The idea of runnable load average (let runnable time contribute to weight)
was proposed by Paul Turner and Ben Segall, and it is still followed by
this rewrite. This rewrite aims to solve the following issues:
1. cfs_rq's load average (namely runnable_load_avg and blocked_load_avg) is
updated at the granularity of an entity at a time, which results in the
cfs_rq's load average is stale or partially updated: at any time, only
one entity is up to date, all other entities are effectively lagging
behind. This is undesirable.
To illustrate, if we have n runnable entities in the cfs_rq, as time
elapses, they certainly become outdated:
t0: cfs_rq { e1_old, e2_old, ..., en_old }
and when we update:
t1: update e1, then we have cfs_rq { e1_new, e2_old, ..., en_old }
t2: update e2, then we have cfs_rq { e1_old, e2_new, ..., en_old }
...
We solve this by combining all runnable entities' load averages together
in cfs_rq's avg, and update the cfs_rq's avg as a whole. This is based
on the fact that if we regard the update as a function, then:
w * update(e) = update(w * e) and
update(e1) + update(e2) = update(e1 + e2), then
w1 * update(e1) + w2 * update(e2) = update(w1 * e1 + w2 * e2)
therefore, by this rewrite, we have an entirely updated cfs_rq at the
time we update it:
t1: update cfs_rq { e1_new, e2_new, ..., en_new }
t2: update cfs_rq { e1_new, e2_new, ..., en_new }
...
2. cfs_rq's load average is different between top rq->cfs_rq and other
task_group's per CPU cfs_rqs in whether or not blocked_load_average
contributes to the load.
The basic idea behind runnable load average (the same for utilization)
is that the blocked state is taken into account as opposed to only
accounting for the currently runnable state. Therefore, the average
should include both the runnable/running and blocked load averages.
This rewrite does that.
In addition, we also combine runnable/running and blocked averages
of all entities into the cfs_rq's average, and update it together at
once. This is based on the fact that:
update(runnable) + update(blocked) = update(runnable + blocked)
This significantly reduces the code as we don't need to separately
maintain/update runnable/running load and blocked load.
3. How task_group entities' share is calculated is complex and imprecise.
We reduce the complexity in this rewrite to allow a very simple rule:
the task_group's load_avg is aggregated from its per CPU cfs_rqs's
load_avgs. Then group entity's weight is simply proportional to its
own cfs_rq's load_avg / task_group's load_avg. To illustrate,
if a task_group has { cfs_rq1, cfs_rq2, ..., cfs_rqn }, then,
task_group_avg = cfs_rq1_avg + cfs_rq2_avg + ... + cfs_rqn_avg, then
cfs_rqx's entity's share = cfs_rqx_avg / task_group_avg * task_group's share
To sum up, this rewrite in principle is equivalent to the current one, but
fixes the issues described above. Turns out, it significantly reduces the
code complexity and hence increases clarity and efficiency. In addition,
the new averages are more smooth/continuous (no spurious spikes and valleys)
and updated more consistently and quickly to reflect the load dynamics.
As a result, we have less load tracking overhead, better performance,
and especially better power efficiency due to more balanced load.
Signed-off-by: Yuyang Du <yuyang.du@intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: arjan@linux.intel.com
Cc: bsegall@google.com
Cc: dietmar.eggemann@arm.com
Cc: fengguang.wu@intel.com
Cc: len.brown@intel.com
Cc: morten.rasmussen@arm.com
Cc: pjt@google.com
Cc: rafael.j.wysocki@intel.com
Cc: umgwanakikbuti@gmail.com
Cc: vincent.guittot@linaro.org
Link: http://lkml.kernel.org/r/1436918682-4971-3-git-send-email-yuyang.du@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull scheduler fixes from Ingo Molnar:
"Debug info and other statistics fixes and related enhancements"
* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched/numa: Fix numa balancing stats in /proc/pid/sched
sched/numa: Show numa_group ID in /proc/sched_debug task listings
sched/debug: Move print_cfs_rq() declaration to kernel/sched/sched.h
sched/stat: Expose /proc/pid/schedstat if CONFIG_SCHED_INFO=y
sched/stat: Simplify the sched_info accounting dependency
Commit 44dba3d5d6 ("sched: Refactor task_struct to use
numa_faults instead of numa_* pointers") modified the way
tsk->numa_faults stats are accounted.
However that commit never touched show_numa_stats() that is displayed
in /proc/pid/sched and thus the numbers displayed in /proc/pid/sched
don't match the actual numbers.
Fix it by making sure that /proc/pid/sched reflects the task
fault numbers. Also add group fault stats too.
Also couple of more modifications are added here:
1. Format changes:
- Previously we would list two entries per node, one for private
and one for shared. Also the home node info was listed in each entry.
- Now preferred node, total_faults and current node are
displayed separately.
- Now there is one entry per node, that lists private,shared task and
group faults.
2. Unit changes:
- p->numa_pages_migrated was getting reset after every read of
/proc/pid/sched. It's more useful to have absolute numbers since
differential migrations between two accesses can be more easily
calculated.
Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Acked-by: Rik van Riel <riel@redhat.com>
Cc: Iulia Manda <iulia.manda21@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1435252903-1081-4-git-send-email-srikar@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Currently print_cfs_rq() is declared in include/linux/sched.h.
However it's not used outside kernel/sched. Hence move the
declaration to kernel/sched/sched.h
Also some functions are only available for CONFIG_SCHED_DEBUG=y.
Hence move the declarations to within the #ifdef.
Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Acked-by: Rik van Riel <riel@redhat.com>
Cc: Iulia Manda <iulia.manda21@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1435252903-1081-2-git-send-email-srikar@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull scheduler updates from Thomas Gleixner:
"This series of scheduler updates depends on sched/core and timers/core
branches, which are already in your tree:
- Scheduler balancing overhaul to plug a hard to trigger race which
causes an oops in the balancer (Peter Zijlstra)
- Lockdep updates which are related to the balancing updates (Peter
Zijlstra)"
* 'sched-hrtimers-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched,lockdep: Employ lock pinning
lockdep: Implement lock pinning
lockdep: Simplify lock_release()
sched: Streamline the task migration locking a little
sched: Move code around
sched,dl: Fix sched class hopping CBS hole
sched, dl: Convert switched_{from, to}_dl() / prio_changed_dl() to balance callbacks
sched,dl: Remove return value from pull_dl_task()
sched, rt: Convert switched_{from, to}_rt() / prio_changed_rt() to balance callbacks
sched,rt: Remove return value from pull_rt_task()
sched: Allow balance callbacks for check_class_changed()
sched: Use replace normalize_task() with __sched_setscheduler()
sched: Replace post_schedule with a balance callback list
Pull timer updates from Thomas Gleixner:
"A rather largish update for everything time and timer related:
- Cache footprint optimizations for both hrtimers and timer wheel
- Lower the NOHZ impact on systems which have NOHZ or timer migration
disabled at runtime.
