pinctrl/at91: Fix mask creation in at91_gpio_dbg_show

pin_to_mask expects a bank pin number. So do not add the chip base.

Without that patch cat /sys/kernel/debug/gpio looks like that:
GPIOs 0-31, platform/fffff200.gpio, fffff200.gpio:
[spi32766.0] GPIOfffff200.gpio5: [gpio] set
[ads7846_pendown] GPIOfffff200.gpio15: [gpio] set
[ohci_vbus] GPIOfffff200.gpio21: [gpio] set
[ohci_vbus] GPIOfffff200.gpio24: [gpio] set
[button1] GPIOfffff200.gpio28: [gpio] clear
[button2] GPIOfffff200.gpio29: [gpio] clear

GPIOs 32-63, platform/fffff400.gpio, fffff400.gpio:
[sda] GPIOfffff400.gpio4: [periph A]
[scl] GPIOfffff400.gpio5: [periph A]
[spi32766.3] GPIOfffff400.gpio11: [periph A]
[error] GPIOfffff400.gpio22: [periph A]
[run] GPIOfffff400.gpio23: [periph A]

GPIOs 64-95, platform/fffff600.gpio, fffff600.gpio:
[reset_pin] GPIOfffff600.gpio29: [periph A]

GPIOs 96-127, platform/fffff800.gpio, fffff800.gpio:
[led1] GPIOfffff800.gpio5: [periph A]
[led2] GPIOfffff800.gpio6: [periph A]
[led3] GPIOfffff800.gpio7: [periph A]
[led4] GPIOfffff800.gpio8: [periph A]

GPIOs 128-159, platform/fffffa00.gpio, fffffa00.gpio:
[button3] GPIOfffffa00.gpio10: [periph A]
[button4] GPIOfffffa00.gpio12: [periph A]

Note that every bank despite bank 0 only shows "periph A" which are
obviously used as GPIOs.

Signed-off-by: Alexander Stein <alexanders83@web.de>
Acked-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl/at91: convert driver to use gpiolib irqchip
Alexander Stein [Tue, 15 Apr 2014 20:09:41 +0000 (22:09 +0200)]
pinctrl/at91: convert driver to use gpiolib irqchip

This converts the AT91 pin control driver to register its
chained irq handler and irqchip using the helpers in the
gpiolib core.

Signed-off-by: Alexander Stein <alexanders83@web.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: tegra: add missing kerneldoc
Stephen Warren [Tue, 15 Apr 2014 17:02:03 +0000 (11:02 -0600)]
pinctrl: tegra: add missing kerneldoc

The kerneldoc for struct tegra_pingroup didn't describe all of the fields
in the struct. Add some extra kerneldoc to fix that.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: tegra: print better error messages
Stephen Warren [Mon, 14 Apr 2014 21:33:42 +0000 (15:33 -0600)]
pinctrl: tegra: print better error messages

When an attempt is made to configure an unsupported option on a pin,
print the DT property name of that option, so it's easier to debug
what the problem is.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Laxman Dewangan <ldewangan@nvidia.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: tegra: reduce size of data table fields
Stephen Warren [Mon, 14 Apr 2014 21:33:41 +0000 (15:33 -0600)]
pinctrl: tegra: reduce size of data table fields

The range of npins and function ID values is small enough to fit into a
u8. Use this type rather than unsigned to shrink the pinmux data tables.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Laxman Dewangan <ldewangan@nvidia.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: tegra: remove fsafe from data tables
Stephen Warren [Mon, 14 Apr 2014 21:33:40 +0000 (15:33 -0600)]
pinctrl: tegra: remove fsafe from data tables

The fsafe value in the pingroup data tables is only used to implement
tegra_pinctrl_disable(). The only reason this function is called is when
dynamically switching between pinmux states, i.e. when disabling the old
state before programming the new state. It's simpler to have the new
target state define the expected value of each pin (and all current DTs
do that). This also gives more flexibility, since it allows individual
boards explicit control over the "inactive" mux function for each pin,
rather than requiring it to be an SoC-specific value. Assuming this, we
can get rid of the fsafe value from the driver completely, thus saving
some more space in the driver tables.

While re-writing the content of tegra124_pingroups[], fix the indentation
to use a TAB instead of spaces.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
Acked-by: Laxman Dewangan <ldewangan@nvidia.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: tegra: remove redundant data table fields
Stephen Warren [Tue, 15 Apr 2014 17:00:50 +0000 (11:00 -0600)]
pinctrl: tegra: remove redundant data table fields

Any SoC which supports the einput, odrain, lock, ioreset, or rcv_sel
options has the relevant HW register fields in the same register as the
mux function selection. Similarly, the drvtype option is always in the
drive register, if it is supported at all. Hence, we don't need to have
struct *_reg fields in the pin group table to define which register and
bank to use for those options. Delete this to save space in the driver's
data tables.

However, many of those options are not supported on all SoCs, or not
supported on some pingroups. We need a way to detect when they are
supported. Previously, this was indicated by setting the struct *_reg
field to -1. With the struct *_reg fields removed, we use the struct
*_bit fields for this purpose instead. The struct *_bit fields need to
be expanded from 5 to 6 bits in order to store a value outside the valid
HW bit range of 0..31.

Even without removing the struct *_reg fields, we still need to add code
to validate the struct *_bit fields, since some struct *_bit fields were
already being set to -1, without an option-specific struct *_reg field to
"guard" them. In other words, before this change, the pinmux driver might
allow some unsupported options to be written to HW.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: pfc: r8a7790: add mux data for IIC(B) cores
Wolfram Sang [Tue, 25 Mar 2014 18:56:26 +0000 (19:56 +0100)]
pinctrl: pfc: r8a7790: add mux data for IIC(B) cores

Signed-off-by: Wolfram Sang <wsa@sang-engineering.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: pfc: r8a7790: add i2c0 muxing
Wolfram Sang [Tue, 25 Mar 2014 18:56:25 +0000 (19:56 +0100)]
pinctrl: pfc: r8a7790: add i2c0 muxing

Add the muxing for the last missing i2c rcar core. Fix the sorting for
SH_PFC_PIN_NAMED while we are here.

