From: Ankur Arora <ankur.a.arora@oracle.com>
To: linux-kernel@vger.kernel.org
Cc: tglx@linutronix.de, peterz@infradead.org,
torvalds@linux-foundation.org, paulmck@kernel.org,
akpm@linux-foundation.org, luto@kernel.org, bp@alien8.de,
dave.hansen@linux.intel.com, hpa@zytor.com, mingo@redhat.com,
juri.lelli@redhat.com, vincent.guittot@linaro.org,
willy@infradead.org, mgorman@suse.de, jpoimboe@kernel.org,
mark.rutland@arm.com, jgross@suse.com, andrew.cooper3@citrix.com,
bristot@kernel.org, mathieu.desnoyers@efficios.com,
geert@linux-m68k.org, glaubitz@physik.fu-berlin.de,
anton.ivanov@cambridgegreys.com, mattst88@gmail.com,
krypton@ulrich-teichert.org, rostedt@goodmis.org,
David.Laight@ACULAB.COM, richard@nod.at, mjguzik@gmail.com,
jon.grimm@amd.com, bharata@amd.com, raghavendra.kt@amd.com,
boris.ostrovsky@oracle.com, konrad.wilk@oracle.com,
Ankur Arora <ankur.a.arora@oracle.com>
Subject: [PATCH 23/30] sched/fair: handle tick expiry under lazy preemption
Date: Mon, 12 Feb 2024 21:55:47 -0800 [thread overview]
Message-ID: <20240213055554.1802415-24-ankur.a.arora@oracle.com> (raw)
In-Reply-To: <20240213055554.1802415-1-ankur.a.arora@oracle.com>
The default policy for lazy scheduling is to schedule in exit-to-user,
assuming that would happen within the remaining time quanta of the
task.
However, that runs into the 'hog' problem -- the target task might
be running in the kernel and might not relinquish CPU on its own.
Handle that by upgrading the ignored tif_resched(NR_lazy) bit to
tif_resched(NR_now) at the next tick.
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Vincent Guittot <vincent.guittot@linaro.org>
Originally-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/lkml/87jzshhexi.ffs@tglx/
Signed-off-by: Ankur Arora <ankur.a.arora@oracle.com>
---
Note:
Instead of special casing the tick, it might be simpler to always
do the upgrade on the second resched_curr().
The theoretical problem with doing that is that the current
approach deterministically provides a well-defined extra unit of
time. Going with a second resched_curr() might mean that the
amount of extra time the task gets depends on the vagaries of
the incoming resched_curr() (which is fine if it's mostly from
the tick; not fine if we could get it due to other reasons.)
Practically, both performed equally well in my tests.
Thoughts?
kernel/sched/core.c | 8 ++++++++
kernel/sched/deadline.c | 2 +-
kernel/sched/fair.c | 2 +-
kernel/sched/rt.c | 2 +-
kernel/sched/sched.h | 6 ++++++
5 files changed, 17 insertions(+), 3 deletions(-)
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index de963e8e2bee..5df59a4548dc 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -1038,6 +1038,10 @@ void wake_up_q(struct wake_q_head *head)
* For PREEMPT_AUTO: schedule idle threads eagerly, allow everything
* else, whether running in user or kernel context, to finish its time
* quanta, and mark for rescheduling at the next exit to user.
+ *
+ * Note: to avoid the hog problem, where the user does not relinquish
+ * CPU even after its time quanta has expired, upgrade lazy to eager
+ * on the second tick.
