| /* |
| * Server-side request handling |
| * |
| * Copyright (C) 1998 Alexandre Julliard |
| * |
| * This library 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. |
| * |
| * This library 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 this library; if not, write to the Free Software |
| * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA |
| */ |
| |
| #include "config.h" |
| #include "wine/port.h" |
| |
| #include <assert.h> |
| #include <errno.h> |
| #include <fcntl.h> |
| #ifdef HAVE_PWD_H |
| #include <pwd.h> |
| #endif |
| #include <signal.h> |
| #include <stdio.h> |
| #include <stdlib.h> |
| #include <stdarg.h> |
| #include <string.h> |
| #include <sys/stat.h> |
| #include <sys/time.h> |
| #include <sys/types.h> |
| #ifdef HAVE_SYS_SOCKET_H |
| # include <sys/socket.h> |
| #endif |
| #ifdef HAVE_SYS_WAIT_H |
| # include <sys/wait.h> |
| #endif |
| #ifdef HAVE_SYS_UIO_H |
| #include <sys/uio.h> |
| #endif |
| #ifdef HAVE_SYS_UN_H |
| #include <sys/un.h> |
| #endif |
| #include <unistd.h> |
| #ifdef HAVE_POLL_H |
| #include <poll.h> |
| #endif |
| #ifdef __APPLE__ |
| # include <mach/mach_time.h> |
| #endif |
| |
| #include "ntstatus.h" |
| #define WIN32_NO_STATUS |
| #include "windef.h" |
| #include "winbase.h" |
| #include "wincon.h" |
| #include "winternl.h" |
| #include "wine/library.h" |
| |
| #include "file.h" |
| #include "process.h" |
| #include "thread.h" |
| #include "security.h" |
| #define WANT_REQUEST_HANDLERS |
| #include "request.h" |
| |
| /* Some versions of glibc don't define this */ |
| #ifndef SCM_RIGHTS |
| #define SCM_RIGHTS 1 |
| #endif |
| |
| /* path names for server master Unix socket */ |
| static const char * const server_socket_name = "socket"; /* name of the socket file */ |
| static const char * const server_lock_name = "lock"; /* name of the server lock file */ |
| |
| struct master_socket |
| { |
| struct object obj; /* object header */ |
| struct fd *fd; /* file descriptor of the master socket */ |
| }; |
| |
| static void master_socket_dump( struct object *obj, int verbose ); |
| static void master_socket_destroy( struct object *obj ); |
| static void master_socket_poll_event( struct fd *fd, int event ); |
| |
| static const struct object_ops master_socket_ops = |
| { |
| sizeof(struct master_socket), /* size */ |
| master_socket_dump, /* dump */ |
| no_get_type, /* get_type */ |
| no_add_queue, /* add_queue */ |
| NULL, /* remove_queue */ |
| NULL, /* signaled */ |
| NULL, /* satisfied */ |
| no_signal, /* signal */ |
| no_get_fd, /* get_fd */ |
| no_map_access, /* map_access */ |
| default_get_sd, /* get_sd */ |
| default_set_sd, /* set_sd */ |
| no_lookup_name, /* lookup_name */ |
| no_link_name, /* link_name */ |
| NULL, /* unlink_name */ |
| no_open_file, /* open_file */ |
| no_close_handle, /* close_handle */ |
| master_socket_destroy /* destroy */ |
| }; |
| |
| static const struct fd_ops master_socket_fd_ops = |
| { |
| NULL, /* get_poll_events */ |
| master_socket_poll_event, /* poll_event */ |
| NULL, /* flush */ |
| NULL, /* get_fd_type */ |
| NULL, /* ioctl */ |
| NULL, /* queue_async */ |
| NULL /* reselect_async */ |
| }; |
| |
| |
| struct thread *current = NULL; /* thread handling the current request */ |
| unsigned int global_error = 0; /* global error code for when no thread is current */ |
