public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Julien Desfossez <jdesfossez@digitalocean.com>
To: Aaron Lu <aaron.lu@linux.alibaba.com>
Cc: Vineeth Remanan Pillai <vpillai@digitalocean.com>,
	Phil Auld <pauld@redhat.com>,
	Nishanth Aravamudan <naravamudan@digitalocean.com>,
	Peter Zijlstra <peterz@infradead.org>,
	Tim Chen <tim.c.chen@linux.intel.com>,
	mingo@kernel.org, tglx@linutronix.de, pjt@google.com,
	torvalds@linux-foundation.org, linux-kernel@vger.kernel.org,
	subhra.mazumdar@oracle.com, fweisbec@gmail.com,
	keescook@chromium.org, kerrnel@google.com,
	Aaron Lu <aaron.lwe@gmail.com>,
	Aubrey Li <aubrey.intel@gmail.com>,
	Valentin Schneider <valentin.schneider@arm.com>,
	Mel Gorman <mgorman@techsingularity.net>,
	Pawan Gupta <pawan.kumar.gupta@linux.intel.com>,
	Paolo Bonzini <pbonzini@redhat.com>
Subject: Re: [RFC PATCH v2 00/17] Core scheduling v2
Date: Mon, 6 May 2019 15:39:37 -0400	[thread overview]
Message-ID: <20190506193937.GA10264@sinkpad> (raw)
In-Reply-To: <20190429035320.GB128241@aaronlu>

On 29-Apr-2019 11:53:21 AM, Aaron Lu wrote:
> On Tue, Apr 23, 2019 at 06:45:27PM +0000, Vineeth Remanan Pillai wrote:
> > >> - Processes with different tags can still share the core
> > 
> > > I may have missed something... Could you explain this statement?
> > 
> > > This, to me, is the whole point of the patch series. If it's not
> > > doing this then ... what?
> > 
> > What I meant was, the patch needs some more work to be accurate.
> > There are some race conditions where the core violation can still
> > happen. In our testing, we saw around 1 to 5% of the time being
> > shared with incompatible processes. One example of this happening
> > is as follows(let cpu 0 and 1 be siblings):
> > - cpu 0 selects a process with a cookie
> > - cpu 1 selects a higher priority process without cookie
> > - Selection process restarts for cpu 0 and it might select a
> >   process with cookie but with lesser priority.
> > - Since it is lesser priority, the logic in pick_next_task
> >   doesn't compare again for the cookie(trusts pick_task) and
> >   proceeds.
> > 
> > This is one of the scenarios that we saw from traces, but there
> > might be other race conditions as well. Fix seems a little
> > involved and We are working on that.
> 
> This is what I have used to make sure no two unmatched tasks being
> scheduled on the same core: (on top of v1, I thinks it's easier to just
> show the diff instead of commenting on various places of the patches :-)

We imported this fix in v2 and made some small changes and optimizations
(with and without Peter’s fix from https://lkml.org/lkml/2019/4/26/658)
and in both cases, the performance problem where the core can end up
idle with tasks in its runqueues came back.

This is pretty easy to reproduce with a multi-file disk write benchmark.

Here is the patch based on your changes applied on v2 (on top of Peter’s
fix):

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 07f3f0c..e09fa25 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -3653,6 +3653,13 @@ static inline bool cookie_match(struct task_struct *a, struct task_struct *b)
 }
 
 // XXX fairness/fwd progress conditions
+/*
+ * Returns
+ * - NULL if there is no runnable task for this class.
+ * - the highest priority task for this runqueue if it matches
+ *   rq->core->core_cookie or its priority is greater than max.
+ * - Else returns idle_task.
+ */
 static struct task_struct *
 pick_task(struct rq *rq, const struct sched_class *class, struct task_struct *max)
 {
@@ -3660,19 +3667,36 @@ pick_task(struct rq *rq, const struct sched_class *class, struct task_struct *ma
 	unsigned long cookie = rq->core->core_cookie;
 
 	class_pick = class->pick_task(rq);
-	if (!cookie)
+	if (!class_pick)
+		return NULL;
+
+	if (!cookie) {
+		/*
+		 * If class_pick is tagged, return it only if it has
+		 * higher priority than max.
+		 */
+		if (max && class_pick->core_cookie &&
+		    core_prio_less(class_pick, max))
+			return idle_sched_class.pick_task(rq);
+
+		return class_pick;
+	}
+
+	/*
+	 * If there is a cooke match here, return early.
+	 */
+	if (class_pick->core_cookie == cookie)
 		return class_pick;
 
 	cookie_pick = sched_core_find(rq, cookie);
-	if (!class_pick)
-		return cookie_pick;
 
