platform/kernel/linux-rpi.git
5 years agospi-mem: fix kernel-doc for spi_mem_dirmap_{read|write}()
Sergei Shtylyov [Sat, 6 Apr 2019 18:33:29 +0000 (21:33 +0300)]
spi-mem: fix kernel-doc for spi_mem_dirmap_{read|write}()

The function names in the kernel-doc comments were mistyped, with a word
"dirmap" being repeated twice, so fix them.

Fixes: aa167f3fed0c ("spi: spi-mem: Add a new API to support direct mapping")
Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: bcm2835aux: Fix build error without CONFIG_DEBUG_FS
YueHaibing [Sat, 6 Apr 2019 15:14:57 +0000 (23:14 +0800)]
spi: bcm2835aux: Fix build error without CONFIG_DEBUG_FS

When building CONFIG_DEBUG_FS is not set
gcc warn this:

drivers/spi/spi-bcm2835aux.c: In function bcm2835aux_spi_probe:
drivers/spi/spi-bcm2835aux.c:591:2: error: too many arguments to function bcm2835aux_debugfs_create
  bcm2835aux_debugfs_create(bs, dev_name(&pdev->dev));
  ^~~~~~~~~~~~~~~~~~~~~~~~~
drivers/spi/spi-bcm2835aux.c:145:13: note: declared here
 static void bcm2835aux_debugfs_create(struct bcm2835aux_spi *bs)

Reported-by: Hulk Robot <hulkci@huawei.com>
Fixes: 8048d151eb4d ("spi: bcm2835aux: add driver stats to debugfs")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Reviewed-by: Mukesh Ojha <mojha@codeaurora.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: kill useless initializer in spi_register_controller()
Sergei Shtylyov [Fri, 5 Apr 2019 15:48:50 +0000 (18:48 +0300)]
spi: kill useless initializer in spi_register_controller()

The 'status' local variable is initialized but this value is never used,
thus kill that initializer.

Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: spi-mem: Add support for Zynq QSPI controller
Naga Sureshkumar Relli [Mon, 1 Apr 2019 07:59:13 +0000 (13:29 +0530)]
spi: spi-mem: Add support for Zynq QSPI controller

Add support for QSPI controller driver used by Xilinx Zynq SOC.

Signed-off-by: Naga Sureshkumar Relli <naga.sureshkumar.relli@xilinx.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agodt-bindings: spi: Add device tree binding documentation for Zynq QSPI controller
Naga Sureshkumar Relli [Mon, 1 Apr 2019 07:58:32 +0000 (13:28 +0530)]
dt-bindings: spi: Add device tree binding documentation for Zynq QSPI controller

This patch adds the dts binding document for Zynq SOC QSPI controller.

Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Naga Sureshkumar Relli <naga.sureshkumar.relli@xilinx.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: gpio: Drop unused pdev field in struct spi_gpio
Andrey Smirnov [Wed, 3 Apr 2019 04:01:30 +0000 (21:01 -0700)]
spi: gpio: Drop unused pdev field in struct spi_gpio

There's no code using 'pdev' field in struct spi_gpio. Drop it.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Chris Healy <cphealy@gmail.com>
Cc: linux-spi@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: gpio: Drop mflags argument from spi_gpio_request()
Andrey Smirnov [Wed, 3 Apr 2019 04:01:29 +0000 (21:01 -0700)]
spi: gpio: Drop mflags argument from spi_gpio_request()

The logic of setting mflags in spi_gpio_request() is very simple and
there isn't much benefit in having it in that function. Move all of
that code outside into spi_gpio_probe() in order to simplify things.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Chris Healy <cphealy@gmail.com>
Cc: linux-spi@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: Don't call spi_get_gpio_descs() before device name is set
Andrey Smirnov [Wed, 3 Apr 2019 04:01:28 +0000 (21:01 -0700)]
spi: Don't call spi_get_gpio_descs() before device name is set

Move code calling spi_get_gpio_descs() to happen after ctlr->dev's
name is set in order to have proper GPIO consumer names.

Before:

cat /sys/kernel/debug/gpio
gpiochip0: GPIOs 0-31, parent: platform/40049000.gpio, vf610-gpio:
 gpio-6   (                    |regulator-usb0-vbus ) out lo

gpiochip1: GPIOs 32-63, parent: platform/4004a000.gpio, vf610-gpio:
 gpio-36  (                    |scl                 ) in  hi
 gpio-37  (                    |sda                 ) in  hi
 gpio-40  (                    |(null) CS1          ) out lo
 gpio-41  (                    |(null) CS0          ) out lo ACTIVE LOW
 gpio-42  (                    |miso                ) in  hi
 gpio-43  (                    |mosi                ) in  lo
 gpio-44  (                    |sck                 ) out lo

After:

cat /sys/kernel/debug/gpio
gpiochip0: GPIOs 0-31, parent: platform/40049000.gpio, vf610-gpio:
 gpio-6   (                    |regulator-usb0-vbus ) out lo

gpiochip1: GPIOs 32-63, parent: platform/4004a000.gpio, vf610-gpio:
 gpio-36  (                    |scl                 ) in  hi
 gpio-37  (                    |sda                 ) in  hi
 gpio-40  (                    |spi0 CS1            ) out lo
 gpio-41  (                    |spi0 CS0            ) out lo ACTIVE LOW
 gpio-42  (                    |miso                ) in  hi
 gpio-43  (                    |mosi                ) in  lo
 gpio-44  (                    |sck                 ) out lo

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Chris Healy <cphealy@gmail.com>
Cc: linux-spi@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: gpio: Don't request CS GPIO in DT use-case
Andrey Smirnov [Wed, 3 Apr 2019 04:01:27 +0000 (21:01 -0700)]
spi: gpio: Don't request CS GPIO in DT use-case

DT use-case already relies on SPI core to control CS (requested by
of_spi_register_master() and controlled spi_set_cs()), so there's no
need to try to request those GPIO in spi-gpio code. Change the code
such that spi-gpio's CS related code is only used if device is probed
via pdata.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Chris Healy <cphealy@gmail.com>
Cc: linux-spi@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: tegra114: set bus number based on id
Sowjanya Komatineni [Fri, 5 Apr 2019 00:14:06 +0000 (17:14 -0700)]
spi: tegra114: set bus number based on id

This patch sets SPI device id from the device tree as the bus number.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: tegra114: use unpacked mode for below 4 bytes
Sowjanya Komatineni [Fri, 5 Apr 2019 00:14:01 +0000 (17:14 -0700)]
spi: tegra114: use unpacked mode for below 4 bytes

Packed mode expects minimum transfer length of 4 bytes.

This patch fixes this by using unpacked mode for transfers less
than 4 bytes.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: tegra114: set supported bits per word
Sowjanya Komatineni [Fri, 5 Apr 2019 00:14:05 +0000 (17:14 -0700)]
spi: tegra114: set supported bits per word

Tegra SPI supports 4 through 32 bits per word.

This patch sets bits_per_word_mask accordingly to support transfer
with these bits per word.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: tegra114: dump SPI registers during timeout
Sowjanya Komatineni [Fri, 5 Apr 2019 00:14:04 +0000 (17:14 -0700)]
spi: tegra114: dump SPI registers during timeout

This patch dumps SPI registers on transfer error or timeout for debug
purpose.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoMerge branch 'spi-5.1' into spi-5.2
Mark Brown [Fri, 5 Apr 2019 03:19:10 +0000 (10:19 +0700)]
Merge branch 'spi-5.1' into spi-5.2

5 years agospi: tegra114: avoid reset call in atomic context
Sowjanya Komatineni [Fri, 5 Apr 2019 00:14:03 +0000 (17:14 -0700)]
spi: tegra114: avoid reset call in atomic context

This patch moves SPI controller reset out of spin lock.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: tegra114: de-assert CS before SPI mode change
Sowjanya Komatineni [Fri, 5 Apr 2019 00:14:02 +0000 (17:14 -0700)]
spi: tegra114: de-assert CS before SPI mode change

With SW CS, during the transfer completion CS is de-asserted by writing
default command1 register value to SPI_COMMAND1 register. With this both
mode and CS state are set at the same time and if current transfer mode
is different to default SPI mode and if mode change happens prior to CS
de-assert, clock polarity can change while CS is active before transfer
finishes.

