- 12 Aug, 2015 4 commits
-
-
Peter Zijlstra authored
Give every class a set_cpus_allowed() method, this enables some small optimization in the RT,DL implementation by avoiding a double cpumask_weight() call. 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: dedekind1@gmail.com Cc: juri.lelli@arm.com Cc: mgorman@suse.de Cc: riel@redhat.com Cc: rostedt@goodmis.org Link: http://lkml.kernel.org/r/20150515154833.614517487@infradead.orgSigned-off-by:
Ingo Molnar <mingo@kernel.org>
-
Peter Zijlstra authored
Because sched_setscheduler() checks p->flags & PF_NO_SETAFFINITY without locks, a caller might observe an old value and race with the set_cpus_allowed_ptr() call from __kthread_bind() and effectively undo it: __kthread_bind() do_set_cpus_allowed() <SYSCALL> sched_setaffinity() if (p->flags & PF_NO_SETAFFINITIY) set_cpus_allowed_ptr() p->flags |= PF_NO_SETAFFINITY Fix the bug by putting everything under the regular scheduler locks. This also closes a hole in the serialization of task_struct::{nr_,}cpus_allowed. Signed-off-by:Tejun Heo <tj@kernel.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Mike Galbraith <efault@gmx.de> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: dedekind1@gmail.com Cc: juri.lelli@arm.com Cc: mgorman@suse.de Cc: riel@redhat.com Cc: rostedt@goodmis.org Link: http://lkml.kernel.org/r/20150515154833.545640346@infradead.orgSigned-off-by:
-
Byungchul Park authored
Current code ensures that a task has a normalized vruntime when switching away from the fair class, but it does not ensure the task has a non-normalized vruntime when switching back to the fair class. This is an example breaking this consistency: 1. a task is in fair class and !queued 2. changes its class to RT class (still !queued) 3. changes its class to fair class again (still !queued) Signed-off-by:
Byungchul Park <byungchul.park@lge.com> Signed-off-by:
-
Aravind Gopalakrishnan authored
Systems which have all nodes at a distance of at most 1 hop should be identified as 'NUMA_DIRECT'. However, the scheduler incorrectly identifies it as 'NUMA_BACKPLANE'. This is because 'n' is assigned to sched_max_numa_distance but the code (mis)interprets it to mean 'number of hops'. Rik had actually used sched_domains_numa_levels for detecting a 'NUMA_DIRECT' topology: http://marc.info/?l=linux-kernel&m=141279712429834&w=2 But that was changed when he removed the hops table in the subsequent version: http://marc.info/?l=linux-kernel&m=141353106106771&w=2 Fixing the issue here. Signed-off-by:
Aravind Gopalakrishnan <Aravind.Gopalakrishnan@amd.com> Signed-off-by:
Rik van Riel <riel@redhat.com> 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> Link: http://lkml.kernel.org/r/1439256048-3748-1-git-send-email-Aravind.Gopalakrishnan@amd.comSigned-off-by:
-
- 08 Aug, 2015 1 commit
-
-
Chris Metcalf authored
Move the simulator bits into finish_arch_post_lock_switch() and properly call __switch_to() from _switch_to(). Signed-off-by:
Chris Metcalf <cmetcalf@ezchip.com> Cc: <efault@gmx.de> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Link: http://lkml.kernel.org/r/1438783412-10990-1-git-send-email-cmetcalf@ezchip.com [ Made it a delta to: fe363adb ("sched, tile: Remove finish_arch_switch"). ] Signed-off-by:
-
- 04 Aug, 2015 8 commits
-
-
Ingo Molnar authored
There's no finish_arch_switch() anymore in the latest scheduler tree. Also update some other details. Cc: David S. Miller <davem@davemloft.net> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Mike Galbraith <efault@gmx.de> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: linux-kernel@vger.kernel.org Cc: sparclinux@vger.kernel.org Signed-off-by: -
Peter Zijlstra authored
One less arch hook.. Signed-off-by:
-
Chris Metcalf authored
Move the simulator bits into switch_to() and use finish_arch_post_lock_switch() for the homecache migration bits. Signed-off-by:
-
Peter Zijlstra authored
The code looks buggy; why would we be restoring the previous task's DSP state after we've switched to the next task? Fix that and put the restore in switch_to(), removing the need for finish_arch_switch(). Signed-off-by:
-
Peter Zijlstra authored
An empty implementation, make it go away. Signed-off-by:
-
Peter Zijlstra authored
Fold the tracing hook into switch_to() in order to remove finish_arch_switch(). Signed-off-by:
Hans-Christian Egtvedt <egtvedt@samfundet.no> 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: linux-kernel@vger.kernel.org Signed-off-by:
-
Ralf Baechle authored
MIPS was using finish_arch_switch() as a hook to restore and initialize CPU context for all threads, even newly created kernel and user threads. This is however entirely solvable within switch_to() so get rid of finish_arch_switch() which is in the way of scheduler cleanups. Signed-off-by:
Ralf Baechle <ralf@linux-mips.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: linux-kernel@vger.kernel.org Signed-off-by:
-
Will Deacon authored
Fold finish_arch_switch() into switch_to(). Signed-off-by:
