Skip to contentSkip to conversation

Remove use of current->cgns in bpf_cgroup_from_id

PATCH Superseded Kumar Kartikeya Dwivedi 3 participants 7 messages Aug 11, 2025 – Aug 12, 2025

Subsystems (5)
Kselftest BPF BPF selftests cgroup +1 more subsystems

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore

Kumar Kartikeya Dwivedi

author maintainer
#0 lore raw

v2 cover Remove use of current->cgns in bpf_cgroup_from_id

bpf_cgroup_from_id currently ends up doing a check on whether the cgroup being looked up is a descendant of the root cgroup of the current task's cgroup namespace. This leads to unreliable results since this kfunc can be invoked from any arbitrary context, for any arbitrary value of current. Fix this by removing namespace-awarness in the kfunc, and include a test that detects such a case and fails without the fix.

Changelog:
----------
v1 -> v2
v1: https://lore.kernel.org/bpf/[email protected]

* Add Ack from Tejun.
* Fix selftest to perform namespace migration and cgroup setup in a child process to avoid changing test_progs namespace.

Kumar Kartikeya Dwivedi (2):
bpf: Do not limit bpf_cgroup_from_id to current's namespace selftests/bpf: Add a test for bpf_cgroup_from_id lookup in non-root
cgns

 include/linux/cgroup.h                        |  2 +-
 kernel/bpf/cgroup_iter.c                      |  2 +-
 kernel/bpf/helpers.c                          |  2 +-
 kernel/cgroup/cgroup.c                        |  7 +-
 .../selftests/bpf/prog_tests/cgrp_kfunc.c     | 76 +++++++++++++++++++
 .../selftests/bpf/progs/cgrp_kfunc_success.c  | 12 +++
 6 files changed, 97 insertions(+), 4 deletions(-)

base-commit: fa479132845e94b60068fad01c2a9979b3efe2dc

Signature · 1 line
2.47.3
author maintainer
#1 lore raw

PATCH v2 1/2 bpf: Do not limit bpf_cgroup_from_id to current's namespace

The bpf_cgroup_from_id kfunc relies on cgroup_get_from_id to obtain the cgroup corresponding to a given cgroup ID. This helper can be called in a lot of contexts where the current thread can be random. A recent example was its use in sched_ext's ops.tick(), to obtain the root cgroup pointer. Since the current task can be whatever random user space task preempted by the timer tick, this makes the behavior of the helper unreliable.

Resolve this by refactoring cgroup_get_from_id to take a parameter to elide the cgroup_is_descendant check when root_cgns parameter is set to true.

There is no compatibility breakage here, since changing the namespace against which the lookup is being done to the root cgroup namespace only permits a wider set of lookups to succeed now. The cgroup IDs across namespaces are globally unique, and thus don't need to be retranslated.

---
 include/linux/cgroup.h   | 2 +-
 kernel/bpf/cgroup_iter.c | 2 +-
 kernel/bpf/helpers.c     | 2 +-
 kernel/cgroup/cgroup.c   | 7 ++++++-
 4 files changed, 9 insertions(+), 4 deletions(-)
include/linux/cgroup.h +1 -1
index b18fb5fcb38e..da757a496fbe 100644
@@ -650,7 +650,7 @@ static inline void cgroup_kthread_ready(void)
 }
 
 void cgroup_path_from_kernfs_id(u64 id, char *buf, size_t buflen);
-struct cgroup *cgroup_get_from_id(u64 id);
+struct cgroup *cgroup_get_from_id(u64 id, bool root_cgns);
 #else /* !CONFIG_CGROUPS */
 
 struct cgroup_subsys_state;
kernel/bpf/cgroup_iter.c +1 -1
index f04a468cf6a7..49234d035583 100644
@@ -212,7 +212,7 @@ static int bpf_iter_attach_cgroup(struct bpf_prog *prog,
 	if (fd)
 		cgrp = cgroup_v1v2_get_from_fd(fd);
 	else if (id)
-		cgrp = cgroup_get_from_id(id);
+		cgrp = cgroup_get_from_id(id, false);
 	else /* walk the entire hierarchy by default. */
 		cgrp = cgroup_get_from_path("/");
 
kernel/bpf/helpers.c +1 -1
index 6b4877e85a68..12466103917f 100644
@@ -2537,7 +2537,7 @@ __bpf_kfunc struct cgroup *bpf_cgroup_from_id(u64 cgid)
 {
 	struct cgroup *cgrp;
 
