Sweden-Number/server/file.c

519 lines
15 KiB
C
Raw Normal View History

/*
* Server-side file management
*
* 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., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*/
#include "config.h"
#include "wine/port.h"
#include <assert.h>
#include <fcntl.h>
#include <stdarg.h>
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include <errno.h>
#ifdef HAVE_SYS_ERRNO_H
#include <sys/errno.h>
#endif
#include <sys/stat.h>
#include <sys/time.h>
#include <sys/types.h>
#include <time.h>
#include <unistd.h>
2003-02-11 23:27:13 +01:00
#ifdef HAVE_UTIME_H
#include <utime.h>
2003-02-11 23:27:13 +01:00
#endif
#include "winerror.h"
#include "windef.h"
#include "winbase.h"
#include "winreg.h"
#include "winternl.h"
#include "file.h"
#include "handle.h"
#include "thread.h"
#include "request.h"
#include "async.h"
struct file
{
struct object obj; /* object header */
struct fd *fd; /* file descriptor for this file */
unsigned int access; /* file access (GENERIC_READ/WRITE) */
unsigned int options; /* file options (FILE_DELETE_ON_CLOSE, FILE_SYNCHRONOUS...) */
struct async_queue read_q;
struct async_queue write_q;
};
static void file_dump( struct object *obj, int verbose );
static struct fd *file_get_fd( struct object *obj );
static void file_destroy( struct object *obj );
static int file_get_poll_events( struct fd *fd );
static void file_poll_event( struct fd *fd, int event );
static int file_flush( struct fd *fd, struct event **event );
2004-04-08 21:09:04 +02:00
static int file_get_info( struct fd *fd, int *flags );
static void file_queue_async( struct fd *fd, void *ptr, unsigned int status, int type, int count );
static const struct object_ops file_ops =
{
sizeof(struct file), /* size */
file_dump, /* dump */
default_fd_add_queue, /* add_queue */
default_fd_remove_queue, /* remove_queue */
default_fd_signaled, /* signaled */
no_satisfied, /* satisfied */
file_get_fd, /* get_fd */
file_destroy /* destroy */
};
static const struct fd_ops file_fd_ops =
{
file_get_poll_events, /* get_poll_events */
file_poll_event, /* poll_event */
file_flush, /* flush */
file_get_info, /* get_file_info */
file_queue_async /* queue_async */
};
static inline int is_overlapped( const struct file *file )
{
return !(file->options & (FILE_SYNCHRONOUS_IO_ALERT | FILE_SYNCHRONOUS_IO_NONALERT));
}
/* create a file from a file descriptor */
/* if the function fails the fd is closed */
static struct file *create_file_for_fd( int fd, unsigned int access, unsigned int sharing )
{
struct file *file;
if ((file = alloc_object( &file_ops )))
{
file->access = access;
file->options = FILE_SYNCHRONOUS_IO_NONALERT;
if (!(file->fd = create_anonymous_fd( &file_fd_ops, fd, &file->obj )))
{
release_object( file );
return NULL;
}
}
return file;
}
static struct object *create_file( const char *nameptr, size_t len, unsigned int access,
unsigned int sharing, int create, unsigned int options,
unsigned int attrs )
{
struct file *file;
int flags;
char *name;
mode_t mode;
if (!(name = mem_alloc( len + 1 ))) return NULL;
memcpy( name, nameptr, len );
name[len] = 0;
switch(create)
{
case FILE_CREATE: flags = O_CREAT | O_EXCL; break;
case FILE_OVERWRITE_IF: /* FIXME: the difference is whether we trash existing attr or not */
case FILE_SUPERSEDE: flags = O_CREAT | O_TRUNC; break;
case FILE_OPEN: flags = 0; break;
case FILE_OPEN_IF: flags = O_CREAT; break;
case FILE_OVERWRITE: flags = O_TRUNC; break;
default: set_error( STATUS_INVALID_PARAMETER ); goto error;
}
switch(access & (GENERIC_READ | GENERIC_WRITE))
{
case 0: break;
case GENERIC_READ: flags |= O_RDONLY; break;
case GENERIC_WRITE: flags |= O_WRONLY; break;
case GENERIC_READ|GENERIC_WRITE: flags |= O_RDWR; break;
}
mode = (attrs & FILE_ATTRIBUTE_READONLY) ? 0444 : 0666;
if (len >= 4 &&
(!strcasecmp( name + len - 4, ".exe" ) || !strcasecmp( name + len - 4, ".com" )))
mode |= 0111;
if (!(file = alloc_object( &file_ops ))) goto error;
file->access = access;
file->options = options;
if (is_overlapped( file ))
{
init_async_queue (&file->read_q);
init_async_queue (&file->write_q);
}
/* FIXME: should set error to STATUS_OBJECT_NAME_COLLISION if file existed before */
if (!(file->fd = alloc_fd( &file_fd_ops, &file->obj )) ||
!(file->fd = open_fd( file->fd, name, flags | O_NONBLOCK | O_LARGEFILE,
