sin/samba-mirror
1
0
Fork 0
mirror of https://github.com/samba-team/samba.git synced 2024-12-27 03:21:53 +03:00
Code
3d3c50326b
samba-mirror/source3/smbd/open.c

1106 lines
31 KiB
C
Raw Normal View History

this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
/*
Unix SMB/Netbios implementation.
Version 1.9.
file opening and share modes
Copyright (C) Andrew Tridgell 1992-1998
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program 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 General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
*/
#include "includes.h"
extern int DEBUGLEVEL;
extern pstring sesssetup_user;
First cut at kernel oplocks. This should have no effect unless runnin on a machine that supports them in autoconf. Move various functions out of lib/util.c into smbd/process.c and smbd/oplock.c where they belong. Jeremy. (This used to be commit c3c5e13f85c97939746070132dad941e79c546fb)
1998-09-23 05:48:45 +04:00
extern uint16 global_oplock_port;
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
extern BOOL global_client_failed_oplock_break;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
/****************************************************************************
The latest open() code changes broke the NT directory opens. Detect if a read-only open on a directory was done and return an EISDIR from open_file(). Changed interface to fd_close to return error. Jeremy. (This used to be commit df4302f3911447fcebe9342f6cbf3b89bd3bafba)
2000-04-12 01:38:45 +04:00
fd support routines - attempt to do a dos_open.
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
****************************************************************************/
The latest open() code changes broke the NT directory opens. Detect if a read-only open on a directory was done and return an EISDIR from open_file(). Changed interface to fd_close to return error. Jeremy. (This used to be commit df4302f3911447fcebe9342f6cbf3b89bd3bafba)
2000-04-12 01:38:45 +04:00
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
static int fd_open(struct connection_struct *conn, char *fname,
int flags, mode_t mode)
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
{
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
int fd = conn->vfs_ops.open(dos_to_unix(fname,False),flags,mode);
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
/* Fix for files ending in '.' */
if((fd == -1) && (errno == ENOENT) &&
(strchr(fname,'.')==NULL)) {
pstrcat(fname,".");
fd = conn->vfs_ops.open(dos_to_unix(fname,False),flags,mode);
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
}
The latest open() code changes broke the NT directory opens. Detect if a read-only open on a directory was done and return an EISDIR from open_file(). Changed interface to fd_close to return error. Jeremy. (This used to be commit df4302f3911447fcebe9342f6cbf3b89bd3bafba)
2000-04-12 01:38:45 +04:00
DEBUG(10,("fd_open: name %s, mode = %d, fd = %d. %s\n", fname, (int)mode, fd,
(fd == -1) ? strerror(errno) : "" ));
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
return fd;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
}
/****************************************************************************
The latest open() code changes broke the NT directory opens. Detect if a read-only open on a directory was done and return an EISDIR from open_file(). Changed interface to fd_close to return error. Jeremy. (This used to be commit df4302f3911447fcebe9342f6cbf3b89bd3bafba)
2000-04-12 01:38:45 +04:00
Close the file associated with a fsp.
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
****************************************************************************/
The latest open() code changes broke the NT directory opens. Detect if a read-only open on a directory was done and return an EISDIR from open_file(). Changed interface to fd_close to return error. Jeremy. (This used to be commit df4302f3911447fcebe9342f6cbf3b89bd3bafba)
2000-04-12 01:38:45 +04:00
int fd_close(struct connection_struct *conn, files_struct *fsp)
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
{
Added the code that keeps fd's open across a close if there are other fsp's open on the same dev/inode pair with existing POSIX locks. This is done at the smbd/open layer, so smbd just calls fd_close() and the transfer of any open fd's is done under the covers of fd_close(). When an fsp is closed and no other fsp's open on the same dev/inode pair have existing POSIX locks then all fd's associated with this fsp are closed. Now only the hard part of doing the POSIX range unlock code when read locks overlap remains for full POSIX/SMB lock integration.... Jeremy. (This used to be commit 1df48ed55ee303b6d84d7277fd79761cfe5f7052)
2000-04-25 00:48:40 +04:00
int ret = 0;
/*
* Deal with transferring any pending fd's if there
* are POSIX locks outstanding.
*/
Did the rewrite Andrew wanted where all knowledge of POSIX locking is removed from the smbd/open.c code. We now use a dlink list of structures indexed by dev/inode to store all pending fd's for close. This could be rewritten to use lib/hash.c if this is discovered to be too slow in use. Andrew, please take a look and let me know if this is what you had in mind. Jeremy. (This used to be commit 0487841120a7584da9a2b83b9574562c415d7024)
2000-04-28 01:12:33 +04:00
if(!fd_close_posix_locks(conn,fsp))
Added the code that keeps fd's open across a close if there are other fsp's open on the same dev/inode pair with existing POSIX locks. This is done at the smbd/open layer, so smbd just calls fd_close() and the transfer of any open fd's is done under the covers of fd_close(). When an fsp is closed and no other fsp's open on the same dev/inode pair have existing POSIX locks then all fd's associated with this fsp are closed. Now only the hard part of doing the POSIX range unlock code when read locks overlap remains for full POSIX/SMB lock integration.... Jeremy. (This used to be commit 1df48ed55ee303b6d84d7277fd79761cfe5f7052)
2000-04-25 00:48:40 +04:00
return -1;
if(fsp->fd != -1)
ret = conn->vfs_ops.close(fsp->fd);
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
fsp->fd = -1;
The latest open() code changes broke the NT directory opens. Detect if a read-only open on a directory was done and return an EISDIR from open_file(). Changed interface to fd_close to return error. Jeremy. (This used to be commit df4302f3911447fcebe9342f6cbf3b89bd3bafba)
2000-04-12 01:38:45 +04:00
return ret;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
}
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
/****************************************************************************
The latest open() code changes broke the NT directory opens. Detect if a read-only open on a directory was done and return an EISDIR from open_file(). Changed interface to fd_close to return error. Jeremy. (This used to be commit df4302f3911447fcebe9342f6cbf3b89bd3bafba)
2000-04-12 01:38:45 +04:00
Check a filename for the pipe string.
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
****************************************************************************/
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
static void check_for_pipe(char *fname)
{
/* special case of pipe opens */
char s[10];
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
StrnCpy(s,fname,sizeof(s)-1);
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
strlower(s);
if (strstr(s,"pipe/")) {
DEBUG(3,("Rejecting named pipe open for %s\n",fname));
unix_ERR_class = ERRSRV;
unix_ERR_code = ERRaccess;
}
}
/****************************************************************************
The latest open() code changes broke the NT directory opens. Detect if a read-only open on a directory was done and return an EISDIR from open_file(). Changed interface to fd_close to return error. Jeremy. (This used to be commit df4302f3911447fcebe9342f6cbf3b89bd3bafba)
2000-04-12 01:38:45 +04:00
Open a file.
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
****************************************************************************/
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
static BOOL open_file(files_struct *fsp,connection_struct *conn,
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
char *fname1,int flags,mode_t mode)
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
{
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
extern struct current_user current_user;
pstring fname;
int accmode = (flags & O_ACCMODE);
SMB_STRUCT_STAT sbuf;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
Now that fsp's are created on successful file open, the structure member fsp->open is no longer needed (if an fsp pointer is valid, then it's open :-). NB for Luke, this patch also did not apply to TNG. TNG is not yet identical w.r.t file serving with HEAD. This makes it impossible for me to help maintain TNG. Please fix asap. lib/substitute.c: Removed unused variable (pidstr). Jeremy. (This used to be commit 389b700a26e8a308a0dff6fc038c38068aa0119a)
2000-04-24 23:23:51 +04:00
fsp->fd = -1;
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
fsp->oplock_type = NO_OPLOCK;
errno = EPERM;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
pstrcpy(fname,fname1);
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
The latest open() code changes broke the NT directory opens. Detect if a read-only open on a directory was done and return an EISDIR from open_file(). Changed interface to fd_close to return error. Jeremy. (This used to be commit df4302f3911447fcebe9342f6cbf3b89bd3bafba)
2000-04-12 01:38:45 +04:00
/* Check permissions */
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
/*
* This code was changed after seeing a client open request
* containing the open mode of (DENY_WRITE/read-only) with
* the 'create if not exist' bit set. The previous code
* would fail to open the file read only on a read-only share
* as it was checking the flags parameter directly against O_RDONLY,
* this was failing as the flags parameter was set to O_RDONLY|O_CREAT.
* JRA.
*/
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
if (!CAN_WRITE(conn)) {
/* It's a read-only share - fail if we wanted to write. */
if(accmode != O_RDONLY) {
DEBUG(3,("Permission denied opening %s\n",fname));
check_for_pipe(fname);
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
return False;
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
} else if(flags & O_CREAT) {
/* We don't want to write - but we must make sure that O_CREAT
doesn't create the file if we have write access into the
directory.
*/
flags &= ~O_CREAT;
}
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
/* actually do the open */
fsp->fd = fd_open(conn, fname, flags, mode);
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
The latest open() code changes broke the NT directory opens. Detect if a read-only open on a directory was done and return an EISDIR from open_file(). Changed interface to fd_close to return error. Jeremy. (This used to be commit df4302f3911447fcebe9342f6cbf3b89bd3bafba)
2000-04-12 01:38:45 +04:00
if (fsp->fd == -1) {
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
DEBUG(3,("Error opening file %s (%s) (flags=%d)\n",
fname,strerror(errno),flags));
check_for_pipe(fname);
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
return False;
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
conn->vfs_ops.fstat(fsp->fd, &sbuf);
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
The latest open() code changes broke the NT directory opens. Detect if a read-only open on a directory was done and return an EISDIR from open_file(). Changed interface to fd_close to return error. Jeremy. (This used to be commit df4302f3911447fcebe9342f6cbf3b89bd3bafba)
2000-04-12 01:38:45 +04:00
/*
* POSIX allows read-only opens of directories. We don't
* want to do this (we use a different code path for this)
* so catch a directory open and return an EISDIR. JRA.
