- 10 Feb, 2017 1 commit
-
-
Steven Rostedt (VMware) authored
The check for 'running' in sched_move_task() has an unlikely() around it. That is, it is unlikely that the task being moved is running. That use to be true. But with a couple of recent updates, it is now likely that the task will be running. The first change came from ea86cb4b ("sched/cgroup: Fix cpu_cgroup_fork() handling") that moved around the use case of sched_move_task() in do_fork() where the call is now done after the task is woken (hence it is running). The second change came from 8e5bfa8c ("sched/autogroup: Do not use autogroup->tg in zombie threads") where sched_move_task() is called by the exit path, by the task that is exiting. Hence it too is running. Signed-off-by:
Steven Rostedt (VMware) <rostedt@goodmis.org> Signed-off-by:
Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Vincent Guittot <vincent.guittot@linaro.org> Link: http://lkml.kernel.org/r/20170206110426.27ca6426@gandalf.local.homeSigned-off-by:
Ingo Molnar <mingo@kernel.org>
-
- 08 Feb, 2017 1 commit
-
-
Ingo Molnar authored
The names are all 'autogroup', not 'auto_group' - so rename the kernel/sched/auto_group.[ch] to match the existing nomenclature. Cc: Peter Zijlstra <peterz@infradead.org> Cc: Mike Galbraith <efault@gmx.de> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: linux-kernel@vger.kernel.org Signed-off-by:
-
- 07 Feb, 2017 5 commits
-
-
Ingo Molnar authored
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: -
Ingo Molnar authored
Over the years sched/core.c accumulated over 50 #include lines, 40 of which are superfluous. (!) Removing them decreases the preprocessed .c file (.i) size noticeably: triton:~/tip> wc -l kernel/sched/core.i Before: 76387 kernel/sched/core.i After: 75896 kernel/sched/core.i 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: -
Ingo Molnar authored
update_rq_clock_task() and update_rq_clock() we unnecessarily spread across core.c, requiring an extra prototype line. Move them next to each other and in the proper order. 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: -
Ingo Molnar authored
include/linux/delayacct.h relies on 'struct taskstats' but does not include the header that defines it. This worked so far because files that included <linux/taskstats.h> also happened to include other headers that included uapi/linux/taskstats.h. Fix it. 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: -
Ingo Molnar authored
Refresh the comments in the core scheduler code: - Capitalize sentences consistently - Capitalize 'CPU' consistently - ... and other small details. 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:
-
- 01 Feb, 2017 33 commits
-
-
Shile Zhang authored
We added the 'sched_rr_timeslice_ms' SCHED_RR tuning knob in this commit: ce0dbbbb ("sched/rt: Add a tuning knob to allow changing SCHED_RR timeslice") ... which name suggests to users that it's in milliseconds, while in reality it's being set in milliseconds but the result is shown in jiffies. This is obviously confusing when HZ is not 1000, it makes it appear like the value set failed, such as HZ=100: root# echo 100 > /proc/sys/kernel/sched_rr_timeslice_ms root# cat /proc/sys/kernel/sched_rr_timeslice_ms 10 Fix this to be milliseconds all around. Signed-off-by:
Shile Zhang <shile.zhang@nokia.com> Signed-off-by:
-
Paul Gortmaker authored
In commit: acb04058 ("sched/clock: Fix hotplug crash") the PARISC code gained a call to this function. However the prototype for it is within a CONFIG_HAVE_UNSTABLE_SCHED_CLOCK=y #ifdef/#endif. That, combined with this: arch/parisc/Kconfig: select HAVE_UNSTABLE_SCHED_CLOCK if SMP means that PARISC can have it either enabled or disabled, resulting in the following build fail: arch/parisc/kernel/setup.c:180:2: error: implicit declaration of function 'clear_sched_clock_stable' [-Werror=implicit-function-declaration] Add a no-op stub for the non-SMP case to prevent this. Signed-off-by:
Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by:
-
Frederic Weisbecker authored
cputime_t is now only used by two architectures: * powerpc (when CONFIG_VIRT_CPU_ACCOUNTING_NATIVE=y) * s390 And since the core doesn't use it anymore, we don't need any arch support from the others. So we can remove their stub implementations. A final cleanup would be to provide an efficient pure arch implementation of cputime_to_nsec() for s390 and powerpc and finally remove include/linux/cputime.h . Signed-off-by:
Frederic Weisbecker <fweisbec@gmail.com> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Fenghua Yu <fenghua.yu@intel.com> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Michael Ellerman <mpe@ellerman.id.au> Cc: Paul Mackerras <paulus@samba.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Rik van Riel <riel@redhat.com> Cc: Stanislaw Gruszka <sgruszka@redhat.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Tony Luck <tony.luck@intel.com> Cc: Wanpeng Li <wanpeng.li@hotmail.com> Link: http://lkml.kernel.org/r/1485832191-26889-36-git-send-email-fweisbec@gmail.comSigned-off-by:
-
Frederic Weisbecker authored
It's unused now. Signed-off-by:
-
Frederic Weisbecker authored
Since the core doesn't deal with cputime_t anymore, most of these APIs have been left unused. Lets remove these. Signed-off-by:
-
Frederic Weisbecker authored