| timeout_t server_start_time = 0; /* server startup time */ |
| int server_dir_fd = -1; /* file descriptor for the server dir */ |
| int config_dir_fd = -1; /* file descriptor for the config dir */ |
| |
| static struct master_socket *master_socket; /* the master socket object */ |
| static struct timeout_user *master_timeout; |
| |
| /* complain about a protocol error and terminate the client connection */ |
| void fatal_protocol_error( struct thread *thread, const char *err, ... ) |
| { |
| va_list args; |
| |
| va_start( args, err ); |
| fprintf( stderr, "Protocol error:%04x: ", thread->id ); |
| vfprintf( stderr, err, args ); |
| va_end( args ); |
| thread->exit_code = 1; |
| kill_thread( thread, 1 ); |
| } |
| |
| /* die on a fatal error */ |
| void fatal_error( const char *err, ... ) |
| { |
| va_list args; |
| |
| va_start( args, err ); |
| fprintf( stderr, "wineserver: " ); |
| vfprintf( stderr, err, args ); |
| va_end( args ); |
| exit(1); |
| } |
| |
| /* allocate the reply data */ |
| void *set_reply_data_size( data_size_t size ) |
| { |
| assert( size <= get_reply_max_size() ); |
| if (size && !(current->reply_data = mem_alloc( size ))) size = 0; |
| current->reply_size = size; |
| return current->reply_data; |
| } |
| |
| /* return object attributes from the current request */ |
| const struct object_attributes *get_req_object_attributes( const struct security_descriptor **sd, |
| struct unicode_str *name, |
| struct object **root ) |
| { |
| static const struct object_attributes empty_attributes; |
| const struct object_attributes *attr = get_req_data(); |
| data_size_t size = get_req_data_size(); |
| |
| if (root) *root = NULL; |
| |
| if (!size) |
| { |
| *sd = NULL; |
| name->len = 0; |
| return &empty_attributes; |
| } |
| |
| if ((size < sizeof(*attr)) || (size - sizeof(*attr) < attr->sd_len) || |
| (size - sizeof(*attr) - attr->sd_len < attr->name_len)) |
| { |
| set_error( STATUS_ACCESS_VIOLATION ); |
| return NULL; |
| } |
| if (attr->sd_len && !sd_is_valid( (const struct security_descriptor *)(attr + 1), attr->sd_len )) |
| { |
| set_error( STATUS_INVALID_SECURITY_DESCR ); |
| return NULL; |
| } |
| if ((attr->name_len & (sizeof(WCHAR) - 1)) || attr->name_len >= 65534) |
| { |
| set_error( STATUS_OBJECT_NAME_INVALID ); |
| return NULL; |
| } |
| if (root && attr->rootdir && attr->name_len) |
| { |
| if (!(*root = get_directory_obj( current->process, attr->rootdir ))) return NULL; |
| } |
| *sd = attr->sd_len ? (const struct security_descriptor *)(attr + 1) : NULL; |
| name->len = attr->name_len; |
| name->str = (const WCHAR *)(attr + 1) + attr->sd_len / sizeof(WCHAR); |
| return attr; |
| } |
| |
| /* return a pointer to the request data following an object attributes structure */ |
| const void *get_req_data_after_objattr( const struct object_attributes *attr, data_size_t *len ) |
| { |
| const void *ptr = (const WCHAR *)((const struct object_attributes *)get_req_data() + 1) + |
| attr->sd_len / sizeof(WCHAR) + attr->name_len / sizeof(WCHAR); |
| *len = get_req_data_size() - ((const char *)ptr - (const char *)get_req_data()); |
| return ptr; |
| } |
| |
| /* write the remaining part of the reply */ |
| void write_reply( struct thread *thread ) |
| { |
| int ret; |
| |
| if ((ret = write( get_unix_fd( thread->reply_fd ), |
