sin/samba-mirror
1
0
Fork 0
mirror of https://github.com/samba-team/samba.git synced 2025-01-21 18:04:06 +03:00
Code
samba-mirror/source4/ntvfs/posix/pvfs_search.c

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

866 lines
26 KiB
C
Raw Normal View History

r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
/*
Unix SMB/CIFS implementation.
POSIX NTVFS backend - directory search functions
Copyright (C) Andrew Tridgell 2004
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
r23792: convert Samba4 to GPLv3 There are still a few tidyups of old FSF addresses to come (in both s3 and s4). More commits soon. (This used to be commit fcf38a38ac691abd0fa51b89dc951a08e89fdafa)
2007-07-10 02:07:03 +00:00
the Free Software Foundation; either version 3 of the License, or
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
(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
r23792: convert Samba4 to GPLv3 There are still a few tidyups of old FSF addresses to come (in both s3 and s4). More commits soon. (This used to be commit fcf38a38ac691abd0fa51b89dc951a08e89fdafa)
2007-07-10 02:07:03 +00:00
along with this program. If not, see <http://www.gnu.org/licenses/>.
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
*/
r3544: fixed some #include lines to make them more consistent (This used to be commit a1a0118bd3f1cae034a1c564d05c3aa8019ff932)
2004-11-05 07:24:25 +00:00
#include "includes.h"
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
#include "vfs_posix.h"
r3447: more include/system/XXX.h include files (This used to be commit 264ce9181089922547e8f6f67116f2d7277a5105)
2004-11-02 00:24:21 +00:00
#include "system/time.h"
r14464: Don't include ndr_BASENAME.h files unless strictly required, instead try to include just the BASENAME.h files (containing only structs) (This used to be commit 3dd477ca5147f28a962b8437e2611a8222d706bd)
2006-03-16 00:23:11 +00:00
#include "librpc/gen_ndr/security.h"
s4: rename source4/smbd/ to source4/samba/ Signed-off-by: Ralph Boehme <slow@samba.org> Reviewed-by: Andrew Bartlett <abartlet@samba.org> Autobuild-User(master): Ralph Böhme <slow@samba.org> Autobuild-Date(master): Fri Nov 27 10:07:18 UTC 2020 on sn-devel-184
2020-11-20 15:27:17 +01:00
#include "samba/service_stream.h"
r8535: no longer rely on seekdir working after a closedir. Instead, keep directories open, but close search states based on an inactivity timer, with a default of a 5 minute timeout (This used to be commit 2e8d154e7dfb9b320a1344e957a39e96e1eefadd)
2005-07-18 03:35:52 +00:00
#include "lib/events/events.h"
Fix include paths to new location of libutil.
2008-10-11 21:31:42 +02:00
#include "../lib/util/dlinklist.h"
lib: Remove idtree from samba_util.h No need to recompile the world when only a few files need this. Signed-off-by: Volker Lendecke <vl@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
2023-01-05 10:04:23 +01:00
#include "lib/util/idtree.h"
r3276: - allow for more than 256 open old style searches (limit currently set at an arbitrary 5000) - auto-cleanup old searches that the client forgot to close (common with old searches) - expanded the RAW-SEARCH test to test more than 256 old searches, and old search rewind (w2k3 fails this - it appears to not support rewind on old style searches) (This used to be commit bc83d823b2140a10007490bf0101843a886f99a6)
2004-10-27 01:11:44 +00:00
/* place a reasonable limit on old-style searches as clients tend to
not send search close requests */
#define MAX_OLD_SEARCHES 2000
r17752: using -1 in a uint16_t value causes compiler warnings, use UINT16_MAX as invalid search handle metze (This used to be commit a4e19d6cc4cce611241c502de485dbd2496be0d3)
2006-08-23 13:06:31 +00:00
#define MAX_SEARCH_HANDLES (UINT16_MAX - 1)
#define INVALID_SEARCH_HANDLE UINT16_MAX
r3276: - allow for more than 256 open old style searches (limit currently set at an arbitrary 5000) - auto-cleanup old searches that the client forgot to close (common with old searches) - expanded the RAW-SEARCH test to test more than 256 old searches, and old search rewind (w2k3 fails this - it appears to not support rewind on old style searches) (This used to be commit bc83d823b2140a10007490bf0101843a886f99a6)
2004-10-27 01:11:44 +00:00
r3056: added a id -> pointer data structure (a type of radix tree). This is an extremely efficient way of mapping from an integer handle (such as an open file handle) to a pointer (such as the structure containing the open file information). The code is taken from lib/idr.c in the 2.6 Linux kernel, and is very fast and space efficient. By using talloc it even has auto cleanup. This commit converts the handling of open file handles and open directory search handles to use the idtree routines. In combination with talloc destructors, this simplifies the structure handling in the pvfs backend a lot. For example, we no longer need to keep a linked list of open directory searches at all, and we no longer need to do linear scans of the list of open files on most operations. The end result is that the pvfs code is now extremely scalable. You can have 10s of thousands of open files and open searches and the code still runs very fast. I have also added a small optimisation into the file close path, to avoid looking in the byte range locking database if we know that there are no locks outstanding. (This used to be commit 16835a0ef91a16fa01145b773aad8d43da215dbf)
2004-10-19 06:39:51 +00:00
/*
destroy an open search
*/
r15855: more talloc_set_destructor() typesafe fixes. nearly done ... (This used to be commit 396d82a231b6e3a91178db08b706626d4d4b420c)