&mode, access, sharing,
(options & FILE_DELETE_ON_CLOSE) ? name : "" )))
{
free( name );
release_object( file );
return NULL;
}
free( name );
/* refuse to open a directory */
if (S_ISDIR(mode) && !(options & FILE_OPEN_FOR_BACKUP_INTENT))
{
set_error( STATUS_ACCESS_DENIED );
release_object( file );
return NULL;
}
/* check for serial port */
if (S_ISCHR(mode) && is_serial_fd( file->fd ))
{
struct object *obj = create_serial( file->fd, file->options );
release_object( file );
return obj;
}
return &file->obj;
error:
free( name );
return NULL;
}
2001-04-14 00:38:39 +02:00
/* check if two file objects point to the same file */
int is_same_file( struct file *file1, struct file *file2 )
{
return is_same_file_fd( file1->fd, file2->fd );
2001-04-14 00:38:39 +02:00
}
/* create a temp file for anonymous mappings */
struct file *create_temp_file( int access )
{
char tmpfn[16];
int fd;
sprintf( tmpfn, "anonmap.XXXXXX" ); /* create it in the server directory */
fd = mkstemps( tmpfn, 0 );
if (fd == -1)
{
file_set_error();
return NULL;
}
unlink( tmpfn );
return create_file_for_fd( fd, access, 0 );
}
static void file_dump( struct object *obj, int verbose )
{
struct file *file = (struct file *)obj;
assert( obj->ops == &file_ops );
fprintf( stderr, "File fd=%p options=%08x\n", file->fd, file->options );
}
static int file_get_poll_events( struct fd *fd )
{
struct file *file = get_fd_user( fd );
int events = 0;
assert( file->obj.ops == &file_ops );
if (file->access & GENERIC_READ) events |= POLLIN;
if (file->access & GENERIC_WRITE) events |= POLLOUT;
return events;
}
static void file_poll_event( struct fd *fd, int event )
{
struct file *file = get_fd_user( fd );
assert( file->obj.ops == &file_ops );
if (is_overlapped( file ))
{
if( IS_READY(file->read_q) && (POLLIN & event) )
{
async_notify(file->read_q.head, STATUS_ALERTED);
return;
}
if( IS_READY(file->write_q) && (POLLOUT & event) )
{
async_notify(file->write_q.head, STATUS_ALERTED);
return;
}
}
default_poll_event( fd, event );
}
static int file_flush( struct fd *fd, struct event **event )
{
int ret = (fsync( get_unix_fd(fd) ) != -1);
if (!ret) file_set_error();
return ret;
}
2004-04-08 21:09:04 +02:00
static int file_get_info( struct fd *fd, int *flags )
{
struct file *file = get_fd_user( fd );
*flags = 0;
if (is_overlapped( file )) *flags |= FD_FLAG_OVERLAPPED;
return FD_TYPE_DEFAULT;
}
static void file_queue_async(struct fd *fd, void *ptr, unsigned int status, int type, int count)
{
struct file *file = get_fd_user( fd );
struct async *async;
struct async_queue *q;
assert( file->obj.ops == &file_ops );
if (!is_overlapped( file ))
{
set_error ( STATUS_INVALID_HANDLE );
return;
}
switch(type)
{
case ASYNC_TYPE_READ:
q = &file->read_q;
break;
case ASYNC_TYPE_WRITE:
q = &file->write_q;
break;
default:
set_error( STATUS_INVALID_PARAMETER );
return;
}
async = find_async ( q, current, ptr );
if ( status == STATUS_PENDING )
{
int events;
if ( !async )
async = create_async ( &file->obj, current, ptr );
if ( !async )
return;
async->status = STATUS_PENDING;
if ( !async->q )
async_insert( q, async );
/* Check if the new pending request can be served immediately */
events = check_fd_events( fd, file_get_poll_events( fd ) );
if (events) file_poll_event ( fd, events );
}
else if ( async ) destroy_async ( async );
else set_error ( STATUS_INVALID_PARAMETER );
set_fd_events( fd, file_get_poll_events( fd ));
}
static struct fd *file_get_fd( struct object *obj )
{
struct file *file = (struct file *)obj;
assert( obj->ops == &file_ops );
return (struct fd *)grab_object( file->fd );
}
static void file_destroy( struct object *obj )
{
struct file *file = (struct file *)obj;
assert( obj->ops == &file_ops );
if (is_overlapped( file ))
{
destroy_async_queue (&file->read_q);
destroy_async_queue (&file->write_q);
}
if (file->fd) release_object( file->fd );
}
/* set the last error depending on errno */
void file_set_error(void)
{
switch (errno)
{
case EAGAIN: set_error( STATUS_SHARING_VIOLATION ); break;
case EBADF: set_error( STATUS_INVALID_HANDLE ); break;
case ENOSPC: set_error( STATUS_DISK_FULL ); break;
case EACCES:
case ESRCH:
case EPERM: set_error( STATUS_ACCESS_DENIED ); break;
case EROFS: set_error( STATUS_MEDIA_WRITE_PROTECTED ); break;