- Optimize run time overhead of hrtimer interrupt by making the clock
offset updates smarter
- hrtimer cleanups and removal of restrictions to tackle some
problems in sched/perf
- Some more leap second tweaks
- Another round of changes addressing the 2038 problem
- First step to change the internals of clock event devices by
introducing the necessary infrastructure
- Allow constant folding for usecs/msecs_to_jiffies()
- The usual pile of clockevent/clocksource driver updates
The hrtimer changes contain updates to sched, perf and x86 as they
depend on them plus changes all over the tree to cleanup API changes
and redundant code, which got copied all over the place. The y2038
changes touch s390 to remove the last non 2038 safe code related to
boot/persistant clock"
* 'timers-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (114 commits)
clocksource: Increase dependencies of timer-stm32 to limit build wreckage
timer: Minimize nohz off overhead
timer: Reduce timer migration overhead if disabled
timer: Stats: Simplify the flags handling
timer: Replace timer base by a cpu index
timer: Use hlist for the timer wheel hash buckets
timer: Remove FIFO "guarantee"
timers: Sanitize catchup_timer_jiffies() usage
hrtimer: Allow hrtimer::function() to free the timer
seqcount: Introduce raw_write_seqcount_barrier()
seqcount: Rename write_seqcount_barrier()
hrtimer: Fix hrtimer_is_queued() hole
hrtimer: Remove HRTIMER_STATE_MIGRATE
selftest: Timers: Avoid signal deadlock in leap-a-day
timekeeping: Copy the shadow-timekeeper over the real timekeeper last
clockevents: Check state instead of mode in suspend/resume path
selftests: timers: Add leap-second timer edge testing to leap-a-day.c
ntp: Do leapsecond adjustment in adjtimex read path
time: Prevent early expiry of hrtimers[CLOCK_REALTIME] at the leap second edge
ntp: Introduce and use SECS_PER_DAY macro instead of 86400
...
There are two init_sched_dl_class() declarations, this patch drops
the duplicate.
Signed-off-by: Wanpeng Li <wanpeng.li@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Juri Lelli <juri.lelli@arm.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1431496867-4194-5-git-send-email-wanpeng.li@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
In the below two commits (see Fixes) we have periodic timers that can
stop themselves when they're no longer required, but need to be
(re)-started when their idle condition changes.
Further complications is that we want the timer handler to always do
the forward such that it will always correctly deal with the overruns,
and we do not want to race such that the handler has already decided
to stop, but the (external) restart sees the timer still active and we
end up with a 'lost' timer.
The problem with the current code is that the re-start can come before
the callback does the forward, at which point the forward from the
callback will WARN about forwarding an enqueued timer.
Now, conceptually its easy to detect if you're before or after the fwd
by comparing the expiration time against the current time. Of course,
that's expensive (and racy) because we don't have the current time.
Alternatively one could cache this state inside the timer, but then
everybody pays the overhead of maintaining this extra state, and that
is undesired.
The only other option that I could see is the external timer_active
variable, which I tried to kill before. I would love a nicer interface
for this seemingly simple 'problem' but alas.
Fixes: 272325c482 ("perf: Fix mux_interval hrtimer wreckage")
Fixes: 77a4d1a1b9 ("sched: Cleanup bandwidth timers")
Cc: pjt@google.com
Cc: tglx@linutronix.de
Cc: klamm@yandex-team.ru
Cc: mingo@kernel.org
Cc: bsegall@google.com
Cc: hpa@zytor.com
Cc: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20150514102311.GX21418@twins.programming.kicks-ass.net
ACCESS_ONCE doesn't work reliably on non-scalar types. This patch removes
the rest of the existing usages of ACCESS_ONCE() in the scheduler, and use
the new READ_ONCE() and WRITE_ONCE() APIs as appropriate.
Signed-off-by: Jason Low <jason.low2@hp.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Rik van Riel <riel@redhat.com>
Acked-by: Waiman Long <Waiman.Long@hp.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Aswin Chandramouleeswaran <aswin@hp.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Mike Galbraith <umgwanakikbuti@gmail.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Preeti U Murthy <preeti@linux.vnet.ibm.com>
Cc: Scott J Norton <scott.norton@hp.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Link: http://lkml.kernel.org/r/1430251224-5764-2-git-send-email-jason.low2@hp.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
I could not find the loadavg code.. turns out it was hidden in a file
called proc.c. It further got mingled up with the cruft per rq load
indexes (which we really want to get rid of).
Move the per rq load indexes into the fair.c load-balance code (that's
the only thing that uses them) and rename proc.c to loadavg.c so we
can find it again.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Paul Gortmaker <paul.gortmaker@windriver.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
[ Did minor cleanups to the code. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Roman reported a 3 cpu lockup scenario involving __start_cfs_bandwidth().
The more I look at that code the more I'm convinced its crack, that
entire __start_cfs_bandwidth() thing is brain melting, we don't need to
cancel a timer before starting it, *hrtimer_start*() will happily remove
the timer for you if its still enqueued.
Removing that, removes a big part of the problem, no more ugly cancel
loop to get stuck in.
So now, if I understand things right, the entire reason you have this
cfs_b->lock guarded ->timer_active nonsense is to make sure we don't
accidentally lose the timer.
It appears to me that it should be possible to guarantee that same by
unconditionally (re)starting the timer when !queued. Because regardless
what hrtimer::function will return, if we beat it to (re)enqueue the
timer, it doesn't matter.
Now, because hrtimers don't come with any serialization guarantees we
must ensure both handler and (re)start loop serialize their access to
the hrtimer to avoid both trying to forward the timer at the same
time.
Update the rt bandwidth timer to match.
This effectively reverts: 09dc4ab039 ("sched/fair: Fix
tg_set_cfs_bandwidth() deadlock on rq->lock").
Reported-by: Roman Gushchin <klamm@yandex-team.ru>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Ben Segall <bsegall@google.com>
Cc: Paul Turner <pjt@google.com>
Link: http://lkml.kernel.org/r/20150415095011.804589208@infradead.org
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Obviously, 'rq' is not used in these two functions, therefore,
there is no reason for it to be passed as an argument.