| (char *)thread->reply_data + thread->reply_size - thread->reply_towrite, |
| thread->reply_towrite )) >= 0) |
| { |
| if (!(thread->reply_towrite -= ret)) |
| { |
| free( thread->reply_data ); |
| thread->reply_data = NULL; |
| /* sent everything, can go back to waiting for requests */ |
| set_fd_events( thread->request_fd, POLLIN ); |
| set_fd_events( thread->reply_fd, 0 ); |
| } |
| return; |
| } |
| if (errno == EPIPE) |
| kill_thread( thread, 0 ); /* normal death */ |
| else if (errno != EWOULDBLOCK && (EWOULDBLOCK == EAGAIN || errno != EAGAIN)) |
| fatal_protocol_error( thread, "reply write: %s\n", strerror( errno )); |
| } |
| |
| /* send a reply to the current thread */ |
| static void send_reply( union generic_reply *reply ) |
| { |
| int ret; |
| |
| if (!current->reply_size) |
| { |
| if ((ret = write( get_unix_fd( current->reply_fd ), |
| reply, sizeof(*reply) )) != sizeof(*reply)) goto error; |
| } |
| else |
| { |
| struct iovec vec[2]; |
| |
| vec[0].iov_base = (void *)reply; |
| vec[0].iov_len = sizeof(*reply); |
| vec[1].iov_base = current->reply_data; |
| vec[1].iov_len = current->reply_size; |
| |
| if ((ret = writev( get_unix_fd( current->reply_fd ), vec, 2 )) < sizeof(*reply)) goto error; |
| |
| if ((current->reply_towrite = current->reply_size - (ret - sizeof(*reply)))) |
| { |
| /* couldn't write it all, wait for POLLOUT */ |
| set_fd_events( current->reply_fd, POLLOUT ); |
| set_fd_events( current->request_fd, 0 ); |
| return; |
| } |
| } |
| free( current->reply_data ); |
| current->reply_data = NULL; |
| return; |
| |
| error: |
| if (ret >= 0) |
| fatal_protocol_error( current, "partial write %d\n", ret ); |
| else if (errno == EPIPE) |
| kill_thread( current, 0 ); /* normal death */ |
| else |
| fatal_protocol_error( current, "reply write: %s\n", strerror( errno )); |
| } |
| |
| /* call a request handler */ |
| static void call_req_handler( struct thread *thread ) |
| { |
| union generic_reply reply; |
| enum request req = thread->req.request_header.req; |
| |
| current = thread; |
| current->reply_size = 0; |
| clear_error(); |
| memset( &reply, 0, sizeof(reply) ); |
| |
| if (debug_level) trace_request(); |
| |
| if (req < REQ_NB_REQUESTS) |
| req_handlers[req]( ¤t->req, &reply ); |
| else |
| set_error( STATUS_NOT_IMPLEMENTED ); |
| |
| if (current) |
| { |
| if (current->reply_fd) |
| { |
| reply.reply_header.error = current->error; |
| reply.reply_header.reply_size = current->reply_size; |
| if (debug_level) trace_reply( req, &reply ); |
| send_reply( &reply ); |
| } |
| else |
| { |
| current->exit_code = 1; |
| kill_thread( current, 1 ); /* no way to continue without reply fd */ |
| } |
| } |
| current = NULL; |
| } |
| |
| /* read a request from a thread */ |
| void read_request( struct thread *thread ) |
| { |
| int ret; |
| |
| if (!thread->req_toread) /* no pending request */ |
| { |
| if ((ret = read( get_unix_fd( thread->request_fd ), &thread->req, |
| sizeof(thread->req) )) != sizeof(thread->req)) goto error; |
| if (!(thread->req_toread = thread->req.request_header.request_size)) |
| { |
| /* no data, handle request at once */ |
| call_req_handler( thread ); |
| return; |
| } |
| if (!(thread->req_data = malloc( thread->req_toread ))) |
| { |
| fatal_protocol_error( thread, "no memory for %u bytes request %d\n", |
