[PATCH v8 0/4] sched: Fix missing irq time when CONFIG_IRQ_TIME_ACCOUNTING is enabled

[PATCH v8 0/4] sched: Fix missing irq time when CONFIG_IRQ_TIME_ACCOUNTING is enabled

[Yafang Shao]

After enabling CONFIG_IRQ_TIME_ACCOUNTING to track IRQ pressure in our
container environment, we encountered several user-visible behavioral
changes:

- Interrupted IRQ/softirq time is excluded in the cpuacct cgroup

  This breaks userspace applications that rely on CPU usage data from
  cgroups to monitor CPU pressure. This patchset resolves the issue by
  ensuring that IRQ/softirq time is included in the cgroup of the
  interrupted tasks.

- getrusage(2) does not include time interrupted by IRQ/softirq

  Some services use getrusage(2) to check if workloads are experiencing CPU
  pressure. Since IRQ/softirq time is no longer included in task runtime,
  getrusage(2) can no longer reflect the CPU pressure caused by heavy
  interrupts.

This patchset addresses the first issue, which is relatively
straightforward. Once this solution is accepted, I will address the second
issue in a follow-up patchset.

Enabling CONFIG_IRQ_TIME_ACCOUNTING results in the CPU
utilization metric excluding the time spent in IRQs. This means we
lose visibility into how long the CPU was actually interrupted in
comparison to its total utilization. The user will misleadlingly
consider the *interrupted irq time* as *sleep time* as follows,

  |<----Runtime---->|<----Sleep---->|<----Runtime---->|<---Sleep-->|

While actually it is:

  |<----Runtime---->|<--Interrupted time-->|<----Runtime---->|<---Sleep-->|

Currently, the only ways to monitor interrupt time are through IRQ PSI or
the IRQ time recorded in delay accounting. However, these metrics are
independent of CPU utilization, which makes it difficult to combine them
into a single, unified measure

CPU utilization is a critical metric for almost all workloads, and
it's problematic if it fails to reflect the full extent of system
pressure. This situation is similar to iowait: when a task is in
iowait, it could be due to other tasks performing I/O. It doesn’t
matter if the I/O is being done by one of your tasks or by someone
else's; what matters is that your task is stalled and waiting on I/O.
Similarly, a comprehensive CPU utilization metric should reflect all
sources of pressure, including IRQ time, to provide a more accurate
representation of workload behavior.

One of the applications impacted by this issue is our Redis load-balancing
service. The setup operates as follows:

                   ----------------
                   | Load Balancer|
                   ----------------
                /    |      |        \
               /     |      |         \ 
          Server1 Server2 Server3 ... ServerN

Although the load balancer's algorithm is complex, it follows some core
principles:

- When server CPU utilization increases, it adds more servers and deploys
  additional instances to meet SLA requirements.
- When server CPU utilization decreases, it scales down by decommissioning
  servers and reducing the number of instances to save on costs.

On our servers, the majority of IRQ/softIRQ activity originates from
network traffic, and we consistently enable Receive Flow Steering
(RFS) [0]. This configuration ensures that softIRQs are more likely to
interrupt the tasks responsible for processing the corresponding
packets. As a result, the distribution of softIRQs is not random but
instead closely aligned with the packet-handling tasks.

The load balancer is malfunctioning due to the exclusion of IRQ time from
CPU utilization calculations. What's worse, there is no effective way to
add the irq time back into the CPU utilization based on current
available metrics. Therefore, we have to change the kernel code.

Link: https://lwn.net/Articles/381955/ [0]

Changes:
v7->v8:
- Fix a build failure report by kernel test robot

v6->v7: https://lore.kernel.org/all/20241215032315.43698-1-laoar.shao@gmail.com/
- Fix psi_show() (Michal)

v5->v6: https://lore.kernel.org/all/20241211131729.43996-1-laoar.shao@gmail.com/
- Return EOPNOTSUPP in psi_show() if irqtime is disabled (Michal)
- Collect Reviewed-by from Michal 

v4->v5: https://lore.kernel.org/all/20241108132904.6932-1-laoar.shao@gmail.com/
- Don't use static key in the IRQ_TIME_ACCOUNTING=n case (Peter)
- Rename psi_irq_time to irq_time (Peter)
- Use CPUTIME_IRQ instead of CPUTIME_SOFTIRQ (Peter)

