Merge branch 'for-next/of-overlay'

firmware: add support to load firmware from dt overlay ...

fpga-region device tree nodes have the firmware-name property that
contains the file name of firmware in the firmware search path (but not
the path) that shall be loaded before the overlay is applied.

Add the of_firmware_load_overlay() function that accepts an overlay and a
firmware search path, finds the responsible firmware_mgr and loads the
firmware.

Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: add iterator for overlays ...

Device tree overlays (the dto files) may contain multiple fragments for
different target nodes. Each fragment contains a __overlay__ node that
is applied to target node specified in the fragment.

Add a helper to call a function for each fragment in a device tree
overlay to avoid having device tree overlay internal information in
other modules.

Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: add support for devicetree overlays ...

The devicetree overlay support is based on the Linux driver for device
tree overlays, but many features that are not required in Barebox are
left out.

Unlike Linux, which applies the overlay to the live devicetree, Barebox
registers a fixup for the overlay which is applied with other fixups to
whatever tree is fixed. This is necessary to apply the overlay to
devicetrees that are passed to Linux, which might differ from the
devicetree that is currently live in Barebox.

Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: add of_diff() ...

of_diff compares two device trees against each other and prints a
diff-like result.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

drivers: of: Add function to enable and register a device by alias ...

In some cases node names are not unique and passing the full path is really long.
So make add a new device by passing the alias string possible.

Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

mdio_bus: Allow for non PHY-devices on MDIO buses ...

Instead of just creating a simple PHY device for every child of MDIO
bus node, add code to check if any of them have "compatible" property
set, as well as code to create a proper platform device for such
cases.

This change is useful for when MDIO bus and some of Ethernet ports are
connected to a switch or some other MDIO device that doesn't behave
like a generic PHY and can't be probed via its PHY ID.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: Make of_property_get_value() public ...

Make of_property_get_value() public, so it can be used in other part
of the system.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

Merge branch 'for-next/of'

of: Add function to find the same node in another tree ...

For our device tree fixups we have to find a node corresponding to
another node in another device tree. We used to use the full name
to match the nodes, but this falls apart when nodes get renamed
or for example a new bus hierarchy is introduced. To make this
more robust we create reproducible names from device nodes which
mostly depend on the address in MMIO space, the reg property and
as a last resort the name of the device node.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: fdt: add of_unflatten_dtb_const ...

This adds a variant of of_unflatten_dtb() which uses the property data
directly from the input tree rather than copying it. This is mainly
useful for a single user: FIT images.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: Allow for const properties ...

This adds support for const properties. We add a const void *value_const
to struct property which will not be modified. The value of a property
should no longer be used directly, but with the of_property_get_value()
accessor which picks the right value.
With this we can unflatten dtbs and use the property values directly
from the dtb instead of copying them. This is useful for device trees
which (ab)use properties to store huge data files, aka FIT images.
To create a property whose value is not copied but used from the
original input data we introduce of_new_property_const().

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: base: use root_node compatible as suggestion for a hostname ...

on some SoCs we can use generic PLL and RAM initialization. In this
cases we create board file only to provide a host name.
With this patch host name will be created from device tree compatible.
For example:
 compatible = "board_vendor,board", "chip_vendor,soc"
the host name will be:
 "board"

This function will not overwrite a host name which is already set by
board or machine code.

Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

common: oftree: Add autoenable functionality ...

This patch adds an API to automatically enable either hardware components
with existing device drivers or i2c clients. All functions take a device
tree path to find the hardware and will fix up the node status in the
kernel device tree, if it's accessible.

Signed-off-by: Daniel Schultz <d.schultz@phytec.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: base: add function to copy a device tree node ...

Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: base: port of_device_is_big_endian from linux ...

|commit 37786c7
|Author: Kevin Cernekee <cernekee@gmail.com>
|Date:   Thu Apr 9 13:05:14 2015 -0700
|
|    of: Add helper function to check MMIO register endianness
|
|    SoC peripherals can come in several different flavors:
|
|     - little-endian: registers always need to be accessed in LE mode (so the
|       kernel should perform a swap if the CPU is running BE)
|
|     - big-endian: registers always need to be accessed in BE mode (so the
|       kernel should perform a swap if the CPU is running LE)
|
|     - native-endian: the bus will automatically swap accesses, so the kernel
|       should never swap
|
|    Introduce a function that checks an OF device node to see whether it
|    contains a "big-endian" or "native-endian" property.  For the former case,
|    always return true.  For the latter case, return true iff the kernel was
|    built for BE (implying that the BE MMIO accessors do not perform a swap).
|    Otherwise return false, assuming LE registers.
|
|    LE registers are assumed by default because most existing drivers (libahci,
|    serial8250, usb) always use readl/writel in the absence of instructions
|    to the contrary, so that will be our fallback.

