25 Jul, 2012
2 commits
-
When the map operation fails log the error code we get and add a WARN_ON()
so we get a backtrace (which should help work out which interrupt is the
source of the issue).Signed-off-by: Mark Brown
Signed-off-by: Grant Likely -
With the current state of irq_domain, the reverse map is always updated
when new IRQs get mapped. This means that the irq_find_mapping() function
can be simplified to execute the revmap lookup functions unconditionallyThis patch adds lookup functions for the revmaps that don't yet have one
and removes the slow path lookup code path.v8: Broke out unrelated changes into separate patches. Rebased on Paul's irq
association patches.
v7: Rebased to irqdomain/next for v3.4 and applied before the removal of 'hint'
v6: Remove the slow path entirely. The only place where the slow path
could get called is for a linear mapping if the hwirq number is larger
than the linear revmap size. There shouldn't be any interrupt
controllers that do that.
v5: rewrite to not use a ->revmap() callback. It is simpler, smaller,
safer and faster to open code each of the revmap lookups directly into
irq_find_mapping() via a switch statement.
v4: Fix build failure on incorrect variable reference.Signed-off-by: Grant Likely
Cc: Benjamin Herrenschmidt
Cc: Thomas Gleixner
Cc: Milton Miller
Cc: Paul Mundt
Cc: Rob Herring
11 Jul, 2012
7 commits
-
irq_create_direct_mapping can only be used with the NOMAP type. Make
the function test to ensure it is passed the correct type of
irq_domain.Signed-off-by: Grant Likely
Cc: Paul Mundt
Cc: Benjamin Herrenschmidt
Cc: Thomas Gleixner
Cc: Rob Herring -
In preparation to remove the slow revmap path, eliminate the public
radix revmap lookup functions. This simplifies the code and makes the
slowpath removal patch a lot simpler.Signed-off-by: Grant Likely
Cc: Paul Mundt
Cc: Benjamin Herrenschmidt
Cc: Thomas Gleixner
Cc: Rob Herring -
This adds a new strict mapping API for supporting creation of linux IRQs
at existing positions within the domain. The new routines are as follows:For dynamic allocation and insertion to specified ranges:
- irq_create_identity_mapping()
- irq_create_strict_mappings()These will allocate and associate a range of linux IRQs at the specified
location. This can be used by controllers that have their own static linux IRQ
definitions to map a hwirq range to, as well as for platforms that wish to
establish 1:1 identity mapping between linux and hwirq space.For insertion to specified ranges by platforms that do their own irq_desc
management:- irq_domain_associate()
- irq_domain_associate_many()These in turn call back in to the domain's ->map() routine, for further
processing by the platform. Disassociation of IRQs get handled through
irq_dispose_mapping() as normal.With these in place it should be possible to begin migration of legacy IRQ
domains to linear ones, without requiring special handling for static vs
dynamic IRQ definitions in DT vs non-DT paths. This also makes it possible
for domains with static mappings to adopt whichever tree model best fits
their needs, rather than simply restricting them to linear revmaps.Signed-off-by: Paul Mundt
[grant.likely: Reorganized irq_domain_associate{,_many} to have all logic in one place]
[grant.likely: Add error checking for unallocated irq_descs at associate time]
Signed-off-by: Grant Likely
Cc: Benjamin Herrenschmidt
Cc: Thomas Gleixner
Cc: Rob Herring -
At irq_setup_virq() time all of the data needed to update the reverse
map is available, but the current code ignores it and relies upon the
slow path to insert revmap records. This patch adds revmap updating
to the setup path so the slow path will no longer be necessary.Signed-off-by: Grant Likely
Cc: Paul Mundt
Cc: Benjamin Herrenschmidt
Cc: Thomas Gleixner
Cc: Rob Herring -
This patch moves the irq disassociation code out into a separate
function in preparation to extend irq_setup_virq to handle multiple
irqs and rename it for use by interrupt controller drivers. The new
function will be used by irq_setup_virq() in its error path.Signed-off-by: Grant Likely
Acked-by: Benjamin Herrenschmidt
Cc: Paul Mundt
Cc: Thomas Gleixner
Cc: Rob Herring -
The revmap type should be linear for irq_domain_add_linear function.