-	cgrp = cgroup_get_from_id(cgid);
+	cgrp = cgroup_get_from_id(cgid, true);
 	if (IS_ERR(cgrp))
 		return NULL;
 	return cgrp;
kernel/cgroup/cgroup.c +6 -1
index 312c6a8b55bb..b490e1e0d2c4 100644
@@ -6345,10 +6345,11 @@ void cgroup_path_from_kernfs_id(u64 id, char *buf, size_t buflen)
 /*
  * cgroup_get_from_id : get the cgroup associated with cgroup id
  * @id: cgroup id
+ * @root_cgns: Select root cgroup namespace instead of current's.
  * On success return the cgrp or ERR_PTR on failure
  * Only cgroups within current task's cgroup NS are valid.
  */
-struct cgroup *cgroup_get_from_id(u64 id)
+struct cgroup *cgroup_get_from_id(u64 id, bool root_cgns)
 {
 	struct kernfs_node *kn;
 	struct cgroup *cgrp, *root_cgrp;
@@ -6374,6 +6375,10 @@ struct cgroup *cgroup_get_from_id(u64 id)
 	if (!cgrp)
 		return ERR_PTR(-ENOENT);
 
+	/* We don't need to namespace this operation against current. */
+	if (root_cgns)
+		return cgrp;
+
 	root_cgrp = current_cgns_cgroup_dfl();
 	if (!cgroup_is_descendant(cgrp, root_cgrp)) {
 		cgroup_put(cgrp);
Signature · 1 line
2.47.3
#2 lore raw

Re: [PATCH bpf-next v2 1/2] bpf: Do not limit bpf_cgroup_from_id to current's namespace

Quoted 18 lines · Kumar Kartikeya Dwivedi

The bpf_cgroup_from_id kfunc relies on cgroup_get_from_id to obtain the cgroup corresponding to a given cgroup ID. This helper can be called in a lot of contexts where the current thread can be random. A recent example was its use in sched_ext's ops.tick(), to obtain the root cgroup pointer. Since the current task can be whatever random user space task preempted by the timer tick, this makes the behavior of the helper unreliable.

Resolve this by refactoring cgroup_get_from_id to take a parameter to elide the cgroup_is_descendant check when root_cgns parameter is set to true.

There is no compatibility breakage here, since changing the namespace against which the lookup is being done to the root cgroup namespace only permits a wider set of lookups to succeed now. The cgroup IDs across namespaces are globally unique, and thus don't need to be retranslated.

Quoted 8 lines
---
 include/linux/cgroup.h   | 2 +-
 kernel/bpf/cgroup_iter.c | 2 +-
 kernel/bpf/helpers.c     | 2 +-
 kernel/cgroup/cgroup.c   | 7 ++++++-

hmm... I see mm/memcontrol.c and block/blk-cgroup-fc-appid.c using this function as well, but you didn't update them, so under some kernel config this will break?

but I'm also wondering if it wouldn't be cleaner to keep cgroup_get_from_id() as is, and add __cgroup_get_from_id(u64 id) that would not perform a cgroup_is_descendant() check. BPF helpers that don't need this root descendancy would call __cgroup_get_from_id() directly, while cgroup_get_from_id() would use __cgroup_get_from_id() first, and then (if successful) would perform addition root_cgrp check?