 	/*
 	 * If class > max && class > cookie, it is the highest priority task on
 	 * the core (so far) and it must be selected, otherwise we must go with
 	 * the cookie pick in order to satisfy the constraint.
 	 */
-	if (cpu_prio_less(cookie_pick, class_pick) && core_prio_less(max, class_pick))
+	if (cpu_prio_less(cookie_pick, class_pick) &&
+	    (!max || core_prio_less(max, class_pick)))
 		return class_pick;
 
 	return cookie_pick;
@@ -3742,8 +3766,16 @@ pick_next_task(struct rq *rq, struct task_struct *prev, struct rq_flags *rf)
 
 		rq_i->core_pick = NULL;
 
-		if (i != cpu)
+		if (i != cpu) {
 			update_rq_clock(rq_i);
+
+			/*
+			 * If a sibling is idle, we can initiate an
+			 * unconstrained pick.
+			 */
+			if (is_idle_task(rq_i->curr) && prev_cookie)
+				prev_cookie = 0UL;
+		}
 	}
 
 	/*
@@ -3820,12 +3852,14 @@ pick_next_task(struct rq *rq, struct task_struct *prev, struct rq_flags *rf)
 			/*
 			 * If this new candidate is of higher priority than the
 			 * previous; and they're incompatible; we need to wipe
-			 * the slate and start over.
+			 * the slate and start over. pick_task makes sure that
+			 * p's priority is more than max if it doesn't match
+			 * max's cookie.
 			 *
 			 * NOTE: this is a linear max-filter and is thus bounded
 			 * in execution time.
 			 */
-			if (!max || core_prio_less(max, p)) {
+			if (!max || !cookie_match(max, p)) {
 				struct task_struct *old_max = max;
 
 				rq->core->core_cookie = p->core_cookie;
@@ -3833,7 +3867,7 @@ pick_next_task(struct rq *rq, struct task_struct *prev, struct rq_flags *rf)
 
 				trace_printk("max: %s/%d %lx\n", max->comm, max->pid, max->core_cookie);
 
-				if (old_max && !cookie_match(old_max, p)) {
+				if (old_max) {
 					for_each_cpu(j, smt_mask) {
 						if (j == i)
 							continue;
@@ -3879,6 +3913,23 @@ next_class:;
 
 	trace_printk("picked: %s/%d %lx\n", next->comm, next->pid, next->core_cookie);
 
+	/* make sure we didn't break L1TF */
+	for_each_cpu(i, smt_mask) {
+		struct rq *rq_i = cpu_rq(i);
+		if (i == cpu)
+			continue;
+
+		if (likely(cookie_match(next, rq_i->core_pick)))
+			continue;
+
+		trace_printk("[%d]: cookie mismatch. %s/%d/0x%lx/0x%lx\n",
+			     rq_i->cpu, rq_i->core_pick->comm,
+			     rq_i->core_pick->pid,
+			     rq_i->core_pick->core_cookie,
+			     rq_i->core->core_cookie);
+		WARN_ON_ONCE(1);
+	}
+
 done:
 	set_next_task(rq, next);
 	return next;


  reply	other threads:[~2019-05-06 19:39 UTC|newest]