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: of_path: add of_find_node_by_devpath() ...

We already have of_find_path_by_node() which finds a device path
for a given device node. This is used by the state framework to
find its backend path. This path has to be translated back to
a device node when Linux is started. The current approach turned
out to be too simple: We assumed that the node path is the same
in the tree Linux is started with. This is not true in several
cases:
- partition nodes should have the name "partition@<offset>", but
  in several trees they have "partition@<num>"
- There are two different partition bindings: The legacy binding
  and the new one with an additional partitions subnode which has
  a compatible = "fixed-partitions" property.
The node path only stays the same when the internal tree uses the
same bindings and node names as the tree Linux is started with.

To overcome this limitation we create of_find_node_by_devpath()
which converts the device path back to a device node. It does
so by finding the device node of the hardware device rather than
the partition node. It then parses over the partitions in this
device node with the known bindings looking for a partition which
matches in offset and size to the barebox partition.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: partition: Move of_mtd_fixup to drivers/of/ ...

Move the fixup code where the parser code is already.
Since the code will not only be used for mtd in the future
drivers/of/ is a better place than drivers/mtd/.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: Add of_property_write_string() ...

Setting a property to a string is used many times. Create a convenience
function for it.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: provide for_each_matching_node_from ...

This is for_each_matching_node for a given root similar to the other
..._from functions.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: base: import of_get_cpu_node() from linux-v4.7-rc2 ...

Signed-off-by: Antony Pavlov <antonynpavlov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

common: oftree: Add function to register set status fixup ...

Added a function to register a fixup to enable or disable
device tree nodes.

Signed-off-by: Teresa Remmet <t.remmet@phytec.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: Add of_property_for_each_phandle() iterator ...

This is like of_property_for_each_{string,u32} but loops over a list
of phandles instead of strings or ints.

Returns the node the handle points to as that's generally more useful
than the handle value itself.

Signed-off-by: Trent Piepho <tpiepho@kymetacorp.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: Add for_each_compatible_node_from iterator ...

The existing iterator for_each_compatible_node() searches for each
compatible node starting from the root of the loaded device tree.
This means it only works on the barebox device tree and not the tree
to be passed to the Linux kernel, which is what an of_fixup would
probably want to use.

This adds for_each_compatible_node_from(), which takes an additional
parameter of a root to search from.  This could be the device tree to
be used for the kernel.  It could also be used to search just a
subtree.

Signed-off-by: Trent Piepho <tpiepho@kymetacorp.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of_path: add of_find_path_by_node() ...

This patch adds the function of_find_path_by_node(), which is similar to
of_find_path(), but it translates a device tree node into a barebox device path
directly.

Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: base: implement of_get_next_child ...

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: export of_platform_device_create ...

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: add a function to remove an of_fixup ...

This function is needed when a device that already registered a fixup in
the probe routine fails later to probe completely. Without unregistering
the fixup the function might later be called with invalid data.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

Merge branch 'for-next/of-device-id'

of: use 'const void *' for struct of_device_id.data ...

Since 2011 barebox' of_device_id struct uses unsigned long type for data field:

    struct of_device_id {
            char *compatible;
            unsigned long data;
    };

Almost always struct of_device_id.data field are used as pointer
and need 'unsigned long' casting.

E.g. see 'git grep -A 4 of_device_id drivers/' output:

    drivers/ata/sata-imx.c:static __maybe_unused struct of_device_id imx_sata_dt_ids[] = {
    drivers/ata/sata-imx.c- {
    drivers/ata/sata-imx.c-         .compatible = "fsl,imx6q-ahci",
    drivers/ata/sata-imx.c-         .data = (unsigned long)&data_imx6,
    drivers/ata/sata-imx.c- }, {

Here is of_device_id struct in linux kernel v4.0:

    struct of_device_id {
            char name[32];
            char type[32];
            char compatible[128];
            const void *data;
    };

Changing of_device_id.data type to 'const void *data' will increase
barebox' linux kernel compatibility and decrease number of 'unsigned
long' casts.

Part of the patch was done using the 'coccinelle' tool with the
following semantic patch:

    @rule1@
    identifier dev;
    identifier type;
    identifier func;
    @@
    func(...) {
    <...
    - dev_get_drvdata(dev, (unsigned long *)&type)
    + dev_get_drvdata(dev, (const void **)&type)
    ...>
    }
    @rule2@
    identifier dev;
    identifier type;
    identifier func;
    identifier data;
    @@
    func(...) {
    <...
    - dev_get_drvdata(dev, (unsigned long *)&type->data)
    + dev_get_drvdata(dev, (const void **)&type->data)
    ...>
    }

Signed-off-by: Antony Pavlov <antonynpavlov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>