nix-super/src/nix/daemon.cc

389 lines
12 KiB
C++
Raw Normal View History

2021-01-14 01:05:04 +02:00
#include "command.hh"
#include "shared.hh"
#include "local-store.hh"
#include "remote-store.hh"
#include "util.hh"
#include "serialise.hh"
2006-11-30 22:45:20 +02:00
#include "archive.hh"
#include "globals.hh"
#include "derivations.hh"
#include "finally.hh"
#include "legacy.hh"
#include "daemon.hh"
#include <algorithm>
2019-11-06 17:53:02 +02:00
#include <climits>
#include <cstring>
2019-11-06 17:53:02 +02:00
#include <unistd.h>
#include <signal.h>
#include <sys/types.h>
#include <sys/wait.h>
#include <sys/stat.h>
#include <sys/socket.h>
#include <sys/un.h>
#include <errno.h>
#include <pwd.h>
#include <grp.h>
#include <fcntl.h>
#if __APPLE__ || __FreeBSD__
#include <sys/ucred.h>
#endif
using namespace nix;
using namespace nix::daemon;
struct AuthorizationSettings : Config {
Move `trustedUsers` and `allowedUsers` to separate config struct These settings are not needed for libstore at all, they are just used by the nix daemon *command* for authorization on unix domain sockets. My moving them to a new configuration struct just in that file, we avoid them leaking anywhere else. Also, it is good to break up the mammoth `Settings` struct in general. Issue #5638 tracks this. The message is not changed because I do not want to regress in convenience to the user. Just saying "this connection is not trusted" doesn't tell them out to fix the issue. The ideal thing to do would be to somehow parameterize `processCommand` on how the error should be displayed, so different sorts of connections can display different information to the user based on how authentication is performed for the connection in question. This, however, is a good bit more work, so it is left for the future. This came up with me thinking about the tcp:// store (#5265). The larger project is not TCP *per se*, but the idea that it should be possible for something else to manage access control to services like the Nix Daemon, and those services simply trust or trust the incoming connection as they are told. This is a more capability-oriented way of thinking about trust than "every server implements its own auth separately" as we are used to today. Its very great that libstore itself already implements just this model, and so via this refactor I basically want to "enshrine" that so it continues to be the case.
2023-02-02 20:20:44 +02:00
Setting<Strings> trustedUsers{
this, {"root"}, "trusted-users",
R"(
A list of names of users (separated by whitespace) that have
additional rights when connecting to the Nix daemon, such as the
ability to specify additional binary caches, or to import unsigned
NARs. You can also specify groups by prefixing them with `@`; for
instance, `@wheel` means all users in the `wheel` group. The default
is `root`.
> **Warning**
>
> Adding a user to `trusted-users` is essentially equivalent to
> giving that user root access to the system. For example, the user
> can set `sandbox-paths` and thereby obtain read access to
> directories that are otherwise inacessible to them.
)"};
/* ?Who we trust to use the daemon in safe ways */
Setting<Strings> allowedUsers{
this, {"*"}, "allowed-users",
R"(
A list of names of users (separated by whitespace) that are allowed
to connect to the Nix daemon. As with the `trusted-users` option,
you can specify groups by prefixing them with `@`. Also, you can
allow all users by specifying `*`. The default is `*`.
Note that trusted users are always allowed to connect.
)"};
};
AuthorizationSettings authorizationSettings;
Move `trustedUsers` and `allowedUsers` to separate config struct These settings are not needed for libstore at all, they are just used by the nix daemon *command* for authorization on unix domain sockets. My moving them to a new configuration struct just in that file, we avoid them leaking anywhere else. Also, it is good to break up the mammoth `Settings` struct in general. Issue #5638 tracks this. The message is not changed because I do not want to regress in convenience to the user. Just saying "this connection is not trusted" doesn't tell them out to fix the issue. The ideal thing to do would be to somehow parameterize `processCommand` on how the error should be displayed, so different sorts of connections can display different information to the user based on how authentication is performed for the connection in question. This, however, is a good bit more work, so it is left for the future. This came up with me thinking about the tcp:// store (#5265). The larger project is not TCP *per se*, but the idea that it should be possible for something else to manage access control to services like the Nix Daemon, and those services simply trust or trust the incoming connection as they are told. This is a more capability-oriented way of thinking about trust than "every server implements its own auth separately" as we are used to today. Its very great that libstore itself already implements just this model, and so via this refactor I basically want to "enshrine" that so it continues to be the case.