Will Deacon <will.deacon@arm.com> Signed-off-by:
-
- 03 Aug, 2015 22 commits
-
-
Yuyang Du authored
For cfs_rq, we have load.weight, runnable_load_avg, and load_avg. Clean up how they are used: - First, as group sched_entity already largely uses load_avg, we now expand to use load_avg in all cases. - Second, for CPU-wide load balancing, we choose to use runnable_load_avg in all cases, which is the same as before this series. Signed-off-by:Yuyang Du <yuyang.du@intel.com> Signed-off-by:
-
Yuyang Du authored
The cfs_rq's load_avg is composed of runnable_load_avg and blocked_load_avg. Before this series, sometimes the runnable_load_avg is used, and sometimes the load_avg is used. Completely replacing all uses of runnable_load_avg with load_avg may be too big a leap, i.e., the blocked_load_avg is concerned to result in overrated load. Therefore, we get runnable_load_avg back. The new cfs_rq's runnable_load_avg is improved to be updated with all of the runnable sched_eneities at the same time, so the one sched_entity updated and the others stale problem is solved. Signed-off-by:
-
Yuyang Du authored
When task exits or group is destroyed, the entity's load should be removed from its parent cfs_rq's load. Otherwise, it will take time for the parent cfs_rq to decay the dead entity's load to 0, which is not desired. Signed-off-by:
-
Yuyang Du authored
The runnable load and utilization averages of cfs_rq's sched_entity were not initiated. Like done to a task, give new cfs_rq' sched_entity start values to heavy its load in infant time. Signed-off-by:
-
Vincent Guittot authored
The load and the utilization of idle CPUs must be updated periodically in order to decay the blocked part. If CONFIG_FAIR_GROUP_SCHED is not set, the load and util of idle cpus are not decayed and stay at the values set before becoming idle. Signed-off-by:
Vincent Guittot <vincent.guittot@linaro.org> Signed-off-by:
-
Yuyang Du authored
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 authored
The current rq->avg is not used at all since its merge into the kernel, and the code is in the scheduler's hot path, so remove it. Tested-by:
Dietmar Eggemann <dietmar.eggemann@arm.com> Signed-off-by:
-
Oleg Nesterov authored
cpu_stop_park() does cpu_stop_signal_done() but leaves the work on stopper->works. The owner of this work can free/reuse this memory right after that and corrupt the list, so if this CPU becomes online again cpu_stopper_thread() will crash. Signed-off-by:
Oleg Nesterov <oleg@redhat.com> Signed-off-by:
-
Oleg Nesterov authored
Cosmetic, but 'cpu_stop_fn_t' actually makes the code more readable and it doesn't break cscope. And most of the declarations already use it. Signed-off-by:
-
Oleg Nesterov authored
The only caller outside of stop_machine.c is _cpu_down(), it can use stop_machine(). get_online_cpus() is fine under cpu_hotplug_begin(). Signed-off-by:
-
Oleg Nesterov authored
queue_stop_cpus_work() can do everything in one for_each_cpu() loop. Signed-off-by:
-
Oleg Nesterov authored
Multpiple DEFINE_PER_CPU's do not make sense, move all the per-cpu variables into 'struct cpu_stopper'. Signed-off-by:
-
Konstantin Khlebnikov authored
These functions check should_resched() before unlocking spinlock/bh-enable: preempt_count always non-zero => should_resched() always returns false. cond_resched_lock() worked iff spin_needbreak is set. This patch adds argument "preempt_offset" to should_resched(). preempt_count offset constants for that: PREEMPT_DISABLE_OFFSET - offset after preempt_disable() PREEMPT_LOCK_OFFSET - offset after spin_lock() SOFTIRQ_DISABLE_OFFSET - offset after local_bh_distable() SOFTIRQ_LOCK_OFFSET - offset after spin_lock_bh() Signed-off-by:
Konstantin Khlebnikov <khlebnikov@yandex-team.ru> Signed-off-by:
-
Konstantin Khlebnikov authored
Function should_resched() is equal to (!preempt_count() && need_resched()). In preemptive kernel preempt_count here is non-zero because of vc->lock. Signed-off-by:
-
Konstantin Khlebnikov authored
This code is used only when CONFIG_PREEMPT=n and only in non-atomic context: xen_in_preemptible_hcall is set only in privcmd_ioctl_hypercall(). Thus preempt_count is zero and should_resched() is equal to need_resched(). Signed-off-by:
-
Mike Galbraith authored
Josef Bacik reported that Facebook sees better performance with their 1:N load (1 dispatch/node, N workers/node) when carrying an old patch to try very hard to wake to an idle CPU. While looking at wake_wide(), I noticed that it doesn't pay attention to the wakeup of a many partner waker, returning 1 only when waking one of its many partners. Correct that, letting explicit domain flags override the heuristic. While at it, adjust task_struct bits, we don't need a 64-bit counter. Tested-by:
Josef Bacik <jbacik@fb.com> Signed-off-by:
Mike Galbraith <umgwanakikbuti@gmail.com> [ Tidy things up. ] Signed-off-by:
-
Peter Zijlstra authored
