[PATCH v3 3/4] selftests/bpf: Add tests verifying bpf lsm userns_create hook

[Frederick Lawler]

On 7/22/22 1:07 AM, Martin KaFai Lau wrote:
> On Thu, Jul 21, 2022 at 12:28:07PM -0500, Frederick Lawler wrote:
>> The LSM hook userns_create was introduced to provide LSM's an
>> opportunity to block or allow unprivileged user namespace creation. This
>> test serves two purposes: it provides a test eBPF implementation, and
>> tests the hook successfully blocks or allows user namespace creation.
>>
>> This tests 4 cases:
>>
>>          1. Unattached bpf program does not block unpriv user namespace
>>             creation.
>>          2. Attached bpf program allows user namespace creation given
>>             CAP_SYS_ADMIN privileges.
>>          3. Attached bpf program denies user namespace creation for a
>>             user without CAP_SYS_ADMIN.
>>          4. The sleepable implementation loads
>>
>> Signed-off-by: Frederick Lawler <fred at cloudflare.com>
>>
>> ---
>> The generic deny_namespace file name is used for future namespace
>> expansion. I didn't want to limit these files to just the create_user_ns
>> hook.
>> Changes since v2:
>> - Rename create_user_ns hook to userns_create
>> Changes since v1:
>> - Introduce this patch
>> ---
>>   .../selftests/bpf/prog_tests/deny_namespace.c | 88 +++++++++++++++++++
>>   .../selftests/bpf/progs/test_deny_namespace.c | 39 ++++++++
>>   2 files changed, 127 insertions(+)
>>   create mode 100644 tools/testing/selftests/bpf/prog_tests/deny_namespace.c
>>   create mode 100644 tools/testing/selftests/bpf/progs/test_deny_namespace.c
>>
>> diff --git a/tools/testing/selftests/bpf/prog_tests/deny_namespace.c b/tools/testing/selftests/bpf/prog_tests/deny_namespace.c
>> new file mode 100644
>> index 000000000000..9e4714295008
>> --- /dev/null
>> +++ b/tools/testing/selftests/bpf/prog_tests/deny_namespace.c
>> @@ -0,0 +1,88 @@
>> +// SPDX-License-Identifier: GPL-2.0
>> +#define _GNU_SOURCE
>> +#include <test_progs.h>
>> +#include "test_deny_namespace.skel.h"
>> +#include <sched.h>
>> +#include "cap_helpers.h"
>> +
>> +#define STACK_SIZE (1024 * 1024)
> Does the child need 1M stack space ?

No, I can reduce that.

> 
>> +static char child_stack[STACK_SIZE];
>> +
>> +int clone_callback(void *arg)
> static
> 
>> +{
>> +	return 0;
>> +}
>> +
>> +static int create_new_user_ns(void)
>> +{
>> +	int status;
>> +	pid_t cpid;
>> +
>> +	cpid = clone(clone_callback, child_stack + STACK_SIZE,
>> +		     CLONE_NEWUSER | SIGCHLD, NULL);
>> +
>> +	if (cpid == -1)
>> +		return errno;
>> +
>> +	if (cpid == 0)
> Not an expert in clone() call and it is not clear what 0
> return value mean from the man page.  Could you explain ?
> 

Good catch. This is using the libc clone().

>> +		return 0;
>> +
>> +	waitpid(cpid, &status, 0);
>> +	if (WIFEXITED(status))
>> +		return WEXITSTATUS(status);
>> +
>> +	return -1;
>> +}
>> +
>> +static void test_userns_create_bpf(void)
>> +{
>> +	__u32 cap_mask = 1ULL << CAP_SYS_ADMIN;
>> +	__u64 old_caps = 0;
>> +
>> +	ASSERT_OK(create_new_user_ns(), "priv new user ns");
> Does it need to enable CAP_SYS_ADMIN first ?
> 

You're right, this should be more explicitly set. I ran tests with the 
vmtest.sh script supplied with sefltests/bpf which run under root. I 
should always set CAP_SYS_ADMIN here to be consistent.