Signed-off-by: Abel Vesa <abelvesa@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1425383427-26244-1-git-send-email-abelvesa@gmail.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Currently the freq invariant accounting (in
__update_entity_runnable_avg() and sched_rt_avg_update()) get the
scale factor from a weak function call, this means that even for archs
that default on their implementation the compiler cannot see into this
function and optimize the extra scaling math away.
This is sad, esp. since its a 64-bit multiplication which can be quite
costly on some platforms.
So replace the weak function with #ifdef and __always_inline goo. This
is not quite as nice from an arch support PoV but should at least
result in compile time errors if done wrong.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Ben Segall <bsegall@google.com>
Cc: Morten.Rasmussen@arm.com
Cc: Paul Turner <pjt@google.com>
Cc: Vincent Guittot <vincent.guittot@linaro.org>
Cc: dietmar.eggemann@arm.com
Cc: efault@gmx.de
Cc: kamalesh@linux.vnet.ibm.com
Cc: nicolas.pitre@linaro.org
Cc: preeti@linux.vnet.ibm.com
Cc: riel@redhat.com
Link: http://lkml.kernel.org/r/20150323131905.GF23123@twins.programming.kicks-ass.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This new field 'cpu_capacity_orig' reflects the original capacity of a CPU
before being altered by rt tasks and/or IRQ
The cpu_capacity_orig will be used:
- to detect when the capacity of a CPU has been noticeably reduced so we can
trig load balance to look for a CPU with better capacity. As an example, we
can detect when a CPU handles a significant amount of irq
(with CONFIG_IRQ_TIME_ACCOUNTING) but this CPU is seen as an idle CPU by
scheduler whereas CPUs, which are really idle, are available.
- evaluate the available capacity for CFS tasks
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Kamalesh Babulal <kamalesh@linux.vnet.ibm.com>
Acked-by: Morten Rasmussen <morten.rasmussen@arm.com>
Cc: Morten.Rasmussen@arm.com
Cc: dietmar.eggemann@arm.com
Cc: efault@gmx.de
Cc: linaro-kernel@lists.linaro.org
Cc: nicolas.pitre@linaro.org
Cc: preeti@linux.vnet.ibm.com
Cc: riel@redhat.com
Link: http://lkml.kernel.org/r/1425052454-25797-7-git-send-email-vincent.guittot@linaro.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The average running time of RT tasks is used to estimate the remaining compute
capacity for CFS tasks. This remaining capacity is the original capacity scaled
down by a factor (aka scale_rt_capacity). This estimation of available capacity
must also be invariant with frequency scaling.
A frequency scaling factor is applied on the running time of the RT tasks for
computing scale_rt_capacity.
In sched_rt_avg_update(), we now scale the RT execution time like below:
rq->rt_avg += rt_delta * arch_scale_freq_capacity() >> SCHED_CAPACITY_SHIFT
Then, scale_rt_capacity can be summarized by:
scale_rt_capacity = SCHED_CAPACITY_SCALE * available / total
with available = total - rq->rt_avg
This has been been optimized in current code by:
scale_rt_capacity = available / (total >> SCHED_CAPACITY_SHIFT)
But we can also developed the equation like below:
scale_rt_capacity = SCHED_CAPACITY_SCALE - ((rq->rt_avg << SCHED_CAPACITY_SHIFT) / total)
and we can optimize the equation by removing SCHED_CAPACITY_SHIFT shift in
the computation of rq->rt_avg and scale_rt_capacity().
so rq->rt_avg += rt_delta * arch_scale_freq_capacity()
and
scale_rt_capacity = SCHED_CAPACITY_SCALE - (rq->rt_avg / total)
arch_scale_frequency_capacity() will be called in the hot path of the scheduler
which implies to have a short and efficient function.
As an example, arch_scale_frequency_capacity() should return a cached value that
is updated periodically outside of the hot path.
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Morten Rasmussen <morten.rasmussen@arm.com>
Cc: Morten.Rasmussen@arm.com
Cc: dietmar.eggemann@arm.com
Cc: efault@gmx.de
Cc: kamalesh@linux.vnet.ibm.com
Cc: linaro-kernel@lists.linaro.org
Cc: nicolas.pitre@linaro.org
Cc: preeti@linux.vnet.ibm.com
Cc: riel@redhat.com
Link: http://lkml.kernel.org/r/1425052454-25797-6-git-send-email-vincent.guittot@linaro.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Add new statistics which reflect the average time a task is running on the CPU
and the sum of these running time of the tasks on a runqueue. The latter is
named utilization_load_avg.
This patch is based on the usage metric that was proposed in the 1st
versions of the per-entity load tracking patchset by Paul Turner
<pjt@google.com> but that has be removed afterwards. This version differs from
the original one in the sense that it's not linked to task_group.
The rq's utilization_load_avg will be used to check if a rq is overloaded or
not instead of trying to compute how many tasks a group of CPUs can handle.
Rename runnable_avg_period into avg_period as it is now used with both
runnable_avg_sum and running_avg_sum.
Add some descriptions of the variables to explain their differences.
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Morten Rasmussen <morten.rasmussen@arm.com>
Cc: Paul Turner <pjt@google.com>
Cc: Ben Segall <bsegall@google.com>
Cc: Ben Segall <bsegall@google.com>
Cc: Morten.Rasmussen@arm.com
Cc: Paul Turner <pjt@google.com>
Cc: dietmar.eggemann@arm.com
Cc: efault@gmx.de
Cc: kamalesh@linux.vnet.ibm.com
Cc: linaro-kernel@lists.linaro.org
Cc: nicolas.pitre@linaro.org
Cc: preeti@linux.vnet.ibm.com
Cc: riel@redhat.com
Link: http://lkml.kernel.org/r/1425052454-25797-2-git-send-email-vincent.guittot@linaro.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
When debugging the latencies on a 40 core box, where we hit 300 to
500 microsecond latencies, I found there was a huge contention on the
runqueue locks.
Investigating it further, running ftrace, I found that it was due to
the pulling of RT tasks.
The test that was run was the following:
cyclictest --numa -p95 -m -d0 -i100
This created a thread on each CPU, that would set its wakeup in iterations
of 100 microseconds. The -d0 means that all the threads had the same
interval (100us). Each thread sleeps for 100us and wakes up and measures
its latencies.
cyclictest is maintained at:
git://git.kernel.org/pub/scm/linux/kernel/git/clrkwllms/rt-tests.git
What happened was another RT task would be scheduled on one of the CPUs
that was running our test, when the other CPU tests went to sleep and
scheduled idle. This caused the "pull" operation to execute on all
these CPUs. Each one of these saw the RT task that was overloaded on
the CPU of the test that was still running, and each one tried
to grab that task in a thundering herd way.
