@@ -37,6 +37,11 @@ struct landlock_ruleset_attr {
* rule explicitly allow them.
*/
__u64 handled_access_net;
+ /**
+ * scoped: Bitmask of actions (cf. `Scope access flags`_)
+ * which are confined to only affect the current Landlock domain.
+ */
+ __u64 scoped;
};
/*
@@ -266,4 +271,26 @@ struct landlock_net_port_attr {
#define LANDLOCK_ACCESS_NET_BIND_TCP (1ULL << 0)
#define LANDLOCK_ACCESS_NET_CONNECT_TCP (1ULL << 1)
/* clang-format on */
+
+/**
+ * DOC: scope
+ *
+ * .scoped attribute handles a set of restrictions on kernel IPCs through
+ * the following flags.
+ *
+ * Scope access flags
+ * ~~~~~~~~~~~~~~~~~~~~
+ *
+ * These flags enable to restrict a sandboxed process from a set of IPC
+ * actions. Setting a flag in a landlock domain will isolate the Landlock
+ * domain to forbid connections to resources outside the domain.
+ *
+ * IPCs with allowed actions:
+ * - %LANDLOCK_SCOPED_ABSTRACT_UNIX_SOCKET: Restrict a sandbox process to
+ * connect to a process outside of the sandbox domain through abstract
+ * unix sockets.
+ */
+/* clang-format off */
+#define LANDLOCK_SCOPED_ABSTRACT_UNIX_SOCKET (1ULL << 0)
+/* clang-format on*/
#endif /* _UAPI_LINUX_LANDLOCK_H */
@@ -26,6 +26,9 @@
#define LANDLOCK_MASK_ACCESS_NET ((LANDLOCK_LAST_ACCESS_NET << 1) - 1)
#define LANDLOCK_NUM_ACCESS_NET __const_hweight64(LANDLOCK_MASK_ACCESS_NET)
+#define LANDLOCK_LAST_ACCESS_SCOPE LANDLOCK_SCOPED_ABSTRACT_UNIX_SOCKET
+#define LANDLOCK_MASK_ACCESS_SCOPE ((LANDLOCK_LAST_ACCESS_SCOPE << 1) - 1)
+#define LANDLOCK_NUM_ACCESS_SCOPE __const_hweight64(LANDLOCK_MASK_ACCESS_SCOPE)
/* clang-format on */
#endif /* _SECURITY_LANDLOCK_LIMITS_H */
@@ -52,12 +52,13 @@ static struct landlock_ruleset *create_ruleset(const u32 num_layers)
struct landlock_ruleset *
landlock_create_ruleset(const access_mask_t fs_access_mask,
- const access_mask_t net_access_mask)
+ const access_mask_t net_access_mask,
+ const access_mask_t scope_mask)
{
struct landlock_ruleset *new_ruleset;
/* Informs about useless ruleset. */
- if (!fs_access_mask && !net_access_mask)
+ if (!fs_access_mask && !net_access_mask && !scope_mask)
return ERR_PTR(-ENOMSG);
new_ruleset = create_ruleset(1);
if (IS_ERR(new_ruleset))
@@ -66,6 +67,8 @@ landlock_create_ruleset(const access_mask_t fs_access_mask,
landlock_add_fs_access_mask(new_ruleset, fs_access_mask, 0);
if (net_access_mask)
landlock_add_net_access_mask(new_ruleset, net_access_mask, 0);
+ if (scope_mask)
+ landlock_add_scope_mask(new_ruleset, scope_mask, 0);
return new_ruleset;
}
@@ -35,6 +35,8 @@ typedef u16 access_mask_t;
static_assert(BITS_PER_TYPE(access_mask_t) >= LANDLOCK_NUM_ACCESS_FS);
/* Makes sure all network access rights can be stored. */
static_assert(BITS_PER_TYPE(access_mask_t) >= LANDLOCK_NUM_ACCESS_NET);
+/* Makes sure all scoped rights can be stored*/
+static_assert(BITS_PER_TYPE(access_mask_t) >= LANDLOCK_NUM_ACCESS_SCOPE);
/* Makes sure for_each_set_bit() and for_each_clear_bit() calls are OK. */
static_assert(sizeof(unsigned long) >= sizeof(access_mask_t));