Quoted 72 lines
 4 files changed, 9 insertions(+), 4 deletions(-)
include/linux/cgroup.h +1 -1
index b18fb5fcb38e..da757a496fbe 100644
@@ -650,7 +650,7 @@ static inline void cgroup_kthread_ready(void)
 }
 
 void cgroup_path_from_kernfs_id(u64 id, char *buf, size_t buflen);
-struct cgroup *cgroup_get_from_id(u64 id);
+struct cgroup *cgroup_get_from_id(u64 id, bool root_cgns);
 #else /* !CONFIG_CGROUPS */
 
 struct cgroup_subsys_state;
kernel/bpf/cgroup_iter.c +1 -1
index f04a468cf6a7..49234d035583 100644
@@ -212,7 +212,7 @@ static int bpf_iter_attach_cgroup(struct bpf_prog *prog,
        if (fd)
                cgrp = cgroup_v1v2_get_from_fd(fd);
        else if (id)
-               cgrp = cgroup_get_from_id(id);
+               cgrp = cgroup_get_from_id(id, false);
        else /* walk the entire hierarchy by default. */
                cgrp = cgroup_get_from_path("/");
 
kernel/bpf/helpers.c +1 -1
index 6b4877e85a68..12466103917f 100644
@@ -2537,7 +2537,7 @@ __bpf_kfunc struct cgroup *bpf_cgroup_from_id(u64 cgid)
 {
        struct cgroup *cgrp;
 
-       cgrp = cgroup_get_from_id(cgid);
+       cgrp = cgroup_get_from_id(cgid, true);
        if (IS_ERR(cgrp))
                return NULL;
        return cgrp;
kernel/cgroup/cgroup.c +6 -2
index 312c6a8b55bb..b490e1e0d2c4 100644
@@ -6345,10 +6345,11 @@ void cgroup_path_from_kernfs_id(u64 id, char *buf, size_t buflen)
 /*
  * cgroup_get_from_id : get the cgroup associated with cgroup id
  * @id: cgroup id
+ * @root_cgns: Select root cgroup namespace instead of current's.
  * On success return the cgrp or ERR_PTR on failure
  * Only cgroups within current task's cgroup NS are valid.
  */
-struct cgroup *cgroup_get_from_id(u64 id)
+struct cgroup *cgroup_get_from_id(u64 id, bool root_cgns)
 {
        struct kernfs_node *kn;
        struct cgroup *cgrp, *root_cgrp;
@@ -6374,6 +6375,10 @@ struct cgroup *cgroup_get_from_id(u64 id)
        if (!cgrp)
                return ERR_PTR(-ENOENT);
 
+       /* We don't need to namespace this operation against current. */
+       if (root_cgns)
+               return cgrp;
+
        root_cgrp = current_cgns_cgroup_dfl();
        if (!cgroup_is_descendant(cgrp, root_cgrp)) {
                cgroup_put(cgrp);
--

2.47.3

author maintainer
#3 lore raw

Re: [PATCH bpf-next v2 1/2] bpf: Do not limit bpf_cgroup_from_id to current's namespace

Quoted 30 lines · Andrii Nakryiko
Quoted 18 lines · Kumar Kartikeya Dwivedi

The bpf_cgroup_from_id kfunc relies on cgroup_get_from_id to obtain the cgroup corresponding to a given cgroup ID. This helper can be called in a lot of contexts where the current thread can be random. A recent example was its use in sched_ext's ops.tick(), to obtain the root cgroup pointer. Since the current task can be whatever random user space task preempted by the timer tick, this makes the behavior of the helper unreliable.

Resolve this by refactoring cgroup_get_from_id to take a parameter to elide the cgroup_is_descendant check when root_cgns parameter is set to true.

There is no compatibility breakage here, since changing the namespace against which the lookup is being done to the root cgroup namespace only permits a wider set of lookups to succeed now. The cgroup IDs across namespaces are globally unique, and thus don't need to be retranslated.

Quoted 8 lines
---
 include/linux/cgroup.h   | 2 +-
 kernel/bpf/cgroup_iter.c | 2 +-
 kernel/bpf/helpers.c     | 2 +-
 kernel/cgroup/cgroup.c   | 7 ++++++-
Quoted 3 lines

hmm... I see mm/memcontrol.c and block/blk-cgroup-fc-appid.c using this function as well, but you didn't update them, so under some kernel config this will break?

Yeah, addressed in upcoming v3, I also got a kernel test robot error privately about it.

Quoted 8 lines

but I'm also wondering if it wouldn't be cleaner to keep cgroup_get_from_id() as is, and add __cgroup_get_from_id(u64 id) that would not perform a cgroup_is_descendant() check. BPF helpers that don't need this root descendancy would call __cgroup_get_from_id() directly, while cgroup_get_from_id() would use __cgroup_get_from_id() first, and then (if successful) would perform addition root_cgrp check?