Signed-off-by: Wolfram Sang <wsa@sang-engineering.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: pinctrl-imx: Print the mux_mode field in hex format
Fabio Estevam [Sun, 13 Apr 2014 15:09:05 +0000 (12:09 -0300)]
pinctrl: pinctrl-imx: Print the mux_mode field in hex format

With debug enabled we get better readability dumps of the mux_mode register if
we use hexadecimal format instead:

imx6sl-pinctrl 20e0000.iomuxc: MX6SL_PAD_FEC_REF_CLK: 0x10 0x0001b0a8

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Acked-by: Shawn Guo <shawn.guo@freescale.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agosh-pfc: r8a7791: Add Audio pin support
Kuninori Morimoto [Mon, 14 Apr 2014 00:24:04 +0000 (17:24 -0700)]
sh-pfc: r8a7791: Add Audio pin support

Reviewed-by: Geert Uytterhoeven <geert@linux-m68k.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agosh-pfc: r8a7791: Add SSI pin support
Kuninori Morimoto [Mon, 14 Apr 2014 00:23:35 +0000 (17:23 -0700)]
sh-pfc: r8a7791: Add SSI pin support

Reviewed-by: Geert Uytterhoeven <geert@linux-m68k.org>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: sunxi: fix typo in module author address
Antoine Ténart [Wed, 16 Apr 2014 12:57:48 +0000 (14:57 +0200)]
pinctrl: sunxi: fix typo in module author address

According to the MODULE_AUTHOR() comments, the author name should be
"Name <email>" or just "Name". Add the missing '>'.

Signed-off-by: Antoine Ténart <antoine.tenart@free-electrons.com>
Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: allows not to define the get_group_pins operation
Antoine Ténart [Thu, 10 Apr 2014 13:07:50 +0000 (15:07 +0200)]
pinctrl: allows not to define the get_group_pins operation

When using a group only pinctrl driver, which does not have any
information on the pins it is useless to define a get_group_pins
always returning an empty list of pins.

When not using get_group_pin[1], a driver must implement it so
pins = NULL and num_pins = 0. This patch makes it the default
behaviour if not defined in the pinctrl driver when used in
pinmux enable and disable funtions and in pinctrl_groups_show.

It also adds a check in pinctrl_get_group_pins and return -EINVAL if
not defined. This function is called in the gpiolib when adding when
pingroup range. It cannot be used if no group is defined, so this seams
reasonable.

[1] get_group_pin(struct pinctrl_dev *pctldev,
  unsigned selector,
  const unsigned **pins,
  unsigned *num_pins);

Signed-off-by: Antoine Ténart <antoine.tenart@free-electrons.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agoARM: sunxi: update the default ARCH_NR_GPIO for sunxi arch
Boris BREZILLON [Thu, 10 Apr 2014 13:52:46 +0000 (15:52 +0200)]
ARM: sunxi: update the default ARCH_NR_GPIO for sunxi arch

The A31 SoC has PL and PM banks and thus increase the default ARCH_NR_GPIO.

Signed-off-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: sunxi: add reset control support
Boris BREZILLON [Thu, 10 Apr 2014 13:52:45 +0000 (15:52 +0200)]
pinctrl: sunxi: add reset control support

The A31 SoC define a reset line for the R_PIO block which needs to be
deasserted.

Try to retrieve a reset control and deassert if one was found.

Signed-off-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: sunxi: define A31 R_PIO pin functions
Boris BREZILLON [Thu, 10 Apr 2014 13:52:44 +0000 (15:52 +0200)]
pinctrl: sunxi: define A31 R_PIO pin functions

The A31 SoC provides both PL and PM pio bank through the R_PIO block.

These pins all support gpio function and can bbe assigned to system
peripherals (like TWI, P2WI, JTAG, ...)

Signed-off-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: sunxi: support multiple pin controller
Boris BREZILLON [Thu, 10 Apr 2014 13:52:43 +0000 (15:52 +0200)]
pinctrl: sunxi: support multiple pin controller

Add support for multiple pin controller instances.

First remove the static definition of the sunxi gpio chip struct and fill
the dynamically struct instead.
Then define a new pin_base field in the sunxi_pinctrl_desc which will be
used to specify the gpiochip base pin.

Signed-off-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: sunxi: add PL and PM pin definitions
Boris BREZILLON [Thu, 10 Apr 2014 13:52:42 +0000 (15:52 +0200)]
pinctrl: sunxi: add PL and PM pin definitions

Define PL and PM pin macros.

Signed-off-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: sunxi: disable clk when failing to probe pin controller
Boris BREZILLON [Thu, 10 Apr 2014 13:52:41 +0000 (15:52 +0200)]
pinctrl: sunxi: disable clk when failing to probe pin controller

Disable the clk when failing to probe the pin controller device.

Signed-off-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: sunxi: check clk_prepare_enable return value
Boris BREZILLON [Thu, 10 Apr 2014 13:52:40 +0000 (15:52 +0200)]
pinctrl: sunxi: check clk_prepare_enable return value

Check the clk_prepare_enable return value to avoid false positive probe.

Signed-off-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: st: Use ARRAY_SIZE for STiH415 data
Maxime COQUELIN [Tue, 8 Apr 2014 15:21:49 +0000 (17:21 +0200)]
pinctrl: st: Use ARRAY_SIZE for STiH415 data

This patch completes the one that used ARRAY_SIZE for STiH407 and STiH416
for setting ninput_delays and noutput_delays fields.

Signed-off-by: Maxime Coquelin <maxime.coquelin@st.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: st: Use const qualifier when required
Maxime COQUELIN [Tue, 8 Apr 2014 15:21:48 +0000 (17:21 +0200)]
pinctrl: st: Use const qualifier when required

This patch adds const qualifier where applicable.

Reported-by: Joe Perches <joe@perches.com>
Signed-off-by: Maxime Coquelin <maxime.coquelin@st.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: lantiq: Fix header file include guard
Axel Lin [Thu, 3 Apr 2014 13:47:53 +0000 (21:47 +0800)]
pinctrl: lantiq: Fix header file include guard

Define __PINCTRL_LANTIQ_H to prevent multiple inclusion.

Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: adi2: Statize adi_gpio_irq_domain_ops
Axel Lin [Thu, 3 Apr 2014 13:44:47 +0000 (21:44 +0800)]
pinctrl: adi2: Statize adi_gpio_irq_domain_ops

It's only referenced in this file, make it static.

Signed-off-by: Axel Lin <axel.lin@ingics.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: sh-pfc: r8a7791: Split the DU sync and cde/disp groups
Laurent Pinchart [Tue, 1 Apr 2014 10:59:09 +0000 (12:59 +0200)]
pinctrl: sh-pfc: r8a7791: Split the DU sync and cde/disp groups

The DU parallel interface ODDF signal is optional, move it out of the
HSYNC/VSYNC group into a group of its down. The CDE and DISP signals are
independent, split them to two different groups.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: msm8x74: make Kconfig dependency more strict
Uwe Kleine-König [Tue, 1 Apr 2014 20:25:59 +0000 (22:25 +0200)]
pinctrl: msm8x74: make Kconfig dependency more strict

This driver is only useful on MSM8x74, so let the driver depend on
ARCH_QCOM but allow compile coverage testing.
The main benefit is that the driver isn't available to be selected for
machines that don't have the matching hardware.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: msm: Add definitions for the APQ8064 platform
Bjorn Andersson [Mon, 31 Mar 2014 21:49:57 +0000 (14:49 -0700)]
pinctrl: msm: Add definitions for the APQ8064 platform

This adds pinctrl definitions for the GPIO pins of the TLMM v2 block in the
Qualcomm APQ8064 platform.

Signed-off-by: Bjorn Andersson <bjorn.andersson@sonymobile.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: msm: Add documentation for pinctrl-apq8064 binding
Bjorn Andersson [Mon, 31 Mar 2014 21:49:56 +0000 (14:49 -0700)]
pinctrl: msm: Add documentation for pinctrl-apq8064 binding

DT bindingdocumentation for qcom,apq8064-pinctrl driver.

Signed-off-by: Bjorn Andersson <bjorn.andersson@sonymobile.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: msm: Make number of functions variable
Bjorn Andersson [Mon, 31 Mar 2014 21:49:55 +0000 (14:49 -0700)]
pinctrl: msm: Make number of functions variable

The various pins may have different number of functions defined, so make this
number definable per pin instead of just increasing it to the largest one for
all of the platforms.

Signed-off-by: Bjorn Andersson <bjorn.andersson@sonymobile.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agopinctrl: st: use gpiolib irqchip helpers
Linus Walleij [Tue, 8 Apr 2014 12:45:47 +0000 (14:45 +0200)]
pinctrl: st: use gpiolib irqchip helpers

This lets the gpiolib core handle the irqchip set-up and
chained IRQ on the primary (behind the mux) IRQ chip in
the st pinctrl driver.

Default irq type is set to level low at irqchip add time.

The v1 was sent by Linus
(https://lkml.org/lkml/2014/4/4/287).

Two changes were necessary to make it to work properly
on STiH416:
  1 - dev reference was not passed to the gpio_chip
      struct, causing a panic.
  2 - gpiochip_irqchip_add passed IRQ_TYPE_NONE as
      default type, which caused lot of warnings at
      init time. I choose IRQ_TYPE_LEVEL_LOW as default.

Cc: Srinivas Kandagatla <srinivas.kandagatla@gmail.com>
Cc: Giuseppe Cavallaro <peppe.cavallaro@st.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Maxime COQUELIN <maxime.coquelin@st.com>
10 years agopinctrl: st: switch IRQ locking to resource callbacks
Linus Walleij [Fri, 4 Apr 2014 14:02:21 +0000 (16:02 +0200)]
pinctrl: st: switch IRQ locking to resource callbacks

In the mass-conversion to the new irqchip callbacks, this
in-transit IRQ support was missed. Fix it.

Cc: Srinivas Kandagatla <srinivas.kandagatla@st.com>
Cc: Maxime COQUELIN <maxime.coquelin@st.com>
Cc: Giuseppe Cavallaro <peppe.cavallaro@st.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
10 years agoLinux 3.15-rc2
Linus Torvalds [Sun, 20 Apr 2014 18:08:50 +0000 (11:08 -0700)]
Linux 3.15-rc2

10 years agoMerge branch 'fixes' of git://git.infradead.org/users/vkoul/slave-dma
Linus Torvalds [Sun, 20 Apr 2014 17:35:31 +0000 (10:35 -0700)]
Merge branch 'fixes' of git://git.infradead.org/users/vkoul/slave-dma

Pull slave-dmaengine fixes from Vinod Koul:
 "Back from long weekend here in India and now the time to send fixes
  for slave dmaengine.
   - Dan's fix of sirf xlate code
   - Jean's fix for timberland
   - edma fixes by Sekhar for SG handling and Yuan for changing init
     call"

* 'fixes' of git://git.infradead.org/users/vkoul/slave-dma:
  dma: fix eDMA driver as a subsys_initcall
  dmaengine: sirf: off by one in of_dma_sirfsoc_xlate()
  platform: Fix timberdale dependencies
  dma: edma: fix incorrect SG list handling

10 years agoMerge tag 'iommu-fixes-v3.15-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 20 Apr 2014 17:33:49 +0000 (10:33 -0700)]
Merge tag 'iommu-fixes-v3.15-rc1' of git://git./linux/kernel/git/joro/iommu

Pull iommu fixes from Joerg Roedel:
 "Fixes for regressions:

   - fix wrong IOMMU enumeration causing some SCSI device drivers
     initialization failures
   - ARM-SMMU fixes for a panic condition and a wrong return value"

* tag 'iommu-fixes-v3.15-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu:
  iommu/arm-smmu: fix panic in arm_smmu_alloc_init_pte
  iommu/arm-smmu: Return 0 on unmap failure
  iommu/vt-d: fix bug in matching PCI devices with DRHD/RMRR descriptors
  iommu/vt-d: Fix get_domain_for_dev() handling of upstream PCIe bridges
  iommu/vt-d: fix memory leakage caused by commit ea8ea46

