/* w32-io.c - W32 API I/O functions. Copyright (C) 2000 Werner Koch (dd9jn) Copyright (C) 2001, 2002, 2003, 2004 g10 Code GmbH This file is part of GPGME. GPGME 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. GPGME 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 program; if not, write to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02111-1307, USA. */ #ifdef HAVE_CONFIG_H #include #endif #include #include #include #include #include #include #include #include #include #include #include #include "util.h" #include "sema.h" #include "priv-io.h" #include "debug.h" /* We assume that a HANDLE can be represented by an int which should be true for all i386 systems (HANDLE is defined as void *) and these are the only systems for which Windows is available. Further we assume that -1 denotes an invalid handle. */ #define fd_to_handle(a) ((HANDLE)(a)) #define handle_to_fd(a) ((int)(a)) #define pid_to_handle(a) ((HANDLE)(a)) #define handle_to_pid(a) ((int)(a)) #define READBUF_SIZE 4096 #define WRITEBUF_SIZE 4096 #define PIPEBUF_SIZE 4096 #define MAX_READERS 20 #define MAX_WRITERS 20 static struct { int inuse; int fd; void (*handler)(int,void*); void *value; } notify_table[256]; DEFINE_STATIC_LOCK (notify_table_lock); struct reader_context_s { HANDLE file_hd; HANDLE thread_hd; DECLARE_LOCK (mutex); int stop_me; int eof; int eof_shortcut; int error; int error_code; HANDLE have_data_ev; /* manually reset */ HANDLE have_space_ev; /* auto reset */ HANDLE stopped; size_t readpos, writepos; char buffer[READBUF_SIZE]; }; static struct { volatile int used; int fd; struct reader_context_s *context; } reader_table[MAX_READERS]; static int reader_table_size= MAX_READERS; DEFINE_STATIC_LOCK (reader_table_lock); struct writer_context_s { HANDLE file_hd; HANDLE thread_hd; DECLARE_LOCK (mutex); int stop_me; int error; int error_code; HANDLE have_data; /* manually reset */ HANDLE is_empty; HANDLE stopped; size_t nbytes; char buffer[WRITEBUF_SIZE]; }; static struct { volatile int used; int fd; struct writer_context_s *context; } writer_table[MAX_WRITERS]; static int writer_table_size= MAX_WRITERS; DEFINE_STATIC_LOCK (writer_table_lock); static int get_desired_thread_priority (void) { int value; if (!_gpgme_get_conf_int ("IOThreadPriority", &value)) { value = THREAD_PRIORITY_HIGHEST; DEBUG1 ("** Using standard IOThreadPriority of %d\n", value); } else DEBUG1 ("** Configured IOThreadPriority is %d\n", value); return value; } static HANDLE set_synchronize (HANDLE h) { HANDLE tmp; /* For NT we have to set the sync flag. It seems that the only * way to do it is by duplicating the handle. Tsss.. */ if (!DuplicateHandle( GetCurrentProcess(), h, GetCurrentProcess(), &tmp, EVENT_MODIFY_STATE|SYNCHRONIZE, FALSE, 0 ) ) { DEBUG1 ("** Set SYNCRONIZE failed: ec=%d\n", (int)GetLastError()); } else { CloseHandle (h); h = tmp; } return h; } static DWORD CALLBACK reader (void *arg) { struct reader_context_s *c = arg; int nbytes; DWORD nread; DEBUG2 ("reader thread %p for file %p started", c->thread_hd, c->file_hd ); for (;;) { LOCK (c->mutex); /* leave a 1 byte gap so that we can see whether it is empty or full*/ if ((c->writepos + 1) % READBUF_SIZE == c->readpos) { /* wait for