[PATCH 0/4] cputime: Virtual cputime accounting small cleanups and consolidation v2

[PATCH 0/4] cputime: Virtual cputime accounting small cleanups and consolidation v2

[Frederic Weisbecker]

Hi,

No fundamental change in this release but a rebase to solve conflicts
against latest tip:/sched/core commits.

Thanks.

Frederic Weisbecker (4):
  cputime: Generalize CONFIG_VIRT_CPU_ACCOUNTING
  sched: Move cputime code to its own file
  cputime: Consolidate vtime handling on context switch
  s390: Remove leftover account_tick_vtime() header

 arch/Kconfig                           |    3 +
 arch/ia64/Kconfig                      |   12 +-
 arch/ia64/include/asm/switch_to.h      |    8 -
 arch/ia64/kernel/time.c                |    4 +-
 arch/powerpc/include/asm/time.h        |    6 -
 arch/powerpc/kernel/process.c          |    3 -
 arch/powerpc/kernel/time.c             |    6 +
 arch/powerpc/platforms/Kconfig.cputype |   16 +-
 arch/s390/Kconfig                      |    5 +-
 arch/s390/include/asm/switch_to.h      |    4 -
 arch/s390/kernel/vtime.c               |    4 +-
 include/linux/kernel_stat.h            |    6 +
 init/Kconfig                           |   13 +
 kernel/sched/Makefile                  |    2 +-
 kernel/sched/core.c                    |  558 +-------------------------------
 kernel/sched/cputime.c                 |  503 ++++++++++++++++++++++++++++
 kernel/sched/sched.h                   |   63 ++++
 17 files changed, 606 insertions(+), 610 deletions(-)
 create mode 100644 kernel/sched/cputime.c

-- 
1.7.5.4

[PATCH 1/4] cputime: Generalize CONFIG_VIRT_CPU_ACCOUNTING

[Frederic Weisbecker]

S390, ia64 and powerpc all define their own version
of CONFIG_VIRT_CPU_ACCOUNTING. Generalize the config
and its description to a single place to avoid
duplication.

Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
---
 arch/Kconfig                           |    3 +++
 arch/ia64/Kconfig                      |   12 +-----------
 arch/powerpc/platforms/Kconfig.cputype |   16 +---------------
 arch/s390/Kconfig                      |    5 ++---
 init/Kconfig                           |   13 +++++++++++++
 5 files changed, 20 insertions(+), 29 deletions(-)

diff --git a/arch/Kconfig b/arch/Kconfig
index 72f2fa1..f78de57 100644
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -281,4 +281,7 @@ config SECCOMP_FILTER
 
 	  See Documentation/prctl/seccomp_filter.txt for details.
 
+config HAVE_VIRT_CPU_ACCOUNTING
+	bool
+
 source "kernel/gcov/Kconfig"
diff --git a/arch/ia64/Kconfig b/arch/ia64/Kconfig
index 310cf57..3c720ef 100644
--- a/arch/ia64/Kconfig
+++ b/arch/ia64/Kconfig
@@ -25,6 +25,7 @@ config IA64
 	select HAVE_GENERIC_HARDIRQS
 	select HAVE_MEMBLOCK
 	select HAVE_MEMBLOCK_NODE_MAP
+	select HAVE_VIRT_CPU_ACCOUNTING
 	select ARCH_DISCARD_MEMBLOCK
 	select GENERIC_IRQ_PROBE
 	select GENERIC_PENDING_IRQ if SMP
@@ -340,17 +341,6 @@ config FORCE_MAX_ZONEORDER
 	default "17" if HUGETLB_PAGE
 	default "11"
 
-config VIRT_CPU_ACCOUNTING
-	bool "Deterministic task and CPU time accounting"
-	default n
-	help
-	  Select this option to enable more accurate task and CPU time
-	  accounting.  This is done by reading a CPU counter on each
-	  kernel entry and exit and on transitions within the kernel
-	  between system, softirq and hardirq state, so there is a
-	  small performance impact.
-	  If in doubt, say N here.
-
 config SMP
 	bool "Symmetric multi-processing support"
 	select USE_GENERIC_SMP_HELPERS
diff --git a/arch/powerpc/platforms/Kconfig.cputype b/arch/powerpc/platforms/Kconfig.cputype
index 30fd01d..72afd28 100644
--- a/arch/powerpc/platforms/Kconfig.cputype
+++ b/arch/powerpc/platforms/Kconfig.cputype
@@ -1,6 +1,7 @@
 config PPC64
 	bool "64-bit kernel"
 	default n
+	select HAVE_VIRT_CPU_ACCOUNTING
 	help
 	  This option selects whether a 32-bit or a 64-bit kernel
 	  will be built.
@@ -337,21 +338,6 @@ config PPC_MM_SLICES
 	default y if (!PPC_FSL_BOOK3E && PPC64 && HUGETLB_PAGE) || (PPC_STD_MMU_64 && PPC_64K_PAGES)
 	default n
 
-config VIRT_CPU_ACCOUNTING
-	bool "Deterministic task and CPU time accounting"
-	depends on PPC64
-	default y
-	help
-	  Select this option to enable more accurate task and CPU time
-	  accounting.  This is done by reading a CPU counter on each
-	  kernel entry and exit and on transitions within the kernel
-	  between system, softirq and hardirq state, so there is a
-	  small performance impact.  This also enables accounting of
-	  stolen time on logically-partitioned systems running on
-	  IBM POWER5-based machines.
-
-	  If in doubt, say Y here.
-
 config PPC_HAVE_PMU_SUPPORT
        bool
 
diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig
index 76de6b6..49ebfb6 100644
--- a/arch/s390/Kconfig
+++ b/arch/s390/Kconfig
@@ -49,9 +49,6 @@ config GENERIC_LOCKBREAK
 config PGSTE
 	def_bool y if KVM
 
-config VIRT_CPU_ACCOUNTING
-	def_bool y
-
 config ARCH_SUPPORTS_DEBUG_PAGEALLOC
 	def_bool y
 
@@ -89,6 +86,8 @@ config S390
 	select HAVE_MEMBLOCK
 	select HAVE_MEMBLOCK_NODE_MAP
 	select HAVE_CMPXCHG_LOCAL
+	select HAVE_VIRT_CPU_ACCOUNTING
+	select VIRT_CPU_ACCOUNTING
 	select ARCH_DISCARD_MEMBLOCK
 	select BUILDTIME_EXTABLE_SORT
 	select ARCH_INLINE_SPIN_TRYLOCK
diff --git a/init/Kconfig b/init/Kconfig
index af6c7f8..894b073 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -267,6 +267,19 @@ config POSIX_MQUEUE_SYSCTL
 	depends on SYSCTL
 	default y
 
+config VIRT_CPU_ACCOUNTING
+	bool "Deterministic task and CPU time accounting"
+	depends on HAVE_VIRT_CPU_ACCOUNTING
+	default y if PPC64
+	help
+	  Select this option to enable more accurate task and CPU time
+	  accounting.  This is done by reading a CPU counter on each
+	  kernel entry and exit and on transitions within the kernel
+	  between system, softirq and hardirq state, so there is a
+	  small performance impact.  This also enables accounting of
+	  stolen time on logically-partitioned systems running on
+	  IBM POWER5-based machines.
+
 config BSD_PROCESS_ACCT
 	bool "BSD Process Accounting"
 	help
-- 
1.7.5.4

[PATCH 2/4] sched: Move cputime code to its own file

[Frederic Weisbecker]

Extract cputime code from the giant sched/core.c and
put it in its own file. This make it easier to deal with
this particular area and de-bloat a bit more core.c

Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
---
 kernel/sched/Makefile  |    2 +-
 kernel/sched/core.c    |  557 +-----------------------------------------------
 kernel/sched/cputime.c |  503 +++++++++++++++++++++++++++++++++++++++++++
 kernel/sched/sched.h   |   63 ++++++
 4 files changed, 569 insertions(+), 556 deletions(-)
 create mode 100644 kernel/sched/cputime.c

diff --git a/kernel/sched/Makefile b/kernel/sched/Makefile
index 173ea52..f06d249 100644
--- a/kernel/sched/Makefile
+++ b/kernel/sched/Makefile
@@ -11,7 +11,7 @@ ifneq ($(CONFIG_SCHED_OMIT_FRAME_POINTER),y)
 CFLAGS_core.o := $(PROFILING) -fno-omit-frame-pointer
 endif
 
-obj-y += core.o clock.o idle_task.o fair.o rt.o stop_task.o
+obj-y += core.o clock.o cputime.o idle_task.o fair.o rt.o stop_task.o
 obj-$(CONFIG_SMP) += cpupri.o
 obj-$(CONFIG_SCHED_AUTOGROUP) += auto_group.o
 obj-$(CONFIG_SCHEDSTATS) += stats.o
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 4376c9f..ae3bcaa 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -740,126 +740,6 @@ void deactivate_task(struct rq *rq, struct task_struct *p, int flags)
 	dequeue_task(rq, p, flags);
 }
 
-#ifdef CONFIG_IRQ_TIME_ACCOUNTING
-
-/*
- * There are no locks covering percpu hardirq/softirq time.
- * They are only modified in account_system_vtime, on corresponding CPU
- * with interrupts disabled. So, writes are safe.
- * They are read and saved off onto struct rq in update_rq_clock().
- * This may result in other CPU reading this CPU's irq time and can
- * race with irq/account_system_vtime on this CPU. We would either get old
- * or new value with a side effect of accounting a slice of irq time to wrong
- * task when irq is in progress while we read rq->clock. That is a worthy
- * compromise in place of having locks on each irq in account_system_time.
- */
-static DEFINE_PER_CPU(u64, cpu_hardirq_time);
-static DEFINE_PER_CPU(u64, cpu_softirq_time);
-
-static DEFINE_PER_CPU(u64, irq_start_time);
-static int sched_clock_irqtime;
-
-void enable_sched_clock_irqtime(void)
-{
-	sched_clock_irqtime = 1;
-}
-
-void disable_sched_clock_irqtime(void)
-{
-	sched_clock_irqtime = 0;
-}
-
-#ifndef CONFIG_64BIT
-static DEFINE_PER_CPU(seqcount_t, irq_time_seq);
-
-static inline void irq_time_write_begin(void)
-{
-	__this_cpu_inc(irq_time_seq.sequence);
-	smp_wmb();
-}
-
-static inline void irq_time_write_end(void)
-{
-	smp_wmb();
-	__this_cpu_inc(irq_time_seq.sequence);
-}
-
-static inline u64 irq_time_read(int cpu)
-{
-	u64 irq_time;
-	unsigned seq;
-
-	do {
-		seq = read_seqcount_begin(&per_cpu(irq_time_seq, cpu));
-		irq_time = per_cpu(cpu_softirq_time, cpu) +
-			   per_cpu(cpu_hardirq_time, cpu);
-	} while (read_seqcount_retry(&per_cpu(irq_time_seq, cpu), seq));
-
-	return irq_time;
-}
-#else /* CONFIG_64BIT */
-static inline void irq_time_write_begin(void)
-{
-}
-
-static inline void irq_time_write_end(void)
-{
-}
-
-static inline u64 irq_time_read(int cpu)
-{
-	return per_cpu(cpu_softirq_time, cpu) + per_cpu(cpu_hardirq_time, cpu);
-}
-#endif /* CONFIG_64BIT */
-
-/*
- * Called before incrementing preempt_count on {soft,}irq_enter
- * and before decrementing preempt_count on {soft,}irq_exit.
- */
-void account_system_vtime(struct task_struct *curr)
-{
-	unsigned long flags;
-	s64 delta;
-	int cpu;
-
-	if (!sched_clock_irqtime)
-		return;
-
-	local_irq_save(flags);
-
-	cpu = smp_processor_id();
-	delta = sched_clock_cpu(cpu) - __this_cpu_read(irq_start_time);
-	__this_cpu_add(irq_start_time, delta);
-
-	irq_time_write_begin();
-	/*
-	 * We do not account for softirq time from ksoftirqd here.
-	 * We want to continue accounting softirq time to ksoftirqd thread
-	 * in that case, so as not to confuse scheduler with a special task
-	 * that do not consume any time, but still wants to run.
-	 */
-	if (hardirq_count())
-		__this_cpu_add(cpu_hardirq_time, delta);
-	else if (in_serving_softirq() && curr != this_cpu_ksoftirqd())
-		__this_cpu_add(cpu_softirq_time, delta);
-
-	irq_time_write_end();
-	local_irq_restore(flags);
-}
-EXPORT_SYMBOL_GPL(account_system_vtime);
-
-#endif /* CONFIG_IRQ_TIME_ACCOUNTING */
-
-#ifdef CONFIG_PARAVIRT
-static inline u64 steal_ticks(u64 steal)
-{
-	if (unlikely(steal > NSEC_PER_SEC))
-		return div_u64(steal, TICK_NSEC);
-
-	return __iter_div_u64_rem(steal, TICK_NSEC, &steal);
-}
-#endif
-
 static void update_rq_clock_task(struct rq *rq, s64 delta)
 {
 /*
@@ -920,43 +800,6 @@ static void update_rq_clock_task(struct rq *rq, s64 delta)
 #endif
 }
 
