Having device_nodes be kobjects is only needed if sysfs or OF_DYNAMIC is
enabled. Otherwise, having a kobject in struct device_node is
unnecessary bloat in minimal kernel configurations.

Likewise, bin_attribute is only needed in struct property when sysfs is
enabled, so we can make it configurable too.
Tested-by: NNicolas Pitre <nico@linaro.org>
Reviewed-by: NFrank Rowand <frowand.list@gmail.com>
Acked-by: NGrant Likely <grant.likely@secretlab.ca>
Signed-off-by: NRob Herring <robh@kernel.org>

The only user of kobj_to_device_node() is in dynamic.c, so move it
there. This avoids having to make it conditional once kobject is
configurable.
Tested-by: NNicolas Pitre <nico@linaro.org>
Reviewed-by: NFrank Rowand <frowand.list@gmail.com>
Acked-by: NGrant Likely <grant.likely@secretlab.ca>
Signed-off-by: NRob Herring <robh@kernel.org>

The ref counting is broken for OF_DYNAMIC when sysfs is disabled because
the kobject initialization is skipped. Only the properties
add/remove/update should be skipped for !SYSFS config.
Tested-by: NNicolas Pitre <nico@linaro.org>
Reviewed-by: NFrank Rowand <frowand.list@gmail.com>
Acked-by: NGrant Likely <grant.likely@secretlab.ca>
Signed-off-by: NRob Herring <robh@kernel.org>

With dependencies on a statically allocated full path name converted to
use %pOF format specifier, we can store just the basename of node, and
the unflattening of the FDT can be simplified.

This commit will affect the remaining users of full_name. After
analyzing these users, the remaining cases should only change some print
messages. The main users of full_name are providing a name for struct
resource. The resource names shouldn't be important other than providing
/proc/iomem names.

We no longer distinguish between pre and post 0x10 dtb formats as either
a full path or basename will work. However, less than 0x10 formats have
been broken since the conversion to use libfdt (and no one has cared).
The conversion of the unflattening code to be non-recursive also broke
pre 0x10 formats as the populate_node function would return 0 in that
case.
Signed-off-by: NRob Herring <robh@kernel.org>

Tweak a small number of status "ok" lines in a single file under
/drivers/of/ to use the proper DTSpec 0.1 spelling of "okay".
Signed-off-by: NRobert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: NRob Herring <robh@kernel.org>

Moving DMA configuration to happen later at driver probe time had the
unnoticed side-effect that we now perform DMA configuration for *every*
device represented in DT, rather than only those explicitly created by
the of_platform and PCI code.

As Christoph points out, this is not really the best thing to do. Whilst
there may well be other DMA-capable buses that can benefit from having
their children automatically configured after the bridge has probed,
there are also plenty of others like USB, MDIO, etc. that definitely do
not support DMA and should not be indiscriminately processed.

The good news is that in most cases the DT "dma-ranges" property serves
as an appropriate indicator - per a strict interpretation of the spec,
anything lacking a "dma-ranges" property should be considered not to
have a mapping of DMA address space from its children to its parent,
thus anything for which of_dma_get_range() does not succeed does not
need DMA configuration. Certain bus types have a general expectation of
DMA capability and carry a well-established precedent that an absent
"dma-ranges" implies the same as the empty property, so we automatically
opt those in to DMA configuration regardless, to avoid regressing most
existing platforms.

Fixes: 09515ef5 ("of/acpi: Configure dma operations at probe time for platform/amba/pci bus devices")
Reported-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NRobin Murphy <robin.murphy@arm.com>
Acked-by: NRob Herring <robh@kernel.org>
Signed-off-by: NChristoph Hellwig <hch@lst.de>

Use dedicated definition instead of plain -1 where it's appropriate.

No functional change intended.
Signed-off-by: NAndy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: NRob Herring <robh@kernel.org>

of_device_request_module() calls of_device_get_modalias() with "len" 0,
to calculate the size of the buffer needed to store the result, but due
to integer promotion the ssize_t "len" will be compared as unsigned with
strlen(compat) and the loop will generally never break. This results in
a call to snprintf() with a negative len, which triggers below warning,
followed by a dereference of a invalid pointer:

  [    3.060067] WARNING: CPU: 0 PID: 51 at lib/vsprintf.c:2122 vsnprintf+0x348/0x6d8
  ...
  [    3.060301] [<ffffff800891ede8>] vsnprintf+0x348/0x6d8
  [    3.060308] [<ffffff800891f248>] snprintf+0x48/0x50
  [    3.060316] [<ffffff80086a7c80>] of_device_get_modalias+0x108/0x160
  [    3.060322] [<ffffff80086a7cf8>] of_device_request_module+0x20/0x88
  ...