*/
if(S_ISDIR(sbuf.st_mode)) {
fd_close(conn, fsp);
errno = EISDIR;
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
return False;
The latest open() code changes broke the NT directory opens. Detect if a read-only open on a directory was done and return an EISDIR from open_file(). Changed interface to fd_close to return error. Jeremy. (This used to be commit df4302f3911447fcebe9342f6cbf3b89bd3bafba)
2000-04-12 01:38:45 +04:00
}
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
fsp->mode = sbuf.st_mode;
fsp->inode = sbuf.st_ino;
fsp->dev = sbuf.st_dev;
GetTimeOfDay(&fsp->open_time);
fsp->vuid = current_user.vuid;
fsp->size = 0;
fsp->pos = -1;
fsp->can_lock = True;
fsp->can_read = ((flags & O_WRONLY)==0);
fsp->can_write = ((flags & (O_WRONLY|O_RDWR))!=0);
fsp->share_mode = 0;
fsp->print_file = False;
fsp->modified = False;
fsp->oplock_type = NO_OPLOCK;
fsp->sent_oplock_break = NO_BREAK_SENT;
fsp->is_directory = False;
fsp->stat_open = False;
fsp->directory_delete_on_close = False;
fsp->conn = conn;
/*
* Note that the file name here is the *untranslated* name
* ie. it is still in the DOS codepage sent from the client.
* All use of this filename will pass though the sys_xxxx
* functions which will do the dos_to_unix translation before
* mapping into a UNIX filename. JRA.
*/
string_set(&fsp->fsp_name,fname);
fsp->wbmpx_ptr = NULL;
fsp->wcp = NULL; /* Write cache pointer. */
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
DEBUG(2,("%s opened file %s read=%s write=%s (numopen=%d)\n",
*sesssetup_user ? sesssetup_user : conn->user,fsp->fsp_name,
BOOLSTR(fsp->can_read), BOOLSTR(fsp->can_write),
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
conn->num_files_open + 1));
return True;
Makefile.in: Fixed bug with continuation line causing proto to fail. Added $(PROGS) $(SPROGS) as targets for make clean. acconfig.h: Added HAVE_IRIX_SPECIFIC_CAPABILITIES. configure.in: Added sys/capability.h header check. Added function checks for srandom random srand rand. Added HAVE_IRIX_SPECIFIC_CAPABILITIES test. includes.h: Added #include <sys/capability.h>. ntdomain.h: Moved struct acct_info into here from smb.h smb.h: Added KERNEL_OPLOCK_CAPABILITY define. Moved enum action_type into rpcclient.h Moved struct cli_state into client.h Moved struct nt_client_info, struct tar_client_info, struct client_info into rpcclient.h lib/genrand.c: Changed to use sys_random() & friends. lib/smbrun.c: Lose capabilities after fork. lib/system.c: Added set_process_capability(), set_inherited_process_capability() sys_random(), sys_srandom(). lib/util.c: Added Ander's EFBIG lock check to fcntl_lock for 64 bit access to an 32 bit mounted NFS filesystem. nmbd/nmbd.c: Changed to use sys_random() & friends. nmbd/nmbd_browsesync.c: Changed to use sys_random() & friends. passdb/ldap.c: Missed one pdb_encode_acct_ctrl call. passdb/passdb.c: Changed to Ander's code for ' ' characters. passdb/smbpass.c: Added Ander's code to reset ACB_PWNOTREQ. script/mkproto.awk: Added 'long' to prototypes. smbd/chgpasswd.c: Lose capabilities after fork. smbd/open.c: Do the mmap *after* the kernel oplock. smbd/oplock.c: Removed stub code from kernel oplock path. Added set_process_capability(), set_inherited_process_capability() calls. smbd/reply.c: Initialize count = 0, offset = 0. smbd/server.c: Added set_process_capability(), set_inherited_process_capability() calls. tests/summary.c: Ensure we have RANDOM or RAND. utils/smbpasswd.c: Added Ander's code to reset ACB_PWNOTREQ. utils/torture.c: Changed to use sys_random() & friends. Jeremy. (This used to be commit e8be306f23963ac00b1a383ebe0cc1421529fb02)
1998-09-26 03:40:49 +04:00
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
/****************************************************************************
C. Hoch 11/22/95
Helper for open_file_shared.
Truncate a file after checking locking; close file if locked.
**************************************************************************/
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
static int truncate_unless_locked(struct connection_struct *conn, files_struct *fsp)
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
{
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
SMB_BIG_UINT mask = (SMB_BIG_UINT)-1;
if (!fsp->can_write)
return -1;
if (is_locked(fsp,fsp->conn,mask,0,WRITE_LOCK)){
errno = EACCES;
unix_ERR_class = ERRDOS;
unix_ERR_code = ERRlock;
return -1;
} else {
return conn->vfs_ops.ftruncate(fsp->fd,0);
removed another use of the LL suffix. Hopefully this is the last one (a grep doesn't show any more) (This used to be commit c1bd188744b0df950f2e00550c25f7d3e148094b)
1998-09-18 06:27:24 +04:00
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
}
made access_table() a pure logic function - makes it simpler to apply maths to (This used to be commit 0ad62f128d1c26b8f7e71bc045c6f4a584f8d374)
2000-01-16 15:53:12 +03:00
/*******************************************************************
return True if the filename is one of the special executable types
********************************************************************/
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
static BOOL is_executable(const char *fname)
made access_table() a pure logic function - makes it simpler to apply maths to (This used to be commit 0ad62f128d1c26b8f7e71bc045c6f4a584f8d374)
2000-01-16 15:53:12 +03:00
{
if ((fname = strrchr(fname,'.'))) {
if (strequal(fname,".com") ||
strequal(fname,".dll") ||
strequal(fname,".exe") ||
strequal(fname,".sym")) {
return True;
}
}
return False;
}
tridge the destroyer returns! prompted by the interpret_security() dead code that Jean-Francois pointed out I added a make target "finddead" that finds potentially dead (ie. unused) code. It spat out 304 function names ... I went through these are deleted many of them, making others static (finddead also reports functions that are used only in the local file). in doing this I have almost certainly deleted some useful code. I may have even prevented compilation with some compile options. I apologise. I decided it was better to get rid of this code now and add back the one or two functions that are needed than to keep all this baggage. So, if I have done a bit too much "destroying" then let me know. Keep the swearing to a minimum :) One bit I didn't do is the ubibt code. Chris, can you look at that? Heaps of unused functions there. Can they be made static? (This used to be commit 2204475c87f3024ea8fd1fbd7385b2def617a46f)
1998-09-05 09:07:05 +04:00
enum {AFAIL,AREAD,AWRITE,AALL};
/*******************************************************************
reproduce the share mode access table
made access_table() a pure logic function - makes it simpler to apply maths to (This used to be commit 0ad62f128d1c26b8f7e71bc045c6f4a584f8d374)
2000-01-16 15:53:12 +03:00
this is horrendoously complex, and really can't be justified on any
rational grounds except that this is _exactly_ what NT does. See
the DENY1 and DENY2 tests in smbtorture for a comprehensive set of
test routines.
