Skip to contentSkip to conversation

cpuset: Allow setscheduler regardless of manipulated task

PATCH Superseded Michal Koutný 2 participants 5 messages Jun 30, 2023

Subsystems (3)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore

Michal Koutný

author maintainer
#0 lore raw

v2 cover cpuset: Allow setscheduler regardless of manipulated task

Changes in v2:
- rebased on mainline
- drop is_in_v2_mode()

Changes in v1 (https://lore.kernel.org/r/20230629091146.28801-1-mkoutny-IBi9RG/[email protected]):
- added selftests
- comments rewording

RFC in https://lore.kernel.org/r/20220623124944.2753-1-mkoutny-IBi9RG/[email protected]

Michal Koutný (3):
cpuset: Allow setscheduler regardless of manipulated task
selftests: cgroup: Minor code reorganizations
selftests: cgroup: Add cpuset migrations testcase

 MAINTAINERS                                   |   2 +
 kernel/cgroup/cpuset.c                        |  13 +-
 tools/testing/selftests/cgroup/.gitignore     |   1 +
 tools/testing/selftests/cgroup/Makefile       |   2 +
 tools/testing/selftests/cgroup/cgroup_util.c  |   2 +
 tools/testing/selftests/cgroup/cgroup_util.h  |   2 +
 tools/testing/selftests/cgroup/test_core.c    |   2 +-
 tools/testing/selftests/cgroup/test_cpuset.c  | 272 ++++++++++++++++++
 .../selftests/cgroup/test_cpuset_prs.sh       |   2 +-
 9 files changed, 293 insertions(+), 5 deletions(-)
 create mode 100644 tools/testing/selftests/cgroup/test_cpuset.c

base-commit: e55e5df193d247a38a5e1ac65a5316a0adcc22fa

Signature · 1 line
2.41.0
author maintainer
#1 lore raw

PATCH v2 1/3 cpuset: Allow setscheduler regardless of manipulated task

When we migrate a task between two cgroups, one of the checks is a verification whether we can modify task's scheduler settings (cap_task_setscheduler()).

An implicit migration occurs also when enabling a controller on the unified hierarchy (think of parent to child migration). The aforementioned check may be problematic if the caller of the migration (enabling a controller) has no permissions over migrated tasks. For instance, a user's cgroup that ends up running a process of a different user. Although cgroup permissions are configured favorably, the enablement fails due to the foreign process [1].

Change the behavior by relaxing the permissions check on the unified hierarchy (or in v2 mode). This is in accordance with unified hierarchy attachment behavior when permissions of the source to target cgroups are decisive whereas the migrated task is opaque (as opposed to more restrictive check in __cgroup1_procs_write()).

[1] https://github.com/systemd/systemd/issues/18293#issuecomment-831205649

---
 kernel/cgroup/cpuset.c | 13 ++++++++++---
 1 file changed, 10 insertions(+), 3 deletions(-)
kernel/cgroup/cpuset.c +10 -3
index 58e6f18f01c1..41d3ed14b0f4 100644
@@ -2505,9 +2505,16 @@ static int cpuset_can_attach(struct cgroup_taskset *tset)
 		ret = task_can_attach(task);
 		if (ret)
 			goto out_unlock;
-		ret = security_task_setscheduler(task);
-		if (ret)
-			goto out_unlock;
+
+		/*
+		 * Skip rights over task check in v2, migration permission derives
+		 * from hierarchy ownership in cgroup_procs_write_permission()).
+		 */
+		if (!cgroup_subsys_on_dfl(cpuset_cgrp_subsys)) {
+			ret = security_task_setscheduler(task);
+			if (ret)
+				goto out_unlock;
+		}
 
 		if (dl_task(task)) {
 			cs->nr_migrate_dl_tasks++;
Signature · 1 line
2.41.0
maintainer
#2 lore raw
Quoted 36 lines · Michal Koutný

When we migrate a task between two cgroups, one of the checks is a verification whether we can modify task's scheduler settings (cap_task_setscheduler()).