To grab the task, each thread would do a double rq lock grab, grabbing
its own lock as well as the rq of the overloaded CPU. As the sched
domains on this box was rather flat for its size, I saw up to 12 CPUs
block on this lock at once. This caused a ripple affect with the
rq locks especially since the taking was done via a double rq lock, which
means that several of the CPUs had their own rq locks held while trying
to take this rq lock. As these locks were blocked, any wakeups or load
balanceing on these CPUs would also block on these locks, and the wait
time escalated.
I've tried various methods to lessen the load, but things like an
atomic counter to only let one CPU grab the task wont work, because
the task may have a limited affinity, and we may pick the wrong
CPU to take that lock and do the pull, to only find out that the
CPU we picked isn't in the task's affinity.
Instead of doing the PULL, I now have the CPUs that want the pull to
send over an IPI to the overloaded CPU, and let that CPU pick what
CPU to push the task to. No more need to grab the rq lock, and the
push/pull algorithm still works fine.
With this patch, the latency dropped to just 150us over a 20 hour run.
Without the patch, the huge latencies would trigger in seconds.
I've created a new sched feature called RT_PUSH_IPI, which is enabled
by default.
When RT_PUSH_IPI is not enabled, the old method of grabbing the rq locks
and having the pulling CPU do the work is implemented. When RT_PUSH_IPI
is enabled, the IPI is sent to the overloaded CPU to do a push.
To enabled or disable this at run time:
# mount -t debugfs nodev /sys/kernel/debug
# echo RT_PUSH_IPI > /sys/kernel/debug/sched_features
or
# echo NO_RT_PUSH_IPI > /sys/kernel/debug/sched_features
Update: This original patch would send an IPI to all CPUs in the RT overload
list. But that could theoretically cause the reverse issue. That is, there
could be lots of overloaded RT queues and one CPU lowers its priority. It would
then send an IPI to all the overloaded RT queues and they could then all try
to grab the rq lock of the CPU lowering its priority, and then we have the
same problem.
The latest design sends out only one IPI to the first overloaded CPU. It tries to
push any tasks that it can, and then looks for the next overloaded CPU that can
push to the source CPU. The IPIs stop when all overloaded CPUs that have pushable
tasks that have priorities greater than the source CPU are covered. In case the
source CPU lowers its priority again, a flag is set to tell the IPI traversal to
restart with the first RT overloaded CPU after the source CPU.
Parts-suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Joern Engel <joern@purestorage.com>
Cc: Clark Williams <williams@redhat.com>
Cc: Mike Galbraith <umgwanakikbuti@gmail.com>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20150318144946.2f3cc982@gandalf.local.home
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Kirill reported that a dl task can be throttled and dequeued at the
same time. This happens, when it becomes throttled in schedule(),
which is called to go to sleep:
current->state = TASK_INTERRUPTIBLE;
schedule()
deactivate_task()
dequeue_task_dl()
update_curr_dl()
start_dl_timer()
__dequeue_task_dl()
prev->on_rq = 0;
This invalidates the assumption from commit 0f397f2c90 ("sched/dl:
Fix race in dl_task_timer()"):
"The only reason we don't strictly need ->pi_lock now is because
we're guaranteed to have p->state == TASK_RUNNING here and are
thus free of ttwu races".
And therefore we have to use the full task_rq_lock() here.
This further amends the fact that we forgot to update the rq lock loop
for TASK_ON_RQ_MIGRATE, from commit cca26e8009 ("sched: Teach
scheduler to understand TASK_ON_RQ_MIGRATING state").
Reported-by: Kirill Tkhai <ktkhai@parallels.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Juri Lelli <juri.lelli@arm.com>
Link: http://lkml.kernel.org/r/20150217123139.GN5029@twins.programming.kicks-ass.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The original purpose of rq::skip_clock_update was to avoid 'costly' clock
updates for back to back wakeup-preempt pairs. The big problem with it
has always been that the rq variable is unaware of the context and
causes indiscrimiate clock skips.
Rework the entire thing and create a sense of context by only allowing
schedule() to skip clock updates. (XXX can we measure the cost of the
added store?)
By ensuring only schedule can ever skip an update, we guarantee we're
never more than 1 tick behind on the update.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: umgwanakikbuti@gmail.com
Link: http://lkml.kernel.org/r/20150105103554.432381549@infradead.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
rq->clock{,_task} are serialized by rq->lock, verify this.
One immediate fail is the usage in scale_rt_capability, so 'annotate'
that for now, there's more 'funny' there. Maybe change rq->lock into a
raw_seqlock_t?
(Only 32-bit is affected)
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20150105103554.361872747@infradead.org
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: umgwanakikbuti@gmail.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Commit d670ec1317 "posix-cpu-timers: Cure SMP wobbles" fixes one glibc
test case in cost of breaking another one. After that commit, calling
clock_nanosleep(TIMER_ABSTIME, X) and then clock_gettime(&Y) can result
of Y time being smaller than X time.
Reproducer/tester can be found further below, it can be compiled and ran by:
gcc -o tst-cpuclock2 tst-cpuclock2.c -pthread
while ./tst-cpuclock2 ; do : ; done
This reproducer, when running on a buggy kernel, will complain
about "clock_gettime difference too small".
Issue happens because on start in thread_group_cputimer() we initialize
sum_exec_runtime of cputimer with threads runtime not yet accounted and
then add the threads runtime to running cputimer again on scheduler
tick, making it's sum_exec_runtime bigger than actual threads runtime.
KOSAKI Motohiro posted a fix for this problem, but that patch was never
applied: https://lkml.org/lkml/2013/5/26/191 .
This patch takes different approach to cure the problem. It calls
update_curr() when cputimer starts, that assure we will have updated
stats of running threads and on the next schedule tick we will account
only the runtime that elapsed from cputimer start. That also assure we
have consistent state between cpu times of individual threads and cpu
time of the process consisted by those threads.
