File 0020-core-when-creating-the-socket-fds-for-a-so... of Package systemd

Overview Repositories Revisions Requests Users Attributes Meta

File 0020-core-when-creating-the-socket-fds-for-a-socket-unit-.patch of Package systemd (Revision 2a0ef622b860d89b09df9bc0f43c3432)

Currently displaying revision 2a0ef622b860d89b09df9bc0f43c3432 , Show latest

From c33ee30f6fbf44c109cb3b015dbb4b723cf53b19 Mon Sep 17 00:00:00 2001
From: Lennart Poettering <lennart@poettering.net>
Date: Thu, 7 Sep 2017 11:17:43 +0200
Subject: [PATCH 20/51] core: when creating the socket fds for a socket unit,
 join socket's cgroup first

Let's make sure that a socket unit's IPAddressAllow=/IPAddressDeny=
settings are in effect on all socket fds associated with it. In order to
make this happen we need to make sure the cgroup the fds are associated
with are the socket unit's cgroup. The only way to do that is invoking
socket()+accept() in them. Since we really don't want to migrate PID 1
around we do this by forking off a helper process, which invokes
socket()/accept() and sends the newly created fd to PID 1. Ugly, but
works, and there's apparently no better way right now.

This generalizes forking off per-unit helper processes in a new function
unit_fork_helper_process(), which is then also used by the NSS chown()
code of socket units.

(cherry picked from commit a79279c7fd9f43680da7d5cac382981cf7714f52)

[fbui: adjust context]
---
 src/core/socket.c | 257 +++++++++++++++++++++++++++++++++++++++++++-----------
 src/core/unit.c   |  40 +++++++++
 src/core/unit.h   |   2 +
 3 files changed, 248 insertions(+), 51 deletions(-)

diff --git a/src/core/socket.c b/src/core/socket.c
index 63238ac198..a69e2283e5 100644
--- a/src/core/socket.c
+++ b/src/core/socket.c
@@ -29,6 +29,7 @@
 #include <linux/sctp.h>
 
 #include "alloc-util.h"
+#include "bpf-firewall.h"
 #include "bus-error.h"
 #include "bus-util.h"
 #include "copy.h"
@@ -37,6 +38,7 @@
 #include "exit-status.h"
 #include "fd-util.h"
 #include "format-util.h"
+#include "in-addr-util.h"
 #include "io-util.h"
 #include "label.h"
 #include "log.h"
@@ -56,7 +58,6 @@
 #include "unit-name.h"
 #include "unit.h"
 #include "user-util.h"
-#include "in-addr-util.h"
 
 struct SocketPeer {
         unsigned n_ref;
@@ -1435,6 +1436,102 @@ no_label:
         return 0;
 }
 
