/* * \brief Unix emulation environment for Genode * \author Norman Feske * \date 2011-02-14 */ /* * Copyright (C) 2011-2012 Genode Labs GmbH * * This file is part of the Genode OS framework, which is distributed * under the terms of the GNU General Public License version 2. */ /* Genode includes */ #include #include #include #include /* Noux includes */ #include #include #include #include #include #include #include #include #include #include static bool trace_syscalls = false; namespace Noux { static Noux::Child *init_child; bool is_init_process(Child *child) { return child == init_child; } void init_process_exited() { init_child = 0; } }; extern void init_network(); /** * Timeout thread for SYSCALL_SELECT */ namespace Noux { using namespace Genode; class Timeout_scheduler : Thread<4096>, public Alarm_scheduler { private: Timer::Connection _timer; Alarm::Time _curr_time; enum { TIMER_GRANULARITY_MSEC = 10 }; void entry() { for (;;) { _timer.msleep(TIMER_GRANULARITY_MSEC); Alarm_scheduler::handle(_curr_time); _curr_time += TIMER_GRANULARITY_MSEC; } } public: Timeout_scheduler(unsigned long curr_time) : _curr_time(curr_time) { start(); } Alarm::Time curr_time() const { return _curr_time; } }; struct Timeout_state { bool timed_out; Timeout_state() : timed_out(false) { } }; class Timeout_alarm : public Alarm { private: Timeout_state *_state; Semaphore *_blocker; Timeout_scheduler *_scheduler; public: Timeout_alarm(Timeout_state *st, Semaphore *blocker, Timeout_scheduler *scheduler, Time timeout) : _state(st), _blocker(blocker), _scheduler(scheduler) { _scheduler->schedule_absolute(this, _scheduler->curr_time() + timeout); _state->timed_out = false; } void discard() { _scheduler->discard(this); } protected: bool on_alarm() { _state->timed_out = true; _blocker->up(); return false; } }; }; /***************************** ** Noux syscall dispatcher ** *****************************/ bool Noux::Child::syscall(Noux::Session::Syscall sc) { if (trace_syscalls) Genode::printf("PID %d -> SYSCALL %s\n", pid(), Noux::Session::syscall_name(sc)); try { switch (sc) { case SYSCALL_WRITE: { size_t const count_in = _sysio->write_in.count; for (size_t count = 0; count != count_in; ) { Shared_pointer io = _lookup_channel(_sysio->write_in.fd); if (!io->is_nonblocking()) if (!io->check_unblock(false, true, false)) _block_for_io_channel(io); /* * 'io->write' is expected to update 'write_out.count' */ if (io->write(_sysio, count) == false) return false; } return true; } case SYSCALL_READ: { Shared_pointer io = _lookup_channel(_sysio->read_in.fd); if (!io->is_nonblocking()) while (!io->check_unblock(true, false, false)) _block_for_io_channel(io); return io->read(_sysio); } case SYSCALL_FTRUNCATE: { Shared_pointer io = _lookup_channel(_sysio->ftruncate_in.fd); while (!io->check_unblock(true, false, false)) _block_for_io_channel(io); return io->ftruncate(_sysio); } case SYSCALL_STAT: case SYSCALL_LSTAT: /* XXX implement difference between 'lstat' and 'stat' */ { bool result = _root_dir->stat(_sysio, _sysio->stat_in.path); /** * Instead of using the uid/gid given by the actual file system * we use the ones specificed in the config. */ if (result) { _sysio->stat_out.st.uid = user_info()->uid; _sysio->stat_out.st.gid = user_info()->gid; } return result; } case SYSCALL_FSTAT: return _lookup_channel(_sysio->fstat_in.fd)->fstat(_sysio); case SYSCALL_FCNTL: if (_sysio->fcntl_in.cmd == Sysio::FCNTL_CMD_SET_FD_FLAGS) { /* we assume that there is only the close-on-execve flag */ _lookup_channel(_sysio->fcntl_in.fd)->close_on_execve = !!