tridge the destroyer returns! prompted by the interpret_security() dead code that Jean-Francois pointed out I added a make target "finddead" that finds potentially dead (ie. unused) code. It spat out 304 function names ... I went through these are deleted many of them, making others static (finddead also reports functions that are used only in the local file). in doing this I have almost certainly deleted some useful code. I may have even prevented compilation with some compile options. I apologise. I decided it was better to get rid of this code now and add back the one or two functions that are needed than to keep all this baggage. So, if I have done a bit too much "destroying" then let me know. Keep the swearing to a minimum :) One bit I didn't do is the ubibt code. Chris, can you look at that? Heaps of unused functions there. Can they be made static? (This used to be commit 2204475c87f3024ea8fd1fbd7385b2def617a46f)
1998-09-05 09:07:05 +04:00
********************************************************************/
static int access_table(int new_deny,int old_deny,int old_mode,
made access_table() a pure logic function - makes it simpler to apply maths to (This used to be commit 0ad62f128d1c26b8f7e71bc045c6f4a584f8d374)
2000-01-16 15:53:12 +03:00
BOOL same_pid, BOOL isexe)
tridge the destroyer returns! prompted by the interpret_security() dead code that Jean-Francois pointed out I added a make target "finddead" that finds potentially dead (ie. unused) code. It spat out 304 function names ... I went through these are deleted many of them, making others static (finddead also reports functions that are used only in the local file). in doing this I have almost certainly deleted some useful code. I may have even prevented compilation with some compile options. I apologise. I decided it was better to get rid of this code now and add back the one or two functions that are needed than to keep all this baggage. So, if I have done a bit too much "destroying" then let me know. Keep the swearing to a minimum :) One bit I didn't do is the ubibt code. Chris, can you look at that? Heaps of unused functions there. Can they be made static? (This used to be commit 2204475c87f3024ea8fd1fbd7385b2def617a46f)
1998-09-05 09:07:05 +04:00
{
as obelix would say "these romans are crazy" I've finally got the access table code right for the case where the two opens are on the same connection. It is _incredibly_ complex, but now all 1296 test cases pass. I'll be very surprised if anyone by MS and us gets this right at CIFS2000 (This used to be commit 31a5857ce4f03eb53ff8ea96c5f1ce335941336d)
2000-01-16 15:26:42 +03:00
if (new_deny == DENY_ALL || old_deny == DENY_ALL) return(AFAIL);
made access_table() a pure logic function - makes it simpler to apply maths to (This used to be commit 0ad62f128d1c26b8f7e71bc045c6f4a584f8d374)
2000-01-16 15:53:12 +03:00
if (same_pid) {
fixed some typos in access_table() which, amaziingly enough, make no difference to the result. (This used to be commit 9e28b7220bc293e386fec30c2d2e7e401b61b1f7)
2000-01-16 16:17:05 +03:00
if (isexe && old_mode == DOS_OPEN_RDONLY &&
as obelix would say "these romans are crazy" I've finally got the access table code right for the case where the two opens are on the same connection. It is _incredibly_ complex, but now all 1296 test cases pass. I'll be very surprised if anyone by MS and us gets this right at CIFS2000 (This used to be commit 31a5857ce4f03eb53ff8ea96c5f1ce335941336d)
2000-01-16 15:26:42 +03:00
old_deny == DENY_DOS && new_deny == DENY_READ) {
return AFAIL;
}
fixed some typos in access_table() which, amaziingly enough, make no difference to the result. (This used to be commit 9e28b7220bc293e386fec30c2d2e7e401b61b1f7)
2000-01-16 16:17:05 +03:00
if (!isexe && old_mode == DOS_OPEN_RDONLY &&
as obelix would say "these romans are crazy" I've finally got the access table code right for the case where the two opens are on the same connection. It is _incredibly_ complex, but now all 1296 test cases pass. I'll be very surprised if anyone by MS and us gets this right at CIFS2000 (This used to be commit 31a5857ce4f03eb53ff8ea96c5f1ce335941336d)
2000-01-16 15:26:42 +03:00
old_deny == DENY_DOS && new_deny == DENY_DOS) {
return AREAD;
}
if (new_deny == DENY_FCB && old_deny == DENY_DOS) {
if (isexe) return AFAIL;
if (old_mode == DOS_OPEN_RDONLY) return AFAIL;
return AALL;
}
fixed some typos in access_table() which, amaziingly enough, make no difference to the result. (This used to be commit 9e28b7220bc293e386fec30c2d2e7e401b61b1f7)
2000-01-16 16:17:05 +03:00
if (old_mode == DOS_OPEN_RDONLY && old_deny == DENY_DOS) {
as obelix would say "these romans are crazy" I've finally got the access table code right for the case where the two opens are on the same connection. It is _incredibly_ complex, but now all 1296 test cases pass. I'll be very surprised if anyone by MS and us gets this right at CIFS2000 (This used to be commit 31a5857ce4f03eb53ff8ea96c5f1ce335941336d)
2000-01-16 15:26:42 +03:00
if (new_deny == DENY_FCB || new_deny == DENY_READ) {
if (isexe) return AREAD;
return AFAIL;
}
}
if (old_deny == DENY_FCB) {
if (new_deny == DENY_DOS || new_deny == DENY_FCB) return AALL;
return AFAIL;
}
}
if (old_deny == DENY_DOS || new_deny == DENY_DOS ||
old_deny == DENY_FCB || new_deny == DENY_FCB) {
if (isexe) {
rewrote the access_table() code to get it right for *.exe, *.dll, *.sym and *.com files. I still find it incredible that SMB treats file locking differently depending on the file name - that is so ugly it stinks. anyway, we now behave the same way as NT, except for the same-connection case which I'm still working on. (This used to be commit 5b9be3feca26f1f20fe3f32ba90769cc78f9bb6b)
2000-01-16 14:21:33 +03:00
if (old_deny == DENY_FCB || new_deny == DENY_FCB) {
return AFAIL;
}
if (old_deny == DENY_DOS) {
if (new_deny == DENY_READ &&
(old_mode == DOS_OPEN_RDONLY ||
old_mode == DOS_OPEN_RDWR)) {
return AFAIL;
}
if (new_deny == DENY_WRITE &&
(old_mode == DOS_OPEN_WRONLY ||
old_mode == DOS_OPEN_RDWR)) {
return AFAIL;
}
return AALL;
}
if (old_deny == DENY_NONE) return AALL;
if (old_deny == DENY_READ) return AWRITE;
if (old_deny == DENY_WRITE) return AREAD;
}
as obelix would say "these romans are crazy" I've finally got the access table code right for the case where the two opens are on the same connection. It is _incredibly_ complex, but now all 1296 test cases pass. I'll be very surprised if anyone by MS and us gets this right at CIFS2000 (This used to be commit 31a5857ce4f03eb53ff8ea96c5f1ce335941336d)
2000-01-16 15:26:42 +03:00
/* it isn't a exe, dll, sym or com file */
made access_table() a pure logic function - makes it simpler to apply maths to (This used to be commit 0ad62f128d1c26b8f7e71bc045c6f4a584f8d374)
2000-01-16 15:53:12 +03:00
if (old_deny == new_deny && same_pid)
rewrote the access_table() code to get it right for *.exe, *.dll, *.sym and *.com files. I still find it incredible that SMB treats file locking differently depending on the file name - that is so ugly it stinks. anyway, we now behave the same way as NT, except for the same-connection case which I'm still working on. (This used to be commit 5b9be3feca26f1f20fe3f32ba90769cc78f9bb6b)
2000-01-16 14:21:33 +03:00
return(AALL);
if (old_deny == DENY_READ || new_deny == DENY_READ) return AFAIL;
if (old_mode == DOS_OPEN_RDONLY) return(AREAD);
return(AFAIL);
}
as obelix would say "these romans are crazy" I've finally got the access table code right for the case where the two opens are on the same connection. It is _incredibly_ complex, but now all 1296 test cases pass. I'll be very surprised if anyone by MS and us gets this right at CIFS2000 (This used to be commit 31a5857ce4f03eb53ff8ea96c5f1ce335941336d)
2000-01-16 15:26:42 +03:00
switch (new_deny)
{
case DENY_WRITE:
if (old_deny==DENY_WRITE && old_mode==DOS_OPEN_RDONLY) return(AREAD);
if (old_deny==DENY_READ && old_mode==DOS_OPEN_RDONLY) return(AWRITE);
if (old_deny==DENY_NONE && old_mode==DOS_OPEN_RDONLY) return(AALL);
return(AFAIL);
case DENY_READ:
if (old_deny==DENY_WRITE && old_mode==DOS_OPEN_WRONLY) return(AREAD);
if (old_deny==DENY_READ && old_mode==DOS_OPEN_WRONLY) return(AWRITE);
if (old_deny==DENY_NONE && old_mode==DOS_OPEN_WRONLY) return(AALL);
return(AFAIL);
case DENY_NONE:
if (old_deny==DENY_WRITE) return(AREAD);
if (old_deny==DENY_READ) return(AWRITE);
if (old_deny==DENY_NONE) return(AALL);
return(AFAIL);
}
return(AFAIL);
tridge the destroyer returns! prompted by the interpret_security() dead code that Jean-Francois pointed out I added a make target "finddead" that finds potentially dead (ie. unused) code. It spat out 304 function names ... I went through these are deleted many of them, making others static (finddead also reports functions that are used only in the local file). in doing this I have almost certainly deleted some useful code. I may have even prevented compilation with some compile options. I apologise. I decided it was better to get rid of this code now and add back the one or two functions that are needed than to keep all this baggage. So, if I have done a bit too much "destroying" then let me know. Keep the swearing to a minimum :) One bit I didn't do is the ubibt code. Chris, can you look at that? Heaps of unused functions there. Can they be made static? (This used to be commit 2204475c87f3024ea8fd1fbd7385b2def617a46f)
1998-09-05 09:07:05 +04:00
}
made access_table() a pure logic function - makes it simpler to apply maths to (This used to be commit 0ad62f128d1c26b8f7e71bc045c6f4a584f8d374)
2000-01-16 15:53:12 +03:00
tridge the destroyer returns! prompted by the interpret_security() dead code that Jean-Francois pointed out I added a make target "finddead" that finds potentially dead (ie. unused) code. It spat out 304 function names ... I went through these are deleted many of them, making others static (finddead also reports functions that are used only in the local file). in doing this I have almost certainly deleted some useful code. I may have even prevented compilation with some compile options. I apologise. I decided it was better to get rid of this code now and add back the one or two functions that are needed than to keep all this baggage. So, if I have done a bit too much "destroying" then let me know. Keep the swearing to a minimum :) One bit I didn't do is the ubibt code. Chris, can you look at that? Heaps of unused functions there. Can they be made static? (This used to be commit 2204475c87f3024ea8fd1fbd7385b2def617a46f)
1998-09-05 09:07:05 +04:00
/****************************************************************************
check if we can open a file with a share mode
****************************************************************************/
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
tridge the destroyer returns! prompted by the interpret_security() dead code that Jean-Francois pointed out I added a make target "finddead" that finds potentially dead (ie. unused) code. It spat out 304 function names ... I went through these are deleted many of them, making others static (finddead also reports functions that are used only in the local file). in doing this I have almost certainly deleted some useful code. I may have even prevented compilation with some compile options. I apologise. I decided it was better to get rid of this code now and add back the one or two functions that are needed than to keep all this baggage. So, if I have done a bit too much "destroying" then let me know. Keep the swearing to a minimum :) One bit I didn't do is the ubibt code. Chris, can you look at that? Heaps of unused functions there. Can they be made static? (This used to be commit 2204475c87f3024ea8fd1fbd7385b2def617a46f)
1998-09-05 09:07:05 +04:00
static int check_share_mode( share_mode_entry *share, int deny_mode,
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
const char *fname, BOOL fcbopen, int *flags)
{
int old_open_mode = GET_OPEN_MODE(share->share_mode);
int old_deny_mode = GET_DENY_MODE(share->share_mode);
/*
* Don't allow any open once the delete on close flag has been
* set.