Yeah, I can refactor as such.

Quoted 74 lines
Quoted 74 lines

 4 files changed, 9 insertions(+), 4 deletions(-)
include/linux/cgroup.h +1 -1
index b18fb5fcb38e..da757a496fbe 100644
@@ -650,7 +650,7 @@ static inline void cgroup_kthread_ready(void)
 }
 
 void cgroup_path_from_kernfs_id(u64 id, char *buf, size_t buflen);
-struct cgroup *cgroup_get_from_id(u64 id);
+struct cgroup *cgroup_get_from_id(u64 id, bool root_cgns);
 #else /* !CONFIG_CGROUPS */
 
 struct cgroup_subsys_state;
kernel/bpf/cgroup_iter.c +1 -1
index f04a468cf6a7..49234d035583 100644
@@ -212,7 +212,7 @@ static int bpf_iter_attach_cgroup(struct bpf_prog *prog,
        if (fd)
                cgrp = cgroup_v1v2_get_from_fd(fd);
        else if (id)
-               cgrp = cgroup_get_from_id(id);
+               cgrp = cgroup_get_from_id(id, false);
        else /* walk the entire hierarchy by default. */
                cgrp = cgroup_get_from_path("/");
 
kernel/bpf/helpers.c +1 -1
index 6b4877e85a68..12466103917f 100644
@@ -2537,7 +2537,7 @@ __bpf_kfunc struct cgroup *bpf_cgroup_from_id(u64 cgid)
 {
        struct cgroup *cgrp;
 
-       cgrp = cgroup_get_from_id(cgid);
+       cgrp = cgroup_get_from_id(cgid, true);
        if (IS_ERR(cgrp))
                return NULL;
        return cgrp;
kernel/cgroup/cgroup.c +6 -2
index 312c6a8b55bb..b490e1e0d2c4 100644
@@ -6345,10 +6345,11 @@ void cgroup_path_from_kernfs_id(u64 id, char *buf, size_t buflen)
 /*
  * cgroup_get_from_id : get the cgroup associated with cgroup id
  * @id: cgroup id
+ * @root_cgns: Select root cgroup namespace instead of current's.
  * On success return the cgrp or ERR_PTR on failure
  * Only cgroups within current task's cgroup NS are valid.
  */
-struct cgroup *cgroup_get_from_id(u64 id)
+struct cgroup *cgroup_get_from_id(u64 id, bool root_cgns)
 {
        struct kernfs_node *kn;
        struct cgroup *cgrp, *root_cgrp;
@@ -6374,6 +6375,10 @@ struct cgroup *cgroup_get_from_id(u64 id)
        if (!cgrp)
                return ERR_PTR(-ENOENT);
 
+       /* We don't need to namespace this operation against current. */
+       if (root_cgns)
+               return cgrp;
+
        root_cgrp = current_cgns_cgroup_dfl();
        if (!cgroup_is_descendant(cgrp, root_cgrp)) {
                cgroup_put(cgrp);
--

2.47.3

author maintainer
#4 lore raw

PATCH v2 2/2 selftests/bpf: Add a test for bpf_cgroup_from_id lookup in non-root cgns

In reply to Kumar Kartikeya Dwivedi

Make sure that we only switch the cgroup namespace and enter a new cgroup in a child process separate from test_progs, to not mess up the environment for subsequent tests.

---
 .../selftests/bpf/prog_tests/cgrp_kfunc.c     | 76 +++++++++++++++++++
 .../selftests/bpf/progs/cgrp_kfunc_success.c  | 12 +++
 2 files changed, 88 insertions(+)
tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c +76 -0
index adda85f97058..e75a29728f9c 100644
@@ -4,6 +4,7 @@
 #define _GNU_SOURCE
 #include <cgroup_helpers.h>
 #include <test_progs.h>
+#include <sched.h>
 
 #include "cgrp_kfunc_failure.skel.h"
 #include "cgrp_kfunc_success.skel.h"
@@ -87,6 +88,78 @@ static const char * const success_tests[] = {
 	"test_cgrp_from_id",
 };
 
