linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Srivatsa S. Bhat" <srivatsa.bhat@linux.vnet.ibm.com>
To: paul@paulmenage.org, a.p.zijlstra@chello.nl, mingo@elte.hu,
	rjw@sisk.pl, tj@kernel.org
Cc: frank.rowand@am.sony.com, pjt@google.com, tglx@linutronix.de,
	lizf@cn.fujitsu.com, prashanth@linux.vnet.ibm.com,
	paulmck@linux.vnet.ibm.com, vatsa@linux.vnet.ibm.com,
	srivatsa.bhat@linux.vnet.ibm.com, linux-kernel@vger.kernel.org,
	linux-pm@vger.kernel.org
Subject: [PATCH 3/4] cpuset: Add function to introduce CPUs to cpusets during CPU online
Date: Wed, 08 Feb 2012 00:27:09 +0530	[thread overview]
Message-ID: <20120207185651.7482.18642.stgit@srivatsabhat.in.ibm.com> (raw)
In-Reply-To: <20120207185411.7482.43576.stgit@srivatsabhat.in.ibm.com>

During a CPU online operation, at present, that CPU is added only to
top_cpuset.cpus_allowed, but not to any of the other cpusets, which
is a mistake.

So, add a function to go through all the cpusets and add this CPU to
each cpuset which contained this CPU previously. In other words, add this
CPU to those cpusets, whose cpus_allowed_before_hotplug mask contains
this CPU.

Note that this function should not modify cpus_allowed_before_hotplug.
Hence, we move that particular piece of code from do_update_cpumask()
to update_cpumask().

Reported-by: Prashanth K. Nageshappa <prashanth@linux.vnet.ibm.com>
Signed-off-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Cc: stable@vger.kernel.org
---

 kernel/cpuset.c |   71 +++++++++++++++++++++++++++++++++++++++++++++++++++++--
 1 files changed, 69 insertions(+), 2 deletions(-)

diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index 2be71da..c56e705 100644
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -900,7 +900,6 @@ static int do_update_cpumask(struct cpuset *cs, struct cpuset *trialcs)
 
 	mutex_lock(&callback_mutex);
 	cpumask_copy(cs->cpus_allowed, trialcs->cpus_allowed);
-	cpumask_copy(cs->cpus_allowed_before_hotplug, cs->cpus_allowed);
 	mutex_unlock(&callback_mutex);
 
 	/*
@@ -948,7 +947,10 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs,
 			return -EINVAL;
 	}
 
-	return do_update_cpumask(cs, trialcs);
+	retval = do_update_cpumask(cs, trialcs);
+
+	cpumask_copy(cs->cpus_allowed_before_hotplug, cs->cpus_allowed);
+	return retval;
 }
 
 /*
@@ -2079,6 +2081,71 @@ static void remove_tasks_in_empty_cpuset(struct cpuset *cs)
 	move_member_tasks_to_cpuset(cs, parent);
 }
 
+
+/*
+ * Walk the specified cpuset subtree and add the newly onlined CPU to the
+ * cpuset, if that CPU was in this cpuset previously (IOW, if this CPU
+ * is still in the cpuset_allowed_before_hotplug mask of this cpuset).
+ *
+ * Called with cgroup_mutex held.
+ *
+ * This walk processes the tree from top to bottom, completing one layer
+ * before dropping down to the next.  It always processes a node before
+ * any of its children.
+ */
+static void scan_cpusets_during_cpu_online(struct cpuset *root)
+{
+	LIST_HEAD(queue);
+	struct cpuset *cp;	/* scans cpusets being updated */
+	struct cpuset *child;	/* scans child cpusets of cp */
+	struct cgroup *cont;
+	struct cpuset *trialcs;
+
+	trialcs = alloc_trial_cpuset(root);
+	if (!trialcs)
+		return;
+
+	list_add_tail((struct list_head *)&root->stack_list, &queue);
+
+	while (!list_empty(&queue)) {
+		cp = list_first_entry(&queue, struct cpuset, stack_list);
+		list_del(queue.next);
+		list_for_each_entry(cont, &cp->css.cgroup->children, sibling) {
+			child = cgroup_cs(cont);
+			list_add_tail(&child->stack_list, &queue);
+		}
+
+		/*
+		 * Continue past cpusets which didn't have this CPU previously
+		 * (Note that when this CPU went offline, it had been removed
+		 * from all cpusets.)
+		 */
+		if (cpumask_equal(cp->cpus_allowed,
+					cp->cpus_allowed_before_hotplug))
+			continue;
+
+		/*
+		 * Add newly onlined CPUs to this cpuset if they were part of
+		 * this cpuset previously.
+		 */
+		cpumask_and(trialcs->cpus_allowed,
+				cp->cpus_allowed_before_hotplug,
+				cpu_active_mask);
+
+		/* Update the cpuset */
+		do_update_cpumask(cp, trialcs);
+
+		/*
+		 * Do NOT update cpus_allowed_before_hotplug here. We want it
+		 * to reflect the value of cpus_allowed as seen *before* the
+		 * hotplug event.
+		 */
+	}
+
+	free_trial_cpuset(trialcs);
+}
+
+
 /*
  * Walk the specified cpuset subtree and look for empty cpusets.
  * The tasks of such cpuset must be moved to a parent cpuset.