case EBUSY: set_error( STATUS_FILE_LOCK_CONFLICT ); break;
case ENOENT: set_error( STATUS_NO_SUCH_FILE ); break;
case EISDIR: set_win32_error( ERROR_CANNOT_MAKE ); break;
case ENFILE:
case EMFILE: set_error( STATUS_NO_MORE_FILES ); break;
case EEXIST: set_error( STATUS_OBJECT_NAME_COLLISION ); break;
case EINVAL: set_error( STATUS_INVALID_PARAMETER ); break;
case ESPIPE: set_win32_error( ERROR_SEEK ); break;
case ENOTEMPTY: set_error( STATUS_DIRECTORY_NOT_EMPTY ); break;
case EIO: set_error( STATUS_ACCESS_VIOLATION ); break;
case ENOTDIR: set_error( STATUS_NOT_A_DIRECTORY ); break;
2004-01-02 21:11:35 +01:00
#ifdef EOVERFLOW
case EOVERFLOW: set_error( STATUS_INVALID_PARAMETER ); break;
2004-01-02 21:11:35 +01:00
#endif
default: perror("file_set_error"); set_error( STATUS_UNSUCCESSFUL ); break;
}
}
struct file *get_file_obj( struct process *process, obj_handle_t handle, unsigned int access )
{
return (struct file *)get_handle_obj( process, handle, access, &file_ops );
}
int get_file_unix_fd( struct file *file )
{
return get_unix_fd( file->fd );
}
/* extend a file beyond the current end of file */
static int extend_file( struct file *file, off_t size )
{
static const char zero;
int unix_fd = get_file_unix_fd( file );
/* extend the file one byte beyond the requested size and then truncate it */
/* this should work around ftruncate implementations that can't extend files */
if ((lseek( unix_fd, size, SEEK_SET ) != -1) &&
(write( unix_fd, &zero, 1 ) != -1))
{
ftruncate( unix_fd, size );
return 1;
}
file_set_error();
return 0;
}
/* truncate file at current position */
static int truncate_file( struct file *file )
{
int ret = 0;
int unix_fd = get_file_unix_fd( file );
off_t pos = lseek( unix_fd, 0, SEEK_CUR );
off_t eof = lseek( unix_fd, 0, SEEK_END );
if (eof < pos) ret = extend_file( file, pos );
else
{
if (ftruncate( unix_fd, pos ) != -1) ret = 1;
else file_set_error();
}
lseek( unix_fd, pos, SEEK_SET ); /* restore file pos */
return ret;
}
/* try to grow the file to the specified size */
int grow_file( struct file *file, int size_high, int size_low )
{
int ret = 0;
struct stat st;
int unix_fd = get_file_unix_fd( file );
off_t old_pos, size = size_low + (((off_t)size_high)<<32);
if (fstat( unix_fd, &st ) == -1)
{
file_set_error();
return 0;
}
if (st.st_size >= size) return 1; /* already large enough */
old_pos = lseek( unix_fd, 0, SEEK_CUR ); /* save old pos */
ret = extend_file( file, size );
lseek( unix_fd, old_pos, SEEK_SET ); /* restore file pos */
return ret;
}
/* create a file */
DECL_HANDLER(create_file)
{
struct object *file;
reply->handle = 0;
if ((file = create_file( get_req_data(), get_req_data_size(), req->access,
req->sharing, req->create, req->options, req->attrs )))
{
reply->handle = alloc_handle( current->process, file, req->access, req->inherit );
release_object( file );
}
}
/* allocate a file handle for a Unix fd */
DECL_HANDLER(alloc_file_handle)
{
struct file *file;
int fd;
reply->handle = 0;
if ((fd = thread_get_inflight_fd( current, req->fd )) == -1)
{
set_error( STATUS_INVALID_HANDLE );
return;
}
if ((file = create_file_for_fd( fd, req->access, FILE_SHARE_READ | FILE_SHARE_WRITE )))
{
reply->handle = alloc_handle( current->process, file, req->access, req->inherit );
release_object( file );
}
}
/* truncate (or extend) a file */
DECL_HANDLER(truncate_file)
{
struct file *file;
if ((file = get_file_obj( current->process, req->handle, GENERIC_WRITE )))
{
truncate_file( file );
release_object( file );
}
}
/* lock a region of a file */
DECL_HANDLER(lock_file)
{
struct file *file;
file_pos_t offset = ((file_pos_t)req->offset_high << 32) | req->offset_low;
file_pos_t count = ((file_pos_t)req->count_high << 32) | req->count_low;
if ((file = get_file_obj( current->process, req->handle, 0 )))
{
reply->handle = lock_fd( file->fd, offset, count, req->shared, req->wait );
reply->overlapped = is_overlapped( file );
release_object( file );
}
}
/* unlock a region of a file */
DECL_HANDLER(unlock_file)
{
struct file *file;
file_pos_t offset = ((file_pos_t)req->offset_high << 32) | req->offset_low;
file_pos_t count = ((file_pos_t)req->count_high << 32) | req->count_low;
if ((file = get_file_obj( current->process, req->handle, 0 )))
{
unlock_fd( file->fd, offset, count );
release_object( file );
}
}