Currently we try to do task placement in wake_up_new_task() after we do
the load-balance pass in sched_fork(). This yields complicated semantics
in that we have to deal with tasks on different RQs and the
set_task_cpu() calls in copy_process() and sched_fork()
Rename ->task_new() to ->task_fork() and call it from sched_fork()
before the balancing, this gives the policy a clear point to place the
task.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Since set_task_clock() doesn't rely on rq->clock anymore we can simplyfy
the mess in ttwu().
Optimize things a bit by not fiddling with the IRQ state there.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
set_task_cpu() should be rq invariant and only touch task state, it
currently fails to do so, which opens up a few races, since not all
callers hold both rq->locks.
Remove the relyance on rq->clock, as any site calling set_task_cpu()
should also do a remote clock update, which should ensure the observed
time between these two cpus is monotonic, as per
kernel/sched_clock.c:sched_clock_remote().
Therefore we can simply remove the clock_offset bits and be happy.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
sched_rr_get_param calls
task->sched_class->get_rr_interval(task) without protection
against a concurrent sched_setscheduler() call which modifies
task->sched_class.
Serialize the access with task_rq_lock(task) and hand the rq
pointer into get_rr_interval() as it's needed at least in the
sched_fair implementation.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Peter Zijlstra <peterz@infradead.org>
LKML-Reference: <alpine.LFD.2.00.0912090930120.3089@localhost.localdomain>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
sched_getaffinity() is not protected against a concurrent
modification of the tasks affinity.
Serialize the access with task_rq_lock(task).
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Peter Zijlstra <peterz@infradead.org>
LKML-Reference: <20091208202026.769251187@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Since (e761b77: cpu hotplug, sched: Introduce cpu_active_map and redo
sched domain managment) we have cpu_active_mask which is suppose to rule
scheduler migration and load-balancing, except it never (fully) did.
The particular problem being solved here is a crash in try_to_wake_up()
where select_task_rq() ends up selecting an offline cpu because
select_task_rq_fair() trusts the sched_domain tree to reflect the
current state of affairs, similarly select_task_rq_rt() trusts the
root_domain.
However, the sched_domains are updated from CPU_DEAD, which is after the
cpu is taken offline and after stop_machine is done. Therefore it can
race perfectly well with code assuming the domains are right.
Cure this by building the domains from cpu_active_mask on
CPU_DOWN_PREPARE.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
* 'core-rcu-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (31 commits)
rcu: Make RCU's CPU-stall detector be default
rcu: Add expedited grace-period support for preemptible RCU
rcu: Enable fourth level of TREE_RCU hierarchy
rcu: Rename "quiet" functions
rcu: Re-arrange code to reduce #ifdef pain
rcu: Eliminate unneeded function wrapping
rcu: Fix grace-period-stall bug on large systems with CPU hotplug
rcu: Eliminate __rcu_pending() false positives
rcu: Further cleanups of use of lastcomp
rcu: Simplify association of forced quiescent states with grace periods
rcu: Accelerate callback processing on CPUs not detecting GP end
rcu: Mark init-time-only rcu_bootup_announce() as __init
rcu: Simplify association of quiescent states with grace periods
rcu: Rename dynticks_completed to completed_fqs
rcu: Enable synchronize_sched_expedited() fastpath
rcu: Remove inline from forward-referenced functions
rcu: Fix note_new_gpnum() uses of ->gpnum
rcu: Fix synchronization for rcu_process_gp_end() uses of ->completed counter
rcu: Prepare for synchronization fixes: clean up for non-NO_HZ handling of ->completed counter
rcu: Cleanup: balance rcu_irq_enter()/rcu_irq_exit() calls
...
We don't need to build mutex_spin_on_owner() if we have
CONFIG_DEBUG_MUTEXES or CONFIG_HAVE_DEFAULT_NO_SPIN_MUTEXES as
it won't be used under such configs.
Use CONFIG_MUTEX_SPIN_ON_OWNER as it gathers all the necessary
checks before building it.
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
LKML-Reference: <1259783357-8542-2-git-send-regression-fweisbec@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
This is a real fix for problem of utime/stime values decreasing
described in the thread:
http://lkml.org/lkml/2009/11/3/522
Now cputime is accounted in the following way:
- {u,s}time in task_struct are increased every time when the thread
is interrupted by a tick (timer interrupt).
