From: Waiman Long <longman@redhat.com>
To: Tejun Heo <tj@kernel.org>, Li Zefan <lizefan@huawei.com>,
Johannes Weiner <hannes@cmpxchg.org>,
Peter Zijlstra <peterz@infradead.org>,
Ingo Molnar <mingo@redhat.com>
Cc: cgroups@vger.kernel.org, linux-kernel@vger.kernel.org,
linux-doc@vger.kernel.org, kernel-team@fb.com, pjt@google.com,
luto@amacapital.net, Mike Galbraith <efault@gmx.de>,
torvalds@linux-foundation.org, Roman Gushchin <guro@fb.com>,
Juri Lelli <juri.lelli@redhat.com>,
Patrick Bellasi <patrick.bellasi@arm.com>,
Waiman Long <longman@redhat.com>
Subject: [PATCH v13 06/11] cpuset: Track cpusets that use parent's effective_cpus
Date: Fri, 12 Oct 2018 13:55:46 -0400 [thread overview]
Message-ID: <1539366951-8498-7-git-send-email-longman@redhat.com> (raw)
In-Reply-To: <1539366951-8498-1-git-send-email-longman@redhat.com>
In the default hierarchy, a cpuset will use the parent's effective_cpus
if none of the requested CPUs can be granted from the parent. That can
be a problem if a parent is a partition root with children partition
roots. Changes to a parent's effective_cpus list due to changes in a
child partition root may not be properly reflected in a child cpuset
that use parent's effective_cpus because the cpu_exclusive rule of a
partition root will not guard against that.
In order to avoid the mismatch, two new tracking variables are added to
the cpuset structure to track if a cpuset uses parent's effective_cpus
and the number of children cpusets that use its effective_cpus. So
whenever cpumask changes are made to a parent, it will also check to
see if it has other children cpusets that use its effective_cpus and
call update_cpumasks_hier() if that is the case.
Signed-off-by: Waiman Long <longman@redhat.com>
---
kernel/cgroup/cpuset.c | 71 +++++++++++++++++++++++++++++++++++++++++++++++++-
1 file changed, 70 insertions(+), 1 deletion(-)
diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
index 12ec61d..463c583 100644
--- a/kernel/cgroup/cpuset.c
+++ b/kernel/cgroup/cpuset.c
@@ -147,6 +147,14 @@ struct cpuset {
/* partition root state */
int partition_root_state;
+
+ /*
+ * Default hierarchy only:
+ * use_parent_ecpus - set if using parent's effective_cpus
+ * child_ecpus_count - # of children with use_parent_ecpus set
+ */
+ int use_parent_ecpus;
+ int child_ecpus_count;
};
/*
@@ -1221,8 +1229,17 @@ static void update_cpumasks_hier(struct cpuset *cs, struct tmpmasks *tmp)
* If it becomes empty, inherit the effective mask of the
* parent, which is guaranteed to have some CPUs.
*/
- if (is_in_v2_mode() && cpumask_empty(tmp->new_cpus))
+ if (is_in_v2_mode() && cpumask_empty(tmp->new_cpus)) {
cpumask_copy(tmp->new_cpus, parent->effective_cpus);
+ if (!cp->use_parent_ecpus) {
+ cp->use_parent_ecpus = true;
+ parent->child_ecpus_count++;
+ }
+ } else if (cp->use_parent_ecpus) {
+ cp->use_parent_ecpus = false;
+ WARN_ON_ONCE(!parent->child_ecpus_count);
+ parent->child_ecpus_count--;
+ }
/*
* Skip the whole subtree if the cpumask remains the same
@@ -1335,6 +1352,35 @@ static void update_cpumasks_hier(struct cpuset *cs, struct tmpmasks *tmp)
}
/**
+ * update_sibling_cpumasks - Update siblings cpumasks
+ * @parent: Parent cpuset
+ * @cs: Current cpuset
+ * @tmp: Temp variables
+ */
+static void update_sibling_cpumasks(struct cpuset *parent, struct cpuset *cs,
+ struct tmpmasks *tmp)
+{
+ struct cpuset *sibling;
+ struct cgroup_subsys_state *pos_css;
+
+ /*
+ * Check all its siblings and call update_cpumasks_hier()
+ * if their use_parent_ecpus flag is set in order for them
+ * to use the right effective_cpus value.