space */ if (!ResetEvent (c->have_space_ev) ) DEBUG1 ("ResetEvent failed: ec=%d", (int)GetLastError ()); UNLOCK (c->mutex); DEBUG1 ("reader thread %p: waiting for space ...", c->thread_hd ); WaitForSingleObject (c->have_space_ev, INFINITE); DEBUG1 ("reader thread %p: got space", c->thread_hd ); LOCK (c->mutex); } if ( c->stop_me ) { UNLOCK (c->mutex); break; } nbytes = (c->readpos + READBUF_SIZE - c->writepos-1) % READBUF_SIZE; if ( nbytes > READBUF_SIZE - c->writepos ) nbytes = READBUF_SIZE - c->writepos; UNLOCK (c->mutex); DEBUG2 ("reader thread %p: reading %d bytes", c->thread_hd, nbytes ); if ( !ReadFile ( c->file_hd, c->buffer+c->writepos, nbytes, &nread, NULL) ) { c->error_code = (int)GetLastError (); if (c->error_code == ERROR_BROKEN_PIPE ) { c->eof=1; DEBUG1 ("reader thread %p: got eof (broken pipe)", c->thread_hd ); } else { c->error = 1; DEBUG2 ("reader thread %p: read error: ec=%d", c->thread_hd, c->error_code ); } break; } if ( !nread ) { c->eof = 1; DEBUG1 ("reader thread %p: got eof", c->thread_hd ); break; } DEBUG2 ("reader thread %p: got %d bytes", c->thread_hd, (int)nread ); LOCK (c->mutex); if (c->stop_me) { UNLOCK (c->mutex); break; } c->writepos = (c->writepos + nread) % READBUF_SIZE; if ( !SetEvent (c->have_data_ev) ) DEBUG1 ("SetEvent failed: ec=%d", (int)GetLastError ()); UNLOCK (c->mutex); } /* indicate that we have an error or eof */ if ( !SetEvent (c->have_data_ev) ) DEBUG1 ("SetEvent failed: ec=%d", (int)GetLastError ()); DEBUG1 ("reader thread %p ended", c->thread_hd ); SetEvent (c->stopped); return 0; } static struct reader_context_s * create_reader (HANDLE fd) { struct reader_context_s *c; SECURITY_ATTRIBUTES sec_attr; DWORD tid; DEBUG1 ("creating new read thread for file handle %p", fd ); memset (&sec_attr, 0, sizeof sec_attr ); sec_attr.nLength = sizeof sec_attr; sec_attr.bInheritHandle = FALSE; c = calloc (1, sizeof *c ); if (!c) return NULL; c->file_hd = fd; c->have_data_ev = CreateEvent (&sec_attr, TRUE, FALSE, NULL); c->have_space_ev = CreateEvent (&sec_attr, FALSE, TRUE, NULL); c->stopped = CreateEvent (&sec_attr, TRUE, FALSE, NULL); if (!c->have_data_ev || !c->have_space_ev || !c->stopped ) { DEBUG1 ("** CreateEvent failed: ec=%d\n", (int)GetLastError ()); if (c->have_data_ev) CloseHandle (c->have_data_ev); if (c->have_space_ev) CloseHandle (c->have_space_ev); if (c->stopped) CloseHandle (c->stopped); free (c); return NULL; } c->have_data_ev = set_synchronize (c->have_data_ev); INIT_LOCK (c->mutex); c->thread_hd = CreateThread (&sec_attr, 0, reader, c, 0, &tid ); if (!c->thread_hd) { DEBUG1 ("** failed to create reader thread: ec=%d\n", (int)GetLastError ()); DESTROY_LOCK (c->mutex); if (c->have_data_ev) CloseHandle (c->have_data_ev); if (c->have_space_ev) CloseHandle (c->have_space_ev); if (c->stopped) CloseHandle (c->stopped); free (c); return NULL; } else { /* We set the priority of the thread higher because we know that it only runs for a short time. This greatly helps to increase the performance of the I/O. */ SetThreadPriority (c->thread_hd, get_desired_thread_priority ()); } return c; } static void destroy_reader (struct reader_context_s *c) { LOCK (c->mutex); c->stop_me = 1; if (c->have_space_ev) SetEvent (c->have_space_ev); UNLOCK (c->mutex); DEBUG1 ("waiting for thread %p termination ...", c->thread_hd ); WaitForSingleObject (c->stopped, INFINITE); DEBUG1 ("thread %p has terminated", c->thread_hd ); if (c->stopped) CloseHandle (c->stopped); if (c->have_data_ev) CloseHandle (c->have_data_ev); if (c->have_space_ev) CloseHandle (c->have_space_ev); CloseHandle (c->thread_hd); DESTROY_LOCK (c->mutex); free (c); } /* * Find a reader context or create a new one * Note that the reader context will last until a io_close. */ static struct reader_context_s * find_reader (int fd, int start_it) { int i; for (i=0; i < reader_table_size ; i++ ) { if ( reader_table[i].used && reader_table[i].fd == fd ) return reader_table[i].context; } if (!start_it) return NULL; LOCK (reader_table_lock); for (i=0; i < reader_table_size; i++ ) { if (!reader_table[i].used) { reader_table[i].fd = fd; reader_table[i].context = create_reader (fd_to_handle (fd)); reader_table[i].used = 1; UNLOCK (reader_table_lock); return reader_table[i].context; } } UNLOCK (reader_table_lock); return NULL; } static void kill_reader (int fd) { int i; LOCK (reader_table_lock); for (i=0; i < reader_table_size; i++ ) { if (reader_table[i].used && reader_table[i].fd == fd ) { destroy_reader (reader_table[i].context); reader_table[i].context = NULL; reader_table[i].used = 0; break; } } UNLOCK (reader_table_lock); } int _gpgme_io_read ( int fd, void *buffer, size_t count ) { int nread; struct reader_context_s *c = find_reader (fd,1); DEBUG2 ("fd %d: about to read %d bytes\n", fd, (int)count ); if ( !c ) { DEBUG0 ( "no reader thread\n"); return -1; } if (c->eof_shortcut) { DEBUG1 ("fd %d: EOF (again)", fd ); return 0; } LOCK (c->mutex); if (c->readpos == c->writepos && !c->error) { /*no data avail*/ UNLOCK (c->mutex); DEBUG2 ("fd %d: waiting for data from thread %p", fd, c->thread_hd); WaitForSingleObject (c->have_data_ev, INFINITE); DEBUG2 ("fd %d: data from thread %p available", fd, c->thread_hd); LOCK (c->mutex); } if (c->readpos == c->writepos || c->error) { UNLOCK (c->mutex); c->eof_shortcut = 1; if (c->eof) { DEBUG1 ("fd %d: EOF", fd ); return 0; } if (!c->error) { DEBUG1 ("fd %d: EOF but eof flag not set", fd ); return 0; } DEBUG1 ("fd %d: read error", fd ); return -1; } nread = c->readpos < c->writepos? c->writepos - c->readpos : READBUF_SIZE - c->readpos; if (nread > count) nread = count; memcpy (buffer, c->buffer+c->readpos, nread); c->readpos = (c->readpos + nread) % READBUF_SIZE; if (c->readpos == c->writepos && !c->eof) { if ( !ResetEvent (c->have_data_ev) ) DEBUG1 ("ResetEvent failed: ec=%d", (int)GetLastError ()); } if (!SetEvent (c->have_space_ev)) DEBUG1 ("SetEvent failed: ec=%d", (int)GetLastError ()); UNLOCK (c->mutex); DEBUG2 ("fd %d: got %d bytes\n", fd, nread ); if (nread > 0) _gpgme_debug (2, "fd %d: got `%.