From: Chengming Zhou <zhouchengming@bytedance.com>
To: hannes@cmpxchg.org, mingo@redhat.com, peterz@infradead.org,
juri.lelli@redhat.com, vincent.guittot@linaro.org,
dietmar.eggemann@arm.com, rostedt@goodmis.org
Cc: linux-kernel@vger.kernel.org, songmuchun@bytedance.com,
zhouchengming@bytedance.com
Subject: [PATCH v2 4/4] psi: Optimize task switch inside shared cgroups
Date: Wed, 3 Mar 2021 11:46:59 +0800 [thread overview]
Message-ID: <20210303034659.91735-5-zhouchengming@bytedance.com> (raw)
In-Reply-To: <20210303034659.91735-1-zhouchengming@bytedance.com>
The commit 36b238d57172 ("psi: Optimize switching tasks inside shared
cgroups") only update cgroups whose state actually changes during a
task switch only in task preempt case, not in task sleep case.
We actually don't need to clear and set TSK_ONCPU state for common cgroups
of next and prev task in sleep case, that can save many psi_group_change
especially when most activity comes from one leaf cgroup.
sleep before:
psi_dequeue()
while ((group = iterate_groups(prev))) # all ancestors
psi_group_change(prev, .clear=TSK_RUNNING|TSK_ONCPU)
psi_task_switch()
while ((group = iterate_groups(next))) # all ancestors
psi_group_change(next, .set=TSK_ONCPU)
sleep after:
psi_dequeue()
nop
psi_task_switch()
while ((group = iterate_groups(next))) # until (prev & next)
psi_group_change(next, .set=TSK_ONCPU)
while ((group = iterate_groups(prev))) # all ancestors
psi_group_change(prev, .clear=common?TSK_RUNNING:TSK_RUNNING|TSK_ONCPU)
When a voluntary sleep switches to another task, we remove one call of
psi_group_change() for every common cgroup ancestor of the two tasks.
Signed-off-by: Muchun Song <songmuchun@bytedance.com>
Signed-off-by: Chengming Zhou <zhouchengming@bytedance.com>
---
Updates since v1:
- Many improvements in the comments and code from Johannes Weiner.
kernel/sched/psi.c | 35 +++++++++++++++++++++++++----------
kernel/sched/stats.h | 28 ++++++++++++----------------
2 files changed, 37 insertions(+), 26 deletions(-)
diff --git a/kernel/sched/psi.c b/kernel/sched/psi.c
index 3907a6b847aa..ee3c5b48622f 100644
--- a/kernel/sched/psi.c
+++ b/kernel/sched/psi.c
@@ -840,20 +840,35 @@ void psi_task_switch(struct task_struct *prev, struct task_struct *next,
}
}
- /*
- * If this is a voluntary sleep, dequeue will have taken care
- * of the outgoing TSK_ONCPU alongside TSK_RUNNING already. We
- * only need to deal with it during preemption.
- */
- if (sleep)
- return;
-
if (prev->pid) {
- psi_flags_change(prev, TSK_ONCPU, 0);
+ int clear = TSK_ONCPU, set = 0;
+
+ /*
+ * When we're going to sleep, psi_dequeue() lets us handle
+ * TSK_RUNNING and TSK_IOWAIT here, where we can combine it
+ * with TSK_ONCPU and save walking common ancestors twice.
+ */
+ if (sleep) {
+ clear |= TSK_RUNNING;
+ if (prev->in_iowait)
+ set |= TSK_IOWAIT;
+ }
+
+ psi_flags_change(prev, clear, set);
iter = NULL;
while ((group = iterate_groups(prev, &iter)) && group != common)
- psi_group_change(group, cpu, TSK_ONCPU, 0, true);
+ psi_group_change(group, cpu, clear, set, true);
+
+ /*
+ * TSK_ONCPU is handled up to the common ancestor. If we're tasked
+ * with dequeuing too, finish that for the rest of the hierarchy.
