[PATCH] cgroup/cpuset: Fix wrong check in update_parent_subparts_cpumask()

[PATCH] cgroup/cpuset: Fix wrong check in update_parent_subparts_cpumask()

[Waiman Long]

It was found that the check to see if a partition could use up all
the cpus from the parent cpuset in update_parent_subparts_cpumask()
was incorrect. As a result, it is possible to leave parent with no
effective cpu left even if there are tasks in the parent cpuset. This
can lead to system panic as reported in [1].

Fix this probem by updating the check to fail the enabling the partition
if parent's effective_cpus is a subset of the child's cpus_allowed.

Also record the error code when an error happens in update_prstate()
and add a test case where parent partition and child have the same cpu
list and parent has task. Enabling partition in the child will fail in
this case.

[1] https://www.spinics.net/lists/cgroups/msg36254.html

Fixes: f0af1bfc27b5 ("cgroup/cpuset: Relax constraints to partition & cpus changes")
Reported-by: Srinivas Pandruvada <srinivas.pandruvada-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Signed-off-by: Waiman Long <longman-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
---
 kernel/cgroup/cpuset.c                            | 3 ++-
 tools/testing/selftests/cgroup/test_cpuset_prs.sh | 1 +
 2 files changed, 3 insertions(+), 1 deletion(-)

diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
index a29c0b13706b..205dc9edcaa9 100644
--- a/kernel/cgroup/cpuset.c
+++ b/kernel/cgroup/cpuset.c
@@ -1346,7 +1346,7 @@ static int update_parent_subparts_cpumask(struct cpuset *cs, int cmd,
 		 * A parent can be left with no CPU as long as there is no
 		 * task directly associated with the parent partition.
 		 */
-		if (!cpumask_intersects(cs->cpus_allowed, parent->effective_cpus) &&
+		if (cpumask_subset(parent->effective_cpus, cs->cpus_allowed) &&
 		    partition_is_populated(parent, cs))
 			return PERR_NOCPUS;
 
@@ -2324,6 +2324,7 @@ static int update_prstate(struct cpuset *cs, int new_prs)
 		new_prs = -new_prs;
 	spin_lock_irq(&callback_lock);
 	cs->partition_root_state = new_prs;
+	WRITE_ONCE(cs->prs_err, err);
 	spin_unlock_irq(&callback_lock);
 	/*
 	 * Update child cpusets, if present.
diff --git a/tools/testing/selftests/cgroup/test_cpuset_prs.sh b/tools/testing/selftests/cgroup/test_cpuset_prs.sh
index 186e1c26867e..75c100de90ff 100755
--- a/tools/testing/selftests/cgroup/test_cpuset_prs.sh
+++ b/tools/testing/selftests/cgroup/test_cpuset_prs.sh
@@ -268,6 +268,7 @@ TEST_MATRIX=(
 	# Taking away all CPUs from parent or itself if there are tasks
 	# will make the partition invalid.
 	"  S+ C2-3:P1:S+  C3:P1  .      .      T     C2-3    .      .     0 A1:2-3,A2:2-3 A1:P1,A2:P-1"
+	"  S+  C3:P1:S+    C3    .      .      T      P1     .      .     0 A1:3,A2:3 A1:P1,A2:P-1"
 	"  S+ $SETUP_A123_PARTITIONS    .    T:C2-3   .      .      .     0 A1:2-3,A2:2-3,A3:3 A1:P1,A2:P-1,A3:P-1"
 	"  S+ $SETUP_A123_PARTITIONS    . T:C2-3:C1-3 .      .      .     0 A1:1,A2:2,A3:3 A1:P1,A2:P1,A3:P1"
 
-- 
2.31.1

Re: [PATCH] cgroup/cpuset: Fix wrong check in update_parent_subparts_cpumask()

[Tejun Heo]

On Tue, Jan 31, 2023 at 10:48:03AM -0500, Waiman Long wrote:
> It was found that the check to see if a partition could use up all
> the cpus from the parent cpuset in update_parent_subparts_cpumask()
> was incorrect. As a result, it is possible to leave parent with no
> effective cpu left even if there are tasks in the parent cpuset. This
> can lead to system panic as reported in [1].
> 
> Fix this probem by updating the check to fail the enabling the partition
> if parent's effective_cpus is a subset of the child's cpus_allowed.
> 
> Also record the error code when an error happens in update_prstate()
> and add a test case where parent partition and child have the same cpu
> list and parent has task. Enabling partition in the child will fail in
> this case.
> 
> [1] https://www.spinics.net/lists/cgroups/msg36254.html
> 
> Fixes: f0af1bfc27b5 ("cgroup/cpuset: Relax constraints to partition & cpus changes")
> Reported-by: Srinivas Pandruvada <srinivas.pandruvada-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
> Signed-off-by: Waiman Long <longman-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>

Applied to cgroup/for-6.2-fixes w/ stable cc added.

Thanks.

-- 
tejun