6.18-stable review patch. If anyone has any objections, please let me know.
------------------
From: Chen Ridong chenridong@huawei.com
[ Upstream commit b1bcaed1e39a9e0dfbe324a15d2ca4253deda316 ]
Currently, the check for whether a partition is populated does not account for tasks in the cpuset of attaching. This is a corner case that can leave a task stuck in a partition with no effective CPUs.
The race condition occurs as follows:
cpu0 cpu1 //cpuset A with cpu N migrate task p to A cpuset_can_attach // with effective cpus // check ok
// cpuset_mutex is not held // clear cpuset.cpus.exclusive // making effective cpus empty update_exclusive_cpumask // tasks_nocpu_error check ok // empty effective cpus, partition valid cpuset_attach ... // task p stays in A, with non-effective cpus.
To fix this issue, this patch introduces cs_is_populated, which considers tasks in the attaching cpuset. This new helper is used in validate_change and partition_is_populated.
Fixes: e2d59900d936 ("cgroup/cpuset: Allow no-task partition to have empty cpuset.cpus.effective") Signed-off-by: Chen Ridong chenridong@huawei.com Reviewed-by: Waiman Long longman@redhat.com Signed-off-by: Tejun Heo tj@kernel.org Signed-off-by: Sasha Levin sashal@kernel.org --- kernel/cgroup/cpuset.c | 35 +++++++++++++++++++++++++++-------- 1 file changed, 27 insertions(+), 8 deletions(-)
diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c index 52468d2c178a3..4dcd633fd6df5 100644 --- a/kernel/cgroup/cpuset.c +++ b/kernel/cgroup/cpuset.c @@ -352,6 +352,15 @@ static inline bool is_in_v2_mode(void) (cpuset_cgrp_subsys.root->flags & CGRP_ROOT_CPUSET_V2_MODE); }
+static inline bool cpuset_is_populated(struct cpuset *cs) +{ + lockdep_assert_held(&cpuset_mutex); + + /* Cpusets in the process of attaching should be considered as populated */ + return cgroup_is_populated(cs->css.cgroup) || + cs->attach_in_progress; +} + /** * partition_is_populated - check if partition has tasks * @cs: partition root to be checked @@ -364,21 +373,31 @@ static inline bool is_in_v2_mode(void) static inline bool partition_is_populated(struct cpuset *cs, struct cpuset *excluded_child) { - struct cgroup_subsys_state *css; - struct cpuset *child; + struct cpuset *cp; + struct cgroup_subsys_state *pos_css;
- if (cs->css.cgroup->nr_populated_csets) + /* + * We cannot call cs_is_populated(cs) directly, as + * nr_populated_domain_children may include populated + * csets from descendants that are partitions. + */ + if (cs->css.cgroup->nr_populated_csets || + cs->attach_in_progress) return true; if (!excluded_child && !cs->nr_subparts) return cgroup_is_populated(cs->css.cgroup);
rcu_read_lock(); - cpuset_for_each_child(child, css, cs) { - if (child == excluded_child) + cpuset_for_each_descendant_pre(cp, pos_css, cs) { + if (cp == cs || cp == excluded_child) continue; - if (is_partition_valid(child)) + + if (is_partition_valid(cp)) { + pos_css = css_rightmost_descendant(pos_css); continue; - if (cgroup_is_populated(child->css.cgroup)) { + } + + if (cpuset_is_populated(cp)) { rcu_read_unlock(); return true; } @@ -663,7 +682,7 @@ static int validate_change(struct cpuset *cur, struct cpuset *trial) * be changed to have empty cpus_allowed or mems_allowed. */ ret = -ENOSPC; - if ((cgroup_is_populated(cur->css.cgroup) || cur->attach_in_progress)) { + if (cpuset_is_populated(cur)) { if (!cpumask_empty(cur->cpus_allowed) && cpumask_empty(trial->cpus_allowed)) goto out;