*/
if(GET_DELETE_ON_CLOSE_FLAG(share->share_mode)) {
DEBUG(5,("check_share_mode: Failing open on file %s as delete on close flag is set.\n",
fname ));
unix_ERR_class = ERRDOS;
unix_ERR_code = ERRnoaccess;
return False;
}
{
int access_allowed = access_table(deny_mode,old_deny_mode,old_open_mode,
(share->pid == getpid()),is_executable(fname));
if ((access_allowed == AFAIL) ||
(!fcbopen && (access_allowed == AREAD && *flags == O_RDWR)) ||
(access_allowed == AREAD && *flags != O_RDONLY) ||
(access_allowed == AWRITE && *flags != O_WRONLY)) {
DEBUG(2,("Share violation on file (%d,%d,%d,%d,%s,fcbopen = %d, flags = %d) = %d\n",
deny_mode,old_deny_mode,old_open_mode,
(int)share->pid,fname, fcbopen, *flags, access_allowed));
unix_ERR_class = ERRDOS;
unix_ERR_code = ERRbadshare;
return False;
}
if (access_allowed == AREAD)
*flags = O_RDONLY;
if (access_allowed == AWRITE)
*flags = O_WRONLY;
}
return True;
}
/****************************************************************************
Deal with open deny mode and oplock break processing.
Invarient: Share mode must be locked on entry and exit.
Returns -1 on error, or number of share modes on success (may be zero).
****************************************************************************/
static int open_mode_check(connection_struct *conn, const char *fname, SMB_DEV_T dev,
SMB_INO_T inode, int share_mode, int *p_flags, int *p_oplock_request,
BOOL *p_all_current_opens_are_level_II)
tridge the destroyer returns! prompted by the interpret_security() dead code that Jean-Francois pointed out I added a make target "finddead" that finds potentially dead (ie. unused) code. It spat out 304 function names ... I went through these are deleted many of them, making others static (finddead also reports functions that are used only in the local file). in doing this I have almost certainly deleted some useful code. I may have even prevented compilation with some compile options. I apologise. I decided it was better to get rid of this code now and add back the one or two functions that are needed than to keep all this baggage. So, if I have done a bit too much "destroying" then let me know. Keep the swearing to a minimum :) One bit I didn't do is the ubibt code. Chris, can you look at that? Heaps of unused functions there. Can they be made static? (This used to be commit 2204475c87f3024ea8fd1fbd7385b2def617a46f)
1998-09-05 09:07:05 +04:00
{
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
int i;
int num_share_modes;
int oplock_contention_count = 0;
share_mode_entry *old_shares = 0;
BOOL fcbopen = False;
int deny_mode = GET_DENY_MODE(share_mode);
BOOL broke_oplock;
if(GET_OPEN_MODE(share_mode) == DOS_OPEN_FCB)
fcbopen = True;
num_share_modes = get_share_modes(conn, dev, inode, &old_shares);
if(num_share_modes == 0)
return 0;
tridge the destroyer returns! prompted by the interpret_security() dead code that Jean-Francois pointed out I added a make target "finddead" that finds potentially dead (ie. unused) code. It spat out 304 function names ... I went through these are deleted many of them, making others static (finddead also reports functions that are used only in the local file). in doing this I have almost certainly deleted some useful code. I may have even prevented compilation with some compile options. I apologise. I decided it was better to get rid of this code now and add back the one or two functions that are needed than to keep all this baggage. So, if I have done a bit too much "destroying" then let me know. Keep the swearing to a minimum :) One bit I didn't do is the ubibt code. Chris, can you look at that? Heaps of unused functions there. Can they be made static? (This used to be commit 2204475c87f3024ea8fd1fbd7385b2def617a46f)
1998-09-05 09:07:05 +04:00
Reasonably large change to give us *exactly* correct NT delete on close semantics. This was trickier than it looks :-). Check out the new DELETE_ON_CLOSE flag in the share modes and the new code that iterates through all open files on the same device and inode in files.c and trans2.c Also changed the code that modifies share mode entries to take generic function pointers rather than doing a specific thing so this sort of change should be easier in the future. Jeremy. (This used to be commit 5e6a7cd99d29d1cf068fc517272559c1cf47ea3a)
1998-10-23 07:34:50 +04:00
/*
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
* Check if the share modes will give us access.
Reasonably large change to give us *exactly* correct NT delete on close semantics. This was trickier than it looks :-). Check out the new DELETE_ON_CLOSE flag in the share modes and the new code that iterates through all open files on the same device and inode in files.c and trans2.c Also changed the code that modifies share mode entries to take generic function pointers rather than doing a specific thing so this sort of change should be easier in the future. Jeremy. (This used to be commit 5e6a7cd99d29d1cf068fc517272559c1cf47ea3a)
1998-10-23 07:34:50 +04:00
*/
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
do {
Reasonably large change to give us *exactly* correct NT delete on close semantics. This was trickier than it looks :-). Check out the new DELETE_ON_CLOSE flag in the share modes and the new code that iterates through all open files on the same device and inode in files.c and trans2.c Also changed the code that modifies share mode entries to take generic function pointers rather than doing a specific thing so this sort of change should be easier in the future. Jeremy. (This used to be commit 5e6a7cd99d29d1cf068fc517272559c1cf47ea3a)
1998-10-23 07:34:50 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
broke_oplock = False;
*p_all_current_opens_are_level_II = True;
for(i = 0; i < num_share_modes; i++) {
share_mode_entry *share_entry = &old_shares[i];
/*
* By observation of NetBench, oplocks are broken *before* share
* modes are checked. This allows a file to be closed by the client
* if the share mode would deny access and the client has an oplock.
* Check if someone has an oplock on this file. If so we must break
* it before continuing.
*/
if((*p_oplock_request && EXCLUSIVE_OPLOCK_TYPE(share_entry->op_type)) ||
(!*p_oplock_request && (share_entry->op_type != NO_OPLOCK))) {
BOOL opb_ret;
DEBUG(5,("open_mode_check: breaking oplock (%x) on file %s, \
dev = %x, inode = %.0f\n", share_entry->op_type, fname, (unsigned int)dev, (double)inode));
/* Oplock break - unlock to request it. */
unlock_share_entry(conn, dev, inode);
opb_ret = request_oplock_break(share_entry, dev, inode);
/* Now relock. */
lock_share_entry(conn, dev, inode);
if(opb_ret == False) {
free((char *)old_shares);
DEBUG(0,("open_mode_check: FAILED when breaking oplock (%x) on file %s, \
dev = %x, inode = %.0f\n", old_shares[i].op_type, fname, (unsigned int)dev, (double)inode));
errno = EACCES;
unix_ERR_class = ERRDOS;
unix_ERR_code = ERRbadshare;
return -1;
}
broke_oplock = True;
*p_all_current_opens_are_level_II = False;
break;
} else if (!LEVEL_II_OPLOCK_TYPE(share_entry->op_type)) {
*p_all_current_opens_are_level_II = False;
}
/* someone else has a share lock on it, check to see
if we can too */
if(check_share_mode(share_entry, deny_mode, fname, fcbopen, p_flags) == False) {
free((char *)old_shares);
errno = EACCES;
return -1;
}
} /* end for */
if(broke_oplock) {
free((char *)old_shares);
num_share_modes = get_share_modes(conn, dev, inode, &old_shares);
oplock_contention_count++;
tridge the destroyer returns! prompted by the interpret_security() dead code that Jean-Francois pointed out I added a make target "finddead" that finds potentially dead (ie. unused) code. It spat out 304 function names ... I went through these are deleted many of them, making others static (finddead also reports functions that are used only in the local file). in doing this I have almost certainly deleted some useful code. I may have even prevented compilation with some compile options. I apologise. I decided it was better to get rid of this code now and add back the one or two functions that are needed than to keep all this baggage. So, if I have done a bit too much "destroying" then let me know. Keep the swearing to a minimum :) One bit I didn't do is the ubibt code. Chris, can you look at that? Heaps of unused functions there. Can they be made static? (This used to be commit 2204475c87f3024ea8fd1fbd7385b2def617a46f)
1998-09-05 09:07:05 +04:00
}
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
} while(broke_oplock);
tridge the destroyer returns! prompted by the interpret_security() dead code that Jean-Francois pointed out I added a make target "finddead" that finds potentially dead (ie. unused) code. It spat out 304 function names ... I went through these are deleted many of them, making others static (finddead also reports functions that are used only in the local file). in doing this I have almost certainly deleted some useful code. I may have even prevented compilation with some compile options. I apologise. I decided it was better to get rid of this code now and add back the one or two functions that are needed than to keep all this baggage. So, if I have done a bit too much "destroying" then let me know. Keep the swearing to a minimum :) One bit I didn't do is the ubibt code. Chris, can you look at that? Heaps of unused functions there. Can they be made static? (This used to be commit 2204475c87f3024ea8fd1fbd7385b2def617a46f)
1998-09-05 09:07:05 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
if(old_shares != 0)
free((char *)old_shares);
tridge the destroyer returns! prompted by the interpret_security() dead code that Jean-Francois pointed out I added a make target "finddead" that finds potentially dead (ie. unused) code. It spat out 304 function names ... I went through these are deleted many of them, making others static (finddead also reports functions that are used only in the local file). in doing this I have almost certainly deleted some useful code. I may have even prevented compilation with some compile options. I apologise. I decided it was better to get rid of this code now and add back the one or two functions that are needed than to keep all this baggage. So, if I have done a bit too much "destroying" then let me know. Keep the swearing to a minimum :) One bit I didn't do is the ubibt code. Chris, can you look at that? Heaps of unused functions there. Can they be made static? (This used to be commit 2204475c87f3024ea8fd1fbd7385b2def617a46f)
1998-09-05 09:07:05 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
/*
* Refuse to grant an oplock in case the contention limit is
* reached when going through the lock list multiple times.