- When a thread exits, its {u,s}time are added to signal->{u,s}time,
after adjusted by task_times().
- When all threads in a thread_group exits, accumulated {u,s}time
(and also c{u,s}time) in signal struct are added to c{u,s}time
in signal struct of the group's parent.
So {u,s}time in task struct are "raw" tick count, while
{u,s}time and c{u,s}time in signal struct are "adjusted" values.
And accounted values are used by:
- task_times(), to get cputime of a thread:
This function returns adjusted values that originates from raw
{u,s}time and scaled by sum_exec_runtime that accounted by CFS.
- thread_group_cputime(), to get cputime of a thread group:
This function returns sum of all {u,s}time of living threads in
the group, plus {u,s}time in the signal struct that is sum of
adjusted cputimes of all exited threads belonged to the group.
The problem is the return value of thread_group_cputime(),
because it is mixed sum of "raw" value and "adjusted" value:
group's {u,s}time = foreach(thread){{u,s}time} + exited({u,s}time)
This misbehavior can break {u,s}time monotonicity.
Assume that if there is a thread that have raw values greater
than adjusted values (e.g. interrupted by 1000Hz ticks 50 times
but only runs 45ms) and if it exits, cputime will decrease (e.g.
-5ms).
To fix this, we could do:
group's {u,s}time = foreach(t){task_times(t)} + exited({u,s}time)
But task_times() contains hard divisions, so applying it for
every thread should be avoided.
This patch fixes the above problem in the following way:
- Modify thread's exit (= __exit_signal()) not to use task_times().
It means {u,s}time in signal struct accumulates raw values instead
of adjusted values. As the result it makes thread_group_cputime()
to return pure sum of "raw" values.
- Introduce a new function thread_group_times(*task, *utime, *stime)
that converts "raw" values of thread_group_cputime() to "adjusted"
values, in same calculation procedure as task_times().
- Modify group's exit (= wait_task_zombie()) to use this introduced
thread_group_times(). It make c{u,s}time in signal struct to
have adjusted values like before this patch.
- Replace some thread_group_cputime() by thread_group_times().
This replacements are only applied where conveys the "adjusted"
cputime to users, and where already uses task_times() near by it.
(i.e. sys_times(), getrusage(), and /proc/<PID>/stat.)
This patch have a positive side effect:
- Before this patch, if a group contains many short-life threads
(e.g. runs 0.9ms and not interrupted by ticks), the group's
cputime could be invisible since thread's cputime was accumulated
after adjusted: imagine adjustment function as adj(ticks, runtime),
{adj(0, 0.9) + adj(0, 0.9) + ....} = {0 + 0 + ....} = 0.
After this patch it will not happen because the adjustment is
applied after accumulated.
v2:
- remove if()s, put new variables into signal_struct.
Signed-off-by: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Cc: Spencer Candland <spencer@bluehost.com>
Cc: Americo Wang <xiyou.wangcong@gmail.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Stanislaw Gruszka <sgruszka@redhat.com>
LKML-Reference: <4B162517.8040909@jp.fujitsu.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
- Remove if({u,s}t)s because no one call it with NULL now.
- Use cputime_{add,sub}().
- Add ifndef-endif for prev_{u,s}time since they are used
only when !VIRT_CPU_ACCOUNTING.
Signed-off-by: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Spencer Candland <spencer@bluehost.com>
Cc: Americo Wang <xiyou.wangcong@gmail.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Stanislaw Gruszka <sgruszka@redhat.com>
LKML-Reference: <4B1624C7.7040302@jp.fujitsu.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Anton Blanchard wrote:
> We allocate and zero cpu_isolated_map after the isolcpus
> __setup option has run. This means cpu_isolated_map always
> ends up empty and if CPUMASK_OFFSTACK is enabled we write to a
> cpumask that hasn't been allocated.
I introduced this regression in 49557e6203 (sched: Fix
boot crash by zalloc()ing most of the cpu masks).
Use the bootmem allocator if they set isolcpus=, otherwise
allocate and zero like normal.