*/
static resched_t resched_opt_translate(struct task_struct *curr,
enum resched_opt opt)
@@ -1051,6 +1055,10 @@ static resched_t resched_opt_translate(struct task_struct *curr,
if (is_idle_task(curr))
return NR_now;
+ if (opt == RESCHED_TICK &&
+ unlikely(test_tsk_need_resched(curr, NR_lazy)))
+ return NR_now;
+
return NR_lazy;
}
diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
index b4e68cfc3c3a..b935e634fbf8 100644
--- a/kernel/sched/deadline.c
+++ b/kernel/sched/deadline.c
@@ -1379,7 +1379,7 @@ static void update_curr_dl_se(struct rq *rq, struct sched_dl_entity *dl_se, s64
}
if (!is_leftmost(dl_se, &rq->dl))
- resched_curr(rq);
+ resched_curr_tick(rq);
}
/*
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 278eebe6656a..92910b721adb 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -12621,7 +12621,7 @@ static void task_tick_fair(struct rq *rq, struct task_struct *curr, int queued)
}
if (resched)
- resched_curr(rq);
+ resched_curr_tick(rq);
if (static_branch_unlikely(&sched_numa_balancing))
task_tick_numa(rq, curr);
diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
index c57cc8427a57..1a2f3524d0eb 100644
--- a/kernel/sched/rt.c
+++ b/kernel/sched/rt.c
@@ -1023,7 +1023,7 @@ static void update_curr_rt(struct rq *rq)
rt_rq->rt_time += delta_exec;
exceeded = sched_rt_runtime_exceeded(rt_rq);
if (exceeded)
- resched_curr(rq);
+ resched_curr_tick(rq);
raw_spin_unlock(&rt_rq->rt_runtime_lock);
if (exceeded)
do_start_rt_bandwidth(sched_rt_bandwidth(rt_rq));
diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
index 3600a8673d08..c7e7acab1022 100644
--- a/kernel/sched/sched.h
+++ b/kernel/sched/sched.h
@@ -2465,6 +2465,7 @@ extern void reweight_task(struct task_struct *p, int prio);
enum resched_opt {
RESCHED_DEFAULT,
RESCHED_FORCE,
+ RESCHED_TICK,
};
extern void __resched_curr(struct rq *rq, enum resched_opt opt);
@@ -2474,6 +2475,11 @@ static inline void resched_curr(struct rq *rq)
__resched_curr(rq, RESCHED_DEFAULT);
}
+static inline void resched_curr_tick(struct rq *rq)
+{
+ __resched_curr(rq, RESCHED_TICK);
+}
+
extern void resched_cpu(int cpu);
extern struct rt_bandwidth def_rt_bandwidth;
--
2.31.1
next prev parent reply other threads:[~2024-02-13 5:57 UTC|newest]
Thread overview: 155+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-02-13 5:55 [PATCH 00/30] PREEMPT_AUTO: support lazy rescheduling Ankur Arora
2024-02-13 5:55 ` [PATCH 01/30] preempt: introduce CONFIG_PREEMPT_AUTO Ankur Arora
2024-02-13 5:55 ` [PATCH 02/30] thread_info: selector for TIF_NEED_RESCHED[_LAZY] Ankur Arora
2024-02-19 15:16 ` Thomas Gleixner
2024-02-20 22:50 ` Ankur Arora
2024-02-21 17:05 ` Thomas Gleixner
2024-02-21 18:26 ` Steven Rostedt
2024-02-21 20:03 ` Thomas Gleixner
2024-02-13 5:55 ` [PATCH 03/30] thread_info: tif_need_resched() now takes resched_t as param Ankur Arora
2024-02-14 3:17 ` kernel test robot
2024-02-14 14:08 ` Mark Rutland
2024-02-15 4:08 ` Ankur Arora
2024-02-19 12:30 ` Mark Rutland
2024-02-20 22:09 ` Ankur Arora
2024-02-19 15:21 ` Thomas Gleixner
2024-02-20 22:21 ` Ankur Arora
2024-02-21 17:07 ` Thomas Gleixner
2024-02-21 21:22 ` Ankur Arora
2024-02-13 5:55 ` [PATCH 04/30] sched: make test_*_tsk_thread_flag() return bool Ankur Arora
2024-02-14 14:12 ` Mark Rutland
2024-02-15 2:04 ` Ankur Arora
2024-02-13 5:55 ` [PATCH 05/30] sched: *_tsk_need_resched() now takes resched_t as param Ankur Arora
2024-02-19 15:26 ` Thomas Gleixner
2024-02-20 22:37 ` Ankur Arora
2024-02-21 17:10 ` Thomas Gleixner
2024-02-13 5:55 ` [PATCH 06/30] entry: handle lazy rescheduling at user-exit Ankur Arora