Full reproducer (tst-cpuclock2.c):
#define _GNU_SOURCE
#include <unistd.h>
#include <sys/syscall.h>
#include <stdio.h>
#include <time.h>
#include <pthread.h>
#include <stdint.h>
#include <inttypes.h>
/* Parameters for the Linux kernel ABI for CPU clocks. */
#define CPUCLOCK_SCHED 2
#define MAKE_PROCESS_CPUCLOCK(pid, clock) \
((~(clockid_t) (pid) << 3) | (clockid_t) (clock))
static pthread_barrier_t barrier;
/* Help advance the clock. */
static void *chew_cpu(void *arg)
{
pthread_barrier_wait(&barrier);
while (1) ;
return NULL;
}
/* Don't use the glibc wrapper. */
static int do_nanosleep(int flags, const struct timespec *req)
{
clockid_t clock_id = MAKE_PROCESS_CPUCLOCK(0, CPUCLOCK_SCHED);
return syscall(SYS_clock_nanosleep, clock_id, flags, req, NULL);
}
static int64_t tsdiff(const struct timespec *before, const struct timespec *after)
{
int64_t before_i = before->tv_sec * 1000000000ULL + before->tv_nsec;
int64_t after_i = after->tv_sec * 1000000000ULL + after->tv_nsec;
return after_i - before_i;
}
int main(void)
{
int result = 0;
pthread_t th;
pthread_barrier_init(&barrier, NULL, 2);
if (pthread_create(&th, NULL, chew_cpu, NULL) != 0) {
perror("pthread_create");
return 1;
}
pthread_barrier_wait(&barrier);
/* The test. */
struct timespec before, after, sleeptimeabs;
int64_t sleepdiff, diffabs;
const struct timespec sleeptime = {.tv_sec = 0,.tv_nsec = 100000000 };
/* The relative nanosleep. Not sure why this is needed, but its presence
seems to make it easier to reproduce the problem. */
if (do_nanosleep(0, &sleeptime) != 0) {
perror("clock_nanosleep");
return 1;
}
/* Get the current time. */
if (clock_gettime(CLOCK_PROCESS_CPUTIME_ID, &before) < 0) {
perror("clock_gettime[2]");
return 1;
}
/* Compute the absolute sleep time based on the current time. */
uint64_t nsec = before.tv_nsec + sleeptime.tv_nsec;
sleeptimeabs.tv_sec = before.tv_sec + nsec / 1000000000;
sleeptimeabs.tv_nsec = nsec % 1000000000;
/* Sleep for the computed time. */
if (do_nanosleep(TIMER_ABSTIME, &sleeptimeabs) != 0) {
perror("absolute clock_nanosleep");
return 1;
}
/* Get the time after the sleep. */
if (clock_gettime(CLOCK_PROCESS_CPUTIME_ID, &after) < 0) {
perror("clock_gettime[3]");
return 1;
}
/* The time after sleep should always be equal to or after the absolute sleep
time passed to clock_nanosleep. */
sleepdiff = tsdiff(&sleeptimeabs, &after);
if (sleepdiff < 0) {
printf("absolute clock_nanosleep woke too early: %" PRId64 "\n", sleepdiff);
result = 1;
printf("Before %llu.%09llu\n", before.tv_sec, before.tv_nsec);
printf("After %llu.%09llu\n", after.tv_sec, after.tv_nsec);
printf("Sleep %llu.%09llu\n", sleeptimeabs.tv_sec, sleeptimeabs.tv_nsec);
}
/* The difference between the timestamps taken before and after the
clock_nanosleep call should be equal to or more than the duration of the
sleep. */
diffabs = tsdiff(&before, &after);
if (diffabs < sleeptime.tv_nsec) {
printf("clock_gettime difference too small: %" PRId64 "\n", diffabs);
result = 1;
}
pthread_cancel(th);
return result;
}
Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/20141112155843.GA24803@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This patch simplifies task_struct by removing the four numa_* pointers
in the same array and replacing them with the array pointer. By doing this,
on x86_64, the size of task_struct is reduced by 3 ulong pointers (24 bytes on
x86_64).
A new parameter is added to the task_faults_idx function so that it can return
an index to the correct offset, corresponding with the old precalculated
pointers.
All of the code in sched/ that depended on task_faults_idx and numa_* was
changed in order to match the new logic.
Signed-off-by: Iulia Manda <iulia.manda21@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: mgorman@suse.de
Cc: dave@stgolabs.net
Cc: riel@redhat.com
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/20141031001331.GA30662@winterfell
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Currently used hrtimer_try_to_cancel() is racy:
raw_spin_lock(&rq->lock)
... dl_task_timer raw_spin_lock(&rq->lock)
... raw_spin_lock(&rq->lock) ...
switched_from_dl() ... ...
hrtimer_try_to_cancel() ... ...
switched_to_fair() ... ...
... ... ...
... ... ...
raw_spin_unlock(&rq->lock) ... (asquired)
... ... ...
... ... ...
do_exit() ... ...
schedule() ... ...
raw_spin_lock(&rq->lock) ... raw_spin_unlock(&rq->lock)
... ... ...
raw_spin_unlock(&rq->lock) ... raw_spin_lock(&rq->lock)
... ... (asquired)
put_task_struct() ... ...
free_task_struct() ... ...
... ... raw_spin_unlock(&rq->lock)
... (asquired) ...
... ... ...
... (use after free) ...
So, let's implement 100% guaranteed way to cancel the timer and let's
be sure we are safe even in very unlikely situations.
rq unlocking does not limit the area of switched_from_dl() use, because
this has already been possible in pull_dl_task() below.
Let's consider the safety of of this unlocking. New code in the patch
is working when hrtimer_try_to_cancel() fails. This means the callback
is running. In this case hrtimer_cancel() is just waiting till the
callback is finished. Two
1) Since we are in switched_from_dl(), new class is not dl_sched_class and
new prio is not less MAX_DL_PRIO. So, the callback returns early; it's
right after !dl_task() check. After that hrtimer_cancel() returns back too.
The above is:
raw_spin_lock(rq->lock); ...
... dl_task_timer()
... raw_spin_lock(rq->lock);
switched_from_dl() ...
hrtimer_try_to_cancel() ...
raw_spin_unlock(rq->lock); ...
hrtimer_cancel() ...
... raw_spin_unlock(rq->lock);
... return HRTIMER_NORESTART;
... ...
raw_spin_lock(rq->lock); ...
2) But the below is also possible:
dl_task_timer()
raw_spin_lock(rq->lock);
...
raw_spin_unlock(rq->lock);
raw_spin_lock(rq->lock); ...
switched_from_dl() ...
hrtimer_try_to_cancel() ...
... return HRTIMER_NORESTART;
raw_spin_unlock(rq->lock); ...
hrtimer_cancel(); ...
raw_spin_lock(rq->lock); ...
In this case hrtimer_cancel() returns immediately. Very unlikely case,
just to mention.
