The cpufreq documentation specifies
policy->cpuinfo.transition_latency the time it takes on this CPU to
switch between two frequencies in
nanoseconds (if appropriate, else
specify CPUFREQ_ETERNAL)
currently pcc-cpufreq does not expose the value and sets it to zero. I
changed the pcc-cpufreq driver and it's documentation to conform to the
default value specified in Documentation/cpu-freq/cpu-drivers.txt
Signed-off-by: Jacob Tanenbaum <jtanenba@redhat.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Register passive cooling devices when initialising cpufreq on
big.LITTLE systems. If the device tree provides a dynamic power
coefficient for the CPUs then the bound cooling device will support
the extensions that allow it to be used with all the existing thermal
governors including the power allocator governor.
A cooling device will be created per individual frequency domain and
can be bound to thermal zones via the thermal DT bindings.
Signed-off-by: Punit Agrawal <punit.agrawal@arm.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Support registering cooling devices with dynamic power coefficient
where provided by the device tree. This allows OF registered cooling
devices driver to be used with the power_allocator thermal governor.
Signed-off-by: Punit Agrawal <punit.agrawal@arm.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Javi Merino <javi.merino@arm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The dynamic power consumption of a device is proportional to the
square of voltage (V) and the clock frequency (f). It can be expressed as
Pdyn = dynamic-power-coefficient * V^2 * f.
The coefficient represents the running time dynamic power consumption in
units of mw/MHz/uVolt^2 and can be used in the above formula to
calculate the dynamic power in mW.
Signed-off-by: Punit Agrawal <punit.agrawal@arm.com>
Acked-by: Rob Herring <robh@kernel.org>
Reviewed-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Add the dt-bindings header for the rk3228, that gets shared between
the clock controller and the clock references in the dts.
Signed-off-by: Jeffy Chen <jeffy.chen@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
If partition parsers need to clean up their resources, we shouldn't
assume that all memory will fit in a single kmalloc() that the caller
can kfree(). We should allow the parser to provide a proper cleanup
routine.
Note that this means we need to keep a hold on the parser's module for a
bit longer, and release it later with mtd_part_parser_put().
Alongside this, define a default callback that we'll automatically use
if the parser doesn't provide one, so we can still retain the old
behavior.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Reviewed-by: Boris Brezillon <boris.brezillon@free-electrons.com>
If the CONFIG_ACPI Kconfig symbol is not enabled and a partial build is
attempted, compile errors will happen due missing types and identifiers.
This can be easily reproduced with the following commands:
$ export CROSS_COMPILE="arm-linux-gnueabihf-" ARCH=arm
$ make allmodconfig
$ make M=drivers/acpi/
CC drivers/acpi//tables.o
drivers/acpi//tables.c:235:3: warning: 'struct acpi_subtable_proc' declared inside parameter list
unsigned int max_entries)
^
drivers/acpi//tables.c:235:3: warning: its scope is only this definition or declaration, which is probably not what you want
drivers/acpi//tables.c: In function 'acpi_parse_entries_array':
drivers/acpi//tables.c:269:4: error: invalid use of undefined type 'struct acpi_subtable_proc'
...
scripts/Makefile.build:258: recipe for target 'drivers/acpi//tables.o' failed
make[1]: *** [drivers/acpi//tables.o] Error 1
Makefile:1401: recipe for target '_module_drivers/acpi/' failed
make: *** [_module_drivers/acpi/] Error 2
This is because objects are tried to be built unconditionally even when
CONFIG_ACPI is not enabled. This is usually not a problem since arches'
Kconfig sources drivers/acpi/Kconfig directly and also selects ACPI but
the Makefile should conditionally build the objects as well to prevent
these build errors.
Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
D3cold is only regarded as valid if the "_PR3" object is
present for the given device after the commit 20dacb71ad
("ACPI/PM: Rework device power management to follow ACPI 6").
