After the move to having device nodes be proper kobjects the lifecycle
of the node needs to be controlled better.

At first convert of_add_node() in the unflattened functions to
of_init_node() which initializes the kobject so that of_node_get/put
work correctly even before of_init is called.

Afterwards introduce of_node_is_initialized & of_node_is_attached that
query the underlying kobject about the state (attached means kobj
is visible in sysfs)

Using that make sure the lifecycle of the tree is correct at all
times.
Signed-off-by: NPantelis Antoniou <panto@antoniou-consulting.com>
[grant.likely: moved of_node_init() calls, fixed up locking, and
               dropped __of_populate() hunks]
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

The same data is now available in sysfs, so we can remove the code
that exports it in /proc and replace it with a symlink to the sysfs
version.

Tested on versatile qemu model and mpc5200 eval board. More testing
would be appreciated.

v5: Fixed up conflicts with mainline changes
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>
Cc: Rob Herring <rob.herring@calxeda.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: David S. Miller <davem@davemloft.net>
Cc: Nathan Fontenot <nfont@linux.vnet.ibm.com>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>

Device tree nodes are already treated as objects, and we already want to
expose them to userspace which is done using the /proc filesystem today.
Right now the kernel has to do a lot of work to keep the /proc view in
sync with the in-kernel representation. If device_nodes are switched to
be kobjects then the device tree code can be a whole lot simpler. It
also turns out that switching to using /sysfs from /proc results in
smaller code and data size, and the userspace ABI won't change if
/proc/device-tree symlinks to /sys/firmware/devicetree/base.

v7: Add missing sysfs_bin_attr_init()
v6: Add __of_add_property() early init fixes from Pantelis
v5: Rename firmware/ofw to firmware/devicetree
    Fix updating property values in sysfs
v4: Fixed build error on Powerpc
    Fixed handling of dynamic nodes on powerpc
v3: Fixed handling of duplicate attribute and child node names
v2: switch to using sysfs bin_attributes which solve the problem of
    reporting incorrect property size.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>
Tested-by: NSascha Hauer <s.hauer@pengutronix.de>
Cc: Rob Herring <rob.herring@calxeda.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: David S. Miller <davem@davemloft.net>
Cc: Nathan Fontenot <nfont@linux.vnet.ibm.com>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>

This reverts commit 02ed594e.

The change is completely broken. It attempt to get the previous item in
a linked list by grabbing the address of a stack variable. Outright
wrong.
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

In the current implementation of __of_match_node(), it will compare
each given match entry against all the node's compatible strings
with of_device_is_compatible().

To achieve multiple compatible strings per node with ordering from
specific to generic, this requires given matches to be ordered from
specific to generic. For most of the drivers this is not true and
also an alphabetical ordering is more sane there.

Therefore, we define a following priority order for the match, and
then scan all the entries to find the best match.
  1. specific compatible && type && name
  2. specific compatible && type
  3. specific compatible && name
  4. specific compatible
  5. general compatible && type && name
  6. general compatible && type
  7. general compatible && name
  8. general compatible
  9. type && name
  10. type
  11. name

v5: Fix nested locking bug
v4: Short-circuit failure cases instead of mucking with score, and
    remove extra __of_device_is_compatible() wrapper stub.
    Move scoring logic directly into __of_device_is_compatible()
v3: Also need to bail out when there does have a compatible member in match
    entry, but it doesn't match with the device node's compatible.
v2: Fix the bug such as we get the same score for the following two match
    entries with the empty node 'name2 { };'
	struct of_device_id matches[] = {
		{.name = "name2", },
		{.name = "name2", .type = "type1", },
		{}
	};
Signed-off-by: NKevin Hao <haokexin@gmail.com>
[grant.likely: added v4 changes]
Signed-off-by: NGrant Likely <grant.likely@linaro.org>
Tested-by: NPaul Gortmaker <paul.gortmaker@windriver.com>
Tested-by: NStephen Chivers <schivers@csc.com>
Tested-by: NSachin Kamat <sachin.kamat@linaro.org>

This reverts commit 06b29e76.
As pointed out by Grant Likely, we should also take the type and name
into account when searching the best compatible match. That means the
match with compatible, type and name should be better than the match
just with the same compatible string. So revert this and we will
implement another method to find the best match entry.
Signed-off-by: NKevin Hao <haokexin@gmail.com>
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

Currently, of_match_node compares each given match against all node's
compatible strings with of_device_is_compatible.