Nobody can manipulate the task, because check_class_changed() is
always called with pi_lock locked. Nobody can force the task to
participate in (concurrent) priority inheritance schemes (the same reason).
All concurrent task operations require pi_lock, which is held by us.
No deadlocks with dl_task_timer() are possible, because it returns
right after !dl_task() check (it does nothing).
If we receive a new dl_task during the time of unlocked rq, we just
don't have to do pull_dl_task() in switched_from_dl() further.
Signed-off-by: Kirill Tkhai <ktkhai@parallels.com>
[ Added comments]
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Juri Lelli <juri.lelli@arm.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1414420852.19914.186.camel@tkhai
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Exclusive cpusets are the only way users can restrict SCHED_DEADLINE tasks
affinity (performing what is commonly called clustered scheduling).
Unfortunately, such thing is currently broken for two reasons:
- No check is performed when the user tries to attach a task to
an exlusive cpuset (recall that exclusive cpusets have an
associated maximum allowed bandwidth).
- Bandwidths of source and destination cpusets are not correctly
updated after a task is migrated between them.
This patch fixes both things at once, as they are opposite faces
of the same coin.
The check is performed in cpuset_can_attach(), as there aren't any
points of failure after that function. The updated is split in two
halves. We first reserve bandwidth in the destination cpuset, after
we pass the check in cpuset_can_attach(). And we then release
bandwidth from the source cpuset when the task's affinity is
actually changed. Even if there can be time windows when sched_setattr()
may erroneously fail in the source cpuset, we are fine with it, as
we can't perfom an atomic update of both cpusets at once.
Reported-by: Daniel Wagner <daniel.wagner@bmw-carit.de>
Reported-by: Vincent Legout <vincent@legout.info>
Signed-off-by: Juri Lelli <juri.lelli@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Dario Faggioli <raistlin@linux.it>
Cc: Michael Trimarchi <michael@amarulasolutions.com>
Cc: Fabio Checconi <fchecconi@gmail.com>
Cc: michael@amarulasolutions.com
Cc: luca.abeni@unitn.it
Cc: Li Zefan <lizefan@huawei.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: cgroups@vger.kernel.org
Link: http://lkml.kernel.org/r/1411118561-26323-3-git-send-email-juri.lelli@arm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Smaller NUMA systems tend to have all NUMA nodes directly connected
to each other. This includes the degenerate case of a system with just
one node, ie. a non-NUMA system.
Larger systems can have two kinds of NUMA topology, which affects how
tasks and memory should be placed on the system.
On glueless mesh systems, nodes that are not directly connected to
each other will bounce traffic through intermediary nodes. Task groups
can be run closer to each other by moving tasks from a node to an
intermediary node between it and the task's preferred node.
On NUMA systems with backplane controllers, the intermediary hops
are incapable of running programs. This creates "islands" of nodes
that are at an equal distance to anywhere else in the system.
Each kind of topology requires a slightly different placement
algorithm; this patch provides the mechanism to detect the kind
of NUMA topology of a system.
Signed-off-by: Rik van Riel <riel@redhat.com>
Tested-by: Chegu Vinod <chegu_vinod@hp.com>
[ Changed to use kernel/sched/sched.h ]
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: mgorman@suse.de
Cc: chegu_vinod@hp.com
Link: http://lkml.kernel.org/r/1413530994-9732-3-git-send-email-riel@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Export some information that is necessary to do placement of
tasks on systems with multi-level NUMA topologies.
Signed-off-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: mgorman@suse.de
Cc: chegu_vinod@hp.com
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1413530994-9732-2-git-send-email-riel@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull percpu consistent-ops changes from Tejun Heo:
"Way back, before the current percpu allocator was implemented, static
and dynamic percpu memory areas were allocated and handled separately
and had their own accessors. The distinction has been gone for many
years now; however, the now duplicate two sets of accessors remained
with the pointer based ones - this_cpu_*() - evolving various other
operations over time. During the process, we also accumulated other
inconsistent operations.
This pull request contains Christoph's patches to clean up the
duplicate accessor situation. __get_cpu_var() uses are replaced with
with this_cpu_ptr() and __this_cpu_ptr() with raw_cpu_ptr().
Unfortunately, the former sometimes is tricky thanks to C being a bit
messy with the distinction between lvalues and pointers, which led to
a rather ugly solution for cpumask_var_t involving the introduction of
this_cpu_cpumask_var_ptr().
This converts most of the uses but not all. Christoph will follow up
with the remaining conversions in this merge window and hopefully
remove the obsolete accessors"
* 'for-3.18-consistent-ops' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/percpu: (38 commits)
irqchip: Properly fetch the per cpu offset
percpu: Resolve ambiguities in __get_cpu_var/cpumask_var_t -fix
ia64: sn_nodepda cannot be assigned to after this_cpu conversion. Use __this_cpu_write.
percpu: Resolve ambiguities in __get_cpu_var/cpumask_var_t
Revert "powerpc: Replace __get_cpu_var uses"
percpu: Remove __this_cpu_ptr
clocksource: Replace __this_cpu_ptr with raw_cpu_ptr
sparc: Replace __get_cpu_var uses
avr32: Replace __get_cpu_var with __this_cpu_write
blackfin: Replace __get_cpu_var uses
tile: Use this_cpu_ptr() for hardware counters
tile: Replace __get_cpu_var uses
powerpc: Replace __get_cpu_var uses
alpha: Replace __get_cpu_var
ia64: Replace __get_cpu_var uses
s390: cio driver &__get_cpu_var replacements
s390: Replace __get_cpu_var uses
mips: Replace __get_cpu_var uses
MIPS: Replace __get_cpu_var uses in FPU emulator.
arm: Replace __this_cpu_ptr with raw_cpu_ptr
...
When the cpu enters idle, it stores the cpuidle state pointer in its
struct rq instance which in turn could be used to make a better decision
when balancing tasks.
As soon as the cpu exits its idle state, the struct rq reference is
cleared.
There are a couple of situations where the idle state pointer could be changed
while it is being consulted:
1. For x86/acpi with dynamic c-states, when a laptop switches from battery
to AC that could result on removing the deeper idle state. The acpi driver
triggers:
'acpi_processor_cst_has_changed'
'cpuidle_pause_and_lock'
'cpuidle_uninstall_idle_handler'
'kick_all_cpus_sync'.
All cpus will exit their idle state and the pointed object will be set to
NULL.
2. The cpuidle driver is unloaded. Logically that could happen but not
in practice because the drivers are always compiled in and 95% of them are
not coded to unregister themselves. In any case, the unloading code must
call 'cpuidle_unregister_device', that calls 'cpuidle_pause_and_lock'
leading to 'kick_all_cpus_sync' as mentioned above.