This causes Slave to see spurious clock edges resulting in data mismatch.

This patch fixes this by de-asserting CS before writing SPI_COMMAND1 to
its default value so through out the transfer it will be in same SPI mode.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: sh-msiof: Convert to use GPIO descriptors
Geert Uytterhoeven [Wed, 3 Apr 2019 15:08:52 +0000 (17:08 +0200)]
spi: sh-msiof: Convert to use GPIO descriptors

Convert GPIO chip selects in the Renesas MSIOF SPI driver from legacy
GPIO numbers to GPIO descriptors.

Notes:
  - The board file for the SH7724-based Ecovec24 development board now
    registers a GPIO descriptor lookup, instead of passing a GPIO number
    through controller_data,
  - sh_msiof_get_cs_gpios() must release all GPIOs, else
    spi_get_gpio_descs() cannot claim them during SPI controller
    registration.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: bcm2835aux: polling_limit_us can be static
kbuild test robot [Wed, 3 Apr 2019 18:16:51 +0000 (02:16 +0800)]
spi: bcm2835aux: polling_limit_us can be static

Fixes: 5fd917afc4bf ("spi: bcm2835aux: make the polling duration limits configurable")
Signed-off-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: Add missing error handling for CS GPIOs
Geert Uytterhoeven [Wed, 3 Apr 2019 14:46:56 +0000 (16:46 +0200)]
spi: Add missing error handling for CS GPIOs

While devm_gpiod_get_index_optional() returns NULL if the GPIO is not
present (i.e. -ENOENT), it may still return other error codes, like
-EPROBE_DEFER.  Currently these are not handled, leading to
unrecoverable failures later in case of probe deferral:

    gpiod_set_consumer_name: invalid GPIO (errorpointer)
    gpiod_direction_output: invalid GPIO (errorpointer)
    gpiod_set_value_cansleep: invalid GPIO (errorpointer)
    gpiod_set_value_cansleep: invalid GPIO (errorpointer)
    gpiod_set_value_cansleep: invalid GPIO (errorpointer)

Detect and propagate errors to fix this.

Fixes: f3186dd876697e69 ("spi: Optionally use GPIO descriptors for CS GPIOs")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: spi-gpio: Remove spi->controller_data comment
Geert Uytterhoeven [Wed, 3 Apr 2019 14:46:02 +0000 (16:46 +0200)]
spi: spi-gpio: Remove spi->controller_data comment

The conversion from GPIO numbers to GPIO descriptors removed the use of
spi->controller_data, but forgot to update a comment referring to it.

Fixes: 9b00bc7b901ff672 ("spi: spi-gpio: Rewrite to use GPIO descriptors")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: bcm2835aux: add driver stats to debugfs
Martin Sperl [Sat, 30 Mar 2019 10:13:53 +0000 (10:13 +0000)]
spi: bcm2835aux: add driver stats to debugfs

To estimate efficiency add statistics on transfer types
(polling and interrupt) used to debugfs.

Signed-off-by: Martin Sperl <kernel@martin.sperl.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: sh-msiof: Add reset of registers before starting transfer
Geert Uytterhoeven [Tue, 2 Apr 2019 14:40:23 +0000 (16:40 +0200)]
spi: sh-msiof: Add reset of registers before starting transfer

In accordance with hardware specification Ver 1.0, reset register
transmission / reception setting before transfer.

Signed-off-by: Hiromitsu Yamasaki <hiromitsu.yamasaki.ym@renesas.com>
[geert: Use readl_poll_timeout_atomic()]
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: sh-msiof: Use readl_poll_timeout_atomic() instead of open-coding
Geert Uytterhoeven [Tue, 2 Apr 2019 14:40:22 +0000 (16:40 +0200)]
spi: sh-msiof: Use readl_poll_timeout_atomic() instead of open-coding

Replace the open-coded loop in sh_msiof_modify_ctr_wait() by a call to
the readl_poll_timeout_atomic() helper macro.

Suggested-by: Wolfram Sang <wsa@the-dreams.de>
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: lpspi: add missing complete in abort func at dma mode
Clark Wang [Tue, 2 Apr 2019 12:45:53 +0000 (12:45 +0000)]
spi: lpspi: add missing complete in abort func at dma mode

Add the missing complete operations for dma_completion to fix the
problem of blocking at the wait_for_completion_interruptible()
function when use spi_slave_abort().

Signed-off-by: Clark Wang <xiaoning.wang@nxp.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: gpio: Drop unused pdata copy in struct spi_gpio
Andrey Smirnov [Wed, 3 Apr 2019 04:01:26 +0000 (21:01 -0700)]
spi: gpio: Drop unused pdata copy in struct spi_gpio

Drop unused pdata copy in struct spi_gpio. No functional change
intended.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Chris Healy <cphealy@gmail.com>
Cc: linux-spi@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: gpio: Simplify SPI_MASTER_NO_TX check in spi_gpio_probe()
Andrey Smirnov [Wed, 3 Apr 2019 04:01:25 +0000 (21:01 -0700)]
spi: gpio: Simplify SPI_MASTER_NO_TX check in spi_gpio_probe()

Swap branches of the if statement in order to simplify it's logical
condition being checked. No functional change intended.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Chris Healy <cphealy@gmail.com>
Cc: linux-spi@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: gpio: Add local struct spi_bitbang pointer in spi_gpio_probe()
Andrey Smirnov [Wed, 3 Apr 2019 04:01:24 +0000 (21:01 -0700)]
spi: gpio: Add local struct spi_bitbang pointer in spi_gpio_probe()

Use a local "struct spi_bitbang *bb" in spi_gpio_probe() for
brevity. No functional change intended.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Chris Healy <cphealy@gmail.com>
Cc: linux-spi@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: gpio: Add local struct device pointer in spi_gpio_probe()
Andrey Smirnov [Wed, 3 Apr 2019 04:01:23 +0000 (21:01 -0700)]
spi: gpio: Add local struct device pointer in spi_gpio_probe()

Use a local "struct device *dev" in spi_gpio_probe() for brevity. No
functional change intended.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Chris Healy <cphealy@gmail.com>
Cc: linux-spi@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: gpio: Drop unused spi_to_pdata()
Andrey Smirnov [Wed, 3 Apr 2019 04:01:22 +0000 (21:01 -0700)]
spi: gpio: Drop unused spi_to_pdata()

Spi_to_pdata() is not used anywhere in the code. Drop it.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Chris Healy <cphealy@gmail.com>
Cc: linux-spi@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: bcm2835aux: make the polling duration limits configurable
Martin Sperl [Sat, 30 Mar 2019 09:31:06 +0000 (09:31 +0000)]
spi: bcm2835aux: make the polling duration limits configurable

Under some circumstances the default 30 us polling limit is not optimal
and may lead to long delays because we are waiting on an interrupt.
with this patch we have the possibility to influence this policy.

So make this limit (in us) configurable via a module parameters
(but also modifyable via /sys/modules/...)