To achieve multiple compatible strings per node with ordering from
specific to generic, this requires given matches to be ordered from
specific to generic. For most of the drivers this is not true and also
an alphabetical ordering is more sane there.

Therefore, this patch introduces a function to match each of the node's
compatible strings against all given compatible matches without type and
name first, before checking the next compatible string. This implies
that node's compatibles are ordered from specific to generic while
given matches can be in any order. If we fail to find such a match
entry, then fall-back to the old method in order to keep compatibility.

Cc: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: NKevin Hao <haokexin@gmail.com>
Tested-by: NStephen Chivers <schivers@csc.com>
Signed-off-by: NRob Herring <robh@kernel.org>

This reverts commit 10535314.
Stephen Chivers reported this is broken as we will get a match
entry '.type = "serial"' instead of the '.compatible = "ns16550"'
in the following scenario:
	serial0: serial@4500 {
		compatible = "fsl,ns16550", "ns16550";
	}

	struct of_device_id of_platform_serial_table[] = {
		{ .compatible = "ns8250",   .data = (void *)PORT_8250, },
		{ .compatible = "ns16450",  .data = (void *)PORT_16450, },
		{ .compatible = "ns16550a", .data = (void *)PORT_16550A, },
		{ .compatible = "ns16550",  .data = (void *)PORT_16550, },
		{ .compatible = "ns16750",  .data = (void *)PORT_16750, },
		{ .compatible = "ns16850",  .data = (void *)PORT_16850, },
		...
		{ .type = "serial",         .data = (void *)PORT_UNKNOWN, },
		{ /* end of list */ },
	};

So just revert this patch, we will use another implementation to find
the best compatible match in a follow-on patch.
Reported-by: NStephen N Chivers <schivers@csc.com.au>
Cc: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: NKevin Hao <haokexin@gmail.com>
Signed-off-by: NRob Herring <robh@kernel.org>

The of_update_property() is intented to update a property in a node
and if the property does not exist, will add it.

The second search of the property is possibly won't be found, that
maybe removed by other thread just before the second search begain.

Using the __of_find_property() and __of_add_property() instead and
move them into lock operations.
Signed-off-by: NXiubo Li <Li.Xiubo@freescale.com>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

There two places will use the same code for adding one new property to
the DT node. Adding __of_add_property() and prepare for fixing
of_update_property()'s bug.
Signed-off-by: NXiubo Li <Li.Xiubo@freescale.com>
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

When attaching a node always clear the detach flag. Without this change
the sequence detach, attach fails.
Signed-off-by: NPantelis Antoniou <panto@antoniou-consulting.com>
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

From IEEE 1275, there defined a standard 'status' property indicating the
operational status of one device. The 'status' property has four possible
values: 'okay/ok', 'disabled', 'fail' and 'fail-xxx'.

If it is absent, that means the status of the device is unknown or okay.

The __of_device_is_available checks the state of the 'status' property of
a device. If the property is absent or set to 'okay/ok', it returns 1.
Otherwise it returns 0.

While in __of_device_is_available:
 >	status = of_get_property(device, "status", &statlen);
 >	if (status == NULL)
 >		return 1;
The status value returned from 'of_get_property()' will be NULL in two cases:
Firstly: the 'device' value (device node) is NULL.
Secondly: the 'status' property is actaully not exist.

If the device node is NULL, the __of_device_is_available will return true,
that will mean the absent state of the 'status' property.

So this add the device node check before checking the 'status' property's
state, and if the device node is not exist, 0 will be returned.
Signed-off-by: NXiubo Li <Li.Xiubo@freescale.com>
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

Currently, of_match_node compares each given match against all node's
compatible strings with of_device_is_compatible.

To achieve multiple compatible strings per node with ordering from
specific to generic, this requires given matches to be ordered from
specific to generic. For most of the drivers this is not true and also
an alphabetical ordering is more sane there.