-#ifdef CONFIG_IRQ_TIME_ACCOUNTING
-static int irqtime_account_hi_update(void)
-{
-	u64 *cpustat = kcpustat_this_cpu->cpustat;
-	unsigned long flags;
-	u64 latest_ns;
-	int ret = 0;
-
-	local_irq_save(flags);
-	latest_ns = this_cpu_read(cpu_hardirq_time);
-	if (nsecs_to_cputime64(latest_ns) > cpustat[CPUTIME_IRQ])
-		ret = 1;
-	local_irq_restore(flags);
-	return ret;
-}
-
-static int irqtime_account_si_update(void)
-{
-	u64 *cpustat = kcpustat_this_cpu->cpustat;
-	unsigned long flags;
-	u64 latest_ns;
-	int ret = 0;
-
-	local_irq_save(flags);
-	latest_ns = this_cpu_read(cpu_softirq_time);
-	if (nsecs_to_cputime64(latest_ns) > cpustat[CPUTIME_SOFTIRQ])
-		ret = 1;
-	local_irq_restore(flags);
-	return ret;
-}
-
-#else /* CONFIG_IRQ_TIME_ACCOUNTING */
-
-#define sched_clock_irqtime	(0)
-
-#endif
-
 void sched_set_stop_task(int cpu, struct task_struct *stop)
 {
 	struct sched_param param = { .sched_priority = MAX_RT_PRIO - 1 };
@@ -2809,404 +2652,6 @@ unsigned long long task_sched_runtime(struct task_struct *p)
 	return ns;
 }
 
-#ifdef CONFIG_CGROUP_CPUACCT
-struct cgroup_subsys cpuacct_subsys;
-struct cpuacct root_cpuacct;
-#endif
-
-static inline void task_group_account_field(struct task_struct *p, int index,
-					    u64 tmp)
-{
-#ifdef CONFIG_CGROUP_CPUACCT
-	struct kernel_cpustat *kcpustat;
-	struct cpuacct *ca;
-#endif
-	/*
-	 * Since all updates are sure to touch the root cgroup, we
-	 * get ourselves ahead and touch it first. If the root cgroup
-	 * is the only cgroup, then nothing else should be necessary.
-	 *
-	 */
-	__get_cpu_var(kernel_cpustat).cpustat[index] += tmp;
-
-#ifdef CONFIG_CGROUP_CPUACCT
-	if (unlikely(!cpuacct_subsys.active))
-		return;
-
-	rcu_read_lock();
-	ca = task_ca(p);
-	while (ca && (ca != &root_cpuacct)) {
-		kcpustat = this_cpu_ptr(ca->cpustat);
-		kcpustat->cpustat[index] += tmp;
-		ca = parent_ca(ca);
-	}
-	rcu_read_unlock();
-#endif
-}
-
-
-/*
- * Account user cpu time to a process.
- * @p: the process that the cpu time gets accounted to
- * @cputime: the cpu time spent in user space since the last update
- * @cputime_scaled: cputime scaled by cpu frequency
- */
-void account_user_time(struct task_struct *p, cputime_t cputime,
-		       cputime_t cputime_scaled)
-{
-	int index;
-
-	/* Add user time to process. */
-	p->utime += cputime;
-	p->utimescaled += cputime_scaled;
-	account_group_user_time(p, cputime);
-
-	index = (TASK_NICE(p) > 0) ? CPUTIME_NICE : CPUTIME_USER;
-
-	/* Add user time to cpustat. */
-	task_group_account_field(p, index, (__force u64) cputime);
-
-	/* Account for user time used */
-	acct_update_integrals(p);
-}
-
-/*
- * Account guest cpu time to a process.
- * @p: the process that the cpu time gets accounted to
- * @cputime: the cpu time spent in virtual machine since the last update
- * @cputime_scaled: cputime scaled by cpu frequency
- */
-static void account_guest_time(struct task_struct *p, cputime_t cputime,
-			       cputime_t cputime_scaled)
-{
-	u64 *cpustat = kcpustat_this_cpu->cpustat;
-
-	/* Add guest time to process. */
-	p->utime += cputime;
-	p->utimescaled += cputime_scaled;
-	account_group_user_time(p, cputime);
-	p->gtime += cputime;
-
-	/* Add guest time to cpustat. */
-	if (TASK_NICE(p) > 0) {
-		cpustat[CPUTIME_NICE] += (__force u64) cputime;
-		cpustat[CPUTIME_GUEST_NICE] += (__force u64) cputime;
-	} else {
-		cpustat[CPUTIME_USER] += (__force u64) cputime;
-		cpustat[CPUTIME_GUEST] += (__force u64) cputime;
-	}
-}
-
-/*
- * Account system cpu time to a process and desired cpustat field
- * @p: the process that the cpu time gets accounted to
- * @cputime: the cpu time spent in kernel space since the last update
- * @cputime_scaled: cputime scaled by cpu frequency
- * @target_cputime64: pointer to cpustat field that has to be updated
- */
-static inline
-void __account_system_time(struct task_struct *p, cputime_t cputime,
-			cputime_t cputime_scaled, int index)
-{
-	/* Add system time to process. */
-	p->stime += cputime;
-	p->stimescaled += cputime_scaled;
-	account_group_system_time(p, cputime);
-
-	/* Add system time to cpustat. */
-	task_group_account_field(p, index, (__force u64) cputime);
-
-	/* Account for system time used */
-	acct_update_integrals(p);
-}
-
-/*
- * Account system cpu time to a process.
- * @p: the process that the cpu time gets accounted to
- * @hardirq_offset: the offset to subtract from hardirq_count()
- * @cputime: the cpu time spent in kernel space since the last update
- * @cputime_scaled: cputime scaled by cpu frequency
- */
-void account_system_time(struct task_struct *p, int hardirq_offset,
-			 cputime_t cputime, cputime_t cputime_scaled)
-{
-	int index;
-
-	if ((p->flags & PF_VCPU) && (irq_count() - hardirq_offset == 0)) {
-		account_guest_time(p, cputime, cputime_scaled);
-		return;
-	}
-
-	if (hardirq_count() - hardirq_offset)
-		index = CPUTIME_IRQ;
-	else if (in_serving_softirq())
-		index = CPUTIME_SOFTIRQ;
-	else
-		index = CPUTIME_SYSTEM;
-
-	__account_system_time(p, cputime, cputime_scaled, index);
-}
-
-/*
- * Account for involuntary wait time.
- * @cputime: the cpu time spent in involuntary wait
- */
-void account_steal_time(cputime_t cputime)
-{
-	u64 *cpustat = kcpustat_this_cpu->cpustat;
-
-	cpustat[CPUTIME_STEAL] += (__force u64) cputime;
-}
-
-/*
- * Account for idle time.
- * @cputime: the cpu time spent in idle wait
- */
-void account_idle_time(cputime_t cputime)
-{
-	u64 *cpustat = kcpustat_this_cpu->cpustat;
-	struct rq *rq = this_rq();
-
-	if (atomic_read(&rq->nr_iowait) > 0)
-		cpustat[CPUTIME_IOWAIT] += (__force u64) cputime;
-	else
-		cpustat[CPUTIME_IDLE] += (__force u64) cputime;
-}
-
-static __always_inline bool steal_account_process_tick(void)
-{
-#ifdef CONFIG_PARAVIRT
-	if (static_key_false(&paravirt_steal_enabled)) {
-		u64 steal, st = 0;
-
-		steal = paravirt_steal_clock(smp_processor_id());
-		steal -= this_rq()->prev_steal_time;
-
-		st = steal_ticks(steal);
-		this_rq()->prev_steal_time += st * TICK_NSEC;
-
-		account_steal_time(st);
-		return st;
-	}
-#endif
-	return false;
-}
-
-#ifndef CONFIG_VIRT_CPU_ACCOUNTING
-
-#ifdef CONFIG_IRQ_TIME_ACCOUNTING
-/*
- * Account a tick to a process and cpustat
- * @p: the process that the cpu time gets accounted to
- * @user_tick: is the tick from userspace
- * @rq: the pointer to rq
- *
- * Tick demultiplexing follows the order
- * - pending hardirq update
- * - pending softirq update
- * - user_time
- * - idle_time
- * - system time
- *   - check for guest_time
- *   - else account as system_time
- *
- * Check for hardirq is done both for system and user time as there is
- * no timer going off while we are on hardirq and hence we may never get an
- * opportunity to update it solely in system time.
- * p->stime and friends are only updated on system time and not on irq
- * softirq as those do not count in task exec_runtime any more.
- */
-static void irqtime_account_process_tick(struct task_struct *p, int user_tick,
-						struct rq *rq)
-{
-	cputime_t one_jiffy_scaled = cputime_to_scaled(cputime_one_jiffy);
-	u64 *cpustat = kcpustat_this_cpu->cpustat;
-
-	if (steal_account_process_tick())
-		return;
-
-	if (irqtime_account_hi_update()) {
-		cpustat[CPUTIME_IRQ] += (__force u64) cputime_one_jiffy;
-	} else if (irqtime_account_si_update()) {
-		cpustat[CPUTIME_SOFTIRQ] += (__force u64) cputime_one_jiffy;
-	} else if (this_cpu_ksoftirqd() == p) {
-		/*
-		 * ksoftirqd time do not get accounted in cpu_softirq_time.
-		 * So, we have to handle it separately here.
-		 * Also, p->stime needs to be updated for ksoftirqd.
-		 */
-		__account_system_time(p, cputime_one_jiffy, one_jiffy_scaled,
-					CPUTIME_SOFTIRQ);
-	} else if (user_tick) {
-		account_user_time(p, cputime_one_jiffy, one_jiffy_scaled);
-	} else if (p == rq->idle) {
-		account_idle_time(cputime_one_jiffy);
-	} else if (p->flags & PF_VCPU) { /* System time or guest time */
-		account_guest_time(p, cputime_one_jiffy, one_jiffy_scaled);
-	} else {
-		__account_system_time(p, cputime_one_jiffy, one_jiffy_scaled,
-					CPUTIME_SYSTEM);
-	}
-}
-
-static void irqtime_account_idle_ticks(int ticks)
-{
-	int i;
-	struct rq *rq = this_rq();
-
-	for (i = 0; i < ticks; i++)
-		irqtime_account_process_tick(current, 0, rq);
-}
-#else /* CONFIG_IRQ_TIME_ACCOUNTING */
-static void irqtime_account_idle_ticks(int ticks) {}
-static void irqtime_account_process_tick(struct task_struct *p, int user_tick,
-						struct rq *rq) {}
-#endif /* CONFIG_IRQ_TIME_ACCOUNTING */
-
-/*
- * Account a single tick of cpu time.
- * @p: the process that the cpu time gets accounted to
- * @user_tick: indicates if the tick is a user or a system tick
- */
-void account_process_tick(struct task_struct *p, int user_tick)
-{
-	cputime_t one_jiffy_scaled = cputime_to_scaled(cputime_one_jiffy);
-	struct rq *rq = this_rq();
-
-	if (sched_clock_irqtime) {
-		irqtime_account_process_tick(p, user_tick, rq);
-		return;
-	}
-
-	if (steal_account_process_tick())
-		return;
-
-	if (user_tick)
-		account_user_time(p, cputime_one_jiffy, one_jiffy_scaled);
-	else if ((p != rq->idle) || (irq_count() != HARDIRQ_OFFSET))
-		account_system_time(p, HARDIRQ_OFFSET, cputime_one_jiffy,
-				    one_jiffy_scaled);
-	else
-		account_idle_time(cputime_one_jiffy);
-}
-
-/*
- * Account multiple ticks of steal time.
- * @p: the process from which the cpu time has been stolen
- * @ticks: number of stolen ticks
- */
-void account_steal_ticks(unsigned long ticks)
-{
-	account_steal_time(jiffies_to_cputime(ticks));
-}
-
-/*
- * Account multiple ticks of idle time.
- * @ticks: number of stolen ticks
- */
-void account_idle_ticks(unsigned long ticks)
-{
-
-	if (sched_clock_irqtime) {
-		irqtime_account_idle_ticks(ticks);
-		return;
-	}
-
-	account_idle_time(jiffies_to_cputime(ticks));
-}
-
-#endif
-
-/*
- * Use precise platform statistics if available:
- */
-#ifdef CONFIG_VIRT_CPU_ACCOUNTING
-void task_times(struct task_struct *p, cputime_t *ut, cputime_t *st)
-{
-	*ut = p->utime;
-	*st = p->stime;
-}
-
-void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t *st)
-{
-	struct task_cputime cputime;
-
-	thread_group_cputime(p, &cputime);
-
-	*ut = cputime.utime;
-	*st = cputime.stime;
-}
-#else
-
-#ifndef nsecs_to_cputime
-# define nsecs_to_cputime(__nsecs)	nsecs_to_jiffies(__nsecs)
-#endif
-
-static cputime_t scale_utime(cputime_t utime, cputime_t rtime, cputime_t total)
-{
-	u64 temp = (__force u64) rtime;
-
-	temp *= (__force u64) utime;
-
-	if (sizeof(cputime_t) == 4)
-		temp = div_u64(temp, (__force u32) total);
-	else
-		temp = div64_u64(temp, (__force u64) total);
-
-	return (__force cputime_t) temp;
-}
-
-void task_times(struct task_struct *p, cputime_t *ut, cputime_t *st)
-{
-	cputime_t rtime, utime = p->utime, total = utime + p->stime;
-
-	/*
-	 * Use CFS's precise accounting:
-	 */
-	rtime = nsecs_to_cputime(p->se.sum_exec_runtime);
-
-	if (total)
-		utime = scale_utime(utime, rtime, total);
-	else
-		utime = rtime;
-
-	/*
-	 * Compare with previous values, to keep monotonicity:
-	 */
-	p->prev_utime = max(p->prev_utime, utime);
-	p->prev_stime = max(p->prev_stime, rtime - p->prev_utime);
-
-	*ut = p->prev_utime;
-	*st = p->prev_stime;
-}
-
-/*
- * Must be called with siglock held.
- */
-void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t *st)
-{
-	struct signal_struct *sig = p->signal;
-	struct task_cputime cputime;
-	cputime_t rtime, utime, total;
-
-	thread_group_cputime(p, &cputime);
-
-	total = cputime.utime + cputime.stime;
-	rtime = nsecs_to_cputime(cputime.sum_exec_runtime);
-
-	if (total)
-		utime = scale_utime(cputime.utime, rtime, total);
-	else
-		utime = rtime;
-
-	sig->prev_utime = max(sig->prev_utime, utime);
-	sig->prev_stime = max(sig->prev_stime, rtime - sig->prev_utime);
-
-	*ut = sig->prev_utime;
-	*st = sig->prev_stime;
-}
-#endif
-
 /*
  * This function gets called by the timer code, with HZ frequency.
  * We call it with interrupts disabled.
@@ -8419,6 +7864,8 @@ struct cgroup_subsys cpu_cgroup_subsys = {
  * (balbir@in.ibm.com).
  */
 
