* pm-cpuidle:
cpuidle: add MAINTAINERS entry for ARM Exynos cpuidle driver
drivers: cpuidle: Remove cpuidle-arm64 duplicate error messages
drivers: cpuidle: Add idle-state-name description to ARM idle states
drivers: cpuidle: Add status property to ARM idle states
cpuidle: Invert CPUIDLE_FLAG_TIME_VALID logic
* pm-domains:
ARM: shmobile: Convert to genpd flags for PM clocks for R-mobile
ARM: shmobile: Convert to genpd flags for PM clocks for r8a7779
PM / Domains: Initial PM clock support for genpd
PM / Domains: Power on the PM domain right after attach completes
PM / Domains: Move struct pm_domain_data to pm_domain.h
PM / Domains: Extract code to power off/on a PM domain
PM / Domains: Make genpd parameter of pm_genpd_present() const
* pm-sleep:
PM / hibernate: Deletion of an unnecessary check before the function call "vfree"
PM / Hibernate: Migrate to ktime_t
* pm-tools:
tools: cpupower: fix return checks for sysfs_get_idlestate_count()
* powercap:
powercap / RAPL: fix build dependency on iosf_mbi
powercap / RAPL: add new model ids
powercap / RAPL: handle atom and core differences
powercap / RAPL: abstract per cpu type functions
* pm-clk:
PM / clock_ops: make __pm_clk_enable more generic
PM / clock_ops: Add pm_clk_add_clk()
* pm-config:
PM: Kconfig: fix unmet dependency for CPU_PM
* pm-opp:
PM / OPP replace kfree_rcu() with call_srcu() in opp_set_availability()
PM / OPP Introduce APIs to remove OPPs
PM / OPP mark OPPs as 'static' or 'dynamic'
PM / OPP don't match for existing OPPs when list is empty
PM / OPP rename 'head' as 'rcu_head' or 'srcu_head' based on its type
As inline_{dir,inode} stat is increased/decreased concurrently by multi threads,
so the value is not so accurate, let's use atomic type for counting accurately.
Signed-off-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* acpi-video:
ACPI / video: Run _BCL before deciding registering backlight
* acpi-pmic:
ACPI / PMIC: AXP288: support virtual GPIO in ACPI table
ACPI / PMIC: support PMIC operation region for XPower AXP288
ACPI / PMIC: support PMIC operation region for CrystalCove
iio/axp288_adc: remove THIS_MODULE owner
mfd/axp20x: avoid irq numbering collision
iio: adc: Add module device table for autoloading
iio: adc: Add support for axp288 adc
mfd: axp20x: Extend axp20x to support axp288 pmic
* acpi-scan:
ACPI: Add _DEP support to fix battery issue on Asus T100TA
* acpi-pm:
ACPI / sleep: Drain outstanding events after disabling multiple GPEs
ACPI / PM: Fixed a typo in a comment
* acpi-lpss:
dmaengine: dw: enable runtime PM
ACPI / LPSS: introduce a 'proxy' device to power on LPSS for DMA
ACPI / LPSS: allow to use specific PM domain during ->probe()
ACPI / LPSS: add all LPSS devices to the specific power domain
* acpi-processor:
ACPI / cpuidle: avoid assigning signed errno to acpi_status
ACPI / processor: remove unused variabled from acpi_processor_power structure
ACPI / processor: Update the comments in processor.h
* acpica:
ACPICA: Events: Always modify GPE registers under the GPE lock
ACPICA: Save current masks of enabled GPEs after enable register writes
ACPICA: Update version to 20141107.
ACPICA: Disassembler: Emit correct string for 0 stop bits.
ACPICA: Disassembler: Update for C-style expressions.
ACPICA: Disassembler: Add support for C-style operators and expressions.
ACPICA: acpiexec: Add option to specify an object initialization file.
ACPICA: iASL: Add support for to_PLD macro.