+static int socket_address_listen_do(
+                Socket *s,
+                const SocketAddress *address,
+                const char *label) {
+
+        assert(s);
+        assert(address);
+
+        return socket_address_listen(
+                        address,
+                        SOCK_CLOEXEC|SOCK_NONBLOCK,
+                        s->backlog,
+                        s->bind_ipv6_only,
+                        s->bind_to_device,
+                        s->reuse_port,
+                        s->free_bind,
+                        s->transparent,
+                        s->directory_mode,
+                        s->socket_mode,
+                        label);
+}
+
+static int socket_address_listen_in_cgroup(
+                Socket *s,
+                const SocketAddress *address,
+                const char *label) {
+
+        _cleanup_close_pair_ int pair[2] = { -1, -1 };
+        int fd, r;
+        pid_t pid;
+
+        assert(s);
+        assert(address);
+
+        /* This is a wrapper around socket_address_listen(), that forks off a helper process inside the socket's cgroup
+         * in which the socket is actually created. This way we ensure the socket is actually properly attached to the
+         * unit's cgroup for the purpose of BPF filtering and such. */
+
+        if (!IN_SET(address->sockaddr.sa.sa_family, AF_INET, AF_INET6))
+                goto shortcut; /* BPF filtering only applies to IPv4 + IPv6, shortcut things for other protocols */
+
+        r = bpf_firewall_supported();
+        if (r < 0)
+                return r;
+        if (r == 0) /* If BPF firewalling isn't supported anyway — there's no point in this forking complexity */
+                goto shortcut;
+
+        if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, pair) < 0)
+                return log_unit_error_errno(UNIT(s), errno, "Failed to create communication channel: %m");
+
+        r = unit_fork_helper_process(UNIT(s), &pid);
+        if (r < 0)
+                return log_unit_error_errno(UNIT(s), r, "Failed to fork off listener stub process: %m");
+        if (r == 0) {
+                /* Child */
+
+                pair[0] = safe_close(pair[0]);
+
+                fd = socket_address_listen_do(s, address, label);
+                if (fd < 0) {
+                        log_unit_error_errno(UNIT(s), fd, "Failed to create listening socket: %m");
+                        _exit(EXIT_FAILURE);
+                }
+
+                r = send_one_fd(pair[1], fd, 0);
+                if (r < 0) {
+                        log_unit_error_errno(UNIT(s), r, "Failed to send listening socket to parent: %m");
+                        _exit(EXIT_FAILURE);
+                }
+
+                _exit(EXIT_SUCCESS);
+        }
+
+        pair[1] = safe_close(pair[1]);
+        fd = receive_one_fd(pair[0], 0);
+
+        /* We synchronously wait for the helper, as it shouldn't be slow */
+        r = wait_for_terminate_and_warn("listen-cgroup-helper", pid, false);
+        if (r < 0) {
+                safe_close(fd);
+                return r;
+        }
+
+        if (fd < 0)
+                return log_unit_error_errno(UNIT(s), fd, "Failed to receive listening socket: %m");
+
+        return fd;
+
+shortcut:
+        fd = socket_address_listen_do(s, address, label);
+        if (fd < 0)
+                return log_error_errno(fd, "Failed to create listening socket: %m");
+
+        return fd;
+}
+
 static int socket_open_fds(Socket *s) {
         _cleanup_(mac_selinux_freep) char *label = NULL;
         bool know_label = false;
@@ -1478,18 +1575,7 @@ static int socket_open_fds(Socket *s) {
                                 break;
                         }
 
-                        r = socket_address_listen(
-                                        &p->address,
-                                        SOCK_CLOEXEC|SOCK_NONBLOCK,
-                                        s->backlog,
-                                        s->bind_ipv6_only,
-                                        s->bind_to_device,
-                                        s->reuse_port,
-                                        s->free_bind,
-                                        s->transparent,
-                                        s->directory_mode,
-                                        s->socket_mode,
-                                        label);
+                        r = socket_address_listen_in_cgroup(s, &p->address, label);
                         if (r < 0)
                                 goto rollback;
 
@@ -1829,27 +1915,23 @@ static int socket_chown(Socket *s, pid_t *_pid) {
         /* We have to resolve the user names out-of-process, hence
          * let's fork here. It's messy, but well, what can we do? */
 
-        pid = fork();
-        if (pid < 0)
-                return -errno;
-
-        if (pid == 0) {
-                SocketPort *p;
+        r = unit_fork_helper_process(UNIT(s), &pid);
+        if (r < 0)
+                return r;
+        if (r == 0) {
                 uid_t uid = UID_INVALID;
                 gid_t gid = GID_INVALID;
-                int ret;
+                SocketPort *p;
 
-                (void) default_signals(SIGNALS_CRASH_HANDLER, SIGNALS_IGNORE, -1);
-                (void) ignore_signals(SIGPIPE, -1);
-                log_forget_fds();
+                /* Child */
 
                 if (!isempty(s->user)) {
                         const char *user = s->user;
 
