We're going to have separate user-configured masks and effective ones.
Eventually configured masks can only be changed by writing cpuset.cpus
and cpuset.mems, and they won't be restricted by parent cpuset. While
effective masks reflect cpu/memory hotplug and hierachical restriction,
and these are the real masks that apply to the tasks in the cpuset.
We calculate effective mask this way:
- top cpuset's effective_mask == online_mask, otherwise
- cpuset's effective_mask == configured_mask & parent effective_mask,
if the result is empty, it inherits parent effective mask.
Those behavior changes are for sane_behavior only. For !sane_behavior
effective_mask and configured_mask are the same, so we won't break old
interfaces.
To make cs->effective_{cpus,mems} to be effective masks, we need to
- change the effective masks at hotplug
- change the effective masks at config change
- take on ancestor's mask when the effective mask is empty
This won't introduce behavior change.
v2:
- Add comments to explain effective masks are the same with configured
masks for !sane_behavior.
Signed-off-by: Li Zefan <lizefan@xxxxxxxxxx>
---
kernel/cpuset.c | 48 ++++++++++++++++++++++++++++++++++--------------
1 file changed, 34 insertions(+), 14 deletions(-)
diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index bdc6047..6723b88 100644
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -899,16 +899,22 @@ static void update_cpumasks_hier(struct cpuset *cs, struct cpuset *trialcs,
struct cpuset *parent = parent_cs(cp);
struct cpumask *new_cpus = trialcs->effective_cpus;
- cpumask_and(new_cpus, cp->cpus_allowed,
- parent->effective_cpus);
+ /*
+ * If !sane_behavior, new_cpus will equals cpus_allowed,
+ * which is not empty, so it's guaranteed the effective mask
+ * is the same with the configured mask.
+ */
+ cpumask_and(new_cpus, cp->cpus_allowed, parent->effective_cpus);
/*
- * Skip the whole subtree if the cpumask remains the same
- * and isn't empty. If it's empty, we need to update tasks
- * to take on an ancestor's cpumask.
+ * If it becomes empty, inherit the effective mask of the
+ * parent, which is guaranteed to have some CPUs.
*/
- if (cpumask_equal(new_cpus, cp->effective_cpus) &&
- ((cp == cs) || !cpumask_empty(new_cpus))) {
+ if (cpumask_empty(new_cpus))
+ cpumask_copy(new_cpus, parent->effective_cpus);
+
+ /* Skip the whole subtree if the cpumask remains the same. */
+ if (cpumask_equal(new_cpus, cp->effective_cpus)) {
pos_css = css_rightmost_descendant(pos_css);
continue;
}
@@ -1163,16 +1169,22 @@ static void update_nodemasks_hier(struct cpuset *cs, struct cpuset *trialcs,
struct cpuset *parent = parent_cs(cp);
nodemask_t *new_mems = &trialcs->effective_mems;
- nodes_and(*new_mems, cp->mems_allowed,
- parent->effective_mems);
+ /*
+ * If !sane_behavior, new_mems will equal mems_allowed,
+ * which is not empty, so it's guaranteed the effective mask
+ * is the same with the configured mask.
+ */
+ nodes_and(*new_mems, cp->mems_allowed, parent->effective_mems);
/*
- * Skip the whole subtree if the nodemask remains the same
- * and isn't empty. If it's empty, we need to update tasks
- * to take on an ancestor's nodemask.
+ * If it becomes empty, inherit the effective mask of the
+ * parent, which is guaranteed to have some MEMs.
*/
- if (nodes_equal(*new_mems, cp->effective_mems) &&
- ((cp == cs) || !nodes_empty(*new_mems))) {
+ if (nodes_empty(*new_mems))
+ *new_mems = parent->effective_mems;
+
+ /* Skip the whole subtree if the nodemask is not changed. */
+ if (nodes_equal(*new_mems, cp->effective_mems)) {
pos_css = css_rightmost_descendant(pos_css);
continue;
}
@@ -2203,7 +2215,11 @@ retry:
mutex_lock(&callback_mutex);
cpumask_andnot(cs->cpus_allowed, cs->cpus_allowed, &off_cpus);
+
+ /* Inherit the effective mask of the parent, if it becomes empty. */
cpumask_andnot(cs->effective_cpus, cs->effective_cpus, &off_cpus);
+ if (sane && cpumask_empty(cs->effective_cpus))
+ cpumask_copy(cs->effective_cpus, parent_cs(cs)->effective_cpus);
mutex_unlock(&callback_mutex);
/*
@@ -2218,7 +2234,11 @@ retry:
mutex_lock(&callback_mutex);
nodes_andnot(cs->mems_allowed, cs->mems_allowed, off_mems);
+
+ /* Inherit the effective mask of the parent, if it becomes empty */
nodes_andnot(cs->effective_mems, cs->effective_mems, off_mems);
+ if (sane && nodes_empty(cs->effective_mems))
+ cs->effective_mems = parent_cs(cs)->effective_mems;
mutex_unlock(&callback_mutex);
/*
--
1.8.0.2
--
To unsubscribe from this list: send the line "unsubscribe cgroups" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