Further more of_device_get_modalias() is supposed to return the number
of bytes needed to store the entire modalias, so the loop needs to
continue accumulate the total size even though the buffer is full.

Finally the function is not expected to ensure space for the NUL, nor
include it in the returned size, so only 1 should be added to the length
of "compat" in the loop (to account for the character 'C').

Fixes: bc575064 ("of/device: use of_property_for_each_string to parse compatible strings")
Cc: Rob Herring <robh@kernel.org>
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: NRob Herring <robh@kernel.org>

As of_device_get_modalias() returns the number of bytes that would have
been written to the target string, regardless of how much did fit in the
buffer, it's possible that the returned index points beyond the buffer
passed to of_device_modalias() - causing memory beyond the buffer to be
null terminated.

Fixes: 0634c295 ("of: Add function for generating a DT modalias with a newline")
Cc: Rob Herring <robh@kernel.org>
Cc: stable@vger.kernel.org
Signed-off-by: NBjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: NRob Herring <robh@kernel.org>

Historically, DMA masks have suffered some ambiguity between whether
they represent the range of physical memory a device can access, or the
address bits a device is capable of driving, particularly since on many
platforms the two are equivalent. Whilst there are some stragglers left
(dma_max_pfn(), I'm looking at you...), the majority of DMA code has
been cleaned up to follow the latter definition, not least since it is
the only one which makes sense once IOMMUs are involved.

In this respect, of_dma_configure() has always done the wrong thing in
how it generates initial masks based on "dma-ranges". Although rounding
down did not affect the TI Keystone platform where dma_addr + size is
already a power of two, in any other case it results in a mask which is
at best unnecessarily constrained and at worst unusable.

BCM2837 illustrates the problem nicely, where we have a DMA base of 3GB
and a size of 1GB - 16MB, giving dma_addr + size = 0xff000000 and a
resultant mask of 0x7fffffff, which is then insufficient to even cover
the necessary offset, effectively making all DMA addresses out-of-range.
This has been hidden until now (mostly because we don't yet prevent
drivers from simply overwriting this initial mask later upon probe), but
due to recent changes elsewhere now shows up as USB being broken on
Raspberry Pi 3.

Make it right by rounding up instead of down, such that the mask
correctly correctly describes all possisble bits the device needs to
emit.

Fixes: 9a6d7298 ("of: Calculate device DMA masks based on DT dma-range size")
Reported-by: NStefan Wahren <stefan.wahren@i2se.com>
Reported-by: NAndreas Färber <afaerber@suse.de>
Reported-by: NHans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: NRobin Murphy <robin.murphy@arm.com>
Acked-by: NRob Herring <robh@kernel.org>
Signed-off-by: NChristoph Hellwig <hch@lst.de>

Now, we can use of_graph_get_remote_endpoint(). Let's use it.
Signed-off-by: NKuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: NRob Herring <robh@kernel.org>

Looks  like gcc isn't always able to figure  out that 3 *if* branches in
of_phy_register_fixed_link() calling fixed_phy_register() at their ends
are similar enough and thus can be merged. The "manual" merge saves 40
bytes of the object code (AArch64 gcc 4.8.5), and still saves 12 bytes
even  if gcc was able to merge the branch tails (ARM gcc 4.8.5)...
Signed-off-by: NSergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The "fixed-link" prop support predated of_property_read_u32_array(), so
basically had to open-code it. Using the modern API saves 24 bytes of the
object code (ARM gcc 4.8.5); the only behavior change would be that the
prop length check is now less strict (however the strict pre-check done
in of_phy_is_fixed_link() is left intact anyway)...
Signed-off-by: NSergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Reviewed-by: NAndrew Lunn <andrew@lunn.ch>
Reviewed-by: NRob Herring <robh@kernel.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Convert the interrupts property parsing to use the OF property API
instead of open coding the parsing of the raw property value. This saves
a number of LoC, and the result is easier to read.
Signed-off-by: NRob Herring <robh@kernel.org>

Instead of directly parsing the compatible property, use the
of_property_for_each_string() helper to iterate over each compatible
string. This reduces the LoC and makes the functions easier to
understand.
Signed-off-by: NRob Herring <robh@kernel.org>

Fix inconsistent use of of_graph_get_port_parent() where
asoc_simple_card_parse_graph_dai() does of_node_get() before
calling it while other callers do not. We can fix this by
not trashing the node passed to of_graph_get_port_parent().