Reported-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: peterz@infradead.org
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: <stable@kernel.org>
LKML-Reference: <200912021409.17013.rusty@rustcorp.com.au>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Tested-by: Anton Blanchard <anton@samba.org>
498657a478 incorrectly assumed
that preempt wasn't disabled around context_switch() and thus
was fixing imaginary problem. It also broke KVM because it
depended on ->sched_in() to be called with irq enabled so that
it can do smp calls from there.
Revert the incorrect commit and add comment describing different
contexts under with the two callbacks are invoked.
Avi: spotted transposed in/out in the added comment.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Avi Kivity <avi@redhat.com>
Cc: peterz@infradead.org
Cc: efault@gmx.de
Cc: rusty@rustcorp.com.au
LKML-Reference: <1259726212-30259-2-git-send-email-tj@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Use of msecs_to_jiffies() for nsecs_to_cputime() have some
problems:
- The type of msecs_to_jiffies()'s argument is unsigned int, so
it cannot convert msecs greater than UINT_MAX = about 49.7 days.
- msecs_to_jiffies() returns MAX_JIFFY_OFFSET if MSB of argument
is set, assuming that input was negative value. So it cannot
convert msecs greater than INT_MAX = about 24.8 days too.
This patch defines a new function nsecs_to_jiffies() that can
deal greater values, and that can deal all incoming values as
unsigned.
Signed-off-by: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Cc: Stanislaw Gruszka <sgruszka@redhat.com>
Cc: Spencer Candland <spencer@bluehost.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Amrico Wang <xiyou.wangcong@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: John Stultz <johnstul@linux.vnet.ibm.com>
LKML-Reference: <4B0E16E7.5070307@jp.fujitsu.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Now all task_{u,s}time() pairs are replaced by task_times().
And task_gtime() is too simple to be an inline function.
Cleanup them all.
Signed-off-by: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Cc: Stanislaw Gruszka <sgruszka@redhat.com>
Cc: Spencer Candland <spencer@bluehost.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Americo Wang <xiyou.wangcong@gmail.com>
LKML-Reference: <4B0E16D1.70902@jp.fujitsu.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Functions task_{u,s}time() are called in pair in almost all
cases. However task_stime() is implemented to call task_utime()
from its inside, so such paired calls run task_utime() twice.
It means we do heavy divisions (div_u64 + do_div) twice to get
utime and stime which can be obtained at same time by one set
of divisions.
This patch introduces a function task_times(*tsk, *utime,
*stime) to retrieve utime and stime at once in better, optimized
way.
Signed-off-by: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Cc: Stanislaw Gruszka <sgruszka@redhat.com>
Cc: Spencer Candland <spencer@bluehost.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Americo Wang <xiyou.wangcong@gmail.com>
LKML-Reference: <4B0E16AE.906@jp.fujitsu.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
In commit v2.6.21-691-g39bc89f ("make SysRq-T show all tasks
again") the interface of show_state_filter() was changed: zero
valued 'state_filter' specifies "dump all tasks" (instead of -1).
However, the condition for calling debug_show_all_locks() ("show
locks if all tasks are dumped") was not updated accordingly.
Signed-off-by: Shmulik Ladkani <shmulik.ladkani@gmail.com>
Cc: peterz@infradead.org
LKML-Reference: <4b0d2fe4.0ab6660a.6437.3cfc@mx.google.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Heiko reported a case where a timer interrupt managed to
reference a root_domain structure that was already freed by a
concurrent hot-un-plug operation.
Solve this like the regular sched_domain stuff is also
synchronized, by adding a synchronize_sched() stmt to the free
path, this ensures that a root_domain stays present for any
atomic section that could have observed it.
Reported-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Gregory Haskins <ghaskins@novell.com>
Cc: Siddha Suresh B <suresh.b.siddha@intel.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
LKML-Reference: <1258363873.26714.83.camel@laptop>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
In finish_task_switch(), fire_sched_in_preempt_notifiers() is
called after finish_lock_switch().
However, depending on architecture, preemption can be enabled after
finish_lock_switch() which breaks the semantics of preempt
notifiers.
So move it before finish_arch_switch(). This also makes the in-
notifiers symmetric to out- notifiers in terms of locking - now
both are called under rq lock.