+struct cpuacct root_cpuacct;
+
 /* create a new cpu accounting group */
 static struct cgroup_subsys_state *cpuacct_create(struct cgroup *cgrp)
 {
diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c
new file mode 100644
index 0000000..ea9b4b6
--- /dev/null
+++ b/kernel/sched/cputime.c
@@ -0,0 +1,503 @@
+#include <linux/export.h>
+#include <linux/sched.h>
+#include <linux/tsacct_kern.h>
+#include <linux/kernel_stat.h>
+#include "sched.h"
+
+
+#ifdef CONFIG_IRQ_TIME_ACCOUNTING
+
+/*
+ * There are no locks covering percpu hardirq/softirq time.
+ * They are only modified in account_system_vtime, on corresponding CPU
+ * with interrupts disabled. So, writes are safe.
+ * They are read and saved off onto struct rq in update_rq_clock().
+ * This may result in other CPU reading this CPU's irq time and can
+ * race with irq/account_system_vtime on this CPU. We would either get old
+ * or new value with a side effect of accounting a slice of irq time to wrong
+ * task when irq is in progress while we read rq->clock. That is a worthy
+ * compromise in place of having locks on each irq in account_system_time.
+ */
+DEFINE_PER_CPU(u64, cpu_hardirq_time);
+DEFINE_PER_CPU(u64, cpu_softirq_time);
+
+static DEFINE_PER_CPU(u64, irq_start_time);
+static int sched_clock_irqtime;
+
+void enable_sched_clock_irqtime(void)
+{
+	sched_clock_irqtime = 1;
+}
+
+void disable_sched_clock_irqtime(void)
+{
+	sched_clock_irqtime = 0;
+}
+
+#ifndef CONFIG_64BIT
+DEFINE_PER_CPU(seqcount_t, irq_time_seq);
+#endif /* CONFIG_64BIT */
+
+/*
+ * Called before incrementing preempt_count on {soft,}irq_enter
+ * and before decrementing preempt_count on {soft,}irq_exit.
+ */
+void account_system_vtime(struct task_struct *curr)
+{
+	unsigned long flags;
+	s64 delta;
+	int cpu;
+
+	if (!sched_clock_irqtime)
+		return;
+
+	local_irq_save(flags);
+
+	cpu = smp_processor_id();
+	delta = sched_clock_cpu(cpu) - __this_cpu_read(irq_start_time);
+	__this_cpu_add(irq_start_time, delta);
+
+	irq_time_write_begin();
+	/*
+	 * We do not account for softirq time from ksoftirqd here.
+	 * We want to continue accounting softirq time to ksoftirqd thread
+	 * in that case, so as not to confuse scheduler with a special task
+	 * that do not consume any time, but still wants to run.
+	 */
+	if (hardirq_count())
+		__this_cpu_add(cpu_hardirq_time, delta);
+	else if (in_serving_softirq() && curr != this_cpu_ksoftirqd())
+		__this_cpu_add(cpu_softirq_time, delta);
+
+	irq_time_write_end();
+	local_irq_restore(flags);
+}
+EXPORT_SYMBOL_GPL(account_system_vtime);
+
+static int irqtime_account_hi_update(void)
+{
+	u64 *cpustat = kcpustat_this_cpu->cpustat;
+	unsigned long flags;
+	u64 latest_ns;
+	int ret = 0;
+
+	local_irq_save(flags);
+	latest_ns = this_cpu_read(cpu_hardirq_time);
+	if (nsecs_to_cputime64(latest_ns) > cpustat[CPUTIME_IRQ])
+		ret = 1;
+	local_irq_restore(flags);
+	return ret;
+}
+
+static int irqtime_account_si_update(void)
+{
+	u64 *cpustat = kcpustat_this_cpu->cpustat;
+	unsigned long flags;
+	u64 latest_ns;
+	int ret = 0;
+
+	local_irq_save(flags);
+	latest_ns = this_cpu_read(cpu_softirq_time);
+	if (nsecs_to_cputime64(latest_ns) > cpustat[CPUTIME_SOFTIRQ])
+		ret = 1;
+	local_irq_restore(flags);
+	return ret;
+}
+
+#else /* CONFIG_IRQ_TIME_ACCOUNTING */
+
+#define sched_clock_irqtime	(0)
+
+#endif /* !CONFIG_IRQ_TIME_ACCOUNTING */
+
+static inline void task_group_account_field(struct task_struct *p, int index,
+					    u64 tmp)
+{
+#ifdef CONFIG_CGROUP_CPUACCT
+	struct kernel_cpustat *kcpustat;
+	struct cpuacct *ca;
+#endif
+	/*
+	 * Since all updates are sure to touch the root cgroup, we
+	 * get ourselves ahead and touch it first. If the root cgroup
+	 * is the only cgroup, then nothing else should be necessary.
+	 *
+	 */
+	__get_cpu_var(kernel_cpustat).cpustat[index] += tmp;
+
+#ifdef CONFIG_CGROUP_CPUACCT
+	if (unlikely(!cpuacct_subsys.active))
+		return;
+
+	rcu_read_lock();
+	ca = task_ca(p);
+	while (ca && (ca != &root_cpuacct)) {
+		kcpustat = this_cpu_ptr(ca->cpustat);
+		kcpustat->cpustat[index] += tmp;
+		ca = parent_ca(ca);
+	}
+	rcu_read_unlock();
+#endif
+}
+
+/*
+ * Account user cpu time to a process.
+ * @p: the process that the cpu time gets accounted to
+ * @cputime: the cpu time spent in user space since the last update
+ * @cputime_scaled: cputime scaled by cpu frequency
+ */
+void account_user_time(struct task_struct *p, cputime_t cputime,
+		       cputime_t cputime_scaled)
+{
+	int index;
+
+	/* Add user time to process. */
+	p->utime += cputime;
+	p->utimescaled += cputime_scaled;
+	account_group_user_time(p, cputime);
+
+	index = (TASK_NICE(p) > 0) ? CPUTIME_NICE : CPUTIME_USER;
+
+	/* Add user time to cpustat. */
+	task_group_account_field(p, index, (__force u64) cputime);
+
+	/* Account for user time used */
+	acct_update_integrals(p);
+}
+
+/*
+ * Account guest cpu time to a process.
+ * @p: the process that the cpu time gets accounted to
+ * @cputime: the cpu time spent in virtual machine since the last update
+ * @cputime_scaled: cputime scaled by cpu frequency
+ */
+static void account_guest_time(struct task_struct *p, cputime_t cputime,
+			       cputime_t cputime_scaled)
+{
+	u64 *cpustat = kcpustat_this_cpu->cpustat;
+
+	/* Add guest time to process. */
+	p->utime += cputime;
+	p->utimescaled += cputime_scaled;
+	account_group_user_time(p, cputime);
+	p->gtime += cputime;
+
+	/* Add guest time to cpustat. */
+	if (TASK_NICE(p) > 0) {
+		cpustat[CPUTIME_NICE] += (__force u64) cputime;
+		cpustat[CPUTIME_GUEST_NICE] += (__force u64) cputime;
+	} else {
+		cpustat[CPUTIME_USER] += (__force u64) cputime;
+		cpustat[CPUTIME_GUEST] += (__force u64) cputime;
+	}
+}
+
+/*
+ * Account system cpu time to a process and desired cpustat field
+ * @p: the process that the cpu time gets accounted to
+ * @cputime: the cpu time spent in kernel space since the last update
+ * @cputime_scaled: cputime scaled by cpu frequency
+ * @target_cputime64: pointer to cpustat field that has to be updated
+ */
+static inline
+void __account_system_time(struct task_struct *p, cputime_t cputime,
+			cputime_t cputime_scaled, int index)
+{
+	/* Add system time to process. */
+	p->stime += cputime;
+	p->stimescaled += cputime_scaled;
+	account_group_system_time(p, cputime);
+
+	/* Add system time to cpustat. */
+	task_group_account_field(p, index, (__force u64) cputime);
+
+	/* Account for system time used */
+	acct_update_integrals(p);
+}
+
+/*
+ * Account system cpu time to a process.
+ * @p: the process that the cpu time gets accounted to
+ * @hardirq_offset: the offset to subtract from hardirq_count()
+ * @cputime: the cpu time spent in kernel space since the last update
+ * @cputime_scaled: cputime scaled by cpu frequency
+ */
+void account_system_time(struct task_struct *p, int hardirq_offset,
+			 cputime_t cputime, cputime_t cputime_scaled)
+{
+	int index;
+
+	if ((p->flags & PF_VCPU) && (irq_count() - hardirq_offset == 0)) {
+		account_guest_time(p, cputime, cputime_scaled);
+		return;
+	}
+
+	if (hardirq_count() - hardirq_offset)
+		index = CPUTIME_IRQ;
+	else if (in_serving_softirq())
