[thirdparty/systemd.git] / src / coredump / coredump.c

/***
  This file is part of systemd.

  Copyright 2012 Lennart Poettering

  systemd is free software; you can redistribute it and/or modify it
  under the terms of the GNU Lesser General Public License as published by
  the Free Software Foundation; either version 2.1 of the License, or
  (at your option) any later version.

  systemd is distributed in the hope that it will be useful, but
  WITHOUT ANY WARRANTY; without even the implied warranty of
  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
  Lesser General Public License for more details.

  You should have received a copy of the GNU Lesser General Public License
  along with systemd; If not, see <http://www.gnu.org/licenses/>.
***/

#include <errno.h>
#include <stdio.h>
#include <sys/prctl.h>
#include <sys/xattr.h>
#include <unistd.h>

#ifdef HAVE_ELFUTILS
#include <dwarf.h>
#include <elfutils/libdwfl.h>
#endif

#include "sd-daemon.h"
#include "sd-journal.h"
#include "sd-login.h"
#include "sd-messages.h"

#include "acl-util.h"
#include "alloc-util.h"
#include "capability-util.h"
#include "cgroup-util.h"
#include "compress.h"
#include "conf-parser.h"
#include "copy.h"
#include "coredump-vacuum.h"
#include "dirent-util.h"
#include "escape.h"
#include "fd-util.h"
#include "fileio.h"
#include "fs-util.h"
#include "io-util.h"
#include "journald-native.h"
#include "log.h"
#include "macro.h"
#include "missing.h"
#include "mkdir.h"
#include "parse-util.h"
#include "process-util.h"
#include "socket-util.h"
#include "special.h"
#include "stacktrace.h"
#include "string-table.h"
#include "string-util.h"
#include "strv.h"
#include "user-util.h"
#include "util.h"

/* The maximum size up to which we process coredumps */
#define PROCESS_SIZE_MAX ((uint64_t) (2LLU*1024LLU*1024LLU*1024LLU))

/* The maximum size up to which we leave the coredump around on disk */
#define EXTERNAL_SIZE_MAX PROCESS_SIZE_MAX

/* The maximum size up to which we store the coredump in the journal */
#define JOURNAL_SIZE_MAX ((size_t) (767LU*1024LU*1024LU))

/* Make sure to not make this larger than the maximum journal entry
 * size. See DATA_SIZE_MAX in journald-native.c. */
assert_cc(JOURNAL_SIZE_MAX <= DATA_SIZE_MAX);

enum {
        /* We use this as array indexes for a couple of special fields we use for naming coredumping files, and
         * attaching xattrs */
        CONTEXT_PID,
        CONTEXT_UID,
        CONTEXT_GID,
        CONTEXT_SIGNAL,
        CONTEXT_TIMESTAMP,
        CONTEXT_RLIMIT,
        CONTEXT_COMM,
        CONTEXT_EXE,
        _CONTEXT_MAX
};

typedef enum CoredumpStorage {
        COREDUMP_STORAGE_NONE,
        COREDUMP_STORAGE_EXTERNAL,
        COREDUMP_STORAGE_JOURNAL,
        _COREDUMP_STORAGE_MAX,
        _COREDUMP_STORAGE_INVALID = -1
} CoredumpStorage;

static const char* const coredump_storage_table[_COREDUMP_STORAGE_MAX] = {
        [COREDUMP_STORAGE_NONE] = "none",
        [COREDUMP_STORAGE_EXTERNAL] = "external",
        [COREDUMP_STORAGE_JOURNAL] = "journal",
};

DEFINE_PRIVATE_STRING_TABLE_LOOKUP(coredump_storage, CoredumpStorage);
static DEFINE_CONFIG_PARSE_ENUM(config_parse_coredump_storage, coredump_storage, CoredumpStorage, "Failed to parse storage setting");

static CoredumpStorage arg_storage = COREDUMP_STORAGE_EXTERNAL;
static bool arg_compress = true;
static uint64_t arg_process_size_max = PROCESS_SIZE_MAX;
static uint64_t arg_external_size_max = EXTERNAL_SIZE_MAX;
static uint64_t arg_journal_size_max = JOURNAL_SIZE_MAX;
static uint64_t arg_keep_free = (uint64_t) -1;
static uint64_t arg_max_use = (uint64_t) -1;

static int parse_config(void) {
        static const ConfigTableItem items[] = {
                { "Coredump", "Storage",          config_parse_coredump_storage,  0, &arg_storage           },
                { "Coredump", "Compress",         config_parse_bool,              0, &arg_compress          },
                { "Coredump", "ProcessSizeMax",   config_parse_iec_uint64,        0, &arg_process_size_max  },
                { "Coredump", "ExternalSizeMax",  config_parse_iec_uint64,        0, &arg_external_size_max },
                { "Coredump", "JournalSizeMax",   config_parse_iec_size,          0, &arg_journal_size_max  },
                { "Coredump", "KeepFree",         config_parse_iec_uint64,        0, &arg_keep_free         },
                { "Coredump", "MaxUse",           config_parse_iec_uint64,        0, &arg_max_use           },
                {}
        };

        return config_parse_many_nulstr(PKGSYSCONFDIR "/coredump.conf",
                                 CONF_PATHS_NULSTR("systemd/coredump.conf.d"),
                                 "Coredump\0",
                                 config_item_table_lookup, items,
                                 false, NULL);
}

static inline uint64_t storage_size_max(void) {
        return arg_storage == COREDUMP_STORAGE_EXTERNAL ? arg_external_size_max : arg_journal_size_max;
}

static int fix_acl(int fd, uid_t uid) {

#ifdef HAVE_ACL
        _cleanup_(acl_freep) acl_t acl = NULL;
        acl_entry_t entry;
        acl_permset_t permset;
        int r;

        assert(fd >= 0);

        if (uid <= SYSTEM_UID_MAX)
                return 0;