* device-properties:
leds: leds-gpio: Fix multiple instances registration without 'label' property
leds: leds-gpio: Fix legacy GPIO number case
ACPI / property: Drop size_prop from acpi_dev_get_property_reference()
leds: leds-gpio: Convert gpio_blink_set() to use GPIO descriptors
ACPI / GPIO: Document ACPI GPIO mappings API
net: rfkill: gpio: Add default GPIO driver mappings for ACPI
ACPI / GPIO: Driver GPIO mappings for ACPI GPIOs
input: gpio_keys_polled: Make use of device property API
leds: leds-gpio: Make use of device property API
gpio: Support for unified device properties interface
Driver core: Unified interface for firmware node properties
input: gpio_keys_polled: Add support for GPIO descriptors
leds: leds-gpio: Add support for GPIO descriptors
gpio: sch: Consolidate core and resume banks
gpio / ACPI: Add support for _DSD device properties
misc: at25: Make use of device property API
ACPI: Allow drivers to match using Device Tree compatible property
Driver core: Unified device properties interface for platform firmware
ACPI: Add support for device specific properties
Added some commentaries for code readability and cleaned up if-statement
clearly.
Signed-off-by: Changman Lee <cm224.lee@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
If inode state is dirty, go straight to write.
Suggested-by: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Changman Lee <cm224.lee@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
The inmemory pages should be handled by invalidate_page since it needs to be
released int the truncation path.
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
In do_read_inode, if we failed __recover_inline_status, the inode has inline
flag without increasing its count.
Later, f2fs_evict_inode will decrease the count, which causes -1.
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
This patch revists retrial paths in f2fs.
The basic idea is to use cond_resched instead of retrying from the very early
stage.
Suggested-by: Gu Zheng <guz.fnst@cn.fujitsu.com>
Reviewed-by: Chao Yu <chao2.yu@samsung.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Adding my copyright information for two purposes:
- To get cc'd for future patches to review (Only if people read this header
while sending mail)
- Have done enough changes to earn a place here?
Cc: Amit Daniel Kachhap <amit.daniel@samsung.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
get_property() was an over complicated beast with BUGs. It used to believe that
cpufreq table is present in ascending or descending order, which might not
always be true.
Previous patch has created another freq table in descending order for us and we
better use it now. With that get_property() simply goes away and another helper
get_level() comes in.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
CPUFreq framework *doesn't* guarantee that frequencies present in cpufreq table
will be in ascending or descending order. But cpu_cooling somehow assumes that.
Probably because most of current users are creating this list from DT, which is
done with the help of OPP layer. And OPP layer creates the list in ascending
order of frequencies.
But cpu_cooling can be used for other platforms too, which don't have
frequencies arranged in any order.
This patch tries to fix this issue by creating another list of valid frequencies
in descending order. Care is also taken to throw warnings for duplicate entries.
Later patches would use it to simplify code.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
We already know the value of 'cpufreq_dev->max_level' and so there is no need
calculating that once again. For this, we need to send 'cpufreq_dev' to
get_property().
Make all necessary changes for this change. Because cpufreq_cooling_get_level()
doesn't have access to 'cpufreq_dev', it is updated to iterate over the list of
cpufreq_cooling_devices to get cooling device for the cpu number passed to it.
This also makes it robust to return levels only for the CPU registered via a
cooling device. We don't have to support anything that isn't registered yet.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
As we already have a list of cpufreq_cooling_devices now, lets use it instead of
a local counter.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
We don't use get_property() to find max levels anymore as it is done at boot
now. So, don't support GET_MAXL in get_property().
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
CPU frequency tables don't update after the driver is registered and so we don't
need to iterate over them to find total number of states every time
cpufreq_get_max_state() is called. Do it once at boot time.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
get_cpu_frequency() isn't doing much by itself, just calling get_property(). And
so this wrapper isn't required at all. Get rid of it.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
cpufreq_apply_cooling() has a single caller, cpufreq_set_cur_state() and
cpufreq_set_cur_state() is an unnecessary wrapper over cpufreq_apply_cooling().
Get rid of it by merging both routines.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
When there is serious memory pressure, all workers in a pool could be
blocked, and a new thread cannot be created because it requires memory
allocation.