2023-02-02 20:20:44 +02:00
static GlobalConfig::Register rSettings(&authorizationSettings);
Move `trustedUsers` and `allowedUsers` to separate config struct These settings are not needed for libstore at all, they are just used by the nix daemon *command* for authorization on unix domain sockets. My moving them to a new configuration struct just in that file, we avoid them leaking anywhere else. Also, it is good to break up the mammoth `Settings` struct in general. Issue #5638 tracks this. The message is not changed because I do not want to regress in convenience to the user. Just saying "this connection is not trusted" doesn't tell them out to fix the issue. The ideal thing to do would be to somehow parameterize `processCommand` on how the error should be displayed, so different sorts of connections can display different information to the user based on how authentication is performed for the connection in question. This, however, is a good bit more work, so it is left for the future. This came up with me thinking about the tcp:// store (#5265). The larger project is not TCP *per se*, but the idea that it should be possible for something else to manage access control to services like the Nix Daemon, and those services simply trust or trust the incoming connection as they are told. This is a more capability-oriented way of thinking about trust than "every server implements its own auth separately" as we are used to today. Its very great that libstore itself already implements just this model, and so via this refactor I basically want to "enshrine" that so it continues to be the case.
2023-02-02 20:20:44 +02:00
#ifndef __linux__
#define SPLICE_F_MOVE 0
static ssize_t splice(int fd_in, void *off_in, int fd_out, void *off_out, size_t len, unsigned int flags)
{
2020-06-15 15:12:39 +03:00
// We ignore most parameters, we just have them for conformance with the linux syscall
std::vector<char> buf(8192);
auto read_count = read(fd_in, buf.data(), buf.size());
if (read_count == -1)
return read_count;
auto write_count = decltype(read_count)(0);
while (write_count < read_count) {
auto res = write(fd_out, buf.data() + write_count, read_count - write_count);
if (res == -1)
return res;
write_count += res;
}
return read_count;
}
#endif
static void sigChldHandler(int sigNo)
{
// Ensure we don't modify errno of whatever we've interrupted
auto saved_errno = errno;
2020-06-15 15:12:39 +03:00
// Reap all dead children.
while (waitpid(-1, 0, WNOHANG) > 0) ;
errno = saved_errno;
}
static void setSigChldAction(bool autoReap)
{
struct sigaction act, oact;
act.sa_handler = autoReap ? sigChldHandler : SIG_DFL;
sigfillset(&act.sa_mask);
act.sa_flags = 0;
if (sigaction(SIGCHLD, &act, &oact))
throw SysError("setting SIGCHLD handler");
}
bool matchUser(const std::string & user, const std::string & group, const Strings & users)
{
if (find(users.begin(), users.end(), "*") != users.end())
return true;
if (find(users.begin(), users.end(), user) != users.end())
return true;
for (auto & i : users)
if (i.substr(0, 1) == "@") {
if (group == i.substr(1)) return true;
struct group * gr = getgrnam(i.c_str() + 1);
if (!gr) continue;
for (char * * mem = gr->gr_mem; *mem; mem++)
if (user == std::string(*mem)) return true;
}
return false;
}
struct PeerInfo
{
bool pidKnown;
pid_t pid;
bool uidKnown;
uid_t uid;
bool gidKnown;
gid_t gid;
};
2020-06-15 15:12:39 +03:00
// Get the identity of the caller, if possible.
static PeerInfo getPeerInfo(int remote)
{
PeerInfo peer = { false, 0, false, 0, false, 0 };
#if defined(SO_PEERCRED)
ucred cred;
socklen_t credLen = sizeof(cred);
if (getsockopt(remote, SOL_SOCKET, SO_PEERCRED, &cred, &credLen) == -1)
throw SysError("getting peer credentials");
peer = { true, cred.pid, true, cred.uid, true, cred.gid };
#elif defined(LOCAL_PEERCRED)
#if !defined(SOL_LOCAL)
#define SOL_LOCAL 0
#endif
xucred cred;
socklen_t credLen = sizeof(cred);
if (getsockopt(remote, SOL_LOCAL, LOCAL_PEERCRED, &cred, &credLen) == -1)
throw SysError("getting peer credentials");
peer = { false, 0, true, cred.cr_uid, false, 0 };
#endif
return peer;
}
#define SD_LISTEN_FDS_START 3
static ref<Store> openUncachedStore()
{
Store::Params params; // FIXME: get params from somewhere
// Disable caching since the client already does that.