+		index = CPUTIME_SOFTIRQ;
+	else
+		index = CPUTIME_SYSTEM;
+
+	__account_system_time(p, cputime, cputime_scaled, index);
+}
+
+/*
+ * Account for involuntary wait time.
+ * @cputime: the cpu time spent in involuntary wait
+ */
+void account_steal_time(cputime_t cputime)
+{
+	u64 *cpustat = kcpustat_this_cpu->cpustat;
+
+	cpustat[CPUTIME_STEAL] += (__force u64) cputime;
+}
+
+/*
+ * Account for idle time.
+ * @cputime: the cpu time spent in idle wait
+ */
+void account_idle_time(cputime_t cputime)
+{
+	u64 *cpustat = kcpustat_this_cpu->cpustat;
+	struct rq *rq = this_rq();
+
+	if (atomic_read(&rq->nr_iowait) > 0)
+		cpustat[CPUTIME_IOWAIT] += (__force u64) cputime;
+	else
+		cpustat[CPUTIME_IDLE] += (__force u64) cputime;
+}
+
+static __always_inline bool steal_account_process_tick(void)
+{
+#ifdef CONFIG_PARAVIRT
+	if (static_key_false(&paravirt_steal_enabled)) {
+		u64 steal, st = 0;
+
+		steal = paravirt_steal_clock(smp_processor_id());
+		steal -= this_rq()->prev_steal_time;
+
+		st = steal_ticks(steal);
+		this_rq()->prev_steal_time += st * TICK_NSEC;
+
+		account_steal_time(st);
+		return st;
+	}
+#endif
+	return false;
+}
+
+#ifndef CONFIG_VIRT_CPU_ACCOUNTING
+
+#ifdef CONFIG_IRQ_TIME_ACCOUNTING
+/*
+ * Account a tick to a process and cpustat
+ * @p: the process that the cpu time gets accounted to
+ * @user_tick: is the tick from userspace
+ * @rq: the pointer to rq
+ *
+ * Tick demultiplexing follows the order
+ * - pending hardirq update
+ * - pending softirq update
+ * - user_time
+ * - idle_time
+ * - system time
+ *   - check for guest_time
+ *   - else account as system_time
+ *
+ * Check for hardirq is done both for system and user time as there is
+ * no timer going off while we are on hardirq and hence we may never get an
+ * opportunity to update it solely in system time.
+ * p->stime and friends are only updated on system time and not on irq
+ * softirq as those do not count in task exec_runtime any more.
+ */
+static void irqtime_account_process_tick(struct task_struct *p, int user_tick,
+						struct rq *rq)
+{
+	cputime_t one_jiffy_scaled = cputime_to_scaled(cputime_one_jiffy);
+	u64 *cpustat = kcpustat_this_cpu->cpustat;
+
+	if (steal_account_process_tick())
+		return;
+
+	if (irqtime_account_hi_update()) {
+		cpustat[CPUTIME_IRQ] += (__force u64) cputime_one_jiffy;
+	} else if (irqtime_account_si_update()) {
+		cpustat[CPUTIME_SOFTIRQ] += (__force u64) cputime_one_jiffy;
+	} else if (this_cpu_ksoftirqd() == p) {
+		/*
+		 * ksoftirqd time do not get accounted in cpu_softirq_time.
+		 * So, we have to handle it separately here.
+		 * Also, p->stime needs to be updated for ksoftirqd.
+		 */
+		__account_system_time(p, cputime_one_jiffy, one_jiffy_scaled,
+					CPUTIME_SOFTIRQ);
+	} else if (user_tick) {
+		account_user_time(p, cputime_one_jiffy, one_jiffy_scaled);
+	} else if (p == rq->idle) {
+		account_idle_time(cputime_one_jiffy);
+	} else if (p->flags & PF_VCPU) { /* System time or guest time */
+		account_guest_time(p, cputime_one_jiffy, one_jiffy_scaled);
+	} else {
+		__account_system_time(p, cputime_one_jiffy, one_jiffy_scaled,
+					CPUTIME_SYSTEM);
+	}
+}
+
+static void irqtime_account_idle_ticks(int ticks)
+{
+	int i;
+	struct rq *rq = this_rq();
+
+	for (i = 0; i < ticks; i++)
+		irqtime_account_process_tick(current, 0, rq);
+}
+#else /* CONFIG_IRQ_TIME_ACCOUNTING */
+static void irqtime_account_idle_ticks(int ticks) {}
+static void irqtime_account_process_tick(struct task_struct *p, int user_tick,
+						struct rq *rq) {}
+#endif /* CONFIG_IRQ_TIME_ACCOUNTING */
+
+/*
+ * Account a single tick of cpu time.
+ * @p: the process that the cpu time gets accounted to
+ * @user_tick: indicates if the tick is a user or a system tick
+ */
+void account_process_tick(struct task_struct *p, int user_tick)
+{
+	cputime_t one_jiffy_scaled = cputime_to_scaled(cputime_one_jiffy);
+	struct rq *rq = this_rq();
+
+	if (sched_clock_irqtime) {
+		irqtime_account_process_tick(p, user_tick, rq);
+		return;
+	}
+
+	if (steal_account_process_tick())
+		return;
+
+	if (user_tick)
+		account_user_time(p, cputime_one_jiffy, one_jiffy_scaled);
+	else if ((p != rq->idle) || (irq_count() != HARDIRQ_OFFSET))
+		account_system_time(p, HARDIRQ_OFFSET, cputime_one_jiffy,
+				    one_jiffy_scaled);
+	else
+		account_idle_time(cputime_one_jiffy);
+}
+
+/*
+ * Account multiple ticks of steal time.
+ * @p: the process from which the cpu time has been stolen
+ * @ticks: number of stolen ticks
+ */
+void account_steal_ticks(unsigned long ticks)
+{
+	account_steal_time(jiffies_to_cputime(ticks));
+}
+
+/*
+ * Account multiple ticks of idle time.
+ * @ticks: number of stolen ticks
+ */
+void account_idle_ticks(unsigned long ticks)
+{
+
+	if (sched_clock_irqtime) {
+		irqtime_account_idle_ticks(ticks);
+		return;
+	}
+
+	account_idle_time(jiffies_to_cputime(ticks));
+}
+
+#endif
+
+/*
+ * Use precise platform statistics if available:
+ */
+#ifdef CONFIG_VIRT_CPU_ACCOUNTING
+void task_times(struct task_struct *p, cputime_t *ut, cputime_t *st)
+{
+	*ut = p->utime;
+	*st = p->stime;
+}
+
+void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t *st)
+{
+	struct task_cputime cputime;
+
+	thread_group_cputime(p, &cputime);
+
+	*ut = cputime.utime;
+	*st = cputime.stime;
+}
+#else
+
+#ifndef nsecs_to_cputime
+# define nsecs_to_cputime(__nsecs)	nsecs_to_jiffies(__nsecs)
+#endif
+
+static cputime_t scale_utime(cputime_t utime, cputime_t rtime, cputime_t total)
+{
+	u64 temp = (__force u64) rtime;
+
+	temp *= (__force u64) utime;
+
+	if (sizeof(cputime_t) == 4)
+		temp = div_u64(temp, (__force u32) total);
+	else
+		temp = div64_u64(temp, (__force u64) total);
+
+	return (__force cputime_t) temp;
+}
+
+void task_times(struct task_struct *p, cputime_t *ut, cputime_t *st)
+{
+	cputime_t rtime, utime = p->utime, total = utime + p->stime;
+
+	/*
+	 * Use CFS's precise accounting:
+	 */
+	rtime = nsecs_to_cputime(p->se.sum_exec_runtime);
+
+	if (total)
+		utime = scale_utime(utime, rtime, total);
+	else
+		utime = rtime;
+
+	/*
+	 * Compare with previous values, to keep monotonicity:
+	 */
+	p->prev_utime = max(p->prev_utime, utime);
+	p->prev_stime = max(p->prev_stime, rtime - p->prev_utime);
+
+	*ut = p->prev_utime;
+	*st = p->prev_stime;
+}
+
+/*
+ * Must be called with siglock held.
+ */
+void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t *st)
+{
+	struct signal_struct *sig = p->signal;
+	struct task_cputime cputime;
+	cputime_t rtime, utime, total;
+
+	thread_group_cputime(p, &cputime);
+
+	total = cputime.utime + cputime.stime;
+	rtime = nsecs_to_cputime(cputime.sum_exec_runtime);
+
+	if (total)
+		utime = scale_utime(cputime.utime, rtime, total);
+	else
+		utime = rtime;
+
+	sig->prev_utime = max(sig->prev_utime, utime);
+	sig->prev_stime = max(sig->prev_stime, rtime - sig->prev_utime);
+
+	*ut = sig->prev_utime;
+	*st = sig->prev_stime;
+}
+#endif
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index f6714d0..62f9850 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -891,6 +891,9 @@ struct cpuacct {
 	struct kernel_cpustat __percpu *cpustat;
 };
 