*/
tridge the destroyer returns! prompted by the interpret_security() dead code that Jean-Francois pointed out I added a make target "finddead" that finds potentially dead (ie. unused) code. It spat out 304 function names ... I went through these are deleted many of them, making others static (finddead also reports functions that are used only in the local file). in doing this I have almost certainly deleted some useful code. I may have even prevented compilation with some compile options. I apologise. I decided it was better to get rid of this code now and add back the one or two functions that are needed than to keep all this baggage. So, if I have done a bit too much "destroying" then let me know. Keep the swearing to a minimum :) One bit I didn't do is the ubibt code. Chris, can you look at that? Heaps of unused functions there. Can they be made static? (This used to be commit 2204475c87f3024ea8fd1fbd7385b2def617a46f)
1998-09-05 09:07:05 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
if(oplock_contention_count >= lp_oplock_contention_limit(SNUM(conn))) {
*p_oplock_request = 0;
DEBUG(4,("open_mode_check: oplock contention = %d. Not granting oplock.\n",
oplock_contention_count ));
tridge the destroyer returns! prompted by the interpret_security() dead code that Jean-Francois pointed out I added a make target "finddead" that finds potentially dead (ie. unused) code. It spat out 304 function names ... I went through these are deleted many of them, making others static (finddead also reports functions that are used only in the local file). in doing this I have almost certainly deleted some useful code. I may have even prevented compilation with some compile options. I apologise. I decided it was better to get rid of this code now and add back the one or two functions that are needed than to keep all this baggage. So, if I have done a bit too much "destroying" then let me know. Keep the swearing to a minimum :) One bit I didn't do is the ubibt code. Chris, can you look at that? Heaps of unused functions there. Can they be made static? (This used to be commit 2204475c87f3024ea8fd1fbd7385b2def617a46f)
1998-09-05 09:07:05 +04:00
}
Reasonably large change to give us *exactly* correct NT delete on close semantics. This was trickier than it looks :-). Check out the new DELETE_ON_CLOSE flag in the share modes and the new code that iterates through all open files on the same device and inode in files.c and trans2.c Also changed the code that modifies share mode entries to take generic function pointers rather than doing a specific thing so this sort of change should be easier in the future. Jeremy. (This used to be commit 5e6a7cd99d29d1cf068fc517272559c1cf47ea3a)
1998-10-23 07:34:50 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
return num_share_modes;
tridge the destroyer returns! prompted by the interpret_security() dead code that Jean-Francois pointed out I added a make target "finddead" that finds potentially dead (ie. unused) code. It spat out 304 function names ... I went through these are deleted many of them, making others static (finddead also reports functions that are used only in the local file). in doing this I have almost certainly deleted some useful code. I may have even prevented compilation with some compile options. I apologise. I decided it was better to get rid of this code now and add back the one or two functions that are needed than to keep all this baggage. So, if I have done a bit too much "destroying" then let me know. Keep the swearing to a minimum :) One bit I didn't do is the ubibt code. Chris, can you look at that? Heaps of unused functions there. Can they be made static? (This used to be commit 2204475c87f3024ea8fd1fbd7385b2def617a46f)
1998-09-05 09:07:05 +04:00
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
/****************************************************************************
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
Open a file with a share mode.
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
****************************************************************************/
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
files_struct *open_file_shared(connection_struct *conn,char *fname,int share_mode,int ofun,
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
mode_t mode,int oplock_request, int *Access,int *action)
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
{
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
int flags=0;
int flags2=0;
int deny_mode = GET_DENY_MODE(share_mode);
BOOL allow_share_delete = GET_ALLOW_SHARE_DELETE(share_mode);
SMB_STRUCT_STAT sbuf;
BOOL file_existed = vfs_file_exist(conn, fname, &sbuf);
BOOL fcbopen = False;
SMB_DEV_T dev = 0;
SMB_INO_T inode = 0;
int num_share_modes = 0;
BOOL all_current_opens_are_level_II = False;
Now that fsp's are created on successful file open, the structure member fsp->open is no longer needed (if an fsp pointer is valid, then it's open :-). NB for Luke, this patch also did not apply to TNG. TNG is not yet identical w.r.t file serving with HEAD. This makes it impossible for me to help maintain TNG. Please fix asap. lib/substitute.c: Removed unused variable (pidstr). Jeremy. (This used to be commit 389b700a26e8a308a0dff6fc038c38068aa0119a)
2000-04-24 23:23:51 +04:00
BOOL fsp_open = False;
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
files_struct *fsp = NULL;
int open_mode=0;
uint16 port = 0;
if (conn->printer) {
/* printers are handled completely differently. Most of the passed parameters are
ignored */
*Access = DOS_OPEN_WRONLY;
*action = FILE_WAS_CREATED;
return print_fsp_open(conn, fname);
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
fsp = file_new();
if(!fsp)
return NULL;
libsmb/smbdes.c: #ifdef'ed out code prior to removal. rpc_client/cli_pipe.c: Inlined code removed from smbdes.c rpc_server/srv_samr.c: Fixed unused variable warning. rpc_server/srv_util.c: Inlined code removed from smbdes.c Luke - the above changes are the first part of the changes you and I discussed as being neccessary at the CIFS conference. *PLEASE REVIEW THESE CHANGES* - make sure I haven't broken any of the authenticated DCE/RPC code. smbd/nttrans.c: Fixed to allow NT5.0beta2 to use Samba shares with NT SMB support. smbd/open.c: Fixed mkdir when called from nttrans calls. smbd/server.c: Set correct size for strcpy of global_myworkgroup. Jeremy. (This used to be commit d891421d16ff80998dee429227bd391455f9d1a1)
1998-10-22 20:55:03 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
fsp->fd = -1;
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
DEBUG(10,("open_file_shared: fname = %s, share_mode = %x, ofun = %x, mode = %o, oplock request = %d\n",
fname, share_mode, ofun, (int)mode, oplock_request ));
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
if (!check_name(fname,conn)) {
file_free(fsp);
return NULL;
}
/* ignore any oplock requests if oplocks are disabled */
if (!lp_oplocks(SNUM(conn)) || global_client_failed_oplock_break) {
oplock_request = 0;
}
/* this is for OS/2 EAs - try and say we don't support them */
if (strstr(fname,".+,;=[].")) {
unix_ERR_class = ERRDOS;
/* OS/2 Workplace shell fix may be main code stream in a later release. */
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
#if 1 /* OS2_WPS_FIX - Recent versions of OS/2 need this. */
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
unix_ERR_code = ERRcannotopen;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
#else /* OS2_WPS_FIX */
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
unix_ERR_code = ERROR_EAS_NOT_SUPPORTED;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
#endif /* OS2_WPS_FIX */
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
DEBUG(5,("open_file_shared: OS/2 EA's are not supported.\n"));
file_free(fsp);
return NULL;
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
if ((GET_FILE_OPEN_DISPOSITION(ofun) == FILE_EXISTS_FAIL) && file_existed) {
DEBUG(5,("open_file_shared: create new requested for file %s and file already exists.\n",
fname ));
file_free(fsp);
errno = EEXIST;
return NULL;
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
if (GET_FILE_CREATE_DISPOSITION(ofun) == FILE_CREATE_IF_NOT_EXIST)
flags2 |= O_CREAT;
if (GET_FILE_OPEN_DISPOSITION(ofun) == FILE_EXISTS_TRUNCATE)
flags2 |= O_TRUNC;
if (GET_FILE_OPEN_DISPOSITION(ofun) == FILE_EXISTS_FAIL)
flags2 |= O_EXCL;
/* note that we ignore the append flag as
append does not mean the same thing under dos and unix */
switch (GET_OPEN_MODE(share_mode)) {
case DOS_OPEN_WRONLY:
flags = O_WRONLY;
break;
case DOS_OPEN_FCB:
fcbopen = True;
flags = O_RDWR;
break;
case DOS_OPEN_RDWR:
flags = O_RDWR;
break;
default:
flags = O_RDONLY;
break;
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
#if defined(O_SYNC)
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
if (GET_FILE_SYNC_OPENMODE(share_mode)) {
flags2 |= O_SYNC;
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
#endif /* O_SYNC */
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
if (flags != O_RDONLY && file_existed &&
(!CAN_WRITE(conn) || IS_DOS_READONLY(dos_mode(conn,fname,&sbuf)))) {
if (!fcbopen) {
DEBUG(5,("open_file_shared: read/write access requested for file %s on read only %s\n",
fname, !CAN_WRITE(conn) ? "share" : "file" ));
file_free(fsp);
errno = EACCES;
return NULL;
}
flags = O_RDONLY;
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
if (deny_mode > DENY_NONE && deny_mode!=DENY_FCB) {
DEBUG(2,("Invalid deny mode %d on file %s\n",deny_mode,fname));
file_free(fsp);
errno = EINVAL;
return NULL;
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
if (file_existed) {
dev = sbuf.st_dev;
inode = sbuf.st_ino;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
lock_share_entry(conn, dev, inode);
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
num_share_modes = open_mode_check(conn, fname, dev, inode, share_mode,
&flags, &oplock_request, &all_current_opens_are_level_II);
if(num_share_modes == -1) {
unlock_share_entry(conn, dev, inode);
file_free(fsp);
return NULL;
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
/*
* We exit this block with the share entry *locked*.....