@@ -42,8 +44,15 @@ static_assert(sizeof(unsigned long) >= sizeof(access_mask_t));
struct access_masks {
access_mask_t fs : LANDLOCK_NUM_ACCESS_FS;
access_mask_t net : LANDLOCK_NUM_ACCESS_NET;
+ access_mask_t scoped : LANDLOCK_NUM_ACCESS_SCOPE;
};
+typedef u32 access_masks_t;
+/* Makes sure all ruleset access rights can be stored. */
+static_assert(BITS_PER_TYPE(access_masks_t) >=
+ LANDLOCK_NUM_ACCESS_FS + LANDLOCK_NUM_ACCESS_NET +
+ LANDLOCK_NUM_ACCESS_SCOPE);
+
typedef u16 layer_mask_t;
/* Makes sure all layers can be checked. */
static_assert(BITS_PER_TYPE(layer_mask_t) >= LANDLOCK_MAX_NUM_LAYERS);
@@ -233,7 +242,8 @@ struct landlock_ruleset {
struct landlock_ruleset *
landlock_create_ruleset(const access_mask_t access_mask_fs,
- const access_mask_t access_mask_net);
+ const access_mask_t access_mask_net,
+ const access_mask_t scope_mask);
void landlock_put_ruleset(struct landlock_ruleset *const ruleset);
void landlock_put_ruleset_deferred(struct landlock_ruleset *const ruleset);
@@ -280,6 +290,16 @@ landlock_add_net_access_mask(struct landlock_ruleset *const ruleset,
ruleset->access_masks[layer_level].net |= net_mask;
}
+static inline void
+landlock_add_scope_mask(struct landlock_ruleset *const ruleset,
+ const access_mask_t scope_mask, const u16 layer_level)
+{
+ access_mask_t scoped_mask = scope_mask & LANDLOCK_MASK_ACCESS_SCOPE;
+
+ WARN_ON_ONCE(scope_mask != scoped_mask);
+ ruleset->access_masks[layer_level].scoped |= scoped_mask;
+}
+
static inline access_mask_t
landlock_get_raw_fs_access_mask(const struct landlock_ruleset *const ruleset,
const u16 layer_level)
@@ -303,6 +323,13 @@ landlock_get_net_access_mask(const struct landlock_ruleset *const ruleset,
return ruleset->access_masks[layer_level].net;
}
+static inline access_mask_t
+landlock_get_scope_mask(const struct landlock_ruleset *const ruleset,
+ const u16 layer_level)
+{
+ return ruleset->access_masks[layer_level].scoped;
+}
+
bool landlock_unmask_layers(const struct landlock_rule *const rule,
const access_mask_t access_request,
layer_mask_t (*const layer_masks)[],
@@ -97,8 +97,9 @@ static void build_check_abi(void)
*/
ruleset_size = sizeof(ruleset_attr.handled_access_fs);
ruleset_size += sizeof(ruleset_attr.handled_access_net);
+ ruleset_size += sizeof(ruleset_attr.scoped);
BUILD_BUG_ON(sizeof(ruleset_attr) != ruleset_size);
- BUILD_BUG_ON(sizeof(ruleset_attr) != 16);
+ BUILD_BUG_ON(sizeof(ruleset_attr) != 24);
path_beneath_size = sizeof(path_beneath_attr.allowed_access);
path_beneath_size += sizeof(path_beneath_attr.parent_fd);
@@ -170,7 +171,7 @@ static const struct file_operations ruleset_fops = {
* Possible returned errors are:
*
* - %EOPNOTSUPP: Landlock is supported by the kernel but disabled at boot time;
- * - %EINVAL: unknown @flags, or unknown access, or too small @size;
+ * - %EINVAL: unknown @flags, or unknown access, or uknown scope, or too small @size;
* - %E2BIG or %EFAULT: @attr or @size inconsistencies;
* - %ENOMSG: empty &landlock_ruleset_attr.handled_access_fs.