10 years agoMerge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 20 Apr 2014 17:32:33 +0000 (10:32 -0700)]
Merge branch 'perf-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull perf tooling fixes from Ingo Molnar:
 "Three small tooling fixes"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf tools: Improve error reporting
  perf tools: Adjust symbols in VDSO
  perf kvm: Fix 'Min time' counting in report command

10 years agoMerge tag 'perf-urgent-for-mingo' of git://git.kernel.org/pub/scm/linux/kernel/git...
Ingo Molnar [Sun, 20 Apr 2014 07:53:55 +0000 (09:53 +0200)]
Merge tag 'perf-urgent-for-mingo' of git://git./linux/kernel/git/jolsa/perf into perf/urgent

Pull perf/urgent fixes from Jiri Olsa:

User visible changes:

  * Adjust symbols in VDSO to properly resolve its function names (Vladimir Nikulichev)

  * Improve error reporting for record session failure (Adrien BAK)

  * Fix 'Min time' counting in report command (Alexander Yarygin)

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
10 years agoperf tools: Improve error reporting
Adrien BAK [Fri, 18 Apr 2014 02:00:43 +0000 (11:00 +0900)]
perf tools: Improve error reporting

In the current version, when using perf record, if something goes
wrong in tools/perf/builtin-record.c:375
  session = perf_session__new(file, false, NULL);

The error message:
"Not enough memory for reading per file header"

is issued. This error message seems to be outdated and is not very
helpful. This patch proposes to replace this error message by
"Perf session creation failed"

I believe this issue has been brought to lkml:
https://lkml.org/lkml/2014/2/24/458
although this patch only tackles a (small) part of the issue.

Additionnaly, this patch improves error reporting in
tools/perf/util/data.c open_file_write.

Currently, if the call to open fails, the user is unaware of it.
This patch logs the error, before returning the error code to
the caller.

Reported-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Adrien BAK <adrien.bak@metascale.org>
Link: http://lkml.kernel.org/r/1397786443.3093.4.camel@beast
[ Reorganize the changelog into paragraphs ]
[ Added empty line after fd declaration in open_file_write ]
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
10 years agoperf tools: Adjust symbols in VDSO
Vladimir Nikulichev [Thu, 17 Apr 2014 15:27:01 +0000 (08:27 -0700)]
perf tools: Adjust symbols in VDSO

pert-report doesn't resolve function names in VDSO:

$ perf report --stdio -g flat,0.0,15,callee --sort pid
...
            8.76%
               0x7fff6b1fe861
               __gettimeofday
               ACE_OS::gettimeofday()
...

In this case symbol values should be adjusted the same way as for executables,
relocatable objects and prelinked libraries.

After fix:

$ perf report --stdio -g flat,0.0,15,callee --sort pid
...
            8.76%
               __vdso_gettimeofday
               __gettimeofday
               ACE_OS::gettimeofday()

Signed-off-by: Vladimir Nikulichev <nvs@tbricks.com>
Tested-by: Namhyung Kim <namhyung@kernel.org>
Reviewed-by: Adrian Hunter <adrian.hunter@intel.com>
Link: http://lkml.kernel.org/r/969812.163009436-sendEmail@nvs
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
10 years agoperf kvm: Fix 'Min time' counting in report command
Alexander Yarygin [Wed, 9 Apr 2014 14:21:59 +0000 (16:21 +0200)]
perf kvm: Fix 'Min time' counting in report command

Every event in the perf-kvm has a 'stats' structure, which contains
max/min/average/etc times of handling this event.
The problem is that the 'perf-kvm stat report' command always shows
that 'min time' is 0us for every event. Example:

 # perf kvm stat report

 Analyze events for all VCPUs:

    VM-EXIT    Samples  Samples%     Time%   Min Time   Max Time Avg time
  [..]
  0xB2 MSCH         12     0.07%     0.00%        0us        8us 7.31us ( +-   2.11% )
  0xB2 CHSC         12     0.07%     0.00%        0us       18us 9.39us ( +-   9.49% )
  0xB2 STPX          8     0.05%     0.00%        0us        2us 1.88us ( +-   7.18% )
  0xB2 STSI          7     0.04%     0.00%        0us       44us 16.49us ( +-  38.20% )
  [..]

This happens because the 'stats' structure is not initialized and
stats->min equals to 0. Lets initialize the structure for every
event after its allocation using init_stats() function. This initializes
stats->min to -1 and makes 'Min time' statistics counting work:

 # perf kvm stat report

 Analyze events for all VCPUs:

    VM-EXIT    Samples  Samples%     Time%   Min Time   Max Time Avg time
  [..]
  0xB2 MSCH         12     0.07%     0.00%        6us        8us 7.31us ( +-   2.11% )
  0xB2 CHSC         12     0.07%     0.00%        7us       18us 9.39us ( +-   9.49% )
  0xB2 STPX          8     0.05%     0.00%        1us        2us 1.88us ( +-   7.18% )
  0xB2 STSI          7     0.04%     0.00%        1us       44us 16.49us ( +-  38.20% )
  [..]

Signed-off-by: Alexander Yarygin <yarygin@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Link: http://lkml.kernel.org/r/1397053319-2130-3-git-send-email-borntraeger@de.ibm.com
[ Fixing the perf examples changelog output ]
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
10 years agocoredump: fix va_list corruption
Eric Dumazet [Sat, 19 Apr 2014 17:15:07 +0000 (10:15 -0700)]
coredump: fix va_list corruption

A va_list needs to be copied in case it needs to be used twice.

Thanks to Hugh for debugging this issue, leading to various panics.