An implicit migration occurs also when enabling a controller on the unified hierarchy (think of parent to child migration). The aforementioned check may be problematic if the caller of the migration (enabling a controller) has no permissions over migrated tasks. For instance, a user's cgroup that ends up running a process of a different user. Although cgroup permissions are configured favorably, the enablement fails due to the foreign process [1].

Change the behavior by relaxing the permissions check on the unified hierarchy (or in v2 mode). This is in accordance with unified hierarchy attachment behavior when permissions of the source to target cgroups are decisive whereas the migrated task is opaque (as opposed to more restrictive check in __cgroup1_procs_write()).

[1] https://github.com/systemd/systemd/issues/18293#issuecomment-831205649

---
  kernel/cgroup/cpuset.c | 13 ++++++++++---
  1 file changed, 10 insertions(+), 3 deletions(-)
kernel/cgroup/cpuset.c +0 -3
index 58e6f18f01c1..41d3ed14b0f4 100644
@@ -2505,9 +2505,16 @@ static int cpuset_can_attach(struct cgroup_taskset *tset)
  		ret = task_can_attach(task);
  		if (ret)
  			goto out_unlock;
-		ret = security_task_setscheduler(task);
-		if (ret)
-			goto out_unlock;

kernel/cgroup/cpuset.c static int cpuset_can_attach(struct cgroup_taskset *tset) · line 2517

Quoted 10 lines
kernel/cgroup/cpuset.c +10 -0
+
+		/*
+		 * Skip rights over task check in v2, migration permission derives
+		 * from hierarchy ownership in cgroup_procs_write_permission()).
+		 */
+		if (!cgroup_subsys_on_dfl(cpuset_cgrp_subsys)) {
+			ret = security_task_setscheduler(task);
+			if (ret)
+				goto out_unlock;
+		}

I am somewhat hesitant to skip the security_task_setscheduler() check for all cgroup v2 task migrations. The check is controlled by SElinux which is a different subsystem. I believe the scheduler property here refer's to the task cpu affinity and node mask. If you look at cpuset_attach(), we have actually skipped the task iteration process to change them if cpu affinity and node mask aren't changed at all.

I don't want to introduce a possible security vulnerability because of this relaxation. I would suggest you skip it under the same condition of no change to cpu affinity and node mask for cgroup v2.

Thanks,
Longman

author maintainer
#3 lore raw

PATCH v2 2/3 selftests: cgroup: Minor code reorganizations

In reply to Michal Koutný

No functional change intended, these small changes are merged into one commit and they serve as a preparation for an upcoming new testcase.

---
 MAINTAINERS                                       | 1 +
 tools/testing/selftests/cgroup/cgroup_util.c      | 2 ++
 tools/testing/selftests/cgroup/cgroup_util.h      | 2 ++
 tools/testing/selftests/cgroup/test_core.c        | 2 +-
 tools/testing/selftests/cgroup/test_cpuset_prs.sh | 2 +-
 5 files changed, 7 insertions(+), 2 deletions(-)
MAINTAINERS +1 -0
index e0976ae2a523..03bec83944c4 100644
@@ -5260,6 +5260,7 @@ T:	git git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git
 F:	Documentation/admin-guide/cgroup-v1/cpusets.rst
 F:	include/linux/cpuset.h
 F:	kernel/cgroup/cpuset.c
+F:	tools/testing/selftests/cgroup/test_cpuset_prs.sh
 
 CONTROL GROUP - MEMORY RESOURCE CONTROLLER (MEMCG)
 M:	Johannes Weiner <[email protected]>
tools/testing/selftests/cgroup/cgroup_util.c +2 -0
index e8bbbdb77e0d..0340d4ca8f51 100644
@@ -286,6 +286,8 @@ int cg_destroy(const char *cgroup)
 {
 	int ret;
 
+	if (!cgroup)
+		return 0;
 retry:
 	ret = rmdir(cgroup);
 	if (ret && errno == EBUSY) {
tools/testing/selftests/cgroup/cgroup_util.h +2 -0
index c92df4e5d395..1df7f202214a 100644
@@ -11,6 +11,8 @@
 #define USEC_PER_SEC	1000000L
 #define NSEC_PER_SEC	1000000000L
 