But some old BIOS only defined "_PS3" for the D3COLD device,
such as ZPODD device. And old kernel also believes the device with
"_PS3" is a D3COLD device.
So, add some logics for supporting D3 COLD device with old BIOS
which is compatible with earlier ACPI spec and kernel behavior.
Link: http://marc.info/?l=linux-acpi&m=144946938709759&w=2
Signed-off-by: Ken Xue <Ken.Xue@amd.com>
Reported-and-tested-by: Gang Long <Gang.Long@amd.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
A trivial change suggested by checkpatch.pl.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Rob Herring <robh@kernel.org>
Currently, of_irq_init() iterates over interrupt controller nodes
with for_each_matching_node(), and then gets each init function with
of_match_node() later.
This routine can be optimized with for_each_matching_node_and_match().
It allows to get the interrupt controller node and its init function
at the same time, saving __of_match_node() callings.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Rob Herring <robh@kernel.org>
All the users of the tda998x driver are component based and bind the
driver via the device graph method described in
Documentation/devicetree/bindings/graph.txt. Add the fact that the
'port' node is required to the bindings.
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
Signed-off-by: Rob Herring <robh@kernel.org>
Merge two bindings for the same driver to together.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Acked-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Acked-by: Moritz Fischer <moritz.fischer@ettus.com>
Signed-off-by: Rob Herring <robh@kernel.org>
Fix a few typos and reword the description of the
'#qca,ddr-wb-channel-cells' property.
Signed-off-by: Alban Bedel <albeu@free.fr>
CC: trivial@kernel.org
Signed-off-by: Rob Herring <robh@kernel.org>
The existing device tree bindings assume that we are only trying to
describe a single address space with a device tree (for ARM, either
the Normal or the Secure world). Some uses for device tree need to
describe both Normal and Secure worlds in a single device tree. Add
documentation of how to do this, by adding extra properties which
describe when a device appears differently in the two worlds or when
it only appears in one of them.
The binding describes the general principles for adding new
properties describing the secure world, but for now we only need a
single new property, "secure-status", which can be used to annotate
devices to indicate that they are only visible in one of the two
worlds.
The primary expected use of this binding is for a virtual machine
like QEMU to describe the VM layout to a TrustZone aware firmware
(which would then use the secure-only devices itself, and pass the DT
on to a kernel running in the non-secure world, which ignores the
secure-only devices and uses the rest).
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Rob Herring <robh@kernel.org>
RK3228's armclk has 3 parents, so allow cpuclk to have
more than 2 parents.
Signed-off-by: Jeffy Chen <jeffy.chen@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
This company already provided some products, so add them to the
vendor prefix list.
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: Rob Herring <robh@kernel.org>
Document device-tree bindings for the USB controller on older
OCTEON SOCs (OCTEON, OCTEON+).
Signed-off-by: Aaro Koskinen <aaro.koskinen@iki.fi>
Signed-off-by: Rob Herring <robh@kernel.org>
SRAM bindings for various SoCs, using the mmio-sram genalloc
API, are spread over different places - per SoC vendor. Since all of
these are quite similar (they depend on mmio-sram) move them to a common
place.
Suggested-by: Rob Herring <robh+dt@kernel.org>
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Cc: Heiko Stuebner <heiko@sntech.de>
Cc: Maxime Ripard <maxime.ripard@free-electrons.com>
Cc: Chen-Yu Tsai <wens@csie.org>
Cc: Kukjin Kim <kgene@kernel.org>
Acked-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Rob Herring <robh@kernel.org>
Fix the incorrect interrupt documentation file path in binding docs.
Signed-off-by: Andrew F. Davis <afd@ti.com>
Acked-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Rob Herring <robh@kernel.org>
It is possible to get rid of the timer_lock spinlock used by the
governor timer function for synchronization, but a couple of races
need to be avoided.