Tested:

  lpq84:~# echo "|/foobar12345 %h %h %h %h %h %h %h %h %h %h %h %h %h %h %h %h %h %h %h %h" >/proc/sys/kernel/core_pattern

'produce_core' is simply : main() { *(int *)0 = 1;}

  lpq84:~# ./produce_core
  Segmentation fault (core dumped)
  lpq84:~# dmesg | tail -1
  [  614.352947] Core dump to |/foobar12345 lpq84 lpq84 lpq84 lpq84 lpq84 lpq84 lpq84 lpq84 lpq84 lpq84 lpq84 lpq84 lpq84 lpq84 lpq84 lpq84 lpq84 lpq84 lpq84 (null) pipe failed

Notice the last argument was replaced by a NULL (we were lucky enough to
not crash, but do not try this on your production machine !)

After fix :

  lpq83:~# echo "|/foobar12345 %h %h %h %h %h %h %h %h %h %h %h %h %h %h %h %h %h %h %h %h" >/proc/sys/kernel/core_pattern
  lpq83:~# ./produce_core
  Segmentation fault
  lpq83:~# dmesg | tail -1
  [  740.800441] Core dump to |/foobar12345 lpq83 lpq83 lpq83 lpq83 lpq83 lpq83 lpq83 lpq83 lpq83 lpq83 lpq83 lpq83 lpq83 lpq83 lpq83 lpq83 lpq83 lpq83 lpq83 lpq83 pipe failed

Fixes: 5fe9d8ca21cc ("coredump: cn_vprintf() has no reason to call vsnprintf() twice")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Diagnosed-by: Hugh Dickins <hughd@google.com>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Cc: Neil Horman <nhorman@tuxdriver.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: stable@vger.kernel.org # 3.11+
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agoMerge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 19 Apr 2014 17:41:43 +0000 (10:41 -0700)]
Merge branch 'x86-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull x86 fix from Ingo Molnar:
 "This fixes the preemption-count imbalance crash reported by Owen
  Kibel"

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/mce: Fix CMCI preemption bugs

10 years agoMerge branch 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 19 Apr 2014 17:40:51 +0000 (10:40 -0700)]
Merge branch 'sched-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull scheduler fixes from Ingo Molnar:
 "Two fixes:

   - a SCHED_DEADLINE task selection fix
   - a sched/numa related lockdep splat fix"

* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched: Check for stop task appearance when balancing happens
  sched/numa: Fix task_numa_free() lockdep splat

10 years agoMerge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 19 Apr 2014 17:40:11 +0000 (10:40 -0700)]
Merge branch 'perf-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull perf fixes from Ingo Molnar:
 "Two kernel side fixes:

   - an Intel uncore PMU driver potential crash fix
   - a kprobes/perf-call-graph interaction fix"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/x86/intel: Use rdmsrl_safe() when initializing RAPL PMU
  kprobes/x86: Fix page-fault handling logic

10 years agoMerge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux
Linus Torvalds [Sat, 19 Apr 2014 17:35:30 +0000 (10:35 -0700)]
Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux

Pull drm fixes from Dave Airlie:
 "Unfortunately this contains no easter eggs, its a bit larger than I'd
  like, but I included a patch that just moves code from one file to
  another and I'd like to avoid merge conflicts with that later, so it
  makes it seem worse than it is,

  Otherwise:
   - radeon: fixes to use new microcode to stabilise some cards, use
     some common displayport code, some runtime pm fixes, pll regression
     fixes
   - i915: fix for some context oopses, a warn in a used path, backlight
     fixes
   - nouveau: regression fix
   - omap: a bunch of fixes"

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux: (51 commits)
  drm: bochs: drop unused struct fields
  drm: bochs: add power management support
  drm: cirrus: add power management support
  drm: Split out drm_probe_helper.c from drm_crtc_helper.c
  drm/plane-helper: Don't fake-implement primary plane disabling
  drm/ast: fix value check in cbr_scan2
  drm/nouveau/bios: fix a bit shift error introduced by 457e77b
  drm/radeon/ci: make sure mc ucode is loaded before checking the size
  drm/radeon/si: make sure mc ucode is loaded before checking the size
  drm/radeon: improve PLL params if we don't match exactly v2
  drm/radeon: memory leak on bo reservation failure. v2
  drm/radeon: fix VCE fence command
  drm/radeon: re-enable mclk dpm on R7 260X asics
  drm/radeon: add support for newer mc ucode on CI (v2)
  drm/radeon: add support for newer mc ucode on SI (v2)
  drm/radeon: apply more strict limits for PLL params v2
  drm/radeon: update CI DPM powertune settings
  drm/radeon: fix runpm handling on APUs (v4)
  drm/radeon: disable mclk dpm on R7 260X
  drm/tegra: Remove gratuitous pad field
  ...

10 years agoMerge branch 'drm-next-3.15-wip' of git://people.freedesktop.org/~deathsimple/linux...
Dave Airlie [Sat, 19 Apr 2014 01:16:02 +0000 (11:16 +1000)]
Merge branch 'drm-next-3.15-wip' of git://people.freedesktop.org/~deathsimple/linux into drm-next

Some i2c fixes over DisplayPort.

* 'drm-next-3.15-wip' of git://people.freedesktop.org/~deathsimple/linux:
  drm/radeon: Improve vramlimit module param documentation
  drm/radeon: fix audio pin counts for DCE6+ (v2)
  drm/radeon/dp: switch to the common i2c over aux code
  drm/dp/i2c: Update comments about common i2c over dp assumptions (v3)
  drm/dp/i2c: send bare addresses to properly reset i2c connections (v4)
  drm/radeon/dp: handle zero sized i2c over aux transactions (v2)
  drm/i915: support address only i2c-over-aux transactions
  drm/tegra: dp: Support address-only I2C-over-AUX transactions

10 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Sat, 19 Apr 2014 00:53:46 +0000 (17:53 -0700)]
Merge git://git./linux/kernel/git/davem/net

Pull more networking fixes from David Miller:

 1) Fix mlx4_en_netpoll implementation, it needs to schedule a NAPI
    context, not synchronize it.  From Chris Mason.

 2) Ipv4 flow input interface should never be zero, it should be
    LOOPBACK_IFINDEX instead.  From Cong Wang and Julian Anastasov.

 3) Properly configure MAC to PHY connection in mvneta devices, from
    Thomas Petazzoni.