*/
@@ -212,10 +213,16 @@ SYSCALL_DEFINE3(landlock_create_ruleset,
if ((ruleset_attr.handled_access_net | LANDLOCK_MASK_ACCESS_NET) !=
LANDLOCK_MASK_ACCESS_NET)
return -EINVAL;
+
+ /* Checks IPC scoping content (and 32-bits cast). */
+ if ((ruleset_attr.scoped | LANDLOCK_MASK_ACCESS_SCOPE) !=
+ LANDLOCK_MASK_ACCESS_SCOPE)
+ return -EINVAL;
/* Checks arguments and transforms to kernel struct. */
ruleset = landlock_create_ruleset(ruleset_attr.handled_access_fs,
- ruleset_attr.handled_access_net);
+ ruleset_attr.handled_access_net,
+ ruleset_attr.scoped);
if (IS_ERR(ruleset))
return PTR_ERR(ruleset);
@@ -13,6 +13,8 @@
#include <linux/lsm_hooks.h>
#include <linux/rcupdate.h>
#include <linux/sched.h>
+#include <net/sock.h>
+#include <net/af_unix.h>
#include "common.h"
#include "cred.h"
@@ -108,9 +110,65 @@ static int hook_ptrace_traceme(struct task_struct *const parent)
return task_ptrace(parent, current);
}
+static access_mask_t
+get_scoped_accesses(const struct landlock_ruleset *const domain)
+{
+ access_mask_t access_dom = 0;
+ size_t layer_level;
+
+ for (layer_level = 0; layer_level < domain->num_layers; layer_level++)
+ access_dom |= landlock_get_scope_mask(domain, layer_level);
+ return access_dom;
+}
+
+static bool sock_is_scoped(struct sock *const other)
+{
+ bool is_scoped = true;
+ const struct landlock_ruleset *dom_other;
+ const struct landlock_ruleset *const dom =
+ landlock_get_current_domain();
+
+ /* quick return if there is no domain or .scoped is not set */
+ if (!dom || !get_scoped_accesses(dom))
+ return true;
+
+ /* the credentials will not change */
+ lockdep_assert_held(&unix_sk(other)->lock);
+ if (other->sk_type != SOCK_DGRAM){
+ dom_other = landlock_cred(other->sk_peer_cred)->domain;
+ } else {
+ dom_other = landlock_cred(other->sk_socket->file->f_cred)->domain;
+ }
+ is_scoped = domain_scope_le(dom, dom_other);
+ return is_scoped;
+}
+
+static int hook_unix_stream_connect(struct sock *const sock,
+ struct sock *const other,
+ struct sock *const newsk)
+{
+ if (sock_is_scoped(other))
+ return 0;
+
+ return -EPERM;
+}
+
+static int hook_unix_may_send(struct socket *const sock,
+ struct socket *const other)
+{
+ pr_warn("XXX %s:%d sock->file:%p other->file:%p\n", __func__, __LINE__,
+ sock->file, other->file);
+ if (sock_is_scoped(other->sk))
+ return 0;
+
+ return -EPERM;
+}
+
static struct security_hook_list landlock_hooks[] __ro_after_init = {
LSM_HOOK_INIT(ptrace_access_check, hook_ptrace_access_check),
LSM_HOOK_INIT(ptrace_traceme, hook_ptrace_traceme),
+ LSM_HOOK_INIT(unix_stream_connect, hook_unix_stream_connect),
+ LSM_HOOK_INIT(unix_may_send, hook_unix_may_send),
};
__init void landlock_add_task_hooks(void)
@@ -17,6 +17,10 @@
#include <sys/wait.h>
#include <unistd.h>
+#include <sys/socket.h>
+#include <sys/un.h>
+#include <stddef.h>
+
#include "common.h"
/* Copied from security/yama/yama_lsm.c */
@@ -436,4 +440,261 @@ TEST_F(hierarchy, trace)
_metadata->exit_code = KSFT_FAIL;
}
+static void create_unix_domain(struct __test_metadata *const _metadata)
+{
+ int ruleset_fd;
+ const struct landlock_ruleset_attr ruleset_attr = {
+ .scoped = LANDLOCK_SCOPED_ABSTRACT_UNIX_SOCKET,
+ };
+
+ ruleset_fd =
+ landlock_create_ruleset(&ruleset_attr, sizeof(ruleset_attr), 0);
+ EXPECT_LE(0, ruleset_fd)
+ {
+ TH_LOG("Failed to create a ruleset: %s", strerror(errno));
+ }
+ enforce_ruleset(_metadata, ruleset_fd);
+ EXPECT_EQ(0, close(ruleset_fd));
+}
+
+/* clang-format off */
+FIXTURE(unix_socket)
+{
+ int server, client;
+};
+/* clang-format on */
+
+FIXTURE_VARIANT(unix_socket)