Thread overview: 109+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-04-23 16:18 [RFC PATCH v2 00/17] Core scheduling v2 Vineeth Remanan Pillai
2019-04-23 16:18 ` [RFC PATCH v2 01/17] stop_machine: Fix stop_cpus_in_progress ordering Vineeth Remanan Pillai
2019-04-23 16:18 ` [RFC PATCH v2 02/17] sched: Fix kerneldoc comment for ia64_set_curr_task Vineeth Remanan Pillai
2019-04-23 16:18 ` [RFC PATCH v2 03/17] sched: Wrap rq::lock access Vineeth Remanan Pillai
2019-04-23 16:18 ` [RFC PATCH v2 04/17] sched/{rt,deadline}: Fix set_next_task vs pick_next_task Vineeth Remanan Pillai
2019-04-23 16:18 ` [RFC PATCH v2 05/17] sched: Add task_struct pointer to sched_class::set_curr_task Vineeth Remanan Pillai
2019-04-23 16:18 ` [RFC PATCH v2 06/17] sched/fair: Export newidle_balance() Vineeth Remanan Pillai
2019-04-23 16:18 ` [RFC PATCH v2 07/17] sched: Allow put_prev_task() to drop rq->lock Vineeth Remanan Pillai
2019-04-23 16:18 ` [RFC PATCH v2 08/17] sched: Rework pick_next_task() slow-path Vineeth Remanan Pillai
2019-04-23 16:18 ` [RFC PATCH v2 09/17] sched: Introduce sched_class::pick_task() Vineeth Remanan Pillai
2019-04-26 14:02   ` Peter Zijlstra
2019-04-26 16:10     ` Vineeth Remanan Pillai
2019-04-29  5:38   ` Aaron Lu
2019-04-23 16:18 ` [RFC PATCH v2 10/17] sched: Core-wide rq->lock Vineeth Remanan Pillai
2019-04-23 16:18 ` [RFC PATCH v2 11/17] sched: Basic tracking of matching tasks Vineeth Remanan Pillai
2019-04-24  0:08   ` Tim Chen
2019-04-24 20:43     ` Vineeth Remanan Pillai
2019-04-24 22:12       ` Tim Chen
2019-04-25 14:35       ` Phil Auld
2019-05-22 19:52         ` Vineeth Remanan Pillai
2019-04-24  0:17   ` Tim Chen
2019-04-24 20:43     ` Vineeth Remanan Pillai
2019-04-29  3:36   ` Aaron Lu
2019-05-10 13:06     ` Peter Zijlstra
2019-04-29  6:15   ` Aaron Lu
2019-05-01 23:27     ` Tim Chen
2019-05-03  0:06       ` Tim Chen
2019-05-08 15:49         ` Aubrey Li
2019-05-08 18:19           ` Subhra Mazumdar
2019-05-08 18:37             ` Subhra Mazumdar
2019-05-09  0:01               ` Aubrey Li
2019-05-09  0:25                 ` Subhra Mazumdar
2019-05-09  1:38                   ` Aubrey Li
2019-05-09  2:14                     ` Subhra Mazumdar
2019-05-09 15:10                       ` Aubrey Li
2019-05-09 17:50                         ` Subhra Mazumdar
2019-05-10  0:09                           ` Tim Chen
2019-04-23 16:18 ` [RFC PATCH v2 12/17] sched: A quick and dirty cgroup tagging interface Vineeth Remanan Pillai
2019-04-25 14:26   ` Phil Auld
2019-04-26 14:13     ` Peter Zijlstra
2019-04-26 14:19       ` Phil Auld
2019-05-10 15:12   ` Julien Desfossez
2019-04-23 16:18 ` [RFC PATCH v2 13/17] sched: Add core wide task selection and scheduling Vineeth Remanan Pillai
2019-04-29  7:13   ` Aaron Lu
2019-05-18 15:37   ` Aubrey Li
2019-05-20 13:04     ` Phil Auld
2019-05-20 14:04       ` Vineeth Pillai
2019-05-21  8:19         ` Aubrey Li
2019-05-21 13:24           ` Vineeth Pillai
2019-04-23 16:18 ` [RFC PATCH v2 14/17] sched/fair: Add a few assertions Vineeth Remanan Pillai
2019-04-23 16:18 ` [RFC PATCH v2 15/17] sched: Trivial forced-newidle balancer Vineeth Remanan Pillai
2019-04-23 23:46   ` Aubrey Li
2019-04-24 14:03     ` Vineeth Remanan Pillai
2019-04-24 14:05     ` Vineeth Remanan Pillai
2019-04-23 16:18 ` [RFC PATCH v2 16/17] sched: Wake up sibling if it has something to run Vineeth Remanan Pillai
2019-04-26 15:03   ` Peter Zijlstra