 4) sys_recv should use SYSCALL_DEFINE.  From Jan Glauber.

 5) Tunnel driver ioctls do not use the correct namespace, fix from
    Nicolas Dichtel.

 6) Fix memory leak on seccomp filter attach, from Kees Cook.

 7) Fix lockdep warning for nested vlans, from Ding Tianhong.

 8) Crashes can happen in SCTP due to how the auth_enable value is
    managed, fix from Vlad Yasevich.

 9) Wireless fixes from John W Linville and co.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (45 commits)
  net: sctp: cache auth_enable per endpoint
  tg3: update rx_jumbo_pending ring param only when jumbo frames are enabled
  vlan: Fix lockdep warning when vlan dev handle notification
  seccomp: fix memory leak on filter attach
  isdn: icn: buffer overflow in icn_command()
  ip6_tunnel: use the right netns in ioctl handler
  sit: use the right netns in ioctl handler
  ip_tunnel: use the right netns in ioctl handler
  net: use SYSCALL_DEFINEx for sys_recv
  net: mdio-gpio: Add support for separate MDI and MDO gpio pins
  net: mdio-gpio: Add support for active low gpio pins
  net: mdio-gpio: Use devm_ functions where possible
  ipv4, route: pass 0 instead of LOOPBACK_IFINDEX to fib_validate_source()
  ipv4, fib: pass LOOPBACK_IFINDEX instead of 0 to flowi4_iif
  mlx4_en: don't use napi_synchronize inside mlx4_en_netpoll
  net: mvneta: properly configure the MAC <-> PHY connection in all situations
  net: phy: add minimal support for QSGMII PHY
  sfc:On MCDI timeout, issue an FLR (and mark MCDI to fail-fast)
  mwifiex: fix hung task on command timeout
  mwifiex: process event before command response
  ...

10 years agoMerge branch 'for-next' of git://git.samba.org/sfrench/cifs-2.6
Linus Torvalds [Sat, 19 Apr 2014 00:52:39 +0000 (17:52 -0700)]
Merge branch 'for-next' of git://git.samba.org/sfrench/cifs-2.6

Pull cifs fixes from Steve French:
 "A set of 5 small cifs fixes"

* 'for-next' of git://git.samba.org/sfrench/cifs-2.6:
  cif: fix dead code
  cifs: fix error handling cifs_user_readv
  fs: cifs: remove unused variable.
  Return correct error on query of xattr on file with empty xattrs
  cifs: Wait for writebacks to complete before attempting write.

10 years agoMerge tag 'char-misc-3.15-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregk...
Linus Torvalds [Sat, 19 Apr 2014 00:02:35 +0000 (17:02 -0700)]
Merge tag 'char-misc-3.15-rc2' of git://git./linux/kernel/git/gregkh/char-misc

Pull char/misc driver fixes from Greg KH:
 "Here are a few driver fixes for char/misc drivers that resolve
  reported issues.

  All have been in linux-next successfully for a few days"

* tag 'char-misc-3.15-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc:
  Drivers: hv: vmbus: Negotiate version 3.0 when running on ws2012r2 hosts
  Tools: hv: Handle the case when the target file exists correctly
  vme_tsi148: Utilize to_pci_dev() macro
  vme_tsi148: Fix PCI address mapping assumption
  vme_tsi148: Fix typo in tsi148_slave_get()
  w1: avoid recursive device_add
  w1: fix netlink refcnt leak on error path
  misc: Grammar s/addition/additional/
  drivers: mcb: fix memory leak in chameleon_parse_cells() error path
  mei: ignore client writing state during cb completion
  mei: me: do not load the driver if the FW doesn't support MEI interface
  GenWQE: Increase driver version number
  GenWQE: Fix multithreading problems
  GenWQE: Ensure rc is not returning an uninitialized value
  GenWQE: Add wmb before DDCB is started
  GenWQE: Enable access to VPD flash area

10 years agoMerge tag 'driver-core-3.15-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 18 Apr 2014 23:59:52 +0000 (16:59 -0700)]
Merge tag 'driver-core-3.15-rc2' of git://git./linux/kernel/git/gregkh/driver-core

Pull driver core fixes from Greg KH:
 "Here are some driver core fixes for 3.15-rc2.  Also in here are some
  documentation updates, as well as an API removal that had to wait for
  after -rc1 due to the cleanups coming into you from multiple developer
  trees (this one and the PPC tree.)

  All have been in linux next successfully"

* tag 'driver-core-3.15-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core:
  drivers/base/dd.c incorrect pr_debug() parameters
  Documentation: Update stable address in Chinese and Japanese translations
  topology: Fix compilation warning when not in SMP
  Chinese: add translation of io_ordering.txt
  stable_kernel_rules: spelling/word usage
  sysfs, driver-core: remove unused {sysfs|device}_schedule_callback_owner()
  kernfs: protect lazy kernfs_iattrs allocation with mutex
  fs: Don't return 0 from get_anon_bdev

10 years agoMerge tag 'staging-3.15-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Fri, 18 Apr 2014 23:58:47 +0000 (16:58 -0700)]
Merge tag 'staging-3.15-rc2' of git://git./linux/kernel/git/gregkh/staging

Pull staging driver fixes from Greg KH:
 "Here are a few staging driver fixes for issues that have been reported
  for 3.15-rc2.