Signed-off-by: Martin Sperl <kernel@martin.sperl.org>
Acked-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: bcm2835aux: setup gpio-cs to output and correct level during setup
Martin Sperl [Sat, 30 Mar 2019 09:31:04 +0000 (09:31 +0000)]
spi: bcm2835aux: setup gpio-cs to output and correct level during setup

Setup gpio-cs to the correct levels during setup and also make the
gpio definitely an output GPIO.

This is transparently fixing some badly configured DTs in the process
where cs-gpio is set but the gpios are still configured with native cs.

It also makes 100% sure that the initial CS levels are as expected -
especially on systems with devices on a bus with mixed CS_HIGH/CS_LOW
settings.

Fixes: 1ea29b39f4c812ec ("spi: bcm2835aux: add bcm2835 auxiliary spi device...")
Signed-off-by: Martin Sperl <kernel@martin.sperl.org>
Acked-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: bcm2835aux: warn in dmesg that native cs is not really supported
Martin Sperl [Sat, 30 Mar 2019 09:31:03 +0000 (09:31 +0000)]
spi: bcm2835aux: warn in dmesg that native cs is not really supported

From personal bad experience (even as the author of the original driver)
it shows that native-cs is "somewhat" supported by the spi bus driver
when using a buggy device tree.

So make sure that the driver is warning in dmesg about this fact
that we are running in a not supported mode that may have surprizing
limitations.

Fixes: 1ea29b39f4c812ec ("spi: bcm2835aux: add bcm2835 auxiliary spi device...")
Signed-off-by: Martin Sperl <kernel@martin.sperl.org>
Acked-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: bcm2835aux: fix driver to not allow 65535 (=-1) cs-gpios
Martin Sperl [Sat, 30 Mar 2019 09:31:02 +0000 (09:31 +0000)]
spi: bcm2835aux: fix driver to not allow 65535 (=-1) cs-gpios

The original driver by default defines num_chipselects as -1.
This actually allicates an array of 65535 entries in
of_spi_register_master.

There is a side-effect for buggy device trees that (contrary to
dt-binding documentation) have no cs-gpio defined.

This mode was never supported by the driver due to limitations
of native cs and additional code complexity and is explicitly
not stated to be implemented.

To keep backwards compatibility with such buggy DTs we limit
the number of chip_selects to 1, as for all practical purposes
it is only ever realistic to use a single chip select in
native cs mode without negative side-effects.

Fixes: 1ea29b39f4c812ec ("spi: bcm2835aux: add bcm2835 auxiliary spi device...")
Signed-off-by: Martin Sperl <kernel@martin.sperl.org>
Acked-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: bcm2835aux: remove dead code
Martin Sperl [Sat, 30 Mar 2019 09:31:01 +0000 (09:31 +0000)]
spi: bcm2835aux: remove dead code

Remove dead code that never can get reached, as we limit count to
a max of 3.

Suggested-by: Hubert Denkmair <h.denkmair@intence.de>
Signed-off-by: Martin Sperl <kernel@martin.sperl.org>
Acked-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: bcm2835aux: fix corruptions for longer spi transfers
Martin Sperl [Sat, 30 Mar 2019 09:31:00 +0000 (09:31 +0000)]
spi: bcm2835aux: fix corruptions for longer spi transfers

On long running tests with a mcp2517fd can controller it showed that
on rare occations the data read shows corruptions for longer spi transfers.

Example of a 22 byte transfer:

expected (as captured on logic analyzer):
FF FF 78 00 00 00 08 06 00 00 91 20 77 56 84 85 86 87 88 89 8a 8b

read by the driver:
FF FF 78 00 00 00 08 06 00 00 91 20 77 56 84 88 89 8a 00 00 8b 9b

To fix this use BCM2835_AUX_SPI_STAT_RX_LVL to determine when we may
read data from the fifo reliably without any corruption.

Surprisingly the only values ever empirically read in
BCM2835_AUX_SPI_STAT_RX_LVL are 0x00, 0x10, 0x20 and 0x30.
So whenever the mask is not 0 we can read from the fifo in a safe manner.

The patch has now been tested intensively and we are no longer
able to reproduce the "RX" issue any longer.

Fixes: 1ea29b39f4c812ec ("spi: bcm2835aux: add bcm2835 auxiliary spi device...")
Reported-by: Hubert Denkmair <h.denkmair@intence.de>
Signed-off-by: Martin Sperl <kernel@martin.sperl.org>
Acked-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: bcm2835aux: remove dangerous uncontrolled read of fifo
Martin Sperl [Sat, 30 Mar 2019 09:30:59 +0000 (09:30 +0000)]
spi: bcm2835aux: remove dangerous uncontrolled read of fifo

This read of the fifo is a potential candidate for a race condition
as the spi transfer is not necessarily finished and so can lead to
an early read of the fifo that still misses data.

So it has been removed.

Fixes: 1ea29b39f4c812ec ("spi: bcm2835aux: add bcm2835 auxiliary spi device...")
Suggested-by: Hubert Denkmair <h.denkmair@intence.de>
Signed-off-by: Martin Sperl <kernel@martin.sperl.org>
Acked-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: bcm2835aux: unifying code between polling and interrupt driven code
Martin Sperl [Sat, 30 Mar 2019 09:30:58 +0000 (09:30 +0000)]
spi: bcm2835aux: unifying code between polling and interrupt driven code

Sharing more code between polling and interrupt-driven mode.

Signed-off-by: Martin Sperl <kernel@martin.sperl.org>
Acked-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: spi-mem: export spi_mem_default_supports_op()
Naga Sureshkumar Relli [Mon, 1 Apr 2019 07:59:00 +0000 (13:29 +0530)]
spi: spi-mem: export spi_mem_default_supports_op()

Export spi_mem_default_supports_op(), so that controller drivers
can use this.
spi-mem driver already exports this using EXPORT_SYMBOL,
but not declared it in spi-mem.h.
This patch declares spi_mem_default_supports_op() in spi-mem.h and
also removes the static from the function prototype.

Signed-off-by: Naga Sureshkumar Relli <naga.sureshkumar.relli@xilinx.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: sh-msiof: Use BIT() and GENMASK()
Geert Uytterhoeven [Tue, 2 Apr 2019 14:40:21 +0000 (16:40 +0200)]
spi: sh-msiof: Use BIT() and GENMASK()

Improve maintainability by converting the register bit, bitmask, and
bitfield definitions from hexadecimal constants to constructs using
BIT(), GENMASK(), or "val << shift".

Suggested-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: spi-fsl-spi: automatically adapt bits-per-word in cpu mode
Rasmus Villemoes [Wed, 27 Mar 2019 14:30:52 +0000 (14:30 +0000)]
spi: spi-fsl-spi: automatically adapt bits-per-word in cpu mode

Taking one interrupt for every byte is rather slow. Since the
controller is perfectly capable of transmitting 32 bits at a time,
change t->bits_per-word to 32 when the length is divisible by 4 and
large enough that the reduced number of interrupts easily compensates
for the one or two extra fsl_spi_setup_transfer() calls this causes.

Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: spi-fsl-spi: allow changing bits_per_word while CS is still active
Rasmus Villemoes [Wed, 27 Mar 2019 14:30:51 +0000 (14:30 +0000)]
spi: spi-fsl-spi: allow changing bits_per_word while CS is still active

Commit c9bfcb315104 (spi_mpc83xx: much improved driver) introduced
logic to ensure bits_per_word and speed_hz stay the same for a series
of spi_transfers with CS active, arguing that

    The current driver may cause glitches on SPI CLK line since one
    must disable the SPI controller before changing any HW settings.

This sounds quite reasonable. So this is a quite naive attempt at
relaxing this sanity checking to only ensure that speed_hz is
constant - in the faint hope that if we do not causes changes to the
clock-related fields of the SPMODE register (DIV16 and PM), those
glitches won't appear.