Let's also make sure the callers have correct refcounts and remove
related incorrect of_node_put() calls for of_for_each_phandle
as that's done by of_phandle_iterator_next() except when
we break out of the loop early.

Let's fix both issues with a single patch to avoid kobject
refcounts getting messed up more if two patches are merged
separately.

Otherwise strange issues can happen caused by memory corruption
caused by too many kobject_del() calls such as:

BUG: sleeping function called from invalid context at
kernel/locking/mutex.c:747
...
(___might_sleep)
(__mutex_lock)
(mutex_lock_nested)
(kernfs_remove)
(kobject_del)
(kobject_put)
(of_get_next_parent)
(of_graph_get_port_parent)
(asoc_simple_card_parse_graph_dai [snd_soc_simple_card_utils])
(asoc_graph_card_probe [snd_soc_audio_graph_card])

Fixes: 0ef472a9 ("of_graph: add of_graph_get_port_parent()")
Fixes: 2692c1c6 ("ASoC: add audio-graph-card support")
Fixes: 1689333f ("ASoC: simple-card-utils: add asoc_simple_card_parse_graph_dai()")
Signed-off-by: NTony Lindgren <tony@atomide.com>
Reviewed-by: NRob Herring <robh@kernel.org>
Tested-by: NAntonio Borneo <borneo.antonio@gmail.com>
Tested-by: NKuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: NMark Brown <broonie@kernel.org>

of_phy_register_fixed_link() declares the 'err' variable to hold the result
of of_property_read_string() but only uses it once after that, while that
function can be called directly from the *if* statement...

Remove that variable and move/regroup 'link_gpio' and 'len' variables in
order to sort the declarations in the reverse Xmas tree order -- to please
DaveM. ;-)
Signed-off-by: NSergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

of_pci_get_devfn() and of_pci_parse_bus_range() somehow didn't use
of_property_read_u32_array() though it was long available, basically
open-coding it.  Using the modern DT API saves several bytes and
also adds some prop sanity checks as a bonus...
Signed-off-by: NSergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: NRob Herring <robh@kernel.org>

of_get_pci_domain_nr() somehow didn't use of_property_read_u32() though it
was long available, basically open-coding it. Using the modern DT API saves
several LoCs/bytes and also adds some prop sanity checks as a bonus...
Signed-off-by: NSergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: NRob Herring <robh@kernel.org>

of_alias_scan() can use of_property_read_string() -- using the modern DT API
adds some prop sanity checks as a bonus; it does add couple LoCs but only
because the original code violated the 80-column limit...
Signed-off-by: NSergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: NRob Herring <robh@kernel.org>

of_n_{addr|size}_cells() predate of_property_read_u32(), so they have to
basically open-code it. Using the modern DT API saves several LoCs and also
adds some prop sanity checks as a bonus...
Signed-off-by: NSergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: NRob Herring <robh@kernel.org>

The new fwnode_property_get_reference_args() interface amends the fwnode
property API with the functionality of both of_parse_phandle_with_args()
and __acpi_node_get_property_reference().

The semantics is slightly different: the cells property is ignored on ACPI
as the number of arguments can be explicitly obtained from the firmware
interface.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Make fwnode arguments to the fwnode property API const.
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NRob Herring <robh@kernel.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Instead of relying on the struct fwnode_handle type field, define
fwnode_operations structs for all separate types of fwnodes. To find out
the type, compare to the ops field to relevant ops structs.

This change has two benefits:

1. it avoids adding the type field to each and every instance of struct
fwnode_handle, thus saving memory and

2. makes the ops field the single factor that defines both the types of
the fwnode as well as defines the implementation of its operations,
decreasing the possibility of bugs when developing code dealing with
fwnode internals.
Suggested-by: NRob Herring <robh@kernel.org>
Signed-off-by: NSakari Ailus <sakari.ailus@linux.intel.com>
Reviewed-by: NMika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

of_irq_to_resource() has recently been fixed to return negative error #'s
along with 0, however of_irq_to_resource_table() still only regards 0 as
invalid IRQ -- fix it up.