  Also dominating the diffstat for the pull request is the removal of
  the rtl8187se driver.  It's no longer needed in staging as a "real"
  driver for this hardware is now merged in the tree in the "correct"
  location in drivers/net/

  All of these patches have been tested in linux-next"

* tag 'staging-3.15-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
  staging: r8188eu: Fix case where ethtype was never obtained and always be checked against 0
  staging: r8712u: Fix case where ethtype was never obtained and always be checked against 0
  staging: r8188eu: Calling rtw_get_stainfo() with a NULL sta_addr will return NULL
  staging: comedi: fix circular locking dependency in comedi_mmap()
  staging: r8723au: Add missing initialization of change_inx in sort algorithm
  Staging: unisys: use after free in list_for_each()
  staging: unisys: use after free in error messages
  staging: speakup: fix misuse of kstrtol() in handle_goto()
  staging: goldfish: Call free_irq in error path
  staging: delete rtl8187se wireless driver
  staging: rtl8723au: Fix buffer overflow in rtw_get_wfd_ie()
  staging: gs_fpgaboot: remove __TIMESTAMP__ macro
  staging: vme: fix memory leak in vme_user_probe()
  staging: fpgaboot: clean up Makefile
  staging/usbip: fix store_attach() sscanf return value check
  staging/usbip: userspace - fix usbipd SIGSEGV from refresh_exported_devices()
  staging: rtl8188eu: remove spaces, correct counts to unbreak P2P ioctls
  staging/rtl8821ae: Fix OOM handling in _rtl_init_deferred_work()

10 years agoMerge tag 'tty-3.15-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty
Linus Torvalds [Fri, 18 Apr 2014 23:57:53 +0000 (16:57 -0700)]
Merge tag 'tty-3.15-rc2' of git://git./linux/kernel/git/gregkh/tty

Pull tty/serial driver fixes from Greg KH:
 "Here are a number of small tty/serial driver fixes for 3.15-rc2.  Also
  in here are some Documentation file removals for drivers that we
  removed a long time ago, no need to keep it around any longer.

  All of these have been in linux-next for a bit"

* tag 'tty-3.15-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
  Revert "serial: 8250, disable "too much work" messages"
  serial: amba-pl011: fix regression, causing an Oops on rmmod
  tty: Fix help text of SYNCLINK_CS
  tty: fix memleak in alloc_pid
  ttyprintk: Allow built as a module
  ttyprintk: Fix wrong tty_unregister_driver() call in the error path
  serial: 8250, disable "too much work" messages
  Documentation/serial: Delete obsolete driver documentation
  serial: omap: Fix missing pm_runtime_resume handling by simplifying code
  serial_core: Fix pm imbalance on unbind
  serial: pl011: change Rx burst size to half of trigger level
  serial: timberdale: Depend on X86_32
  serial: st-asc: Fix SysRq char handling
  Revert "serial: clps711x: Give a chance to perform useful tasks during wait loop"
  serial_core: Fix conditional start_tx on ring buffer not empty
  serial: efm32: use $vendor,$device scheme for compatible string
  serial: omap: free the wakeup settings in remove

10 years agoMerge tag 'usb-3.15-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Linus Torvalds [Fri, 18 Apr 2014 23:57:00 +0000 (16:57 -0700)]
Merge tag 'usb-3.15-rc2' of git://git./linux/kernel/git/gregkh/usb

Pull USB fixes from Greg KH:
 "Here are a number of tiny USB fixes and new device ids for 3.15-rc2.
  Nothing major, just issues some people have reported.

  All of these have been in linux-next"

* tag 'usb-3.15-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb:
  uas: fix deadlocky memory allocations
  uas: fix error handling during scsi_scan()
  uas: fix GFP_NOIO under spinlock
  uwb: adds missing error handling
  USB: cdc-acm: Remove Motorola/Telit H24 serial interfaces from ACM driver
  USB: ohci-jz4740: FEAT_POWER is a port feature, not a hub feature
  USB: ohci-jz4740: Fix uninitialized variable warning
  USB: EHCI: tegra: set txfill_tuning
  usb: ehci-platform: Return immediately from suspend if ehci_suspend fails
  usb: ehci-exynos: Return immediately from suspend if ehci_suspend fails
  USB: fix crash during hotplug of PCI USB controller card
  USB: cdc-acm: fix double usb_autopm_put_interface() in acm_port_activate()
  usb: usb-common: fix typo for usb_state_string
  USB: usb_wwan: fix handling of missing bulk endpoints
  USB: pl2303: add ids for Hewlett-Packard HP POS pole displays
  USB: cp210x: Add 8281 (Nanotec Plug & Drive)
  usb: option driver, add support for Telit UE910v2
  Revert "USB: serial: add usbid for dell wwan card to sierra.c"
  USB: serial: ftdi_sio: add id for Brainboxes serial cards

10 years agoMerge branch 'akpm' (incoming from Andrew)
Linus Torvalds [Fri, 18 Apr 2014 23:40:31 +0000 (16:40 -0700)]
Merge branch 'akpm' (incoming from Andrew)

Merge misc fixes from Andrew Morton:
 "13 fixes"

* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
  thp: close race between split and zap huge pages
  mm: fix new kernel-doc warning in filemap.c
  mm: fix CONFIG_DEBUG_VM_RB description
  mm: use paravirt friendly ops for NUMA hinting ptes
  mips: export flush_icache_range
  mm/hugetlb.c: add cond_resched_lock() in return_unused_surplus_pages()
  wait: explain the shadowing and type inconsistencies
  Shiraz has moved
  Documentation/vm/numa_memory_policy.txt: fix wrong document in numa_memory_policy.txt
  powerpc/mm: fix ".__node_distance" undefined
  kernel/watchdog.c:touch_softlockup_watchdog(): use raw_cpu_write()
  init/Kconfig: move the trusted keyring config option to general setup
  vmscan: reclaim_clean_pages_from_list() must use mod_zone_page_state()

10 years agothp: close race between split and zap huge pages
Kirill A. Shutemov [Fri, 18 Apr 2014 22:07:25 +0000 (15:07 -0700)]
thp: close race between split and zap huge pages

Sasha Levin has reported two THP BUGs[1][2].  I believe both of them
have the same root cause.  Let's look to them one by one.

The first bug[1] is "kernel BUG at mm/huge_memory.c:1829!".  It's
BUG_ON(mapcount != page_mapcount(page)) in __split_huge_page().  From my
testing I see that page_mapcount() is higher than mapcount here.

