mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Chengming Zhou <zhouchengming@bytedance.com>
To: hannes@cmpxchg.org, tj@kernel.org, corbet@lwn.net,
	surenb@google.com, mingo@redhat.com, peterz@infradead.org,
	vincent.guittot@linaro.org, dietmar.eggemann@arm.com,
	rostedt@goodmis.org, bsegall@google.com
Cc: cgroups@vger.kernel.org, linux-doc@vger.kernel.org,
	linux-kernel@vger.kernel.org, songmuchun@bytedance.com,
	Chengming Zhou <zhouchengming@bytedance.com>
Subject: [PATCH v2 09/10] sched/psi: per-cgroup PSI stats disable/re-enable interface
Date: Mon,  8 Aug 2022 19:03:40 +0800	[thread overview]
Message-ID: <20220808110341.15799-10-zhouchengming@bytedance.com> (raw)
In-Reply-To: <20220808110341.15799-1-zhouchengming@bytedance.com>

PSI accounts stalls for each cgroup separately and aggregates it
at each level of the hierarchy. This may cause non-negligible overhead
for some workloads when under deep level of the hierarchy.

commit 3958e2d0c34e ("cgroup: make per-cgroup pressure stall tracking configurable")
make PSI to skip per-cgroup stall accounting, only account system-wide
to avoid this each level overhead.

But for our use case, we also want leaf cgroup PSI stats accounted for
userspace adjustment on that cgroup, apart from only system-wide adjustment.

So this patch introduce a per-cgroup PSI stats disable/re-enable
interface "cgroup.psi", which is a read-write single value file that
allowed values are "0" and "1", the defaults is "1" so per-cgroup
PSI stats is enabled by default.

Implementation details:

It should be relatively straight-forward to disable and re-enable
state aggregation, time tracking, averaging on a per-cgroup level,
if we can live with losing history from while it was disabled.
I.e. the avgs will restart from 0, total= will have gaps.

But it's hard or complex to stop/restart groupc->tasks[] updates,
which is not implemented in this patch. So we always update
groupc->tasks[] and PSI_ONCPU bit in psi_group_change() even when
the cgroup PSI stats is disabled.

Suggested-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Chengming Zhou <zhouchengming@bytedance.com>
---
 Documentation/admin-guide/cgroup-v2.rst |  7 ++++
 include/linux/psi.h                     |  2 ++
 include/linux/psi_types.h               |  2 ++
 kernel/cgroup/cgroup.c                  | 43 +++++++++++++++++++++++++
 kernel/sched/psi.c                      | 40 +++++++++++++++++++----
 5 files changed, 87 insertions(+), 7 deletions(-)

diff --git a/Documentation/admin-guide/cgroup-v2.rst b/Documentation/admin-guide/cgroup-v2.rst
index dd84e34bc051..ade40506ab80 100644
--- a/Documentation/admin-guide/cgroup-v2.rst
+++ b/Documentation/admin-guide/cgroup-v2.rst
@@ -968,6 +968,13 @@ All cgroup core files are prefixed with "cgroup."
 	killing cgroups is a process directed operation, i.e. it affects
 	the whole thread-group.
 
+  cgroup.psi
+	A read-write single value file that allowed values are "0" and "1".
+	The default is "1".
+
+	Writing "0" to the file will disable the cgroup PSI stats accounting.
+	Writing "1" to the file will re-enable the cgroup PSI stats accounting.
+
   irq.pressure
 	A read-write nested-keyed file.
 
diff --git a/include/linux/psi.h b/include/linux/psi.h
index aa168a038242..1138ccffd76b 100644
--- a/include/linux/psi.h
+++ b/include/linux/psi.h
@@ -33,6 +33,7 @@ __poll_t psi_trigger_poll(void **trigger_ptr, struct file *file,
 int psi_cgroup_alloc(struct cgroup *cgrp);
 void psi_cgroup_free(struct cgroup *cgrp);
 void cgroup_move_task(struct task_struct *p, struct css_set *to);
+void psi_cgroup_enable(struct psi_group *group, bool enable);
 #endif
 
 #else /* CONFIG_PSI */
@@ -54,6 +55,7 @@ static inline void cgroup_move_task(struct task_struct *p, struct css_set *to)
 {
 	rcu_assign_pointer(p->cgroups, to);
 }
+static inline void psi_cgroup_enable(struct psi_group *group, bool enable) {}
 #endif
 