+static void test_cgrp_from_id_ns(void)
+{
+	LIBBPF_OPTS(bpf_test_run_opts, opts);
+	struct cgrp_kfunc_success *skel;
+	struct bpf_program *prog;
+	int fd, pid, pipe_fd[2];
+
+	skel = open_load_cgrp_kfunc_skel();
+	if (!ASSERT_OK_PTR(skel, "open_load_skel"))
+		return;
+
+	if (!ASSERT_OK(skel->bss->err, "pre_mkdir_err"))
+		goto cleanup;
+
+	prog = bpf_object__find_program_by_name(skel->obj, "test_cgrp_from_id_ns");
+	if (!ASSERT_OK_PTR(prog, "bpf_object__find_program_by_name"))
+		goto cleanup;
+
+	if (!ASSERT_OK(pipe(pipe_fd), "pipe"))
+		goto cleanup;
+
+	pid = fork();
+	if (!ASSERT_GE(pid, 0, "fork result"))
+		goto pipe_cleanup;
+
+	if (pid == 0) {
+		int ret = 1;
+
+		close(pipe_fd[0]);
+		fd = create_and_get_cgroup("cgrp_from_id_ns");
+		if (!ASSERT_GE(fd, 0, "cgrp_fd"))
+			_exit(1);
+
+		if (!ASSERT_OK(join_cgroup("cgrp_from_id_ns"), "join cgrp"))
+			goto fail;
+
+		if (!ASSERT_OK(unshare(CLONE_NEWCGROUP), "unshare cgns"))
+			goto fail;
+
+		ret = bpf_prog_test_run_opts(bpf_program__fd(prog), &opts);
+		if (!ASSERT_OK(ret, "test run ret"))
+			goto fail;
+
+		remove_cgroup("cgrp_from_id_ns");
+
+		if (!ASSERT_OK(opts.retval, "test run retval"))
+			_exit(1);
+		ret = 0;
+		close(fd);
+		if (!ASSERT_EQ(write(pipe_fd[1], &ret, sizeof(ret)), sizeof(ret), "write pipe"))
+			_exit(1);
+
+		_exit(0);
+fail:
+		remove_cgroup("cgrp_from_id_ns");
+		_exit(1);
+	} else {
+		int res;
+
+		close(pipe_fd[1]);
+		if (!ASSERT_EQ(read(pipe_fd[0], &res, sizeof(res)), sizeof(res), "read res"))
+			goto pipe_cleanup;
+		if (!ASSERT_OK(res, "result from run"))
+			goto pipe_cleanup;
+	}
+
+pipe_cleanup:
+	close(pipe_fd[1]);
+cleanup:
+	cgrp_kfunc_success__destroy(skel);
+}
+
 void test_cgrp_kfunc(void)
 {
 	int i, err;
@@ -102,6 +175,9 @@ void test_cgrp_kfunc(void)
 		run_success_test(success_tests[i]);
 	}
 
+	if (test__start_subtest("test_cgrp_from_id_ns"))
+		test_cgrp_from_id_ns();
+
 	RUN_TESTS(cgrp_kfunc_failure);
 
 cleanup:
tools/testing/selftests/bpf/progs/cgrp_kfunc_success.c +12 -0
index 5354455a01be..02d8f160ca0e 100644
@@ -221,3 +221,15 @@ int BPF_PROG(test_cgrp_from_id, struct cgroup *cgrp, const char *path)
 
 	return 0;
 }
+
+SEC("syscall")
+int test_cgrp_from_id_ns(void *ctx)
+{
+	struct cgroup *cg;
+
+	cg = bpf_cgroup_from_id(1);
+	if (!cg)
+		return 42;
+	bpf_cgroup_release(cg);
+	return 0;
+}
Signature · 1 line
2.47.3
maintainer
#5 lore raw

Re: [PATCH bpf-next v2 2/2] selftests/bpf: Add a test for bpf_cgroup_from_id lookup in non-root cgns

On Mon, 2025-08-11 at 12:59 -0700, Kumar Kartikeya Dwivedi wrote:

[...]