| thread->req_toread, thread->req.request_header.req ); |
| return; |
| } |
| } |
| |
| /* read the variable sized data */ |
| for (;;) |
| { |
| ret = read( get_unix_fd( thread->request_fd ), |
| (char *)thread->req_data + thread->req.request_header.request_size |
| - thread->req_toread, |
| thread->req_toread ); |
| if (ret <= 0) break; |
| if (!(thread->req_toread -= ret)) |
| { |
| call_req_handler( thread ); |
| free( thread->req_data ); |
| thread->req_data = NULL; |
| return; |
| } |
| } |
| |
| error: |
| if (!ret) /* closed pipe */ |
| kill_thread( thread, 0 ); |
| else if (ret > 0) |
| fatal_protocol_error( thread, "partial read %d\n", ret ); |
| else if (errno != EWOULDBLOCK && (EWOULDBLOCK == EAGAIN || errno != EAGAIN)) |
| fatal_protocol_error( thread, "read: %s\n", strerror( errno )); |
| } |
| |
| /* receive a file descriptor on the process socket */ |
| int receive_fd( struct process *process ) |
| { |
| struct iovec vec; |
| struct send_fd data; |
| struct msghdr msghdr; |
| int fd = -1, ret; |
| |
| #ifdef HAVE_STRUCT_MSGHDR_MSG_ACCRIGHTS |
| msghdr.msg_accrightslen = sizeof(int); |
| msghdr.msg_accrights = (void *)&fd; |
| #else /* HAVE_STRUCT_MSGHDR_MSG_ACCRIGHTS */ |
| char cmsg_buffer[256]; |
| msghdr.msg_control = cmsg_buffer; |
| msghdr.msg_controllen = sizeof(cmsg_buffer); |
| msghdr.msg_flags = 0; |
| #endif /* HAVE_STRUCT_MSGHDR_MSG_ACCRIGHTS */ |
| |
| msghdr.msg_name = NULL; |
| msghdr.msg_namelen = 0; |
| msghdr.msg_iov = &vec; |
| msghdr.msg_iovlen = 1; |
| vec.iov_base = (void *)&data; |
| vec.iov_len = sizeof(data); |
| |
| ret = recvmsg( get_unix_fd( process->msg_fd ), &msghdr, 0 ); |
| |
| #ifndef HAVE_STRUCT_MSGHDR_MSG_ACCRIGHTS |
| if (ret > 0) |
| { |
| struct cmsghdr *cmsg; |
| for (cmsg = CMSG_FIRSTHDR( &msghdr ); cmsg; cmsg = CMSG_NXTHDR( &msghdr, cmsg )) |
| { |
| if (cmsg->cmsg_level != SOL_SOCKET) continue; |
| if (cmsg->cmsg_type == SCM_RIGHTS) fd = *(int *)CMSG_DATA(cmsg); |
| } |
| } |
| #endif /* HAVE_STRUCT_MSGHDR_MSG_ACCRIGHTS */ |
| |
| if (ret == sizeof(data)) |
| { |
| struct thread *thread; |
| |
| if (data.tid) thread = get_thread_from_id( data.tid ); |
| else thread = (struct thread *)grab_object( get_process_first_thread( process )); |
| |
| if (!thread || thread->process != process || thread->state == TERMINATED) |
| { |
| if (debug_level) |
| fprintf( stderr, "%04x: *fd* %d <- %d bad thread id\n", |
| data.tid, data.fd, fd ); |
| close( fd ); |
| } |
| else |
| { |
| if (debug_level) |
| fprintf( stderr, "%04x: *fd* %d <- %d\n", |
| thread->id, data.fd, fd ); |
| thread_add_inflight_fd( thread, data.fd, fd ); |
| } |
| if (thread) release_object( thread ); |
| return 0; |
| } |
| |
| if (!ret) |
| { |
| kill_process( process, 0 ); |
| } |
| else if (ret > 0) |
| { |
| fprintf( stderr, "Protocol error: process %04x: partial recvmsg %d for fd\n", |
| process->id, ret ); |
| if (fd != -1) close( fd ); |
| kill_process( process, 1 ); |
| } |
| else |
| { |
| if (errno != EWOULDBLOCK && (EWOULDBLOCK == EAGAIN || errno != EAGAIN)) |
| { |
| fprintf( stderr, "Protocol error: process %04x: ", process->id ); |
| perror( "recvmsg" ); |
| kill_process( process, 1 ); |
| } |
| } |
| return -1; |
| } |
| |
| /* send an fd to a client */ |
| int send_client_fd( struct process *process, int fd, obj_handle_t handle ) |
| { |