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Avi Kivity <avi@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <4AFD2801.7020900@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Originally task_s/utime() were designed to return clock_t but
later changed to return cputime_t by following commit:
commit efe567fc82
Author: Christian Borntraeger <borntraeger@de.ibm.com>
Date: Thu Aug 23 15:18:02 2007 +0200
It only changed the type of return value, but not the
implementation. As the result the granularity of task_s/utime()
is still that of clock_t, not that of cputime_t.
So using task_s/utime() in __exit_signal() makes values
accumulated to the signal struct to be rounded and coarse
grained.
This patch removes casts to clock_t in task_u/stime(), to keep
granularity of cputime_t over the calculation.
v2:
Use div_u64() to avoid error "undefined reference to `__udivdi3`"
on some 32bit systems.
Signed-off-by: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Cc: xiyou.wangcong@gmail.com
Cc: Spencer Candland <spencer@bluehost.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Stanislaw Gruszka <sgruszka@redhat.com>
LKML-Reference: <4AFB9029.9000208@jp.fujitsu.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
kthread_bind(), migrate_task() and sched_fork were missing
updates, and try_to_wake_up() was updating after having already
used the stale clock.
Aside from preventing potential latency hits, there' a side
benefit in that early boot printk time stamps become monotonic.
Signed-off-by: Mike Galbraith <efault@gmx.de>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1258020464.6491.2.camel@marge.simson.net>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
LKML-Reference: <new-submission>
From the code in rt_mutex_setprio(), it is evident that the
intention is that task's with a RT 'prio' value as a consequence
of receiving a PI boost also have their 'sched_class' field set
to '&rt_sched_class'.
However, Peter noticed that the code in __setscheduler() could
result in this intention being frustrated. Fix it.
Reported-by: Peter Williams <pwil3058@bigpond.net.au>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
LKML-Reference: <1257880321.4108.457.camel@laptop>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Commit 1b9508f, "Rate-limit newidle" has been confirmed to fix
the netperf UDP loopback regression reported by Alex Shi.
This is a cleanup and a fix:
- moved to a more out of the way spot
- fix to ensure that balancing doesn't try to balance
runqueues which haven't gone online yet, which can
mess up CPU enumeration during boot.
Reported-by: Alex Shi <alex.shi@intel.com>
Reported-by: Zhang, Yanmin <yanmin_zhang@linux.intel.com>
Signed-off-by: Mike Galbraith <efault@gmx.de>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: <stable@kernel.org> # .32.x: a1f84a3: sched: Check for an idle shared cache
Cc: <stable@kernel.org> # .32.x: 1b9508f: sched: Rate-limit newidle
Cc: <stable@kernel.org> # .32.x: fd21073: sched: Fix affinity logic
Cc: <stable@kernel.org> # .32.x
LKML-Reference: <1257821402.5648.17.camel@marge.simson.net>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
In 15934a3732,
field last_tick_seen is added to struct rq.
But it is unused now.
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Guillaume Chazarain <guichaz@yahoo.fr>
LKML-Reference: <4AE6A513.6010100@cn.fujitsu.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
root_task_group_empty is used only with FAIR_GROUP_SCHED
so if we use other scheduler options we get:
kernel/sched.c:314: warning: 'root_task_group_empty' defined but not used
So move CONFIG_FAIR_GROUP_SCHED up that it covers
root_task_group_empty().
Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Cc: Peter Zijlstra <peterz@infradead.org>
LKML-Reference: <20091026192414.GB5321@lenovo>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Fix variable name in sched.c kernel-doc notation.
Fixes this DocBook warning:
Warning(kernel/sched.c:2008): No description found for parameter
'p' Warning(kernel/sched.c:2008): Excess function parameter 'k'
description in 'kthread_bind'
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
LKML-Reference: <4AF4B1BC.8020604@oracle.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
* 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
sched: Fix kthread_bind() by moving the body of kthread_bind() to sched.c
sched: Disable SD_PREFER_LOCAL at node level
sched: Fix boot crash by zalloc()ing most of the cpu masks
sched: Strengthen buddies and mitigate buddy induced latencies
Rate limit newidle to migration_cost. It's a win for all
stages of sysbench oltp tests.
Signed-off-by: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Currently partition_sched_domains() takes a 'struct cpumask
*doms_new' which is a kmalloc'ed array of cpumask_t. You can't
have such an array if 'struct cpumask' is undefined, as we plan
for CONFIG_CPUMASK_OFFSTACK=y.