Fixes: 7a4228bb ("of: irq: use of_irq_get() in of_irq_to_resource()")
Signed-off-by: NSergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: NRob Herring <robh@kernel.org>

It was never used and could be removed, otherwise
we could see a warning:

drivers/of/address.c: In function 'of_pci_range_parser_one':
drivers/of/address.c:277:14: warning: variable 'pci_space' set but not
used [-Wunused-but-set-variable]
Signed-off-by: NShawn Lin <shawn.lin@rock-chips.com>
Signed-off-by: NRob Herring <robh@kernel.org>

The DT interrupt parsing code predates of_property_read_u32(), so it has to
basically open-code it. Using the modern DT API saves several LoCs and also
adds some prop sanity checks as a bonus...
Signed-off-by: NSergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: NRob Herring <robh@kernel.org>

The "interrupt-controller" property is boolean, i.e. has no value.  The DT
interrupt parsing code predates of_property_read_bool(), so it uses either
of_get_property() or of_find_property() -- the former isn't quite correct
for the boolean props (but works somehow). Use the modern boolean prop API
instead.
Signed-off-by: NSergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: NRob Herring <robh@kernel.org>

Add overlay __symbols__ properties to live tree when an overlay
is added to the live tree so that the symbols are available to
subsequent overlays.

Expected test result is new __symbols__ entries for labels from
the overlay after this commit.

Before this commit:

   Console error message near end of unittest:
      ### dt-test ### FAIL of_unittest_overlay_high_level():2296 Adding overlay 'overlay_bad_symbol' failed
      ### dt-test ### end of unittest - 190 passed, 1 failed

   The new unittest "fails" because the expected result of loading the
   new overlay is an error instead of success.

   $ # node hvac-medium-2 exists because the overlay loaded
   $ # since the duplicate symbol was not detected
   $ cd /proc/device-tree/testcase-data-2/substation@100/
   $ ls
   compatible     hvac-medium-2  motor-8        reg
   hvac-large-1   linux,phandle  name           status
   hvac-medium-1  motor-1        phandle

   $ cd /proc/device-tree/__symbols__/
   $ ls
   electric_1   lights_1     name         rides_1      spin_ctrl_2
   hvac_1       lights_2     retail_1     spin_ctrl_1

After this commit:

   Previous console error message no longer occurs, but expected error
   occurs:
      OF: overlay: Failed to apply prop @/__symbols__/hvac_1
      OF: overlay: apply failed '/__symbols__'
      ### dt-test ### end of unittest - 191 passed, 0 failed

   $ # node hvac-medium-2 does not exist because the overlay
   $ # properly failed to load due to the duplicate symbol
   $ cd /proc/device-tree/testcase-data-2/substation@100/
   $ ls
   compatible     hvac-medium-1  motor-1        name           reg
   hvac-large-1   linux,phandle  motor-8        phandle        status

   $ cd /proc/device-tree/__symbols__/
   $ ls
   electric_1      lights_1        retail_1        ride_200_right  spin_ctrl_2
   hvac_1          lights_2        ride_200        rides_1
   hvac_2          name            ride_200_left   spin_ctrl_1
   $ cat ride_200; echo
   /testcase-data-2/fairway-1/ride@200
   $ cat ride_200_left ; echo
   /testcase-data-2/fairway-1/ride@200/track@10
   $ cat ride_200_right ; echo
   /testcase-data-2/fairway-1/ride@200/track@20
Signed-off-by: NFrank Rowand <frank.rowand@sony.com>
Signed-off-by: NRob Herring <robh@kernel.org>

Correct existing node name detection when overlay node name has
a unit-address.

Expected test result is overlay will update the nodes and properties
for /testcase-data-2/fairway-1/ride@100/ after this commit.

Before this commit:

   Console error message near end of unittest:
      OF: Duplicate name in fairway-1, renamed to "ride@100#1"

   $ cd /proc/device-tree/testcase-data-2/fairway-1/
   $ # extra node: ride@100#1
   $ ls
   #address-cells  linux,phandle   phandle         ride@200
   #size-cells     name            ride@100        status
   compatible      orientation     ride@100#1
   $ cd /proc/device-tree/testcase-data-2/fairway-1/ride@100/
   $ ls track@30/incline-up
   ls: track@30/incline-up: No such file or directory
   $ ls track@40/incline-up
   ls: track@40/incline-up: No such file or directory

After this commit:

