Currently we get the following kind of errors if we try to use interrupt
phandles to irqchips that have not yet initialized:

irq: no irq domain found for /ocp/pinmux@48002030 !
------------[ cut here ]------------
WARNING: CPU: 0 PID: 1 at drivers/of/platform.c:171 of_device_alloc+0x144/0x184()
Modules linked in:
CPU: 0 PID: 1 Comm: swapper/0 Not tainted 3.12.0-00038-g42a9708 #1012
(show_stack+0x14/0x1c)
(dump_stack+0x6c/0xa0)
(warn_slowpath_common+0x64/0x84)
(warn_slowpath_null+0x1c/0x24)
(of_device_alloc+0x144/0x184)
(of_platform_device_create_pdata+0x44/0x9c)
(of_platform_bus_create+0xd0/0x170)
(of_platform_bus_create+0x12c/0x170)
(of_platform_populate+0x60/0x98)

This is because we're wrongly trying to populate resources that are not
yet available. It's perfectly valid to create irqchips dynamically, so
let's fix up the issue by resolving the interrupt resources when
platform_get_irq is called.

And then we also need to accept the fact that some irqdomains do not
exist that early on, and only get initialized later on. So we can
make the current WARN_ON into just into a pr_debug().

We still attempt to populate irq resources when we create the devices.
This allows current drivers which don't use platform_get_irq to continue
to function. Once all drivers are fixed, this code can be removed.
Suggested-by: NRussell King <linux@arm.linux.org.uk>
Signed-off-by: NRob Herring <robh@kernel.org>
Signed-off-by: NTony Lindgren <tony@atomide.com>
Tested-by: NTony Lindgren <tony@atomide.com>
Cc: stable@vger.kernel.org # v3.10+
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

Platform devices created by DT code don't initialize dma_mask pointer to
anything. Set it to coherent_dma_mask by default if the architecture
code has not set it.
Signed-off-by: NRob Herring <rob.herring@calxeda.com>

AMBA Primecell devices always treat streaming and coherent DMA exactly
the same, so there's no point in having the masks separated.
Acked-by: NGrant Likely <grant.likely@linaro.org>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

This reverts commit 9d8eab7a. There is
still no consensus on the bindings for the reserved memory and various
drawbacks of the proposed solution has been shown, so the best now is to
revert it completely and start again from scratch later.
Signed-off-by: NMarek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

Add error reporting to of_amba_device_create() so the user knows
when (and why) some device tree nodes fail to initialize.

[ The issue was spotted on Universal C210 board (using revision 0 of
  ARM Exynos4210 SoC) on which initialization was silently failing
  for PL330 MDMA1 device tree node (it was using the wrong addres
  resulting in amba_device_add() returning -ENODEV). ]
Signed-off-by: NBartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Signed-off-by: NKyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

This patch adds device tree support for contiguous and reserved memory
regions defined in device tree.

Large memory blocks can be reliably reserved only during early boot.
This must happen before the whole memory management subsystem is
initialized, because we need to ensure that the given contiguous blocks
are not yet allocated by kernel. Also it must happen before kernel
mappings for the whole low memory are created, to ensure that there will
be no mappings (for reserved blocks) or mapping with special properties
can be created (for CMA blocks). This all happens before device tree
structures are unflattened, so we need to get reserved memory layout
directly from fdt.

Later, those reserved memory regions are assigned to devices on each
device structure initialization.
Signed-off-by: NMarek Szyprowski <m.szyprowski@samsung.com>
Acked-by: NKyungmin Park <kyungmin.park@samsung.com>
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Acked-by: NTomasz Figa <t.figa@samsung.com>
Acked-by: NStephen Warren <swarren@nvidia.com>
Reviewed-by: NRob Herring <rob.herring@calxeda.com>

It is not used outside of this file so doesn't need to be in the global
namespace.
Signed-off-by: NMark Brown <broonie@linaro.org>
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

This reverts commit aac73f34. That
commit causes two kinds of breakage; it breaks registration of AMBA
devices when one of the parent nodes already contains overlapping
resource regions, and it breaks calls to request_region() by device
drivers in certain conditions where there are overlapping memory
regions. Both of these problems can probably be fixed, but it is better
to back out the commit and get a proper fix designed before trying again.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

This allows platform_device_add a chance to call insert_resource on all
of the resources from OF. At a minimum this fills in proc/iomem and
presumably makes resource tracking and conflict detection work better.
However, it has the side effect of moving all OF generated platform
devices from /sys/devices to /sys/devices/platform/. It /shouldn't/
break userspace because userspace is not supposed to depend on the full
path (because userspace always does what it is supposed to, right?).