 #endif /* CONFIG_PSI */
diff --git a/include/linux/psi_types.h b/include/linux/psi_types.h
index 4677655f6ca1..fced39e255aa 100644
--- a/include/linux/psi_types.h
+++ b/include/linux/psi_types.h
@@ -147,6 +147,8 @@ struct psi_trigger {
 };
 
 struct psi_group {
+	bool enabled;
+
 	/* Protects data used by the aggregator */
 	struct mutex avgs_lock;
 
diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c
index 91de8ff7fa50..6ba56983b5a5 100644
--- a/kernel/cgroup/cgroup.c
+++ b/kernel/cgroup/cgroup.c
@@ -3709,6 +3709,43 @@ static ssize_t cgroup_irq_pressure_write(struct kernfs_open_file *of,
 }
 #endif
 
+static int cgroup_psi_show(struct seq_file *seq, void *v)
+{
+	struct cgroup *cgrp = seq_css(seq)->cgroup;
+	struct psi_group *psi = cgroup_ino(cgrp) == 1 ? &psi_system : &cgrp->psi;
+
+	seq_printf(seq, "%d\n", psi->enabled);
+
+	return 0;
+}
+
+static ssize_t cgroup_psi_write(struct kernfs_open_file *of,
+				char *buf, size_t nbytes, loff_t off)
+{
+	ssize_t ret;
+	int enable;
+	struct cgroup *cgrp;
+	struct psi_group *psi;
+
+	ret = kstrtoint(strstrip(buf), 0, &enable);
+	if (ret)
+		return ret;
+
+	if (enable < 0 || enable > 1)
+		return -ERANGE;
+
+	cgrp = cgroup_kn_lock_live(of->kn, false);
+	if (!cgrp)
+		return -ENOENT;
+
+	psi = cgroup_ino(cgrp) == 1 ? &psi_system : &cgrp->psi;
+	psi_cgroup_enable(psi, enable);
+
+	cgroup_kn_unlock(of->kn);
+
+	return nbytes;
+}
+
 static __poll_t cgroup_pressure_poll(struct kernfs_open_file *of,
 					  poll_table *pt)
 {
@@ -5115,6 +5152,12 @@ static struct cftype cgroup_base_files[] = {
 		.release = cgroup_pressure_release,
 	},
 #endif
+	{
+		.name = "cgroup.psi",
+		.flags = CFTYPE_PRESSURE,
+		.seq_show = cgroup_psi_show,
+		.write = cgroup_psi_write,
+	},
 #endif /* CONFIG_PSI */
 	{ }	/* terminate */
 };
diff --git a/kernel/sched/psi.c b/kernel/sched/psi.c
index 58f8092c938f..9df1686ee02d 100644
--- a/kernel/sched/psi.c
+++ b/kernel/sched/psi.c
@@ -181,6 +181,7 @@ static void group_init(struct psi_group *group)
 {
 	int cpu;
 
+	group->enabled = true;
 	for_each_possible_cpu(cpu)
 		seqcount_init(&per_cpu_ptr(group->pcpu, cpu)->seq);
 	group->avg_last_update = sched_clock();
@@ -700,17 +701,16 @@ static void psi_group_change(struct psi_group *group, int cpu,
 	groupc = per_cpu_ptr(group->pcpu, cpu);
 
 	/*
-	 * First we assess the aggregate resource states this CPU's
-	 * tasks have been in since the last change, and account any
-	 * SOME and FULL time these may have resulted in.
-	 *
-	 * Then we update the task counts according to the state
+	 * First we update the task counts according to the state
 	 * change requested through the @clear and @set bits.
+	 *
+	 * Then if the cgroup PSI stats accounting enabled, we
+	 * assess the aggregate resource states this CPU's tasks
+	 * have been in since the last change, and account any
+	 * SOME and FULL time these may have resulted in.
 	 */
 	write_seqcount_begin(&groupc->seq);
 
-	record_times(groupc, now);
-
 	/*
 	 * Start with TSK_ONCPU, which doesn't have a corresponding
 	 * task count - it's just a boolean flag directly encoded in
@@ -750,6 +750,14 @@ static void psi_group_change(struct psi_group *group, int cpu,
 		if (set & (1 << t))
 			groupc->tasks[t]++;
 
+	if (!group->enabled) {
+		if (groupc->state_mask & (1 << PSI_NONIDLE))
+			record_times(groupc, now);
+		groupc->state_mask = state_mask;
+		write_seqcount_end(&groupc->seq);
+		return;
+	}
+
 	for (s = 0; s < NR_PSI_STATES; s++) {
 		if (test_state(groupc->tasks, s, state_mask & PSI_ONCPU))
 			state_mask |= (1 << s);
@@ -766,6 +774,7 @@ static void psi_group_change(struct psi_group *group, int cpu,
 	if (unlikely((state_mask & PSI_ONCPU) && cpu_curr(cpu)->in_memstall))
 		state_mask |= (1 << PSI_MEM_FULL);
 