params["path-info-cache-size"] = "0";
return openStore(settings.storeUri, params);
}
2021-01-14 01:05:04 +02:00
static void daemonLoop()
{
2014-12-12 18:14:28 +02:00
if (chdir("/") == -1)
throw SysError("cannot change current directory");
2014-08-13 04:50:44 +03:00
AutoCloseFD fdSocket;
2020-06-15 15:12:39 +03:00
// Handle socket-based activation by systemd.
auto listenFds = getEnv("LISTEN_FDS");
if (listenFds) {
if (getEnv("LISTEN_PID") != std::to_string(getpid()) || listenFds != "1")
throw Error("unexpected systemd environment variables");
fdSocket = SD_LISTEN_FDS_START;
closeOnExec(fdSocket.get());
}
2020-06-15 15:12:39 +03:00
// Otherwise, create and bind to a Unix domain socket.
else {
createDirs(dirOf(settings.nixDaemonSocketFile));
fdSocket = createUnixDomainSocket(settings.nixDaemonSocketFile, 0666);
}
// Get rid of children automatically; don't let them become zombies.
setSigChldAction(true);
2020-06-15 15:12:39 +03:00
// Loop accepting connections.
while (1) {
try {
2020-06-15 15:12:39 +03:00
// Accept a connection.
struct sockaddr_un remoteAddr;
socklen_t remoteAddrLen = sizeof(remoteAddr);
2016-07-11 22:44:44 +03:00
AutoCloseFD remote = accept(fdSocket.get(),
(struct sockaddr *) &remoteAddr, &remoteAddrLen);
checkInterrupt();
2016-07-11 22:44:44 +03:00
if (!remote) {
if (errno == EINTR) continue;
throw SysError("accepting connection");
}
2016-07-11 22:44:44 +03:00
closeOnExec(remote.get());
2012-07-31 00:13:25 +03:00
Recursive Nix support This allows Nix builders to call Nix to build derivations, with some limitations. Example: let nixpkgs = fetchTarball channel:nixos-18.03; in with import <nixpkgs> {}; runCommand "foo" { buildInputs = [ nix jq ]; NIX_PATH = "nixpkgs=${nixpkgs}"; } '' hello=$(nix-build -E '(import <nixpkgs> {}).hello.overrideDerivation (args: { name = "hello-3.5"; })') $hello/bin/hello mkdir -p $out/bin ln -s $hello/bin/hello $out/bin/hello nix path-info -r --json $hello | jq . '' This derivation makes a recursive Nix call to build GNU Hello and symlinks it from its $out, i.e. # ll ./result/bin/ lrwxrwxrwx 1 root root 63 Jan 1 1970 hello -> /nix/store/s0awxrs71gickhaqdwxl506hzccb30y5-hello-3.5/bin/hello # nix-store -qR ./result /nix/store/hwwqshlmazzjzj7yhrkyjydxamvvkfd3-glibc-2.26-131 /nix/store/s0awxrs71gickhaqdwxl506hzccb30y5-hello-3.5 /nix/store/sgmvvyw8vhfqdqb619bxkcpfn9lvd8ss-foo This is implemented as follows: * Before running the outer builder, Nix creates a Unix domain socket '.nix-socket' in the builder's temporary directory and sets $NIX_REMOTE to point to it. It starts a thread to process connections to this socket. (Thus you don't need to have nix-daemon running.) * The daemon thread uses a wrapper store (RestrictedStore) to keep track of paths added through recursive Nix calls, to implement some restrictions (see below), and to do some censorship (e.g. for purity, queryPathInfo() won't return impure information such as signatures and timestamps). * After the build finishes, the output paths are scanned for references to the paths added through recursive Nix calls (in addition to the inputs closure). Thus, in the example above, $out has a reference to $hello. The main restriction on recursive Nix calls is that they cannot do arbitrary substitutions. For example, doing nix-store -r /nix/store/kmwd1hq55akdb9sc7l3finr175dajlby-hello-2.10 is forbidden unless /nix/store/kmwd... is in the inputs closure or previously built by a recursive Nix call. This is to prevent irreproducible derivations that have hidden dependencies on substituters or the current store contents. Building a derivation is fine, however, and Nix will use substitutes if available. In other words, the builder has to present proof that it knows how to build a desired store path from scratch by constructing a derivation graph for that path. Probably we should also disallow instantiating/building fixed-output derivations (specifically, those that access the network, but currently we have no way to mark fixed-output derivations that don't access the network). Otherwise sandboxed derivations can bypass sandbox restrictions and access the network. When sandboxing is enabled, we make paths appear in the sandbox of the builder by entering the mount namespace of the builder and bind-mounting each path. This is tricky because we do a pivot_root() in the builder to change the root directory of its mount namespace, and thus the host /nix/store is not visible in the mount namespace of the builder. To get around this, just before doing pivot_root(), we branch a second mount namespace that shares its /nix/store mountpoint with the parent. Recursive Nix currently doesn't work on macOS in sandboxed mode (because we can't change the sandbox policy of a running build) and in non-root mode (because setns() barfs).