| struct iovec vec; |
| struct msghdr msghdr; |
| int ret; |
| |
| #ifdef HAVE_STRUCT_MSGHDR_MSG_ACCRIGHTS |
| msghdr.msg_accrightslen = sizeof(fd); |
| msghdr.msg_accrights = (void *)&fd; |
| #else /* HAVE_STRUCT_MSGHDR_MSG_ACCRIGHTS */ |
| char cmsg_buffer[256]; |
| struct cmsghdr *cmsg; |
| msghdr.msg_control = cmsg_buffer; |
| msghdr.msg_controllen = sizeof(cmsg_buffer); |
| msghdr.msg_flags = 0; |
| cmsg = CMSG_FIRSTHDR( &msghdr ); |
| cmsg->cmsg_len = CMSG_LEN( sizeof(fd) ); |
| cmsg->cmsg_level = SOL_SOCKET; |
| cmsg->cmsg_type = SCM_RIGHTS; |
| *(int *)CMSG_DATA(cmsg) = fd; |
| msghdr.msg_controllen = cmsg->cmsg_len; |
| #endif /* HAVE_STRUCT_MSGHDR_MSG_ACCRIGHTS */ |
| |
| msghdr.msg_name = NULL; |
| msghdr.msg_namelen = 0; |
| msghdr.msg_iov = &vec; |
| msghdr.msg_iovlen = 1; |
| |
| vec.iov_base = (void *)&handle; |
| vec.iov_len = sizeof(handle); |
| |
| if (debug_level) |
| fprintf( stderr, "%04x: *fd* %04x -> %d\n", current ? current->id : process->id, handle, fd ); |
| |
| ret = sendmsg( get_unix_fd( process->msg_fd ), &msghdr, 0 ); |
| |
| if (ret == sizeof(handle)) return 0; |
| |
| if (ret >= 0) |
| { |
| fprintf( stderr, "Protocol error: process %04x: partial sendmsg %d\n", process->id, ret ); |
| kill_process( process, 1 ); |
| } |
| else if (errno == EPIPE) |
| { |
| kill_process( process, 0 ); |
| } |
| else |
| { |
| fprintf( stderr, "Protocol error: process %04x: ", process->id ); |
| perror( "sendmsg" ); |
| kill_process( process, 1 ); |
| } |
| return -1; |
| } |
| |
| /* get current tick count to return to client */ |
| unsigned int get_tick_count(void) |
| { |
| #ifdef __APPLE__ |
| static mach_timebase_info_data_t timebase; |
| |
| if (!timebase.denom) mach_timebase_info( &timebase ); |
| return mach_absolute_time() * timebase.numer / timebase.denom / 1000000; |
| #elif defined(HAVE_CLOCK_GETTIME) |
| struct timespec ts; |
| #ifdef CLOCK_MONOTONIC_RAW |
| if (!clock_gettime( CLOCK_MONOTONIC_RAW, &ts )) |
| return ts.tv_sec * 1000 + ts.tv_nsec / 1000000; |
| #endif |
| if (!clock_gettime( CLOCK_MONOTONIC, &ts )) |
| return ts.tv_sec * 1000 + ts.tv_nsec / 1000000; |
| #endif |
| return (current_time - server_start_time) / 10000; |
| } |
| |
| static void master_socket_dump( struct object *obj, int verbose ) |
| { |
| struct master_socket *sock = (struct master_socket *)obj; |
| assert( obj->ops == &master_socket_ops ); |
| fprintf( stderr, "Master socket fd=%p\n", sock->fd ); |
| } |
| |
| static void master_socket_destroy( struct object *obj ) |
| { |
| struct master_socket *sock = (struct master_socket *)obj; |
| assert( obj->ops == &master_socket_ops ); |
| release_object( sock->fd ); |
| } |
| |
| /* handle a socket event */ |
| static void master_socket_poll_event( struct fd *fd, int event ) |
| { |
| struct master_socket *sock = get_fd_user( fd ); |
| assert( master_socket->obj.ops == &master_socket_ops ); |
| |
| assert( sock == master_socket ); /* there is only one master socket */ |
| |
| if (event & (POLLERR | POLLHUP)) |
| { |
| /* this is not supposed to happen */ |
| fprintf( stderr, "wineserver: Error on master socket\n" ); |
| set_fd_events( sock->fd, -1 ); |
| } |
| else if (event & POLLIN) |
| { |
| struct sockaddr_un dummy; |
| socklen_t len = sizeof(dummy); |
| int client = accept( get_unix_fd( master_socket->fd ), (struct sockaddr *) &dummy, &len ); |