I think it happens due to race between zap_huge_pmd() and
page_check_address_pmd().  page_check_address_pmd() misses PMD which is
under zap:

CPU0 CPU1
zap_huge_pmd()
  pmdp_get_and_clear()
__split_huge_page()
  anon_vma_interval_tree_foreach()
    __split_huge_page_splitting()
      page_check_address_pmd()
        mm_find_pmd()
  /*
   * We check if PMD present without taking ptl: no
   * serialization against zap_huge_pmd(). We miss this PMD,
   * it's not accounted to 'mapcount' in __split_huge_page().
   */
  pmd_present(pmd) == 0

  BUG_ON(mapcount != page_mapcount(page)) // CRASH!!!

  page_remove_rmap(page)
    atomic_add_negative(-1, &page->_mapcount)

The second bug[2] is "kernel BUG at mm/huge_memory.c:1371!".
It's VM_BUG_ON_PAGE(!PageHead(page), page) in zap_huge_pmd().

This happens in similar way:

CPU0 CPU1
zap_huge_pmd()
  pmdp_get_and_clear()
  page_remove_rmap(page)
    atomic_add_negative(-1, &page->_mapcount)
__split_huge_page()
  anon_vma_interval_tree_foreach()
    __split_huge_page_splitting()
      page_check_address_pmd()
        mm_find_pmd()
  pmd_present(pmd) == 0 /* The same comment as above */
  /*
   * No crash this time since we already decremented page->_mapcount in
   * zap_huge_pmd().
   */
  BUG_ON(mapcount != page_mapcount(page))

  /*
   * We split the compound page here into small pages without
   * serialization against zap_huge_pmd()
   */
  __split_huge_page_refcount()
VM_BUG_ON_PAGE(!PageHead(page), page); // CRASH!!!

So my understanding the problem is pmd_present() check in mm_find_pmd()
without taking page table lock.

The bug was introduced by me commit with commit 117b0791ac42. Sorry for
that. :(

Let's open code mm_find_pmd() in page_check_address_pmd() and do the
check under page table lock.

Note that __page_check_address() does the same for PTE entires
if sync != 0.

I've stress tested split and zap code paths for 36+ hours by now and
don't see crashes with the patch applied. Before it took <20 min to
trigger the first bug and few hours for second one (if we ignore
first).

[1] https://lkml.kernel.org/g/<53440991.9090001@oracle.com>
[2] https://lkml.kernel.org/g/<5310C56C.60709@oracle.com>

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reported-by: Sasha Levin <sasha.levin@oracle.com>
Tested-by: Sasha Levin <sasha.levin@oracle.com>
Cc: Bob Liu <lliubbo@gmail.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Michel Lespinasse <walken@google.com>
Cc: Dave Jones <davej@redhat.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: <stable@vger.kernel.org> [3.13+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm: fix new kernel-doc warning in filemap.c
Randy Dunlap [Fri, 18 Apr 2014 22:07:23 +0000 (15:07 -0700)]
mm: fix new kernel-doc warning in filemap.c

Fix new kernel-doc warning in mm/filemap.c:

  Warning(mm/filemap.c:2600): Excess function parameter 'ppos' description in '__generic_file_aio_write'

Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm: fix CONFIG_DEBUG_VM_RB description
Davidlohr Bueso [Fri, 18 Apr 2014 22:07:22 +0000 (15:07 -0700)]
mm: fix CONFIG_DEBUG_VM_RB description

This appears to be a copy/paste error.  Update the description to
reflect extra rbtree debug and checks for the config option instead of
duplicating CONFIG_DEBUG_VM.

Signed-off-by: Davidlohr Bueso <davidlohr@hp.com>
Cc: Aswin Chandramouleeswaran <aswin@hp.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomm: use paravirt friendly ops for NUMA hinting ptes
Mel Gorman [Fri, 18 Apr 2014 22:07:21 +0000 (15:07 -0700)]
mm: use paravirt friendly ops for NUMA hinting ptes

David Vrabel identified a regression when using automatic NUMA balancing
under Xen whereby page table entries were getting corrupted due to the
use of native PTE operations.  Quoting him

Xen PV guest page tables require that their entries use machine
addresses if the preset bit (_PAGE_PRESENT) is set, and (for
successful migration) non-present PTEs must use pseudo-physical
addresses.  This is because on migration MFNs in present PTEs are
translated to PFNs (canonicalised) so they may be translated back
to the new MFN in the destination domain (uncanonicalised).

pte_mknonnuma(), pmd_mknonnuma(), pte_mknuma() and pmd_mknuma()
set and clear the _PAGE_PRESENT bit using pte_set_flags(),
pte_clear_flags(), etc.

In a Xen PV guest, these functions must translate MFNs to PFNs
when clearing _PAGE_PRESENT and translate PFNs to MFNs when setting
_PAGE_PRESENT.

His suggested fix converted p[te|md]_[set|clear]_flags to using
paravirt-friendly ops but this is overkill.  He suggested an alternative
of using p[te|md]_modify in the NUMA page table operations but this is
does more work than necessary and would require looking up a VMA for
protections.

This patch modifies the NUMA page table operations to use paravirt
friendly operations to set/clear the flags of interest.  Unfortunately
this will take a performance hit when updating the PTEs on
CONFIG_PARAVIRT but I do not see a way around it that does not break
Xen.

Signed-off-by: Mel Gorman <mgorman@suse.de>
Acked-by: David Vrabel <david.vrabel@citrix.com>
Tested-by: David Vrabel <david.vrabel@citrix.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Anvin <hpa@zytor.com>
Cc: Fengguang Wu <fengguang.wu@intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Steven Noonan <steven@uplinklabs.net>
Cc: Rik van Riel <riel@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 years agomips: export flush_icache_range
Kees Cook [Fri, 18 Apr 2014 22:07:19 +0000 (15:07 -0700)]
mips: export flush_icache_range

The lkdtm module performs tests against executable memory ranges, so it
needs to flush the icache for proper behaviors.  Other architectures
already export this, so do the same for MIPS.

[akpm@linux-foundation.org: relocate export sites]
Signed-off-by: Kees Cook <keescook@chromium.org>
Cc: Paul Gortmaker <paul.gortmaker@windriver.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Sanjay Lal <sanjayl@kymasys.com>
Cc: John Crispin <blogic@openwrt.org>
Cc: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>