                         r = get_user_creds(&user, &uid, &gid, NULL, NULL);
                         if (r < 0) {
-                                ret = EXIT_USER;
-                                goto fail_child;
+                                log_unit_error_errno(UNIT(s), r, "Failed to resolve user %s: %m", user);
+                                _exit(EXIT_USER);
                         }
                 }
 
@@ -1858,8 +1940,8 @@ static int socket_chown(Socket *s, pid_t *_pid) {
 
                         r = get_group_creds(&group, &gid);
                         if (r < 0) {
-                                ret = EXIT_GROUP;
-                                goto fail_child;
+                                log_unit_error_errno(UNIT(s), r, "Failed to resolve group %s: %m", group);
+                                _exit(EXIT_GROUP);
                         }
                 }
 
@@ -1875,19 +1957,12 @@ static int socket_chown(Socket *s, pid_t *_pid) {
                                 continue;
 
                         if (chown(path, uid, gid) < 0) {
-                                r = -errno;
-                                ret = EXIT_CHOWN;
-                                goto fail_child;
+                                log_unit_error_errno(UNIT(s), errno, "Failed to chown(): %m");
+                                _exit(EXIT_CHOWN);
                         }
                 }
 
-                _exit(0);
-
-        fail_child:
-                log_open();
-                log_error_errno(r, "Failed to chown socket at step %s: %m", exit_status_to_string(ret, EXIT_STATUS_SYSTEMD));
-
-                _exit(ret);
+                _exit(EXIT_SUCCESS);
         }
 
         r = unit_watch_pid(UNIT(s), pid);
@@ -2699,6 +2774,97 @@ _pure_ static bool socket_check_gc(Unit *u) {
         return s->n_connections > 0;
 }
 
+static int socket_accept_do(Socket *s, int fd) {
+        int cfd;
+
+        assert(s);
+        assert(fd >= 0);
+
+        for (;;) {
+                cfd = accept4(fd, NULL, NULL, SOCK_NONBLOCK);
+                if (cfd < 0) {
+                        if (errno == EINTR)
+                                continue;
+
+                        return -errno;
+                }
+
+                break;
+        }
+
+        return cfd;
+}
+
+static int socket_accept_in_cgroup(Socket *s, SocketPort *p, int fd) {
+        _cleanup_close_pair_ int pair[2] = { -1, -1 };
+        int cfd, r;
+        pid_t pid;
+
+        assert(s);
+        assert(p);
+        assert(fd >= 0);
+
+        /* Similar to socket_address_listen_in_cgroup(), but for accept() rathern than socket(): make sure that any
+         * connection socket is also properly associated with the cgroup. */
+
+        if (!IN_SET(p->address.sockaddr.sa.sa_family, AF_INET, AF_INET6))
+                goto shortcut;
+
+        r = bpf_firewall_supported();
+        if (r < 0)
+                return r;
+        if (r == 0)
+                goto shortcut;
+
+        if (socketpair(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0, pair) < 0)
+                return log_unit_error_errno(UNIT(s), errno, "Failed to create communication channel: %m");
+
+        r = unit_fork_helper_process(UNIT(s), &pid);
+        if (r < 0)
+                return log_unit_error_errno(UNIT(s), r, "Failed to fork off accept stub process: %m");
+        if (r == 0) {
+                /* Child */
+
+                pair[0] = safe_close(pair[0]);
+
+                cfd = socket_accept_do(s, fd);
+                if (cfd < 0) {
+                        log_unit_error_errno(UNIT(s), cfd, "Failed to accept connection socket: %m");
+                        _exit(EXIT_FAILURE);
+                }
+
+                r = send_one_fd(pair[1], cfd, 0);
+                if (r < 0) {
+                        log_unit_error_errno(UNIT(s), r, "Failed to send connection socket to parent: %m");
+                        _exit(EXIT_FAILURE);
+                }
+
+                _exit(EXIT_SUCCESS);
+        }
+
+        pair[1] = safe_close(pair[1]);
+        cfd = receive_one_fd(pair[0], 0);
+
+        /* We synchronously wait for the helper, as it shouldn't be slow */
+        r = wait_for_terminate_and_warn("accept-cgroup-helper", pid, false);
+        if (r < 0) {
+                safe_close(cfd);
+                return r;
+        }
+
+        if (cfd < 0)
+                return log_unit_error_errno(UNIT(s), cfd, "Failed to receive connection socket: %m");
+
+        return cfd;
+
+shortcut:
+        cfd = socket_accept_do(s, fd);
+        if (cfd < 0)
+                return log_unit_error_errno(UNIT(s), cfd, "Failed to accept connection socket: %m");
+
+        return cfd;
+}
+
 static int socket_dispatch_io(sd_event_source *source, int fd, uint32_t revents, void *userdata) {
         SocketPort *p = userdata;
         int cfd = -1;
@@ -2724,20 +2890,9 @@ static int socket_dispatch_io(sd_event_source *source, int fd, uint32_t revents,
             p->type == SOCKET_SOCKET &&
             socket_address_can_accept(&p->address)) {
 