| if (client == -1) return; |
| fcntl( client, F_SETFL, O_NONBLOCK ); |
| create_process( client, NULL, 0 ); |
| } |
| } |
| |
| /* remove the socket upon exit */ |
| static void socket_cleanup(void) |
| { |
| static int do_it_once; |
| if (!do_it_once++) unlink( server_socket_name ); |
| } |
| |
| /* create a directory and check its permissions */ |
| static void create_dir( const char *name, struct stat *st ) |
| { |
| if (lstat( name, st ) == -1) |
| { |
| if (errno != ENOENT) |
| fatal_error( "lstat %s: %s", name, strerror( errno )); |
| if (mkdir( name, 0700 ) == -1 && errno != EEXIST) |
| fatal_error( "mkdir %s: %s\n", name, strerror( errno )); |
| if (lstat( name, st ) == -1) |
| fatal_error( "lstat %s: %s\n", name, strerror( errno )); |
| } |
| if (!S_ISDIR(st->st_mode)) fatal_error( "%s is not a directory\n", name ); |
| if (st->st_uid != getuid()) fatal_error( "%s is not owned by you\n", name ); |
| if (st->st_mode & 077) fatal_error( "%s must not be accessible by other users\n", name ); |
| } |
| |
| /* create the server directory and chdir to it */ |
| static void create_server_dir( const char *dir ) |
| { |
| char *p, *server_dir; |
| struct stat st, st2; |
| |
| if (!(server_dir = strdup( dir ))) fatal_error( "out of memory\n" ); |
| |
| /* first create the base directory if needed */ |
| |
| p = strrchr( server_dir, '/' ); |
| *p = 0; |
| create_dir( server_dir, &st ); |
| |
| /* now create the server directory */ |
| |
| *p = '/'; |
| create_dir( server_dir, &st ); |
| |
| if (chdir( server_dir ) == -1) |
| fatal_error( "chdir %s: %s\n", server_dir, strerror( errno )); |
| if ((server_dir_fd = open( ".", O_RDONLY )) == -1) |
| fatal_error( "open %s: %s\n", server_dir, strerror( errno )); |
| if (fstat( server_dir_fd, &st2 ) == -1) |
| fatal_error( "stat %s: %s\n", server_dir, strerror( errno )); |
| if (st.st_dev != st2.st_dev || st.st_ino != st2.st_ino) |
| fatal_error( "chdir did not end up in %s\n", server_dir ); |
| |
| free( server_dir ); |
| } |
| |
| /* create the lock file and return its file descriptor */ |
| static int create_server_lock(void) |
| { |
| struct stat st; |
| int fd; |
| |
| if (lstat( server_lock_name, &st ) == -1) |
| { |
| if (errno != ENOENT) |
| fatal_error( "lstat %s/%s: %s", wine_get_server_dir(), server_lock_name, strerror( errno )); |
| } |
| else |
| { |
| if (!S_ISREG(st.st_mode)) |
| fatal_error( "%s/%s is not a regular file\n", wine_get_server_dir(), server_lock_name ); |
| } |
| |
| if ((fd = open( server_lock_name, O_CREAT|O_TRUNC|O_WRONLY, 0600 )) == -1) |
| fatal_error( "error creating %s/%s: %s", wine_get_server_dir(), server_lock_name, strerror( errno )); |
| return fd; |
| } |
| |
| /* wait for the server lock */ |
| int wait_for_lock(void) |
| { |
| const char *server_dir = wine_get_server_dir(); |
| int fd, r; |
| struct flock fl; |
| |
| if (!server_dir) return 0; /* no server dir, so no lock to wait on */ |
| |
| create_server_dir( server_dir ); |
| fd = create_server_lock(); |
| |
| fl.l_type = F_WRLCK; |
| fl.l_whence = SEEK_SET; |
| fl.l_start = 0; |
| fl.l_len = 1; |
| r = fcntl( fd, F_SETLKW, &fl ); |
| close(fd); |
| |
| return r; |
| } |
| |
| /* kill the wine server holding the lock */ |
| int kill_lock_owner( int sig ) |
| { |
| const char *server_dir = wine_get_server_dir(); |