This may cause breakage if either:
1) any two nodes in a given device tree have overlapping & staggered
   regions (ie. 0x80..0xbf and 0xa0..0xdf; where one is not contained
   within the other). In this case one of the devices will fail to
   register and an exception will be needed in platform_device_add() to
   complain but not fail.
2) any device calls request_mem_region() on a region larger than
   specified in the device tree. In this case the device node may be
   wrong, or the driver is overreaching. In either case I'd like to know
   about any problems and fix them.

Please test. Despite the above, I'm still fairly confident that this
patch is in good shape. I'd like to put it into linux-next, but would
appreciate some bench testing from others before I do; particularly on
PowerPC machines.

v2: Remove powerpc special-case

Cc: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Rob Herring <rob.herring@calxeda.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

15c3597d (dt/platform: allow device name to be overridden) added a
lookup parameter to of_platform_populate() but did not update the
documentation.  This patch adds the missing documentation entry.

Cc: Grant Likely <grant.likely@secretlab.ca>
Cc: Jiri Kosina <trivial@kernel.org>
Signed-off-by: NJavi Merino <javi.merino@arm.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

drivers/of/platform.c:110:59: warning: incorrect type in argument 2 (different base types)
drivers/of/platform.c:110:59:    expected restricted __be32 const [usertype] *addr
drivers/of/platform.c:110:59:    got unsigned int const [usertype] *[assigned] reg
Signed-off-by: NKim Phillips <kim.phillips@freescale.com>
Signed-off-by: NRob Herring <rob.herring@calxeda.com>

It's quite legitimate for a DT node to specify #size-cells=0. One example
is a node that's used to collect a number of non-memory-mapped devices.
In that scenario, there may be multiple child nodes with the same name
(type) thus necessitating the use of unit addresses in node names, and
reg properties:

/ {
	regulators {
		compatible = "simple-bus";
		#address-cells = <1>;
		#size-cells = <0>;

		regulator@0 {
			compatible = "regulator-fixed";
			reg = <0>;
			...
		};

		regulator@1 {
			compatible = "regulator-fixed";
			reg = <1>;
			...
		};

		...
	};
};

However, #size-cells=0 prevents translation of reg property values into
the parent node's address space. In turn, this triggers the kernel to
emit error messages during boot, such as:

    prom_parse: Bad cell count for /regulators/regulator@0

To prevent printing these error messages for legitimate DT content, a
number of changes are made:

1) of_get_address()/of_get_pci_address() are modified only to validate
   the value of #address-cells, and not #size-cells.

2) of_can_translate_address() is added to indicate whether address
   translation is possible.

3) of_device_make_bus_id() is modified to name devices based on the
   translated address only where possible, and otherwise fall back to
   using the (first cell of the) raw untranslated address.

4) of_device_alloc() is modified to create memory resources for a device
   only if the address can be translated into the CPU's address space.
Signed-off-by: NStephen Warren <swarren@nvidia.com>
Signed-off-by: NRob Herring <rob.herring@calxeda.com>

Sometimes it doesn't make any sense for a node to have an address.
In this case device lookup will always be unsuccessful because we
currently assume every node will have a reg property. This patch
changes the semantics so that the resource address and the lookup
address will only be compared if one exists.