In this situation a WQ_MEM_RECLAIM workqueue will wake up the
rescuer thread to do some work.
The rescuer will only handle requests that are already on ->worklist.
If max_requests is 1, that means it will handle a single request.
The rescuer will be woken again in 100ms to handle another max_requests
requests.
I've seen a machine (running a 3.0 based "enterprise" kernel) with
thousands of requests queued for xfslogd, which has a max_requests of
1, and is needed for retiring all 'xfs' write requests. When one of
the worker pools gets into this state, it progresses extremely slowly
and possibly never recovers (only waited an hour or two).
With this patch we leave a pool_workqueue on mayday list
until it is clearly no longer in need of assistance. This allows
all requests to be handled in a timely fashion.
We keep each pool_workqueue on the mayday list until
need_to_create_worker() is false, and no work for this workqueue is
found in the pool.
I have tested this in combination with a (hackish) patch which forces
all work items to be handled by the rescuer thread. In that context
it significantly improves performance. A similar patch for a 3.0
kernel significantly improved performance on a heavy work load.
Thanks to Jan Kara for some design ideas, and to Dongsu Park for
some comments and testing.
tj: Inverted the lock order between wq_mayday_lock and pool->lock with
a preceding patch and simplified this patch. Added comment and
updated changelog accordingly. Dongsu spotted missing get_pwq()
in the simplified code.
Cc: Dongsu Park <dongsu.park@profitbricks.com>
Cc: Jan Kara <jack@suse.cz>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: Tejun Heo <tj@kernel.org>
Currently, pool->lock nests inside pool->lock. There's no inherent
reason for this order. The only place where the two locks are held
together is pool_mayday_timeout() and it just got decided that way.
This nesting order turns out to complicate things with the planned
rescuer_thread() update. Let's invert them. This doesn't cause any
behavior differences.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: NeilBrown <neilb@suse.de>
Cc: Dongsu Park <dongsu.park@profitbricks.com>
The at91 cleanups changed a lot of files, this merges in the
latest cleanups to resolve the conflicts
Conflicts:
arch/arm/mach-at91/at91sam9260.c
arch/arm/mach-at91/at91sam9261.c
arch/arm/mach-at91/at91sam9263.c
arch/arm/mach-at91/clock.c
arch/arm/mach-at91/clock.h
drivers/rtc/Kconfig
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
This resolves a nonobvious merge conflict that I got wrong the
first time.
* brcm/stb-smp-uart:
bus: brcmstb_gisb: save and restore GISB timeout
bus: brcmstb_gisb: register the fault code hook
ARM: brcmstb: Kconfig: drop unneeded symbol selections
ARM: brcmstb: reintroduce SMP support
ARM: brcmstb: add debug UART for earlyprintk support
Conflicts:
drivers/bus/brcmstb_gisb.c
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reported-by: Florian Fainelli <f.fainelli@gmail.com>
There is no point checking for validity of 'cpufreq_val' from
cpufreq_thermal_notifier() every time the routine is called. Its guaranteed to
be 0 on the first call but will be valid otherwise.
Lets update it once while the device registers.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
This makes life easy and bug free. And is scalable for future resource
allocations.
Acked-by: Javi Merino <javi.merino@arm.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
Because get_cpu_frequency() has returned a valid frequency, it means that the
cpufreq policy is surely valid and so no point checking that again with
is_cpufreq_valid(). Get rid of the routine as well as there are no more users.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
All CPUs present in 'allowed_cpus' share the same 'struct cpufreq_policy'
structure and so calling cpufreq_update_policy() for each of them doesn't make
sense.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
In __cpufreq_cooling_register() we try to match min/max frequencies for all CPUs
passed in 'clip_cpus' mask. This mask is the cpumask of cpus where the frequency
constraints will be applied.
Same frequency constraint can be applied only to the CPUs belonging to the same
cluster (i.e. CPUs sharing clock line). For all such CPUs we have a single
'struct cpufreq_policy' structure managing them and so getting policies for all
CPUs wouldn't make any sense as they will all return the same pointer.
