Already existing property flags are filled wrong for properties created from
initial FDT. This could cause problems if this DYNAMIC device-tree functions
are used later, i.e. properties are attached/detached/replaced. Simply dumping
flags from the running system show, that some initial static (not allocated via
kzmalloc()) nodes are marked as dynamic.

I putted some debug extensions to property_proc_show(..) :
..
+       if (OF_IS_DYNAMIC(pp))
+               pr_err("DEBUG: xxx : OF_IS_DYNAMIC\n");
+       if (OF_IS_DETACHED(pp))
+               pr_err("DEBUG: xxx : OF_IS_DETACHED\n");

when you operate on the nodes (e.g.: ~$ cat /proc/device-tree/*some_node*) you
will see that those flags are filled wrong, basically in most cases it will dump
a DYNAMIC or DETACHED status, which is in not true.
(BTW. this OF_IS_DETACHED is a own define for debug purposes which which just
make a test_bit(OF_DETACHED, &x->_flags)

If nodes are dynamic kernel is allowed to kfree() them. But it will crash
attempting to do so on the nodes from FDT -- they are not allocated via
kzmalloc().
Signed-off-by: NWladislav Wiebe <wladislav.kw@gmail.com>
Acked-by: NAlexander Sverdlin <alexander.sverdlin@nsn.com>
Cc: stable@vger.kernel.org
Signed-off-by: NRob Herring <rob.herring@calxeda.com>

On some PAE architectures, the entire range of physical memory could reside
outside the 32-bit limit.  These systems need the ability to specify the
initrd location using 64-bit numbers.

This patch globally modifies the early_init_dt_setup_initrd_arch() function to
use 64-bit numbers instead of the current unsigned long.

There has been quite a bit of debate about whether to use u64 or phys_addr_t.
It was concluded to stick to u64 to be consistent with rest of the device
tree code. As summarized by Geert, "The address to load the initrd is decided
by the bootloader/user and set at that point later in time. The dtb should not
be tied to the kernel you are booting"

More details on the discussion can be found here:
https://lkml.org/lkml/2013/6/20/690
https://lkml.org/lkml/2012/9/13/544Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NRob Herring <rob.herring@calxeda.com>
Acked-by: NVineet Gupta <vgupta@synopsys.com>
Acked-by: NJean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

Fix "/aliasas" typo in comments, no functional change.
Signed-off-by: NRobert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

Signed-off-by: NAndy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: NGrant Likely <grant.likely@secretlab.ca>
Cc: Rob Herring <rob.herring@calxeda.com>
Cc: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

ERROR: "allnodes" [drivers/w1/masters/w1-gpio.ko] undefined!
Signed-off-by: NRandy Dunlap <rdunlap@infradead.org>
[grant.likely: allnodes is too generic; rename to of_allnodes]
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>
Cc: Ville Syrjala <syrjala@sci.fi>

Commit 509b7455 (of/fdt: Don't copy garbage after "/" in root node path)
sets the path length to 0 to ignore any garbage after "/" in the root
node path. This has the side effect of also ignoring '\0' at the end of
the root node path. This patch sets the ignores the garbage by setting
the last character to '\0' and length to 1.
Signed-off-by: NCatalin Marinas <catalin.marinas@arm.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NRob Herring <rob.herring@calxeda.com>

Constify 'pathp' in order to get rid of the following warning:

drivers/of/fdt.c:491:10: warning: assignment discards 'const' qualifier from pointer target type [enabled by default]
Signed-off-by: NFabio Estevam <fabio.estevam@freescale.com>
[Rob Herring: also constify np and lp]
Signed-off-by: NRob Herring <rob.herring@calxeda.com>

Neither of these should ever be changed once set. Make them const and
fix up the users that try to modify it in-place. In one case
kmalloc+memcpy is replaced with kstrdup() to avoid modifying the string.

Build tested with defconfigs on ARM, PowerPC, Sparc, MIPS, x86 among
others.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>
Acked-by: NDavid S. Miller <davem@davemloft.net>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Julian Calaby <julian.calaby@gmail.com>

The root node path must be internally converted to "/", or various
pieces of code looking for it that way will fail. The code to do
that however had a bug where we might incorrectly append pieces
of the original path from the fdt to the "/".

We should probably add a proper dedicated accessor for the root node
but in the meantime this patch should fix it.
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NRob Herring <rob.herring@calxeda.com>

drivers/of/fdt.c included 'asm/setup.h' twice, remove the duplicate.
Signed-off-by: NDanny Kukawka <danny.kukawka@bisect.de>
Signed-off-by: NRob Herring <rob.herring@calxeda.com>

In the file drivers/of/fdt.c, it uses the COMMAND_LINE_SIZE which is stated
in asm/setup.h, so asm/setup.h should be included in drivers/of/fdt.c.
Signed-off-by: NYu Yue <yuyue@mprc.pku.edu.cn>
Signed-off-by: NGuan Xuetao <guanxuetao@mprc.pku.edu.cn>

Cc: Grant Likerly <grant.likely@secretlab.ca>
Cc: devicetree-discuss@lists.ozlabs.org
Cc: Arnd Bergmann <arnd@arndb.de>