+#define TEST_UID	65534 /* usually nobody, any !root is fine */
+
 /*
  * Checks if two given values differ by less than err% of their sum.
  */
tools/testing/selftests/cgroup/test_core.c +1 -1
index 600123503063..80aa6b2373b9 100644
@@ -683,7 +683,7 @@ static int test_cgcore_thread_migration(const char *root)
  */
 static int test_cgcore_lesser_euid_open(const char *root)
 {
-	const uid_t test_euid = 65534;	/* usually nobody, any !root is fine */
+	const uid_t test_euid = TEST_UID;
 	int ret = KSFT_FAIL;
 	char *cg_test_a = NULL, *cg_test_b = NULL;
 	char *cg_test_a_procs = NULL, *cg_test_b_procs = NULL;
tools/testing/selftests/cgroup/test_cpuset_prs.sh +1 -1
index 2b5215cc599f..4afb132e4e4f 100755
@@ -10,7 +10,7 @@
 skip_test() {
 	echo "$1"
 	echo "Test SKIPPED"
-	exit 0
+	exit 4 # ksft_skip
 }
 
 [[ $(id -u) -eq 0 ]] || skip_test "Test must be run as root!"
Signature · 1 line
2.41.0
author maintainer
#4 lore raw

PATCH v2 3/3 selftests: cgroup: Add cpuset migrations testcase

In reply to Michal Koutný

Add a separate testfile to verify treating permissions when tasks are migrated on cgroup v2 hierarchy between cpuset cgroups.

In accordance with v2 design, migration should be allowed based on delegation boundaries (i.e. cgroup.procs permissions) and does not depend on the migrated object (i.e. unprivileged process can migrate another process (even privileged) as long as it remains in the original dedicated scope).

---
 MAINTAINERS                                  |   1 +
 tools/testing/selftests/cgroup/.gitignore    |   1 +
 tools/testing/selftests/cgroup/Makefile      |   2 +
 tools/testing/selftests/cgroup/test_cpuset.c | 272 +++++++++++++++++++
 4 files changed, 276 insertions(+)
 create mode 100644 tools/testing/selftests/cgroup/test_cpuset.c
MAINTAINERS +1 -0
index 03bec83944c4..5c55de000ee3 100644
@@ -5260,6 +5260,7 @@ T:	git git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup.git
 F:	Documentation/admin-guide/cgroup-v1/cpusets.rst
 F:	include/linux/cpuset.h
 F:	kernel/cgroup/cpuset.c
+F:	tools/testing/selftests/cgroup/test_cpuset.c
 F:	tools/testing/selftests/cgroup/test_cpuset_prs.sh
 
 CONTROL GROUP - MEMORY RESOURCE CONTROLLER (MEMCG)
tools/testing/selftests/cgroup/.gitignore +1 -0
index c4a57e69f749..8443a8d46a1c 100644
@@ -5,4 +5,5 @@ test_freezer
 test_kmem
 test_kill
 test_cpu
+test_cpuset
 wait_inotify
tools/testing/selftests/cgroup/Makefile +2 -0
index 3d263747d2ad..dee0f013c7f4 100644
@@ -12,6 +12,7 @@ TEST_GEN_PROGS += test_core
 TEST_GEN_PROGS += test_freezer
 TEST_GEN_PROGS += test_kill
 TEST_GEN_PROGS += test_cpu
+TEST_GEN_PROGS += test_cpuset
 
 LOCAL_HDRS += $(selfdir)/clone3/clone3_selftests.h $(selfdir)/pidfd/pidfd.h
 
@@ -23,3 +24,4 @@ $(OUTPUT)/test_core: cgroup_util.c
 $(OUTPUT)/test_freezer: cgroup_util.c
 $(OUTPUT)/test_kill: cgroup_util.c
 $(OUTPUT)/test_cpu: cgroup_util.c
+$(OUTPUT)/test_cpuset: cgroup_util.c
tools/testing/selftests/cgroup/test_cpuset.c new file +272 -0

272 lines not on this page · raw message

Signature · 1 line
2.41.0