Signed-off-by: Dong Aisheng
Signed-off-by: Grant Likely -
A large proportion of interrupt controllers that support legacy mappings
do so because non-DT systems need to use fixed IRQ numbers when registering
devices via buses but can otherwise use a linear mapping. The interrupt
controller itself typically is not affected by the mapping used and best
practice is to use a linear mapping where possible so drivers frequently
select at runtime depending on if a legacy range has been allocated to
them.Standardise this behaviour by providing irq_domain_register_simple() which
will allocate a linear mapping unless a positive first_irq is provided in
which case it will fall back to a legacy mapping. This helps make best
practice for irq_domain adoption clearer.Signed-off-by: Mark Brown
Signed-off-by: Grant Likely
18 Jun, 2012
1 commit
-
There isn't a really compelling reason to force ->map to be populated,
so allow it to be left unset.Signed-off-by: Grant Likely
Acked-by: Benjamin Herrenschmidt
Cc: Paul Mundt
Cc: Thomas Gleixner
Cc: Rob Herring
16 Jun, 2012
3 commits
-
Where irq_domain_associate() is called in irq_create_mapping, there is
no need to test for IRQ_DOMAIN_MAP_LEGACY because it is already tested
for earlier in the routine.Signed-off-by: Grant Likely
Cc: Paul Mundt
Cc: Benjamin Herrenschmidt
Cc: Thomas Gleixner
Cc: Rob Herring -
While common irqdesc allocation is node aware, the irqdomain code is not.
Presently we observe a number of regressions/inconsistencies on
NUMA-capable platforms:- Platforms using irqdomains with legacy mappings, where the
irq_descs are allocated node-local and the irqdomain data
structure is not.- Drivers implementing irqdomains will lose node locality
regardless of the underlying struct device's node id.This plugs in NUMA node id proliferation across the various allocation
callsites by way of_node_to_nid() node lookup. While of_node_to_nid()
does the right thing for OF-capable platforms it doesn't presently handle
the non-DT case. This is trivially dealt with by simply wraping in to
numa_node_id() unconditionally.Signed-off-by: Paul Mundt
Cc: Benjamin Herrenschmidt
Cc: Thomas Gleixner
Cc: Rob Herring
Signed-off-by: Grant Likely -
The pattern (np ? np->full_name : "") is rather common in the
kernel, but can also make for quite long lines. This patch adds a new
inline function, of_node_full_name() so that the test for a valid node
pointer doesn't need to be open coded at all call sites.Signed-off-by: Grant Likely
Cc: Paul Mundt
Cc: Benjamin Herrenschmidt
Cc: Thomas Gleixner
20 May, 2012
5 commits
-
Signed-off-by: Mark Brown
Signed-off-by: Grant Likely -
Convert to pr_fmt before things start to get out of hand and some
janitors start getting overly excited.Signed-off-by: Paul Mundt
Signed-off-by: Grant Likely -
Presently irq_domain_simple_map() isn't labelled as static, but there's
no definition for it in the public irqdomain header either. At present
all in-tree ->map users have meaningful work to do, and all others are
using irq_domain_simple_ops directly. Make it static for now, as it can
always be exported and added to the public API later.Signed-off-by: Paul Mundt
Signed-off-by: Grant Likely -
modules making use of irq domains at the very least need access to the
add/remove/lookup routines, though there's nothing preventing them from
using the remainder of the public API, either.The current set of exports seem primarily geared at DT-enabled platforms
using DT-backed IRQ domains, where many of the API accesses are hidden
away in OF code. The non-DT cases need to do most of this on their own.Signed-off-by: Paul Mundt
Signed-off-by: Grant Likely -
Now that IRQ domains are being used by modules it's necessary to support
removing them, too. This adds a new irq_domain_remove() routine for doing
the bulk of the heavy lifting. It's left as an exercise to the caller to
ensure all mappings have been appropriatey disposed of before attempting
to remove the domain.Signed-off-by: Paul Mundt
Signed-off-by: Grant Likely
13 Apr, 2012
1 commit
-
sizeof(void*) returns an unsigned long, but it was being used as a width parameter to a "%-*s" format string which requires an int. On 64 bit platforms this causes a type mismatch:
linux/kernel/irq/irqdomain.c:575: warning: field width should have type
'int', but argument 6 has type 'long unsigned int'This change casts the size to an int so printf gets the right data type.