The first race is between multiple dbs_timer_handler() instances
that may be running in parallel with each other on different
CPUs. Namely, one of them has to queue up the work item, but it
cannot be queued up more than once. To achieve that,
atomic_inc_return() can be used on the skip_work field of
struct cpu_common_dbs_info.
The second race is between an already running dbs_timer_handler()
and gov_cancel_work(). In that case the dbs_timer_handler() might
not notice the skip_work incrementation in gov_cancel_work() and
it might queue up its work item after gov_cancel_work() had
returned (and that work item would corrupt skip_work going
forward). To prevent that from happening, gov_cancel_work()
can be made wait for the timer function to complete (on all CPUs)
right after skip_work has been incremented.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Currently update_sampling_rate() runs over each online CPU and
cancels/queues timers on all policy->cpus every time. This should be
done just once for any cpu belonging to a policy.
Create a cpumask and keep on clearing it as and when we process
policies, so that we don't have to traverse through all CPUs of the same
policy.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
cpufreq governors evaluate load at sampling rate and based on that they
update frequency for a group of CPUs belonging to the same cpufreq
policy.
This is required to be done in a single thread for all policy->cpus, but
because we don't want to wakeup idle CPUs to do just that, we use
deferrable work for this. If we would have used a single delayed
deferrable work for the entire policy, there were chances that the CPU
required to run the handler can be in idle and we might end up not
changing the frequency for the entire group with load variations.
And so we were forced to keep per-cpu works, and only the one that
expires first need to do the real work and others are rescheduled for
next sampling time.
We have been using the more complex solution until now, where we used a
delayed deferrable work for this, which is a combination of a timer and
a work.
This could be made lightweight by keeping per-cpu deferred timers with a
single work item, which is scheduled by the first timer that expires.
This patch does just that and here are important changes:
- The timer handler will run in irq context and so we need to use a
spin_lock instead of the timer_mutex. And so a separate timer_lock is
created. This also makes the use of the mutex and lock quite clear, as
we know what exactly they are protecting.
- A new field 'skip_work' is added to track when the timer handlers can
queue a work. More comments present in code.
Suggested-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Ashwin Chaugule <ashwin.chaugule@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
This is using completely the wrong mask and value when updating the
register. Since the correct values are already defined in the header,
switch to using a table with explicit constants rather than shifting the
array index.
Signed-off-by: John Keeping <john@metanate.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Cc: stable@vger.kernel.org
The previous commit removed the only use of these variables.
Signed-off-by: John Keeping <john@metanate.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
If we only clear the tx/rx state when both are disabled it is not
possible to start/stop one multiple times while the other is running.
Since the two are independently controlled, treat them as such and
remove the false dependency between capture and playback.
Signed-off-by: John Keeping <john@metanate.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Since the mmio has support the 64-bit has been supported for the
64-bit platform, so should the regcache core too.
Signed-off-by: Xiubo Li <lixiubo@cmss.chinamobile.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
There will be some warning like the following when checking new
patches near this code:
"WARNING: Missing a blank line after declarations"
This patch will suppress this warning.
Signed-off-by: Xiubo Li <lixiubo@cmss.chinamobile.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
sock_cgroup_data is a struct containing an anonymous union.
sock_cgroup_set_prioidx() and sock_cgroup_set_classid() were
initializing a field inside the anonymous union as follows.
struct sock_ccgroup_data skcd_buf = { .val = VAL };
While this is fine on more recent compilers, gcc-4.4.7 triggers the
following errors.