The purpose of this change is to allow automatically optimizing large
transfers to use 32 bits-per-word; taking one interrupt for every byte
is extremely slow.

Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: pxa2xx: Unify remaing prints in pxa2xx_spi_transfer_one()
Jarkko Nikula [Fri, 29 Mar 2019 13:00:46 +0000 (15:00 +0200)]
spi: pxa2xx: Unify remaing prints in pxa2xx_spi_transfer_one()

Use SPI device pointer in the remaining two error and warning prints in
pxa2xx_spi_transfer_one() instead of platform device of the controller

It make prints in the function uniform and more useful especially the
error print here as it can reveal the driver that has mapped the DMA
itself and attempts to transfer more than the maximum supported DMA
transfer length.

Signed-off-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: pxa2xx: Use struct spi_device directly in pxa2xx_spi_transfer_one()
Jarkko Nikula [Fri, 29 Mar 2019 13:00:45 +0000 (15:00 +0200)]
spi: pxa2xx: Use struct spi_device directly in pxa2xx_spi_transfer_one()

Pointer to a SPI device is passed to pxa2xx_spi_transfer_one() so there
is no need to access it through the current SPI message pointer.

Signed-off-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: tegra114: add SPI_LSB_FIRST support
Sowjanya Komatineni [Wed, 27 Mar 2019 05:56:33 +0000 (22:56 -0700)]
spi: tegra114: add SPI_LSB_FIRST support

Tegra SPI controller supports lsb first mode. Default is MSB bit first
and on selection of SPI_LSB_FIRST through SPI mode transmission happens
with LSB bit first.

This patch adds SPI_LSB_FIRST flag to mode_bits and also configures it
on request.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: tegra114: use packed mode for 32 bits per word
Sowjanya Komatineni [Wed, 27 Mar 2019 05:56:25 +0000 (22:56 -0700)]
spi: tegra114: use packed mode for 32 bits per word

Fixes: Use packed mode for 32 bits per word transfers to increase
performance as each packet is a full 32-bit word.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: spi-fsl-spi: relax message sanity checking a little
Rasmus Villemoes [Wed, 27 Mar 2019 14:30:51 +0000 (14:30 +0000)]
spi: spi-fsl-spi: relax message sanity checking a little

The comment says that we should not allow changes (to
bits_per_word/speed_hz) while CS is active, and indeed the code below
does fsl_spi_setup_transfer() when the ->cs_change of the previous
spi_transfer was set (and for the very first transfer).

So the sanity checking is a bit too strict - we can change it to
follow the same logic as is used by the actual transfer loop.

Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: spi-fsl-spi: remove always-true conditional in fsl_spi_do_one_msg
Rasmus Villemoes [Wed, 27 Mar 2019 14:30:50 +0000 (14:30 +0000)]
spi: spi-fsl-spi: remove always-true conditional in fsl_spi_do_one_msg

__spi_validate() in the generic SPI code sets ->speed_hz and
->bits_per_word to non-zero values, so this condition is always true.

Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: tegra114: reset controller on probe
Sowjanya Komatineni [Wed, 27 Mar 2019 05:56:32 +0000 (22:56 -0700)]
spi: tegra114: reset controller on probe

Fixes: SPI driver can be built as module so perform SPI controller reset
on probe to make sure it is in valid state before initiating transfer.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: tegra114: configure dma burst size to fifo trig level
Sowjanya Komatineni [Wed, 27 Mar 2019 05:56:29 +0000 (22:56 -0700)]
spi: tegra114: configure dma burst size to fifo trig level

Fixes: Configure DMA burst size to be same as SPI TX/RX trigger levels
to avoid mismatch.

SPI FIFO trigger levels are calculated based on the transfer length.
So this patch moves DMA slave configuration to happen before start
of DMAs.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: tegra114: flush fifos
Sowjanya Komatineni [Wed, 27 Mar 2019 05:56:28 +0000 (22:56 -0700)]
spi: tegra114: flush fifos

Fixes: Flush TX and RX FIFOs before start of new transfer and on FIFO
overflow or underrun errors.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: tegra114: terminate dma and reset on transfer timeout
Sowjanya Komatineni [Wed, 27 Mar 2019 05:56:27 +0000 (22:56 -0700)]
spi: tegra114: terminate dma and reset on transfer timeout

Fixes: terminate DMA and perform controller reset on transfer timeout
to clear the FIFO's and errors.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: tegra114: fix for unpacked mode transfers
Sowjanya Komatineni [Wed, 27 Mar 2019 05:56:24 +0000 (22:56 -0700)]
spi: tegra114: fix for unpacked mode transfers

Fixes: computation of actual bytes to fill/receive in/from FIFO in unpacked
mode when transfer length is not a multiple of requested bits per word.

unpacked mode transfers fails when the transfer includes partial bytes in
the last word.

Total words to be written/read to/from FIFO is computed based on transfer
length and bits per word. Unpacked mode includes 0 padding bytes for partial
words to align with bits per word and these extra bytes are also accounted
for calculating bytes left to transfer in the current driver.

This causes extra bytes access of tx/rx buffers along with buffer index
position crossing actual length where remain_len becomes negative and due to
unsigned type, negative value is a 32 bit representation of signed value
and transferred bytes never meets the actual transfer length resulting in
transfer timeout and a hang.

This patch fixes this with proper computation of the actual bytes to fill in
FIFO during transmit and the actual bytes to read from FIFO during receive
ignoring 0 padded bytes.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: tegra114: clear packed bit for unpacked mode
Sowjanya Komatineni [Wed, 27 Mar 2019 05:56:23 +0000 (22:56 -0700)]
spi: tegra114: clear packed bit for unpacked mode

Fixes: Clear packed bit when not using packed mode.

Packed bit is not cleared when not using packed mode. This results
in transfer timeouts for the unpacked mode transfers followed by the
packed mode transfers.

Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: pic32: fix dma channels termination
Cezary Gapinski [Tue, 26 Mar 2019 21:48:59 +0000 (22:48 +0100)]
spi: pic32: fix dma channels termination

When timeout occurs DMA TX and RX channels should be stopped
instead of stopping RX channel twice time.

Signed-off-by: Cezary Gapinski <cezary.gapinski@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: tegra20-slink: change chip select action order
Randolph Maaßen [Tue, 26 Mar 2019 14:30:50 +0000 (15:30 +0100)]
spi: tegra20-slink: change chip select action order

To transfer via SPI the tegra20-slink driver first sets the command
register, which contains the chip select value, and after that the
command2 register, which contains the chip select line. This leads to a
small spike in the chip selct 0 line between the set of the value and
the selection of the chip select line.

This commit changes the order of the register writes so that first the
chip select line is chosen and then the value is set, removing the
spike.

Signed-off-by: Randolph Maaßen <gaireg@gaireg.de>
Reviewed-by: Sowjanya Komatineni <skomatineni@nvidia.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: stm32-qspi: add dma support
Ludovic Barre [Mon, 25 Mar 2019 17:01:40 +0000 (18:01 +0100)]
spi: stm32-qspi: add dma support

This patch adds the dma support for the stm32-qspi hardware.
The memory buffer constraints (lowmem, vmalloc, kmap) are taken into
account by framework. In read mode, the memory map is preferred vs
dma (due to better throughput). If the dma transfer fails the buffer
is sent by polling.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: stm32-qspi: add spi_master_put in release function
Ludovic Barre [Mon, 25 Mar 2019 17:01:39 +0000 (18:01 +0100)]
spi: stm32-qspi: add spi_master_put in release function

This patch adds spi_master_put in release function
to drop the controller's refcount.