  parent reply	other threads:[~2012-02-07 18:57 UTC|newest]

Thread overview: 32+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-02-07 18:55 [PATCH 0/4] CPU hotplug, cpusets: Fix CPU online handling related to cpusets Srivatsa S. Bhat
2012-02-07 18:56 ` [PATCH 1/4] CPU hotplug, cpuset: Maintain a copy of the cpus_allowed mask before CPU hotplug Srivatsa S. Bhat
2012-02-07 18:56 ` [PATCH 2/4] cpuset: Split up update_cpumask() so that its functionality can be reused Srivatsa S. Bhat
2012-02-07 18:57 ` Srivatsa S. Bhat [this message]
2012-02-07 18:57 ` [PATCH 4/4] CPU hotplug, cpusets: Differentiate the CPU online and CPU offline callbacks Srivatsa S. Bhat
2012-02-08  3:22 ` [PATCH 0/4] CPU hotplug, cpusets: Fix CPU online handling related to cpusets Peter Zijlstra
2012-02-08  6:33   ` Srivatsa S. Bhat
2012-02-09  7:57     ` Ingo Molnar
2012-02-09  8:42       ` Srivatsa S. Bhat
2012-02-09 15:11         ` Ingo Molnar
2012-02-10 15:52           ` Peter Zijlstra
2012-02-10 16:53             ` Paul E. McKenney
2012-02-10 17:34               ` Peter Zijlstra
2012-02-10 21:51                 ` Alan Stern
2012-02-10 22:39                   ` Rafael J. Wysocki
2012-02-11  2:07                     ` Peter Zijlstra
2012-02-11  4:26                       ` Srivatsa S. Bhat
2012-02-13 17:47                         ` Srivatsa S. Bhat
2012-02-17 12:15                           ` Srivatsa S. Bhat
2012-02-20 12:49                             ` Peter Zijlstra
2012-02-20 12:59                               ` Srivatsa S. Bhat
2012-02-23  9:57                                 ` Srivatsa S. Bhat
2012-02-24 23:24                                   ` Rafael J. Wysocki
2012-02-27 10:18                                   ` Peter Zijlstra
2012-02-27 12:09                                   ` [tip:sched/urgent] CPU hotplug, cpusets, suspend: Don' t touch cpusets during suspend/resume tip-bot for Srivatsa S. Bhat
2012-02-11 16:00                 ` [PATCH 0/4] CPU hotplug, cpusets: Fix CPU online handling related to cpusets Paul E. McKenney
2012-02-13 17:47               ` Srivatsa S. Bhat
2012-02-13 20:39                 ` Paul E. McKenney
2012-02-13 20:49                   ` Srivatsa S. Bhat
2012-02-11 13:39             ` Ingo Molnar
2012-02-10 15:53         ` Peter Zijlstra
2012-02-09 16:43     ` Peter Zijlstra

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=20120207185651.7482.18642.stgit@srivatsabhat.in.ibm.com \
    --to=srivatsa.bhat@linux.vnet.ibm.com \
    --cc=a.p.zijlstra@chello.nl \
    --cc=frank.rowand@am.sony.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pm@vger.kernel.org \
    --cc=lizf@cn.fujitsu.com \
    --cc=mingo@elte.hu \
    --cc=paul@paulmenage.org \
    --cc=paulmck@linux.vnet.ibm.com \
    --cc=pjt@google.com \
    --cc=prashanth@linux.vnet.ibm.com \
    --cc=rjw@sisk.pl \
    --cc=tglx@linutronix.de \
    --cc=tj@kernel.org \
    --cc=vatsa@linux.vnet.ibm.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;
as well as URLs for NNTP newsgroup(s).