>> +
>> +	cap_disable_effective(cap_mask, &old_caps);
>> +
>> +	ASSERT_EQ(create_new_user_ns(), EPERM, "unpriv new user ns");
>> +
>> +	if (cap_mask & old_caps)
>> +		cap_enable_effective(cap_mask, NULL);
>> +}
>> +
>> +static void test_unpriv_userns_create_no_bpf(void)
>> +{
>> +	__u32 cap_mask = 1ULL << CAP_SYS_ADMIN;
>> +	__u64 old_caps = 0;
>> +
>> +	cap_disable_effective(cap_mask, &old_caps);
>> +
>> +	ASSERT_OK(create_new_user_ns(), "no-bpf unpriv new user ns");
>> +
>> +	if (cap_mask & old_caps)
>> +		cap_enable_effective(cap_mask, NULL);
>> +}
>> +
>> +void test_deny_namespace(void)
>> +{
>> +	struct test_deny_namespace *skel = NULL;
>> +	int err;
>> +
>> +	if (test__start_subtest("unpriv_userns_create_no_bpf"))
>> +		test_unpriv_userns_create_no_bpf();
>> +
>> +	skel = test_deny_namespace__open_and_load();
>> +	if (!ASSERT_OK_PTR(skel, "skel load"))
>> +		goto close_prog;
>> +
>> +	err = test_deny_namespace__attach(skel);
>> +	if (!ASSERT_OK(err, "attach"))
>> +		goto close_prog;
>> +
>> +	if (test__start_subtest("userns_create_bpf"))
>> +		test_userns_create_bpf();
>> +
>> +	test_deny_namespace__detach(skel);
>> +
>> +close_prog:
>> +	test_deny_namespace__destroy(skel);
>> +}
>> diff --git a/tools/testing/selftests/bpf/progs/test_deny_namespace.c b/tools/testing/selftests/bpf/progs/test_deny_namespace.c
>> new file mode 100644
>> index 000000000000..9ec9dabc8372
>> --- /dev/null
>> +++ b/tools/testing/selftests/bpf/progs/test_deny_namespace.c
>> @@ -0,0 +1,39 @@
>> +// SPDX-License-Identifier: GPL-2.0
>> +#include <linux/bpf.h>
>> +#include <bpf/bpf_helpers.h>
>> +#include <bpf/bpf_tracing.h>
>> +#include <errno.h>
>> +#include <linux/capability.h>
>> +
>> +struct kernel_cap_struct {
>> +	__u32 cap[_LINUX_CAPABILITY_U32S_3];
>> +} __attribute__((preserve_access_index));
>> +
>> +struct cred {
>> +	struct kernel_cap_struct cap_effective;
>> +} __attribute__((preserve_access_index));
>> +
>> +char _license[] SEC("license") = "GPL";
>> +
>> +SEC("lsm/userns_create")
>> +int BPF_PROG(test_userns_create, const struct cred *cred, int ret)
>> +{
>> +	struct kernel_cap_struct caps = cred->cap_effective;
>> +	int cap_index = CAP_TO_INDEX(CAP_SYS_ADMIN);
>> +	__u32 cap_mask = CAP_TO_MASK(CAP_SYS_ADMIN);
>> +
>> +	if (ret)
>> +		return 0;
>> +
>> +	ret = -EPERM;
>> +	if (caps.cap[cap_index] & cap_mask)
>> +		return 0;
>> +
>> +	return -EPERM;
>> +}
>> +
>> +SEC("lsm.s/userns_create")
>> +int BPF_PROG(test_sleepable_userns_create, const struct cred *cred, int ret)
>> +{
> An empty program is weird.  If the intention is
> to ensure a sleepable program can attach to userns_create,
> move the test logic here and remove the non-sleepable
> program above.
> 

Sure, I can do that.

>> +	return 0;
>> +}
>> -- 
>> 2.30.2
>>