Fix OF selftest compile on sparc which does not enable OF_IRQ.

drivers/of/selftest.c:177: undefined reference to `of_irq_parse_one'
drivers/of/selftest.c:197: undefined reference to `of_irq_parse_one'
drivers/of/selftest.c:248: undefined reference to `of_irq_parse_one'
Signed-off-by: NRob Herring <rob.herring@calxeda.com>
Acked-by: NGrant Likely <grant.likely@linaro.org>

This patch fixes interrupt-map entry matching code to properly match all
specifier cells with interrupt map entries.
Signed-off-by: NTomasz Figa <t.figa@samsung.com>
Tested-by: NSachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: NRob Herring <rob.herring@calxeda.com>

Commit 23616132, "of/irq: Refactor interrupt-map parsing" introduced
a potential buffer overflow bug because it doesn't do sufficient range
checking on the input data. This patch adds the appropriate checking and
buffer size adjustments. If the bounds are out of range then warn
loudly. MAX_PHANDLE_ARGS should be sufficient. If it is not then the
value can be increased.
Signed-off-by: NGrant Likely <grant.likely@linaro.org>
Cc: Rob Herring <rob.herring@calxeda.com>

Commit 23616132, "of/irq: Refactor interrupt-map parsing" introduced
a bug. The irq parsing will fail for some nodes that don't have a reg
property. It is fixed by deferring the check for reg until it is
actually needed. Also adjust the testcase data to catch the bug.
Signed-off-by: NGrant Likely <grant.likely@linaro.org>
Tested-by: NStephen Warren <swarren@nvidia.com>
Tested-by: NMing Lei <tom.leiming@gmail.com>
Tested-by: NStephen Warren <swarren@nvidia.com>
Cc: Rob Herring <rob.herring@calxeda.com>

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>

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>

The standard interrupts property in device tree can only handle
interrupts coming from a single interrupt parent. If a device is wired
to multiple interrupt controllers, then it needs to be attached to a
node with an interrupt-map property to demux the interrupt specifiers
which is confusing. It would be a lot easier if there was a form of the
interrupts property that allows for a separate interrupt phandle for
each interrupt specifier.

This patch does exactly that by creating a new interrupts-extended
property which reuses the phandle+arguments pattern used by GPIOs and
other core bindings.
Signed-off-by: NGrant Likely <grant.likely@linaro.org>
Acked-by: NTony Lindgren <tony@atomide.com>
Acked-by: NKumar Gala <galak@codeaurora.org>
[grant.likely: removed versatile platform hunks into separate patch]
Cc: Rob Herring <rob.herring@calxeda.com>

Several architectures open code effectively the same code block for
finding and mapping PCI irqs. This patch consolidates it down to a
single function.
Signed-off-by: NGrant Likely <grant.likely@linaro.org>
Acked-by: NMichal Simek <monstr@monstr.eu>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>

The of_irq_to_resource() helper that is used to implement of_irq_count()
tries to resolve interrupts and in fact creates a mapping for resolved
interrupts. That's pretty heavy lifting for something that claims to
just return the number of interrupts requested by a given device node.

Instead, use the more lightweight of_irq_map_one(), which, despite the
name, doesn't create an actual mapping. Perhaps a better name would be
of_irq_translate_one().
Signed-off-by: NThierry Reding <treding@nvidia.com>
Acked-by: NRob Herring <rob.herring@calxeda.com>
[grant.likely: fixup s/of_irq_map_one/of_irq_parse_one/]
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

This patch extends the DT selftest code with some test cases for the
interrupt parsing functions.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

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>

All the users of of_irq_parse_raw pass in a raw interrupt specifier from
the device tree and expect it to be returned (possibly modified) in an
of_phandle_args structure. However, the primary function of
of_irq_parse_raw() is to check for translations due to the presence of
one or more interrupt-map properties. The actual placing of the data
into an of_phandle_args structure is trivial. If it is refactored to
accept an of_phandle_args structure directly, then it becomes possible
to consume of_phandle_args from other sources. This is important for an
upcoming patch that allows a device to be connected to more than one
interrupt parent. It also simplifies the code a bit.