+	record_times(groupc, now);
 	groupc->state_mask = state_mask;
 
 	write_seqcount_end(&groupc->seq);
@@ -1088,6 +1097,23 @@ void cgroup_move_task(struct task_struct *task, struct css_set *to)
 
 	task_rq_unlock(rq, task, &rf);
 }
+
+void psi_cgroup_enable(struct psi_group *group, bool enable)
+{
+	struct psi_group_cpu *groupc;
+	int cpu;
+	u64 now;
+
+	if (group->enabled == enable)
+		return;
+	group->enabled = enable;
+
+	for_each_possible_cpu(cpu) {
+		groupc = per_cpu_ptr(group->pcpu, cpu);
+		now = cpu_clock(cpu);
+		psi_group_change(group, cpu, 0, 0, now, true);
+	}
+}
 #endif /* CONFIG_CGROUPS */
 
 int psi_show(struct seq_file *m, struct psi_group *group, enum psi_res res)
-- 
2.36.1


  parent reply	other threads:[~2022-08-08 11:06 UTC|newest]

Thread overview: 32+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-08-08 11:03 [PATCH v2 00/10] sched/psi: some optimization and extension Chengming Zhou
2022-08-08 11:03 ` [PATCH v2 01/10] sched/psi: fix periodic aggregation shut off Chengming Zhou
2022-08-08 11:03 ` [PATCH v2 02/10] sched/psi: optimize task switch inside shared cgroups again Chengming Zhou
2022-08-08 11:03 ` [PATCH v2 03/10] sched/psi: move private helpers to sched/stats.h Chengming Zhou
2022-08-08 11:03 ` [PATCH v2 04/10] sched/psi: don't change task psi_flags when migrate CPU/group Chengming Zhou
2022-08-08 11:03 ` [PATCH v2 05/10] sched/psi: don't create cgroup PSI files when psi_disabled Chengming Zhou
2022-08-08 11:03 ` [PATCH v2 06/10] sched/psi: save percpu memory when !psi_cgroups_enabled Chengming Zhou
2022-08-08 11:03 ` [PATCH v2 07/10] sched/psi: remove NR_ONCPU task accounting Chengming Zhou
2022-08-16 10:40   ` Chengming Zhou
2022-08-08 11:03 ` [PATCH v2 08/10] sched/psi: add PSI_IRQ to track IRQ/SOFTIRQ pressure Chengming Zhou
2022-08-08 11:03 ` Chengming Zhou [this message]
2022-08-09 17:48   ` [PATCH v2 09/10] sched/psi: per-cgroup PSI stats disable/re-enable interface Tejun Heo
2022-08-10  0:39     ` Chengming Zhou
2022-08-10  1:30       ` Chengming Zhou
2022-08-10 15:25         ` Johannes Weiner
2022-08-10 17:27           ` Tejun Heo
2022-08-11  2:09             ` Chengming Zhou
2022-08-15 13:23           ` Michal Koutný
2022-08-23  6:18             ` Chengming Zhou
2022-08-23 15:35               ` Johannes Weiner
2022-08-23 15:43                 ` Chengming Zhou
2022-08-23 16:20                 ` Tejun Heo
2022-08-12 10:14   ` Michal Koutný
2022-08-12 12:36     ` Chengming Zhou
2022-08-15 13:23       ` Michal Koutný
2022-08-15 15:49   ` Johannes Weiner
2022-08-15 19:50     ` Tejun Heo
2022-08-16 13:06     ` Chengming Zhou
2022-08-08 11:03 ` [PATCH v2 10/10] sched/psi: cache parent psi_group to speed up groups iterate Chengming Zhou
2022-08-15 13:25 ` [PATCH v2 00/10] sched/psi: some optimization and extension Michal Koutný
2022-08-16 14:01   ` Chengming Zhou
2022-08-17 15:19     ` Chengming Zhou

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20220808110341.15799-10-zhouchengming@bytedance.com \
    --to=zhouchengming@bytedance.com \
    --cc=bsegall@google.com \
    --cc=cgroups@vger.kernel.org \
    --cc=corbet@lwn.net \
    --cc=dietmar.eggemann@arm.com \
    --cc=hannes@cmpxchg.org \
    --cc=linux-doc@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@redhat.com \
    --cc=peterz@infradead.org \
    --cc=rostedt@goodmis.org \
    --cc=songmuchun@bytedance.com \
    --cc=surenb@google.com \
    --cc=tj@kernel.org \
    --cc=vincent.guittot@linaro.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®