Fix the following build error:

  CC [M]  fs/udf/balloc.o
In file included from /home/fabio/next/linux-next/arch/arm/include/asm/prom.h:16,
                 from include/linux/of.h:140,
                 from include/asm-generic/gpio.h:7,
                 from arch/arm/plat-mxc/include/mach/irqs.h:14,
                 from /home/fabio/next/linux-next/arch/arm/include/asm/irq.h:4,
                 from /home/fabio/next/linux-next/arch/arm/include/asm/hardirq.h:6,
                 from include/linux/hardirq.h:7,
                 from include/linux/highmem.h:8,
                 from include/linux/pagemap.h:10,
                 from include/linux/buffer_head.h:13,
                 from fs/udf/udfdecl.h:11,
                 from fs/udf/balloc.c:22:
/home/fabio/next/linux-next/arch/arm/include/asm/setup.h:146: error: redefinition of 'struct tag'
Signed-off-by: NFabio Estevam <fabio.estevam@freescale.com>
[grant.likely: fix build failure on drivers/of/fdt.c]
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

This allows dt_compat to point to a constant list of compatible strings.
Signed-off-by: NUwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: NRob Herring <rob.herring@calxeda.com>

The patch adds function of_alias_scan to populate a global lookup
table with the properties of 'aliases' node and function
of_alias_get_id for drivers to find alias id from the lookup table.

v3: Split out automatic addition of aliases on id lookup so that it can be
    debated separately from the core functionality.
v2: - Add of_chosen/of_aliases populating and of_alias_scan() invocation
    for OF_PROMTREE.
    - Add locking
    - rework parse loop
Signed-off-by: NShawn Guo <shawn.guo@linaro.org>
Acked-by: NDavid S. Miller <davem@davemloft.net>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

We used to overwrite with CONFIG_CMDLINE if we found a chosen
node but failed to get bootargs out of it or they were empty,
unless CONFIG_CMDLINE_FORCE is set.

Instead change that to overwrite if "data" is non empty after
the bootargs check. It allows arch code to have other mechanisms
to retrieve the command line prior to parsing the device-tree.

Note: CONFIG_CMDLINE_FORCE case should ideally be handled elsewhere
as it won't work as it-is if the device-tree has no /chosen node
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>
CC: devicetree-discuss@lists-ozlabs.org
CC: Grant Likely <grant.likely@secretlab.ca>
Acked-by: NGrant Likely <grant.likely@secretlab.ca>

This reverts commit 750f463a.

of_alias_* still needs work to be generalized for 'promtree' dt
platforms, and to no implicitly create entries for available ids.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

The patch adds function of_alias_scan to populate a global lookup
table with the properties of 'aliases' node and function
of_alias_get_id for drivers to find alias id from the lookup table.
Signed-off-by: NShawn Guo <shawn.guo@linaro.org>
[grant.likely: add locking and rework parse loop]
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

This patch drops the reference to a global 'cmd_line' variable from
early_init_dt_scan_chosen, and instead passes the pointer to the command
line string via the *data argument.  Each architecture does something
slightly different with the initial command line, so it makes sense for
the architecture to be able to specify the variable name.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Fixes generated by 'codespell' and manually reviewed.
Signed-off-by: NLucas De Marchi <lucas.demarchi@profusion.mobi>

 - static-ize some functions
 - add some additional comments
Signed-off-by: NAndres Salomon <dilinger@queued.net>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

The physical address is never used by the device tree code when
allocating memory for unflattening.  Change the architecture's alloc
hook to return the virutal address instead.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

This patch adds of_flat_dt_match() which tests a node for
compatibility with a list of values and converts the relevant powerpc
platform code to use it.  This approach simplifies the board support
code a bit.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>
Reviewed-by: NStephen Neuendorffer <stephen.neuendorffer@xilinx.com>

unflatten_device_tree has two dependencies on things that happen
during boot time.  Firstly, it references the initial device tree
directly. Secondly, it allocates memory using the early boot
allocator.  This patch factors out these dependencies and uses
the new __unflatten_device_tree function to implement a driver-visible
fdt_unflatten_tree function, which can be used to unflatten a
blob after boot time.

V2:
- remove extra __va() call
- make dt_alloc functions return void *.  This doesn't fix the general
  strangeness in this code that constantly casts back and forth between
  unsigned long and __be32 *
Signed-off-by: NStephen Neuendorffer <stephen.neuendorffer@xilinx.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Move unflatten_dt_node to be grouped with non-__init functions.
Signed-off-by: NStephen Neuendorffer <stephen.neuendorffer@xilinx.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

unflatten_dt_node is a helper function that does most of the work to
convert a device tree blob into tree of device nodes.  This code
now uses a passed-in blob instead of using the single boot-time blob,
allowing it to be called in more contexts.
Signed-off-by: NStephen Neuendorffer <stephen.neuendorffer@xilinx.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

In preparation for providing run-time handling of device trees, factor
out some of the basic functions so that they take an arbitrary blob,
rather than relying on the single boot-time tree.