2018-10-02 17:01:26 +03:00
TrustedFlag trusted = NotTrusted;
2016-07-11 22:44:44 +03:00
PeerInfo peer = getPeerInfo(remote.get());
struct passwd * pw = peer.uidKnown ? getpwuid(peer.uid) : 0;
std::string user = pw ? pw->pw_name : std::to_string(peer.uid);
struct group * gr = peer.gidKnown ? getgrgid(peer.gid) : 0;
std::string group = gr ? gr->gr_name : std::to_string(peer.gid);
Strings trustedUsers = authorizationSettings.trustedUsers;
Strings allowedUsers = authorizationSettings.allowedUsers;
2014-08-04 19:13:14 +03:00
2014-08-05 11:19:57 +03:00
if (matchUser(user, group, trustedUsers))
Recursive Nix support This allows Nix builders to call Nix to build derivations, with some limitations. Example: let nixpkgs = fetchTarball channel:nixos-18.03; in with import <nixpkgs> {}; runCommand "foo" { buildInputs = [ nix jq ]; NIX_PATH = "nixpkgs=${nixpkgs}"; } '' hello=$(nix-build -E '(import <nixpkgs> {}).hello.overrideDerivation (args: { name = "hello-3.5"; })') $hello/bin/hello mkdir -p $out/bin ln -s $hello/bin/hello $out/bin/hello nix path-info -r --json $hello | jq . '' This derivation makes a recursive Nix call to build GNU Hello and symlinks it from its $out, i.e. # ll ./result/bin/ lrwxrwxrwx 1 root root 63 Jan 1 1970 hello -> /nix/store/s0awxrs71gickhaqdwxl506hzccb30y5-hello-3.5/bin/hello # nix-store -qR ./result /nix/store/hwwqshlmazzjzj7yhrkyjydxamvvkfd3-glibc-2.26-131 /nix/store/s0awxrs71gickhaqdwxl506hzccb30y5-hello-3.5 /nix/store/sgmvvyw8vhfqdqb619bxkcpfn9lvd8ss-foo This is implemented as follows: * Before running the outer builder, Nix creates a Unix domain socket '.nix-socket' in the builder's temporary directory and sets $NIX_REMOTE to point to it. It starts a thread to process connections to this socket. (Thus you don't need to have nix-daemon running.) * The daemon thread uses a wrapper store (RestrictedStore) to keep track of paths added through recursive Nix calls, to implement some restrictions (see below), and to do some censorship (e.g. for purity, queryPathInfo() won't return impure information such as signatures and timestamps). * After the build finishes, the output paths are scanned for references to the paths added through recursive Nix calls (in addition to the inputs closure). Thus, in the example above, $out has a reference to $hello. The main restriction on recursive Nix calls is that they cannot do arbitrary substitutions. For example, doing nix-store -r /nix/store/kmwd1hq55akdb9sc7l3finr175dajlby-hello-2.10 is forbidden unless /nix/store/kmwd... is in the inputs closure or previously built by a recursive Nix call. This is to prevent irreproducible derivations that have hidden dependencies on substituters or the current store contents. Building a derivation is fine, however, and Nix will use substitutes if available. In other words, the builder has to present proof that it knows how to build a desired store path from scratch by constructing a derivation graph for that path. Probably we should also disallow instantiating/building fixed-output derivations (specifically, those that access the network, but currently we have no way to mark fixed-output derivations that don't access the network). Otherwise sandboxed derivations can bypass sandbox restrictions and access the network. When sandboxing is enabled, we make paths appear in the sandbox of the builder by entering the mount namespace of the builder and bind-mounting each path. This is tricky because we do a pivot_root() in the builder to change the root directory of its mount namespace, and thus the host /nix/store is not visible in the mount namespace of the builder. To get around this, just before doing pivot_root(), we branch a second mount namespace that shares its /nix/store mountpoint with the parent. Recursive Nix currently doesn't work on macOS in sandboxed mode (because we can't change the sandbox policy of a running build) and in non-root mode (because setns() barfs).