Therefore, this patch modifies of_match_node to match each of the node's
compatible strings against all given matches first, before checking the
next compatible string. This implies that node's compatibles are ordered
from specific to generic while given matches can be in any order.
Signed-off-by: NSebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Tested-by: NMeelis Roos <mroos@linux.ee>
Signed-off-by: NRob Herring <rob.herring@calxeda.com>

Since the definition of_find_next_cache_node is architecture independent,
the existing definition in powerpc can be moved to driver/of/base.c

Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Grant Likely <grant.likely@linaro.org>
Cc: Rob Herring <rob.herring@calxeda.com>
Signed-off-by: NSudeep KarkadaNagesha <sudeep.karkadanagesha@arm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

It is sometimes useful for debug to get the contents of an
of_phandle_args structure out into the kernel log.
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

The size of each hwid in a cpu nodes 'reg' property is defined by the
parents #address-cells property in the normal way. The cpu parsing code
has a bug where it will overrun the end of the property if
address-cells is greater than one. This commit fixes the problem by
adjusting the array size by the number of address cells. It also makes
sure address-cells isn't zero for that would cause an infinite loop.

v2: bail if #address-cells is zero instead of forcing to
    OF_ROOT_NODE_ADDR_CELLS_DEFAULT. Forcing it will cause the reg
    property to be parsed incorrectly.
Signed-off-by: NGrant Likely <grant.likely@linaro.org>
Cc: Rob Herring <rob.herring@calxeda.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>

Use for_each_node_by_type() to iterate all cpu nodes in the
system.

Provide and overridable function arch_find_n_match_cpu_physical_id,
which sees if the given device node matches 'cpu' and if so sets
'*thread' when non-NULL to the cpu thread number within the core.

The default implementation behaves the same as the existing code.

Add a sparc64 implementation.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
Tested-by: NSudeep KarkadaNagesha <Sudeep.KarkadaNagesha@arm.com>
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

Not all DT platforms have all the cpus collected under a /cpus node.
That just happens to be a details of FDT, ePAPR and PowerPC platforms.
Sparc does something different, but unfortunately the current code
complains with a warning if /cpus isn't there. This became a problem
with commit f86e4718, "driver/core cpu: initialize of_node in cpu's
device structure", which caused the function to get called for all
architectures.

This commit is a temporary fix to fail silently if the cpus node isn't
present. A proper fix will come later to allow arch code to provide a
custom mechanism for decoding the CPU hwid if the 'reg' property isn't
appropriate.
Signed-off-by: NGrant Likely <grant.likely@linaro.org>
Cc: David Miller <davem@davemloft.net>
Cc: Sudeep KarkadaNagesha <Sudeep.KarkadaNagesha@arm.com>
Cc: Rob Herring <rob.herring@calxeda.com>

Implement of_node_to_nid as weak function to remove the dependency on
asm/prom.h. This is in preparation to make prom.h optional.
Signed-off-by: NRob Herring <rob.herring@calxeda.com>
Acked-by: NGrant Likely <grant.likely@linaro.org>

The simplest case of __of_parse_phandle_with_args() now implements the
semantics of of_parse_phandle(). Rewrite of_parse_phandle() to call
__of_parse_phandle_with_args() rather than open-coding the simple case.

Optimize __of_parse_phandle_with_args() so that it doesn't call
of_find_node_by_phandle() except when it's strictly needed. This avoids
introducing too much overhead when replacing of_parse_phandle().
Signed-off-by: NStephen Warren <swarren@nvidia.com>
Acked-by: NMark Rutland <mark.rutland@arm.com>
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

This is identical to of_parse_phandle_with_args(), except that the
number of argument cells is fixed, rather than being parsed out of the
node referenced by each phandle.
Signed-off-by: NStephen Warren <swarren@nvidia.com>
Acked-by: NMark Rutland <mark.rutland@arm.com>
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

Move of_parse_phandle() after __of_parse_phandle_with_args(), since a
future patch will call __of_parse_phandle_with_args() from
of_parse_phandle(). Moving the function avoids adding a prototype. Doing
the move separately highlights the code changes separately.
Signed-off-by: NStephen Warren <swarren@nvidia.com>
Acked-by: NMark Rutland <mark.rutland@arm.com>
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

Commit bd69f73f "of: Create function for counting number of phandles in
a property" renamed of_parse_phandle_with_args(), and created a wrapper
function that implemented the original name. However, the documentation
of the original function was not moved, leaving it apparently documenting
the newly renamed function.