*/
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
DEBUG(4,("calling open_file with flags=0x%X flags2=0x%X mode=0%o\n",
flags,flags2,(int)mode));
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
Now that fsp's are created on successful file open, the structure member fsp->open is no longer needed (if an fsp pointer is valid, then it's open :-). NB for Luke, this patch also did not apply to TNG. TNG is not yet identical w.r.t file serving with HEAD. This makes it impossible for me to help maintain TNG. Please fix asap. lib/substitute.c: Removed unused variable (pidstr). Jeremy. (This used to be commit 389b700a26e8a308a0dff6fc038c38068aa0119a)
2000-04-24 23:23:51 +04:00
fsp_open = open_file(fsp,conn,fname,flags|(flags2&~(O_TRUNC)),mode);
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
Now that fsp's are created on successful file open, the structure member fsp->open is no longer needed (if an fsp pointer is valid, then it's open :-). NB for Luke, this patch also did not apply to TNG. TNG is not yet identical w.r.t file serving with HEAD. This makes it impossible for me to help maintain TNG. Please fix asap. lib/substitute.c: Removed unused variable (pidstr). Jeremy. (This used to be commit 389b700a26e8a308a0dff6fc038c38068aa0119a)
2000-04-24 23:23:51 +04:00
if (!fsp_open && (flags == O_RDWR) && (errno != ENOENT) && fcbopen) {
if((fsp_open = open_file(fsp,conn,fname,O_RDONLY,mode)) == True)
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
flags = O_RDONLY;
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
Now that fsp's are created on successful file open, the structure member fsp->open is no longer needed (if an fsp pointer is valid, then it's open :-). NB for Luke, this patch also did not apply to TNG. TNG is not yet identical w.r.t file serving with HEAD. This makes it impossible for me to help maintain TNG. Please fix asap. lib/substitute.c: Removed unused variable (pidstr). Jeremy. (This used to be commit 389b700a26e8a308a0dff6fc038c38068aa0119a)
2000-04-24 23:23:51 +04:00
if (!fsp_open) {
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
if(file_existed)
unlock_share_entry(conn, dev, inode);
file_free(fsp);
return NULL;
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
/*
* Deal with the race condition where two smbd's detect the file doesn't
* exist and do the create at the same time. One of them will win and
* set a share mode, the other (ie. this one) should check if the
* requested share mode for this create is allowed.
*/
Modified dev_t and ino_t code to be 64 bit clean (including changes to oplock break message passing). I think that smbd/nmbd are now inode and offset size independent (at least for 32 bit and 64 bit systems). Now to expose all this new functionality to NT clients..... Jeremy. (This used to be commit 5910d07bbf45a34d3c901461f74704c029a79474)
1998-09-05 00:53:58 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
if (!file_existed) {
Modified dev_t and ino_t code to be 64 bit clean (including changes to oplock break message passing). I think that smbd/nmbd are now inode and offset size independent (at least for 32 bit and 64 bit systems). Now to expose all this new functionality to NT clients..... Jeremy. (This used to be commit 5910d07bbf45a34d3c901461f74704c029a79474)
1998-09-05 00:53:58 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
lock_share_entry_fsp(fsp);
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
num_share_modes = open_mode_check(conn, fname, dev, inode, share_mode,
&flags, &oplock_request, &all_current_opens_are_level_II);
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
if(num_share_modes == -1) {
unlock_share_entry_fsp(fsp);
fd_close(conn,fsp);
file_free(fsp);
return NULL;
}
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
/*
* We exit this block with the share entry *locked*.....
*/
}
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
/*
* At this point onwards, we can guarentee that the share entry
* is locked, whether we created the file or not, and that the
* deny mode is compatible with all current opens.
*/
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
/*
* If requested, truncate the file.
*/
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
if ((flags2&O_TRUNC) && (truncate_unless_locked(conn,fsp) == -1)) {
unlock_share_entry_fsp(fsp);
fd_close(conn,fsp);
file_free(fsp);
return NULL;
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
switch (flags) {
case O_RDONLY:
open_mode = DOS_OPEN_RDONLY;
break;
case O_RDWR:
open_mode = DOS_OPEN_RDWR;
break;
case O_WRONLY:
open_mode = DOS_OPEN_WRONLY;
break;
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
fsp->share_mode = SET_DENY_MODE(deny_mode) |
SET_OPEN_MODE(open_mode) |
SET_ALLOW_SHARE_DELETE(allow_share_delete);
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
if (Access)
(*Access) = open_mode;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
if (action) {
if (file_existed && !(flags2 & O_TRUNC))
*action = FILE_WAS_OPENED;
if (!file_existed)
*action = FILE_WAS_CREATED;
if (file_existed && (flags2 & O_TRUNC))
*action = FILE_WAS_OVERWRITTEN;
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
/*
* Setup the oplock info in both the shared memory and
* file structs.
*/
First cut at kernel oplocks. This should have no effect unless runnin on a machine that supports them in autoconf. Move various functions out of lib/util.c into smbd/process.c and smbd/oplock.c where they belong. Jeremy. (This used to be commit c3c5e13f85c97939746070132dad941e79c546fb)
1998-09-23 05:48:45 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
if(oplock_request && (num_share_modes == 0) &&
!IS_VETO_OPLOCK_PATH(conn,fname) && set_file_oplock(fsp, oplock_request) ) {
port = global_oplock_port;
} else if (oplock_request && all_current_opens_are_level_II) {
port = global_oplock_port;
oplock_request = LEVEL_II_OPLOCK;
set_file_oplock(fsp, oplock_request);
} else {
port = 0;
oplock_request = 0;
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
set_share_mode(fsp, port, oplock_request);
First cut at kernel oplocks. This should have no effect unless runnin on a machine that supports them in autoconf. Move various functions out of lib/util.c into smbd/process.c and smbd/oplock.c where they belong. Jeremy. (This used to be commit c3c5e13f85c97939746070132dad941e79c546fb)
1998-09-23 05:48:45 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
unlock_share_entry_fsp(fsp);
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
conn->num_files_open++;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
return fsp;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
}
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
/****************************************************************************
Open a file for permissions read only. Return a pseudo file entry
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
with the 'stat_open' flag set
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
****************************************************************************/
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
files_struct *open_file_stat(connection_struct *conn,
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
char *fname, int smb_ofun, SMB_STRUCT_STAT *pst, int *action)
{
extern struct current_user current_user;
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
files_struct *fsp = file_new();
if(!fsp)
return NULL;
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
Mega-VFS merge. Yeah baby! Synopsis: change every disk access function to work through a vfs_ops structure contained in the connection_struct. (This used to be commit 3aad500c0fb61232ed3431ff4b743b5d18ec852f)
2000-02-03 08:17:25 +03:00
if(conn->vfs_ops.stat(dos_to_unix(fname, False), pst) < 0) {
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
DEBUG(0,("open_file_stat: unable to stat name = %s. Error was %s\n",
fname, strerror(errno) ));
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
file_free(fsp);
return NULL;
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
}
if(S_ISDIR(pst->st_mode)) {
DEBUG(0,("open_file_stat: %s is a directory !\n", fname ));
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
file_free(fsp);
return NULL;
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
}
*action = FILE_WAS_OPENED;
DEBUG(5,("open_file_stat: opening file %s as a stat entry\n", fname));
/*
* Setup the files_struct for it.
*/
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
fsp->fd = -1;
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
fsp->mode = 0;
GetTimeOfDay(&fsp->open_time);
fsp->vuid = current_user.vuid;
fsp->size = 0;
fsp->pos = -1;
fsp->can_lock = False;
fsp->can_read = False;
fsp->can_write = False;
fsp->share_mode = 0;
fsp->print_file = False;
fsp->modified = False;
fsp->oplock_type = NO_OPLOCK;
fsp->sent_oplock_break = NO_BREAK_SENT;
fsp->is_directory = False;
fsp->stat_open = True;
fsp->directory_delete_on_close = False;
fsp->conn = conn;
/*
* Note that the file name here is the *untranslated* name
* ie. it is still in the DOS codepage sent from the client.
* All use of this filename will pass though the sys_xxxx
* functions which will do the dos_to_unix translation before
* mapping into a UNIX filename. JRA.
*/
string_set(&fsp->fsp_name,fname);
fsp->wbmpx_ptr = NULL;
fsp->wcp = NULL; /* Write cache pointer. */
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
conn->num_files_open++;
return fsp;
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
/****************************************************************************
Open a directory from an NT SMB call.
****************************************************************************/
include/smb.h: Added #defines for lots of things - makes our code a *lot* easier to read. lib/util.c: Fixed Luke's set_first_token() function - should return void. smbd/close.c: Move delete_on_close into file_fd_struct structure. smbd/ipc.c: Changed local_machine back to fstring. smbd/nttrans.c: Use defines for mapping share modes. smbd/open.c: Move delete_on_close into file_fd_struct structure, added code for ALLOW_SHARE_DELETE. smbd/reply.c: Use defines for mapping share modes. smbd/trans2.c: Move delete_on_close into file_fd_struct structure. Jeremy. (This used to be commit 8e1ce307bd6a9056b4a95fe6f52ff42dc6e03a08)
1998-10-23 04:58:28 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
files_struct *open_directory(connection_struct *conn,
configure configure.in: Added checks for statvfs64. Last bit of 64 bit widening (I hope :-). include/config.h.in: Added #undef STAT_STATVFS64. include/includes.h: Added SMB_STRUCT_STATVFS type, Changed SMB_BIG_INTEGER to SMB_BIG_UINT and SMB_BIG_INT types. include/smb.h: Added flag defines from CIFS spec. lib/debug.c: Fixed one more mode_t issue. lib/system.c: Added sys_statvfs wrapper. lib/util.c: Changed trim_string to use size_t. param/loadparm.c: Moved "blocking locks" into locking section. Alphabetised locking options. Question - shuld we do this for all options ? passdb/ldap.c: Changed SMB_BIG_INTEGER to SMB_BIG_UINT. passdb/nispass.c: Changed SMB_BIG_INTEGER to SMB_BIG_UINT. passdb/smbpass.c: Changed SMB_BIG_INTEGER to SMB_BIG_UINT. smbd/dfree.c: Changed to use 64 bit types if available. Moved to use unsigned types. smbd/dosmode.c: Fixed one more mode_t issue. smbd/negprot.c: Changed literals to be FLAG_ #defines. smbd/nttrans.c: Removed dead code. smbd/open.c: Changed disk_free call. smbd/process.c: Changed literals to be FLAG_ #defines. smbd/reply.c: Changed disk_free call. smbd/trans2.c: Fixed but in SMB_QUERY_FS_VOLUME_INFO call. Was using UNICODE - should use ascii. tests/summary.c: Added STAT_STATVFS64 check. Jeremy. (This used to be commit c512b1b91fb7f2a7a93b9033a33e06d966daadb4)
1998-09-18 03:06:57 +04:00
char *fname, int smb_ofun, mode_t unixmode, int *action)
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
{
extern struct current_user current_user;
More abstraction of file system data types, to move to a 64 bit file interface for the NT SMB's. Created a new define, SMB_STRUCT_STAT that currently is defined to be struct stat - this wil change to a user defined type containing 64 bit info when the correct wrappers are written for 64 bit stat(), fstat() and lstat() calls. Also changed all sys_xxxx() calls that were previously just wrappers to the same call prefixed by a dos_to_unix() call into dos_xxxx() calls. This makes it explicit when a pathname translation is being done, and when it is not. Now, all sys_xxx() calls are meant to be wrappers to mask OS differences, and not silently converting filenames on the fly. Jeremy. (This used to be commit 28aa182dbffaa4ffd86047e608400de4b26e80eb)
1998-09-02 00:11:54 +04:00
SMB_STRUCT_STAT st;
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
BOOL got_stat = False;
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
files_struct *fsp = file_new();
if(!fsp)
return NULL;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
Mega-VFS merge. Yeah baby! Synopsis: change every disk access function to work through a vfs_ops structure contained in the connection_struct. (This used to be commit 3aad500c0fb61232ed3431ff4b743b5d18ec852f)
2000-02-03 08:17:25 +03:00
if(conn->vfs_ops.stat(dos_to_unix(fname, False), &st) == 0) {
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
got_stat = True;
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
if (got_stat && (GET_FILE_OPEN_DISPOSITION(smb_ofun) == FILE_EXISTS_FAIL)) {
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
file_free(fsp);
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
errno = EEXIST; /* Setup so correct error is returned to client. */
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
return NULL;
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
}
if (GET_FILE_CREATE_DISPOSITION(smb_ofun) == FILE_CREATE_IF_NOT_EXIST) {
if (got_stat) {
if(!S_ISDIR(st.st_mode)) {
DEBUG(0,("open_directory: %s is not a directory !\n", fname ));
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
file_free(fsp);
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
errno = EACCES;
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
return NULL;
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
}
*action = FILE_WAS_OPENED;
} else {
/*
* Try and create the directory.