Quoted 5 lines
tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c +5 -0
+static void test_cgrp_from_id_ns(void)
+{
+	LIBBPF_OPTS(bpf_test_run_opts, opts);
+	struct cgrp_kfunc_success *skel;
+	struct bpf_program *prog;

tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c static const char * const success_tests[] = { · line 105

Quoted 10 lines
tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c +10 -0
+	int fd, pid, pipe_fd[2];
+
+	skel = open_load_cgrp_kfunc_skel();
+	if (!ASSERT_OK_PTR(skel, "open_load_skel"))
+		return;
+
+	if (!ASSERT_OK(skel->bss->err, "pre_mkdir_err"))
+		goto cleanup;
+
+	prog = bpf_object__find_program_by_name(skel->obj, "test_cgrp_from_id_ns");

Nit: skel->test_cgrp_from_id_ns ?

Quoted 19 lines
tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c +19 -0
+	if (!ASSERT_OK_PTR(prog, "bpf_object__find_program_by_name"))
+		goto cleanup;
+
+	if (!ASSERT_OK(pipe(pipe_fd), "pipe"))
+		goto cleanup;
+
+	pid = fork();
+	if (!ASSERT_GE(pid, 0, "fork result"))
+		goto pipe_cleanup;
+
+	if (pid == 0) {
+		int ret = 1;
+
+		close(pipe_fd[0]);
+		fd = create_and_get_cgroup("cgrp_from_id_ns");
+		if (!ASSERT_GE(fd, 0, "cgrp_fd"))
+			_exit(1);
+
+		if (!ASSERT_OK(join_cgroup("cgrp_from_id_ns"), "join cgrp"))

tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c static const char * const success_tests[] = { · line 134

Quoted 10 lines
tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c +10 -0
+			goto fail;
+
+		if (!ASSERT_OK(unshare(CLONE_NEWCGROUP), "unshare cgns"))
+			goto fail;
+
+		ret = bpf_prog_test_run_opts(bpf_program__fd(prog), &opts);
+		if (!ASSERT_OK(ret, "test run ret"))
+			goto fail;
+
+		remove_cgroup("cgrp_from_id_ns");

If this test is executed in -vvv mode, the following is printed:

(cgroup_helpers.c:412: errno: Device or resource busy) rmdiring cgroup cgrp_from_id_ns ...

And cgroup is still in place after exit. As far as I understand, child process needs to change cgroup again or remove_cgroup needs to be called in the parent process.

tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c static const char * const success_tests[] = { · line 137

Quoted 3 lines
tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c +3 -0
+
+		if (!ASSERT_OK(opts.retval, "test run retval"))
+			_exit(1);

Nit: why not 'exit'? '_exit' does not flush file descriptors.

Quoted 11 lines
tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c +11 -0
+		ret = 0;
+		close(fd);
+		if (!ASSERT_EQ(write(pipe_fd[1], &ret, sizeof(ret)), sizeof(ret), "write pipe"))
+			_exit(1);
+
+		_exit(0);
+fail:
+		remove_cgroup("cgrp_from_id_ns");
+		_exit(1);
+	} else {
+		int res;

tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c static const char * const success_tests[] = { · line 158

Quoted 10 lines
tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c +10 -0
+
+		close(pipe_fd[1]);
+		if (!ASSERT_EQ(read(pipe_fd[0], &res, sizeof(res)), sizeof(res), "read res"))
+			goto pipe_cleanup;
+		if (!ASSERT_OK(res, "result from run"))
+			goto pipe_cleanup;
+	}
+
+pipe_cleanup:
+	close(pipe_fd[1]);

Nit: should this be pipe_fd[0]?
in case of a fork() failure, should this be both?

Quoted 7 lines
tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c +4 -0
+cleanup:
+	cgrp_kfunc_success__destroy(skel);
+}
+
 void test_cgrp_kfunc(void)
 {
 	int i, err;

[...]

author maintainer
#6 lore raw

Re: [PATCH bpf-next v2 2/2] selftests/bpf: Add a test for bpf_cgroup_from_id lookup in non-root cgns

Quoted 14 lines · Eduard Zingerman

On Mon, 2025-08-11 at 12:59 -0700, Kumar Kartikeya Dwivedi wrote:

[...]