include/linux/cgroup-defs.h: In function ‘sock_cgroup_set_prioidx’:
include/linux/cgroup-defs.h:619: error: unknown field ‘val’ specified in initializer
include/linux/cgroup-defs.h:619: warning: missing braces around initializer
include/linux/cgroup-defs.h:619: warning: (near initialization for ‘skcd_buf.<anonymous>’)
This is because .val belongs to the anonymous union nested inside the
struct but the initializer is missing the nesting. Fix it by adding
an extra pair of braces.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Alaa Hleihel <alaa@dev.mellanox.co.il>
Fixes: bd1060a1d6 ("sock, cgroup: add sock->sk_cgroup")
Signed-off-by: David S. Miller <davem@davemloft.net>
some hardware.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iEYEABECAAYFAlZof7kACgkQIXnXXONXERehTQCeI3y8x6wza8XwhNFUxrCtG5wc
0ioAoJS0ybm+wN5tLUjXLUlhwqK9j9jT
=lJdc
-----END PGP SIGNATURE-----
Merge tag 'for-linus-4.4-1' of git://git.code.sf.net/p/openipmi/linux-ipmi
Pull IPMI fix from Corey Minyard:
"Fix an Oops if an interrupt occurs at startup. This can happen on
some hardware"
* tag 'for-linus-4.4-1' of git://git.code.sf.net/p/openipmi/linux-ipmi:
ipmi: move timer init to before irq is setup
We encountered a panic on boot in ipmi_si on a dell per320 due to an
uninitialized timer as follows.
static int smi_start_processing(void *send_info,
ipmi_smi_t intf)
{
/* Try to claim any interrupts. */
if (new_smi->irq_setup)
new_smi->irq_setup(new_smi);
--> IRQ arrives here and irq handler tries to modify uninitialized timer
which triggers BUG_ON(!timer->function) in __mod_timer().
Call Trace:
<IRQ>
[<ffffffffa0532617>] start_new_msg+0x47/0x80 [ipmi_si]
[<ffffffffa053269e>] start_check_enables+0x4e/0x60 [ipmi_si]
[<ffffffffa0532bd8>] smi_event_handler+0x1e8/0x640 [ipmi_si]
[<ffffffff810f5584>] ? __rcu_process_callbacks+0x54/0x350
[<ffffffffa053327c>] si_irq_handler+0x3c/0x60 [ipmi_si]
[<ffffffff810efaf0>] handle_IRQ_event+0x60/0x170
[<ffffffff810f245e>] handle_edge_irq+0xde/0x180
[<ffffffff8100fc59>] handle_irq+0x49/0xa0
[<ffffffff8154643c>] do_IRQ+0x6c/0xf0
[<ffffffff8100ba53>] ret_from_intr+0x0/0x11
/* Set up the timer that drives the interface. */
setup_timer(&new_smi->si_timer, smi_timeout, (long)new_smi);
The following patch fixes the problem.
To: Openipmi-developer@lists.sourceforge.net
To: Corey Minyard <minyard@acm.org>
CC: linux-kernel@vger.kernel.org
Signed-off-by: Jan Stancek <jstancek@redhat.com>
Signed-off-by: Tony Camuso <tcamuso@redhat.com>
Signed-off-by: Corey Minyard <cminyard@mvista.com>
Cc: stable@vger.kernel.org # Applies cleanly to 3.10-, needs small rework before
ROL on a 32 bit integer with a shift of 32 or more is undefined and the
result is arch-dependent. Avoid this by handling the trivial case of
roling by 0 correctly.
The trivial solution of checking if shift is 0 breaks gcc's detection
of this code as a ROL instruction, which is unacceptable.
This bug was reported and fixed in GCC
(https://gcc.gnu.org/bugzilla/show_bug.cgi?id=57157):
The standard rotate idiom,
(x << n) | (x >> (32 - n))
is recognized by gcc (for concreteness, I discuss only the case that x
is an uint32_t here).
However, this is portable C only for n in the range 0 < n < 32. For n
== 0, we get x >> 32 which gives undefined behaviour according to the
C standard (6.5.7, Bitwise shift operators). To portably support n ==
0, one has to write the rotate as something like
(x << n) | (x >> ((-n) & 31))
And this is apparently not recognized by gcc.
Note that this is broken on older GCCs and will result in slower ROL.