V2:
- functions have of_fdt_* names
- removed find_flat_dt_string
- blob argument is first
Signed-off-by: NStephen Neuendorffer <stephen.neuendorffer@xilinx.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

The device tree code is now in two pieces: some which can be used generically
on any platform which selects CONFIG_OF_FLATTREE, and some early which is used
at boot time on only a few architectures.  This patch segregates the early
code so that only those architectures which care about it need compile it.
This also means that some of the requirements in the early code (such as
a cmd_line variable) that most architectures (e.g. X86) don't provide
can be ignored.
Signed-off-by: NStephen Neuendorffer <stephen.neuendorffer@xilinx.com>
[grant.likely@secretlab.ca: remove extra blank line addition]
[grant.likely@secretlab.ca: fixed incorrect #ifdef CONFIG_EARLY_FLATTREE check]
[grant.likely@secretlab.ca: Made OF_EARLY_FLATTREE select instead of depend
                            on OF_FLATTREE]
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

This patch refactors the early init parsing of the chosen node so that
architectures aren't forced to provide an empty implementation of
early_init_dt_scan_chosen_arch.  Instead, if an architecture wants to
do something different, it can either use a wrapper function around
early_init_dt_scan_chosen(), or it can replace it altogether.

This patch was written in preparation to adding device tree support to
both x86 ad MIPS.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>
Tested-by: NDavid Daney <ddaney@caviumnetworks.com>

The flat tree code wasn't fixing the endianness on phandle values when
unflattening the tree, and the code in drivers/of wasn't always doing a
be32_to_cpu before trying to dereference the phandle values.  This patch
fixes them.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

With the current string comparison, a device tree compatible of "foo-bar"
would match as compatible with a driver looking for "foo".  This patch
fixes the function to use the of_compat_cmp() macro so that it does the
right thing on all platforms (If sparc ever uses this code, it will still
want the strncasecmp() behaviour).
Signed-off-by: NStuart Yoder <stuart.yoder@freescale.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

There's no reason to use the powerpc-specific _ALIGN macro in the fdt
code.  Replace it with ALIGN() from kernel.h
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>
Acked-By: NJeremy Kerr <jeremy.kerr@canonical.com>
Acked-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

This patch makes unflatten_device_tree() safe to call from any arch
setup code with the following changes:
- Make sure initial_boot_params actually points to a device tree blob
  before unflattening
- Make sure the initial_boot_params->magic field is correct
- If CONFIG_OF_FLATTREE is not set, then make unflatten_device_tree()
  an empty static inline function.

This patch also adds some additional debug output to the top of
unflatten_device_tree().
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

NOPs within the property section are skipped, but NOPs between
OF_DT_END_NODE and OF_DT_BEGIN_NODE were not. My firmware NOPs out
entire nodes depending on various environment parameters.

of_scan_flat_dt already handles NOP more generally.
Signed-off-by: NJason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

We don't always have lmb available, so make arches provide an
early_init_dt_alloc_memory_arch() to handle the allocation of
memory in the fdt code.

When we don't have lmb.h included, we need asm/page.h for __va.
Signed-off-by: NJeremy Kerr <jeremy.kerr@canonical.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>
Acked-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: NMichal Simek <monstr@monstr.eu>

Currently when processing flattened device trees, the kernel expects
the phandle in a property called "linux,phandle".  The ePAPR spec -
not being Linux specific - instead requires phandles to be encoded in
a property named simply "phandle".  This patch makes the kernel accept
either form when unflattening the device tree.
Signed-off-by: NDavid Gibson <david@gibson.dropbear.id.au>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

The boot_param_header has big-endian fields, so change the types to
__be32, and perform endian conversion when we access them.
Signed-off-by: NJeremy Kerr <jeremy.kerr@canonical.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Properties in the device tree are specified as big-endian. At present,
the only platforms to support device trees are also big-endian, so we've
been acessing the properties as raw values.

We'd like to add device tree support to little-endian platforms too, so
add endian conversion to the sites where we access property values in
the common of code.

Compiled on powerpc (ppc44x_defconfig & ppc64_defconfig) and arm (fdt
support only for now).
Signed-off-by: NJeremy Kerr <jeremy.kerr@canonical.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Currently, we're using u32 for cell values, and hence assuming
host-endian device trees.

As we'd like to support little-endian platforms, use a __be32 for cell
values, and convert in the cell accessors.
Signed-off-by: NJeremy Kerr <jeremy.kerr@canonical.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

At present we're using hard-coded values for defaults when parsing the
FDT. This change uses the #defines instead.
Signed-off-by: NJeremy Kerr <jeremy.kerr@canonical.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

At present, the fdt code sets the kernel-wide initrd_start and
initrd_end variables when parsing /chosen. On ARM, we only set these
once the bootmem has been reserved.

This change adds an arch hook to setup the initrd from the device
tree:

 void early_init_dt_setup_initrd_arch(unsigned long start,
				      unsigned long end);

The arch-specific code can then setup the initrd however it likes.

Compiled on powerpc, with CONFIG_BLK_DEV_INITRD=y and =n.
Signed-off-by: NJeremy Kerr <jeremy.kerr@canonical.com>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>