+ */
+ if (sleep) {
+ clear &= ~TSK_ONCPU;
+ for (; group; group = iterate_groups(prev, &iter))
+ psi_group_change(group, cpu, clear, set, true);
+ }
}
}
diff --git a/kernel/sched/stats.h b/kernel/sched/stats.h
index 9e4e67a94731..dc218e9f4558 100644
--- a/kernel/sched/stats.h
+++ b/kernel/sched/stats.h
@@ -84,28 +84,24 @@ static inline void psi_enqueue(struct task_struct *p, bool wakeup)
static inline void psi_dequeue(struct task_struct *p, bool sleep)
{
- int clear = TSK_RUNNING, set = 0;
+ int clear = TSK_RUNNING;
if (static_branch_likely(&psi_disabled))
return;
- if (!sleep) {
- if (p->in_memstall)
- clear |= TSK_MEMSTALL;
- } else {
- /*
- * When a task sleeps, schedule() dequeues it before
- * switching to the next one. Merge the clearing of
- * TSK_RUNNING and TSK_ONCPU to save an unnecessary
- * psi_task_change() call in psi_sched_switch().
- */
- clear |= TSK_ONCPU;
+ /*
+ * A voluntary sleep is a dequeue followed by a task switch. To
+ * avoid walking all ancestors twice, psi_task_switch() handles
+ * TSK_RUNNING and TSK_IOWAIT for us when it moves TSK_ONCPU.
+ * Do nothing here.
+ */
+ if (sleep)
+ return;
- if (p->in_iowait)
- set |= TSK_IOWAIT;
- }
+ if (p->in_memstall)
+ clear |= TSK_MEMSTALL;
- psi_task_change(p, clear, set);
+ psi_task_change(p, clear, 0);
}
static inline void psi_ttwu_dequeue(struct task_struct *p)
--
2.11.0
next prev parent reply other threads:[~2021-03-03 11:41 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-03-03 3:46 [PATCH v2 0/4] psi: Add PSI_CPU_FULL state and some code optimization Chengming Zhou
2021-03-03 3:46 ` [PATCH v2 1/4] psi: Add PSI_CPU_FULL state Chengming Zhou
2021-03-04 9:09 ` [tip: sched/core] " tip-bot2 for Chengming Zhou
2021-03-06 11:42 ` tip-bot2 for Chengming Zhou
2021-03-03 3:46 ` [PATCH v2 2/4] psi: Use ONCPU state tracking machinery to detect reclaim Chengming Zhou
2021-03-03 14:56 ` Johannes Weiner
2021-03-03 15:52 ` Peter Zijlstra
2021-03-04 9:09 ` [tip: sched/core] " tip-bot2 for Chengming Zhou
2021-03-06 11:42 ` tip-bot2 for Chengming Zhou
2021-03-03 3:46 ` [PATCH v2 3/4] psi: pressure states are unlikely Chengming Zhou
2021-03-04 9:09 ` [tip: sched/core] psi: Pressure " tip-bot2 for Johannes Weiner
2021-03-06 11:42 ` tip-bot2 for Johannes Weiner
2021-03-03 3:46 ` Chengming Zhou [this message]
2021-03-03 14:57 ` [PATCH v2 4/4] psi: Optimize task switch inside shared cgroups Johannes Weiner
2021-03-03 15:53 ` Peter Zijlstra
2021-03-03 16:01 ` [External] " Muchun Song
2021-03-04 9:09 ` [tip: sched/core] " tip-bot2 for Chengming Zhou
2021-03-06 11:42 ` tip-bot2 for Chengming Zhou
2021-03-03 14:59 ` [PATCH v2 0/4] psi: Add PSI_CPU_FULL state and some code optimization Johannes Weiner
2021-03-03 15:32 ` Peter Zijlstra
2021-03-03 16:05 ` Peter Zijlstra
2021-03-03 19:38 ` Johannes Weiner
2021-03-04 0:14 ` [External] " 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=20210303034659.91735-5-zhouchengming@bytedance.com \
--to=zhouchengming@bytedance.com \
--cc=dietmar.eggemann@arm.com \
--cc=hannes@cmpxchg.org \
--cc=juri.lelli@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=rostedt@goodmis.org \
--cc=songmuchun@bytedance.com \
--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;
as well as URLs for NNTP newsgroup(s).