*/
if(!CAN_WRITE(conn)) {
DEBUG(2,("open_directory: failing create on read-only share\n"));
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
file_free(fsp);
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
errno = EACCES;
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
return NULL;
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
Put back lots of missing calls to dos_to_unix(). Thanks to aono@cc.osaka-kyoiku.ac.jp (Tomoki AONO) (This used to be commit 176c405d2702a4245561ff56c8eac3c754a0dea3)
2000-02-04 02:08:24 +03:00
if(conn->vfs_ops.mkdir(dos_to_unix(fname, False),
unix_mode(conn,aDIR, fname)) < 0) {
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
DEBUG(0,("open_directory: unable to create %s. Error was %s\n",
fname, strerror(errno) ));
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
file_free(fsp);
return NULL;
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
}
*action = FILE_WAS_CREATED;
}
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
} else {
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
/*
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
* Don't create - just check that it *was* a directory.
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
*/
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
if(!got_stat) {
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
DEBUG(0,("open_directory: unable to stat name = %s. Error was %s\n",
fname, strerror(errno) ));
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
file_free(fsp);
return NULL;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
}
if(!S_ISDIR(st.st_mode)) {
DEBUG(0,("open_directory: %s is not a directory !\n", fname ));
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
file_free(fsp);
return NULL;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
}
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
*action = FILE_WAS_OPENED;
}
DEBUG(5,("open_directory: opening directory %s\n",
fname));
/*
* Setup the files_struct for it.
*/
the bulk of the changes to get rid of fd_ptr and move print open handling to printing/printing.c most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd the changes in open.c are quite dramatic. Most of it is removing all the functions that handled the fd multiplexing (This used to be commit d1827a3648009fd0a0d165055015d9aeda7a1037)
2000-04-10 17:05:23 +04:00
fsp->fd = -1;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
fsp->mode = 0;
GetTimeOfDay(&fsp->open_time);
fsp->vuid = current_user.vuid;
fsp->size = 0;
fsp->pos = -1;
fsp->can_lock = True;
fsp->can_read = False;
fsp->can_write = False;
fsp->share_mode = 0;
fsp->print_file = False;
fsp->modified = False;
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
fsp->oplock_type = NO_OPLOCK;
fsp->sent_oplock_break = NO_BREAK_SENT;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
fsp->is_directory = True;
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
fsp->directory_delete_on_close = False;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
fsp->conn = conn;
/*
* Note that the file name here is the *untranslated* name
* ie. it is still in the DOS codepage sent from the client.
* All use of this filename will pass though the sys_xxxx
* functions which will do the dos_to_unix translation before
* mapping into a UNIX filename. JRA.
*/
string_set(&fsp->fsp_name,fname);
fsp->wbmpx_ptr = NULL;
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
conn->num_files_open++;
return fsp;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
}
/*******************************************************************
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
Check if the share mode on a file allows it to be deleted or unlinked.
Return True if sharing doesn't prevent the operation.
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
********************************************************************/
include/smb.h: Added #defines for lots of things - makes our code a *lot* easier to read. lib/util.c: Fixed Luke's set_first_token() function - should return void. smbd/close.c: Move delete_on_close into file_fd_struct structure. smbd/ipc.c: Changed local_machine back to fstring. smbd/nttrans.c: Use defines for mapping share modes. smbd/open.c: Move delete_on_close into file_fd_struct structure, added code for ALLOW_SHARE_DELETE. smbd/reply.c: Use defines for mapping share modes. smbd/trans2.c: Move delete_on_close into file_fd_struct structure. Jeremy. (This used to be commit 8e1ce307bd6a9056b4a95fe6f52ff42dc6e03a08)
1998-10-23 04:58:28 +04:00
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
BOOL check_file_sharing(connection_struct *conn,char *fname, BOOL rename_op)
{
int i;
int ret = False;
share_mode_entry *old_shares = 0;
int num_share_modes;
More abstraction of file system data types, to move to a 64 bit file interface for the NT SMB's. Created a new define, SMB_STRUCT_STAT that currently is defined to be struct stat - this wil change to a user defined type containing 64 bit info when the correct wrappers are written for 64 bit stat(), fstat() and lstat() calls. Also changed all sys_xxxx() calls that were previously just wrappers to the same call prefixed by a dos_to_unix() call into dos_xxxx() calls. This makes it explicit when a pathname translation is being done, and when it is not. Now, all sys_xxx() calls are meant to be wrappers to mask OS differences, and not silently converting filenames on the fly. Jeremy. (This used to be commit 28aa182dbffaa4ffd86047e608400de4b26e80eb)
1998-09-02 00:11:54 +04:00
SMB_STRUCT_STAT sbuf;
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
pid_t pid = getpid();
configure.in, configure: include/config.h.in: Added stropts and poll. include/smb.h: Moved old typedefs of uint8 etc. into include/includes.h where all the other defines live (changed them from typedefs to defines). Other changes : changed from using uint32 to SMB_DEV_T and SMB_INO_T in preparation for moving to size independed (ie. 64 bit clean) device and inode access. Stat call wrapper comes next :-). Jeremy. (This used to be commit 3d9ec96de5e04e83abafe9c5d980bd39eee856ea)
1998-09-01 00:20:54 +04:00
SMB_DEV_T dev;
SMB_INO_T inode;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
This is a *big* checkin that may break some things, but implements the new open mechanism Andrew & I discussed. config.sub: configure: Included the QNX patch. include/vfs.h: smbd/vfs-wrap.c: smbd/vfs.c: Added ftruncate vfs call (needed). Note that we will also need locking calls in the vfs (to be added). lib/util_unistr.c: nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump. Also fix for sidsize DOS bug. locking/locking.c: Check value of ret before using it for memdup. printing/printing.c: Convert print_fsp_open to return an allocated fsp. rpc_server/srv_lsa.c: Fix for NT domain logons. I have removed all use of lp_share_modes() from the code (although I left the parameter in the table for backwards compatibility). It no longer makes sense for this to exist. smbd/close.c: Removed lp_share_modes(). smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code. smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail. smbd/nttrans.c: smbd/reply.c: smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. smbd/open.c: Changed all occurrences of open_file_shared/open_directory/ open_file_stat to return an fsp from the call. In addition I have fixed a long standing race condition in the deny mode processing w.r.t. two smbd's creating a file. Andrew, please note that your original idea of using open with O_EXCL in this case would not work (I went over the races very carefully) and so we must re-check deny modes *after* the open() call returns. This is because there is a race between the open with O_EXCL and the lock of the share mode entry. Imagine the case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL, but is pre-empted before it locks the share modes and creates the deny mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY and a deny mode of DENY_NONE and the two opens would be allowed. The *only* way to fix this race is to lock the share modes after the open and then do the deny mode checks *after* this lock in the case where the file did not originally exist. This code will need extensive testing but seems to initially work. Jeremy. (This used to be commit ab0ecc39d688f16b9692fe90b991f0b89287070a)
2000-04-22 04:33:16 +04:00
if (conn->vfs_ops.stat(dos_to_unix(fname,False),&sbuf) == -1)
return(True);
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
configure.in, configure: include/config.h.in: Added stropts and poll. include/smb.h: Moved old typedefs of uint8 etc. into include/includes.h where all the other defines live (changed them from typedefs to defines). Other changes : changed from using uint32 to SMB_DEV_T and SMB_INO_T in preparation for moving to size independed (ie. 64 bit clean) device and inode access. Stat call wrapper comes next :-). Jeremy. (This used to be commit 3d9ec96de5e04e83abafe9c5d980bd39eee856ea)
1998-09-01 00:20:54 +04:00
dev = sbuf.st_dev;
inode = sbuf.st_ino;
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
converted all our existing shared memory code to use a tdb database instead of either sysv or mmap shared memory or lock files. this means we can now completely remove locking_shm.c locking_slow.c shmem.c shmem_sysv.c and lots of other things also got simpler locking.c got a bit larger, but is much better compartmentalised now (This used to be commit e48c2d9937eea0667b8cd3332e49c06314ef31e7)
1999-12-21 12:25:59 +03:00
lock_share_entry(conn, dev, inode);
num_share_modes = get_share_modes(conn, dev, inode, &old_shares);
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
/*
* Check if the share modes will give us access.