Signed-off-by: Ludovic Barre <ludovic.barre@st.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: mt7621: Move SPI driver out of staging
Stefan Roese [Mon, 25 Mar 2019 08:29:25 +0000 (09:29 +0100)]
spi: mt7621: Move SPI driver out of staging

This patch moves the MT7621 SPI driver, which is used on some Ralink /
MediaTek MT76xx MIPS SoC's, out of the staging directory. No changes to
the source code are done in this patch.

This driver version was tested successfully on an MT7688 based platform
with an SPI NOR on CS0 and an SPI NAND on CS1 without any issues (so
far).

This patch also documents the devicetree bindings for the MT7621 SPI
device driver.

Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Rob Herring <robh@kernel.org>
Cc: Mark Brown <broonie@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: NeilBrown <neil@brown.name>
Cc: Sankalp Negi <sankalpnegi2310@gmail.com>
Cc: Chuanhong Guo <gch981213@gmail.com>
Cc: John Crispin <john@phrozen.org>
Cc: Armando Miraglia <arma2ff0@gmail.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: at91-usart: Remove duplicated checking for spi->bits_per_word
Axel Lin [Thu, 21 Mar 2019 15:53:48 +0000 (23:53 +0800)]
spi: at91-usart: Remove duplicated checking for spi->bits_per_word

This checking is already done in __spi_validate_bits_per_word().

Signed-off-by: Axel Lin <axel.lin@ingics.com>
Acked-by: Nicolas Ferre <nicolas.ferre@microchip.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: atmel-quadspi: Make atmel_qspi_get_name static
YueHaibing [Thu, 21 Mar 2019 15:16:56 +0000 (23:16 +0800)]
spi: atmel-quadspi: Make atmel_qspi_get_name static

Fix sparse warning:

drivers/spi/atmel-quadspi.c:369:12: warning:
 symbol 'atmel_qspi_get_name' was not declared. Should it be static?

Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: export tracepoint symbols to modules
Arnd Bergmann [Thu, 21 Mar 2019 12:42:25 +0000 (13:42 +0100)]
spi: export tracepoint symbols to modules

The newly added tracepoints in the spi-mxs driver cause a link
error when the driver is a loadable module:

ERROR: "__tracepoint_spi_transfer_stop" [drivers/spi/spi-mxs.ko] undefined!
ERROR: "__tracepoint_spi_transfer_start" [drivers/spi/spi-mxs.ko] undefined!

I'm not quite sure where to put the export statements, but
directly after the inclusion of the header seems as good as
any other place.

Fixes: f3fdea3af405 ("spi: mxs: add tracing to custom .transfer_one_message callback")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: lpspi: fix dataloss when SS is inactivated between every words
Clark Wang [Thu, 21 Mar 2019 09:57:12 +0000 (09:57 +0000)]
spi: lpspi: fix dataloss when SS is inactivated between every words

If we don't use CONT to keep SS activated or use DMA mode without
cs-gpio, SS will be inactivated between every words. The word here
means the data sent once which length can be set as 1/2/4 bytes.

In the isr function, we read the FSR_RXCOUNT just behind the
fsl_lpspi_read_rx_fifo. This causes the value of FSR_RXCOUNT cannot
reflect whether there is still data not sent timely. So do this
judgement by FSR_TXCOUNT.

Signed-off-by: Clark Wang <xiaoning.wang@nxp.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: pxa2xx: Debug print DMA burst size
Andy Shevchenko [Tue, 19 Mar 2019 15:48:43 +0000 (17:48 +0200)]
spi: pxa2xx: Debug print DMA burst size

It's useful during debug to see what DMA burst size is.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Tested-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Reviewed-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: pxa2xx: Introduce DMA burst size support
Andy Shevchenko [Tue, 19 Mar 2019 15:48:42 +0000 (17:48 +0200)]
spi: pxa2xx: Introduce DMA burst size support

Some masters may have different DMA burst size than hard coded default.
In such case respect the value given by DMA burst size provided via
platform data.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Tested-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Reviewed-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: dw: Add support for an optional interface clock
Phil Edworthy [Tue, 19 Mar 2019 15:52:07 +0000 (15:52 +0000)]
spi: dw: Add support for an optional interface clock

The Synopsys SSI Controller has an interface clock, but most SoCs hide
this away. However, on some SoCs you need to explicitly enable the
interface clock in order to access the registers. Therefore, add support
for an optional interface clock.

Signed-off-by: Phil Edworthy <phil.edworthy@renesas.com>
Signed-off-by: Gareth Williams <gareth.williams.jx@renesas.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agodt-bindings: snps,dw-apb-ssi: Add optional clock bindings documentation
Gareth Williams [Tue, 19 Mar 2019 15:52:06 +0000 (15:52 +0000)]
dt-bindings: snps,dw-apb-ssi: Add optional clock bindings documentation

Add documentation to the Synopsys SPI dt-bindings to support an
optional interface clock that may be used for register access.

Signed-off-by: Phil Edworthy <phil.edworthy@renesas.com>
Signed-off-by: Gareth Williams <gareth.williams.jx@renesas.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agodt-bindings: snps,dw-apb-ssi: Add mandatory clock bindings documentation
Phil Edworthy [Tue, 19 Mar 2019 15:52:05 +0000 (15:52 +0000)]
dt-bindings: snps,dw-apb-ssi: Add mandatory clock bindings documentation

The Synopsys SSI driver uses a mandatory clock that is not documented,
so detail it in the device tree bindings. Also correct the spelling of
"pins" in the "Optional Properties" section for the driver.

Signed-off-by: Phil Edworthy <phil.edworthy@renesas.com>
Signed-off-by: Gareth Williams <gareth.williams.jx@renesas.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: mxic: simplify getting .driver_data
Wolfram Sang [Tue, 19 Mar 2019 16:36:37 +0000 (17:36 +0100)]
spi: mxic: simplify getting .driver_data

We should get 'driver_data' from 'struct device' directly. Going via
platform_device is an unneeded step back and forth.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: spi-mem: stm32-qspi: stm32_qspi_pm_ops can be static
kbuild test robot [Sat, 16 Mar 2019 07:33:41 +0000 (15:33 +0800)]
spi: spi-mem: stm32-qspi: stm32_qspi_pm_ops can be static

Fixes: 2e541b64ee52 ("spi: spi-mem: stm32-qspi: add suspend/resume support")
Signed-off-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoMerge tag 'v5.1-rc1' into spi-5.2
Mark Brown [Tue, 19 Mar 2019 13:12:32 +0000 (13:12 +0000)]
Merge tag 'v5.1-rc1' into spi-5.2

Linux 5.1-rc1

5 years agospi: pxa2xx-pci: Drop unused header inclusion
Andy Shevchenko [Mon, 18 Mar 2019 14:42:37 +0000 (16:42 +0200)]
spi: pxa2xx-pci: Drop unused header inclusion

There is nothing in the driver which requires OF specific header
to be included.

Remove it for good.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: lpspi: fsl_lpspi_runtime_resume() can be static
kbuild test robot [Tue, 19 Mar 2019 01:46:33 +0000 (09:46 +0800)]
spi: lpspi: fsl_lpspi_runtime_resume() can be static

Fixes: 944c01a889d9 ("spi: lpspi: enable runtime pm for lpspi")
Signed-off-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: spi-fsl-spi: support use of the SPISEL_BOOT signal on MPC8309
Rasmus Villemoes [Wed, 6 Mar 2019 10:32:05 +0000 (10:32 +0000)]
spi: spi-fsl-spi: support use of the SPISEL_BOOT signal on MPC8309

The MPC8309 has a dedicated signal, SPISEL_BOOT, usually used as chip
select for the flash device from which the bootloader is loaded. It is
not an ordinary gpio, but is simply controlled via the SPI_CS register
in the system configuration.