Since the core doesn't deal with cputime_t anymore, most of these APIs have been left unused. Lets remove these. Signed-off-by:
-
Frederic Weisbecker authored
This way we don't need to deal with cputime_t details from the core code. Signed-off-by:
-
Frederic Weisbecker authored
This nsec based cputime_t implementation isn't used anymore. We can remove it. Signed-off-by:
-
Frederic Weisbecker authored
There is no need anymore for this cputime_t midlayer. Let's use nsec units directly. Signed-off-by:
-
Frederic Weisbecker authored
Initially, nsec based cputime_t implementation belonged to IA64. It got exported later for CONFIG_VIRT_CPU_ACCOUNTING_GEN but now it is again only used by IA64. So let's move it back there. Signed-off-by:
-
Frederic Weisbecker authored
This cputime_t implementation is now unused, we can remove it. Signed-off-by:
-
Frederic Weisbecker authored
Turn the full dynticks cputime clock source to return nsec while keeping its very internals jiffies based for performance reasons. Signed-off-by:
-
Frederic Weisbecker authored
This is the final step toward tick based cputime conversion. Now that the whole cputime accounting engine accounts in nsecs, we can convert the very source of the cputime to account in nsecs. Signed-off-by:
-
Frederic Weisbecker authored
This is one more step toward converting cputime accounting to pure nsecs. Signed-off-by:
-
Frederic Weisbecker authored
This is one more step toward converting cputime accounting to pure nsecs. Signed-off-by:
-
Frederic Weisbecker authored
This is one more step toward converting cputime accounting to pure nsecs. Signed-off-by:
-
Frederic Weisbecker authored
This is one more step toward converting cputime accounting to pure nsecs. Signed-off-by:
-
Frederic Weisbecker authored
Now that the whole cputime conversion to nsec units is complete, we can remove the compatibility accessors. Signed-off-by:
-
Frederic Weisbecker authored
Use the new nsec based cputime accessors as part of the whole cputime conversion from cputime_t to nsecs. Also convert itimers to use nsec based internal counters. This simplifies it and removes the whole game with error/inc_error which served to deal with cputime_t random granularity. Signed-off-by:
-
Frederic Weisbecker authored
Use the new nsec based cputime accessors as part of the whole cputime conversion from cputime_t to nsecs. Also convert posix-cpu-timers to use nsec based internal counters to simplify it. Signed-off-by:
-
Frederic Weisbecker authored
Signed-off-by:
-
Frederic Weisbecker authored
The irqtime is accounted is nsecs and stored in cpu_irq_time.hardirq_time and cpu_irq_time.softirq_time. Once the accumulated amount reaches a new jiffy, this one gets accounted to the kcpustat. This was necessary when kcpustat was stored in cputime_t, which could at worst have jiffies granularity. But now kcpustat is stored in nsecs so this whole discretization game with temporary irqtime storage has become unnecessary. We can now directly account the irqtime to the kcpustat. Signed-off-by:
-
Frederic Weisbecker authored
Use the new nsec based cputime accessors as part of the whole cputime conversion from cputime_t to nsecs. Signed-off-by:
-
Frederic Weisbecker authored
Use the new nsec based cputime accessors as part of the whole cputime conversion from cputime_t to nsecs. Signed-off-by:
-
Frederic Weisbecker authored
Use the new nsec based cputime accessors as part of the whole cputime conversion from cputime_t to nsecs. Signed-off-by:
-
Frederic Weisbecker authored
Use the new nsec based cputime accessors as part of the whole cputime conversion from cputime_t to nsecs. Signed-off-by:
-
Frederic Weisbecker authored
Use the new nsec based cputime accessors as part of the whole cputime conversion from cputime_t to nsecs. Signed-off-by:
-
Frederic Weisbecker authored
Not sure if MISDN stats are ABI but it displays task cputime in cputime_t raw value regardless of what type cputime_t wraps which could be either jiffies, nsecs, usecs, or whatever random time unit. Plus it wrongly assumes that cputime_t is long. Given that this dump is broken anyway, lets just display the nanosec value and stick with that. Signed-off-by:
-
Frederic Weisbecker authored
Use the new nsec based cputime accessors as part of the whole cputime conversion from cputime_t to nsecs. Signed-off-by:
-
Frederic Weisbecker authored
Use the new nsec based cputime accessors as part of the whole cputime conversion from cputime_t to nsecs. Signed-off-by:
-
Frederic Weisbecker authored
Now that most cputime readers use the transition API which return the task cputime in old style cputime_t, we can safely store the cputime in nsecs. This will eventually make cputime statistics less opaque and more granular. Back and forth convertions between cputime_t and nsecs in order to deal with cputime_t random granularity won't be needed anymore. Signed-off-by:
-
Frederic Weisbecker authored
This API returns a task's cputime in cputime_t in order to ease the conversion of cputime internals to use nsecs units instead. Blindly converting all cputime readers to use this API now will later let us convert more smoothly and step by step all these places to use the new nsec based cputime. Signed-off-by:
-
Frederic Weisbecker authored
cputime_t is being obsolete and replaced by nsecs units in order to make internal timestamps less opaque and more granular. Signed-off-by:
-