Things like AUXDATA() rely on of_dev_lookup to return the lookup
entry of a particular device in order to do things like apply
platform_data to a device. However, this is currently broken for
nodes which do not have a reg property, meaning that platform_data
can not be passed in those cases.
Acked-by: NRob Herring <rob.herring@calxeda.com>
Acked-by: NArnd Bergmann <arnd@arndb.de>
Signed-off-by: NLee Jones <lee.jones@linaro.org>

Without this, modules can't use this API, leading to build failures.
Signed-off-by: NStephen Warren <swarren@nvidia.com>
Signed-off-by: NRob Herring <rob.herring@calxeda.com>

Sparc has its own helpers for translating address ranges when the device
tree is parsed at boot time, and it isn't able to use of_platform_populate().
However, there are some device drivers that want to use that function on
other DT enabled platforms (ie. TWL4030).  This patch adds an empty
of_platform_populate() implementation that returns an error when
CONFIG_OF_ADDRESS is not selected.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>
Cc: Benoit Cousson <b-cousson@ti.com>
Cc: Samuel Ortiz <sameo@linux.intel.com>

Convert DT code to use the new amba_device_alloc APIs.
Acked-by: NRob Herring <rob.herring@calxeda.com>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

There is no requirement to override name entries in auxdata.  Fix the
entry matching to use .compatible instead of .name to find the end of the
list.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

* Correct description of of_platform_bus_create to match implementation
* Remove a level of indentation in of_dev_lookup
Signed-off-by: NOlof Johansson <olof@lixom.net>
Acked-by: NGrant Likely <grant.likely@secretlab.ca>
Signed-off-by: NRob Herring <rob.herring@calxeda.com>

We have a long standing issues with platform devices not have a valid
dma_mask pointer.  This hasn't been an issue to date as no platform
device has tried to set its dma_mask value to a non-default value.
Acked-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NKumar Gala <galak@kernel.crashing.org>

Some platform code has specific requirements on the naming of devices.
This patch allows callers of of_platform_populate() to provide a
device name lookup table.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Add a function to create amba_devices (i.e. primecell peripherals)
from device tree nodes. The device tree scanning is done by the
of_platform_populate() function which can call of_amba_device_create
based on a match table entry.

Nodes with a "arm,primecell-periphid" property can override the h/w
peripheral id value.

Based on the original work by Jeremy Kerr.
Signed-off-by: NJeremy Kerr <jeremy.kerr@canonical.com>
Acked-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NRob Herring <rob.herring@calxeda.com>
Reviewed-by: NArnd Bergmann <arnd@arndb.de>
[grant.likely: add Jeremy's original s-o-b line, changes from review
               comments, and moved all code to drivers/of/platform.c]
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

of_platform_populate() is similar to of_platform_bus_probe() except
that it strictly enforces that all device nodes must have a compatible
property, and it can be used to register devices (not buses) which are
children of the root node.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

No need for most platforms to define their own bus table when calling
of_platform_populate().  Supply a stock one.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

There are no users of OF_NO_DEEP_PROBE, and of_match_node() now
gracefully handles being passed a NULL pointer, so the checks at the
top of of_platform_bus_probe can be dropped.

While at it, consolidate the root node pointer check to be easier to
read and tidy up related comments.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

The current implementation uses three copies of of basically identical
code.  This patch consolidates them to make the code simpler.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Commit eca39301, "of: Merge of_platform_bus_type with
platform_bus_type" added a shim to allow of_platform_drivers to get
registers onto the platform bus so that there was time to migrate the
existing drivers to the platform_bus_type.

This patch removes the shim since there are no more users of the old
interface.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

arch/powerpc/kernel/ibmebus.c is the only remaining user of the
of_bus_type support code for initializing the bus and registering
drivers.  All others have either been switched to the vanilla platform
bus or already have their own infrastructure.

This patch moves the functionality that ibmebus is using out of
drivers/of/{platform,device}.c and into ibmebus.c where it is actually
used.  Also renames the moved symbols from of_platform_* to
ibmebus_bus_* to reflect the actual usage.

This patch is part of moving all of the of_platform_bus_type users
over to the platform_bus_type.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Device nodes with the property status="disabled" are not usable and so
don't register them when parsing the device tree for devices.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>
Cc: Hollis Blanchard <hollis_blanchard@mentor.com>
Cc: Deepak Saxena <deepak_saxena@mentor.com>
Cc: Scott Wood <scottwood@freescale.com>,
Cc: David Gibson <david@gibson.dropbear.id.au>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>

The current code allocates and manages platform_devices created from
the device tree manually.  It also uses an unsafe shortcut for
allocating the platform_device and the resource table at the same
time. (which I added in the last rework; sorry).