+extern struct cgroup_subsys cpuacct_subsys;
+extern struct cpuacct root_cpuacct;
+
 /* return cpu accounting group corresponding to this container */
 static inline struct cpuacct *cgroup_ca(struct cgroup *cgrp)
 {
@@ -917,6 +920,16 @@ extern void cpuacct_charge(struct task_struct *tsk, u64 cputime);
 static inline void cpuacct_charge(struct task_struct *tsk, u64 cputime) {}
 #endif
 
+#ifdef CONFIG_PARAVIRT
+static inline u64 steal_ticks(u64 steal)
+{
+	if (unlikely(steal > NSEC_PER_SEC))
+		return div_u64(steal, TICK_NSEC);
+
+	return __iter_div_u64_rem(steal, TICK_NSEC, &steal);
+}
+#endif
+
 static inline void inc_nr_running(struct rq *rq)
 {
 	rq->nr_running++;
@@ -1156,4 +1169,54 @@ enum rq_nohz_flag_bits {
 };
 
 #define nohz_flags(cpu)	(&cpu_rq(cpu)->nohz_flags)
+
+#ifdef CONFIG_IRQ_TIME_ACCOUNTING
+
+DECLARE_PER_CPU(u64, cpu_hardirq_time);
+DECLARE_PER_CPU(u64, cpu_softirq_time);
+
+#ifndef CONFIG_64BIT
+DECLARE_PER_CPU(seqcount_t, irq_time_seq);
+
+static inline void irq_time_write_begin(void)
+{
+	__this_cpu_inc(irq_time_seq.sequence);
+	smp_wmb();
+}
+
+static inline void irq_time_write_end(void)
+{
+	smp_wmb();
+	__this_cpu_inc(irq_time_seq.sequence);
+}
+
+static inline u64 irq_time_read(int cpu)
+{
+	u64 irq_time;
+	unsigned seq;
+
+	do {
+		seq = read_seqcount_begin(&per_cpu(irq_time_seq, cpu));
+		irq_time = per_cpu(cpu_softirq_time, cpu) +
+			   per_cpu(cpu_hardirq_time, cpu);
+	} while (read_seqcount_retry(&per_cpu(irq_time_seq, cpu), seq));
+
+	return irq_time;
+}
+#else /* CONFIG_64BIT */
+static inline void irq_time_write_begin(void)
+{
+}
+
+static inline void irq_time_write_end(void)
+{
+}
+
+static inline u64 irq_time_read(int cpu)
+{
+	return per_cpu(cpu_softirq_time, cpu) + per_cpu(cpu_hardirq_time, cpu);
+}
+#endif /* CONFIG_64BIT */
+#endif /* CONFIG_IRQ_TIME_ACCOUNTING */
+
 #endif
-- 
1.7.5.4

[PATCH 3/4] cputime: Consolidate vtime handling on context switch

[Frederic Weisbecker]

The archs that implement virtual cputime accounting all
flush the cputime of a task when it gets descheduled
and sometimes set up some ground initialization for the
next task to account its cputime.

These archs all put their own hooks in their context
switch callbacks and handle the off-case themselves.

Consolidate this by creating a new account_switch_vtime()
callback called in generic code right after a context switch
and that these archs must implement to flush the prev task
cputime and initialize the next task cputime related state.

Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
---
 arch/ia64/include/asm/switch_to.h |    8 --------
 arch/ia64/kernel/time.c           |    4 ++--
 arch/powerpc/include/asm/time.h   |    6 ------
 arch/powerpc/kernel/process.c     |    3 ---
 arch/powerpc/kernel/time.c        |    6 ++++++
 arch/s390/include/asm/switch_to.h |    2 --
 arch/s390/kernel/vtime.c          |    4 ++--
 include/linux/kernel_stat.h       |    6 ++++++
 kernel/sched/core.c               |    1 +
 9 files changed, 17 insertions(+), 23 deletions(-)

diff --git a/arch/ia64/include/asm/switch_to.h b/arch/ia64/include/asm/switch_to.h
index cb2412f..d38c7ea 100644
--- a/arch/ia64/include/asm/switch_to.h
+++ b/arch/ia64/include/asm/switch_to.h
@@ -30,13 +30,6 @@ extern struct task_struct *ia64_switch_to (void *next_task);
 extern void ia64_save_extra (struct task_struct *task);
 extern void ia64_load_extra (struct task_struct *task);
 
-#ifdef CONFIG_VIRT_CPU_ACCOUNTING
-extern void ia64_account_on_switch (struct task_struct *prev, struct task_struct *next);
-# define IA64_ACCOUNT_ON_SWITCH(p,n) ia64_account_on_switch(p,n)
-#else
-# define IA64_ACCOUNT_ON_SWITCH(p,n)
-#endif
-
 #ifdef CONFIG_PERFMON
   DECLARE_PER_CPU(unsigned long, pfm_syst_info);
 # define PERFMON_IS_SYSWIDE() (__get_cpu_var(pfm_syst_info) & 0x1)
@@ -49,7 +42,6 @@ extern void ia64_account_on_switch (struct task_struct *prev, struct task_struct
 	 || PERFMON_IS_SYSWIDE())
 
 #define __switch_to(prev,next,last) do {							 \
-	IA64_ACCOUNT_ON_SWITCH(prev, next);							 \
 	if (IA64_HAS_EXTRA_STATE(prev))								 \
 		ia64_save_extra(prev);								 \
 	if (IA64_HAS_EXTRA_STATE(next))								 \
diff --git a/arch/ia64/kernel/time.c b/arch/ia64/kernel/time.c
index ecc904b..6247197 100644
--- a/arch/ia64/kernel/time.c
+++ b/arch/ia64/kernel/time.c
@@ -88,10 +88,10 @@ extern cputime_t cycle_to_cputime(u64 cyc);
  * accumulated times to the current process, and to prepare accounting on
  * the next process.
  */
-void ia64_account_on_switch(struct task_struct *prev, struct task_struct *next)
+void account_switch_vtime(struct task_struct *prev)
 {
 	struct thread_info *pi = task_thread_info(prev);
-	struct thread_info *ni = task_thread_info(next);
+	struct thread_info *ni = task_thread_info(current);
 	cputime_t delta_stime, delta_utime;
 	__u64 now;
 
diff --git a/arch/powerpc/include/asm/time.h b/arch/powerpc/include/asm/time.h
index 3b4b4a8..c1f2676 100644
--- a/arch/powerpc/include/asm/time.h
+++ b/arch/powerpc/include/asm/time.h
@@ -197,12 +197,6 @@ struct cpu_usage {
 
 DECLARE_PER_CPU(struct cpu_usage, cpu_usage_array);
 
-#if defined(CONFIG_VIRT_CPU_ACCOUNTING)
-#define account_process_vtime(tsk)		account_process_tick(tsk, 0)
-#else
-#define account_process_vtime(tsk)		do { } while (0)
-#endif
-
 extern void secondary_cpu_time_init(void);
 
 DECLARE_PER_CPU(u64, decrementers_next_tb);
diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c
index 710f400..d73fa99 100644
--- a/arch/powerpc/kernel/process.c
+++ b/arch/powerpc/kernel/process.c
@@ -514,9 +514,6 @@ struct task_struct *__switch_to(struct task_struct *prev,
 
 	local_irq_save(flags);
 
-	account_system_vtime(current);
-	account_process_vtime(current);
-
 	/*
 	 * We can't take a PMU exception inside _switch() since there is a
 	 * window where the kernel stack SLB and the kernel stack are out
diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c
index be171ee..49da7f0 100644
--- a/arch/powerpc/kernel/time.c
+++ b/arch/powerpc/kernel/time.c
@@ -366,6 +366,12 @@ void account_process_tick(struct task_struct *tsk, int user_tick)
 	account_user_time(tsk, utime, utimescaled);
 }
 
+void account_switch_vtime(struct task_struct *prev)
+{
+	account_system_vtime(prev);
+	account_process_tick(prev, 0);
+}
+
 #else /* ! CONFIG_VIRT_CPU_ACCOUNTING */
 #define calc_cputime_factors()
 #endif
diff --git a/arch/s390/include/asm/switch_to.h b/arch/s390/include/asm/switch_to.h
index f223068..e7f9b3d 100644
--- a/arch/s390/include/asm/switch_to.h
+++ b/arch/s390/include/asm/switch_to.h
@@ -89,12 +89,10 @@ static inline void restore_access_regs(unsigned int *acrs)
 	prev = __switch_to(prev,next);					\
 } while (0)
 
-extern void account_vtime(struct task_struct *, struct task_struct *);
 extern void account_tick_vtime(struct task_struct *);
 
 #define finish_arch_switch(prev) do {					     \
 	set_fs(current->thread.mm_segment);				     \
-	account_vtime(prev, current);					     \
 } while (0)
 
 #endif /* __ASM_SWITCH_TO_H */
diff --git a/arch/s390/kernel/vtime.c b/arch/s390/kernel/vtime.c
index 4fc97b4..449ac22 100644
--- a/arch/s390/kernel/vtime.c
+++ b/arch/s390/kernel/vtime.c
@@ -99,7 +99,7 @@ static int do_account_vtime(struct task_struct *tsk, int hardirq_offset)
 	return virt_timer_forward(user + system);
 }
 
-void account_vtime(struct task_struct *prev, struct task_struct *next)
+void account_switch_vtime(struct task_struct *prev)
 {
 	struct thread_info *ti;
 
@@ -107,7 +107,7 @@ void account_vtime(struct task_struct *prev, struct task_struct *next)
 	ti = task_thread_info(prev);
 	ti->user_timer = S390_lowcore.user_timer;
 	ti->system_timer = S390_lowcore.system_timer;
-	ti = task_thread_info(next);
+	ti = task_thread_info(current);
 	S390_lowcore.user_timer = ti->user_timer;
 	S390_lowcore.system_timer = ti->system_timer;
 }
diff --git a/include/linux/kernel_stat.h b/include/linux/kernel_stat.h
index 2fbd905..bbe5d15 100644
--- a/include/linux/kernel_stat.h
+++ b/include/linux/kernel_stat.h
@@ -130,4 +130,10 @@ extern void account_process_tick(struct task_struct *, int user);
 extern void account_steal_ticks(unsigned long ticks);
 extern void account_idle_ticks(unsigned long ticks);
 
+#ifdef CONFIG_VIRT_CPU_ACCOUNTING
+extern void account_switch_vtime(struct task_struct *prev);
+#else
+static inline void account_switch_vtime(struct task_struct *prev) { }
+#endif
+
 #endif /* _LINUX_KERNEL_STAT_H */
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index ae3bcaa..78d9c96 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -1796,6 +1796,7 @@ static void finish_task_switch(struct rq *rq, struct task_struct *prev)
 	 *		Manfred Spraul <manfred@colorfullife.com>
 	 */
 	prev_state = prev->state;
+	account_switch_vtime(prev);
 	finish_arch_switch(prev);
 #ifdef __ARCH_WANT_INTERRUPTS_ON_CTXSW
 	local_irq_disable();
-- 
1.7.5.4

[PATCH 4/4] s390: Remove leftover account_tick_vtime() header

[Frederic Weisbecker]

The function doesn't seem to exist anymore.

Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
---
 arch/s390/include/asm/switch_to.h |    2 --
 1 files changed, 0 insertions(+), 2 deletions(-)

diff --git a/arch/s390/include/asm/switch_to.h b/arch/s390/include/asm/switch_to.h
index e7f9b3d..314cc94 100644
--- a/arch/s390/include/asm/switch_to.h
+++ b/arch/s390/include/asm/switch_to.h
@@ -89,8 +89,6 @@ static inline void restore_access_regs(unsigned int *acrs)
 	prev = __switch_to(prev,next);					\
 } while (0)
 
-extern void account_tick_vtime(struct task_struct *);
-
 #define finish_arch_switch(prev) do {					     \
 	set_fs(current->thread.mm_segment);				     \
 } while (0)
-- 
1.7.5.4

Re: [PATCH 0/4] cputime: Virtual cputime accounting small cleanups and consolidation v2

[Frederic Weisbecker]

On Tue, Aug 14, 2012 at 04:16:46PM +0200, Frederic Weisbecker wrote:
> Hi,
> 
> No fundamental change in this release but a rebase to solve conflicts
> against latest tip:/sched/core commits.
> 
> Thanks.

This can be pulled from:

git://github.com/fweisbec/linux-dynticks.git
	virt-cputime-v2

This patchset, besides beeing a desired consolidation and
cleanup IMO, is necessary for the adaptive nohz feature
(see: http://comments.gmane.org/gmane.linux.kernel/1337690)

Thanks.

> 
> Frederic Weisbecker (4):
>   cputime: Generalize CONFIG_VIRT_CPU_ACCOUNTING
>   sched: Move cputime code to its own file
>   cputime: Consolidate vtime handling on context switch
>   s390: Remove leftover account_tick_vtime() header
> 
>  arch/Kconfig                           |    3 +
>  arch/ia64/Kconfig                      |   12 +-
>  arch/ia64/include/asm/switch_to.h      |    8 -
>  arch/ia64/kernel/time.c                |    4 +-
>  arch/powerpc/include/asm/time.h        |    6 -
>  arch/powerpc/kernel/process.c          |    3 -
>  arch/powerpc/kernel/time.c             |    6 +
>  arch/powerpc/platforms/Kconfig.cputype |   16 +-
>  arch/s390/Kconfig                      |    5 +-
>  arch/s390/include/asm/switch_to.h      |    4 -
>  arch/s390/kernel/vtime.c               |    4 +-
>  include/linux/kernel_stat.h            |    6 +
>  init/Kconfig                           |   13 +
>  kernel/sched/Makefile                  |    2 +-
>  kernel/sched/core.c                    |  558 +-------------------------------
>  kernel/sched/cputime.c                 |  503 ++++++++++++++++++++++++++++
>  kernel/sched/sched.h                   |   63 ++++
>  17 files changed, 606 insertions(+), 610 deletions(-)
>  create mode 100644 kernel/sched/cputime.c
> 
> -- 
> 1.7.5.4
> 

Re: [PATCH 1/4] cputime: Generalize CONFIG_VIRT_CPU_ACCOUNTING

[Martin Schwidefsky]

On Tue, 14 Aug 2012 16:16:47 +0200
Frederic Weisbecker <fweisbec@gmail.com> wrote:

> S390, ia64 and powerpc all define their own version
> of CONFIG_VIRT_CPU_ACCOUNTING. Generalize the config
> and its description to a single place to avoid
> duplication.

For S390 CONFIG_VIRT_CPU_ACCOUNTING is not configurable, it is always
enabled. With this patch we'd get a config option in the menu, no?

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

Re: [PATCH 2/4] sched: Move cputime code to its own file

[Martin Schwidefsky]

On Tue, 14 Aug 2012 16:16:48 +0200
Frederic Weisbecker <fweisbec@gmail.com> wrote:

> Extract cputime code from the giant sched/core.c and
> put it in its own file. This make it easier to deal with
> this particular area and de-bloat a bit more core.c

To move the cputime accouting code to its own file makes sense.
Acked-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

Re: [PATCH 3/4] cputime: Consolidate vtime handling on context switch

[Martin Schwidefsky]

On Tue, 14 Aug 2012 16:16:49 +0200
Frederic Weisbecker <fweisbec@gmail.com> wrote:

> The archs that implement virtual cputime accounting all
> flush the cputime of a task when it gets descheduled
> and sometimes set up some ground initialization for the
> next task to account its cputime.
> 
> These archs all put their own hooks in their context
> switch callbacks and handle the off-case themselves.
> 
> Consolidate this by creating a new account_switch_vtime()
> callback called in generic code right after a context switch
> and that these archs must implement to flush the prev task
> cputime and initialize the next task cputime related state.

That change requires that the accounting for the previous process
can be done before finish_arch_switch() completed. With the old
code the architecture could to the accounting call in the middle
of finish_arch_switch, that is not possible anymore. Dunno if this
is relevant or not. For s390 the new code should work fine.

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

Re: [PATCH 4/4] s390: Remove leftover account_tick_vtime() header

[Martin Schwidefsky]

On Tue, 14 Aug 2012 16:16:50 +0200
Frederic Weisbecker <fweisbec@gmail.com> wrote:

> The function doesn't seem to exist anymore.
> 
> Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
> Cc: Tony Luck <tony.luck@intel.com>
> Cc: Fenghua Yu <fenghua.yu@intel.com>
> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
> Cc: Paul Mackerras <paulus@samba.org>
> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
> Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
> Cc: Ingo Molnar <mingo@kernel.org>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> Cc: Peter Zijlstra <peterz@infradead.org>
> ---
>  arch/s390/include/asm/switch_to.h |    2 --
>  1 files changed, 0 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/s390/include/asm/switch_to.h b/arch/s390/include/asm/switch_to.h
> index e7f9b3d..314cc94 100644
> --- a/arch/s390/include/asm/switch_to.h
> +++ b/arch/s390/include/asm/switch_to.h
> @@ -89,8 +89,6 @@ static inline void restore_access_regs(unsigned int *acrs)
>  	prev = __switch_to(prev,next);					\
>  } while (0)
> 
> -extern void account_tick_vtime(struct task_struct *);
> -
>  #define finish_arch_switch(prev) do {					     \
>  	set_fs(current->thread.mm_segment);				     \
>  } while (0)

Indeed..

Acked-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

Re: [PATCH 1/4] cputime: Generalize CONFIG_VIRT_CPU_ACCOUNTING

[Frederic Weisbecker]

On Wed, Aug 15, 2012 at 05:03:47PM +0200, Martin Schwidefsky wrote:
> On Tue, 14 Aug 2012 16:16:47 +0200
> Frederic Weisbecker <fweisbec@gmail.com> wrote:
> 
> > S390, ia64 and powerpc all define their own version
> > of CONFIG_VIRT_CPU_ACCOUNTING. Generalize the config
> > and its description to a single place to avoid
> > duplication.
> 
> For S390 CONFIG_VIRT_CPU_ACCOUNTING is not configurable, it is always
> enabled. With this patch we'd get a config option in the menu, no?

Indeed it now appears in the menu but in the case of s390, it's impossible
to turn it off due to:

	config S390
		select VIRT_CPU_ACCOUNTING

This creates a strict dependency that the user can't override. The option
is untoggable.

Re: [PATCH 3/4] cputime: Consolidate vtime handling on context switch

[Frederic Weisbecker]

On Wed, Aug 15, 2012 at 05:22:19PM +0200, Martin Schwidefsky wrote:
> On Tue, 14 Aug 2012 16:16:49 +0200
> Frederic Weisbecker <fweisbec@gmail.com> wrote:
> 
> > The archs that implement virtual cputime accounting all
> > flush the cputime of a task when it gets descheduled
> > and sometimes set up some ground initialization for the
> > next task to account its cputime.
> > 
> > These archs all put their own hooks in their context
> > switch callbacks and handle the off-case themselves.
> > 
> > Consolidate this by creating a new account_switch_vtime()
> > callback called in generic code right after a context switch
> > and that these archs must implement to flush the prev task
> > cputime and initialize the next task cputime related state.
> 
> That change requires that the accounting for the previous process
> can be done before finish_arch_switch() completed. With the old
> code the architecture could to the accounting call in the middle
> of finish_arch_switch, that is not possible anymore. Dunno if this
> is relevant or not. For s390 the new code should work fine.

I'm not sure how this could potentially cause a problem. Interrupts are disabled
between while we switch_to() until finish_lock_switch(). So nothing
should be able to mess up with the accounting of the prev task.

I don't really understand what you mean actually.

Thanks.

Re: [PATCH 3/4] cputime: Consolidate vtime handling on context switch

[Martin Schwidefsky]

On Wed, 15 Aug 2012 21:28:17 +0200
Frederic Weisbecker <fweisbec@gmail.com> wrote:

> On Wed, Aug 15, 2012 at 05:22:19PM +0200, Martin Schwidefsky wrote:
> > On Tue, 14 Aug 2012 16:16:49 +0200
> > Frederic Weisbecker <fweisbec@gmail.com> wrote:
> > 
> > > The archs that implement virtual cputime accounting all
> > > flush the cputime of a task when it gets descheduled
> > > and sometimes set up some ground initialization for the
> > > next task to account its cputime.
> > > 
> > > These archs all put their own hooks in their context
> > > switch callbacks and handle the off-case themselves.
> > > 
> > > Consolidate this by creating a new account_switch_vtime()
> > > callback called in generic code right after a context switch
> > > and that these archs must implement to flush the prev task
> > > cputime and initialize the next task cputime related state.
> > 
> > That change requires that the accounting for the previous process
> > can be done before finish_arch_switch() completed. With the old
> > code the architecture could to the accounting call in the middle
> > of finish_arch_switch, that is not possible anymore. Dunno if this
> > is relevant or not. For s390 the new code should work fine.
> 
> I'm not sure how this could potentially cause a problem. Interrupts are disabled
> between while we switch_to() until finish_lock_switch(). So nothing
> should be able to mess up with the accounting of the prev task.
> 
> I don't really understand what you mean actually.

It is more a theoretical consideration. If the finish_arch_switch code
updates fields that are required to do the cputime accounting then the
order could be important. But then you could move that necessary code
from finish_arch_switch to account_switch_vtime.
As said that change is fine for s390, so I'm good with it.

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

Re: [PATCH 1/4] cputime: Generalize CONFIG_VIRT_CPU_ACCOUNTING

[Martin Schwidefsky]

On Wed, 15 Aug 2012 21:09:04 +0200
Frederic Weisbecker <fweisbec@gmail.com> wrote:

> On Wed, Aug 15, 2012 at 05:03:47PM +0200, Martin Schwidefsky wrote:
> > On Tue, 14 Aug 2012 16:16:47 +0200
> > Frederic Weisbecker <fweisbec@gmail.com> wrote:
> > 
> > > S390, ia64 and powerpc all define their own version
> > > of CONFIG_VIRT_CPU_ACCOUNTING. Generalize the config
> > > and its description to a single place to avoid
> > > duplication.
> > 
> > For S390 CONFIG_VIRT_CPU_ACCOUNTING is not configurable, it is always
> > enabled. With this patch we'd get a config option in the menu, no?
> 
> Indeed it now appears in the menu but in the case of s390, it's impossible
> to turn it off due to:
> 
> 	config S390
> 		select VIRT_CPU_ACCOUNTING
> 
> This creates a strict dependency that the user can't override. The option
> is untoggable.