To allow accessing such a spi slave, we need to teach
fsl_spi_cs_control() how to control the SPISEL_BOOT signal. To
distinguish the gpio-controlled slaves, continue to have those use
chip_select values of 0..ngpios-1, and use chip_select == ngpios for
the boot flash.

I'm not too happy with all the ifdeffery, but it seems to be necessary
for guarding the sysdev/fsl_soc.h and use of
get_immrbase() (spi-fsl-lib.c already contains similar ifdeffery).

Googling suggests that the MPC8306 is similar, with the SPI_CS
register at the same offset.

Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: orion: Support spi_xfer->word_delay_usecs
Jan Kundrát [Thu, 7 Mar 2019 14:38:35 +0000 (15:38 +0100)]
spi: orion: Support spi_xfer->word_delay_usecs

Signed-off-by: Jan Kundrát <jan.kundrat@cesnet.cz>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: spidev: Enable control of inter-word delays
Jan Kundrát [Thu, 7 Mar 2019 14:29:42 +0000 (15:29 +0100)]
spi: spidev: Enable control of inter-word delays

Commit b7bb367afa4b added support for inserting delays in between
individual words within a single SPI transaction. This makes it
accessible from userspace.

WARNING: This delay is silently ignored unless the SPI controller
implements extra support for it. This is similar to how the in-kernel
users handle the other existing property, spi_transfer->word_delay.

Signed-off-by: Jan Kundrát <jan.kundrat@cesnet.cz>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: lpspi: Code cleanup
Clark Wang [Wed, 6 Mar 2019 06:30:49 +0000 (06:30 +0000)]
spi: lpspi: Code cleanup

Delete the extra space.

Signed-off-by: Clark Wang <xiaoning.wang@nxp.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: lpspi: Add the missing NULL check
Clark Wang [Wed, 6 Mar 2019 06:30:47 +0000 (06:30 +0000)]
spi: lpspi: Add the missing NULL check

The spi_transfer *t will be used in one transfer whatever. If t is NULL,
there has no need to try sending data, so add an error return here.

Signed-off-by: Clark Wang <xiaoning.wang@nxp.com>
Acked-by: Fugang Duan <Fugang.duan@nxp.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: lpspi: add dma mode support
Clark Wang [Wed, 6 Mar 2019 06:30:45 +0000 (06:30 +0000)]
spi: lpspi: add dma mode support

Add dma mode support for LPSPI. Any frame longer than half txfifosize will
be sent by dma mode.

For now, there are some limits:
1. The maximum transfer speed in master mode depends on the slave device,
   at least 40MHz(tested by spi-nor on 8qm-lpddr4-arm2 base board);
2. The maximum transfer speed in slave mode is 15MHz(imx7ulp),
   22MHz(8qm/qxp). In order to reach the maximum speed which is mentioned
   in datasheet, the load of connect wires between master and slave
   should be less than 15pF.

Signed-off-by: Clark Wang <xiaoning.wang@nxp.com>
Acked-by: Fugang Duan <Fugang.duan@nxp.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: lpspi: use the core way to implement cs-gpio function
Clark Wang [Wed, 6 Mar 2019 06:30:43 +0000 (06:30 +0000)]
spi: lpspi: use the core way to implement cs-gpio function

Use the default implementation of transfer_one_msg/chipselect/setup
functions in spi core to implement cs-gpio control.
Use fsl_lpspi_prepare_message to init the cs_gpio pin.

Signed-off-by: Clark Wang <xiaoning.wang@nxp.com>
Acked-by: Fugang Duan <Fugang.duan@nxp.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: lpspi: add the error info of transfer speed setting
Clark Wang [Wed, 6 Mar 2019 06:30:41 +0000 (06:30 +0000)]
spi: lpspi: add the error info of transfer speed setting

Add a error info when set a speed which greater than half of per-clk of
spi module.

The minimum SCK period is 2 cycles(CCR[SCKDIV]). So the maximum transfer
speed is half of spi per-clk.

Signed-off-by: Clark Wang <xiaoning.wang@nxp.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: lpspi: enable runtime pm for lpspi
Han Xu [Wed, 6 Mar 2019 06:30:39 +0000 (06:30 +0000)]
spi: lpspi: enable runtime pm for lpspi

Enable the runtime power management for lpspi module.

Do some adaptation work from kernel 4.9 to 4.14.

Signed-off-by: Clark Wang <xiaoning.wang@nxp.com>
Signed-off-by: Han Xu <han.xu@nxp.com>
Reviewed-by: Frank Li <frank.li@nxp.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agodoc: lpspi: Document DT bindings for LPSPI clocks
Clark Wang [Wed, 6 Mar 2019 06:30:36 +0000 (06:30 +0000)]
doc: lpspi: Document DT bindings for LPSPI clocks

Add introductions of clocks and clock-names strings.

Signed-off-by: Clark Wang <xiaoning.wang@nxp.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi: lpspi: Add i.MX8 boards support for lpspi
Clark Wang [Wed, 6 Mar 2019 06:30:34 +0000 (06:30 +0000)]
spi: lpspi: Add i.MX8 boards support for lpspi

Add both ipg and per clock for lpspi to support i.MX8QM/QXP boards.

Signed-off-by: Clark Wang <xiaoning.wang@nxp.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agospi : spi-topcliff-pch: Fix to handle empty DMA buffers
Aditya Pakki [Wed, 13 Mar 2019 16:55:41 +0000 (11:55 -0500)]
spi : spi-topcliff-pch: Fix to handle empty DMA buffers

pch_alloc_dma_buf allocated tx, rx DMA buffers which can fail. Further,
these buffers are used without a check. The patch checks for these
failures and sends the error upstream.

Signed-off-by: Aditya Pakki <pakki001@umn.edu>
Signed-off-by: Mark Brown <broonie@kernel.org>
5 years agoLinux 5.1-rc1
Linus Torvalds [Sun, 17 Mar 2019 21:22:26 +0000 (14:22 -0700)]
Linux 5.1-rc1

5 years agoMerge tag 'kbuild-v5.1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy...
Linus Torvalds [Sun, 17 Mar 2019 20:25:26 +0000 (13:25 -0700)]
Merge tag 'kbuild-v5.1-2' of git://git./linux/kernel/git/masahiroy/linux-kbuild

Pull more Kbuild updates from Masahiro Yamada:

 - add more Build-Depends to Debian source package

 - prefix header search paths with $(srctree)/

 - make modpost show verbose section mismatch warnings

 - avoid hard-coded CROSS_COMPILE for h8300

 - fix regression for Debian make-kpkg command

 - add semantic patch to detect missing put_device()

 - fix some warnings of 'make deb-pkg'

 - optimize NOSTDINC_FLAGS evaluation

 - add warnings about redundant generic-y

 - clean up Makefiles and scripts

* tag 'kbuild-v5.1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
  kconfig: remove stale lxdialog/.gitignore
  kbuild: force all architectures except um to include mandatory-y
  kbuild: warn redundant generic-y
  Revert "modsign: Abort modules_install when signing fails"
  kbuild: Make NOSTDINC_FLAGS a simply expanded variable
  kbuild: deb-pkg: avoid implicit effects
  coccinelle: semantic code search for missing put_device()
  kbuild: pkg: grep include/config/auto.conf instead of $KCONFIG_CONFIG
  kbuild: deb-pkg: introduce is_enabled and if_enabled_echo to builddeb
  kbuild: deb-pkg: add CONFIG_ prefix to kernel config options
  kbuild: add workaround for Debian make-kpkg
  kbuild: source include/config/auto.conf instead of ${KCONFIG_CONFIG}
  unicore32: simplify linker script generation for decompressor
  h8300: use cc-cross-prefix instead of hardcoding h8300-unknown-linux-
  kbuild: move archive command to scripts/Makefile.lib
  modpost: always show verbose warning for section mismatch
  ia64: prefix header search path with $(srctree)/
  libfdt: prefix header search paths with $(srctree)/
  deb-pkg: generate correct build dependencies

5 years agoMerge branch 'x86-asm-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 17 Mar 2019 16:21:48 +0000 (09:21 -0700)]
Merge branch 'x86-asm-for-linus' of git://git./linux/kernel/git/tip/tip

Pull x86 asm updates from Thomas Gleixner:
 "Two cleanup patches removing dead conditionals and unused code"

* 'x86-asm-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/asm: Remove unused __constant_c_x_memset() macro and inlines
  x86/asm: Remove dead __GNUC__ conditionals

5 years agoMerge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 17 Mar 2019 16:19:22 +0000 (09:19 -0700)]
Merge branch 'perf-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull perf fixes from Thomas Gleixner:
 "Three fixes for the fallout from the TSX errata workaround:

   - Prevent memory corruption caused by a unchecked out of bound array
     index.