2019-04-29 12:36     ` Julien Desfossez
2019-04-23 16:18 ` [RFC PATCH v2 17/17] sched: Debug bits Vineeth Remanan Pillai
2019-05-17 17:18   ` Aubrey Li
2019-04-23 18:02 ` [RFC PATCH v2 00/17] Core scheduling v2 Phil Auld
2019-04-23 18:45   ` Vineeth Remanan Pillai
2019-04-29  3:53     ` Aaron Lu
2019-05-06 19:39       ` Julien Desfossez [this message]
2019-05-08  2:30         ` Aaron Lu
2019-05-08 17:49           ` Julien Desfossez
2019-05-09  2:11             ` Aaron Lu
2019-05-15 21:36               ` Vineeth Remanan Pillai
2019-04-23 23:25 ` Aubrey Li
2019-04-24 11:19   ` Vineeth Remanan Pillai
2019-05-15 21:39     ` Vineeth Remanan Pillai
2019-04-24 13:13 ` Aubrey Li
2019-04-24 14:00   ` Julien Desfossez
2019-04-25  3:15     ` Aubrey Li
2019-04-25  9:55       ` Ingo Molnar
2019-04-25 14:46         ` Mel Gorman
2019-04-25 18:53           ` Ingo Molnar
2019-04-25 18:59             ` Thomas Gleixner
2019-04-25 19:34               ` Ingo Molnar
2019-04-25 21:31             ` Mel Gorman
2019-04-26  8:42               ` Ingo Molnar
2019-04-26 10:43                 ` Mel Gorman
2019-04-26 18:37                   ` Subhra Mazumdar
2019-04-26 19:49                     ` Mel Gorman
2019-04-26  9:45               ` Ingo Molnar
2019-04-26 10:19                 ` Mel Gorman
2019-04-27  9:06                   ` Ingo Molnar
2019-04-26  9:51               ` Ingo Molnar
2019-04-26 14:15             ` Phil Auld
2019-04-26  2:18         ` Aubrey Li
2019-04-26  9:51           ` Ingo Molnar
2019-04-27  3:51         ` Aubrey Li
2019-04-27  9:17           ` Ingo Molnar
2019-04-27 14:04             ` Aubrey Li
2019-04-27 14:21               ` Ingo Molnar
2019-04-27 15:54                 ` Aubrey Li
2019-04-28  9:33                   ` Ingo Molnar
2019-04-28 10:29                     ` Aubrey Li
2019-04-28 12:17                       ` Ingo Molnar
2019-04-29  2:17                         ` Li, Aubrey
2019-04-29  6:14                           ` Ingo Molnar
2019-04-29 13:25                             ` Li, Aubrey
2019-04-29 15:39                               ` Phil Auld
2019-04-30  1:24                                 ` Aubrey Li
2019-04-29 16:00                               ` Ingo Molnar
2019-04-30  1:34                                 ` Aubrey Li
2019-04-30  4:42                                   ` Ingo Molnar
2019-05-18  0:58                                     ` Li, Aubrey
2019-05-18  1:08                                       ` Li, Aubrey
2019-04-25 14:36 ` Julien Desfossez

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=20190506193937.GA10264@sinkpad \
    --to=jdesfossez@digitalocean.com \
    --cc=aaron.lu@linux.alibaba.com \
    --cc=aaron.lwe@gmail.com \
    --cc=aubrey.intel@gmail.com \
    --cc=fweisbec@gmail.com \
    --cc=keescook@chromium.org \
    --cc=kerrnel@google.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mgorman@techsingularity.net \
    --cc=mingo@kernel.org \
    --cc=naravamudan@digitalocean.com \
    --cc=pauld@redhat.com \
    --cc=pawan.kumar.gupta@linux.intel.com \
    --cc=pbonzini@redhat.com \
    --cc=peterz@infradead.org \
    --cc=pjt@google.com \
    --cc=subhra.mazumdar@oracle.com \
    --cc=tglx@linutronix.de \
    --cc=tim.c.chen@linux.intel.com \
    --cc=torvalds@linux-foundation.org \
    --cc=valentin.schneider@arm.com \
    --cc=vpillai@digitalocean.com \
    /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