        /* Make sure normal users can read (but not write or delete)
         * their own coredumps */

        acl = acl_get_fd(fd);
        if (!acl)
                return log_error_errno(errno, "Failed to get ACL: %m");

        if (acl_create_entry(&acl, &entry) < 0 ||
            acl_set_tag_type(entry, ACL_USER) < 0 ||
            acl_set_qualifier(entry, &uid) < 0)
                return log_error_errno(errno, "Failed to patch ACL: %m");

        if (acl_get_permset(entry, &permset) < 0 ||
            acl_add_perm(permset, ACL_READ) < 0)
                return log_warning_errno(errno, "Failed to patch ACL: %m");

        r = calc_acl_mask_if_needed(&acl);
        if (r < 0)
                return log_warning_errno(r, "Failed to patch ACL: %m");

        if (acl_set_fd(fd, acl) < 0)
                return log_error_errno(errno, "Failed to apply ACL: %m");
#endif

        return 0;
}

static int fix_xattr(int fd, const char *context[_CONTEXT_MAX]) {

        static const char * const xattrs[_CONTEXT_MAX] = {
                [CONTEXT_PID] = "user.coredump.pid",
                [CONTEXT_UID] = "user.coredump.uid",
                [CONTEXT_GID] = "user.coredump.gid",
                [CONTEXT_SIGNAL] = "user.coredump.signal",
                [CONTEXT_TIMESTAMP] = "user.coredump.timestamp",
                [CONTEXT_COMM] = "user.coredump.comm",
                [CONTEXT_EXE] = "user.coredump.exe",
        };

        int r = 0;
        unsigned i;

        assert(fd >= 0);

        /* Attach some metadata to coredumps via extended
         * attributes. Just because we can. */

        for (i = 0; i < _CONTEXT_MAX; i++) {
                int k;

                if (isempty(context[i]) || !xattrs[i])
                        continue;

                k = fsetxattr(fd, xattrs[i], context[i], strlen(context[i]), XATTR_CREATE);
                if (k < 0 && r == 0)
                        r = -errno;
        }

        return r;
}

#define filename_escape(s) xescape((s), "./ ")

static inline const char *coredump_tmpfile_name(const char *s) {
        return s ? s : "(unnamed temporary file)";
}

static int fix_permissions(
                int fd,
                const char *filename,
                const char *target,
                const char *context[_CONTEXT_MAX],
                uid_t uid) {

        int r;

        assert(fd >= 0);
        assert(target);
        assert(context);

        /* Ignore errors on these */
        (void) fchmod(fd, 0640);
        (void) fix_acl(fd, uid);
        (void) fix_xattr(fd, context);

        if (fsync(fd) < 0)
                return log_error_errno(errno, "Failed to sync coredump %s: %m", coredump_tmpfile_name(filename));

        r = link_tmpfile(fd, filename, target);
        if (r < 0)
                return log_error_errno(r, "Failed to move coredump %s into place: %m", target);

        return 0;
}

static int maybe_remove_external_coredump(const char *filename, uint64_t size) {

        /* Returns 1 if might remove, 0 if will not remove, < 0 on error. */

        if (arg_storage == COREDUMP_STORAGE_EXTERNAL &&
            size <= arg_external_size_max)
                return 0;

        if (!filename)
                return 1;

        if (unlink(filename) < 0 && errno != ENOENT)
                return log_error_errno(errno, "Failed to unlink %s: %m", filename);

        return 1;
}

static int make_filename(const char *context[_CONTEXT_MAX], char **ret) {
        _cleanup_free_ char *c = NULL, *u = NULL, *p = NULL, *t = NULL;
        sd_id128_t boot = {};
        int r;

        assert(context);

        c = filename_escape(context[CONTEXT_COMM]);
        if (!c)
                return -ENOMEM;

        u = filename_escape(context[CONTEXT_UID]);
        if (!u)
                return -ENOMEM;

        r = sd_id128_get_boot(&boot);
        if (r < 0)
                return r;

        p = filename_escape(context[CONTEXT_PID]);
        if (!p)
                return -ENOMEM;

        t = filename_escape(context[CONTEXT_TIMESTAMP]);
        if (!t)
                return -ENOMEM;

        if (asprintf(ret,
                     "/var/lib/systemd/coredump/core.%s.%s." SD_ID128_FORMAT_STR ".%s.%s000000",
                     c,
                     u,
                     SD_ID128_FORMAT_VAL(boot),
                     p,
                     t) < 0)
                return -ENOMEM;

        return 0;
}

static int save_external_coredump(
                const char *context[_CONTEXT_MAX],
                int input_fd,
                char **ret_filename,
                int *ret_node_fd,
                int *ret_data_fd,
                uint64_t *ret_size) {

        _cleanup_free_ char *fn = NULL, *tmp = NULL;
        _cleanup_close_ int fd = -1;
        uint64_t rlimit, max_size;
        struct stat st;
        uid_t uid;
        int r;

        assert(context);
        assert(ret_filename);
        assert(ret_node_fd);
        assert(ret_data_fd);
        assert(ret_size);

        r = parse_uid(context[CONTEXT_UID], &uid);
        if (r < 0)
                return log_error_errno(r, "Failed to parse UID: %m");

        r = safe_atou64(context[CONTEXT_RLIMIT], &rlimit);
        if (r < 0)
                return log_error_errno(r, "Failed to parse resource limit: %s", context[CONTEXT_RLIMIT]);
        if (rlimit < page_size()) {
                /* Is coredumping disabled? Then don't bother saving/processing the coredump.
                 * Anything below PAGE_SIZE cannot give a readable coredump (the kernel uses
                 * ELF_EXEC_PAGESIZE which is not easily accessible, but is usually the same as PAGE_SIZE. */
                log_info("Resource limits disable core dumping for process %s (%s).",
                         context[CONTEXT_PID], context[CONTEXT_COMM]);
                return -EBADSLT;
        }