Mathieu reported that since 317f3941 ("sched: Move the second half of ttwu() to the remote cpu") trace_sched_wakeup() can happen out of context of the waker. This is a problem when you want to analyse wakeup paths because it is now very hard to correlate the wakeup event to whoever issued the wakeup. OTOH trace_sched_wakeup() is issued at the point where we set p->state = TASK_RUNNING, which is right were we hand the task off to the scheduler, so this is an important point when looking at scheduling behaviour, up to here its been the wakeup path everything hereafter is due to scheduler policy. To bridge this gap, introduce a second tracepoint: trace_sched_waking. It is guaranteed to be called in the waker context. Reported-by:
Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Signed-off-by:
-
Peter Zijlstra authored
While the current code guarantees monotonicity for stime and utime independently of one another, it does not guarantee that the sum of both is equal to the total time we started out with. This confuses things (and peoples) who look at this sum, like top, and will report >100% usage followed by a matching period of 0%. Rework the code to provide both individual monotonicity and a coherent sum. Suggested-by:
Fredrik Markstrom <fredrik.markstrom@gmail.com> Reported-by:
-
Markus Elfring authored
The unregister_sysctl_table() function tests whether its argument is NULL and then returns immediately. Thus the test around the call is not needed. This issue was detected by using the Coccinelle software. Signed-off-by:
Markus Elfring <elfring@users.sourceforge.net> Signed-off-by:
-
Xunlei Pang authored
'p' has been already queued at this point, so "!task_running(rq, p)" and "p->nr_cpus_allowed > 1" imply that "has_pushable_dl_tasks(rq)" is true, so it can be removed. Signed-off-by:
Xunlei Pang <pang.xunlei@linaro.org> Signed-off-by:
Steven Rostedt <rostedt@goodmis.org> Cc: Juri Lelli <juri.lelli@gmail.com> 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> Link: http://lkml.kernel.org/r/1435995563-3723-2-git-send-email-xlpang@126.comSigned-off-by:
-
Xunlei Pang authored
'p' has been already queued at this point, so "!task_running(rq, p)" and "p->nr_cpus_allowed > 1" imply that "has_pushable_tasks(rq)" is true, so it can be removed. Signed-off-by:
-
Yuyang Du authored
In idle balancing where a CPU going idle pulls tasks from another CPU, a livelock may happen if the CPU pulls all tasks from another, makes it idle, and this iterates. So just avoid this. Reported-by:
Rabin Vincent <rabin.vincent@axis.com> Signed-off-by:
-
- 21 Jul, 2015 1 commit
-
-
Lucas Stach authored
Make sure to stop tracing only once we are past a point where all latency tracing events have been processed (irqs are not enabled again). This has the slight advantage of capturing more latency related events in the idle path, but most importantly it makes sure that latency tracing doesn't get re-enabled inadvertently when new events are coming in. This makes the irqsoff latency tracer useful again, as we stop capturing CPU sleep time as IRQ latency. Signed-off-by:
Lucas Stach <l.stach@pengutronix.de> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Rafael J. Wysocki <rjw@rjwysocki.net> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: kernel@pengutronix.de Cc: patchwork-lst@pengutronix.de Link: http://lkml.kernel.org/r/1437410090-3747-1-git-send-email-l.stach@pengutronix.deSigned-off-by:
-
- 07 Jul, 2015 4 commits
-
-
Byungchul Park authored
update_cfs_rq_load_contribution() was changed to __update_cfs_rq_tg_load_contrib() - sync up the commit in calc_tg_weight() too. Signed-off-by:
-
Boqun Feng authored
Since commit: 4bf0b771 ("sched: remove do_div() from __sched_slice()") ... the logic of __sched_period() can be implemented as a single if-else without any local variables, so this patch cleans it up with an if-else statement, which expresses the function's logic straightforwardly. Signed-off-by:
Boqun Feng <boqun.feng@gmail.com> Signed-off-by:
-
Srikar Dronamraju authored
This is consistent with all other load balancing instances where we absorb unfairness upto env->imbalance_pct. Absorbing unfairness upto env->imbalance_pct allows to pull and retain task to their preferred nodes. Signed-off-by:
Srikar Dronamraju <srikar@linux.vnet.ibm.com> Signed-off-by:
-
Srikar Dronamraju authored
The current load balancer may not try to prevent a task from moving out of a preferred node to a less preferred node. The reason for this being: - Since sched features NUMA and NUMA_RESIST_LOWER are disabled by default, migrate_degrades_locality() always returns false. - Even if NUMA_RESIST_LOWER were to be enabled, if its cache hot, migrate_degrades_locality() never gets called. The above behaviour can mean that tasks can move out of their preferred node but they may be eventually be brought back to their preferred node by numa balancer (due to higher numa faults). To avoid the above, this commit merges migrate_degrades_locality() and migrate_improves_locality(). It also replaces 3 sched features NUMA, NUMA_FAVOUR_HIGHER and NUMA_RESIST_LOWER by a single sched feature NUMA. Signed-off-by:
-