+{
+ int type;
+ bool domain_both;
+ bool domain_parent;
+ bool domain_child;
+ bool connect_to_parent;
+};
+
+/* clang-format off */
+FIXTURE_VARIANT_ADD(unix_socket, allow_without_domain) {
+ /* clang-format on */
+ .type = SOCK_STREAM, .domain_both = false,
+ .domain_parent = false, .domain_child = false,
+ .connect_to_parent = true,
+};
+
+/* clang-format off */
+FIXTURE_VARIANT_ADD(unix_socket, deny_child_connection_with_domain) {
+ /* clang-format on */
+ .type = SOCK_STREAM, .domain_both = false, .domain_parent = false,
+ .domain_child = true, .connect_to_parent = true,
+};
+
+/* clang-format off */
+FIXTURE_VARIANT_ADD(unix_socket, allow_child_connection_and_parent_domain) {
+ /* clang-format on */
+ .type = SOCK_STREAM, .domain_both = false, .domain_parent = true,
+ .domain_child = false, .connect_to_parent = true,
+};
+
+/* clang-format off */
+FIXTURE_VARIANT_ADD(unix_socket, deny_parent_connection_with_domain) {
+ /* clang-format on */
+ .type = SOCK_STREAM, .domain_both = false,
+ .domain_parent = true, .domain_child = false,
+ .connect_to_parent = false,
+};
+
+/* clang-format off */
+FIXTURE_VARIANT_ADD(unix_socket, allow_parent_connection_without_domain) {
+ /* clang-format on */
+ .type = SOCK_STREAM, .domain_both = false,
+ .domain_parent = false, .domain_child = false,
+ .connect_to_parent = false,
+};
+
+/* clang-format off */
+FIXTURE_VARIANT_ADD(unix_socket, allow_parent_connection_with_sibling_domain) {
+ /* clang-format on */
+ .type = SOCK_STREAM, .domain_both = true,
+ .domain_parent = false, .domain_child = false,
+ .connect_to_parent = false,
+};
+
+/* clang-format off */
+FIXTURE_VARIANT_ADD(unix_socket, dgrm_allow_without_domain) {
+ /* clang-format on */
+ .type = SOCK_DGRAM, .domain_both = false,
+ .domain_parent = false, .domain_child = false,
+ .connect_to_parent = true,
+};
+
+/* clang-format off */
+FIXTURE_VARIANT_ADD(unix_socket, dgrm_deny_child_connection_with_domain) {
+ /* clang-format on */
+ .type = SOCK_DGRAM, .domain_both = false, .domain_parent = false,
+ .domain_child = true, .connect_to_parent = true,
+};
+
+/* clang-format off */
+FIXTURE_VARIANT_ADD(unix_socket, dgrm_allow_child_with_sibling_domain) {
+ /* clang-format on */
+ .type = SOCK_DGRAM, .domain_both = true,
+ .domain_parent = false, .domain_child = false,
+ .connect_to_parent = true,
+};
+
+/* clang-format off */
+FIXTURE_VARIANT_ADD(unix_socket, dgrm_deny_parent_connection_with_domain) {
+ /* clang-format on */
+ .type = SOCK_DGRAM, .domain_both = false,
+ .domain_parent = true, .domain_child = false,
+ .connect_to_parent = false,
+};
+
+/* clang-format off */
+FIXTURE_VARIANT_ADD(unix_socket, dgrm_allow_parent_connection_with_child_domain) {
+ /* clang-format on */
+ .type = SOCK_DGRAM, .domain_both = false,
+ .domain_parent = false, .domain_child = true,
+ .connect_to_parent = false,
+};
+
+FIXTURE_SETUP(unix_socket)
+{
+}
+
+FIXTURE_TEARDOWN(unix_socket)
+{
+ close(self->server);
+ close(self->client);
+}
+
+/* Test UNIX_STREAM_CONNECT for parent and child. */
+TEST_F(unix_socket, abstract_unix_socket)
+{
+ int status;
+ pid_t child;
+ socklen_t addrlen;
+ int sock_len = 5;
+ struct sockaddr_un addr = {
+ .sun_family = AF_UNIX,
+ };
+ const char sun_path[8] = "\0test";
+ bool can_connect_to_parent, can_connect_to_child;
+ int err;
+ int pipe_child[2], pipe_parent[2];
+ char buf_parent;
+ /*
+ * can_connect_to_child is true if a parent process can connect to its
+ * child process. The parent process is not isolated from the child
+ * with a dedicated Landlock domain.