        /* Never store more than the process configured, or than we actually shall keep or process */
        max_size = MIN(rlimit, MAX(arg_process_size_max, storage_size_max()));

        r = make_filename(context, &fn);
        if (r < 0)
                return log_error_errno(r, "Failed to determine coredump file name: %m");

        mkdir_p_label("/var/lib/systemd/coredump", 0755);

        fd = open_tmpfile_linkable(fn, O_RDWR|O_CLOEXEC, &tmp);
        if (fd < 0)
                return log_error_errno(fd, "Failed to create temporary file for coredump %s: %m", fn);

        r = copy_bytes(input_fd, fd, max_size, false);
        if (r < 0) {
                log_error_errno(r, "Cannot store coredump of %s (%s): %m", context[CONTEXT_PID], context[CONTEXT_COMM]);
                goto fail;
        } else if (r == 1)
                log_struct(LOG_INFO,
                           LOG_MESSAGE("Core file was truncated to %zu bytes.", max_size),
                           "SIZE_LIMIT=%zu", max_size,
                           LOG_MESSAGE_ID(SD_MESSAGE_TRUNCATED_CORE),
                           NULL);

        if (fstat(fd, &st) < 0) {
                log_error_errno(errno, "Failed to fstat core file %s: %m", coredump_tmpfile_name(tmp));
                goto fail;
        }

        if (lseek(fd, 0, SEEK_SET) == (off_t) -1) {
                log_error_errno(errno, "Failed to seek on %s: %m", coredump_tmpfile_name(tmp));
                goto fail;
        }

#if defined(HAVE_XZ) || defined(HAVE_LZ4)
        /* If we will remove the coredump anyway, do not compress. */
        if (arg_compress && !maybe_remove_external_coredump(NULL, st.st_size)) {

                _cleanup_free_ char *fn_compressed = NULL, *tmp_compressed = NULL;
                _cleanup_close_ int fd_compressed = -1;

                fn_compressed = strappend(fn, COMPRESSED_EXT);
                if (!fn_compressed) {
                        log_oom();
                        goto uncompressed;
                }

                fd_compressed = open_tmpfile_linkable(fn_compressed, O_RDWR|O_CLOEXEC, &tmp_compressed);
                if (fd_compressed < 0) {
                        log_error_errno(fd_compressed, "Failed to create temporary file for coredump %s: %m", fn_compressed);
                        goto uncompressed;
                }

                r = compress_stream(fd, fd_compressed, -1);
                if (r < 0) {
                        log_error_errno(r, "Failed to compress %s: %m", coredump_tmpfile_name(tmp_compressed));
                        goto fail_compressed;
                }

                r = fix_permissions(fd_compressed, tmp_compressed, fn_compressed, context, uid);
                if (r < 0)
                        goto fail_compressed;

                /* OK, this worked, we can get rid of the uncompressed version now */
                if (tmp)
                        unlink_noerrno(tmp);

                *ret_filename = fn_compressed;     /* compressed */
                *ret_node_fd = fd_compressed;      /* compressed */
                *ret_data_fd = fd;                 /* uncompressed */
                *ret_size = (uint64_t) st.st_size; /* uncompressed */

                fn_compressed = NULL;
                fd = fd_compressed = -1;

                return 0;

        fail_compressed:
                if (tmp_compressed)
                        (void) unlink(tmp_compressed);
        }

uncompressed:
#endif

        r = fix_permissions(fd, tmp, fn, context, uid);
        if (r < 0)
                goto fail;

        *ret_filename = fn;
        *ret_data_fd = fd;
        *ret_node_fd = -1;
        *ret_size = (uint64_t) st.st_size;

        fn = NULL;
        fd = -1;

        return 0;

fail:
        if (tmp)
                (void) unlink(tmp);
        return r;
}

static int allocate_journal_field(int fd, size_t size, char **ret, size_t *ret_size) {
        _cleanup_free_ char *field = NULL;
        ssize_t n;

        assert(fd >= 0);
        assert(ret);
        assert(ret_size);

        if (lseek(fd, 0, SEEK_SET) == (off_t) -1)
                return log_warning_errno(errno, "Failed to seek: %m");

        field = malloc(9 + size);
        if (!field) {
                log_warning("Failed to allocate memory for coredump, coredump will not be stored.");
                return -ENOMEM;
        }

        memcpy(field, "COREDUMP=", 9);

        n = read(fd, field + 9, size);
        if (n < 0)
                return log_error_errno((int) n, "Failed to read core data: %m");
        if ((size_t) n < size) {
                log_error("Core data too short.");
                return -EIO;
        }