Quoted 9 lines
tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c +9 -0
+static void test_cgrp_from_id_ns(void)
+{
+     LIBBPF_OPTS(bpf_test_run_opts, opts);
+     struct cgrp_kfunc_success *skel;
+     struct bpf_program *prog;
+     int fd, pid, pipe_fd[2];
+
+     skel = open_load_cgrp_kfunc_skel();
+     if (!ASSERT_OK_PTR(skel, "open_load_skel"))
Quoted 8 lines

tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c static const char * const success_tests[] = { · line 105

Quoted 6 lines
tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c +6 -0
+             return;
+
+     if (!ASSERT_OK(skel->bss->err, "pre_mkdir_err"))
+             goto cleanup;
+
+     prog = bpf_object__find_program_by_name(skel->obj, "test_cgrp_from_id_ns");

Nit: skel->test_cgrp_from_id_ns ?

Fixed.

Quoted 33 lines
Quoted 20 lines
tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c +19 -0
+     if (!ASSERT_OK_PTR(prog, "bpf_object__find_program_by_name"))
+             goto cleanup;
+
+     if (!ASSERT_OK(pipe(pipe_fd), "pipe"))
+             goto cleanup;
+
+     pid = fork();
+     if (!ASSERT_GE(pid, 0, "fork result"))
+             goto pipe_cleanup;
+
+     if (pid == 0) {
+             int ret = 1;
+
+             close(pipe_fd[0]);
+             fd = create_and_get_cgroup("cgrp_from_id_ns");
+             if (!ASSERT_GE(fd, 0, "cgrp_fd"))
+                     _exit(1);
+
+             if (!ASSERT_OK(join_cgroup("cgrp_from_id_ns"), "join cgrp"))

tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c static const char * const success_tests[] = { · line 134

Quoted 10 lines
tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c +10 -0
+                     goto fail;
+
+             if (!ASSERT_OK(unshare(CLONE_NEWCGROUP), "unshare cgns"))
+                     goto fail;
+
+             ret = bpf_prog_test_run_opts(bpf_program__fd(prog), &opts);
+             if (!ASSERT_OK(ret, "test run ret"))
+                     goto fail;
+
+             remove_cgroup("cgrp_from_id_ns");

If this test is executed in -vvv mode, the following is printed:

Quoted 5 lines

(cgroup_helpers.c:412: errno: Device or resource busy) rmdiring cgroup cgrp_from_id_ns ...

And cgroup is still in place after exit. As far as I understand, child process needs to change cgroup again or remove_cgroup needs to be called in the parent process.

It is probably because the cgroup fd is still open, I will check and fix.

Quoted 6 lines

tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c static const char * const success_tests[] = { · line 137

Quoted 4 lines
tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c +3 -0
+
+             if (!ASSERT_OK(opts.retval, "test run retval"))
+                     _exit(1);

Nit: why not 'exit'? '_exit' does not flush file descriptors.

Fixed.

Quoted 17 lines
Quoted 17 lines
tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c +16 -0
+             ret = 0;
+             close(fd);
+             if (!ASSERT_EQ(write(pipe_fd[1], &ret, sizeof(ret)), sizeof(ret), "write pipe"))
+                     _exit(1);
+
+             _exit(0);
+fail:
+             remove_cgroup("cgrp_from_id_ns");
+             _exit(1);
+     } else {
+             int res;
+
+             close(pipe_fd[1]);
+             if (!ASSERT_EQ(read(pipe_fd[0], &res, sizeof(res)), sizeof(res), "read res"))
+                     goto pipe_cleanup;
+             if (!ASSERT_OK(res, "result from run"))
Quoted 8 lines

tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c static const char * const success_tests[] = { · line 158

Quoted 5 lines
tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c +5 -0
+                     goto pipe_cleanup;
+     }
+
+pipe_cleanup:
+     close(pipe_fd[1]);

Nit: should this be pipe_fd[0]?
in case of a fork() failure, should this be both?

Yeah, fixed.

Quoted 10 lines
Quoted 8 lines
tools/testing/selftests/bpf/prog_tests/cgrp_kfunc.c +4 -0
+cleanup:
+     cgrp_kfunc_success__destroy(skel);
+}
+
 void test_cgrp_kfunc(void)
 {
      int i, err;

[...]