Move the documentation so that it is adjacent to the function it
documents.
Signed-off-by: NStephen Warren <swarren@nvidia.com>
Acked-by: NMark Rutland <mark.rutland@arm.com>
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

Any calls to dt_alloc() need to be zeroed. This is a temporary fix, but
the allocation function itself needs to zero memory before returning
it. This is a follow up to patch 9e401275, "of: fdt: fix memory
initialization for expanded DT" which fixed one call site but missed
another.
Signed-off-by: NGrant Likely <grant.likely@linaro.org>
Acked-by: NWladislav Wiebe <wladislav.kw@gmail.com>
Cc: stable@vger.kernel.org

This patch moves the generalized implementation of of_get_cpu_node from
PowerPC to DT core library, thereby adding support for retrieving cpu
node for a given logical cpu index on any architecture.

The CPU subsystem can now use this function to assign of_node in the
cpu device while registering CPUs.

It is recommended to use these helper function only in pre-SMP/early
initialisation stages to retrieve CPU device node pointers in logical
ordering. Once the cpu devices are registered, it can be retrieved easily
from cpu device of_node which avoids unnecessary parsing and matching.

Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Grant Likely <grant.likely@linaro.org>
Acked-by: NRob Herring <rob.herring@calxeda.com>
Signed-off-by: NSudeep KarkadaNagesha <sudeep.karkadanagesha@arm.com>

devicetrees may have a linux,stdout-path property in the chosen
node describing the console device. This adds a helper function
to match a device against this property so a driver can call
add_preferred_console for a matching device.
Signed-off-by: NSascha Hauer <s.hauer@pengutronix.de>
Acked-by: NRob Herring <rob.herring@calxeda.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The OF code uses irqsafe locks everywhere except in a handful of functions
for no obvious reasons. Since the conversion from the old rwlocks, this
now triggers lockdep warnings when used at interrupt time. At least one
driver (ibmvscsi) seems to be doing that from softirq context.

This converts the few non-irqsafe locks into irqsafe ones, making them
consistent with the rest of the code.
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: NThomas Gleixner <tglx@linutronix.de>
Acked-by: NDavid S. Miller <davem@davemloft.net>
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

the function of_property_read_u8/16/32_array() has a parameter
out_values, but the description mentioned it as out_value. This
patch fixes this typo.
Signed-off-by: NLad, Prabhakar <prabhakar.csengg@gmail.com>
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

Supply a function (proc_remove()) to remove a proc entry (and any subtree
rooted there) by proc_dir_entry pointer rather than by name and (optionally)
root dir entry pointer.  This allows us to eliminate all remaining pde->name
accesses outside of procfs.
Signed-off-by: NDavid Howells <dhowells@redhat.com>
Acked-by: NGrant Likely <grant.likely@linaro.or>
cc: linux-acpi@vger.kernel.org
cc: openipmi-developer@lists.sourceforge.net
cc: devicetree-discuss@lists.ozlabs.org
cc: linux-pci@vger.kernel.org
cc: netdev@vger.kernel.org
cc: netfilter-devel@vger.kernel.org
cc: alsa-devel@alsa-project.org
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Currently modular V4L2 build with enabled OF is broken dur to the
of_get_next_parent() function being unavailable to modules. Export it to
fix the build.
Signed-off-by: NGuennadi Liakhovetski <g.liakhovetski@gmx.de>
Acked-by: NGrant Likely <grant.likely@linaro.org>
Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>

Call of_node_put() only when the out_args is NULL on success,
or the node's reference count will not be correct because the caller
will call of_node_put() again.
Signed-off-by: NTang Yuantian <Yuantian.Tang@freescale.com>
[grant.likely: tightened up the patch]
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

Several functions in of/base.c have the same code duplicated for
finding and validating a property and value.

struct property *prop = of_find_property(np, propname, NULL);
if (!prop)
	return -EINVAL;
if (!prop->value)
	return -ENODATA;
if (<some length> > prop->length)
	return -EOVERFLOW;

This patch adds of_find_property_value_of_size() which performs the
equivalent of the above code and removes the instances where it was
duplicated in several functions.
Reported-by: NRob Herring <robherring2@gmail.com>
Signed-off-by: NTony Prisk <linux@prisktech.co.nz>
Acked-by: NLinus Walleij <linus.walleij@linaro.org>
Acked-by: NRob Herring <rob.herring@calxeda.com>