        *ret = field;
        *ret_size = size + 9;

        field = NULL;

        return 0;
}

/* Joins /proc/[pid]/fd/ and /proc/[pid]/fdinfo/ into the following lines:
 * 0:/dev/pts/23
 * pos:    0
 * flags:  0100002
 *
 * 1:/dev/pts/23
 * pos:    0
 * flags:  0100002
 *
 * 2:/dev/pts/23
 * pos:    0
 * flags:  0100002
 * EOF
 */
static int compose_open_fds(pid_t pid, char **open_fds) {
        _cleanup_closedir_ DIR *proc_fd_dir = NULL;
        _cleanup_close_ int proc_fdinfo_fd = -1;
        _cleanup_free_ char *buffer = NULL;
        _cleanup_fclose_ FILE *stream = NULL;
        const char *fddelim = "", *path;
        struct dirent *dent = NULL;
        size_t size = 0;
        int r = 0;

        assert(pid >= 0);
        assert(open_fds != NULL);

        path = procfs_file_alloca(pid, "fd");
        proc_fd_dir = opendir(path);
        if (!proc_fd_dir)
                return -errno;

        proc_fdinfo_fd = openat(dirfd(proc_fd_dir), "../fdinfo", O_DIRECTORY|O_NOFOLLOW|O_CLOEXEC|O_PATH);
        if (proc_fdinfo_fd < 0)
                return -errno;

        stream = open_memstream(&buffer, &size);
        if (!stream)
                return -ENOMEM;

        FOREACH_DIRENT(dent, proc_fd_dir, return -errno) {
                _cleanup_fclose_ FILE *fdinfo = NULL;
                _cleanup_free_ char *fdname = NULL;
                char line[LINE_MAX];
                int fd;

                r = readlinkat_malloc(dirfd(proc_fd_dir), dent->d_name, &fdname);
                if (r < 0)
                        return r;

                fprintf(stream, "%s%s:%s\n", fddelim, dent->d_name, fdname);
                fddelim = "\n";

                /* Use the directory entry from /proc/[pid]/fd with /proc/[pid]/fdinfo */
                fd = openat(proc_fdinfo_fd, dent->d_name, O_NOFOLLOW|O_CLOEXEC|O_RDONLY);
                if (fd < 0)
                        continue;

                fdinfo = fdopen(fd, "re");
                if (fdinfo == NULL) {
                        close(fd);
                        continue;
                }

                FOREACH_LINE(line, fdinfo, break) {
                        fputs(line, stream);
                        if (!endswith(line, "\n"))
                                fputc('\n', stream);
                }
        }

        errno = 0;
        stream = safe_fclose(stream);

        if (errno > 0)
                return -errno;

        *open_fds = buffer;
        buffer = NULL;

        return 0;
}

static int get_process_ns(pid_t pid, const char *namespace, ino_t *ns) {
        const char *p;
        struct stat stbuf;
        _cleanup_close_ int proc_ns_dir_fd;

        p = procfs_file_alloca(pid, "ns");

        proc_ns_dir_fd = open(p, O_DIRECTORY | O_CLOEXEC | O_RDONLY);
        if (proc_ns_dir_fd < 0)
                return -errno;

        if (fstatat(proc_ns_dir_fd, namespace, &stbuf, /* flags */0) < 0)
                return -errno;

        *ns = stbuf.st_ino;
        return 0;
}

static int get_mount_namespace_leader(pid_t pid, pid_t *container_pid) {
        pid_t cpid = pid, ppid = 0;
        ino_t proc_mntns;
        int r = 0;

        r = get_process_ns(pid, "mnt", &proc_mntns);
        if (r < 0)
                return r;

        for (;;) {
                ino_t parent_mntns;

                r = get_process_ppid(cpid, &ppid);
                if (r < 0)
                        return r;

                r = get_process_ns(ppid, "mnt", &parent_mntns);
                if (r < 0)
                        return r;

                if (proc_mntns != parent_mntns)
                        break;

                if (ppid == 1)
                        return -ENOENT;

                cpid = ppid;
        }

        *container_pid = ppid;
        return 0;
}

/* Returns 1 if the parent was found.
 * Returns 0 if there is not a process we can call the pid's
 * container parent (the pid's process isn't 'containerized').
 * Returns a negative number on errors.
 */
static int get_process_container_parent_cmdline(pid_t pid, char** cmdline) {
        int r = 0;
        pid_t container_pid;
        const char *proc_root_path;
        struct stat root_stat, proc_root_stat;

        /* To compare inodes of / and /proc/[pid]/root */
        if (stat("/", &root_stat) < 0)
                return -errno;

        proc_root_path = procfs_file_alloca(pid, "root");
        if (stat(proc_root_path, &proc_root_stat) < 0)
                return -errno;

        /* The process uses system root. */
        if (proc_root_stat.st_ino == root_stat.st_ino) {
                *cmdline = NULL;
                return 0;
        }

        r = get_mount_namespace_leader(pid, &container_pid);
        if (r < 0)
                return r;

        r = get_process_cmdline(container_pid, 0, false, cmdline);
        if (r < 0)
                return r;

        return 1;
}

static int change_uid_gid(const char *context[]) {
        uid_t uid;
        gid_t gid;
        int r;

        r = parse_uid(context[CONTEXT_UID], &uid);
        if (r < 0)
                return r;

        if (uid <= SYSTEM_UID_MAX) {
                const char *user = "systemd-coredump";

                r = get_user_creds(&user, &uid, &gid, NULL, NULL);
                if (r < 0) {
                        log_warning_errno(r, "Cannot resolve %s user. Proceeding to dump core as root: %m", user);
                        uid = gid = 0;
                }
        } else {
                r = parse_gid(context[CONTEXT_GID], &gid);
                if (r < 0)
                        return r;
        }

        return drop_privileges(uid, gid, 0);
}

static int submit_coredump(
                const char *context[_CONTEXT_MAX],
                struct iovec *iovec,
                size_t n_iovec_allocated,
                size_t n_iovec,
                int input_fd) {