v3->v4: https://lore.kernel.org/all/20241101031750.1471-1-laoar.shao@gmail.com/
- Rebase

v2->v3:
- Add a helper account_irqtime() to avoid redundant code (Johannes)

v1->v2: https://lore.kernel.org/cgroups/20241008061951.3980-1-laoar.shao@gmail.com/
- Fix lockdep issues reported by kernel test robot <oliver.sang@intel.com>

v1: https://lore.kernel.org/all/20240923090028.16368-1-laoar.shao@gmail.com/

Yafang Shao (4):
  sched: Define sched_clock_irqtime as static key
  sched: Don't account irq time if sched_clock_irqtime is disabled
  sched, psi: Don't account irq time if sched_clock_irqtime is disabled
  sched: Fix cgroup irq time for CONFIG_IRQ_TIME_ACCOUNTING

 kernel/sched/core.c    | 77 +++++++++++++++++++++++++++++-------------
 kernel/sched/cputime.c | 16 ++++-----
 kernel/sched/psi.c     | 16 ++++-----
 kernel/sched/sched.h   | 15 +++++++-
 kernel/sched/stats.h   |  7 ++--
 5 files changed, 86 insertions(+), 45 deletions(-)

-- 
2.43.5

[PATCH v8 1/4] sched: Define sched_clock_irqtime as static key

[Yafang Shao]

Since CPU time accounting is a performance-critical path, let's define
sched_clock_irqtime as a static key to minimize potential overhead.

Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
Reviewed-by: Michal Koutný <mkoutny@suse.com>
---
 kernel/sched/cputime.c | 16 +++++++---------
 kernel/sched/sched.h   | 13 +++++++++++++
 2 files changed, 20 insertions(+), 9 deletions(-)

diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c
index 0bed0fa1acd9..5d9143dd0879 100644
--- a/kernel/sched/cputime.c
+++ b/kernel/sched/cputime.c
@@ -9,6 +9,8 @@
 
 #ifdef CONFIG_IRQ_TIME_ACCOUNTING
 
+DEFINE_STATIC_KEY_FALSE(sched_clock_irqtime);
+
 /*
  * There are no locks covering percpu hardirq/softirq time.
  * They are only modified in vtime_account, on corresponding CPU
@@ -22,16 +24,14 @@
  */
 DEFINE_PER_CPU(struct irqtime, cpu_irqtime);
 
-static int sched_clock_irqtime;
-
 void enable_sched_clock_irqtime(void)
 {
-	sched_clock_irqtime = 1;
+	static_branch_enable(&sched_clock_irqtime);
 }
 
 void disable_sched_clock_irqtime(void)
 {
-	sched_clock_irqtime = 0;
+	static_branch_disable(&sched_clock_irqtime);
 }
 
 static void irqtime_account_delta(struct irqtime *irqtime, u64 delta,
@@ -57,7 +57,7 @@ void irqtime_account_irq(struct task_struct *curr, unsigned int offset)
 	s64 delta;
 	int cpu;
 
-	if (!sched_clock_irqtime)
+	if (!irqtime_enabled())
 		return;
 
 	cpu = smp_processor_id();
@@ -90,8 +90,6 @@ static u64 irqtime_tick_accounted(u64 maxtime)
 
 #else /* CONFIG_IRQ_TIME_ACCOUNTING */
 
-#define sched_clock_irqtime	(0)
-
 static u64 irqtime_tick_accounted(u64 dummy)
 {
 	return 0;
@@ -478,7 +476,7 @@ void account_process_tick(struct task_struct *p, int user_tick)
 	if (vtime_accounting_enabled_this_cpu())
 		return;
 
-	if (sched_clock_irqtime) {
+	if (irqtime_enabled()) {
 		irqtime_account_process_tick(p, user_tick, 1);
 		return;
 	}
@@ -507,7 +505,7 @@ void account_idle_ticks(unsigned long ticks)
 {
 	u64 cputime, steal;
 
-	if (sched_clock_irqtime) {
+	if (irqtime_enabled()) {
 		irqtime_account_idle_ticks(ticks);
 		return;
 	}
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index aef716c41edb..7e8c73110884 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -3233,6 +3233,12 @@ struct irqtime {
 };
 