+ */
+ rcu_read_lock();
+ cpuset_for_each_child(sibling, pos_css, parent) {
+ if (sibling == cs)
+ continue;
+ if (!sibling->use_parent_ecpus)
+ continue;
+
+ update_cpumasks_hier(sibling, tmp);
+ }
+ rcu_read_unlock();
+}
+
+/**
* update_cpumask - update the cpus_allowed mask of a cpuset and all tasks in it
* @cs: the cpuset to consider
* @trialcs: trial cpuset
@@ -1406,6 +1452,17 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs,
spin_unlock_irq(&callback_lock);
update_cpumasks_hier(cs, &tmp);
+
+ if (cs->partition_root_state) {
+ struct cpuset *parent = parent_cs(cs);
+
+ /*
+ * For partition root, update the cpumasks of sibling
+ * cpusets if they use parent's effective_cpus.
+ */
+ if (parent->child_ecpus_count)
+ update_sibling_cpumasks(parent, cs, &tmp);
+ }
return 0;
}
@@ -1842,6 +1899,9 @@ static int update_prstate(struct cpuset *cs, int val)
if (parent != &top_cpuset)
update_tasks_cpumask(parent);
+ if (parent->child_ecpus_count)
+ update_sibling_cpumasks(parent, cs, &tmp);
+
rebuild_sched_domains_locked();
out:
free_cpumasks(&tmp);
@@ -2536,6 +2596,8 @@ static int cpuset_css_online(struct cgroup_subsys_state *css)
if (is_in_v2_mode()) {
cpumask_copy(cs->effective_cpus, parent->effective_cpus);
cs->effective_mems = parent->effective_mems;
+ cs->use_parent_ecpus = true;
+ parent->child_ecpus_count++;
}
spin_unlock_irq(&callback_lock);
@@ -2599,6 +2661,13 @@ static void cpuset_css_offline(struct cgroup_subsys_state *css)
is_sched_load_balance(cs))
update_flag(CS_SCHED_LOAD_BALANCE, cs, 0);
+ if (cs->use_parent_ecpus) {
+ struct cpuset *parent = parent_cs(cs);
+
+ cs->use_parent_ecpus = false;
+ parent->child_ecpus_count--;
+ }
+
cpuset_dec();
clear_bit(CS_ONLINE, &cs->flags);
--
1.8.3.1
next prev parent reply other threads:[~2018-10-12 17:56 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-10-12 17:55 [PATCH v13 00/11] cpuset: Enable cpuset controller in default hierarchy Waiman Long
2018-10-12 17:55 ` [PATCH v13 01/11] " Waiman Long
2018-10-15 18:31 ` Tom Hromatka
2018-10-15 18:41 ` Waiman Long
2018-10-15 19:10 ` Tom Hromatka
2018-10-12 17:55 ` [PATCH v13 02/11] cpuset: Define data structures to support scheduling partition Waiman Long
2018-10-12 17:55 ` [PATCH v13 03/11] cpuset: Simply allocation and freeing of cpumasks Waiman Long
2018-10-15 18:35 ` Tom Hromatka
2018-10-15 18:43 ` Waiman Long
2018-10-12 17:55 ` [PATCH v13 04/11] cpuset: Add new v2 cpuset.sched.partition flag Waiman Long
2018-10-13 14:10 ` kbuild test robot
2018-10-12 17:55 ` [PATCH v13 05/11] cpuset: Add an error state to cpuset.sched.partition Waiman Long
2018-10-12 17:55 ` Waiman Long [this message]
2018-10-12 17:55 ` [PATCH v13 07/11] cpuset: Make CPU hotplug work with partition Waiman Long
2018-10-12 17:55 ` [PATCH v13 08/11] cpuset: Make generate_sched_domains() " Waiman Long
2018-10-12 17:55 ` [PATCH v13 09/11] cpuset: Expose cpus.effective and mems.effective on cgroup v2 root Waiman Long
2018-10-12 17:55 ` [PATCH v13 10/11] cpuset: Add documentation about the new "cpuset.sched.partition" flag Waiman Long
2018-10-12 17:55 ` [PATCH v13 11/11] cpuset: Expose cpuset.cpus.subpartitions with cgroup_debug Waiman Long
2018-10-15 16:35 ` [PATCH v13 00/11] cpuset: Enable cpuset controller in default hierarchy Tejun Heo
2018-10-15 17:04 ` Waiman Long
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=1539366951-8498-7-git-send-email-longman@redhat.com \
--to=longman@redhat.com \
--cc=cgroups@vger.kernel.org \
--cc=efault@gmx.de \
--cc=guro@fb.com \
--cc=hannes@cmpxchg.org \
--cc=juri.lelli@redhat.com \
--cc=kernel-team@fb.com \
--cc=linux-doc@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=lizefan@huawei.com \
--cc=luto@amacapital.net \
--cc=mingo@redhat.com \
--cc=patrick.bellasi@arm.com \
--cc=peterz@infradead.org \
--cc=pjt@google.com \
--cc=tj@kernel.org \
--cc=torvalds@linux-foundation.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).