So, we make this an array of cpumask_var_t instead: this is the
same for the CONFIG_CPUMASK_OFFSTACK=n case, but requires
multiple allocations for the CONFIG_CPUMASK_OFFSTACK=y case.
Hence we add alloc_sched_domains() and free_sched_domains()
functions.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Peter Zijlstra <peterz@infradead.org>
LKML-Reference: <200911031453.40668.rusty@rustcorp.com.au>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
cpu_nr_migrations() is not used, remove it.
Signed-off-by: Hiroshi Shimamoto <h-shimamoto@ct.jp.nec.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <4AF12A66.6020609@ct.jp.nec.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Eric Paris reported that commit
f685ceacab causes boot time
PREEMPT_DEBUG complaints.
[ 4.590699] BUG: using smp_processor_id() in preemptible [00000000] code: rmmod/1314
[ 4.593043] caller is task_hot+0x86/0xd0
Since kthread_bind() messes with scheduler internals, move the
body to sched.c, and lock the runqueue.
Reported-by: Eric Paris <eparis@redhat.com>
Signed-off-by: Mike Galbraith <efault@gmx.de>
Tested-by: Eric Paris <eparis@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1256813310.7574.3.camel@marge.simson.net>
[ v2: fix !SMP build and clean up ]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
I got a boot crash when forcing cpumasks offstack on 32 bit,
because find_new_ilb() returned 3 on my UP system (nohz.cpu_mask
wasn't zeroed).
AFAICT the others need to be zeroed too: only
nohz.ilb_grp_nohz_mask is initialized before use.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Peter Zijlstra <peterz@infradead.org>
LKML-Reference: <200911022037.21282.rusty@rustcorp.com.au>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/percpu:
sched: move rq_weight data array out of .percpu
percpu: allow pcpu_alloc() to be called with IRQs off
CPU time of a guest is always accounted in 'user' time
without concern for the nice value of its counterpart
process although the guest is scheduled under the nice
value.
This patch fixes the defect and accounts cpu time of
a niced guest in 'nice' time as same as a niced process.
And also the patch adds 'guest_nice' to cpuacct. The
value provides niced guest cpu time which is like 'nice'
to 'user'.
The original discussions can be found here:
http://www.mail-archive.com/kvm@vger.kernel.org/msg23982.htmlhttp://www.mail-archive.com/kvm@vger.kernel.org/msg23860.html
Signed-off-by: Ryota Ozaki <ozaki.ryota@gmail.com>
Acked-by: Avi Kivity <avi@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1256314810-7897-1-git-send-email-ozaki.ryota@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
This patch restores the effectiveness of LAST_BUDDY in preventing
pgsql+oltp from collapsing due to wakeup preemption. It also
switches LAST_BUDDY to exclusively do what it does best, namely
mitigate the effects of aggressive wakeup preemption, which
improves vmark throughput markedly, and restores mysql+oltp
scalability.
Since buddies are about scalability, enable them beginning at the
point where we begin expanding sched_latency, namely
sched_nr_latency. Previously, buddies were cleared aggressively,
which seriously reduced their effectiveness. Not clearing
aggressively however, produces a small drop in mysql+oltp
throughput immediately after peak, indicating that LAST_BUDDY is
actually doing some harm. This is right at the point where X on the
desktop in competition with another load wants low latency service.
Ergo, do not enable until we need to scale.
To mitigate latency induced by buddies, or by a task just missing
wakeup preemption, check latency at tick time.
Last hunk prevents buddies from stymieing BALANCE_NEWIDLE via
CACHE_HOT_BUDDY.