 DECLARE_PER_CPU(struct irqtime, cpu_irqtime);
+DECLARE_STATIC_KEY_FALSE(sched_clock_irqtime);
+
+static inline int irqtime_enabled(void)
+{
+	return static_branch_likely(&sched_clock_irqtime);
+}
 
 /*
  * Returns the irqtime minus the softirq time computed by ksoftirqd.
@@ -3253,6 +3259,13 @@ static inline u64 irq_time_read(int cpu)
 	return total;
 }
 
+#else
+
+static inline int irqtime_enabled(void)
+{
+	return 0;
+}
+
 #endif /* CONFIG_IRQ_TIME_ACCOUNTING */
 
 #ifdef CONFIG_CPU_FREQ
-- 
2.43.5

[PATCH v8 2/4] sched: Don't account irq time if sched_clock_irqtime is disabled

[Yafang Shao]

sched_clock_irqtime may be disabled due to the clock source, in which case
IRQ time should not be accounted. Let's add a conditional check to avoid
unnecessary logic.

Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
Reviewed-by: Michal Koutný <mkoutny@suse.com>
---
 kernel/sched/core.c | 44 +++++++++++++++++++++++---------------------
 1 file changed, 23 insertions(+), 21 deletions(-)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 84902936a620..22dfcd3e92ed 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -740,29 +740,31 @@ static void update_rq_clock_task(struct rq *rq, s64 delta)
 	s64 __maybe_unused steal = 0, irq_delta = 0;
 
 #ifdef CONFIG_IRQ_TIME_ACCOUNTING
-	irq_delta = irq_time_read(cpu_of(rq)) - rq->prev_irq_time;
+	if (irqtime_enabled()) {
+		irq_delta = irq_time_read(cpu_of(rq)) - rq->prev_irq_time;
 
-	/*
-	 * Since irq_time is only updated on {soft,}irq_exit, we might run into
-	 * this case when a previous update_rq_clock() happened inside a
-	 * {soft,}IRQ region.
-	 *
-	 * When this happens, we stop ->clock_task and only update the
-	 * prev_irq_time stamp to account for the part that fit, so that a next
-	 * update will consume the rest. This ensures ->clock_task is
-	 * monotonic.
-	 *
-	 * It does however cause some slight miss-attribution of {soft,}IRQ
-	 * time, a more accurate solution would be to update the irq_time using
-	 * the current rq->clock timestamp, except that would require using
-	 * atomic ops.
-	 */
-	if (irq_delta > delta)
-		irq_delta = delta;
+		/*
+		 * Since irq_time is only updated on {soft,}irq_exit, we might run into
+		 * this case when a previous update_rq_clock() happened inside a
+		 * {soft,}IRQ region.
+		 *
+		 * When this happens, we stop ->clock_task and only update the
+		 * prev_irq_time stamp to account for the part that fit, so that a next
+		 * update will consume the rest. This ensures ->clock_task is
+		 * monotonic.
+		 *
+		 * It does however cause some slight miss-attribution of {soft,}IRQ
+		 * time, a more accurate solution would be to update the irq_time using
+		 * the current rq->clock timestamp, except that would require using
+		 * atomic ops.
+		 */
+		if (irq_delta > delta)
+			irq_delta = delta;
 
-	rq->prev_irq_time += irq_delta;
-	delta -= irq_delta;
-	delayacct_irq(rq->curr, irq_delta);
+		rq->prev_irq_time += irq_delta;
+		delta -= irq_delta;
+		delayacct_irq(rq->curr, irq_delta);
+	}
 #endif
 #ifdef CONFIG_PARAVIRT_TIME_ACCOUNTING
 	if (static_key_false((&paravirt_steal_rq_enabled))) {
-- 
2.43.5

[PATCH v8 3/4] sched, psi: Don't account irq time if sched_clock_irqtime is disabled

[Yafang Shao]

sched_clock_irqtime may be disabled due to the clock source. When disabled,
irq_time_read() won't change over time, so there is nothing to account. We
can save iterating the whole hierarchy on every tick and context switch.

Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Reviewed-by: Michal Koutný <mkoutny@suse.com>
Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
---
 kernel/sched/psi.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/kernel/sched/psi.c b/kernel/sched/psi.c
index 84dad1511d1e..bb56805e3d47 100644
--- a/kernel/sched/psi.c
+++ b/kernel/sched/psi.c
@@ -998,7 +998,7 @@ void psi_account_irqtime(struct rq *rq, struct task_struct *curr, struct task_st
 	s64 delta;
 	u64 irq;
 
-	if (static_branch_likely(&psi_disabled))
+	if (static_branch_likely(&psi_disabled) || !irqtime_enabled())
 		return;
 
 	if (!curr->pid)
@@ -1240,6 +1240,11 @@ int psi_show(struct seq_file *m, struct psi_group *group, enum psi_res res)
 	if (static_branch_likely(&psi_disabled))
 		return -EOPNOTSUPP;
 
+#ifdef CONFIG_IRQ_TIME_ACCOUNTING
+	if (!irqtime_enabled() && res == PSI_IRQ)
+		return -EOPNOTSUPP;
+#endif
+
 	/* Update averages before reporting them */
 	mutex_lock(&group->avgs_lock);
 	now = sched_clock();
-- 
2.43.5

[PATCH v8 4/4] sched: Fix cgroup irq time for CONFIG_IRQ_TIME_ACCOUNTING

[Yafang Shao]

After enabling CONFIG_IRQ_TIME_ACCOUNTING to monitor IRQ pressure in our
container environment, we observed several noticeable behavioral changes.

One of our IRQ-heavy services, such as Redis, reported a significant
reduction in CPU usage after upgrading to the new kernel with
CONFIG_IRQ_TIME_ACCOUNTING enabled. However, despite adding more threads
to handle an increased workload, the CPU usage could not be raised. In
other words, even though the container’s CPU usage appeared low, it was
unable to process more workloads to utilize additional CPU resources, which
caused issues.

This behavior can be demonstrated using netperf:

  function start_server() {
      for j in `seq 1 3`; do
          netserver -p $[12345+j] > /dev/null &
      done
  }

  server_ip=$1
  function start_client() {
    # That applies to cgroup2 as well.
    mkdir -p /sys/fs/cgroup/cpuacct/test
    echo $$ > /sys/fs/cgroup/cpuacct/test/cgroup.procs
    for j in `seq 1 3`; do
        port=$[12345+j]
        taskset -c 0 netperf -H ${server_ip} -l ${run_time:-30000}   \
                -t TCP_STREAM -p $port -- -D -m 1k -M 1K -s 8k -S 8k \
                > /dev/null &
    done
  }

  start_server
  start_client

We can verify the CPU usage of the test cgroup using cpuacct.stat. The
output shows:

  system: 53
  user: 2

The CPU usage of the cgroup is relatively low at around 55%, but this usage
doesn't increase, even with more netperf tasks. The reason is that CPU0 is
at 100% utilization, as confirmed by mpstat:

  02:56:22 PM  CPU    %usr   %nice    %sys %iowait    %irq   %soft  %steal  %guest  %gnice   %idle
  02:56:23 PM    0    0.99    0.00   55.45    0.00    0.99   42.57    0.00    0.00    0.00    0.00

  02:56:23 PM  CPU    %usr   %nice    %sys %iowait    %irq   %soft  %steal  %guest  %gnice   %idle
  02:56:24 PM    0    2.00    0.00   55.00    0.00    0.00   43.00    0.00    0.00    0.00    0.00

It is clear that the %soft is excluded in the cgroup of the interrupted
task. This behavior is unexpected. We should include IRQ time in the
cgroup to reflect the pressure the group is under.

After a thorough analysis, I discovered that this change in behavior is due
to commit 305e6835e055 ("sched: Do not account irq time to current task"),
which altered whether IRQ time should be charged to the interrupted task.
While I agree that a task should not be penalized by random interrupts, the
task itself cannot progress while interrupted. Therefore, the interrupted
time should be reported to the user.

The system metric in cpuacct.stat is crucial in indicating whether a
container is under heavy system pressure, including IRQ/softirq activity.
Hence, IRQ/softirq time should be included in the cpuacct system usage,
which also applies to cgroup2’s rstat.

The reason it doesn't just add the cgroup_account_*() to
irqtime_account_irq() is that it might result in performance hit to hold
the rq_lock in the critical path. Taking inspiration from
commit ddae0ca2a8fe ("sched: Move psi_account_irqtime() out of
update_rq_clock_task() hotpath"), I've now adapted the approach to handle
it in a non-critical path, reducing the performance impact.

Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Koutný <mkoutny@suse.com>
---
 kernel/sched/core.c  | 33 +++++++++++++++++++++++++++++++--
 kernel/sched/psi.c   | 13 +++----------
 kernel/sched/sched.h |  2 +-
 kernel/sched/stats.h |  7 ++++---
 4 files changed, 39 insertions(+), 16 deletions(-)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 22dfcd3e92ed..7faacf320af9 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -5622,6 +5622,35 @@ __setup("resched_latency_warn_ms=", setup_resched_latency_warn_ms);
 static inline u64 cpu_resched_latency(struct rq *rq) { return 0; }
 #endif /* CONFIG_SCHED_DEBUG */
 
+#ifdef CONFIG_IRQ_TIME_ACCOUNTING
+static void account_irqtime(struct rq *rq, struct task_struct *curr,
+			    struct task_struct *prev)
+{
+	int cpu = smp_processor_id();
+	s64 delta;
+	u64 irq;
+
+	if (!irqtime_enabled())
+		return;
+
+	irq = irq_time_read(cpu);
+	delta = (s64)(irq - rq->irq_time);
+	if (delta < 0)
+		return;
+
+	rq->irq_time = irq;
+	psi_account_irqtime(rq, curr, prev, delta);
+	cgroup_account_cputime(curr, delta);
+	/* We account both softirq and irq into CPUTIME_IRQ */
+	cgroup_account_cputime_field(curr, CPUTIME_IRQ, delta);
+}
+#else
+static inline void account_irqtime(struct rq *rq, struct task_struct *curr,
+				   struct task_struct *prev)
+{
+}
+#endif
+
 /*
  * This function gets called by the timer code, with HZ frequency.
  * We call it with interrupts disabled.
@@ -5644,7 +5673,7 @@ void sched_tick(void)
 	rq_lock(rq, &rf);
 	donor = rq->donor;
 
-	psi_account_irqtime(rq, donor, NULL);
+	account_irqtime(rq, donor, NULL);
 
 	update_rq_clock(rq);
 	hw_pressure = arch_scale_hw_pressure(cpu_of(rq));
@@ -6751,7 +6780,7 @@ static void __sched notrace __schedule(int sched_mode)
 		++*switch_count;
 
 		migrate_disable_switch(rq, prev);
-		psi_account_irqtime(rq, prev, next);
+		account_irqtime(rq, prev, next);
 		psi_sched_switch(prev, next, block);
 
 		trace_sched_switch(preempt, prev, next, prev_state);
diff --git a/kernel/sched/psi.c b/kernel/sched/psi.c
index bb56805e3d47..6fa3233fcba9 100644
--- a/kernel/sched/psi.c
+++ b/kernel/sched/psi.c
@@ -990,15 +990,14 @@ void psi_task_switch(struct task_struct *prev, struct task_struct *next,
 }
 
 #ifdef CONFIG_IRQ_TIME_ACCOUNTING
-void psi_account_irqtime(struct rq *rq, struct task_struct *curr, struct task_struct *prev)
+void psi_account_irqtime(struct rq *rq, struct task_struct *curr, struct task_struct *prev,
+			 s64 delta)
 {
 	int cpu = task_cpu(curr);
 	struct psi_group *group;
 	struct psi_group_cpu *groupc;
-	s64 delta;
-	u64 irq;
 
-	if (static_branch_likely(&psi_disabled) || !irqtime_enabled())
+	if (static_branch_likely(&psi_disabled))
 		return;
 
 	if (!curr->pid)
@@ -1009,12 +1008,6 @@ void psi_account_irqtime(struct rq *rq, struct task_struct *curr, struct task_st
 	if (prev && task_psi_group(prev) == group)
 		return;
 
-	irq = irq_time_read(cpu);
-	delta = (s64)(irq - rq->psi_irq_time);
-	if (delta < 0)
-		return;
-	rq->psi_irq_time = irq;
-
 	do {
 		u64 now;
 
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index 7e8c73110884..570cc19a0060 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -1227,7 +1227,7 @@ struct rq {
 
