virt: rework container detection logic

Instead of accessing /proc/1/environ directly, trying to read the
$container variable from it, let's make PID 1 save the contents of that
variable to /run/systemd/container. This allows us to detect containers
without the need for CAP_SYS_PTRACE, which allows us to drop it from a
number of daemons and from the file capabilities of systemd-detect-virt.

Also, don't consider chroot a container technology anymore. After all,
we don't consider file system namespaces container technology anymore,
and hence chroot() should be considered a container even less.

1 files changed, 30 insertions, 18 deletions

diff --git a/src/shared/virt.c b/src/shared/virt.c
index 0db0514dd3..1e227c5fbd 100644
--- a/src/shared/virt.c
+++ b/src/shared/virt.c
@@ -217,8 +217,8 @@ int detect_container(const char **id) {
         static thread_local int cached_found = -1;
         static thread_local const char *cached_id = NULL;
 
-        _cleanup_free_ char *e = NULL;
-        const char *_id = NULL;
+        _cleanup_free_ char *m = NULL;
+        const char *_id = NULL, *e = NULL;
         int r;
 
         if (_likely_(cached_found >= 0)) {
@@ -229,17 +229,6 @@ int detect_container(const char **id) {
                 return cached_found;
         }
 
-        /* Unfortunately many of these operations require root access
-         * in one way or another */
-
-        r = running_in_chroot();
-        if (r < 0)
-                return r;
-        if (r > 0) {
-                _id = "chroot";
-                goto finish;
-        }
-
         /* /proc/vz exists in container and outside of the container,
          * /proc/bc only outside of the container. */
         if (access("/proc/vz", F_OK) >= 0 &&
@@ -249,11 +238,32 @@ int detect_container(const char **id) {
                 goto finish;
         }
 
-        r = getenv_for_pid(1, "container", &e);
-        if (r < 0)
-                return r;
-        if (r == 0)
-                goto finish;
+        if (getpid() == 1) {
+                /* If we are PID 1 we can just check our own
+                 * environment variable */
+
+                e = getenv("container");
+                if (isempty(e)) {
+                        r = 0;
+                        goto finish;
+                }
+        } else {
+
+                /* Otherwise, PID 1 dropped this information into a
+                 * file in /run. This is better than accessing
+                 * /proc/1/environ, since we don't need CAP_SYS_PTRACE
+                 * for that. */
+
+                r = read_one_line_file("/run/systemd/container", &m);
+                if (r == -ENOENT) {
+                        r = 0;
+                        goto finish;
+                }
+                if (r < 0)
+                        return r;
+
+                e = m;
+        }
 
         /* We only recognize a selected few here, since we want to
          * enforce a redacted namespace */
@@ -266,6 +276,8 @@ int detect_container(const char **id) {
         else
                 _id = "other";
 
+        r = 1;
+
 finish:
         cached_found = r;