2018-10-02 17:01:26 +03:00
trusted = Trusted;
if ((!trusted && !matchUser(user, group, allowedUsers)) || group == settings.buildUsersGroup)
throw Error("user '%1%' is not allowed to connect to the Nix daemon", user);
printInfo(format((std::string) "accepted connection from pid %1%, user %2%" + (trusted ? " (trusted)" : ""))
2015-10-29 14:26:55 +02:00
% (peer.pidKnown ? std::to_string(peer.pid) : "<unknown>")
% (peer.uidKnown ? user : "<unknown>"));
2012-07-31 00:13:25 +03:00
2020-06-15 15:12:39 +03:00
// Fork a child to handle the connection.
2014-12-10 17:35:42 +02:00
ProcessOptions options;
options.errorPrefix = "unexpected Nix daemon error: ";
options.dieWithParent = false;
options.runExitHandlers = true;
options.allowVfork = false;
startProcess([&]() {
2016-07-11 22:44:44 +03:00
fdSocket = -1;
2014-09-25 19:45:43 +03:00
2020-06-15 15:12:39 +03:00
// Background the daemon.
if (setsid() == -1)
throw SysError("creating a new session");
2020-06-15 15:12:39 +03:00
// Restore normal handling of SIGCHLD.
setSigChldAction(false);
2020-06-15 15:12:39 +03:00
// For debugging, stuff the pid into argv[1].
2021-01-14 01:05:04 +02:00
if (peer.pidKnown && savedArgv[1]) {
auto processName = std::to_string(peer.pid);
2021-01-14 01:05:04 +02:00
strncpy(savedArgv[1], processName.c_str(), strlen(savedArgv[1]));
}
2012-07-31 00:13:25 +03:00
2020-06-15 15:12:39 +03:00
// Handle the connection.
FdSource from(remote.get());
FdSink to(remote.get());
processConnection(openUncachedStore(), from, to, trusted, NotRecursive);
2012-07-31 00:13:25 +03:00
exit(0);
2014-12-10 17:35:42 +02:00
}, options);
} catch (Interrupted & e) {
return;
2020-05-05 01:19:57 +03:00
} catch (Error & error) {
auto ei = error.info();
Improve error formatting Changes: * The divider lines are gone. These were in practice a bit confusing, in particular with --show-trace or --keep-going, since then there were multiple lines, suggesting a start/end which wasn't the case. * Instead, multi-line error messages are now indented to align with the prefix (e.g. "error: "). * The 'description' field is gone since we weren't really using it. * 'hint' is renamed to 'msg' since it really wasn't a hint. * The error is now printed *before* the location info. * The 'name' field is no longer printed since most of the time it wasn't very useful since it was just the name of the exception (like EvalError). Ideally in the future this would be a unique, easily googleable error ID (like rustc). * "trace:" is now just "…". This assumes error contexts start with something like "while doing X". Example before: error: --- AssertionError ---------------------------------------------------------------------------------------- nix at: (7:7) in file: /home/eelco/Dev/nixpkgs/pkgs/applications/misc/hello/default.nix 6| 7| x = assert false; 1; | ^ 8| assertion 'false' failed ----------------------------------------------------- show-trace ----------------------------------------------------- trace: while evaluating the attribute 'x' of the derivation 'hello-2.10' at: (192:11) in file: /home/eelco/Dev/nixpkgs/pkgs/stdenv/generic/make-derivation.nix 191| // (lib.optionalAttrs (!(attrs ? name) && attrs ? pname && attrs ? version)) { 192| name = "${attrs.pname}-${attrs.version}"; | ^ 193| } // (lib.optionalAttrs (stdenv.hostPlatform != stdenv.buildPlatform && !dontAddHostSuffix && (attrs ? name || (attrs ? pname && attrs ? version)))) { Example after: error: assertion 'false' failed at: (7:7) in file: /home/eelco/Dev/nixpkgs/pkgs/applications/misc/hello/default.nix 6| 7| x = assert false; 1; | ^ 8| … while evaluating the attribute 'x' of the derivation 'hello-2.10' at: (192:11) in file: /home/eelco/Dev/nixpkgs/pkgs/stdenv/generic/make-derivation.nix 191| // (lib.optionalAttrs (!(attrs ? name) && attrs ? pname && attrs ? version)) { 192| name = "${attrs.pname}-${attrs.version}"; | ^ 193| } // (lib.optionalAttrs (stdenv.hostPlatform != stdenv.buildPlatform && !dontAddHostSuffix && (attrs ? name || (attrs ? pname && attrs ? version)))) {
2021-01-21 01:27:36 +02:00
// FIXME: add to trace?