2024-02-19 15:29 ` Thomas Gleixner
2024-02-20 22:38 ` Ankur Arora
2024-02-13 5:55 ` [PATCH 07/30] entry/kvm: handle lazy rescheduling at guest-entry Ankur Arora
2024-02-13 5:55 ` [PATCH 08/30] entry: irqentry_exit only preempts for TIF_NEED_RESCHED Ankur Arora
2024-02-13 5:55 ` [PATCH 09/30] sched: __schedule_loop() doesn't need to check for need_resched_lazy() Ankur Arora
2024-02-13 5:55 ` [PATCH 10/30] sched: separate PREEMPT_DYNAMIC config logic Ankur Arora
2024-02-13 5:55 ` [PATCH 11/30] sched: runtime preemption config under PREEMPT_AUTO Ankur Arora
2024-02-13 5:55 ` [PATCH 12/30] rcu: limit PREEMPT_RCU to full preemption " Ankur Arora
2024-02-13 5:55 ` [PATCH 13/30] rcu: fix header guard for rcu_all_qs() Ankur Arora
2024-02-13 5:55 ` [PATCH 14/30] preempt,rcu: warn on PREEMPT_RCU=n, preempt=full Ankur Arora
2024-02-13 5:55 ` [PATCH 15/30] rcu: handle quiescent states for PREEMPT_RCU=n, PREEMPT_COUNT=y Ankur Arora
2024-03-10 10:03 ` Joel Fernandes
2024-03-10 18:56 ` Paul E. McKenney
2024-03-11 0:48 ` Joel Fernandes
2024-03-11 3:56 ` Paul E. McKenney
2024-03-11 15:01 ` Joel Fernandes
2024-03-11 20:51 ` Ankur Arora
2024-03-11 22:12 ` Thomas Gleixner
2024-03-11 5:18 ` Ankur Arora
2024-03-11 15:25 ` Joel Fernandes
2024-03-11 19:12 ` Thomas Gleixner
2024-03-11 19:53 ` Paul E. McKenney
2024-03-11 20:29 ` Thomas Gleixner
2024-03-12 0:01 ` Paul E. McKenney
2024-03-12 0:08 ` Joel Fernandes
2024-03-12 3:16 ` Ankur Arora
2024-03-12 3:24 ` Joel Fernandes
2024-03-12 5:23 ` Ankur Arora
2024-02-13 5:55 ` [PATCH 16/30] rcu: force context-switch " Ankur Arora
2024-02-13 5:55 ` [PATCH 17/30] x86/thread_info: define TIF_NEED_RESCHED_LAZY Ankur Arora
2024-02-14 13:25 ` Mark Rutland
2024-02-14 20:31 ` Ankur Arora
2024-02-19 12:32 ` Mark Rutland
2024-02-13 5:55 ` [PATCH 18/30] sched: prepare for lazy rescheduling in resched_curr() Ankur Arora
2024-02-13 5:55 ` [PATCH 19/30] sched: default preemption policy for PREEMPT_AUTO Ankur Arora
2024-02-13 5:55 ` [PATCH 20/30] sched: handle idle preemption " Ankur Arora
2024-02-13 5:55 ` [PATCH 21/30] sched: schedule eagerly in resched_cpu() Ankur Arora
2024-02-13 5:55 ` [PATCH 22/30] sched/fair: refactor update_curr(), entity_tick() Ankur Arora
2024-02-13 5:55 ` Ankur Arora [this message]
2024-02-21 21:38 ` [PATCH 23/30] sched/fair: handle tick expiry under lazy preemption Steven Rostedt
2024-02-28 13:47 ` Juri Lelli
2024-02-29 6:43 ` Ankur Arora
2024-02-29 9:33 ` Juri Lelli
2024-02-29 23:54 ` Ankur Arora
2024-03-01 0:28 ` Paul E. McKenney
2024-02-13 5:55 ` [PATCH 24/30] sched: support preempt=none under PREEMPT_AUTO Ankur Arora
2024-02-13 5:55 ` [PATCH 25/30] sched: support preempt=full " Ankur Arora
2024-02-13 5:55 ` [PATCH 26/30] sched: handle preempt=voluntary " Ankur Arora
2024-03-03 1:08 ` Joel Fernandes
2024-03-05 8:11 ` Ankur Arora
2024-03-06 20:42 ` Joel Fernandes
2024-03-07 19:01 ` Paul E. McKenney
2024-03-08 0:15 ` Joel Fernandes
2024-03-08 0:42 ` Paul E. McKenney
2024-03-08 4:22 ` Ankur Arora
2024-03-08 21:33 ` Paul E. McKenney
2024-03-11 4:50 ` Ankur Arora
2024-03-11 19:26 ` Paul E. McKenney
2024-03-11 20:09 ` Ankur Arora
2024-03-11 20:23 ` Linus Torvalds
2024-03-11 21:03 ` Ankur Arora
2024-03-12 0:03 ` Paul E. McKenney
2024-03-12 12:14 ` Thomas Gleixner
2024-03-12 19:40 ` Paul E. McKenney
2024-03-08 3:49 ` Ankur Arora
2024-03-08 5:29 ` Joel Fernandes
2024-03-08 6:54 ` Juri Lelli
2024-03-11 5:34 ` Ankur Arora
2024-02-13 5:55 ` [PATCH 27/30] sched: latency warn for TIF_NEED_RESCHED_LAZY Ankur Arora
2024-02-13 5:55 ` [PATCH 28/30] tracing: support lazy resched Ankur Arora