Reported-by: Andreas Schwab
Signed-off-by: Grant Likely
Cc: David Daney
12 Apr, 2012
1 commit
-
This patch replaces the old global setting of irq_virq_count that is only
used by the NOMAP mapping and instead uses a revmap_data property so that
the maximum NOMAP allocation can be set per NOMAP irq_domain.There is exactly one user of irq_virq_count in-tree right now: PS3.
Also, irq_virq_count is only useful for the NOMAP mapping. So,
instead of having a single global irq_virq_count values, this change
drops it entirely and added a max_irq argument to irq_domain_add_nomap().
That makes it a property of an individual nomap irq domain instead of
a global system settting.Signed-off-by: Grant Likely
Tested-by: Benjamin Herrenschmidt
Cc: Thomas Gleixner
Cc: Milton Miller
11 Apr, 2012
2 commits
-
This patch fixes the irq_domain_mapping debugfs output to pad pointer
values with leading zeros so that pointer values are displayed
correctly. Otherwise you get output similar to "0x 5e0000000000000".
Also, when the irq_domain is set to 'null'Signed-off-by: Grant Likely
Cc: David Daney
Cc: Mika Westerberg -
In commit 4bbdd45a (irq_domain/powerpc: eliminate irq_map; use
irq_alloc_desc() instead) code was added that ignores error returns
from irq_alloc_desc_from() by (silently) casting the return value to
unsigned. The negitive value error return now suddenly looks like a
valid irq number.Commits cc79ca69 (irq_domain: Move irq_domain code from powerpc to
kernel/irq) and 1bc04f2c (irq_domain: Add support for base irq and
hwirq in legacy mappings) move this code to its current location in
irqdomain.cThe result of all of this is a null pointer dereference OOPS if one of
the error cases is hit.The fix: Don't cast away the negativeness of the return value and then
check for errors.Signed-off-by: David Daney
Acked-by: Rob Herring
[grant.likely: dropped addition of new 'irq' variable]
Signed-off-by: Grant Likely
30 Mar, 2012
1 commit
-
The debugfs code is really generic for all platforms. This patch removes the
powerpc-specific directory reference and makes it available to all
architectures.Signed-off-by: Grant Likely
25 Feb, 2012
1 commit
-
This patch makes IRQ_DOMAIN usable on MIPS. It uses an ugly workaround
to preserve current behaviour so that MIPS has time to add irq_domain
registration to the irq controller drivers. The workaround will be
removed in Linux v3.6Signed-off-by: Grant Likely
Cc: Ralf Baechle
Cc: Rob Herring
Cc: Thomas Gleixner
Cc: linux-mips@linux-mips.org
16 Feb, 2012
9 commits
-
Make irq_domain_ops pointer a constant to make it safer for multiple
instances to share the same ops pointer and change the irq_domain code
so that it does not modify the ops.v4: Fix mismatched type reference in powerpc code
Signed-off-by: Grant Likely
Cc: Thomas Gleixner
Cc: Benjamin Herrenschmidt
Cc: Milton Miller
Tested-by: Olof Johansson -
Rather than having each interrupt controller driver creating its own barely
unique .xlate function for irq_domain, create a library of translators which
any driver can use directly.v5: - Remove irq_domain_xlate_pci(). It was incorrect.