_sysio->fcntl_in.long_arg; return true; } return _lookup_channel(_sysio->fcntl_in.fd)->fcntl(_sysio); case SYSCALL_OPEN: { Vfs_handle *vfs_handle = _root_dir->open(_sysio, _sysio->open_in.path); if (!vfs_handle) return false; char const *leaf_path = _root_dir->leaf_path(_sysio->open_in.path); /* * File descriptors of opened directories are handled by * '_root_dir'. In this case, we use the absolute path as leaf * path because path operations always refer to the global * root. */ if (vfs_handle->ds() == _root_dir) leaf_path = _sysio->open_in.path; Shared_pointer channel(new Vfs_io_channel(_sysio->open_in.path, leaf_path, _root_dir, vfs_handle), Genode::env()->heap()); _sysio->open_out.fd = add_io_channel(channel); return true; } case SYSCALL_CLOSE: { remove_io_channel(_sysio->close_in.fd); return true; } case SYSCALL_IOCTL: return _lookup_channel(_sysio->ioctl_in.fd)->ioctl(_sysio); case SYSCALL_LSEEK: return _lookup_channel(_sysio->lseek_in.fd)->lseek(_sysio); case SYSCALL_DIRENT: return _lookup_channel(_sysio->dirent_in.fd)->dirent(_sysio); case SYSCALL_EXECVE: { Dataspace_capability binary_ds = _root_dir->dataspace(_sysio->execve_in.filename); if (!binary_ds.valid()) throw Child::Binary_does_not_exist(); Child_envexecve_in.args)> child_env(_sysio->execve_in.filename, binary_ds, _sysio->execve_in.args, _sysio->execve_in.env); _root_dir->release(_sysio->execve_in.filename, binary_ds); try { Child *child = new Child(child_env.binary_name(), parent(), pid(), _sig_rec, _root_dir, child_env.args(), child_env.env(), _cap_session, _parent_services, _resources.ep, false); /* replace ourself by the new child at the parent */ parent()->remove(this); parent()->insert(child); _assign_io_channels_to(child); /* signal main thread to remove ourself */ Genode::Signal_transmitter(_execve_cleanup_context_cap).submit(); /* start executing the new process */ child->start(); /* this child will be removed by the execve_finalization_dispatcher */ return true; } catch (Child::Binary_does_not_exist) { _sysio->error.execve = Sysio::EXECVE_NONEXISTENT; } return false; } case SYSCALL_SELECT: { Sysio::Select_fds &in_fds = _sysio->select_in.fds; size_t in_fds_total = in_fds.total_fds(); int _rd_array[in_fds_total]; int _wr_array[in_fds_total]; long timeout_sec = _sysio->select_in.timeout.sec; long timeout_usec = _sysio->select_in.timeout.usec; bool timeout_reached = false; /* * Block for one action of the watched file descriptors */ for (;;) { /* * Check I/O channels of specified file descriptors for * unblock condition. Return if one I/O channel satisfies * the condition. */ size_t unblock_rd = 0; size_t unblock_wr = 0; size_t unblock_ex = 0; /* process read fds */ for (Genode::size_t i = 0; i < in_fds_total; i++) { int fd = in_fds.array[i]; if (!fd_in_use(fd)) continue; Shared_pointer io = io_channel_by_fd(fd); if (in_fds.watch_for_rd(i)) if (io->check_unblock(true, false, false)) { _rd_array[unblock_rd++] = fd; } if (in_fds.watch_for_wr(i)) if (io->check_unblock(false, true, false)) { _wr_array[unblock_wr++] = fd; } if (in_fds.watch_for_ex(i)) if (io->check_unblock(false, false, true)) { unblock_ex++; } } if (unblock_rd || unblock_wr || unblock_ex) { /** * Merge the fd arrays in one output array */ for (size_t i = 0; i < unblock_rd; i++) { _sysio->select_out.fds.array[i] = _rd_array[i]; } _sysio->select_out.fds.num_rd = unblock_rd; /* XXX could use a pointer to select_out.fds.array instead */ for (size_t j = unblock_rd, i = 0; i < unblock_wr; i++, j++) { _sysio->select_out.fds.array[j] = _wr_array[i]; } _sysio->select_out.fds.num_wr = unblock_wr; /* exception fds are currently not considered */ _sysio->select_out.fds.num_ex = unblock_ex; return true; } /* * Return if I/O channel triggered, but timeout exceeded */ if (_sysio->select_in.timeout.zero() || timeout_reached) { /* if (timeout_reached) PINF("timeout_reached"); else PINF("timeout.zero()"); */ _sysio->select_out.fds.num_rd = 0; _sysio->select_out.fds.num_wr = 0; _sysio->select_out.fds.num_ex = 0; return true; } /* * Register ourself at all watched I/O channels * * We instantiate as many notifiers as we have file * descriptors to observe. Each notifier is associated * with the child's blocking