So, remove this useless check of checking min/max for all CPUs. Also update doc
comment to make this more obvious that clip_cpus should be same as
policy->related_cpus.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
We aren't supposed to return our own error type here. Return what we got.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
ret is initialized before it is used, so no need to set it to 0 in its declaration.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
It will be overwritten soon with return value of kzalloc().
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
This wasn't explained well anywhere and should be clearly specified. Lets add a
top level comment for this.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
cooling_cpufreq_lock isn't used to protect this structure and so the comment
over it is outdated. Fix it.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
cpufreq_cooling_register() expects mask of all the CPUs where frequency
constraint is applicable.
This platform has more than one CPU to which these constraints will apply and so
passing mask of only CPU0 wouldn't be sufficient. Also, this platform has a
single cluster of CPUs and the constraint applies to all CPUs.
If CPU0 is hoplugged out then we may face strange BUGs as cpu_cooling framework
isn't aware of any siblings sharing clock line.
Fix it by passing cpu_present_mask to cpufreq_cooling_register().
Cc: Chanwoo Choi <cw00.choi@samsung.com>
Cc: Kyungmin Park <kyungmin.park@samsung.com>
Cc: Amit Daniel Kachhap <amit.daniel@samsung.com>
Cc: Lukasz Majewski <l.majewski@samsung.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
cpufreq_cooling_register() expects mask of all the CPUs where frequency
constraint is applicable.
This platform has more than one CPU to which these constraints will apply and so
passing mask of only CPU0 wouldn't be sufficient. Also, this platform has a
single cluster of CPUs and the constraint applies to all CPUs.
If CPU0 is hoplugged out then we may face strange BUGs as cpu_cooling framework
isn't aware of any siblings sharing clock line.
Fix it by passing cpu_present_mask to cpufreq_cooling_register().
Cc: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
cpufreq_cooling_register() expects mask of all the CPUs where frequency
constraint is applicable.
This platform has more than one CPU to which these constraints will apply and so
passing mask of only CPU0 wouldn't be sufficient. Also, this platform has a
single cluster of CPUs and the constraint applies to all CPUs.
If CPU0 is hoplugged out then we may face strange BUGs as cpu_cooling framework
isn't aware of any siblings sharing clock line.
Fix it by passing cpu_present_mask to cpufreq_cooling_register().
Cc: Hongbo Zhang <hongbo.zhang@linaro.com>
Cc: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
In this patch, the cpu_cooling code checks for the usability of cpufreq
layer before proceeding with the CPU cooling device registration. The
main reason is: CPU cooling device is not usable if cpufreq cannot
switch frequencies.
Similar checks are spread in thermal drivers. Thus, the advantage now
is to have the check in a single place: cpu cooling device registration.
For this reason, this patch also updates the existing drivers that
depend on CPU cooling to simply propagate the error code of the cpu
cooling registration call. Therefore, in case cpufreq is not ready, the
thermal drivers will still return -EPROBE_DEFER, in an attempt to try
again when cpufreq layer gets ready.
Cc: devicetree@vger.kernel.org
Cc: Grant Likely <grant.likely@linaro.org>
Cc: Kukjin Kim <kgene.kim@samsung.com>
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-kernel@vger.kernel.org
Cc: linux-pm@vger.kernel.org
Cc: linux-samsung-soc@vger.kernel.org
Cc: Naveen Krishna Chatradhi <ch.naveen@samsung.com>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: Zhang Rui <rui.zhang@intel.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
If we run out of tags and have to sleep, we run the hardware queue
to kick pending IO into gear. During that run, we may have completed
requests, so re-check if we have free tags before going to sleep.
Signed-off-by: Jens Axboe <axboe@fb.com>
Avoid that if there are fewer hardware queues than CPU threads that
bt_get() can hang. The symptoms of the hang were as follows:
* All tags allocated for a particular hardware queue.
* (nr_tags) pending commands for that hardware queue.
* No pending commands for the software queues associated with that
hardware queue.
Signed-off-by: Jens Axboe <axboe@fb.com>