ei.msg = hintfmt("error processing connection: %1%", ei.msg.str());
2020-05-05 01:19:57 +03:00
logError(ei);
}
}
}
2021-01-14 01:05:04 +02:00
static void runDaemon(bool stdio)
{
if (stdio) {
if (auto store = openUncachedStore().dynamic_pointer_cast<RemoteStore>()) {
auto conn = store->openConnectionWrapper();
int from = conn->from.fd;
int to = conn->to.fd;
auto nfds = std::max(from, STDIN_FILENO) + 1;
while (true) {
fd_set fds;
FD_ZERO(&fds);
FD_SET(from, &fds);
FD_SET(STDIN_FILENO, &fds);
if (select(nfds, &fds, nullptr, nullptr, nullptr) == -1)
throw SysError("waiting for data from client or server");
if (FD_ISSET(from, &fds)) {
auto res = splice(from, nullptr, STDOUT_FILENO, nullptr, SSIZE_MAX, SPLICE_F_MOVE);
if (res == -1)
throw SysError("splicing data from daemon socket to stdout");
else if (res == 0)
throw EndOfFile("unexpected EOF from daemon socket");
}
if (FD_ISSET(STDIN_FILENO, &fds)) {
auto res = splice(STDIN_FILENO, nullptr, to, nullptr, SSIZE_MAX, SPLICE_F_MOVE);
if (res == -1)
throw SysError("splicing data from stdin to daemon socket");
else if (res == 0)
return;
}
}
} else {
FdSource from(STDIN_FILENO);
FdSink to(STDOUT_FILENO);
/* Auth hook is empty because in this mode we blindly trust the
standard streams. Limiting access to those is explicitly
not `nix-daemon`'s responsibility. */
processConnection(openUncachedStore(), from, to, Trusted, NotRecursive);
2021-01-14 01:05:04 +02:00
}
} else
daemonLoop();
}
static int main_nix_daemon(int argc, char * * argv)
{
{
auto stdio = false;
2014-08-13 04:50:44 +03:00
parseCmdLine(argc, argv, [&](Strings::iterator & arg, const Strings::iterator & end) {
if (*arg == "--daemon")
2020-06-15 15:12:39 +03:00
; // ignored for backwards compatibility
2014-08-13 04:50:44 +03:00
else if (*arg == "--help")
showManPage("nix-daemon");
else if (*arg == "--version")
printVersion("nix-daemon");
else if (*arg == "--stdio")
stdio = true;
2014-08-13 04:50:44 +03:00
else return false;
return true;
});
2021-01-14 01:05:04 +02:00
runDaemon(stdio);
return 0;
}
}
static RegisterLegacyCommand r_nix_daemon("nix-daemon", main_nix_daemon);
2021-01-14 01:05:04 +02:00
struct CmdDaemon : StoreCommand
{
std::string description() override
{
return "daemon to perform store operations on behalf of non-root clients";
}
Category category() override { return catUtility; }
std::string doc() override
{
return
#include "daemon.md"
;
}
void run(ref<Store> store) override
{
runDaemon(false);
}
};
static auto rCmdDaemon = registerCommand2<CmdDaemon>({"daemon"});