Signed-off-by: Grant Likely
Cc: Rob Herring
Cc: Benjamin Herrenschmidt
Cc: Mark Salter
Cc: Thomas Gleixner
Cc: Milton Miller
Tested-by: Olof Johansson -
irq_domain_add_simple() was a stop-gap measure until complete irq_domain
support was complete. This patch removes the irq_domain_add_simple()
interface.This patch also drops the explicit irq_domain initialization performed
by the mach-versatile code because the versatile interrupt controller
already has irq_domain support built into it. This was a bug that was
hanging around quietly for a while, but with the full irq_domain which
actually verifies that irq_domain ranges are available it would cause
the registration to fail and the system wouldn't boot.v4: Fixed number of irqs in mx5 gpio code
v2: Updated to pass in host_data pointer on irq_domain allocation.Signed-off-by: Grant Likely
Cc: Rob Herring
Cc: Thomas Gleixner
Cc: Milton Miller
Cc: Russell King
Tested-by: Olof Johansson -
This patch removes the simplistic implementation of irq_domains and enables
the powerpc infrastructure for all irq_domain users. The powerpc
infrastructure includes support for complex mappings between Linux and
hardware irq numbers, and can manage allocation of irq_descs.This patch also converts the few users of irq_domain_add()/irq_domain_del()
to call irq_domain_add_legacy() instead.v3: Fix bug that set up too many irqs in translation range.
v2: Fix removal of irq_alloc_descs() call in gic driverSigned-off-by: Grant Likely
Cc: Rob Herring
Cc: Benjamin Herrenschmidt
Cc: Thomas Gleixner
Cc: Milton Miller
Tested-by: Olof Johansson -
Add support for a legacy mapping where irq = (hwirq - first_hwirq + first_irq)
so that a controller driver can allocate a fixed range of irq_descs and use
a simple calculation to translate back and forth between linux and hw irq
numbers. This is needed to use an irq_domain with many of the ARM interrupt
controller drivers that manage their own irq_desc allocations. Ultimately
the goal is to migrate those drivers to use the linear revmap, but doing it
this way allows each driver to be converted separately which makes the
migration path easier.This patch generalizes the IRQ_DOMAIN_MAP_LEGACY method to use
(first_irq-first_hwirq) as the offset between hwirq and linux irq number,
and adds checks to make sure that the hwirq number does not exceed range
assigned to the controller.Signed-off-by: Grant Likely
Cc: Rob Herring
Cc: Benjamin Herrenschmidt
Cc: Thomas Gleixner
Cc: Milton Miller
Tested-by: Olof Johansson -
Each revmap type has different arguments for setting up the revmap.
This patch splits up the generator functions so that each revmap type
can do its own setup and the user doesn't need to keep track of how
each revmap type handles the arguments.This patch also adds a host_data argument to the generators. There are
cases where the host_data pointer will be needed before the function returns.
ie. the legacy map calls the .map callback for each irq before returning.v2: - Add void *host_data argument to irq_domain_add_*() functions
- fixed failure to compile
- Moved IRQ_DOMAIN_MAP_* defines into irqdomain.cSigned-off-by: Grant Likely
Cc: Rob Herring
Cc: Benjamin Herrenschmidt
Cc: Thomas Gleixner
Cc: Milton Miller
Tested-by: Olof Johansson -
No functional changes. Replaces non-exported references to 'host' with domain.