*/
if(num_share_modes != 0)
{
BOOL broke_oplock;
do
{
broke_oplock = False;
for(i = 0; i < num_share_modes; i++)
{
share_mode_entry *share_entry = &old_shares[i];
/*
* Break oplocks before checking share modes. See comment in
* open_file_shared for details.
* Check if someone has an oplock on this file. If so we must
* break it before continuing.
*/
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
if(BATCH_OPLOCK_TYPE(share_entry->op_type))
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
{
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
#if 0
/* JRA. Try removing this code to see if the new oplock changes
fix the problem. I'm dubious, but Andrew is recommending we
try this....
*/
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
/*
* It appears that the NT redirector may have a bug, in that
* it tries to do an SMBmv on a file that it has open with a
* batch oplock, and then fails to respond to the oplock break
* request. This only seems to occur when the client is doing an
* SMBmv to the smbd it is using - thus we try and detect this
* condition by checking if the file being moved is open and oplocked by
* this smbd process, and then not sending the oplock break in this
* special case. If the file was open with a deny mode that
* prevents the move the SMBmv will fail anyway with a share
* violation error. JRA.
*/
if(rename_op && (share_entry->pid == pid))
{
Modified dev_t and ino_t code to be 64 bit clean (including changes to oplock break message passing). I think that smbd/nmbd are now inode and offset size independent (at least for 32 bit and 64 bit systems). Now to expose all this new functionality to NT clients..... Jeremy. (This used to be commit 5910d07bbf45a34d3c901461f74704c029a79474)
1998-09-05 00:53:58 +04:00
DEBUG(0,("check_file_sharing: NT redirector workaround - rename attempted on \
batch oplocked file %s, dev = %x, inode = %.0f\n", fname, (unsigned int)dev, (double)inode));
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
/*
* This next line is a test that allows the deny-mode
* processing to be skipped. This seems to be needed as
* NT insists on the rename succeeding (in Office 9x no less !).
* This should be removed as soon as (a) MS fix the redirector
* bug or (b) NT SMB support in Samba makes NT not issue the
* call (as is my fervent hope). JRA.
*/
continue;
}
else
first pass at updating head branch to be to be the same as the SAMBA_2_0 branch (This used to be commit 453a822a76780063dff23526c35408866d0c0154)
1999-12-13 16:27:58 +03:00
#endif /* 0 */
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
{
Modified dev_t and ino_t code to be 64 bit clean (including changes to oplock break message passing). I think that smbd/nmbd are now inode and offset size independent (at least for 32 bit and 64 bit systems). Now to expose all this new functionality to NT clients..... Jeremy. (This used to be commit 5910d07bbf45a34d3c901461f74704c029a79474)
1998-09-05 00:53:58 +04:00
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
DEBUG(5,("check_file_sharing: breaking oplock (%x) on file %s, \
Modified dev_t and ino_t code to be 64 bit clean (including changes to oplock break message passing). I think that smbd/nmbd are now inode and offset size independent (at least for 32 bit and 64 bit systems). Now to expose all this new functionality to NT clients..... Jeremy. (This used to be commit 5910d07bbf45a34d3c901461f74704c029a79474)
1998-09-05 00:53:58 +04:00
dev = %x, inode = %.0f\n", share_entry->op_type, fname, (unsigned int)dev, (double)inode));
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
/* Oplock break.... */
converted all our existing shared memory code to use a tdb database instead of either sysv or mmap shared memory or lock files. this means we can now completely remove locking_shm.c locking_slow.c shmem.c shmem_sysv.c and lots of other things also got simpler locking.c got a bit larger, but is much better compartmentalised now (This used to be commit e48c2d9937eea0667b8cd3332e49c06314ef31e7)
1999-12-21 12:25:59 +03:00
unlock_share_entry(conn, dev, inode);
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
if(request_oplock_break(share_entry, dev, inode) == False)
{
free((char *)old_shares);
Modified dev_t and ino_t code to be 64 bit clean (including changes to oplock break message passing). I think that smbd/nmbd are now inode and offset size independent (at least for 32 bit and 64 bit systems). Now to expose all this new functionality to NT clients..... Jeremy. (This used to be commit 5910d07bbf45a34d3c901461f74704c029a79474)
1998-09-05 00:53:58 +04:00
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
DEBUG(0,("check_file_sharing: FAILED when breaking oplock (%x) on file %s, \
Modified dev_t and ino_t code to be 64 bit clean (including changes to oplock break message passing). I think that smbd/nmbd are now inode and offset size independent (at least for 32 bit and 64 bit systems). Now to expose all this new functionality to NT clients..... Jeremy. (This used to be commit 5910d07bbf45a34d3c901461f74704c029a79474)
1998-09-05 00:53:58 +04:00
dev = %x, inode = %.0f\n", old_shares[i].op_type, fname, (unsigned int)dev, (double)inode));
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
return False;
}
converted all our existing shared memory code to use a tdb database instead of either sysv or mmap shared memory or lock files. this means we can now completely remove locking_shm.c locking_slow.c shmem.c shmem_sysv.c and lots of other things also got simpler locking.c got a bit larger, but is much better compartmentalised now (This used to be commit e48c2d9937eea0667b8cd3332e49c06314ef31e7)
1999-12-21 12:25:59 +03:00
lock_share_entry(conn, dev, inode);
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
broke_oplock = True;
break;
}
}
include/smb.h: Added #defines for lots of things - makes our code a *lot* easier to read. lib/util.c: Fixed Luke's set_first_token() function - should return void. smbd/close.c: Move delete_on_close into file_fd_struct structure. smbd/ipc.c: Changed local_machine back to fstring. smbd/nttrans.c: Use defines for mapping share modes. smbd/open.c: Move delete_on_close into file_fd_struct structure, added code for ALLOW_SHARE_DELETE. smbd/reply.c: Use defines for mapping share modes. smbd/trans2.c: Move delete_on_close into file_fd_struct structure. Jeremy. (This used to be commit 8e1ce307bd6a9056b4a95fe6f52ff42dc6e03a08)
1998-10-23 04:58:28 +04:00
/*
* If this is a delete request and ALLOW_SHARE_DELETE is set then allow
* this to proceed. This takes precedence over share modes.
*/
if(!rename_op && GET_ALLOW_SHARE_DELETE(share_entry->share_mode))
continue;
/*
* Someone else has a share lock on it, check to see
* if we can too.
*/
rewrote the access_table() code to get it right for *.exe, *.dll, *.sym and *.com files. I still find it incredible that SMB treats file locking differently depending on the file name - that is so ugly it stinks. anyway, we now behave the same way as NT, except for the same-connection case which I'm still working on. (This used to be commit 5b9be3feca26f1f20fe3f32ba90769cc78f9bb6b)
2000-01-16 14:21:33 +03:00
if ((GET_DENY_MODE(share_entry->share_mode) != DENY_DOS) ||
(share_entry->pid != pid))
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
goto free_and_exit;
} /* end for */
if(broke_oplock)
{
free((char *)old_shares);
converted all our existing shared memory code to use a tdb database instead of either sysv or mmap shared memory or lock files. this means we can now completely remove locking_shm.c locking_slow.c shmem.c shmem_sysv.c and lots of other things also got simpler locking.c got a bit larger, but is much better compartmentalised now (This used to be commit e48c2d9937eea0667b8cd3332e49c06314ef31e7)
1999-12-21 12:25:59 +03:00
num_share_modes = get_share_modes(conn, dev, inode, &old_shares);
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
}
} while(broke_oplock);
}
/* XXXX exactly what share mode combinations should be allowed for
deleting/renaming? */
include/smb.h: Added #defines for lots of things - makes our code a *lot* easier to read. lib/util.c: Fixed Luke's set_first_token() function - should return void. smbd/close.c: Move delete_on_close into file_fd_struct structure. smbd/ipc.c: Changed local_machine back to fstring. smbd/nttrans.c: Use defines for mapping share modes. smbd/open.c: Move delete_on_close into file_fd_struct structure, added code for ALLOW_SHARE_DELETE. smbd/reply.c: Use defines for mapping share modes. smbd/trans2.c: Move delete_on_close into file_fd_struct structure. Jeremy. (This used to be commit 8e1ce307bd6a9056b4a95fe6f52ff42dc6e03a08)
1998-10-23 04:58:28 +04:00
/*
* If we got here then either there were no share modes or
* all share modes were DENY_DOS and the pid == getpid() or
* delete access was requested and all share modes had the
* ALLOW_SHARE_DELETE bit set (takes precedence over other
* share modes).
*/
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
ret = True;
free_and_exit:
converted all our existing shared memory code to use a tdb database instead of either sysv or mmap shared memory or lock files. this means we can now completely remove locking_shm.c locking_slow.c shmem.c shmem_sysv.c and lots of other things also got simpler locking.c got a bit larger, but is much better compartmentalised now (This used to be commit e48c2d9937eea0667b8cd3332e49c06314ef31e7)
1999-12-21 12:25:59 +03:00
unlock_share_entry(conn, dev, inode);
this completes the splitup of server.c. the splitup was done with an axe, not a scalpel, so there are some rough edges. I mostly wanted to get the general form right with fine tuning of what goes where to come later. Still, this is better than what we had before where server.c was a general repository for anything that didn't fit elsewhere. (This used to be commit a6d194886a4a5f7507fa37289ff96c1be56f14a6)
1998-08-17 17:11:34 +04:00
if(old_shares != NULL)
free((char *)old_shares);
return(ret);
}
Copy Permalink