core: block PTRACE_POKE{TEXT,DATA} when RestrictFileSystemAccess= is enabled

PTRACE_POKETEXT and PTRACE_POKEDATA use FOLL_FORCE to forcibly write
into a tracee's memory. That includes executable read-only pages. If we
restrict execution to trusted binaries PID 1 must enforce that it's
impossible to abuse this to get untrusted code execution. Note, that
usermodehelpers are not subject to PID 1's seccomp filters.

Install a ptrace seccomp filter from PID 1 if RestrictFileSystemAccess=
is requested. Every process it spawns will inherit it.

Signed-off-by: Christian Brauner (Amutable) <brauner@kernel.org>

src/core/bpf-restrict-fsaccess.c

@@ -13,6 +13,7 @@
 #include "lsm-util.h"
 #include "manager.h"
 #include "memory-util.h"
+#include "seccomp-util.h"
 #include "serialize.h"
 #include "string-table.h"
 
@@ -123,6 +124,26 @@ int bpf_restrict_fsaccess_check_prerequisites(void) {
         return 0;
 }
 
+/* The seccomp filter survives execve() so ensure that we're not installing another filter. */
+static int restrict_fsaccess_install_ptrace_filter(Manager *m) {
+        int r;
+
+        assert(m);
+
+        if (m->restrict_fsaccess_ptrace_filter)
+                return 0;
+
+        r = seccomp_restrict_ptrace();
+        if (r == -EOPNOTSUPP)
+                return log_error_errno(r, "bpf-restrict-fsaccess: seccomp support is not available, cannot block "
+                                       "ptrace() code injection.");
+        if (r < 0)
+                return log_error_errno(r, "bpf-restrict-fsaccess: Failed to install the ptrace() seccomp filter: %m");
+
+        m->restrict_fsaccess_ptrace_filter = true;
+        return 0;
+}
+
 static int get_root_s_dev(uint32_t *ret) {
         struct stat st;
 
@@ -442,6 +463,10 @@ int bpf_restrict_fsaccess_setup(Manager *m) {
                                 return r;
                 }
 
+                /* v261 serialized the link FDs without installing the filter. Enforcement is active already,
+                 * so only complain. */
+                (void) restrict_fsaccess_install_ptrace_filter(m);
+
                 return 0;
         }
 
@@ -449,6 +474,10 @@ int bpf_restrict_fsaccess_setup(Manager *m) {
         if (r < 0)
                 return r;
 
+        r = restrict_fsaccess_install_ptrace_filter(m);
+        if (r < 0)
+                return r;
+
         r = bpf_restrict_fsaccess_prepare(&obj);
         if (r < 0)
                 return r;
@@ -548,7 +577,15 @@ int bpf_restrict_fsaccess_serialize(Manager *m, FILE *f, FDSet *fds) {
         assert(f);
         assert(fds);
 
-        if (!MANAGER_IS_SYSTEM(m) || m->restrict_filesystem_access <= RESTRICT_FILESYSTEM_ACCESS_NO)
+        if (!MANAGER_IS_SYSTEM(m))
+                return 0;
+
+        /* Process state, not policy state: the filter outlives the setting being turned off. */
+        r = serialize_bool_elide(f, "restrict-fsaccess-ptrace-filter", m->restrict_fsaccess_ptrace_filter);
+        if (r < 0)
+                return r;
+
+        if (m->restrict_filesystem_access <= RESTRICT_FILESYSTEM_ACCESS_NO)
                 return 0;
 
         FOREACH_ELEMENT(fd, m->restrict_fsaccess_link_fds) {

src/core/manager-serialize.c

@@ -519,7 +519,7 @@ static void manager_deserialize_gid_refs_one(Manager *m, const char *value) {
 
 static void deserialize_restrict_fsaccess(Manager *m, const char *l, FDSet *fds) {
         const char *val;
-        int fd;
+        int fd, r;
 
         FOREACH_ELEMENT(name, restrict_fsaccess_link_names) {
                 val = startswith(l, *name);
@@ -537,6 +537,17 @@ static void deserialize_restrict_fsaccess(Manager *m, const char *l, FDSet *fds)
                 return;
         }
 
+        val = startswith(l, "restrict-fsaccess-ptrace-filter=");
+        if (val) {
+                r = parse_boolean(val);
+                if (r < 0)
+                        log_warning_errno(r, "bpf-restrict-fsaccess: Failed to parse ptrace filter state '%s', "
+                                          "ignoring: %m", val);
+                else
+                        m->restrict_fsaccess_ptrace_filter = r;
+                return;
+        }
+
         val = startswith(l, "restrict-fsaccess-bss-map=");
         if (!val)
                 return;

src/core/manager.h

@@ -514,6 +514,9 @@ typedef struct Manager {
         int restrict_fsaccess_link_fds[_RESTRICT_FILESYSTEM_ACCESS_LINK_MAX];
         int restrict_fsaccess_bss_map_fd;
 
+        /* Whether the ptrace seccomp filter is installed in this process; it survives execve() */
+        bool restrict_fsaccess_ptrace_filter;
+
         /* Allow users to configure a rate limit for Reload()/Reexecute() operations */
         RateLimit reload_reexec_ratelimit;
         /* Dump*() are slow, so always rate limit them to 10 per 10 minutes */