cpuset: consider dying css as offline

In most cases, a cgroup controller don't care about the liftimes of
cgroups.  For the controller, a css becomes online when ->css_online()
is called on it and offline when ->css_offline() is called.

However, cpuset is special in that the user interface it exposes cares
whether certain cgroups exist or not.  Combined with the RCU delay
between cgroup removal and css offlining, this can lead to user
visible behavior oddities where operations which should succeed after
cgroup removals fail for some time period.  The effects of cgroup
removals are delayed when seen from userland.

This patch adds css_is_dying() which tests whether offline is pending
and updates is_cpuset_online() so that the function returns false also
while offline is pending.  This gets rid of the userland visible
delays.

Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Daniel Jordan <daniel.m.jordan@oracle.com>
Link:
http://lkml.kernel.org/r/327ca1f5-7957-fbb9-9e5f-9ba149d40ba2@oracle.com
Cc: stable@vger.kernel.org
Signed-off-by: Tejun Heo <tj@kernel.org>
Orabug: 26415290

(backport upstream commit 41c25707d21716826e3c1f60967f5550610ec1c9)

Signed-off-by: Thomas Tai <thomas.tai@oracle.com>
Signed-off-by: Tom Hromatka <tom.hromatka@oracle.com>
Signed-off-by: Allen Pais <allen.pais@oracle.com>
Reviewed-by: Tom Hromatka <tom.hromatka@oracle.com>
Reviewed-by: Dhaval Giani <dhaval.giani@oracle.com>

diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h
index b9cb94c3102a402539368cee3016d81b1cc173f9..b4a564df5a1a3961abf58d1ff810fc2f1ddd89bc 100644 (file)
--- a/include/linux/cgroup.h
+++ b/include/linux/cgroup.h
@@ -160,6 +160,26 @@ static inline bool css_tryget_online(struct cgroup_subsys_state *css)
        return true;
 }
 
+/*
+ * css_is_dying - test whether the specified css is dying
+ * @css: target css
+ *
+ * Test whether @css is in the process of offlining or already offline.  In
+ * most cases, ->css_online() and ->css_offline() callbacks should be
+ * enough; however, the actual offline operations are RCU delayed and this
+ * test returns %true also when @css is scheduled to be offlined.
+ *
+ * This is useful, for example, when the use case requires synchronous
+ * behavior with respect to cgroup removal.  cgroup removal schedules css
+ * offlining but the css can seem alive while the operation is being
+ * delayed.  If the delay affects user visible semantics, this test can be
+ * used to resolve the situation.
+  */
+static inline bool css_is_dying(struct cgroup_subsys_state *css)
+{
+       return !(css->flags & CSS_NO_REF) && percpu_ref_is_dying(&css->refcnt);
+}
+
 /**
  * css_put - put a css reference
  * @css: target css

diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index f0acff0f66c91380412dcbc1c899c94b1d3236b0..f193c46f216d7e6c19fef8c8e3a026e45260c734 100644 (file)
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -176,7 +176,7 @@ typedef enum {
 /* convenient tests for these bits */
 static inline bool is_cpuset_online(const struct cpuset *cs)
 {
-       return test_bit(CS_ONLINE, &cs->flags);
+       return test_bit(CS_ONLINE, &cs->flags) && !css_is_dying(&cs->css);
 }
 
 static inline int is_cpu_exclusive(const struct cpuset *cs)