Does not change any symbol names referenced by other .c files.Signed-off-by: Grant Likely
Cc: Benjamin Herrenschmidt
Cc: Thomas Gleixner
Cc: Milton Miller
Tested-by: Olof Johansson -
zero always means no irq when using irq domains. Get rid of the NO_IRQ
references.Signed-off-by: Grant Likely
Cc: Rob Herring
Cc: Benjamin Herrenschmidt
Cc: Thomas Gleixner
Cc: Milton Miller
Tested-by: Olof Johansson -
This patch only moves the code. It doesn't make any changes, and the
code is still only compiled for powerpc. Follow-on patches will generalize
the code for other architectures.Signed-off-by: Grant Likely
Cc: Benjamin Herrenschmidt
Cc: Thomas Gleixner
Cc: Milton Miller
Tested-by: Olof Johansson
15 Feb, 2012
2 commits
-
Part of the series to unify the irq remapping mechanisms in the
kernel. A follow up patch will copy the powerpc implementation into
kernel/irq/irqdomain.c, which will be a lot easier if the structures
are identical.Where they differ, I've chose to use the powerpc names since there is
a lot more code using those names.Signed-off-by: Grant Likely
Cc: Rob Herring
Cc: Benjamin Herrenschmidt
Cc: Thomas Gleixner
Cc: Milton Miller
Tested-by: Olof Johansson -
irq_domain printk's too much. Drop some output.
Signed-off-by: Grant Likely
Cc: Rob Herring
Cc: Thomas Gleixner
Tested-by: Olof Johansson
08 Jan, 2012
1 commit
-
devicetree/next changes queued for v3.3 merge window
* tag 'devicetree-for-linus-20120104' of git://git.secretlab.ca/git/linux-2.6:
ARM: prom.h: Fix build error by removing unneeded header file
irq: check domain hwirq range for DT translate
dt: add empty of_get_node/of_put_node functions
of/pdt: fix section mismatch warning
i2c-designware: add OF binding support
dt/i2c: Enumerate some of the known trivial i2c devices
dt: reform for_each_property to for_each_property_of_node
ARM/of: allow *machine_desc.dt_compat to be const
of/base: Take NULL string into account for property with multiple strings
OF/device-tree: Add some entries to vendor-prefixes.txtFix up trivial add-add conflicts in include/linux/of.h
28 Dec, 2011
1 commit
-
A DT node may have more than 1 domain associated with it, so make sure
the hwirq number is within range when doing DT translation.Signed-off-by: Rob Herring
Acked-by: Grant Likely
Acked-by: Shawn Guo
Cc: Thomas Gleixner
24 Dec, 2011
1 commit
-
irqdomain support is used in interrupt controller drivers that may not
have device tree support but only need the basic HW->Linux irq
translation. Rather than having each of these implement their own IRQ
domain, allow them to use the simple ops.Acked-by: Thomas Gleixner
Acked-by: Rob Herring
Cc: Grant Likely
Signed-off-by: Jamie Iles
Signed-off-by: Russell King
31 Oct, 2011
1 commit
-
Interrupt controllers can have non-zero starting value for h/w irq numbers.
Adding support in irq_domain allows the domain hwirq numbering to match
the interrupt controllers' numbering.As this makes looping over irqs for a domain more complicated, add loop
iterators to iterate over all hwirqs and irqs for a domain.Signed-off-by: Rob Herring
Reviewed-by: Jamie Iles
Tested-by: Thomas Abraham
Acked-by: Grant Likely
Acked-by: Thomas Gleixner
20 Sep, 2011
1 commit
-
The sanity check in irq_domain_add() tests desc->irq_data != NULL or
irq_data->domain != NULL. This prevents adding an irq_domain to a irq
descriptor when irq_data exists, which true when the irq descriptor
exists.This went unnoticed so far as the simple domain code did not enter
this code path because domain->nr_irqs is always 0 for the simple domains.Split the check for irq_data == NULL out and have a separate warning
for it.[ tglx: Made the check for irq_data == NULL separate ]
Signed-off-by: Rob Herring
Cc: Grant Likely
Cc: marc.zyngier@arm.com
Cc: thomas.abraham@linaro.org
Cc: jamie@jamieiles.com
Cc: b-cousson@ti.com
Cc: shawn.guo@linaro.org
Cc: linux-arm-kernel@lists.infradead.org
Cc: devicetree-discuss@lists.ozlabs.org
Link: http://lkml.kernel.org/r/1316017900-19918-3-git-send-email-robherring2@gmail.com
Signed-off-by: Thomas Gleixner