        _cleanup_close_ int coredump_fd = -1, coredump_node_fd = -1;
        _cleanup_free_ char *core_message = NULL, *filename = NULL, *coredump_data = NULL;
        uint64_t coredump_size = UINT64_MAX;
        int r;

        assert(context);
        assert(iovec);
        assert(n_iovec_allocated >= n_iovec + 3);
        assert(input_fd >= 0);

        /* Vacuum before we write anything again */
        (void) coredump_vacuum(-1, arg_keep_free, arg_max_use);

        /* Always stream the coredump to disk, if that's possible */
        r = save_external_coredump(context, input_fd, &filename, &coredump_node_fd, &coredump_fd, &coredump_size);
        if (r < 0)
                /* Skip whole core dumping part */
                goto log;

        /* If we don't want to keep the coredump on disk, remove it now, as later on we will lack the privileges for
         * it. However, we keep the fd to it, so that we can still process it and log it. */
        r = maybe_remove_external_coredump(filename, coredump_size);
        if (r < 0)
                return r;
        if (r == 0) {
                const char *coredump_filename;

                coredump_filename = strjoina("COREDUMP_FILENAME=", filename);
                IOVEC_SET_STRING(iovec[n_iovec++], coredump_filename);
        } else if (arg_storage == COREDUMP_STORAGE_EXTERNAL)
                log_info("The core will not be stored: size %"PRIu64" is greater than %"PRIu64" (the configured maximum)",
                         coredump_size, arg_external_size_max);

        /* Vacuum again, but exclude the coredump we just created */
        (void) coredump_vacuum(coredump_node_fd >= 0 ? coredump_node_fd : coredump_fd, arg_keep_free, arg_max_use);

        /* Now, let's drop privileges to become the user who owns the segfaulted process and allocate the coredump
         * memory under the user's uid. This also ensures that the credentials journald will see are the ones of the
         * coredumping user, thus making sure the user gets access to the core dump. Let's also get rid of all
         * capabilities, if we run as root, we won't need them anymore. */
        r = change_uid_gid(context);
        if (r < 0)
                return log_error_errno(r, "Failed to drop privileges: %m");

#ifdef HAVE_ELFUTILS
        /* Try to get a strack trace if we can */
        if (coredump_size <= arg_process_size_max) {
                _cleanup_free_ char *stacktrace = NULL;

                r = coredump_make_stack_trace(coredump_fd, context[CONTEXT_EXE], &stacktrace);
                if (r >= 0)
                        core_message = strjoin("MESSAGE=Process ", context[CONTEXT_PID],
                                               " (", context[CONTEXT_COMM], ") of user ",
                                               context[CONTEXT_UID], " dumped core.\n\n",
                                               stacktrace);
                else if (r == -EINVAL)
                        log_warning("Failed to generate stack trace: %s", dwfl_errmsg(dwfl_errno()));
                else
                        log_warning_errno(r, "Failed to generate stack trace: %m");
        } else
                log_debug("Not generating stack trace: core size %"PRIu64" is greater than %"PRIu64" (the configured maximum)",
                          coredump_size, arg_process_size_max);

        if (!core_message)
#endif
log:
        core_message = strjoin("MESSAGE=Process ", context[CONTEXT_PID], " (",
                               context[CONTEXT_COMM], ") of user ",
                               context[CONTEXT_UID], " dumped core.");
        if (core_message)
                IOVEC_SET_STRING(iovec[n_iovec++], core_message);

        /* Optionally store the entire coredump in the journal */
        if (arg_storage == COREDUMP_STORAGE_JOURNAL) {
                if (coredump_size <= arg_journal_size_max) {
                        size_t sz = 0;

                        /* Store the coredump itself in the journal */

                        r = allocate_journal_field(coredump_fd, (size_t) coredump_size, &coredump_data, &sz);
                        if (r >= 0) {
                                iovec[n_iovec].iov_base = coredump_data;
                                iovec[n_iovec].iov_len = sz;
                                n_iovec++;
                        } else
                                log_warning_errno(r, "Failed to attach the core to the journal entry: %m");
                } else
                        log_info("The core will not be stored: size %"PRIu64" is greater than %"PRIu64" (the configured maximum)",
                                 coredump_size, arg_journal_size_max);
        }

        assert(n_iovec <= n_iovec_allocated);

        r = sd_journal_sendv(iovec, n_iovec);
        if (r < 0)
                return log_error_errno(r, "Failed to log coredump: %m");

        return 0;
}

static void map_context_fields(const struct iovec *iovec, const char *context[]) {

        static const char * const context_field_names[_CONTEXT_MAX] = {
                [CONTEXT_PID] = "COREDUMP_PID=",
                [CONTEXT_UID] = "COREDUMP_UID=",
                [CONTEXT_GID] = "COREDUMP_GID=",
                [CONTEXT_SIGNAL] = "COREDUMP_SIGNAL=",
                [CONTEXT_TIMESTAMP] = "COREDUMP_TIMESTAMP=",
                [CONTEXT_COMM] = "COREDUMP_COMM=",
                [CONTEXT_EXE] = "COREDUMP_EXE=",
                [CONTEXT_RLIMIT] = "COREDUMP_RLIMIT=",
        };

        unsigned i;

        assert(iovec);
        assert(context);

        for (i = 0; i < _CONTEXT_MAX; i++) {
                size_t l;

                l = strlen(context_field_names[i]);
                if (iovec->iov_len < l)
                        continue;

                if (memcmp(iovec->iov_base, context_field_names[i], l) != 0)
                        continue;