This patch refactors the code to use platform_device_alloc() for
allocating new devices.  This reduces the amount of custom code
implemented by of_platform, eliminates the unsafe alloc trick, and has
the side benefit of letting the platform_bus code manage freeing the
device data and resources when the device is freed.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Greg Kroah-Hartman <gregkh@suse.de>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Michal Simek <monstr@monstr.eu>

Instead of referencing NO_IRQ in platform.c, define some helper functions
in irq.c to call instead from platform.c.  Keep NO_IRQ usage local to
irq.c, and define NO_IRQ if not defined in headers.
Signed-off-by: NAndres Salomon <dilinger@queued.net>
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Currently there are some drivers in tree which register both a
platform_driver and an of_platform_driver with the same name.  This is
a temporary situation until all the relevant of_platform_drivers are
converted to be normal platform_drivers.  Until then, this patch gives
all the of_platform_drivers an "of:" prefix to protect against bogus
matches and namespace conflicts.

The AMBA bus should also use of_device_make_bus_id() when populating device
out of device tree data.  This patch makes the function non-static, and
adds a suitable prototype in of_device.h
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

This list used was by only two platforms with all other platforms defining an
own list of valid bus id's to pass to of_platform_bus_probe.  This patch:

i)   copies the default list to the two platforms that depended on it (powerpc)
ii)  remove the usage of of_default_bus_ids in of_platform_bus_probe
iii) removes the definition of the list from all architectures that defined it

Passing a NULL 'matches' parameter to of_platform_bus_probe is still valid; the
function returns no error in that case as the NULL value is equivalent to an
empty list.
Signed-off-by: NJonas Bonn <jonas@southpole.se>
[grant.likely@secretlab.ca: added __initdata annotations, warn on and return error on missing match table, and fix whitespace errors]
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

There's no need for this function to be architecture specific and all four
architectures defining it had the same definition.  The function has been
moved to drivers/of/platform.c.
Signed-off-by: NJonas Bonn <jonas@southpole.se>
[grant.likely@secretlab.ca: moved to drivers/of/platform.c, simplified code, and added kerneldoc comment]
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>
Acked-by: NDavid S. Miller <davem@davemloft.net>

of_device is currently just an #define alias to platform_device until it
gets removed entirely.  This patch removes references to it from the
include directories and the core drivers/of code.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>
Acked-by: NDavid S. Miller <davem@davemloft.net>

There is an unlikely chance of this situation is occurring, but it is
easy to protect against.  If a matching entry cannot be found in the
of_match_table, then don't bind the driver.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

of_platform_bus was being used in the same manner as the platform_bus.
The only difference being that of_platform_bus devices are generated
from data in the device tree, and platform_bus devices are usually
statically allocated in platform code.  Having them separate causes
the problem of device drivers having to be registered twice if it
was possible for the same device to appear on either bus.

This patch removes of_platform_bus_type and registers all of_platform
bus devices and drivers on the platform bus instead.  A previous patch
made the of_device structure an alias for the platform_device structure,
and a shim is used to adapt of_platform_drivers to the platform bus.

After all of of_platform_bus drivers are converted to be normal platform
drivers, the shim code can be removed.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>
Acked-by: NDavid S. Miller <davem@davemloft.net>

Only powerpc and microblaze supply (struct device *)->archdata.dma_mask.
This patch stops referencing it on other architectures.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

This patch fixes the condition where device tree support is compiled
in, but no device tree was proved by firmware.  It makes
of_platform_bus_probe() explicitly check for a NULL device tree
pointer.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

Commit 94c09319 (of:
Merge of_device_alloc() and of_device_make_bus_id()) moved code that
does calls a dcr routine without including the correct header which
causes the following build error on some powerpc configurations:

drivers/of/platform.c: In function 'of_device_make_bus_id':
drivers/of/platform.c:437: error: implicit declaration of function 'of_translate_dcr_address'

This patch adds the appropriate header to drivers/of/platform.c
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>