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Add fallback compatibility string for R-Car Gen 2 family. This is in
keeping with the fallback scheme being adopted wherever appropriate for
drivers for Renesas SoCs.
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Add fallback compatibility string for R-Car Gen 2 family. This is in
keeping with the fallback scheme being adopted wherever appropriate for
drivers for Renesas SoCs.
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Rob Herring <robh@kernel.org>
When applying block operations (BOPs) do not remove them from the
uncommitted BOP ring-buffer until after they've been applied -- in case
we recurse.
Also, perform BOP_INC operation, in dm_sm_metadata_create() and
sm_metadata_extend(), in terms of the uncommitted BOP ring-buffer rather
than using direct calls to sm_ll_inc().
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Cc: stable@vger.kernel.org
For some of the core partitioning code, it helps to keep info about the
parsed partition (and who parsed them) together in one place.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
The use of kmemdup() complicates the error handling a bit. We don't
actually need to allocate new memory, since this reference is treated as
const, and it is copied into new memory by the partition registration
code anyway. So remove it.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Reviewed-by: Boris Brezillon <boris.brezillon@free-electrons.com>
We're going to reuse put_partition_parser(), so let's fix up the prefix
naming a bit, to hopefully be more consistent. Also make convert to a
true C function instead of a macro.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Reviewed-by: Boris Brezillon <boris.brezillon@free-electrons.com>
We only want to modify these arrays inside the parser "drivers", so the
drivers should construct them however they like, then return them as
immutable arrays.
This will make other refactorings easier.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Reviewed-by: Boris Brezillon <boris.brezillon@free-electrons.com>
It's easier to refactor these parsers if the return value gets assigned
only once, just like every other MTD partition parser.
This prepares for making the second arg to the parse_fn() const. This is
OK if we construct the partitions completely first, and assign them to
the return pointer only after we're done modifying them.
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
Reviewed-by: Boris Brezillon <boris.brezillon@free-electrons.com>
When you take a metadata snapshot the btree roots for the mapping and
details tree need to have their reference counts incremented so they
persist for the lifetime of the metadata snap.
The roots being incremented were those currently written in the
superblock, which could possibly be out of date if concurrent IO is
triggering new mappings, breaking of sharing, etc.
Fix this by performing a commit with the metadata lock held while taking
a metadata snapshot.
Signed-off-by: Joe Thornber <ejt@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Cc: stable@vger.kernel.org
Hopefully final set of fixes for v4.4 release cycle.
There's a fix for a regression on dwc3 caused by recent changes to how
transfers are started. We're not pre-starting interrupt endpoints
anymore.
A NULL pointer dereference fix for the MSM phy driver.
The UVC gadget got a minor fix for permissions to its configfs
attributes and, finally, two fixes for MUSB. A fix for PM runtime when
MUSB returns EPROBE_DEFER and a fix to actually return an error in case
we can't initialize a DMA engine.
Signed-off-by: Felipe Balbi <balbi@ti.com>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1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=o5zI
-----END PGP SIGNATURE-----
Merge tag 'fixes-for-v4.4-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb into usb-linus
Felipe writes:
usb: fixes for-v4.4-rc5
Hopefully final set of fixes for v4.4 release cycle.
There's a fix for a regression on dwc3 caused by recent changes to how
transfers are started. We're not pre-starting interrupt endpoints
anymore.
A NULL pointer dereference fix for the MSM phy driver.
The UVC gadget got a minor fix for permissions to its configfs
attributes and, finally, two fixes for MUSB. A fix for PM runtime when
MUSB returns EPROBE_DEFER and a fix to actually return an error in case
we can't initialize a DMA engine.
Signed-off-by: Felipe Balbi <balbi@ti.com>
The "|" operator has higher precedence than "?:" so this didn't work as
intended. I had previously fixed this bug, but it we copied the older
unfixed version when we moved the function between files.
Fixes: 1673f1f08c ('nvme: move block_device_operations and ns/ctrl freeing to common code')
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