                /* Note that these strings are NUL terminated, because we made sure that a trailing NUL byte is in the
                 * buffer, though not included in the iov_len count. (see below) */
                context[i] = (char*) iovec->iov_base + l;
                break;
        }
}

static int process_socket(int fd) {
        _cleanup_close_ int coredump_fd = -1;
        struct iovec *iovec = NULL;
        size_t n_iovec = 0, n_iovec_allocated = 0, i;
        const char *context[_CONTEXT_MAX] = {};
        int r;

        assert(fd >= 0);

        log_set_target(LOG_TARGET_AUTO);
        log_parse_environment();
        log_open();

        for (;;) {
                union {
                        struct cmsghdr cmsghdr;
                        uint8_t buf[CMSG_SPACE(sizeof(int))];
                } control = {};
                struct msghdr mh = {
                        .msg_control = &control,
                        .msg_controllen = sizeof(control),
                        .msg_iovlen = 1,
                };
                ssize_t n;
                ssize_t l;

                if (!GREEDY_REALLOC(iovec, n_iovec_allocated, n_iovec + 3)) {
                        r = log_oom();
                        goto finish;
                }

                l = next_datagram_size_fd(fd);
                if (l < 0) {
                        r = log_error_errno(l, "Failed to determine datagram size to read: %m");
                        goto finish;
                }

                assert(l >= 0);

                iovec[n_iovec].iov_len = l;
                iovec[n_iovec].iov_base = malloc(l + 1);
                if (!iovec[n_iovec].iov_base) {
                        r = log_oom();
                        goto finish;
                }

                mh.msg_iov = iovec + n_iovec;

                n = recvmsg(fd, &mh, MSG_NOSIGNAL|MSG_CMSG_CLOEXEC);
                if (n < 0)  {
                        free(iovec[n_iovec].iov_base);
                        r = log_error_errno(errno, "Failed to receive datagram: %m");
                        goto finish;
                }

                if (n == 0) {
                        struct cmsghdr *cmsg, *found = NULL;
                        /* The final zero-length datagram carries the file descriptor and tells us that we're done. */

                        free(iovec[n_iovec].iov_base);

                        CMSG_FOREACH(cmsg, &mh) {
                                if (cmsg->cmsg_level == SOL_SOCKET &&
                                    cmsg->cmsg_type == SCM_RIGHTS &&
                                    cmsg->cmsg_len == CMSG_LEN(sizeof(int))) {
                                        assert(!found);
                                        found = cmsg;
                                }
                        }

                        if (!found) {
                                log_error("Coredump file descriptor missing.");
                                r = -EBADMSG;
                                goto finish;
                        }

                        assert(coredump_fd < 0);
                        coredump_fd = *(int*) CMSG_DATA(found);
                        break;
                }

                /* Add trailing NUL byte, in case these are strings */
                ((char*) iovec[n_iovec].iov_base)[n] = 0;
                iovec[n_iovec].iov_len = (size_t) n;

                cmsg_close_all(&mh);
                map_context_fields(iovec + n_iovec, context);
                n_iovec++;
        }

        if (!GREEDY_REALLOC(iovec, n_iovec_allocated, n_iovec + 3)) {
                r = log_oom();
                goto finish;
        }

        /* Make sure we got all data we really need */
        assert(context[CONTEXT_PID]);
        assert(context[CONTEXT_UID]);
        assert(context[CONTEXT_GID]);
        assert(context[CONTEXT_SIGNAL]);
        assert(context[CONTEXT_TIMESTAMP]);
        assert(context[CONTEXT_RLIMIT]);
        assert(context[CONTEXT_COMM]);
        assert(coredump_fd >= 0);

        r = submit_coredump(context, iovec, n_iovec_allocated, n_iovec, coredump_fd);

finish:
        for (i = 0; i < n_iovec; i++)
                free(iovec[i].iov_base);
        free(iovec);

        return r;
}

static int send_iovec(const struct iovec iovec[], size_t n_iovec, int input_fd) {

        static const union sockaddr_union sa = {
                .un.sun_family = AF_UNIX,
                .un.sun_path = "/run/systemd/coredump",
        };
        _cleanup_close_ int fd = -1;
        size_t i;
        int r;

        assert(iovec || n_iovec <= 0);
        assert(input_fd >= 0);

        fd = socket(AF_UNIX, SOCK_SEQPACKET|SOCK_CLOEXEC, 0);
        if (fd < 0)
                return log_error_errno(errno, "Failed to create coredump socket: %m");

        if (connect(fd, &sa.sa, SOCKADDR_UN_LEN(sa.un)) < 0)
                return log_error_errno(errno, "Failed to connect to coredump service: %m");

        for (i = 0; i < n_iovec; i++) {
                struct msghdr mh = {
                        .msg_iov = (struct iovec*) iovec + i,
                        .msg_iovlen = 1,
                };
                struct iovec copy[2];

                for (;;) {
                        if (sendmsg(fd, &mh, MSG_NOSIGNAL) >= 0)
                                break;

                        if (errno == EMSGSIZE && mh.msg_iov[0].iov_len > 0) {
                                /* This field didn't fit? That's a pity. Given that this is just metadata,
                                 * let's truncate the field at half, and try again. We append three dots, in
                                 * order to show that this is truncated. */

                                if (mh.msg_iov != copy) {
                                        /* We don't want to modify the caller's iovec, hence let's create our
                                         * own array, consisting of two new iovecs, where the first is a
                                         * (truncated) copy of what we want to send, and the second one
                                         * contains the trailing dots. */
                                        copy[0] = iovec[i];
                                        copy[1] = (struct iovec) {
                                                .iov_base = (char[]) { '.', '.', '.' },
                                                .iov_len = 3,
                                        };