   - Two trivial fixes to address compiler warnings"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/x86/intel: Make dev_attr_allow_tsx_force_abort static
  perf/x86: Fixup typo in stub functions
  perf/x86/intel: Fix memory corruption

5 years agoMerge tag 'for-linus-5.1b-rc1b-tag' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 17 Mar 2019 16:16:22 +0000 (09:16 -0700)]
Merge tag 'for-linus-5.1b-rc1b-tag' of git://git./linux/kernel/git/xen/tip

Pull xen fix from Juergen Gross:
 "A fix for a Xen bug introduced by David's series for excluding
  ballooned pages in vmcores"

* tag 'for-linus-5.1b-rc1b-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
  xen/balloon: Fix mapping PG_offline pages to user space

5 years agoMerge tag '9p-for-5.1' of git://github.com/martinetd/linux
Linus Torvalds [Sun, 17 Mar 2019 16:10:56 +0000 (09:10 -0700)]
Merge tag '9p-for-5.1' of git://github.com/martinetd/linux

Pull 9p updates from Dominique Martinet:
 "Here is a 9p update for 5.1; there honestly hasn't been much.

  Two fixes (leak on invalid mount argument and possible deadlock on
  i_size update on 32bit smp) and a fall-through warning cleanup"

* tag '9p-for-5.1' of git://github.com/martinetd/linux:
  9p/net: fix memory leak in p9_client_create
  9p: use inode->i_lock to protect i_size_write() under 32-bit
  9p: mark expected switch fall-through

5 years agoperf/x86/intel: Make dev_attr_allow_tsx_force_abort static
kbuild test robot [Wed, 13 Mar 2019 18:42:43 +0000 (02:42 +0800)]
perf/x86/intel: Make dev_attr_allow_tsx_force_abort static

Fixes: 400816f60c54 ("perf/x86/intel: Implement support for TSX Force Abort")
Signed-off-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: "Peter Zijlstra (Intel)" <peterz@infradead.org>
Cc: kbuild-all@01.org
Cc: Borislav Petkov <bp@alien8.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/20190313184243.GA10820@lkp-sb-ep06
5 years agokconfig: remove stale lxdialog/.gitignore
Masahiro Yamada [Sun, 17 Mar 2019 06:05:03 +0000 (15:05 +0900)]
kconfig: remove stale lxdialog/.gitignore

When this .gitignore was added, lxdialog was an independent hostprogs-y.

Now that all objects in lxdialog/ are directly linked to mconf, the
lxdialog is no longer generated.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
5 years agokbuild: force all architectures except um to include mandatory-y
Masahiro Yamada [Sun, 17 Mar 2019 02:01:09 +0000 (11:01 +0900)]
kbuild: force all architectures except um to include mandatory-y

Currently, every arch/*/include/uapi/asm/Kbuild explicitly includes
the common Kbuild.asm file. Factor out the duplicated include directives
to scripts/Makefile.asm-generic so that no architecture would opt out
of the mandatory-y mechanism.

um is not forced to include mandatory-y since it is a very exceptional
case which does not support UAPI.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
5 years agokbuild: warn redundant generic-y
Masahiro Yamada [Sun, 17 Mar 2019 02:01:08 +0000 (11:01 +0900)]
kbuild: warn redundant generic-y

The generic-y is redundant under the following condition:

 - arch has its own implementation

 - the same header is added to generated-y

 - the same header is added to mandatory-y

If a redundant generic-y is found, the warning like follows is displayed:

  scripts/Makefile.asm-generic:20: redundant generic-y found in arch/arm/include/asm/Kbuild: timex.h

I fixed up arch Kbuild files found by this.

Suggested-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
5 years agoRevert "modsign: Abort modules_install when signing fails"
Douglas Anderson [Fri, 15 Mar 2019 16:25:03 +0000 (09:25 -0700)]
Revert "modsign: Abort modules_install when signing fails"

This reverts commit caf6fe91ddf62a96401e21e9b7a07227440f4185.

The commit was fine but is no longer needed as of commit 3a2429e1faf4
("kbuild: change if_changed_rule for multi-line recipe").  Let's go
back to using ";" to be consistent.

For some discussion, see:

https://lkml.kernel.org/r/CAK7LNASde0Q9S5GKeQiWhArfER4S4wL1=R_FW8q0++_X3T5=hQ@mail.gmail.com

Signed-off-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
5 years agokbuild: Make NOSTDINC_FLAGS a simply expanded variable
Douglas Anderson [Thu, 14 Mar 2019 23:41:59 +0000 (16:41 -0700)]
kbuild: Make NOSTDINC_FLAGS a simply expanded variable

During a simple no-op (nothing changed) build I saw 39 invocations of
the C compiler with the argument "-print-file-name=include".  We don't
need to call the C compiler 39 times for this--one time will suffice.

Let's change NOSTDINC_FLAGS to a simply expanded variable to avoid
this since there doesn't appear to be any reason it should be
recursively expanded.

On my build this shaved ~400 ms off my "no-op" build.

Note that the recursive expansion seems to date back to the (really
old) commit e8f5bdb02ce0 ("[PATCH] Makefile include path ordering").
It's a little unclear to me if the point of that patch was to switch
the variable to be recursively expanded (which it did) or to avoid
directly assigning to NOSTDINC_FLAGS (AKA to switch to +=) because
someone else (out of tree?) was setting it.  I presume later since if
the only goal was to switch to recursive expansion the patch would
have just removed the ":".

Signed-off-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
5 years agokbuild: deb-pkg: avoid implicit effects
Arseny Maslennikov [Sat, 9 Mar 2019 15:43:06 +0000 (18:43 +0300)]
kbuild: deb-pkg: avoid implicit effects

* The man page for dpkg-source(1) notes:

>      -b, --build directory [format-specific-parameters]
>             Build  a  source  package  (--build since dpkg 1.17.14).
>             <...>
>
>             dpkg-source will build the source package with the first
>             format found in this ordered list: the format  indicated
>             with  the  --format  command  line  option,  the  format
>             indicated in debian/source/format, “1.0”.  The  fallback
>             to “1.0” is deprecated and will be removed at some point
>             in the future, you should always  document  the  desired
>             source   format  in  debian/source/format.  See  section
>             SOURCE PACKAGE FORMATS for an extensive  description  of
>             the various source package formats.

  Thus it would be more foolproof to explicitly use 1.0 (as we always
  did) than to rely on dpkg-source's defaults.

* In a similar vein, debian/rules is not made executable by mkdebian,
  and dpkg-source warns about that but still silently fixes the file.
  Let's be explicit once again.