Hmm, ok. But then the description should be reworded not to be specific to
the power architecture (the part of the message about "This also enables
accounting of stolen time on logically-partitioned systems running on IBM
POWER5-based machines.").

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

Re: [PATCH 1/4] cputime: Generalize CONFIG_VIRT_CPU_ACCOUNTING

[Benjamin Herrenschmidt]

On Thu, 2012-08-16 at 09:53 +0200, Martin Schwidefsky wrote:
> Hmm, ok. But then the description should be reworded not to be specific to
> the power architecture (the part of the message about "This also enables
> accounting of stolen time on logically-partitioned systems running on IBM
> POWER5-based machines."). 

Which is not very helpful to somebody running on a POWER6 or 7 (which
also support that option just fine :-)

So yes, the description should definitely be improved.

Cheers,
Ben.

Re: [PATCH 3/4] cputime: Consolidate vtime handling on context switch

[Frederic Weisbecker]

On Thu, Aug 16, 2012 at 09:50:32AM +0200, Martin Schwidefsky wrote:
> On Wed, 15 Aug 2012 21:28:17 +0200
> Frederic Weisbecker <fweisbec@gmail.com> wrote:
> 
> > On Wed, Aug 15, 2012 at 05:22:19PM +0200, Martin Schwidefsky wrote:
> > > On Tue, 14 Aug 2012 16:16:49 +0200
> > > Frederic Weisbecker <fweisbec@gmail.com> wrote:
> > > 
> > > > The archs that implement virtual cputime accounting all
> > > > flush the cputime of a task when it gets descheduled
> > > > and sometimes set up some ground initialization for the
> > > > next task to account its cputime.
> > > > 
> > > > These archs all put their own hooks in their context
> > > > switch callbacks and handle the off-case themselves.
> > > > 
> > > > Consolidate this by creating a new account_switch_vtime()
> > > > callback called in generic code right after a context switch
> > > > and that these archs must implement to flush the prev task
> > > > cputime and initialize the next task cputime related state.
> > > 
> > > That change requires that the accounting for the previous process
> > > can be done before finish_arch_switch() completed. With the old
> > > code the architecture could to the accounting call in the middle
> > > of finish_arch_switch, that is not possible anymore. Dunno if this
> > > is relevant or not. For s390 the new code should work fine.
> > 
> > I'm not sure how this could potentially cause a problem. Interrupts are disabled
> > between while we switch_to() until finish_lock_switch(). So nothing
> > should be able to mess up with the accounting of the prev task.
> > 
> > I don't really understand what you mean actually.
> 
> It is more a theoretical consideration. If the finish_arch_switch code
> updates fields that are required to do the cputime accounting then the
> order could be important. But then you could move that necessary code
> from finish_arch_switch to account_switch_vtime.
> As said that change is fine for s390, so I'm good with it.

Ah ok. Well like you said this is fine for s390. And it looks also fine
to me on ia64 and powerpc as it doesn't look like we depend on something
done in finish_arch_switch() there. They were flush the previous task
cputime from switch_to() anyway.

Thanks.

PS: can I add your ack?

> 
> -- 
> blue skies,
>    Martin.
> 
> "Reality continues to ruin my life." - Calvin.
> 

Re: [PATCH 1/4] cputime: Generalize CONFIG_VIRT_CPU_ACCOUNTING

[Frederic Weisbecker]

On Thu, Aug 16, 2012 at 07:38:17PM +1000, Benjamin Herrenschmidt wrote:
> On Thu, 2012-08-16 at 09:53 +0200, Martin Schwidefsky wrote:
> > Hmm, ok. But then the description should be reworded not to be specific to
> > the power architecture (the part of the message about "This also enables
> > accounting of stolen time on logically-partitioned systems running on IBM
> > POWER5-based machines."). 
> 
> Which is not very helpful to somebody running on a POWER6 or 7 (which
> also support that option just fine :-)
> 
> So yes, the description should definitely be improved.

All right. How about something like the below?

diff --git a/init/Kconfig b/init/Kconfig
index 894b073..5f5f8c2 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -276,9 +276,9 @@ config VIRT_CPU_ACCOUNTING
 	  accounting.  This is done by reading a CPU counter on each
 	  kernel entry and exit and on transitions within the kernel
 	  between system, softirq and hardirq state, so there is a
-	  small performance impact.  This also enables accounting of
-	  stolen time on logically-partitioned systems running on
-	  IBM POWER5-based machines.
+	  small performance impact.  In the case of IBM POWER > 5, this
+	  also enables accounting of stolen time on logically-partitioned
+	  systems.
 
 config BSD_PROCESS_ACCT
 	bool "BSD Process Accounting"

Re: [PATCH 3/4] cputime: Consolidate vtime handling on context switch

[Martin Schwidefsky]

On Thu, 16 Aug 2012 14:50:33 +0200
Frederic Weisbecker <fweisbec@gmail.com> wrote:

> On Thu, Aug 16, 2012 at 09:50:32AM +0200, Martin Schwidefsky wrote:
> > On Wed, 15 Aug 2012 21:28:17 +0200
> > Frederic Weisbecker <fweisbec@gmail.com> wrote:
> > 
> > > On Wed, Aug 15, 2012 at 05:22:19PM +0200, Martin Schwidefsky wrote:
> > > > On Tue, 14 Aug 2012 16:16:49 +0200
> > > > Frederic Weisbecker <fweisbec@gmail.com> wrote:
> > > > 
> > > > > The archs that implement virtual cputime accounting all
> > > > > flush the cputime of a task when it gets descheduled
> > > > > and sometimes set up some ground initialization for the
> > > > > next task to account its cputime.
> > > > > 
> > > > > These archs all put their own hooks in their context
> > > > > switch callbacks and handle the off-case themselves.
> > > > > 
> > > > > Consolidate this by creating a new account_switch_vtime()
> > > > > callback called in generic code right after a context switch
> > > > > and that these archs must implement to flush the prev task
> > > > > cputime and initialize the next task cputime related state.
> > > > 
> > > > That change requires that the accounting for the previous process
> > > > can be done before finish_arch_switch() completed. With the old
> > > > code the architecture could to the accounting call in the middle
> > > > of finish_arch_switch, that is not possible anymore. Dunno if this
> > > > is relevant or not. For s390 the new code should work fine.
> > > 
> > > I'm not sure how this could potentially cause a problem. Interrupts are disabled
> > > between while we switch_to() until finish_lock_switch(). So nothing
> > > should be able to mess up with the accounting of the prev task.
> > > 
> > > I don't really understand what you mean actually.
> > 
> > It is more a theoretical consideration. If the finish_arch_switch code
> > updates fields that are required to do the cputime accounting then the
> > order could be important. But then you could move that necessary code
> > from finish_arch_switch to account_switch_vtime.
> > As said that change is fine for s390, so I'm good with it.
> 
> Ah ok. Well like you said this is fine for s390. And it looks also fine
> to me on ia64 and powerpc as it doesn't look like we depend on something
> done in finish_arch_switch() there. They were flush the previous task
> cputime from switch_to() anyway.
> 
> Thanks.
> 
> PS: can I add your ack?

Sure, feel free to add my Acked-by.

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

Re: [PATCH 1/4] cputime: Generalize CONFIG_VIRT_CPU_ACCOUNTING

[Martin Schwidefsky]

On Thu, 16 Aug 2012 14:55:59 +0200
Frederic Weisbecker <fweisbec@gmail.com> wrote:

> On Thu, Aug 16, 2012 at 07:38:17PM +1000, Benjamin Herrenschmidt wrote:
> > On Thu, 2012-08-16 at 09:53 +0200, Martin Schwidefsky wrote:
> > > Hmm, ok. But then the description should be reworded not to be specific to
> > > the power architecture (the part of the message about "This also enables
> > > accounting of stolen time on logically-partitioned systems running on IBM
> > > POWER5-based machines."). 
> > 
> > Which is not very helpful to somebody running on a POWER6 or 7 (which
> > also support that option just fine :-)
> > 
> > So yes, the description should definitely be improved.
> 
> All right. How about something like the below?
> 
> diff --git a/init/Kconfig b/init/Kconfig
> index 894b073..5f5f8c2 100644
> --- a/init/Kconfig
> +++ b/init/Kconfig
> @@ -276,9 +276,9 @@ config VIRT_CPU_ACCOUNTING
>  	  accounting.  This is done by reading a CPU counter on each
>  	  kernel entry and exit and on transitions within the kernel
>  	  between system, softirq and hardirq state, so there is a
> -	  small performance impact.  This also enables accounting of
> -	  stolen time on logically-partitioned systems running on
> -	  IBM POWER5-based machines.
> +	  small performance impact.  In the case of IBM POWER > 5, this
> +	  also enables accounting of stolen time on logically-partitioned
> +	  systems.
> 
>  config BSD_PROCESS_ACCT
>  	bool "BSD Process Accounting"
> 

VIRT_CPU_ACCOUNTING will enable steal time for s390 as well.

-- 
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.

Re: [PATCH 1/4] cputime: Generalize CONFIG_VIRT_CPU_ACCOUNTING

[Frederic Weisbecker]

On Thu, Aug 16, 2012 at 04:00:44PM +0200, Martin Schwidefsky wrote:
> On Thu, 16 Aug 2012 14:55:59 +0200
> Frederic Weisbecker <fweisbec@gmail.com> wrote:
> 
> > On Thu, Aug 16, 2012 at 07:38:17PM +1000, Benjamin Herrenschmidt wrote:
> > > On Thu, 2012-08-16 at 09:53 +0200, Martin Schwidefsky wrote:
> > > > Hmm, ok. But then the description should be reworded not to be specific to
> > > > the power architecture (the part of the message about "This also enables
> > > > accounting of stolen time on logically-partitioned systems running on IBM
> > > > POWER5-based machines."). 
> > > 
> > > Which is not very helpful to somebody running on a POWER6 or 7 (which
> > > also support that option just fine :-)
> > > 
> > > So yes, the description should definitely be improved.
> > 
> > All right. How about something like the below?
> > 
> > diff --git a/init/Kconfig b/init/Kconfig
> > index 894b073..5f5f8c2 100644
> > --- a/init/Kconfig
> > +++ b/init/Kconfig
> > @@ -276,9 +276,9 @@ config VIRT_CPU_ACCOUNTING
> >  	  accounting.  This is done by reading a CPU counter on each
> >  	  kernel entry and exit and on transitions within the kernel
> >  	  between system, softirq and hardirq state, so there is a
> > -	  small performance impact.  This also enables accounting of
> > -	  stolen time on logically-partitioned systems running on
> > -	  IBM POWER5-based machines.
> > +	  small performance impact.  In the case of IBM POWER > 5, this
> > +	  also enables accounting of stolen time on logically-partitioned
> > +	  systems.
> > 
> >  config BSD_PROCESS_ACCT
> >  	bool "BSD Process Accounting"
> > 
> 
> VIRT_CPU_ACCOUNTING will enable steal time for s390 as well.

Ah right. Fixed below:

diff --git a/init/Kconfig b/init/Kconfig
index 894b073..c40d0fb 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -276,9 +276,9 @@ config VIRT_CPU_ACCOUNTING
 	  accounting.  This is done by reading a CPU counter on each
 	  kernel entry and exit and on transitions within the kernel
 	  between system, softirq and hardirq state, so there is a
-	  small performance impact.  This also enables accounting of
-	  stolen time on logically-partitioned systems running on
-	  IBM POWER5-based machines.
+	  small performance impact.  In the case of s390 or IBM POWER > 5,
+	  this also enables accounting of stolen time on logically-partitioned
+	  systems.
 
 config BSD_PROCESS_ACCT
 	bool "BSD Process Accounting"

[PATCH 1/4] cputime: Generalize CONFIG_VIRT_CPU_ACCOUNTING

[Frederic Weisbecker]

S390, ia64 and powerpc all define their own version
of CONFIG_VIRT_CPU_ACCOUNTING. Generalize the config
and its description to a single place to avoid
duplication.

Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
---
 arch/Kconfig                           |    3 +++
 arch/ia64/Kconfig                      |   12 +-----------
 arch/powerpc/platforms/Kconfig.cputype |   16 +---------------
 arch/s390/Kconfig                      |    5 ++---
 init/Kconfig                           |   13 +++++++++++++
 5 files changed, 20 insertions(+), 29 deletions(-)

diff --git a/arch/Kconfig b/arch/Kconfig
index 72f2fa1..f78de57 100644
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -281,4 +281,7 @@ config SECCOMP_FILTER
 
 	  See Documentation/prctl/seccomp_filter.txt for details.
 
+config HAVE_VIRT_CPU_ACCOUNTING
+	bool
+
 source "kernel/gcov/Kconfig"
diff --git a/arch/ia64/Kconfig b/arch/ia64/Kconfig
index 310cf57..3c720ef 100644
--- a/arch/ia64/Kconfig
+++ b/arch/ia64/Kconfig
@@ -25,6 +25,7 @@ config IA64
 	select HAVE_GENERIC_HARDIRQS
 	select HAVE_MEMBLOCK
 	select HAVE_MEMBLOCK_NODE_MAP
+	select HAVE_VIRT_CPU_ACCOUNTING
 	select ARCH_DISCARD_MEMBLOCK
 	select GENERIC_IRQ_PROBE
 	select GENERIC_PENDING_IRQ if SMP
@@ -340,17 +341,6 @@ config FORCE_MAX_ZONEORDER
 	default "17" if HUGETLB_PAGE
 	default "11"
 
-config VIRT_CPU_ACCOUNTING
-	bool "Deterministic task and CPU time accounting"
-	default n
-	help
-	  Select this option to enable more accurate task and CPU time
-	  accounting.  This is done by reading a CPU counter on each
-	  kernel entry and exit and on transitions within the kernel
-	  between system, softirq and hardirq state, so there is a
-	  small performance impact.
-	  If in doubt, say N here.
-
 config SMP
 	bool "Symmetric multi-processing support"
 	select USE_GENERIC_SMP_HELPERS
diff --git a/arch/powerpc/platforms/Kconfig.cputype b/arch/powerpc/platforms/Kconfig.cputype
index 30fd01d..72afd28 100644
--- a/arch/powerpc/platforms/Kconfig.cputype
+++ b/arch/powerpc/platforms/Kconfig.cputype
@@ -1,6 +1,7 @@
 config PPC64
 	bool "64-bit kernel"
 	default n
+	select HAVE_VIRT_CPU_ACCOUNTING
 	help
 	  This option selects whether a 32-bit or a 64-bit kernel
 	  will be built.
@@ -337,21 +338,6 @@ config PPC_MM_SLICES
 	default y if (!PPC_FSL_BOOK3E && PPC64 && HUGETLB_PAGE) || (PPC_STD_MMU_64 && PPC_64K_PAGES)
 	default n
 
-config VIRT_CPU_ACCOUNTING
-	bool "Deterministic task and CPU time accounting"
-	depends on PPC64
-	default y
-	help
-	  Select this option to enable more accurate task and CPU time
-	  accounting.  This is done by reading a CPU counter on each
-	  kernel entry and exit and on transitions within the kernel
-	  between system, softirq and hardirq state, so there is a
-	  small performance impact.  This also enables accounting of
-	  stolen time on logically-partitioned systems running on
-	  IBM POWER5-based machines.
-
-	  If in doubt, say Y here.
-
 config PPC_HAVE_PMU_SUPPORT
        bool
 
diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig
index 76de6b6..49ebfb6 100644
--- a/arch/s390/Kconfig
+++ b/arch/s390/Kconfig
@@ -49,9 +49,6 @@ config GENERIC_LOCKBREAK
 config PGSTE
 	def_bool y if KVM
 
-config VIRT_CPU_ACCOUNTING
-	def_bool y
-
 config ARCH_SUPPORTS_DEBUG_PAGEALLOC
 	def_bool y
 
@@ -89,6 +86,8 @@ config S390
 	select HAVE_MEMBLOCK
 	select HAVE_MEMBLOCK_NODE_MAP
 	select HAVE_CMPXCHG_LOCAL
+	select HAVE_VIRT_CPU_ACCOUNTING
+	select VIRT_CPU_ACCOUNTING
 	select ARCH_DISCARD_MEMBLOCK
 	select BUILDTIME_EXTABLE_SORT
 	select ARCH_INLINE_SPIN_TRYLOCK
diff --git a/init/Kconfig b/init/Kconfig
index af6c7f8..c40d0fb 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -267,6 +267,19 @@ config POSIX_MQUEUE_SYSCTL
 	depends on SYSCTL
 	default y
 
+config VIRT_CPU_ACCOUNTING
+	bool "Deterministic task and CPU time accounting"
+	depends on HAVE_VIRT_CPU_ACCOUNTING
+	default y if PPC64
+	help
+	  Select this option to enable more accurate task and CPU time
+	  accounting.  This is done by reading a CPU counter on each
+	  kernel entry and exit and on transitions within the kernel
+	  between system, softirq and hardirq state, so there is a
+	  small performance impact.  In the case of s390 or IBM POWER > 5,
+	  this also enables accounting of stolen time on logically-partitioned
+	  systems.
+
 config BSD_PROCESS_ACCT
 	bool "BSD Process Accounting"
 	help
-- 
1.7.5.4

[PATCH 1/4] cputime: Generalize CONFIG_VIRT_CPU_ACCOUNTING

[Frederic Weisbecker]

S390, ia64 and powerpc all define their own version
of CONFIG_VIRT_CPU_ACCOUNTING. Generalize the config
and its description to a single place to avoid
duplication.

Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
---
 arch/Kconfig                           |    3 +++
 arch/ia64/Kconfig                      |   12 +-----------
 arch/powerpc/platforms/Kconfig.cputype |   16 +---------------
 arch/s390/Kconfig                      |    5 ++---
 init/Kconfig                           |   13 +++++++++++++
 5 files changed, 20 insertions(+), 29 deletions(-)

diff --git a/arch/Kconfig b/arch/Kconfig
index 8c3d957..ff712af 100644
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -274,4 +274,7 @@ config SECCOMP_FILTER
 
 	  See Documentation/prctl/seccomp_filter.txt for details.
 
+config HAVE_VIRT_CPU_ACCOUNTING
+	bool
+
 source "kernel/gcov/Kconfig"
diff --git a/arch/ia64/Kconfig b/arch/ia64/Kconfig
index 8186ec5..509af14 100644
--- a/arch/ia64/Kconfig
+++ b/arch/ia64/Kconfig
@@ -25,6 +25,7 @@ config IA64
 	select HAVE_GENERIC_HARDIRQS
 	select HAVE_MEMBLOCK
 	select HAVE_MEMBLOCK_NODE_MAP
+	select HAVE_VIRT_CPU_ACCOUNTING
 	select ARCH_DISCARD_MEMBLOCK
 	select GENERIC_IRQ_PROBE
 	select GENERIC_PENDING_IRQ if SMP
@@ -341,17 +342,6 @@ config FORCE_MAX_ZONEORDER
 	default "17" if HUGETLB_PAGE
 	default "11"
 
-config VIRT_CPU_ACCOUNTING
-	bool "Deterministic task and CPU time accounting"
-	default n
-	help
-	  Select this option to enable more accurate task and CPU time
-	  accounting.  This is done by reading a CPU counter on each
-	  kernel entry and exit and on transitions within the kernel
-	  between system, softirq and hardirq state, so there is a
-	  small performance impact.
-	  If in doubt, say N here.
-
 config SMP
 	bool "Symmetric multi-processing support"
 	select USE_GENERIC_SMP_HELPERS
diff --git a/arch/powerpc/platforms/Kconfig.cputype b/arch/powerpc/platforms/Kconfig.cputype
index 61c9550..2a896de 100644
--- a/arch/powerpc/platforms/Kconfig.cputype
+++ b/arch/powerpc/platforms/Kconfig.cputype
@@ -1,6 +1,7 @@
 config PPC64
 	bool "64-bit kernel"
 	default n
+	select HAVE_VIRT_CPU_ACCOUNTING
 	help
 	  This option selects whether a 32-bit or a 64-bit kernel
 	  will be built.
@@ -333,21 +334,6 @@ config PPC_MM_SLICES
 	default y if (!PPC_FSL_BOOK3E && PPC64 && HUGETLB_PAGE) || (PPC_STD_MMU_64 && PPC_64K_PAGES)
 	default n
 
-config VIRT_CPU_ACCOUNTING
-	bool "Deterministic task and CPU time accounting"
-	depends on PPC64
-	default y
-	help
-	  Select this option to enable more accurate task and CPU time
-	  accounting.  This is done by reading a CPU counter on each
-	  kernel entry and exit and on transitions within the kernel
-	  between system, softirq and hardirq state, so there is a
-	  small performance impact.  This also enables accounting of
-	  stolen time on logically-partitioned systems running on
-	  IBM POWER5-based machines.
-
-	  If in doubt, say Y here.
-
 config PPC_HAVE_PMU_SUPPORT
        bool
 
diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig
index a39b469..351aa40 100644
--- a/arch/s390/Kconfig
+++ b/arch/s390/Kconfig
@@ -49,9 +49,6 @@ config GENERIC_LOCKBREAK
 config PGSTE
 	def_bool y if KVM
 
-config VIRT_CPU_ACCOUNTING
-	def_bool y
-
 config ARCH_SUPPORTS_DEBUG_PAGEALLOC
 	def_bool y
 
@@ -88,6 +85,8 @@ config S390
 	select HAVE_MEMBLOCK
 	select HAVE_MEMBLOCK_NODE_MAP
 	select HAVE_CMPXCHG_LOCAL
+	select HAVE_VIRT_CPU_ACCOUNTING
+	select VIRT_CPU_ACCOUNTING
 	select ARCH_DISCARD_MEMBLOCK
 	select ARCH_INLINE_SPIN_TRYLOCK
 	select ARCH_INLINE_SPIN_TRYLOCK_BH
diff --git a/init/Kconfig b/init/Kconfig
index d07dcf9..d3f04e1 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -267,6 +267,19 @@ config POSIX_MQUEUE_SYSCTL
 	depends on SYSCTL
 	default y
 
+config VIRT_CPU_ACCOUNTING
+	bool "Deterministic task and CPU time accounting"
+	depends on HAVE_VIRT_CPU_ACCOUNTING
+	default y if PPC64
+	help
+	  Select this option to enable more accurate task and CPU time
+	  accounting.  This is done by reading a CPU counter on each
+	  kernel entry and exit and on transitions within the kernel
+	  between system, softirq and hardirq state, so there is a
+	  small performance impact.  This also enables accounting of
+	  stolen time on logically-partitioned systems running on
+	  IBM POWER5-based machines.
+
 config BSD_PROCESS_ACCT
 	bool "BSD Process Accounting"
 	help
-- 
1.7.5.4