The biggest complication with this patch is that the old version works
on the interrupt specifiers in __be32 form, but the of_phandle_args
structure is intended to carry it in the cpu-native version. A bit of
churn was required to make this work. In the end it results in tighter
code, so the churn is worth it.
Signed-off-by: NGrant Likely <grant.likely@linaro.org>
Acked-by: NTony Lindgren <tony@atomide.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>

All the callers of irq_create_of_mapping() pass the contents of a struct
of_phandle_args structure to the function. Since all the callers already
have an of_phandle_args pointer, why not pass it directly to
irq_create_of_mapping()?
Signed-off-by: NGrant Likely <grant.likely@linaro.org>
Acked-by: NMichal Simek <monstr@monstr.eu>
Acked-by: NTony Lindgren <tony@atomide.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>

struct of_irq and struct of_phandle_args are exactly the same structure.
This patch makes the kernel use of_phandle_args everywhere. This in
itself isn't a big deal, but it makes some follow-on patches simpler.
Signed-off-by: NGrant Likely <grant.likely@linaro.org>
Acked-by: NMichal Simek <monstr@monstr.eu>
Acked-by: NTony Lindgren <tony@atomide.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>

The OF irq handling code has been overloading the term 'map' to refer to
both parsing the data in the device tree and mapping it to the internal
linux irq system. This is probably because the device tree does have the
concept of an 'interrupt-map' function for translating interrupt
references from one node to another, but 'map' is still confusing when
the primary purpose of some of the functions are to parse the DT data.

This patch renames all the of_irq_map_* functions to of_irq_parse_*
which makes it clear that there is a difference between the parsing
phase and the mapping phase. Kernel code can make use of just the
parsing or just the mapping support as needed by the subsystem.

The patch was generated mechanically with a handful of sed commands.
Signed-off-by: NGrant Likely <grant.likely@linaro.org>
Acked-by: NMichal Simek <monstr@monstr.eu>
Acked-by: NTony Lindgren <tony@atomide.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>

Some drivers might rely on availability of trigger flags in IRQ
resource, for example to configure the hardware for particular interrupt
type. However current code creating IRQ resources from data in device
tree does not configure trigger flags in resulting resources.

This patch tries to solve the problem, based on the fact that
irq_of_parse_and_map() configures the trigger based on DT interrupt
specifier and IRQD_TRIGGER_* flags are consistent with IORESOURCE_IRQ_*,
and we can get correct trigger flags by calling irqd_get_trigger_type()
after mapping the interrupt.
Signed-off-by: NTomasz Figa <tomasz.figa@gmail.com>
[grant.likely: Merged the two assignments to r->flags]
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

Several locations in the of_address and of_irq code dereference the
full_name parameter from a device_node pointer without checking if the
pointer is valid.  This patch switches to use of_node_full_name() which
always checks the pointer.
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>

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>

This reverts commit 109b6236.

Tim Bird expressed concern that this will have a bad effect on boot
time, and while simple tests have shown it to be okay with simple tree,
a device tree blob can potentially be quite large and
add_device_randomness() is not a fast function. Rather than do this for
all platforms unconditionally, I'm reverting this patch and would like
to see it revisited. Instead of feeding the entire tree into the random
pool, it would probably be appropriate to hash the tree and feed the
hash result into the pool. There really isn't a lot of randomness in a
device tree anyway. In the majority of cases only a handful of
properties are going to be different between machines with the same
baseboard.
Signed-off-by: NGrant Likely <grant.likely@secretlab.ca>

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>

Implement pci_address_to_pio 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>
Cc: Michal Simek <monstr@monstr.eu>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: x86@kernel.org
Cc: Grant Likely <grant.likely@linaro.org>

Remove unnecessary prom.h includes in preparation to make prom.h optional.
Signed-off-by: NRob Herring <rob.herring@calxeda.com>
Acked-by: NGrant Likely <grant.likely@linaro.org>

Introduce common of_flat_dt_match_machine and
of_flat_dt_get_machine_name functions to unify architectures' handling
of machine level model and compatible properties.