Signed-off-by: Arseny Maslennikov <ar@cs.msu.ru>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
5 years agococcinelle: semantic code search for missing put_device()
Wen Yang [Fri, 15 Feb 2019 07:55:19 +0000 (15:55 +0800)]
coccinelle: semantic code search for missing put_device()

The of_find_device_by_node() takes a reference to the underlying device
structure, we should release that reference.
The implementation of this semantic code search is:
In a function, for a local variable returned by calling
of_find_device_by_node(),
a, if it is released by a function such as
   put_device()/of_dev_put()/platform_device_put() after the last use,
   it is considered that there is no reference leak;
b, if it is passed back to the caller via
   dev_get_drvdata()/platform_get_drvdata()/get_device(), etc., the
   reference will be released in other functions, and the current function
   also considers that there is no reference leak;
c, for the rest of the situation, the current function should release the
   reference by calling put_device, this code search will report the
   corresponding error message.

By using this semantic code search, we have found some object reference leaks,
such as:
commit 11907e9d3533 ("ASoC: fsl-asoc-card: fix object reference leaks in
fsl_asoc_card_probe")
commit a12085d13997 ("mtd: rawnand: atmel: fix possible object reference leak")
commit 11493f26856a ("mtd: rawnand: jz4780: fix possible object reference leak")

There are still dozens of reference leaks in the current kernel code.

Further, for the case of b, the object returned to other functions may also
have a reference leak, we will continue to develop other cocci scripts to
further check the reference leak.

Signed-off-by: Wen Yang <wen.yang99@zte.com.cn>
Reviewed-by: Julia Lawall <Julia.Lawall@lip6.fr>
Reviewed-by: Markus Elfring <Markus.Elfring@web.de>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
5 years agoMerge tag 'pidfd-v5.1-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/brauner...
Linus Torvalds [Sat, 16 Mar 2019 20:47:14 +0000 (13:47 -0700)]
Merge tag 'pidfd-v5.1-rc1' of git://git./linux/kernel/git/brauner/linux

Pull pidfd system call from Christian Brauner:
 "This introduces the ability to use file descriptors from /proc/<pid>/
  as stable handles on struct pid. Even if a pid is recycled the handle
  will not change. For a start these fds can be used to send signals to
  the processes they refer to.

  With the ability to use /proc/<pid> fds as stable handles on struct
  pid we can fix a long-standing issue where after a process has exited
  its pid can be reused by another process. If a caller sends a signal
  to a reused pid it will end up signaling the wrong process.

  With this patchset we enable a variety of use cases. One obvious
  example is that we can now safely delegate an important part of
  process management - sending signals - to processes other than the
  parent of a given process by sending file descriptors around via scm
  rights and not fearing that the given process will have been recycled
  in the meantime. It also allows for easy testing whether a given
  process is still alive or not by sending signal 0 to a pidfd which is
  quite handy.

  There has been some interest in this feature e.g. from systems
  management (systemd, glibc) and container managers. I have requested
  and gotten comments from glibc to make sure that this syscall is
  suitable for their needs as well. In the future I expect it to take on
  most other pid-based signal syscalls. But such features are left for
  the future once they are needed.

  This has been sitting in linux-next for quite a while and has not
  caused any issues. It comes with selftests which verify basic
  functionality and also test that a recycled pid cannot be signaled via
  a pidfd.

  Jon has written about a prior version of this patchset. It should
  cover the basic functionality since not a lot has changed since then:

      https://lwn.net/Articles/773459/

  The commit message for the syscall itself is extensively documenting
  the syscall, including it's functionality and extensibility"

* tag 'pidfd-v5.1-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/brauner/linux:
  selftests: add tests for pidfd_send_signal()
  signal: add pidfd_send_signal() syscall

5 years agoMerge tag 'devdax-for-5.1' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm...
Linus Torvalds [Sat, 16 Mar 2019 20:05:32 +0000 (13:05 -0700)]
Merge tag 'devdax-for-5.1' of git://git./linux/kernel/git/nvdimm/nvdimm

Pull device-dax updates from Dan Williams:
 "New device-dax infrastructure to allow persistent memory and other
  "reserved" / performance differentiated memories, to be assigned to
  the core-mm as "System RAM".

  Some users want to use persistent memory as additional volatile
  memory. They are willing to cope with potential performance
  differences, for example between DRAM and 3D Xpoint, and want to use
  typical Linux memory management apis rather than a userspace memory
  allocator layered over an mmap() of a dax file. The administration
  model is to decide how much Persistent Memory (pmem) to use as System
  RAM, create a device-dax-mode namespace of that size, and then assign
  it to the core-mm. The rationale for device-dax is that it is a
  generic memory-mapping driver that can be layered over any "special
  purpose" memory, not just pmem. On subsequent boots udev rules can be
  used to restore the memory assignment.

  One implication of using pmem as RAM is that mlock() no longer keeps
  data off persistent media. For this reason it is recommended to enable
  NVDIMM Security (previously merged for 5.0) to encrypt pmem contents
  at rest. We considered making this recommendation an actively enforced
  requirement, but in the end decided to leave it as a distribution /
  administrator policy to allow for emulation and test environments that
  lack security capable NVDIMMs.

  Summary:

   - Replace the /sys/class/dax device model with /sys/bus/dax, and
     include a compat driver so distributions can opt-in to the new ABI.

   - Allow for an alternative driver for the device-dax address-range

   - Introduce the 'kmem' driver to hotplug / assign a device-dax
     address-range to the core-mm.

   - Arrange for the device-dax target-node to be onlined so that the
     newly added memory range can be uniquely referenced by numa apis"

NOTE! I'm not entirely happy with the whole "PMEM as RAM" model because
we currently have special - and very annoying rules in the kernel about
accessing PMEM only with the "MC safe" accessors, because machine checks
inside the regular repeat string copy functions can be fatal in some
(not described) circumstances.

And apparently the PMEM modules can cause that a lot more than regular
RAM.  The argument is that this happens because PMEM doesn't necessarily
get scrubbed at boot like RAM does, but that is planned to be added for
the user space tooling.

Quoting Dan from another email:
 "The exposure can be reduced in the volatile-RAM case by scanning for
  and clearing errors before it is onlined as RAM. The userspace tooling
  for that can be in place before v5.1-final. There's also runtime
  notifications of errors via acpi_nfit_uc_error_notify() from
  background scrubbers on the DIMM devices. With that mechanism the
  kernel could proactively clear newly discovered poison in the volatile
  case, but that would be additional development more suitable for v5.2.

  I understand the concern, and the need to highlight this issue by
  tapping the brakes on feature development, but I don't see PMEM as RAM
  making the situation worse when the exposure is also there via DAX in
  the PMEM case. Volatile-RAM is arguably a safer use case since it's
  possible to repair pages where the persistent case needs active
  application coordination"

* tag 'devdax-for-5.1' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm:
  device-dax: "Hotplug" persistent memory for use like normal RAM
  mm/resource: Let walk_system_ram_range() search child resources
  mm/memory-hotplug: Allow memory resources to be children
  mm/resource: Move HMM pr_debug() deeper into resource code
  mm/resource: Return real error codes from walk failures
  device-dax: Add a 'modalias' attribute to DAX 'bus' devices
  device-dax: Add a 'target_node' attribute
  device-dax: Auto-bind device after successful new_id
  acpi/nfit, device-dax: Identify differentiated memory with a unique numa-node
  device-dax: Add /sys/class/dax backwards compatibility
  device-dax: Add support for a dax override driver
  device-dax: Move resource pinning+mapping into the common driver
  device-dax: Introduce bus + driver model
  device-dax: Start defining a dax bus model
  device-dax: Remove multi-resource infrastructure
  device-dax: Kill dax_region base
  device-dax: Kill dax_region ida