                                        mh.msg_iov = copy;
                                        mh.msg_iovlen = 2;
                                }

                                copy[0].iov_len /= 2; /* halve it, and try again */
                                continue;
                        }

                        return log_error_errno(errno, "Failed to send coredump datagram: %m");
                }
        }

        r = send_one_fd(fd, input_fd, 0);
        if (r < 0)
                return log_error_errno(r, "Failed to send coredump fd: %m");

        return 0;
}

static int process_special_crash(const char *context[], int input_fd) {
        _cleanup_close_ int coredump_fd = -1, coredump_node_fd = -1;
        _cleanup_free_ char *filename = NULL;
        uint64_t coredump_size;
        int r;

        assert(context);
        assert(input_fd >= 0);

        /* If we are pid1 or journald, we cut things short, don't write to the journal, but still create a coredump. */

        if (arg_storage != COREDUMP_STORAGE_NONE)
                arg_storage = COREDUMP_STORAGE_EXTERNAL;

        r = save_external_coredump(context, input_fd, &filename, &coredump_node_fd, &coredump_fd, &coredump_size);
        if (r < 0)
                return r;

        r = maybe_remove_external_coredump(filename, coredump_size);
        if (r < 0)
                return r;

        log_notice("Detected coredump of the journal daemon or PID 1, diverted to %s.", filename);

        return 0;
}

static char* set_iovec_field(struct iovec iovec[27], size_t *n_iovec, const char *field, const char *value) {
        char *x;

        x = strappend(field, value);
        if (x)
                IOVEC_SET_STRING(iovec[(*n_iovec)++], x);
        return x;
}

static char* set_iovec_field_free(struct iovec iovec[27], size_t *n_iovec, const char *field, char *value) {
        char *x;

        x = set_iovec_field(iovec, n_iovec, field, value);
        free(value);
        return x;
}

static int gather_pid_metadata(
                const char *context[_CONTEXT_MAX],
                char **comm_fallback,
                struct iovec iovec[27], size_t *n_iovec) {

        _cleanup_free_ char *exe = NULL, *comm = NULL;
        uid_t owner_uid;
        pid_t pid;
        char *t;
        const char *p;
        int r;

        r = parse_pid(context[CONTEXT_PID], &pid);
        if (r < 0)
                return log_error_errno(r, "Failed to parse PID.");

        r = get_process_comm(pid, &comm);
        if (r < 0) {
                log_warning_errno(r, "Failed to get COMM, falling back to the command line: %m");
                comm = strv_join(comm_fallback, " ");
                if (!comm)
                        return log_oom();
        }

        r = get_process_exe(pid, &exe);
        if (r < 0)
                log_warning_errno(r, "Failed to get EXE, ignoring: %m");

        if (cg_pid_get_unit(pid, &t) >= 0) {

                /* If this is PID 1 disable coredump collection, we'll unlikely be able to process it later on. */
                if (streq(t, SPECIAL_INIT_SCOPE)) {
                        log_notice("Due to PID 1 having crashed coredump collection will now be turned off.");
                        (void) write_string_file("/proc/sys/kernel/core_pattern", "|/bin/false", 0);
                }

                /* Let's avoid dead-locks when processing journald and init crashes, as socket activation and logging
                 * are unlikely to work then. */
                if (STR_IN_SET(t, SPECIAL_JOURNALD_SERVICE, SPECIAL_INIT_SCOPE)) {
                        free(t);
                        return process_special_crash(context, STDIN_FILENO);
                }

                set_iovec_field_free(iovec, n_iovec, "COREDUMP_UNIT=", t);
        }

        /* OK, now we know it's not the journal, hence we can make use of it now. */
        log_set_target(LOG_TARGET_JOURNAL_OR_KMSG);
        log_open();

        if (cg_pid_get_user_unit(pid, &t) >= 0)
                set_iovec_field_free(iovec, n_iovec, "COREDUMP_USER_UNIT=", t);

        /* The next few are mandatory */
        if (!set_iovec_field(iovec, n_iovec, "COREDUMP_PID=", context[CONTEXT_PID]))
                return log_oom();

        if (!set_iovec_field(iovec, n_iovec, "COREDUMP_UID=", context[CONTEXT_UID]))
                return log_oom();

        if (!set_iovec_field(iovec, n_iovec, "COREDUMP_GID=", context[CONTEXT_GID]))
                return log_oom();

        if (!set_iovec_field(iovec, n_iovec, "COREDUMP_SIGNAL=", context[CONTEXT_SIGNAL]))
                return log_oom();

        if (!set_iovec_field(iovec, n_iovec, "COREDUMP_RLIMIT=", context[CONTEXT_RLIMIT]))
                return log_oom();

        if (!set_iovec_field(iovec, n_iovec, "COREDUMP_COMM=", comm))
                return log_oom();

        if (exe &&
            !set_iovec_field(iovec, n_iovec, "COREDUMP_EXE=", exe))
                return log_oom();

        if (sd_pid_get_session(pid, &t) >= 0)
                set_iovec_field_free(iovec, n_iovec, "COREDUMP_SESSION=", t);

        if (sd_pid_get_owner_uid(pid, &owner_uid) >= 0) {
                r = asprintf(&t, "COREDUMP_OWNER_UID=" UID_FMT, owner_uid);
                if (r > 0)
                        IOVEC_SET_STRING(iovec[(*n_iovec)++], t);
        }

        if (sd_pid_get_slice(pid, &t) >= 0)
                set_iovec_field_free(iovec, n_iovec, "COREDUMP_SLICE=", t);