Several architectures match the root compatible string with an arch
specific list of machine descriptors duplicating the same search
algorithm. Create a common implementation with a simple architecture
specific hook to iterate over each machine's match table.
Signed-off-by: NRob Herring <rob.herring@calxeda.com>
Acked-by: NGrant Likely <grant.likely@linaro.org>

All arches do essentially the same thing now for
early_init_dt_setup_initrd_arch, so it can now be removed.
Signed-off-by: NRob Herring <rob.herring@calxeda.com>
Acked-by: NVineet Gupta <vgupta@synopsys.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Mark Salter <msalter@redhat.com>
Cc: Aurelien Jacquiot <a-jacquiot@ti.com>
Cc: James Hogan <james.hogan@imgtec.com>
Cc: Michal Simek <monstr@monstr.eu>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Jonas Bonn <jonas@southpole.se>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: x86@kernel.org
Cc: Chris Zankel <chris@zankel.net>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Acked-by: NGrant Likely <grant.likely@linaro.org>

Create a weak version of early_init_dt_add_memory_arch which uses
memblock. This will unify all architectures except ones with custom
memory bank structs.
Signed-off-by: NRob Herring <rob.herring@calxeda.com>
Acked-by: NCatalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Michal Simek <monstr@monstr.eu>
Cc: Jonas Bonn <jonas@southpole.se>
Acked-by: NGrant Likely <grant.likely@linaro.org>
Cc: linux-arm-kernel@lists.infradead.org
Cc: microblaze-uclinux@itee.uq.edu.au
Cc: linux@lists.openrisc.net
Cc: devicetree@vger.kernel.org

Most architectures scan the all the same items early in the FDT and none
are really architecture specific. Create a common early_init_dt_scan to
unify the early scan of root, memory, and chosen nodes in the flattened
DT.
Signed-off-by: NRob Herring <rob.herring@calxeda.com>
Acked-by: NGrant Likely <grant.likely@linaro.org>

Several architectures using DT support built-in dtb's in the init
section. These platforms need to copy the dtb from init since the
strings are referenced after unflattening. Every arch has their own
copying routine which do the same thing. Create a common function,
unflatten_and_copy_device_tree, to copy the dtb when unflattening the
dtb.
Signed-off-by: NRob Herring <rob.herring@calxeda.com>
Acked-by: NGrant Likely <grant.likely@linaro.org>

It is not needed to include asm/dma-contiguous.h header to compile
reserved memory initialization code, so remove it to avoid build break
on architectures without CMA support.
Signed-off-by: NMarek Szyprowski <m.szyprowski@samsung.com>
Tested-by: NGuenter Roeck <linux@roeck-us.net>

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>

Fix trivial comment typo for irq_of_parse_and_map().
Signed-off-by: NYijing Wang <wangyijing@huawei.com>
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

We feed the entire DMI table into the random pool to provide
better random data during early boot, so do the same with the
flattened device tree.
Signed-off-by: NAnton Blanchard <anton@samba.org>
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

The flat tree unflatting path is using unsigned longs to carry around
virtual address pointers to the device tree and the allocated memory
used to unpack it. This is a little insane since every access to them
needs to be cast to a pointer type before using it. This patch changes
the data type to void* for the 'start' and 'mem' pointers and reworks
the unflattening functions to use those values directly which results in
slightly simpler code.
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

Patch 9e401275, "of: fdt: fix memory initialization for expanded DT"
fixed incomplete clearing of memory when unflattening the device tree.
However the code was already clearing some of the memory, it just wasn't
doing so for all allocations. Now that the memory is cleared right at
the point of allocation, the memset after unflatten_dt_alloc() is
redundant. Remove it.
Signed-off-by: NGrant Likely <grant.likely@linaro.org>
Acked-by: NWladislav Wiebe <wladislav.kw@gmail.com>

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>