2024-02-13 5:55 ` [PATCH 29/30] Documentation: tracing: add TIF_NEED_RESCHED_LAZY Ankur Arora
2024-02-21 21:43 ` Steven Rostedt
2024-02-21 23:22 ` Ankur Arora
2024-02-21 23:53 ` Steven Rostedt
2024-03-01 23:33 ` Joel Fernandes
2024-03-02 3:09 ` Ankur Arora
2024-03-03 19:32 ` Joel Fernandes
2024-02-13 5:55 ` [PATCH 30/30] osnoise: handle quiescent states for PREEMPT_RCU=n, PREEMPTION=y Ankur Arora
2024-02-13 9:47 ` [PATCH 00/30] PREEMPT_AUTO: support lazy rescheduling Geert Uytterhoeven
2024-02-13 21:46 ` Ankur Arora
2024-02-14 23:57 ` Paul E. McKenney
2024-02-15 2:03 ` Ankur Arora
2024-02-15 3:45 ` Paul E. McKenney
2024-02-15 19:28 ` Paul E. McKenney
2024-02-15 20:04 ` Thomas Gleixner
2024-02-15 20:54 ` Paul E. McKenney
2024-02-15 20:53 ` Ankur Arora
2024-02-15 20:55 ` Paul E. McKenney
2024-02-15 21:24 ` Ankur Arora
2024-02-15 22:54 ` Paul E. McKenney
2024-02-15 22:56 ` Paul E. McKenney
2024-02-16 0:45 ` Ankur Arora
2024-02-16 2:59 ` Paul E. McKenney
2024-02-17 0:55 ` Paul E. McKenney
2024-02-17 3:59 ` Ankur Arora
2024-02-18 18:17 ` Paul E. McKenney
2024-02-19 16:48 ` Paul E. McKenney
2024-02-21 18:19 ` Steven Rostedt
2024-02-21 19:41 ` Paul E. McKenney
2024-02-21 20:11 ` Steven Rostedt
2024-02-21 20:22 ` Paul E. McKenney
2024-02-22 15:50 ` Mark Rutland
2024-02-22 19:11 ` Paul E. McKenney
2024-02-23 11:05 ` Mark Rutland
2024-02-23 15:31 ` Paul E. McKenney
2024-03-02 1:16 ` Paul E. McKenney
2024-03-19 11:45 ` Tasks RCU, ftrace, and trampolines (was: Re: [PATCH 00/30] PREEMPT_AUTO: support lazy rescheduling) Mark Rutland
2024-03-19 23:33 ` Paul E. McKenney
2024-02-21 6:48 ` [PATCH 00/30] PREEMPT_AUTO: support lazy rescheduling Ankur Arora
2024-02-21 17:44 ` Paul E. McKenney
2024-02-16 0:45 ` Ankur Arora
2024-02-21 12:23 ` Raghavendra K T
2024-02-21 17:15 ` Thomas Gleixner
2024-02-21 17:27 ` Raghavendra K T
2024-02-21 21:16 ` Ankur Arora
2024-02-22 4:05 ` Raghavendra K T
2024-02-22 21:23 ` Thomas Gleixner
2024-02-23 3:14 ` Ankur Arora
2024-02-23 6:28 ` Raghavendra K T
2024-02-24 3:15 ` Raghavendra K T
2024-02-27 17:45 ` Ankur Arora
2024-02-22 13:04 ` Raghavendra K T
2024-04-23 15:21 ` Shrikanth Hegde
2024-04-23 16:13 ` Linus Torvalds
2024-04-26 7:46 ` Shrikanth Hegde
2024-04-26 19:00 ` Ankur Arora
2024-05-07 11:16 ` Shrikanth Hegde
2024-05-08 5:18 ` Ankur Arora
2024-05-15 14:31 ` Shrikanth Hegde
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=20240213055554.1802415-24-ankur.a.arora@oracle.com \
--to=ankur.a.arora@oracle.com \
--cc=David.Laight@ACULAB.COM \
--cc=akpm@linux-foundation.org \
--cc=andrew.cooper3@citrix.com \
--cc=anton.ivanov@cambridgegreys.com \
--cc=bharata@amd.com \
--cc=boris.ostrovsky@oracle.com \
--cc=bp@alien8.de \
--cc=bristot@kernel.org \
--cc=dave.hansen@linux.intel.com \
--cc=geert@linux-m68k.org \
--cc=glaubitz@physik.fu-berlin.de \
--cc=hpa@zytor.com \
--cc=jgross@suse.com \
--cc=jon.grimm@amd.com \
--cc=jpoimboe@kernel.org \
--cc=juri.lelli@redhat.com \
--cc=konrad.wilk@oracle.com \
--cc=krypton@ulrich-teichert.org \
--cc=linux-kernel@vger.kernel.org \
--cc=luto@kernel.org \
--cc=mark.rutland@arm.com \
--cc=mathieu.desnoyers@efficios.com \
--cc=mattst88@gmail.com \
--cc=mgorman@suse.de \
--cc=mingo@redhat.com \
--cc=mjguzik@gmail.com \
--cc=paulmck@kernel.org \
--cc=peterz@infradead.org \
--cc=raghavendra.kt@amd.com \
--cc=richard@nod.at \
--cc=rostedt@goodmis.org \
--cc=tglx@linutronix.de \
--cc=torvalds@linux-foundation.org \
--cc=vincent.guittot@linaro.org \
--cc=willy@infradead.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).