   Console error message no longer occurs

   $ cd /proc/device-tree/testcase-data-2/fairway-1/
   $ # no extra node: ride@100#1
   $ ls
   #address-cells  compatible      name            phandle         ride@200
   #size-cells     linux,phandle   orientation     ride@100        status
   $ cd /proc/device-tree/testcase-data-2/fairway-1/ride@100/
   $ ls track@30/incline-up
   track@30/incline-up
   $ ls track@40/incline-up
   track@40/incline-up
Signed-off-by: NFrank Rowand <frank.rowand@sony.com>
Signed-off-by: NRob Herring <robh@kernel.org>

Add nodes and properties to overlay_base and overlay dts files to
test for
   - incorrect existing node name detection when overlay node name
     has a unit-address
   - adding overlay __symbols__ properties to live tree when an
     overlay is added to the live tree

The following console messages will appear near the end of unittest
until the code errors are corrected:

   OF: Duplicate name in fairway-1, renamed to "ride@100#1"

   ### dt-test ### FAIL of_unittest_overlay_high_level():2296 Adding overlay 'overlay_bad_symbol' failed

   ### dt-test ### end of unittest - 190 passed, 1 failed
Signed-off-by: NFrank Rowand <frank.rowand@sony.com>
Signed-off-by: NRob Herring <robh@kernel.org>

Now that we have a custom printf format specifier, convert users of
full_name to use %pOF instead. This is preparation to remove storing
of the full path string for each node.
Signed-off-by: NRob Herring <robh@kernel.org>

With the introduction of struct pci_host_bridge.(*map_irq) function pointer
add another direct usage of of_irq_parse_and_map_pci().

Update the function comment to reflect this behaviour.
Signed-off-by: NMatthew Minter <matt@masarand.com>
Signed-off-by: NLorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>

90% of the usage of device node's full_name is printing it out in a
kernel message. However, storing the full path for every node is
wasteful and redundant. With a custom format specifier, we can generate
the full path at run-time and eventually remove the full path from every
node.

For instance typical use is:
	pr_info("Frobbing node %s\n", node->full_name);

Which can be written now as:
	pr_info("Frobbing node %pOF\n", node);

'%pO' is the base specifier to represent kobjects with '%pOF'
representing struct device_node. Currently, struct device_node is the
only supported type of kobject.

More fine-grained control of formatting includes printing the name,
flags, path-spec name and others, explained in the documentation entry.

Originally written by Pantelis, but pretty much rewrote the core
function using existing string/number functions. The 2 passes were
unnecessary and have been removed. Also, updated the checkpatch.pl
check. The unittest code was written by Grant Likely.
Signed-off-by: NPantelis Antoniou <pantelis.antoniou@konsulko.com>
Acked-by: NJoe Perches <joe@perches.com>
Signed-off-by: NRob Herring <robh@kernel.org>

find_node_by_full_name() does the same thing as of_find_node_by_path(),
but takes no locks and doesn't work on aliases. Refactor
of_find_node_opts_by_path() into __of_find_node_by_full_path() and
replace find_node_by_full_name() with it.
Signed-off-by: NRob Herring <robh@kernel.org>

Several places in DT code open code the equivalent of kbasename.
Replace them.

The behavior for root nodes in node_name_cmp will be slightly different.
Instead of comparing "/", "" will be compared. The comparison will be
the same.
Reviewed-by: NAndy Shevchenko <andy.shevhchenko@gmail.com>
Signed-off-by: NRob Herring <robh@kernel.org>

Overlays are not allowed to modify phandle values of previously existing
nodes because there is no information available to allow fixup up
properties that use the previously existing phandle.
Signed-off-by: NFrank Rowand <frank.rowand@sony.com>
Signed-off-by: NRob Herring <robh@kernel.org>

checkpatch whined about using S_IRUGO instead of octal equivalent
when adding phandle sysfs code, so used octal in that pending patch.
Change other instances of the S_* constants in the same file to
the octal form.
Signed-off-by: NFrank Rowand <frank.rowand@sony.com>
Signed-off-by: NRob Herring <robh@kernel.org>

__of_attach_node() is not used outside of drivers/of/dynamic.c.  Make
it static and remove it from drivers/of/of_private.h.
Signed-off-by: NFrank Rowand <frank.rowand@sony.com>
Signed-off-by: NRob Herring <robh@kernel.org>

Fix typo in header comment for of_address_to_resource()
Signed-off-by: NFrank Rowand <frank.rowand@sony.com>
Signed-off-by: NRob Herring <robh@kernel.org>