        if (get_process_cmdline(pid, 0, false, &t) >= 0)
                set_iovec_field_free(iovec, n_iovec, "COREDUMP_CMDLINE=", t);

        if (cg_pid_get_path_shifted(pid, NULL, &t) >= 0)
                set_iovec_field_free(iovec, n_iovec, "COREDUMP_CGROUP=", t);

        if (compose_open_fds(pid, &t) >= 0)
                set_iovec_field_free(iovec, n_iovec, "COREDUMP_OPEN_FDS=", t);

        p = procfs_file_alloca(pid, "status");
        if (read_full_file(p, &t, NULL) >= 0)
                set_iovec_field_free(iovec, n_iovec, "COREDUMP_PROC_STATUS=", t);

        p = procfs_file_alloca(pid, "maps");
        if (read_full_file(p, &t, NULL) >= 0)
                set_iovec_field_free(iovec, n_iovec, "COREDUMP_PROC_MAPS=", t);

        p = procfs_file_alloca(pid, "limits");
        if (read_full_file(p, &t, NULL) >= 0)
                set_iovec_field_free(iovec, n_iovec, "COREDUMP_PROC_LIMITS=", t);

        p = procfs_file_alloca(pid, "cgroup");
        if (read_full_file(p, &t, NULL) >=0)
                set_iovec_field_free(iovec, n_iovec, "COREDUMP_PROC_CGROUP=", t);

        p = procfs_file_alloca(pid, "mountinfo");
        if (read_full_file(p, &t, NULL) >=0)
                set_iovec_field_free(iovec, n_iovec, "COREDUMP_PROC_MOUNTINFO=", t);

        if (get_process_cwd(pid, &t) >= 0)
                set_iovec_field_free(iovec, n_iovec, "COREDUMP_CWD=", t);

        if (get_process_root(pid, &t) >= 0) {
                bool proc_self_root_is_slash;

                proc_self_root_is_slash = strcmp(t, "/") == 0;

                set_iovec_field_free(iovec, n_iovec, "COREDUMP_ROOT=", t);

                /* If the process' root is "/", then there is a chance it has
                 * mounted own root and hence being containerized. */
                if (proc_self_root_is_slash && get_process_container_parent_cmdline(pid, &t) > 0)
                        set_iovec_field_free(iovec, n_iovec, "COREDUMP_CONTAINER_CMDLINE=", t);
        }

        if (get_process_environ(pid, &t) >= 0)
                set_iovec_field_free(iovec, n_iovec, "COREDUMP_ENVIRON=", t);

        t = strjoin("COREDUMP_TIMESTAMP=", context[CONTEXT_TIMESTAMP], "000000", NULL);
        if (t)
                IOVEC_SET_STRING(iovec[(*n_iovec)++], t);

        return 0;
}

static int process_kernel(int argc, char* argv[]) {

        const char *context[_CONTEXT_MAX];
        struct iovec iovec[27];
        size_t n_iovec = 0, i, n_to_free;
        int r;

        if (argc < CONTEXT_COMM + 1) {
                log_error("Not enough arguments passed from kernel (%i, expected %i).", argc - 1, CONTEXT_COMM + 1 - 1);
                return -EINVAL;
        }

        context[CONTEXT_PID]       = argv[CONTEXT_PID + 1];
        context[CONTEXT_UID]       = argv[CONTEXT_UID + 1];
        context[CONTEXT_GID]       = argv[CONTEXT_GID + 1];
        context[CONTEXT_SIGNAL]    = argv[CONTEXT_SIGNAL + 1];
        context[CONTEXT_TIMESTAMP] = argv[CONTEXT_TIMESTAMP + 1];
        context[CONTEXT_RLIMIT]    = argv[CONTEXT_RLIMIT + 1];

        r = gather_pid_metadata(context, argv + CONTEXT_COMM + 1, iovec, &n_iovec);
        if (r < 0)
                goto finish;
        n_to_free = n_iovec;

        IOVEC_SET_STRING(iovec[n_iovec++], "MESSAGE_ID=fc2e22bc6ee647b6b90729ab34a250b1");

        assert_cc(2 == LOG_CRIT);
        IOVEC_SET_STRING(iovec[n_iovec++], "PRIORITY=2");

        assert(n_iovec <= ELEMENTSOF(iovec));

        r = send_iovec(iovec, n_iovec, STDIN_FILENO);

 finish:
        for (i = 0; i < n_to_free; i++)
                free(iovec[i].iov_base);

        return r;
}

int main(int argc, char *argv[]) {
        int r;

        /* First, log to a safe place, since we don't know what crashed and it might
         * be journald which we'd rather not log to then. */

        log_set_target(LOG_TARGET_KMSG);
        log_open();

        /* Make sure we never enter a loop */
        (void) prctl(PR_SET_DUMPABLE, 0);

        /* Ignore all parse errors */
        (void) parse_config();

        log_debug("Selected storage '%s'.", coredump_storage_to_string(arg_storage));
        log_debug("Selected compression %s.", yes_no(arg_compress));

        r = sd_listen_fds(false);
        if (r < 0) {
                log_error_errno(r, "Failed to determine number of file descriptor: %m");
                goto finish;
        }

        /* If we got an fd passed, we are running in coredumpd mode. Otherwise we
         * are invoked from the kernel as coredump handler. */
        if (r == 0)
                r = process_kernel(argc, argv);
        else if (r == 1)
                r = process_socket(SD_LISTEN_FDS_START);
        else {
                log_error("Received unexpected number of file descriptors.");
                r = -EINVAL;
        }

finish:
        return r < 0 ? EXIT_FAILURE : EXIT_SUCCESS;
}