From: Li Zefan <lizf@cn.fujitsu.com>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: Paul Jackson <pj@sgi.com>, Paul Menage <menage@google.com>,
Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>,
Lai Jiangshan <laijs@cn.fujitsu.com>,
LKML <linux-kernel@vger.kernel.org>
Subject: [PATCH 2/3] cpuset: speed up sched domain partition
Date: Mon, 28 Jul 2008 10:47:09 +0800 [thread overview]
Message-ID: <488D332D.3000501@cn.fujitsu.com> (raw)
From: Lai Jiangshan <laijs@cn.fujitsu.com>
Date: Mon, 28 Jul 2008 10:38:22 +0800
Subject: [PATCH 2/3] cpuset: speed up sched domain partition
All child cpusets contain a subset of the parent's cpus, so we can skip
them when partitioning sched domains. This decreases 'csa' greately for
cpusets with multi-level hierarchy.
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
kernel/cpuset.c | 41 +++++++++++++++++++++++++++++++++++++----
1 files changed, 37 insertions(+), 4 deletions(-)
diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index d0c57ac..d274a94 100644
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -485,13 +485,38 @@ static int cpusets_overlap(struct cpuset *a, struct cpuset *b)
static void
update_domain_attr(struct sched_domain_attr *dattr, struct cpuset *c)
{
- if (!dattr)
- return;
if (dattr->relax_domain_level < c->relax_domain_level)
dattr->relax_domain_level = c->relax_domain_level;
return;
}
+static void
+update_domain_attr_tree(struct sched_domain_attr *dattr, struct cpuset *c)
+{
+ LIST_HEAD(q);
+
+ list_add(&c->stack_list, &q);
+ while (!list_empty(&q)) {
+ struct cpuset *cp;
+ struct cgroup *cont;
+ struct cpuset *child;
+
+ cp = list_first_entry(&q, struct cpuset, stack_list);
+ list_del(q.next);
+
+ if (cpus_empty(cp->cpus_allowed))
+ continue;
+
+ if (is_sched_load_balance(cp))
+ update_domain_attr(dattr, cp);
+
+ list_for_each_entry(cont, &cp->css.cgroup->children, sibling) {
+ child = cgroup_cs(cont);
+ list_add_tail(&child->stack_list, &q);
+ }
+ }
+}
+
/*
* rebuild_sched_domains()
*
@@ -611,8 +636,16 @@ void rebuild_sched_domains(void)
if (cpus_empty(cp->cpus_allowed))
continue;
- if (is_sched_load_balance(cp))
+ /*
+ * All child cpusets contain a subset of the parent's cpus, so
+ * just skip them, and then we call update_domain_attr_tree()
+ * to calc relax_domain_level of the corresponding sched
+ * domain.
+ */
+ if (is_sched_load_balance(cp)) {
csa[csn++] = cp;
+ continue;
+ }
list_for_each_entry(cont, &cp->css.cgroup->children, sibling) {
child = cgroup_cs(cont);
@@ -683,7 +716,7 @@ restart:
cpus_or(*dp, *dp, b->cpus_allowed);
b->pn = -1;
if (dattr)
- update_domain_attr(dattr
+ update_domain_attr_tree(dattr
+ nslot, b);
}
}
--
1.5.4.rc3
next reply other threads:[~2008-07-28 2:53 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-07-28 2:47 Li Zefan [this message]
2008-07-28 5:16 ` [PATCH 2/3] cpuset: speed up sched domain partition Paul Jackson
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=488D332D.3000501@cn.fujitsu.com \
--to=lizf@cn.fujitsu.com \
--cc=akpm@linux-foundation.org \
--cc=laijs@cn.fujitsu.com \
--cc=linux-kernel@vger.kernel.org \
--cc=menage@google.com \
--cc=pj@sgi.com \
--cc=seto.hidetoshi@jp.fujitsu.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