*s'\n", fd, nread, buffer); return nread; } /* * The writer does use a simple buffering strategy so that we are * informed about write errors as soon as possible (i.e. with the the * next call to the write function */ static DWORD CALLBACK writer (void *arg) { struct writer_context_s *c = arg; DWORD nwritten; DEBUG2 ("writer thread %p for file %p started", c->thread_hd, c->file_hd ); for (;;) { LOCK (c->mutex); if ( c->stop_me ) { UNLOCK (c->mutex); break; } if ( !c->nbytes ) { if (!SetEvent (c->is_empty)) DEBUG1 ("SetEvent failed: ec=%d", (int)GetLastError ()); if (!ResetEvent (c->have_data) ) DEBUG1 ("ResetEvent failed: ec=%d", (int)GetLastError ()); UNLOCK (c->mutex); DEBUG1 ("writer thread %p: idle ...", c->thread_hd ); WaitForSingleObject (c->have_data, INFINITE); DEBUG1 ("writer thread %p: got data to send", c->thread_hd ); LOCK (c->mutex); } if ( c->stop_me ) { UNLOCK (c->mutex); break; } UNLOCK (c->mutex); DEBUG2 ("writer thread %p: writing %d bytes", c->thread_hd, c->nbytes ); if ( c->nbytes && !WriteFile ( c->file_hd, c->buffer, c->nbytes, &nwritten, NULL)) { c->error_code = (int)GetLastError (); c->error = 1; DEBUG2 ("writer thread %p: write error: ec=%d", c->thread_hd, c->error_code ); break; } DEBUG2 ("writer thread %p: wrote %d bytes", c->thread_hd, (int)nwritten ); LOCK (c->mutex); c->nbytes -= nwritten; UNLOCK (c->mutex); } /* indicate that we have an error */ if ( !SetEvent (c->is_empty) ) DEBUG1 ("SetEvent failed: ec=%d", (int)GetLastError ()); DEBUG1 ("writer thread %p ended", c->thread_hd ); SetEvent (c->stopped); return 0; } static struct writer_context_s * create_writer (HANDLE fd) { struct writer_context_s *c; SECURITY_ATTRIBUTES sec_attr; DWORD tid; DEBUG1 ("creating new write thread for file handle %p", fd ); memset (&sec_attr, 0, sizeof sec_attr ); sec_attr.nLength = sizeof sec_attr; sec_attr.bInheritHandle = FALSE; c = calloc (1, sizeof *c ); if (!c) return NULL; c->file_hd = fd; c->have_data = CreateEvent (&sec_attr, TRUE, FALSE, NULL); c->is_empty = CreateEvent (&sec_attr, TRUE, TRUE, NULL); c->stopped = CreateEvent (&sec_attr, TRUE, FALSE, NULL); if (!c->have_data || !c->is_empty || !c->stopped ) { DEBUG1 ("** CreateEvent failed: ec=%d\n", (int)GetLastError ()); if (c->have_data) CloseHandle (c->have_data); if (c->is_empty) CloseHandle (c->is_empty); if (c->stopped) CloseHandle (c->stopped); free (c); return NULL; } c->is_empty = set_synchronize (c->is_empty); INIT_LOCK (c->mutex); c->thread_hd = CreateThread (&sec_attr, 0, writer, c, 0, &tid ); if (!c->thread_hd) { DEBUG1 ("** failed to create writer thread: ec=%d\n", (int)GetLastError ()); DESTROY_LOCK (c->mutex); if (c->have_data) CloseHandle (c->have_data); if (c->is_empty) CloseHandle (c->is_empty); if (c->stopped) CloseHandle (c->stopped); free (c); return NULL; } else { /* We set the priority of the thread higher because we know that it only runs for a short time. This greatly helps to increase the performance of the I/O. */ SetThreadPriority (c->thread_hd, get_desired_thread_priority ()); } return c; } static void destroy_writer (struct writer_context_s *c) { LOCK (c->mutex); c->stop_me = 1; if (c->have_data) SetEvent (c->have_data); UNLOCK (c->mutex); DEBUG1 ("waiting for thread %p termination ...", c->thread_hd ); WaitForSingleObject (c->stopped, INFINITE); DEBUG1 ("thread %p has terminated", c->thread_hd ); if (c->stopped) CloseHandle (c->stopped); if (c->have_data) CloseHandle (c->have_data); if (c->is_empty) CloseHandle (c->is_empty); CloseHandle (c->thread_hd); DESTROY_LOCK (c->mutex); free (c); } /* * Find a writer context or create a new