Supporting performance tests:
tip = v2.6.32-rc5-1497-ga525b32
tipx = NO_GENTLE_FAIR_SLEEPERS NEXT_BUDDY granularity knobs = 31 knobs + 31 buddies
tip+x = NO_GENTLE_FAIR_SLEEPERS granularity knobs = 31 knobs
(Three run averages except where noted.)
vmark:
------
tip 108466 messages per second
tip+ 125307 messages per second
tip+x 125335 messages per second
tipx 117781 messages per second
2.6.31.3 122729 messages per second
mysql+oltp:
-----------
clients 1 2 4 8 16 32 64 128 256
..........................................................................................
tip 9949.89 18690.20 34801.24 34460.04 32682.88 30765.97 28305.27 25059.64 19548.08
tip+ 10013.90 18526.84 34900.38 34420.14 33069.83 32083.40 30578.30 28010.71 25605.47
tipx 9698.71 18002.70 34477.56 33420.01 32634.30 31657.27 29932.67 26827.52 21487.18
2.6.31.3 8243.11 18784.20 34404.83 33148.38 31900.32 31161.90 29663.81 25995.94 18058.86
pgsql+oltp:
-----------
clients 1 2 4 8 16 32 64 128 256
..........................................................................................
tip 13686.37 26609.25 51934.28 51347.81 49479.51 45312.65 36691.91 26851.57 24145.35
tip+ (1x) 13907.85 27135.87 52951.98 52514.04 51742.52 50705.43 49947.97 48374.19 46227.94
tip+x 13906.78 27065.81 52951.19 52542.59 52176.11 51815.94 50838.90 49439.46 46891.00
tipx 13742.46 26769.81 52351.99 51891.73 51320.79 50938.98 50248.65 48908.70 46553.84
2.6.31.3 13815.35 26906.46 52683.34 52061.31 51937.10 51376.80 50474.28 49394.47 47003.25
Signed-off-by: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
* 'for-linus' of git://git.kernel.dk/linux-2.6-block:
cciss: Add cciss_allow_hpsa module parameter
cciss: Fix multiple calls to pci_release_regions
blk-settings: fix function parameter kernel-doc notation
writeback: kill space in debugfs item name
writeback: account IO throttling wait as iowait
elv_iosched_store(): fix strstrip() misuse
cfq-iosched: avoid probable slice overrun when idling
cfq-iosched: apply bool value where we return 0/1
cfq-iosched: fix think time allowed for seekers
cfq-iosched: fix the slice residual sign
cfq-iosched: abstract out the 'may this cfqq dispatch' logic
block: use proper BLK_RW_ASYNC in blk_queue_start_tag()
block: Seperate read and write statistics of in_flight requests v2
block: get rid of kblock_schedule_delayed_work()
cfq-iosched: fix possible problem with jiffies wraparound
cfq-iosched: fix issue with rq-rq merging and fifo list ordering
The following htmldocs warnings:
Warning(kernel/sched.c:685): No description found for parameter 'cpu'
Warning(kernel/sched.c:3676): No description found for parameter 'sd'
Trigger because new parameters were added to update_rq_clock() and
update_group_power() without updating the kernel-doc notation.
Signed-off-by: Jaswinder Singh Rajput <jaswinderrajput@gmail.com>
Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <4AD29070.7070002@oracle.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
In try_to_wake_up(), we update the runqueue clock, but
select_task_rq() may select a different runqueue than the one we
updated, leaving the new runqueue's clock stale for a bit.
This patch cures occasional huge latencies reported by latencytop
when coming out of idle on a mostly idle NO_HZ box.
Signed-off-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1255070103.7639.30.camel@marge.simson.net>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
It makes sense to do IOWAIT when someone is blocked
due to IO throttle, as suggested by Kame and Peter.
There is an old comment for not doing IOWAIT on throttle,
however it has been mismatching the code for a long time.
If we stop accounting IOWAIT for 2.6.32, it could be an
undesirable behavior change. So restore the io_schedule.
CC: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
CC: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
Remove the comment about calling alloc_bootmem() as it is not
called here since commit 36b7b6d465.
Signed-off-by: Jayson R. King <dev@jaysonking.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Jiri Kosina <trivial@kernel.org>
LKML-Reference: <4AC9C8A6.6010209@jaysonking.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
normal_prio should be updated if policy changes from RT to
SCHED_MORMAL or if static_prio/nice is changed.
Some paths through sched_fork() ignore this requirement and may
result in normal_prio having an invalid value.
Fixing this issue allows the call to effective_prio() in
wake_up_new_task() to be removed.
Signed-off-by: Peter Williams <pwil3058@bigpond.net.au>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Mike Galbraith <efault@gmx.de>
LKML-Reference: <f8f46736fd4e7f090ac0.1253774830@mudlark.pw.nest>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Peter Zijlstra