 #ifdef CONFIG_IRQ_TIME_ACCOUNTING
 	u64			prev_irq_time;
-	u64			psi_irq_time;
+	u64			irq_time;
 #endif
 #ifdef CONFIG_PARAVIRT
 	u64			prev_steal_time;
diff --git a/kernel/sched/stats.h b/kernel/sched/stats.h
index 8ee0add5a48a..b9ed9fea5ab7 100644
--- a/kernel/sched/stats.h
+++ b/kernel/sched/stats.h
@@ -111,10 +111,11 @@ void psi_task_change(struct task_struct *task, int clear, int set);
 void psi_task_switch(struct task_struct *prev, struct task_struct *next,
 		     bool sleep);
 #ifdef CONFIG_IRQ_TIME_ACCOUNTING
-void psi_account_irqtime(struct rq *rq, struct task_struct *curr, struct task_struct *prev);
+void psi_account_irqtime(struct rq *rq, struct task_struct *curr,
+			 struct task_struct *prev, s64 delta);
 #else
 static inline void psi_account_irqtime(struct rq *rq, struct task_struct *curr,
-				       struct task_struct *prev) {}
+				       struct task_struct *prev, s64 delta) {}
 #endif /*CONFIG_IRQ_TIME_ACCOUNTING */
 /*
  * PSI tracks state that persists across sleeps, such as iowaits and
@@ -224,7 +225,7 @@ static inline void psi_sched_switch(struct task_struct *prev,
 				    struct task_struct *next,
 				    bool sleep) {}
 static inline void psi_account_irqtime(struct rq *rq, struct task_struct *curr,
-				       struct task_struct *prev) {}
+				       struct task_struct *prev, s64 delta) {}
 #endif /* CONFIG_PSI */
 
 #ifdef CONFIG_SCHED_INFO
-- 
2.43.5

[PATCH v8 resend 0/4] sched: Fix missing irq time when CONFIG_IRQ_TIME_ACCOUNTING is enabled

[Yafang Shao]

After enabling CONFIG_IRQ_TIME_ACCOUNTING to track IRQ pressure in our
container environment, we encountered several user-visible behavioral
changes:

- Interrupted IRQ/softirq time is excluded in the cpuacct cgroup

  This breaks userspace applications that rely on CPU usage data from
  cgroups to monitor CPU pressure. This patchset resolves the issue by
  ensuring that IRQ/softirq time is included in the cgroup of the
  interrupted tasks.

- getrusage(2) does not include time interrupted by IRQ/softirq

  Some services use getrusage(2) to check if workloads are experiencing CPU
  pressure. Since IRQ/softirq time is no longer included in task runtime,
  getrusage(2) can no longer reflect the CPU pressure caused by heavy
  interrupts.

This patchset addresses the first issue, which is relatively
straightforward. Once this solution is accepted, I will address the second
issue in a follow-up patchset.

Enabling CONFIG_IRQ_TIME_ACCOUNTING results in the CPU
utilization metric excluding the time spent in IRQs. This means we
lose visibility into how long the CPU was actually interrupted in
comparison to its total utilization. The user will misleadlingly
consider the *interrupted irq time* as *sleep time* as follows,

  |<----Runtime---->|<----Sleep---->|<----Runtime---->|<---Sleep-->|

While actually it is:

  |<----Runtime---->|<--Interrupted time-->|<----Runtime---->|<---Sleep-->|

Currently, the only ways to monitor interrupt time are through IRQ PSI or
the IRQ time recorded in delay accounting. However, these metrics are
independent of CPU utilization, which makes it difficult to combine them
into a single, unified measure

CPU utilization is a critical metric for almost all workloads, and
it's problematic if it fails to reflect the full extent of system
pressure. This situation is similar to iowait: when a task is in
iowait, it could be due to other tasks performing I/O. It doesn’t
matter if the I/O is being done by one of your tasks or by someone
else's; what matters is that your task is stalled and waiting on I/O.
Similarly, a comprehensive CPU utilization metric should reflect all
sources of pressure, including IRQ time, to provide a more accurate
representation of workload behavior.

One of the applications impacted by this issue is our Redis load-balancing
service. The setup operates as follows:

                   ----------------
                   | Load Balancer|
                   ----------------
                /    |      |        \
               /     |      |         \ 
          Server1 Server2 Server3 ... ServerN

Although the load balancer's algorithm is complex, it follows some core
principles:

- When server CPU utilization increases, it adds more servers and deploys
  additional instances to meet SLA requirements.
- When server CPU utilization decreases, it scales down by decommissioning
  servers and reducing the number of instances to save on costs.

On our servers, the majority of IRQ/softIRQ activity originates from
network traffic, and we consistently enable Receive Flow Steering
(RFS) [0]. This configuration ensures that softIRQs are more likely to
interrupt the tasks responsible for processing the corresponding
packets. As a result, the distribution of softIRQs is not random but
instead closely aligned with the packet-handling tasks.