| int fd, i, ret = 0; |
| pid_t pid = 0; |
| struct flock fl; |
| |
| if (!server_dir) return 0; /* no server dir, nothing to do */ |
| |
| create_server_dir( server_dir ); |
| fd = create_server_lock(); |
| |
| for (i = 1; i <= 20; i++) |
| { |
| fl.l_type = F_WRLCK; |
| fl.l_whence = SEEK_SET; |
| fl.l_start = 0; |
| fl.l_len = 1; |
| if (fcntl( fd, F_GETLK, &fl ) == -1) goto done; |
| if (fl.l_type != F_WRLCK) goto done; /* the file is not locked */ |
| if (!pid) /* first time around */ |
| { |
| if (!(pid = fl.l_pid)) goto done; /* shouldn't happen */ |
| if (sig == -1) |
| { |
| if (kill( pid, SIGINT ) == -1) goto done; |
| kill( pid, SIGCONT ); |
| ret = 1; |
| } |
| else /* just send the specified signal and return */ |
| { |
| ret = (kill( pid, sig ) != -1); |
| goto done; |
| } |
| } |
| else if (fl.l_pid != pid) goto done; /* no longer the same process */ |
| usleep( 50000 * i ); |
| } |
| /* waited long enough, now kill it */ |
| kill( pid, SIGKILL ); |
| |
| done: |
| close( fd ); |
| return ret; |
| } |
| |
| /* acquire the main server lock */ |
| static void acquire_lock(void) |
| { |
| struct sockaddr_un addr; |
| struct stat st; |
| struct flock fl; |
| int fd, slen, got_lock = 0; |
| |
| fd = create_server_lock(); |
| |
| fl.l_type = F_WRLCK; |
| fl.l_whence = SEEK_SET; |
| fl.l_start = 0; |
| fl.l_len = 1; |
| if (fcntl( fd, F_SETLK, &fl ) != -1) |
| { |
| /* check for crashed server */ |
| if (stat( server_socket_name, &st ) != -1 && /* there is a leftover socket */ |
| stat( "core", &st ) != -1 && st.st_size) /* and there is a non-empty core file */ |
| { |
| fprintf( stderr, |
| "Warning: a previous instance of the wine server seems to have crashed.\n" |
| "Please run 'gdb %s %s/core',\n" |
| "type 'backtrace' at the gdb prompt and report the results. Thanks.\n\n", |
| server_argv0, wine_get_server_dir() ); |
| } |
| unlink( server_socket_name ); /* we got the lock, we can safely remove the socket */ |
| got_lock = 1; |
| /* in that case we reuse fd without closing it, this ensures |
| * that we hold the lock until the process exits */ |
| } |
| else |
| { |
| switch(errno) |
| { |
| case ENOLCK: |
| break; |
| case EACCES: |
| /* check whether locks work at all on this file system */ |
| if (fcntl( fd, F_GETLK, &fl ) == -1) break; |
| /* fall through */ |
| case EAGAIN: |
| exit(2); /* we didn't get the lock, exit with special status */ |
| default: |
| fatal_error( "fcntl %s/%s: %s", wine_get_server_dir(), server_lock_name, strerror( errno )); |
| } |
| /* it seems we can't use locks on this fs, so we will use the socket existence as lock */ |
| close( fd ); |
| } |
| |
| if ((fd = socket( AF_UNIX, SOCK_STREAM, 0 )) == -1) fatal_error( "socket: %s\n", strerror( errno )); |
| addr.sun_family = AF_UNIX; |
| strcpy( addr.sun_path, server_socket_name ); |
| slen = sizeof(addr) - sizeof(addr.sun_path) + strlen(addr.sun_path) + 1; |
| #ifdef HAVE_STRUCT_SOCKADDR_UN_SUN_LEN |
| addr.sun_len = slen; |
| #endif |
| if (bind( fd, (struct sockaddr *)&addr, slen ) == -1) |
| { |
| if ((errno == EEXIST) || (errno == EADDRINUSE)) |
| { |
| if (got_lock) |
| fatal_error( "couldn't bind to the socket even though we hold the lock\n" ); |
| exit(2); /* we didn't get the lock, exit with special status */ |
| } |
| fatal_error( "bind: %s\n", strerror( errno )); |