This patch adds an of_property_read_u32_index() function to allow
reading a single indexed u32 value from a property containing multiple
u32 values.
Signed-off-by: NTony Prisk <linux@prisktech.co.nz>
Reviewed-by: NStephen Warren <swarren@nvidia.com>
Acked-by: NLinus Walleij <linus.walleij@linaro.org>
Acked-by: NRob Herring <robherring2@gmail.com>

This patch creates of_count_phandle_with_args(), a new function for
counting the number of phandle+argument tuples in a given property. This
is better than the existing method of parsing each phandle individually
until parsing fails which is a horribly slow way to do the count.

Tested on ARM using the selftest code.

v3: - Rebased on top of selftest code cleanup patch
v2: - fix bug where of_parse_phandle_with_args() could behave like _count_.
    - made of_gpio_named_count() into a static inline regardless of CONFIG_OF_GPIO
Tested-by: NAndreas Larsson <andreas@gaisler.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>
Cc: Linus Walleij <linus.walleij@linaro.org>
Cc: Rob Herring <rob.herring@calxeda.com>

Some of the exit paths were not correctly releasing the node. Fix it by
creating an 'err' label for collecting the error paths and releasing the
node.

Cc: Rob Herring <rob.herring@calxeda.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

of_get_next_available_child() acquires devtree_lock, then calls
of_device_is_available() which calls of_get_property() which calls
of_find_property() which tries to re-acquire devtree_lock, thus causing
deadlock.

To avoid this, create a new __of_device_is_available() which calls
__of_get_property() instead, which calls __of_find_property(), which
does not take the lock,. Update of_get_next_available_child() to call
the new __of_device_is_available() since it already owns the lock.
Signed-off-by: NStephen Warren <swarren@nvidia.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

With the locking cleanup in place (from "OF: Fixup resursive
locking code paths"), we can now do the conversion from the
rw_lock to a raw spinlock as required for preempt-rt.

The previous cleanup and this conversion were originally
separate since they predated when mainline got raw spinlock (in
commit c2f21ce2 "locking: Implement new raw_spinlock").

So, at that point in time, the cleanup was considered plausible
for mainline, but not this conversion.  In any case, we've kept
them separate as it makes for easier review and better bisection.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
[PG: taken from preempt-rt, update subject & add a commit log]
Signed-off-by: NPaul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: NRob Herring <rob.herring@calxeda.com>

In some situations, userspace may want to resolve a
device by function and logical number (ie, "serial0")
rather than by the base address or full device path. Being
able to resolve a device by alias frees userspace from the
burden of otherwise having to maintain a mapping between
device addresses and their logical assignments on each
platform when multiple instances of the same hardware block
are present in the system.

Although the uevent device attribute contains devicetree
compatible information and the full device path, the uevent
does not list the alises that may have been defined for the
device.
Signed-off-by: NStepan Moskovchenko <stepanm@codeaurora.org>
[grant.likely: Removed OF_ALIAS_N field; I don't think it's needed]
[grant.likely: Added #ifndef _LINUX_OF_PRIVATE_H wrapper]
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

There is no real reason to use a rwlock for devtree_lock. It even
could be a mutex, but unfortunately it's locked from cpu hotplug
paths which can't schedule :(

So it needs to become a raw lock on rt as well.  The devtree_lock would
be the only user of a raw_rw_lock, so we are better off cleaning up the
recursive locking paths which allows us to convert devtree_lock to a
read_lock.

Here we do the standard thing of introducing __foo() as the "raw"
version of foo(), so that we can take better control of the locking.
The "raw" versions are not exported and are for internal use within
the file itself.
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Signed-off-by: NPaul Gortmaker <paul.gortmaker@windriver.com>
Signed-off-by: NRob Herring <rob.herring@calxeda.com>

The of_find_matching_node_and_match() function incorrectly sets the matched
entry to 'matches' when the compatible value of a node matches one of the
possible values. This results in incorrectly selecting the the first entry in
the 'matches' list as the matched entry. Fix this by noting down the result of
the call to of_match_node() and setting that as the matched entry.
Signed-off-by: NThomas Abraham <thomas.abraham@linaro.org>
Signed-off-by: NRob Herring <rob.herring@calxeda.com>