one * Note that the writer context will last until a io_close. */ static struct writer_context_s * find_writer (int fd, int start_it) { int i; for (i=0; i < writer_table_size ; i++ ) { if ( writer_table[i].used && writer_table[i].fd == fd ) return writer_table[i].context; } if (!start_it) return NULL; LOCK (writer_table_lock); for (i=0; i < writer_table_size; i++ ) { if (!writer_table[i].used) { writer_table[i].fd = fd; writer_table[i].context = create_writer (fd_to_handle (fd)); writer_table[i].used = 1; UNLOCK (writer_table_lock); return writer_table[i].context; } } UNLOCK (writer_table_lock); return NULL; } static void kill_writer (int fd) { int i; LOCK (writer_table_lock); for (i=0; i < writer_table_size; i++ ) { if (writer_table[i].used && writer_table[i].fd == fd ) { destroy_writer (writer_table[i].context); writer_table[i].context = NULL; writer_table[i].used = 0; break; } } UNLOCK (writer_table_lock); } int _gpgme_io_write ( int fd, const void *buffer, size_t count ) { struct writer_context_s *c = find_writer (fd,1); DEBUG2 ("fd %d: about to write %d bytes\n", fd, (int)count ); _gpgme_debug (2, "fd %d: write `%.*s'\n", fd, (int) count, buffer); if ( !c ) { DEBUG0 ( "no writer thread\n"); return -1; } LOCK (c->mutex); if ( c->nbytes ) { /* bytes are pending for send */ /* Reset the is_empty event. Better safe than sorry. */ if (!ResetEvent (c->is_empty)) DEBUG1 ("ResetEvent failed: ec=%d", (int)GetLastError ()); UNLOCK (c->mutex); DEBUG2 ("fd %d: waiting for empty buffer in thread %p", fd, c->thread_hd); WaitForSingleObject (c->is_empty, INFINITE); DEBUG2 ("fd %d: thread %p buffer is empty", fd, c->thread_hd); LOCK (c->mutex); } if ( c->error) { UNLOCK (c->mutex); DEBUG1 ("fd %d: write error", fd ); return -1; } /* If no error occured, the number of bytes in the buffer must be zero. */ assert (!c->nbytes); if (count > WRITEBUF_SIZE) count = WRITEBUF_SIZE; memcpy (c->buffer, buffer, count); c->nbytes = count; /* We have to reset the is_empty event early, because it is also used by the select() implementation to probe the channel. */ if (!ResetEvent (c->is_empty)) DEBUG1 ("ResetEvent failed: ec=%d", (int)GetLastError ()); if (!SetEvent (c->have_data)) DEBUG1 ("SetEvent failed: ec=%d", (int)GetLastError ()); UNLOCK (c->mutex); DEBUG2 ("fd %d: copied %d bytes\n", fd, (int)count ); return (int)count; } int _gpgme_io_pipe ( int filedes[2], int inherit_idx ) { HANDLE r, w; SECURITY_ATTRIBUTES sec_attr; memset (&sec_attr, 0, sizeof sec_attr ); sec_attr.nLength = sizeof sec_attr; sec_attr.bInheritHandle = FALSE; if (!CreatePipe ( &r, &w, &sec_attr, PIPEBUF_SIZE)) return -1; /* Make one end inheritable. */ if ( inherit_idx == 0 ) { HANDLE h; if (!DuplicateHandle( GetCurrentProcess(), r, GetCurrentProcess(), &h, 0, TRUE, DUPLICATE_SAME_ACCESS ) ) { DEBUG1 ("DuplicateHandle failed: ec=%d\n", (int)GetLastError()); CloseHandle (r); CloseHandle (w); return -1; } CloseHandle (r); r = h; } else if ( inherit_idx == 1 ) { HANDLE h; if (!DuplicateHandle( GetCurrentProcess(), w, GetCurrentProcess(), &h, 0, TRUE, DUPLICATE_SAME_ACCESS ) ) { DEBUG1 ("DuplicateHandle failed: ec=%d\n", (int)GetLastError()); CloseHandle (r); CloseHandle (w); return -1; } CloseHandle (w); w = h; } filedes[0] = handle_to_fd (r); filedes[1] = handle_to_fd (w); DEBUG5 ("CreatePipe %p %p %d %d inherit=%d\n", r, w, filedes[0], filedes[1], inherit_idx ); return 0; } int _gpgme_io_close ( int fd ) { int i; void (*handler)(int, void*) = NULL; void *value = NULL; if ( fd == -1 ) return -1; DEBUG1 ("** closing handle for fd %d\n", fd); kill_reader (fd); kill_writer (fd); LOCK (notify_table_lock); for ( i=0; i < DIM (notify_table); i++ ) { if (notify_table[i].inuse && notify_table[i].fd == fd) { handler = notify_table[i].handler; value = notify_table[i].value; notify_table[i].handler = NULL; notify_table[i].value = NULL; notify_table[i].inuse = 0; break; } } UNLOCK (notify_table_lock); if (handler) handler (fd, value); if ( !CloseHandle (fd_to_handle (fd)) ) { DEBUG2 ("CloseHandle for fd %d failed: ec=%d\n", fd, (int)GetLastError ()); return -1; } return 0; } int _gpgme_io_set_close_notify (int fd, void (*handler)(int, void*), void *value) { int i; assert (fd != -1); LOCK (notify_table_lock); for (i=0; i < DIM (notify_table); i++ ) { if ( notify_table[i].inuse && notify_table[i].fd == fd ) break; } if ( i == DIM (notify_table) ) { for (i=0; i < DIM (notify_table); i++ ) { if ( !notify_table[i].inuse ) break; } } if ( i == DIM (notify_table) ) { UNLOCK (notify_table_lock); return -1; } notify_table[i].fd = fd; notify_table[i].handler = handler; notify_table[i].value = value; notify_table[i].inuse = 1; UNLOCK (notify_table_lock); DEBUG2 ("set notification for fd %d (idx=%d)", fd, i ); return 0; } int _gpgme_io_set_nonblocking ( int fd ) { return 0; } static char * build_commandline (char **argv) { int i; int j; int n = 0; char *buf; char *p; /* We have to quote some things because under Windows the program parses the commandline and does some unquoting. We enclose the whole argument in double-quotes, and escape literal double-quotes as well as backslashes with a backslash. We end up with a trailing space at the end of the line, but that is harmless. */ for (i = 0; argv[i]; i++) { p = argv[i]; /* The leading double-quote. */ n++; while (*p) { /* An extra one for each literal that must be escaped. */ if (*p == '\\' || *p == '"') n++; n++; p++; } /* The trailing double-quote and the delimiter. */ n += 2; } /* And a trailing zero. */ n++; buf = p = malloc (n); if (!buf) return NULL; for (i = 0; argv[i]; i++) { char *argvp = argv[i]; *(p++) = '"'; while (*argvp) { if (*argvp == '\\' || *argvp == '"') *(p++) = '\\'; *(p++) = *(argvp++); } *(p++) = '"'; *(p++) = ' '; } *(p++) = 0; return buf; } int _gpgme_io_spawn ( const char *path, char **argv, struct spawn_fd_item_s *fd_child_list, struct spawn_fd_item_s *fd_parent_list ) { SECURITY_ATTRIBUTES sec_attr; PROCESS_INFORMATION pi = { NULL, /* returns process handle */ 0, /* returns primary thread handle */ 0, /* returns pid */ 0 /* returns tid */ }; STARTUPINFO si; char *envblock = NULL; int cr_flags = CREATE_DEFAULT_ERROR_MODE | GetPriorityClass (GetCurrentProcess ()); int i; char *arg_string; int duped_stdin = 0; int duped_stderr = 0; HANDLE hnul = INVALID_HANDLE_VALUE; /* FIXME. */ int debug_me = 0; memset (&sec_attr, 0, sizeof sec_attr ); sec_attr.nLength = sizeof sec_attr; sec_attr.bInheritHandle = FALSE; arg_string = build_commandline ( argv ); if (!arg_string ) return -1; memset (&si, 0, sizeof si); si.cb = sizeof (si); si.dwFlags = STARTF_USESTDHANDLES | STARTF_USESHOWWINDOW; si.wShowWindow = debug_me? SW_SHOW : SW_HIDE; si.hStdInput = GetStdHandle (STD_INPUT_HANDLE); si.hStdOutput = GetStdHandle (STD_OUTPUT_HANDLE); si.hStdError = GetStdHandle (STD_ERROR_HANDLE); for (i=0; fd_child_list[i].fd != -1; i++ ) { if (fd_child_list[i].dup_to == 0 ) { si.hStdInput = fd_to_handle (fd_child_list[i].fd); DEBUG1 ("using %d for stdin", fd_child_list[i].fd ); duped_stdin=1; } else if (fd_child_list[i].dup_to == 1 ) { si.hStdOutput = fd_to_handle (fd_child_list[i].fd); DEBUG1 ("using %d for stdout", fd_child_list[i].fd ); } else if (fd_child_list[i].dup_to == 2 ) { si.hStdError = fd_to_handle (fd_child_list[i].fd); DEBUG1 ("using %d for stderr", fd_child_list[i].fd ); duped_stderr = 1; } } if( !duped_stdin || !duped_stderr ) { SECURITY_ATTRIBUTES sa; memset (&sa, 0, sizeof sa ); sa.nLength = sizeof sa; sa.bInheritHandle = TRUE; hnul = CreateFile ( "nul", GENERIC_READ|GENERIC_WRITE, FILE_SHARE_READ|FILE_SHARE_WRITE, &sa, OPEN_EXISTING, FILE_ATTRIBUTE_NORMAL, NULL ); if ( hnul == INVALID_HANDLE_VALUE ) { DEBUG1 ("can't open `nul': ec=%d\n", (int)GetLastError ()); free (arg_string); return -1; } /* Make sure that the process has a connected stdin */ if ( !duped_stdin ) { si.hStdInput = hnul; DEBUG1 ("using %d for dummy stdin", (int)hnul ); } /* We normally don't want all the normal output */ if ( !duped_stderr ) { si.hStdError = hnul; DEBUG1 ("using %d for dummy stderr", (int)hnul ); } } DEBUG2 ("CreateProcess, path=`%s' args=`%s'", path, arg_string); cr_flags |= CREATE_SUSPENDED; if ( !CreateProcessA (path, arg_string, &sec_attr, /* process security attributes */ &sec_attr, /* thread security attributes */ TRUE, /* inherit handles */ cr_flags, /* creation flags */ envblock, /* environment */ NULL, /* use current drive/directory */ &si, /* startup information */ &pi /* returns process information */ ) ) { DEBUG1 ("CreateProcess failed: ec=%d\n", (int) GetLastError ()); free (arg_string); return -1; } /* Close the /dev/nul handle if used. */ if (hnul != INVALID_HANDLE_VALUE ) { if ( !CloseHandle ( hnul ) ) DEBUG1 ("CloseHandle(hnul) failed: ec=%d\n", (int)GetLastError()); } /* Close the other ends of the pipes. */ for (i = 0; fd_parent_list[i].fd != -1; i++) _gpgme_io_close (fd_parent_list[i].fd); DEBUG4 ("CreateProcess ready\n" "- hProcess=%p hThread=%p\n" "- dwProcessID=%d dwThreadId=%d\n", pi.hProcess, pi.hThread, (int) pi.dwProcessId, (int) pi.dwThreadId); if ( ResumeThread ( pi.hThread ) < 0 ) { DEBUG1 ("ResumeThread failed: ec=%d\n", (int)GetLastError ()); } if ( !CloseHandle (pi.hThread) ) { DEBUG1 ("CloseHandle of thread failed: ec=%d\n", (int)GetLastError ()); } return handle_to_pid (pi.hProcess); } /* * Select on the list of fds. * Returns: -1 = error * 0 = timeout or nothing to select * >0 = number of signaled fds */ int _gpgme_io_select ( struct io_select_fd_s *fds, size_t nfds, int nonblock ) { HANDLE waitbuf[MAXIMUM_WAIT_OBJECTS]; int waitidx[MAXIMUM_WAIT_OBJECTS]; int code, nwait; int i, any; int count; void *dbg_help; restart: DEBUG_BEGIN (dbg_help, 3, "select on [ "); any = 0; nwait = 0; count = 0; for ( i=0; i < nfds; i++ ) { if ( fds[i].fd == -1 ) continue; fds[i].signaled = 0; if ( fds[i].for_read || fds[i].for_write ) { if ( fds[i].frozen ) { DEBUG_ADD1 (dbg_help, "f%d ", fds[i].fd ); } else if ( fds[i].for_read ) { struct reader_context_s *c = find_reader (fds[i].fd,1); if (!c) { DEBUG1 ("oops: no reader thread for fd %d", fds[i].fd); } else { if ( nwait >= DIM (waitbuf) ) { DEBUG_END (dbg_help, "oops ]"); DEBUG0 ("Too many objects for WFMO!" ); return -1; } waitidx[nwait] = i; waitbuf[nwait++] = c->have_data_ev; } DEBUG_ADD1 (dbg_help, "r%d ", fds[i].fd ); any = 1; } else if ( fds[i].for_write ) { struct writer_context_s *c = find_writer (fds[i].fd,1); if (!c) { DEBUG1 ("oops: no writer thread for fd %d", fds[i].fd); } else { if ( nwait >= DIM (waitbuf) ) { DEBUG_END (dbg_help, "oops ]"); DEBUG0 ("Too many objects for WFMO!" ); return -1; } waitidx[nwait] = i; waitbuf[nwait++] = c->is_empty; } DEBUG_ADD1 (dbg_help, "w%d ", fds[i].fd ); any = 1; } } } DEBUG_END (dbg_help, "]"); if (!any) return 0; code = WaitForMultipleObjects ( nwait, waitbuf, 0, nonblock ? 0 : 1000); if ( code >= WAIT_OBJECT_0 && code < WAIT_OBJECT_0 + nwait ) { /* This WFMO is a really silly function: It does return either * the index of the signaled object or if 2 objects have been * signalled at the same time, the index of the object with the * lowest object is returned - so and how do we find out * how many objects have been signaled???. * The only solution I can imagine is to test each object starting * with the returned index individually - how dull. */ any = 0; for (i=code - WAIT_OBJECT_0; i < nwait; i++ ) { if (WaitForSingleObject (waitbuf[i], 0) == WAIT_OBJECT_0) { assert (waitidx[i] >=0 && waitidx[i] < nfds); fds[waitidx[i]].signaled = 1; any = 1; count++; } } if (!any) { DEBUG0 ("Oops: No signaled objects found after WFMO"); count = -1; } } else if ( code == WAIT_TIMEOUT ) { DEBUG0 ("WFMO timed out\n" ); } else if (code == WAIT_FAILED ) { int le = (int)GetLastError (); if ( le == ERROR_INVALID_HANDLE ) { int k, j = handle_to_fd (waitbuf[i]); DEBUG1 ("WFMO invalid handle %d removed\n", j); for (k=0 ; k < nfds; k++ ) { if ( fds[k].fd == j ) { fds[k].for_read = fds[k].for_write = 0; goto restart; } } DEBUG0 (" oops, or not???\n"); } DEBUG1 ("WFMO failed: %d\n", le ); count = -1; } else { DEBUG1 ("WFMO returned %d\n", code ); count = -1; } if ( count ) { DEBUG_BEGIN (dbg_help, 3, " signaled [ "); for ( i=0; i < nfds; i++ ) { if ( fds[i].fd == -1 ) continue; if ( (fds[i].for_read || fds[i].for_write) && fds[i].signaled ) { DEBUG_ADD2 (dbg_help, "%c%d ", fds[i].for_read? 'r':'w',fds[i].fd ); } } DEBUG_END (dbg_help, "]"); } return count; } void _gpgme_io_subsystem_init (void) { } /* Write the printable version of FD to the buffer BUF of length BUFLEN. The printable version is the representation on the command line that the child process expects. */ int _gpgme_io_fd2str (char *buf, int buflen, int fd) { return snprintf (buf, buflen, "%d", fd); } /* The following interface is only useful for GPGME Glib. */ /* Look up the giochannel for file descriptor FD. */ void * gpgme_get_giochannel (int fd) { return NULL; }