| } |
| atexit( socket_cleanup ); |
| chmod( server_socket_name, 0600 ); /* make sure no other user can connect */ |
| if (listen( fd, 5 ) == -1) fatal_error( "listen: %s\n", strerror( errno )); |
| |
| if (!(master_socket = alloc_object( &master_socket_ops )) || |
| !(master_socket->fd = create_anonymous_fd( &master_socket_fd_ops, fd, &master_socket->obj, 0 ))) |
| fatal_error( "out of memory\n" ); |
| set_fd_events( master_socket->fd, POLLIN ); |
| make_object_static( &master_socket->obj ); |
| } |
| |
| /* open the master server socket and start waiting for new clients */ |
| void open_master_socket(void) |
| { |
| const char *server_dir = wine_get_server_dir(); |
| const char *config_dir = wine_get_config_dir(); |
| int fd, pid, status, sync_pipe[2]; |
| char dummy; |
| |
| /* make sure no request is larger than the maximum size */ |
| assert( sizeof(union generic_request) == sizeof(struct request_max_size) ); |
| assert( sizeof(union generic_reply) == sizeof(struct request_max_size) ); |
| |
| /* make sure the stdio fds are open */ |
| fd = open( "/dev/null", O_RDWR ); |
| while (fd >= 0 && fd <= 2) fd = dup( fd ); |
| |
| if (!server_dir) |
| fatal_error( "directory %s cannot be accessed\n", config_dir ); |
| if (chdir( config_dir ) == -1) |
| fatal_error( "chdir to %s: %s\n", config_dir, strerror( errno )); |
| if ((config_dir_fd = open( ".", O_RDONLY )) == -1) |
| fatal_error( "open %s: %s\n", config_dir, strerror( errno )); |
| |
| create_server_dir( server_dir ); |
| |
| if (!foreground) |
| { |
| if (pipe( sync_pipe ) == -1) fatal_error( "pipe: %s\n", strerror( errno )); |
| pid = fork(); |
| switch( pid ) |
| { |
| case 0: /* child */ |
| setsid(); |
| close( sync_pipe[0] ); |
| |
| acquire_lock(); |
| |
| /* close stdin and stdout */ |
| dup2( fd, 0 ); |
| dup2( fd, 1 ); |
| |
| /* signal parent */ |
| dummy = 0; |
| write( sync_pipe[1], &dummy, 1 ); |
| close( sync_pipe[1] ); |
| break; |
| |
| case -1: |
| fatal_error( "fork: %s\n", strerror( errno )); |
| break; |
| |
| default: /* parent */ |
| close( sync_pipe[1] ); |
| |
| /* wait for child to signal us and then exit */ |
| if (read( sync_pipe[0], &dummy, 1 ) == 1) _exit(0); |
| |
| /* child terminated, propagate exit status */ |
| waitpid( pid, &status, 0 ); |
| if (WIFEXITED(status)) _exit( WEXITSTATUS(status) ); |
| _exit(1); |
| } |
| } |
| else /* remain in the foreground */ |
| { |
| acquire_lock(); |
| } |
| |
| /* init the process tracing mechanism */ |
| init_tracing_mechanism(); |
| close( fd ); |
| } |
| |
| /* master socket timer expiration handler */ |
| static void close_socket_timeout( void *arg ) |
| { |
| master_timeout = NULL; |
| flush_registry(); |
| if (debug_level) fprintf( stderr, "wineserver: exiting (pid=%ld)\n", (long) getpid() ); |
| |
| #ifdef DEBUG_OBJECTS |
| close_objects(); /* shut down everything properly */ |
| #endif |
| exit( 0 ); |
| } |
| |
| /* close the master socket and stop waiting for new clients */ |
| void close_master_socket( timeout_t timeout ) |
| { |
| if (master_socket) |
| { |
| release_object( master_socket ); |
| master_socket = NULL; |
| } |
| if (master_timeout) /* cancel previous timeout */ |
| remove_timeout_user( master_timeout ); |
| |
| master_timeout = add_timeout_user( timeout, close_socket_timeout, NULL ); |
| } |