-                for (;;) {
-
-                        cfd = accept4(fd, NULL, NULL, SOCK_NONBLOCK);
-                        if (cfd < 0) {
-
-                                if (errno == EINTR)
-                                        continue;
-
-                                log_unit_error_errno(UNIT(p->socket), errno, "Failed to accept socket: %m");
-                                goto fail;
-                        }
-
-                        break;
-                }
+                cfd = socket_accept_in_cgroup(p->socket, p, fd);
+                if (cfd < 0)
+                        goto fail;
 
                 socket_apply_socket_options(p->socket, cfd);
         }
diff --git a/src/core/unit.c b/src/core/unit.c
index 911e663a57..cacf35d0ea 100644
--- a/src/core/unit.c
+++ b/src/core/unit.c
@@ -4443,3 +4443,43 @@ int unit_acquire_invocation_id(Unit *u) {
 
         return 0;
 }
+
+int unit_fork_helper_process(Unit *u, pid_t *ret) {
+        pid_t pid;
+        int r;
+
+        assert(u);
+        assert(ret);
+
+        /* Forks off a helper process and makes sure it is a member of the unit's cgroup. Returns == 0 in the child,
+         * and > 0 in the parent. The pid parameter is always filled in with the child's PID. */
+
+        (void) unit_realize_cgroup(u);
+
+        pid = fork();
+        if (pid < 0)
+                return -errno;
+
+        if (pid == 0) {
+
+                (void) default_signals(SIGNALS_CRASH_HANDLER, SIGNALS_IGNORE, -1);
+                (void) ignore_signals(SIGPIPE, -1);
+
+                log_close();
+                log_open();
+
+                if (u->cgroup_path) {
+                        r = cg_attach_everywhere(u->manager->cgroup_supported, u->cgroup_path, 0, NULL, NULL);
+                        if (r < 0) {
+                                log_unit_error_errno(u, r, "Failed to join unit cgroup %s: %m", u->cgroup_path);
+                                _exit(EXIT_CGROUP);
+                        }
+                }
+
+                *ret = getpid();
+                return 0;
+        }
+
+        *ret = pid;
+        return 1;
+}
diff --git a/src/core/unit.h b/src/core/unit.h
index 97eb3f3148..9d2b9f4cac 100644
--- a/src/core/unit.h
+++ b/src/core/unit.h
@@ -685,6 +685,8 @@ int unit_acquire_invocation_id(Unit *u);
 
 bool unit_shall_confirm_spawn(Unit *u);
 
+int unit_fork_helper_process(Unit *u, pid_t *ret);
+
 /* Macros which append UNIT= or USER_UNIT= to the message */
 
 #define log_unit_full(unit, level, error, ...)                          \
-- 
2.16.4

Places

File 0020-core-when-creating-the-socket-fds-for-a-socket-unit-.patch of Package systemd (Revision 2a0ef622b860d89b09df9bc0f43c3432)

Places