A race can happen if we use the pointer and then one of these two scenarios
occurs at the same moment.
In order to be safe, the idle state pointer stored in the rq must be
used inside a rcu_read_lock section where we are protected with the
'rcu_barrier' in the 'cpuidle_uninstall_idle_handler' function. The
idle_get_state() and idle_put_state() accessors should be used to that
effect.
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
Cc: linux-pm@vger.kernel.org
Cc: linaro-kernel@lists.linaro.org
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/n/tip-@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
When a task is using SCHED_DEADLINE and the user setschedules it to a
different class its sched_dl_entity static parameters are not cleaned
up. This causes a bug if the user sets it back to SCHED_DEADLINE with
the same parameters again. The problem resides in the check we
perform at the very beginning of dl_overflow():
if (new_bw == p->dl.dl_bw)
return 0;
This condition is met in the case depicted above, so the function
returns and dl_b->total_bw is not updated (the p->dl.dl_bw is not
added to it). After this, admission control is broken.
This patch fixes the thing, properly clearing static parameters for a
task that ceases to use SCHED_DEADLINE.
Reported-by: Daniele Alessandrelli <daniele.alessandrelli@gmail.com>
Reported-by: Daniel Wagner <daniel.wagner@bmw-carit.de>
Reported-by: Vincent Legout <vincent@legout.info>
Tested-by: Luca Abeni <luca.abeni@unitn.it>
Tested-by: Daniel Wagner <daniel.wagner@bmw-carit.de>
Tested-by: Vincent Legout <vincent@legout.info>
Signed-off-by: Juri Lelli <juri.lelli@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Fabio Checconi <fchecconi@gmail.com>
Cc: Dario Faggioli <raistlin@linux.it>
Cc: Michael Trimarchi <michael@amarulasolutions.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1411118561-26323-2-git-send-email-juri.lelli@arm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Convert all uses of __get_cpu_var for address calculation to use
this_cpu_ptr instead.
[Uses of __get_cpu_var with cpumask_var_t are no longer
handled by this patch]
Cc: Peter Zijlstra <peterz@infradead.org>
Acked-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
This is a new p->on_rq state which will be used to indicate that a task
is in a process of migrating between two RQs. It allows to get
rid of double_rq_lock(), which we used to use to change a rq of
a queued task before.
Let's consider an example. To move a task between src_rq and
dst_rq we will do the following:
raw_spin_lock(&src_rq->lock);
/* p is a task which is queued on src_rq */
p = ...;
dequeue_task(src_rq, p, 0);
p->on_rq = TASK_ON_RQ_MIGRATING;
set_task_cpu(p, dst_cpu);
raw_spin_unlock(&src_rq->lock);
/*
* Both RQs are unlocked here.
* Task p is dequeued from src_rq
* but its on_rq value is not zero.
*/
raw_spin_lock(&dst_rq->lock);
p->on_rq = TASK_ON_RQ_QUEUED;
enqueue_task(dst_rq, p, 0);
raw_spin_unlock(&dst_rq->lock);
While p->on_rq is TASK_ON_RQ_MIGRATING, task is considered as
"migrating", and other parallel scheduler actions with it are
not available to parallel callers. The parallel caller is
spining till migration is completed.
The unavailable actions are changing of cpu affinity, changing
of priority etc, in other words all the functionality which used
to require task_rq(p)->lock before (and related to the task).
To implement TASK_ON_RQ_MIGRATING support we primarily are using
the following fact. Most of scheduler users (from which we are
protecting a migrating task) use task_rq_lock() and
__task_rq_lock() to get the lock of task_rq(p). These primitives
know that task's cpu may change, and they are spining while the
lock of the right RQ is not held. We add one more condition into
them, so they will be also spinning until the migration is
finished.
Signed-off-by: Kirill Tkhai <ktkhai@parallels.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Paul Turner <pjt@google.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Mike Galbraith <umgwanakikbuti@gmail.com>
Cc: Kirill Tkhai <tkhai@yandex.ru>
Cc: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Nicolas Pitre <nicolas.pitre@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1408528062.23412.88.camel@tkhai
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Implement task_on_rq_queued() and use it everywhere instead of
on_rq check. No functional changes.
The only exception is we do not use the wrapper in
check_for_tasks(), because it requires to export
task_on_rq_queued() in global header files. Next patch in series
would return it back, so we do not twist it from here to there.
Signed-off-by: Kirill Tkhai <ktkhai@parallels.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Paul Turner <pjt@google.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Mike Galbraith <umgwanakikbuti@gmail.com>
Cc: Kirill Tkhai <tkhai@yandex.ru>
Cc: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Nicolas Pitre <nicolas.pitre@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1408528052.23412.87.camel@tkhai
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Match the declaration of runqueues with the definition.
Signed-off-by: Pranith Kumar <bobby.prani@gmail.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/1407950893-32731-1-git-send-email-bobby.prani@gmail.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
We always use resched_task() with rq->curr argument.
It's not possible to reschedule any task but rq's current.
The patch introduces resched_curr(struct rq *) to
replace all of the repeating patterns. The main aim
is cleanup, but there is a little size profit too:
(before)
$ size kernel/sched/built-in.o
text data bss dec hex filename
155274 16445 7042 178761 2ba49 kernel/sched/built-in.o
$ size vmlinux
text data bss dec hex filename
7411490 1178376 991232 9581098 92322a vmlinux
(after)
$ size kernel/sched/built-in.o
text data bss dec hex filename
155130 16445 7042 178617 2b9b9 kernel/sched/built-in.o
$ size vmlinux
text data bss dec hex filename
7411362 1178376 991232 9580970 9231aa vmlinux
I was choosing between resched_curr() and resched_rq(),
and the first name looks better for me.
A little lie in Documentation/trace/ftrace.txt. I have not
actually collected the tracing again. With a hope the patch
won't make execution times much worse :)
Signed-off-by: Kirill Tkhai <tkhai@yandex.ru>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Link: http://lkml.kernel.org/r/20140628200219.1778.18735.stgit@localhost
Signed-off-by: Ingo Molnar <mingo@kernel.org>
When a system is lightly loaded (i.e. no more than 1 job per cpu),
attempt to pull job to a cpu before putting it to idle is unnecessary and
can be skipped. This patch adds an indicator so the scheduler can know
when there's no more than 1 active job is on any CPU in the system to
skip needless job pulls.