The load balancer is malfunctioning due to the exclusion of IRQ time from
CPU utilization calculations. What's worse, there is no effective way to
add the irq time back into the CPU utilization based on current
available metrics. Therefore, we have to change the kernel code.

Link: https://lwn.net/Articles/381955/ [0]

Changes:
v7->v8:
- Fix a build failure report by kernel test robot

v6->v7: https://lore.kernel.org/all/20241215032315.43698-1-laoar.shao@gmail.com/
- Fix psi_show() (Michal)

v5->v6: https://lore.kernel.org/all/20241211131729.43996-1-laoar.shao@gmail.com/
- Return EOPNOTSUPP in psi_show() if irqtime is disabled (Michal)
- Collect Reviewed-by from Michal 

v4->v5: https://lore.kernel.org/all/20241108132904.6932-1-laoar.shao@gmail.com/
- Don't use static key in the IRQ_TIME_ACCOUNTING=n case (Peter)
- Rename psi_irq_time to irq_time (Peter)
- Use CPUTIME_IRQ instead of CPUTIME_SOFTIRQ (Peter)

v3->v4: https://lore.kernel.org/all/20241101031750.1471-1-laoar.shao@gmail.com/
- Rebase

v2->v3:
- Add a helper account_irqtime() to avoid redundant code (Johannes)

v1->v2: https://lore.kernel.org/cgroups/20241008061951.3980-1-laoar.shao@gmail.com/
- Fix lockdep issues reported by kernel test robot <oliver.sang@intel.com>

v1: https://lore.kernel.org/all/20240923090028.16368-1-laoar.shao@gmail.com/

Yafang Shao (4):
  sched: Define sched_clock_irqtime as static key
  sched: Don't account irq time if sched_clock_irqtime is disabled
  sched, psi: Don't account irq time if sched_clock_irqtime is disabled
  sched: Fix cgroup irq time for CONFIG_IRQ_TIME_ACCOUNTING

 kernel/sched/core.c    | 77 +++++++++++++++++++++++++++++-------------
 kernel/sched/cputime.c | 16 ++++-----
 kernel/sched/psi.c     | 16 ++++-----
 kernel/sched/sched.h   | 15 +++++++-
 kernel/sched/stats.h   |  7 ++--
 5 files changed, 86 insertions(+), 45 deletions(-)

-- 
2.43.5

[PATCH v8 2/4] sched: Don't account irq time if sched_clock_irqtime is disabled

[Yafang Shao]

sched_clock_irqtime may be disabled due to the clock source, in which case
IRQ time should not be accounted. Let's add a conditional check to avoid
unnecessary logic.

Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
Reviewed-by: Michal Koutný <mkoutny@suse.com>
---
 kernel/sched/core.c | 44 +++++++++++++++++++++++---------------------
 1 file changed, 23 insertions(+), 21 deletions(-)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 84902936a620..22dfcd3e92ed 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -740,29 +740,31 @@ static void update_rq_clock_task(struct rq *rq, s64 delta)
 	s64 __maybe_unused steal = 0, irq_delta = 0;
 
 #ifdef CONFIG_IRQ_TIME_ACCOUNTING
-	irq_delta = irq_time_read(cpu_of(rq)) - rq->prev_irq_time;
+	if (irqtime_enabled()) {
+		irq_delta = irq_time_read(cpu_of(rq)) - rq->prev_irq_time;
 
-	/*
-	 * Since irq_time is only updated on {soft,}irq_exit, we might run into
-	 * this case when a previous update_rq_clock() happened inside a
-	 * {soft,}IRQ region.
-	 *
-	 * When this happens, we stop ->clock_task and only update the
-	 * prev_irq_time stamp to account for the part that fit, so that a next
-	 * update will consume the rest. This ensures ->clock_task is
-	 * monotonic.
-	 *
-	 * It does however cause some slight miss-attribution of {soft,}IRQ
-	 * time, a more accurate solution would be to update the irq_time using
-	 * the current rq->clock timestamp, except that would require using
-	 * atomic ops.
-	 */
-	if (irq_delta > delta)
-		irq_delta = delta;
+		/*
+		 * Since irq_time is only updated on {soft,}irq_exit, we might run into
+		 * this case when a previous update_rq_clock() happened inside a
+		 * {soft,}IRQ region.
+		 *
+		 * When this happens, we stop ->clock_task and only update the
+		 * prev_irq_time stamp to account for the part that fit, so that a next
+		 * update will consume the rest. This ensures ->clock_task is
+		 * monotonic.
+		 *
+		 * It does however cause some slight miss-attribution of {soft,}IRQ
+		 * time, a more accurate solution would be to update the irq_time using
+		 * the current rq->clock timestamp, except that would require using
+		 * atomic ops.
+		 */
+		if (irq_delta > delta)
+			irq_delta = delta;
 