+ */
+ can_connect_to_child = !variant->domain_parent;
+ /*
+ * can_connect_to_parent is true if a child process can connect to its parent
+ * process. This depends on the child process is not isolated from
+ * the parent with a dedicated Landlock domain.
+ */
+ can_connect_to_parent = !variant->domain_child;
+
+ ASSERT_EQ(0, pipe2(pipe_child, O_CLOEXEC));
+ ASSERT_EQ(0, pipe2(pipe_parent, O_CLOEXEC));
+ if (variant->domain_both) {
+ create_unix_domain(_metadata);
+ if (!__test_passed(_metadata))
+ return;
+ }
+
+ addrlen = offsetof(struct sockaddr_un, sun_path) + sock_len;
+ memcpy(&addr.sun_path, sun_path, sock_len);
+
+ child = fork();
+ ASSERT_LE(0, child);
+ if (child == 0) {
+ int child_ret;
+ char buf_child;
+
+ ASSERT_EQ(0, close(pipe_parent[1]));
+ ASSERT_EQ(0, close(pipe_child[0]));
+ if (variant->domain_child)
+ create_unix_domain(_metadata);
+
+ /* Waits for the parent to be in a domain, if any. */
+ ASSERT_EQ(1, read(pipe_parent[0], &buf_child, 1));
+
+ /* create a socket for child process */
+ if (variant->connect_to_parent) {
+ self->client = socket(AF_UNIX, variant->type, 0);
+ ASSERT_NE(-1, self->client);
+
+ ASSERT_EQ(1, read(pipe_parent[0], &buf_child, 1));
+ child_ret = connect(self->client,
+ (struct sockaddr *)&addr, addrlen);
+ if (can_connect_to_parent) {
+ EXPECT_EQ(0, child_ret);
+ } else {
+ EXPECT_EQ(-1, child_ret);
+ EXPECT_EQ(EPERM, errno);
+ }
+ } else {
+ /* child process should create a listening socket */
+ self->server = socket(AF_UNIX, variant->type, 0);
+ ASSERT_NE(-1, self->server);
+
+ ASSERT_EQ(0, bind(self->server,
+ (struct sockaddr *)&addr, addrlen));
+ if (variant->type == SOCK_STREAM) {
+ err = listen(self->server, 32);
+ ASSERT_EQ(0, err);
+ }
+ /* signal to parent that child is listening */
+ ASSERT_EQ(1, write(pipe_child[1], ".", 1));
+ /* wait to connect */
+ ASSERT_EQ(1, read(pipe_parent[0], &buf_child, 1));
+ }
+ _exit(_metadata->exit_code);
+ return;
+ }
+
+ ASSERT_EQ(0, close(pipe_child[1]));
+ ASSERT_EQ(0, close(pipe_parent[0]));
+
+ if (variant->domain_parent)
+ create_unix_domain(_metadata);
+
+ /* Signals that the parent is in a domain, if any. */
+ ASSERT_EQ(1, write(pipe_parent[1], ".", 1));
+
+ if (!variant->connect_to_parent) {
+ self->client = socket(AF_UNIX, variant->type, 0);
+ ASSERT_NE(-1, self->client);
+ /* Waits for the child to listen */
+ ASSERT_EQ(1, read(pipe_child[0], &buf_parent, 1));
+ err = connect(self->client, (struct sockaddr *)&addr, addrlen);
+ if (can_connect_to_child) {
+ EXPECT_EQ(0, err);
+ } else {
+ EXPECT_EQ(-1, err);
+ EXPECT_EQ(EPERM, errno);
+ }
+ ASSERT_EQ(1, write(pipe_parent[1], ".", 1));
+ } else {
+ self->server = socket(AF_UNIX, variant->type, 0);
+ ASSERT_NE(-1, self->server);
+ ASSERT_EQ(0, bind(self->server, (struct sockaddr *)&addr,
+ addrlen));
+ if (variant->type == SOCK_STREAM) {
+ err = listen(self->server, 32);
+ ASSERT_EQ(0, err);
+ }
+ /* signal to child that parent is listening */
+ ASSERT_EQ(1, write(pipe_parent[1], ".", 1));
+ }
+
+ ASSERT_EQ(child, waitpid(child, &status, 0));
+
+ if (WIFSIGNALED(status) || !WIFEXITED(status) ||
+ WEXITSTATUS(status) != EXIT_SUCCESS)
+ _metadata->exit_code = KSFT_FAIL;
+}
+
TEST_HARNESS_MAIN