semaphore. When any of the * notifiers get woken up, the semaphore gets unblocked. * * XXX However, the semaphore may get unblocked for other * conditions such as the destruction of the child. * ...to be done. */ Wake_up_notifier notifiers[in_fds_total]; for (Genode::size_t i = 0; i < in_fds_total; i++) { int fd = in_fds.array[i]; if (!fd_in_use(fd)) continue; Shared_pointer io = io_channel_by_fd(fd); notifiers[i].semaphore = &_blocker; io->register_wake_up_notifier(¬ifiers[i]); } /** * Register ourself at the Io_receptor_registry * * Each entry in the registry will be unblocked if an external * event has happend, e.g. network I/O. */ Io_receptor receptor(&_blocker); io_receptor_registry()->register_receptor(&receptor); /* * Block at barrier except when reaching the timeout */ if (!_sysio->select_in.timeout.infinite()) { unsigned long to_msec = (timeout_sec * 1000) + (timeout_usec / 1000); Timeout_state ts; Timeout_alarm ta(&ts, &_blocker, Noux::timeout_scheduler(), to_msec); /* block until timeout is reached or we were unblocked */ _blocker.down(); if (ts.timed_out) { timeout_reached = 1; } else { /* * We woke up before reaching the timeout, * so we discard the alarm */ ta.discard(); } } else { /* let's block infinitely */ _blocker.down(); } /* * Unregister barrier at watched I/O channels */ for (Genode::size_t i = 0; i < in_fds_total; i++) { int fd = in_fds.array[i]; if (!fd_in_use(fd)) continue; Shared_pointer io = io_channel_by_fd(fd); io->unregister_wake_up_notifier(¬ifiers[i]); } /* * Unregister receptor */ io_receptor_registry()->unregister_receptor(&receptor); } return true; } case SYSCALL_FORK: { Genode::addr_t ip = _sysio->fork_in.ip; Genode::addr_t sp = _sysio->fork_in.sp; Genode::addr_t parent_cap_addr = _sysio->fork_in.parent_cap_addr; int const new_pid = pid_allocator()->alloc(); /* * XXX To ease debugging, it would be useful to generate a * unique name that includes the PID instead of just * reusing the name of the parent. */ Child *child = new Child(_child_policy.name(), this, new_pid, _sig_rec, _root_dir, _args, _env.env(), _cap_session, _parent_services, _resources.ep, true); Family_member::insert(child); _assign_io_channels_to(child); /* copy our address space into the new child */ _resources.rm.replay(child->ram(), child->rm(), child->ds_registry(), _resources.ep); /* start executing the main thread of the new process */ child->start_forked_main_thread(ip, sp, parent_cap_addr); /* activate child entrypoint, thereby starting the new process */ child->start(); _sysio->fork_out.pid = new_pid; return true; } case SYSCALL_GETPID: { _sysio->getpid_out.pid = pid(); return true; } case SYSCALL_WAIT4: { Family_member *exited = _sysio->wait4_in.nohang ? poll4() : wait4(); if (exited) { _sysio->wait4_out.pid = exited->pid(); _sysio->wait4_out.status = exited->exit_status(); Family_member::remove(exited); PINF("submit exit signal for PID %d", exited->pid()); static_cast(exited)->submit_exit_signal(); } else { _sysio->wait4_out.pid = 0; _sysio->wait4_out.status = 0; } return true; } case SYSCALL_PIPE: { Shared_pointer pipe(new Pipe, Genode::env()->heap()); Shared_pointer pipe_sink(new Pipe_sink_io_channel(pipe, *_sig_rec), Genode::env()->heap()); Shared_pointer pipe_source(new Pipe_source_io_channel(pipe, *_sig_rec), Genode::env()->heap()); _sysio->pipe_out.fd[0] = add_io_channel(pipe_source); _sysio->pipe_out.fd[1] = add_io_channel(pipe_sink); return true; } case SYSCALL_DUP2: { int fd = add_io_channel(io_channel_by_fd(_sysio->dup2_in.fd), _sysio->dup2_in.to_fd); _sysio->dup2_out.fd = fd; return true; } case SYSCALL_UNLINK: return _root_dir->unlink(_sysio, _sysio->unlink_in.path); case SYSCALL_READLINK: return _root_dir->readlink(_sysio, _sysio->readlink_in.path); case SYSCALL_RENAME: return _root_dir->rename(_sysio, _sysio->rename_in.from_path, _sysio->rename_in.to_path); case SYSCALL_MKDIR: return _root_dir->mkdir(_sysio, _sysio->mkdir_in.path); case SYSCALL_SYMLINK: return _root_dir->symlink(_sysio, _sysio->symlink_in.newpath); case SYSCALL_USERINFO: { if (_sysio->userinfo_in.request == Sysio::USERINFO_GET_UID || _sysio->userinfo_in.request == Sysio::USERINFO_GET_GID) { _sysio->userinfo_out.uid = Noux::user_info()->uid; _sysio->userinfo_out.gid = Noux::user_info()->gid; return true; } /* * Since NOUX supports exactly one user, return false if we * got a unknown uid. */ if (_sysio->userinfo_in.uid != Noux::user_info()->uid) return false; Genode::memcpy(_sysio->userinfo_out.name, Noux::user_info()->name, sizeof(Noux::user_info()->name)); Genode::memcpy(_sysio->userinfo_out.shell, Noux::user_info()->shell, sizeof(Noux::user_info()->shell)); Genode::memcpy(_sysio->userinfo_out.home, Noux::user_info()->home, sizeof(Noux::user_info()->home)); _sysio->userinfo_out.uid = user_info()->uid; _sysio->userinfo_out.gid = user_info()->gid; return true; } case SYSCALL_GETTIMEOFDAY: { /** * Since the timeout_scheduler thread is started after noux it * basicly returns the eleapsed time since noux was started. We * abuse this timer to provide a more useful implemenation of * gettimeofday() to make certain programs (e.g. ping(1)) happy. * Note: this is just a short-term solution because Genode currently * lacks a proper time interface (there is a RTC driver however, but * there is no interface for it). */ unsigned long time = Noux::timeout_scheduler()->curr_time(); _sysio->gettimeofday_out.sec = (time / 1000); _sysio->gettimeofday_out.usec = (time % 1000) * 1000; return true; } case SYSCALL_CLOCK_GETTIME: { /** * It's the same procedure as in SYSCALL_GETTIMEOFDAY. */ unsigned long time = Noux::timeout_scheduler()->curr_time(); switch (_sysio->clock_gettime_in.clock_id) { /* CLOCK_SECOND is used by time(3) in the libc. */ case Sysio::CLOCK_ID_SECOND: { _sysio->clock_gettime_out.sec = (time / 1000); _sysio->clock_gettime_out.nsec = 0; return true; } default: { _sysio->clock_gettime_out.sec = 0; _sysio->clock_gettime_out.nsec = 0; _sysio->error.clock = Sysio::CLOCK_ERR_INVALID; return false; } } return false; } case SYSCALL_UTIMES: { /** * This systemcall is currently not implemented because we lack * the needed mechanisms in most file-systems. * * But we return true anyway to keep certain programs, e.g. make * happy. */ return true; } case SYSCALL_SOCKET: case SYSCALL_GETSOCKOPT: case SYSCALL_SETSOCKOPT: case SYSCALL_ACCEPT: case SYSCALL_BIND: case SYSCALL_LISTEN: case SYSCALL_SEND: case SYSCALL_SENDTO: case SYSCALL_RECV: case SYSCALL_RECVFROM: case SYSCALL_GETPEERNAME: case SYSCALL_SHUTDOWN: case SYSCALL_CONNECT: return _syscall_net(sc); case SYSCALL_INVALID: break; } } catch (Invalid_fd) { _sysio->error.general = Sysio::ERR_FD_INVALID; PERR("Invalid file descriptor"); } catch (...) { PERR("Unexpected exception"); } return false; } /** * Return name of init process as specified in the config */ static char *name_of_init_process() { enum { INIT_NAME_LEN = 128 }; static char buf[INIT_NAME_LEN]; Genode::config()->xml_node().sub_node("start").attribute("name").value(buf, sizeof(buf)); return buf; } /** * Read command-line arguments of init process from config */ static Noux::Args const &args_of_init_process() { static char args_buf[4096]; static Noux::Args args(args_buf, sizeof(args_buf)); Genode::Xml_node start_node = Genode::config()->xml_node().sub_node("start"); try { /* the first argument is the program name */ args.append(name_of_init_process()); Genode::Xml_node arg_node = start_node.sub_node("arg"); for (; ; arg_node = arg_node.next("arg")) { static char buf[512]; arg_node.attribute("value").value(buf, sizeof(buf)); args.append(buf); } } catch (Genode::Xml_node::Nonexistent_sub_node) { } catch (Noux::Args::Overrun) { PERR("Argument buffer overrun"); } return args; } /** * Return string containing the environment variables