-	rq->prev_irq_time += irq_delta;
-	delta -= irq_delta;
-	delayacct_irq(rq->curr, irq_delta);
+		rq->prev_irq_time += irq_delta;
+		delta -= irq_delta;
+		delayacct_irq(rq->curr, irq_delta);
+	}
 #endif
 #ifdef CONFIG_PARAVIRT_TIME_ACCOUNTING
 	if (static_key_false((&paravirt_steal_rq_enabled))) {
-- 
2.43.5

Re: [PATCH v8 2/4] sched: Don't account irq time if sched_clock_irqtime is disabled

[Vincent Guittot]

On Fri, 3 Jan 2025 at 03:24, Yafang Shao <laoar.shao@gmail.com> wrote:
>
> sched_clock_irqtime may be disabled due to the clock source, in which case
> IRQ time should not be accounted. Let's add a conditional check to avoid
> unnecessary logic.
>
> Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
> Reviewed-by: Michal Koutný <mkoutny@suse.com>

Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>


> ---
>  kernel/sched/core.c | 44 +++++++++++++++++++++++---------------------
>  1 file changed, 23 insertions(+), 21 deletions(-)
>
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index 84902936a620..22dfcd3e92ed 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -740,29 +740,31 @@ static void update_rq_clock_task(struct rq *rq, s64 delta)
>         s64 __maybe_unused steal = 0, irq_delta = 0;
>
>  #ifdef CONFIG_IRQ_TIME_ACCOUNTING
> -       irq_delta = irq_time_read(cpu_of(rq)) - rq->prev_irq_time;
> +       if (irqtime_enabled()) {
> +               irq_delta = irq_time_read(cpu_of(rq)) - rq->prev_irq_time;
>
> -       /*
> -        * Since irq_time is only updated on {soft,}irq_exit, we might run into
> -        * this case when a previous update_rq_clock() happened inside a
> -        * {soft,}IRQ region.
> -        *
> -        * When this happens, we stop ->clock_task and only update the
> -        * prev_irq_time stamp to account for the part that fit, so that a next
> -        * update will consume the rest. This ensures ->clock_task is
> -        * monotonic.
> -        *
> -        * It does however cause some slight miss-attribution of {soft,}IRQ
> -        * time, a more accurate solution would be to update the irq_time using
> -        * the current rq->clock timestamp, except that would require using
> -        * atomic ops.
> -        */
> -       if (irq_delta > delta)
> -               irq_delta = delta;
> +               /*
> +                * Since irq_time is only updated on {soft,}irq_exit, we might run into
> +                * this case when a previous update_rq_clock() happened inside a
> +                * {soft,}IRQ region.
> +                *
> +                * When this happens, we stop ->clock_task and only update the
> +                * prev_irq_time stamp to account for the part that fit, so that a next
> +                * update will consume the rest. This ensures ->clock_task is
> +                * monotonic.
> +                *
> +                * It does however cause some slight miss-attribution of {soft,}IRQ
> +                * time, a more accurate solution would be to update the irq_time using
> +                * the current rq->clock timestamp, except that would require using
> +                * atomic ops.
> +                */
> +               if (irq_delta > delta)
> +                       irq_delta = delta;
>
> -       rq->prev_irq_time += irq_delta;
> -       delta -= irq_delta;
> -       delayacct_irq(rq->curr, irq_delta);
> +               rq->prev_irq_time += irq_delta;
> +               delta -= irq_delta;
> +               delayacct_irq(rq->curr, irq_delta);
> +       }
>  #endif
>  #ifdef CONFIG_PARAVIRT_TIME_ACCOUNTING
>         if (static_key_false((&paravirt_steal_rq_enabled))) {
> --
> 2.43.5
>