On a 4 socket machine with a request/response kind of workload from
clients, we saw about 0.13 msec delay when we go through a full load
balance to try pull job from all the other cpus. While 0.1 msec was
spent on processing the request and generating a response, the 0.13 msec
load balance overhead was actually more than the actual work being done.
This overhead can be skipped much of the time for lightly loaded systems.
With this patch, we tested with a netperf request/response workload that
has the server busy with half the cpus in a 4 socket system. We found
the patch eliminated 75% of the load balance attempts before idling a cpu.
The overhead of setting/clearing the indicator is low as we already gather
the necessary info while we call add_nr_running() and update_sd_lb_stats.()
We switch to full load balance load immediately if any cpu got more than
one job on its run queue in add_nr_running. We'll clear the indicator
to avoid load balance when we detect no cpu's have more than one job
when we scan the work queues in update_sg_lb_stats(). We are aggressive
in turning on the load balance and opportunistic in skipping the load
balance.
Signed-off-by: Tim Chen <tim.c.chen@linux.intel.com>
Acked-by: Rik van Riel <riel@redhat.com>
Acked-by: Jason Low <jason.low2@hp.com>
Cc: "Paul E.McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Davidlohr Bueso <davidlohr@hp.com>
Cc: Alex Shi <alex.shi@linaro.org>
Cc: Michel Lespinasse <walken@google.com>
Cc: Peter Hurley <peter@hurleysoftware.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1403551009.2970.613.camel@schen9-DESK
Signed-off-by: Ingo Molnar <mingo@kernel.org>
A full dynticks CPU is allowed to stop its tick when a single task runs.
Meanwhile when a new task gets enqueued, the CPU must be notified so that
it can restart its tick to maintain local fairness and other accounting
details.
This notification is performed by way of an IPI. Then when the target
receives the IPI, we expect it to see the new value of rq->nr_running.
Hence the following ordering scenario:
CPU 0 CPU 1
write rq->running get IPI
smp_wmb() smp_rmb()
send IPI read rq->nr_running
But Paul Mckenney says that nowadays IPIs imply a full barrier on
all architectures. So we can safely remove this pair and rely on the
implicit barriers that come along IPI send/receive. Lets
just comment on this new assumption.
Acked-by: Peter Zijlstra <peterz@infradead.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Kevin Hilman <khilman@linaro.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Now that we have a nohz full remote kick based on irq work, lets use
it to notify a CPU that it's exiting single task mode.
This unbloats a bit the scheduler IPI that the nohz code was abusing
for its cool "callable anywhere/anytime" properties.
Acked-by: Peter Zijlstra <peterz@infradead.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Kevin Hilman <khilman@linaro.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Pull more scheduler updates from Ingo Molnar:
"Second round of scheduler changes:
- try-to-wakeup and IPI reduction speedups, from Andy Lutomirski
- continued power scheduling cleanups and refactorings, from Nicolas
Pitre
- misc fixes and enhancements"
* 'sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched/deadline: Delete extraneous extern for to_ratio()
sched/idle: Optimize try-to-wake-up IPI
sched/idle: Simplify wake_up_idle_cpu()
sched/idle: Clear polling before descheduling the idle thread
sched, trace: Add a tracepoint for IPI-less remote wakeups
cpuidle: Set polling in poll_idle
sched: Remove redundant assignment to "rt_rq" in update_curr_rt(...)
sched: Rename capacity related flags
sched: Final power vs. capacity cleanups
sched: Remove remaining dubious usage of "power"
sched: Let 'struct sched_group_power' care about CPU capacity
sched/fair: Disambiguate existing/remaining "capacity" usage
sched/fair: Change "has_capacity" to "has_free_capacity"
sched/fair: Remove "power" from 'struct numa_stats'
sched: Fix signedness bug in yield_to()
sched/fair: Use time_after() in record_wakee()
sched/balancing: Reduce the rate of needless idle load balancing
sched/fair: Fix unlocked reads of some cfs_b->quota/period
Now that 3.15 is released, this merges the 'next' branch into 'master',
bringing us to the normal situation where my 'master' branch is the
merge window.
* accumulated work in next: (6809 commits)
ufs: sb mutex merge + mutex_destroy
powerpc: update comments for generic idle conversion
cris: update comments for generic idle conversion
idle: remove cpu_idle() forward declarations
nbd: zero from and len fields in NBD_CMD_DISCONNECT.
mm: convert some level-less printks to pr_*
MAINTAINERS: adi-buildroot-devel is moderated
MAINTAINERS: add linux-api for review of API/ABI changes
mm/kmemleak-test.c: use pr_fmt for logging
fs/dlm/debug_fs.c: replace seq_printf by seq_puts
fs/dlm/lockspace.c: convert simple_str to kstr
fs/dlm/config.c: convert simple_str to kstr
mm: mark remap_file_pages() syscall as deprecated
mm: memcontrol: remove unnecessary memcg argument from soft limit functions
mm: memcontrol: clean up memcg zoneinfo lookup
mm/memblock.c: call kmemleak directly from memblock_(alloc|free)
mm/mempool.c: update the kmemleak stack trace for mempool allocations
lib/radix-tree.c: update the kmemleak stack trace for radix tree allocations
mm: introduce kmemleak_update_trace()
mm/kmemleak.c: use %u to print ->checksum
...
[ This series reduces the number of IPIs on Andy's workload by something like
99%. It's down from many hundreds per second to very few.
The basic idea behind this series is to make TIF_POLLING_NRFLAG be a
reliable indication that the idle task is polling. Once that's done,
the rest is reasonably straightforward. ]
When enqueueing tasks on remote LLC domains, we send an IPI to do the
work 'locally' and avoid bouncing all the cachelines over.
However, when the remote CPU is idle (and polling, say x86 mwait), we
don't need to send an IPI, we can simply kick the TIF word to wake it
up and have the 'idle' loop do the work.
So when _TIF_POLLING_NRFLAG is set, but _TIF_NEED_RESCHED is not (yet)
set, set _TIF_NEED_RESCHED and avoid sending the IPI.
Much-requested-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
[Edited by Andy Lutomirski, but this is mostly Peter Zijlstra's code.]
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Cc: nicolas.pitre@linaro.org
Cc: daniel.lezcano@linaro.org
Cc: Mike Galbraith <umgwanakikbuti@gmail.com>
Cc: umgwanakikbuti@gmail.com
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: linux-kernel@vger.kernel.org
Link: http://lkml.kernel.org/r/ce06f8b02e7e337be63e97597fc4b248d3aa6f9b.1401902905.git.luto@amacapital.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Yuyang Du