of init * * The variable definitions are separated by zeros. The end of the string is * marked with another zero. */ static Noux::Sysio::Env &env_string_of_init_process() { static Noux::Sysio::Env env; int index = 0; /* read environment variables for init process from config */ Genode::Xml_node start_node = Genode::config()->xml_node().sub_node("start"); try { Genode::Xml_node arg_node = start_node.sub_node("env"); for (; ; arg_node = arg_node.next("env")) { static char name_buf[256], value_buf[256]; arg_node.attribute("name").value(name_buf, sizeof(name_buf)); arg_node.attribute("value").value(value_buf, sizeof(value_buf)); Genode::size_t env_var_size = Genode::strlen(name_buf) + Genode::strlen("=") + Genode::strlen(value_buf) + 1; if (index + env_var_size < sizeof(env)) { Genode::snprintf(&env[index], env_var_size, "%s=%s", name_buf, value_buf); index += env_var_size; } else { env[index] = 0; break; } } } catch (Genode::Xml_node::Nonexistent_sub_node) { } return env; } Noux::Pid_allocator *Noux::pid_allocator() { static Noux::Pid_allocator inst; return &inst; } Noux::Timeout_scheduler *Noux::timeout_scheduler() { static Noux::Timeout_scheduler inst(0); return &inst; } Noux::User_info* Noux::user_info() { static Noux::User_info inst; return &inst; } Noux::Io_receptor_registry * Noux::io_receptor_registry() { static Noux::Io_receptor_registry _inst; return &_inst; } Terminal::Connection *Noux::terminal() { static Terminal::Connection _inst; return &_inst; } void *operator new (Genode::size_t size) { return Genode::env()->heap()->alloc(size); } int main(int argc, char **argv) { using namespace Noux; PINF("--- noux started ---"); /* look for dynamic linker */ try { static Genode::Rom_connection rom("ld.lib.so"); Genode::Process::dynamic_linker(rom.dataspace()); } catch (...) { } /* whitelist of service requests to be routed to the parent */ static Genode::Service_registry parent_services; char const *service_names[] = { "LOG", "ROM", "Timer", 0 }; for (unsigned i = 0; service_names[i]; i++) parent_services.insert(new Genode::Parent_service(service_names[i])); static Genode::Cap_connection cap; /* obtain global configuration */ try { trace_syscalls = config()->xml_node().attribute("trace_syscalls").has_value("yes"); } catch (Xml_node::Nonexistent_attribute) { } /* initialize virtual file system */ static Dir_file_system root_dir(config()->xml_node().sub_node("fstab")); /* set user information */ try { user_info()->set_info(config()->xml_node().sub_node("user")); } catch (...) { } /* initialize network */ init_network(); /* * Entrypoint used to virtualize child resources such as RAM, RM */ enum { STACK_SIZE = 1024*sizeof(long) }; static Genode::Rpc_entrypoint resources_ep(&cap, STACK_SIZE, "noux_rsc_ep"); /* create init process */ static Genode::Signal_receiver sig_rec; init_child = new Noux::Child(name_of_init_process(), 0, pid_allocator()->alloc(), &sig_rec, &root_dir, args_of_init_process(), env_string_of_init_process(), &cap, parent_services, resources_ep, false); /* * I/O channels must be dynamically allocated to handle cases where the * init program closes one of these. */ typedef Terminal_io_channel Tio; /* just a local abbreviation */ Shared_pointer channel_0(new Tio(*Noux::terminal(), Tio::STDIN, sig_rec), Genode::env()->heap()), channel_1(new Tio(*Noux::terminal(), Tio::STDOUT, sig_rec), Genode::env()->heap()), channel_2(new Tio(*Noux::terminal(), Tio::STDERR, sig_rec), Genode::env()->heap()); init_child->add_io_channel(channel_0, 0); init_child->add_io_channel(channel_1, 1); init_child->add_io_channel(channel_2, 2); init_child->start(); /* handle asynchronous events */ while (init_child) { Genode::Signal signal = sig_rec.wait_for_signal(); Signal_dispatcher *dispatcher = static_cast(signal.context()); for (int i = 0; i < signal.num(); i++) dispatcher->dispatch(); } PINF("-- exiting noux ---"); return 0; }