2006-05-24 07:35:06 +00:00
static int pvfs_search_destructor(struct pvfs_search_state *search)
r3056: added a id -> pointer data structure (a type of radix tree). This is an extremely efficient way of mapping from an integer handle (such as an open file handle) to a pointer (such as the structure containing the open file information). The code is taken from lib/idr.c in the 2.6 Linux kernel, and is very fast and space efficient. By using talloc it even has auto cleanup. This commit converts the handling of open file handles and open directory search handles to use the idtree routines. In combination with talloc destructors, this simplifies the structure handling in the pvfs backend a lot. For example, we no longer need to keep a linked list of open directory searches at all, and we no longer need to do linear scans of the list of open files on most operations. The end result is that the pvfs code is now extremely scalable. You can have 10s of thousands of open files and open searches and the code still runs very fast. I have also added a small optimisation into the file close path, to avoid looking in the byte range locking database if we know that there are no locks outstanding. (This used to be commit 16835a0ef91a16fa01145b773aad8d43da215dbf)
2004-10-19 06:39:51 +00:00
{
r14540: fix a talloc hierachie problem, make sure file and search handles are cleaned up before anything else in the pvfs_state struct, as there destructors reply on a valid pvfs_state struct metze (This used to be commit aaa5d377b9b6145a83c0e686c7fbb7b561ae8988)
2006-03-18 10:38:38 +00:00
DLIST_REMOVE(search->pvfs->search.list, search);
idr_remove(search->pvfs->search.idtree, search->handle);
r3056: added a id -> pointer data structure (a type of radix tree). This is an extremely efficient way of mapping from an integer handle (such as an open file handle) to a pointer (such as the structure containing the open file information). The code is taken from lib/idr.c in the 2.6 Linux kernel, and is very fast and space efficient. By using talloc it even has auto cleanup. This commit converts the handling of open file handles and open directory search handles to use the idtree routines. In combination with talloc destructors, this simplifies the structure handling in the pvfs backend a lot. For example, we no longer need to keep a linked list of open directory searches at all, and we no longer need to do linear scans of the list of open files on most operations. The end result is that the pvfs code is now extremely scalable. You can have 10s of thousands of open files and open searches and the code still runs very fast. I have also added a small optimisation into the file close path, to avoid looking in the byte range locking database if we know that there are no locks outstanding. (This used to be commit 16835a0ef91a16fa01145b773aad8d43da215dbf)
2004-10-19 06:39:51 +00:00
return 0;
}
r8535: no longer rely on seekdir working after a closedir. Instead, keep directories open, but close search states based on an inactivity timer, with a default of a 5 minute timeout (This used to be commit 2e8d154e7dfb9b320a1344e957a39e96e1eefadd)
2005-07-18 03:35:52 +00:00
/*
called when a search timer goes off
*/
s4:lib/tevent: rename structs list="" list="$list event_context:tevent_context" list="$list fd_event:tevent_fd" list="$list timed_event:tevent_timer" for s in $list; do o=`echo $s | cut -d ':' -f1` n=`echo $s | cut -d ':' -f2` r=`git grep "struct $o" |cut -d ':' -f1 |sort -u` files=`echo "$r" | grep -v source3 | grep -v nsswitch | grep -v packaging4` for f in $files; do cat $f | sed -e "s/struct $o/struct $n/g" > $f.tmp mv $f.tmp $f done done metze
2008-12-29 20:24:57 +01:00
static void pvfs_search_timer(struct tevent_context *ev, struct tevent_timer *te,
r8535: no longer rely on seekdir working after a closedir. Instead, keep directories open, but close search states based on an inactivity timer, with a default of a 5 minute timeout (This used to be commit 2e8d154e7dfb9b320a1344e957a39e96e1eefadd)
2005-07-18 03:35:52 +00:00
struct timeval t, void *ptr)
{
struct pvfs_search_state *search = talloc_get_type(ptr, struct pvfs_search_state);
talloc_free(search);
}
/*
setup a timer to destroy a open search after a inactivity period
*/
static void pvfs_search_setup_timer(struct pvfs_search_state *search)
{
s4:lib/tevent: rename structs list="" list="$list event_context:tevent_context" list="$list fd_event:tevent_fd" list="$list timed_event:tevent_timer" for s in $list; do o=`echo $s | cut -d ':' -f1` n=`echo $s | cut -d ':' -f2` r=`git grep "struct $o" |cut -d ':' -f1 |sort -u` files=`echo "$r" | grep -v source3 | grep -v nsswitch | grep -v packaging4` for f in $files; do cat $f | sed -e "s/struct $o/struct $n/g" > $f.tmp mv $f.tmp $f done done metze
2008-12-29 20:24:57 +01:00
struct tevent_context *ev = search->pvfs->ntvfs->ctx->event_ctx;
r17752: using -1 in a uint16_t value causes compiler warnings, use UINT16_MAX as invalid search handle metze (This used to be commit a4e19d6cc4cce611241c502de485dbd2496be0d3)
2006-08-23 13:06:31 +00:00
if (search->handle == INVALID_SEARCH_HANDLE) return;
r8535: no longer rely on seekdir working after a closedir. Instead, keep directories open, but close search states based on an inactivity timer, with a default of a 5 minute timeout (This used to be commit 2e8d154e7dfb9b320a1344e957a39e96e1eefadd)
2005-07-18 03:35:52 +00:00
talloc_free(search->te);
s4:ntvfs: use tevent_ fn names instead of legacy event_ ones
2010-05-25 15:27:41 -04:00
search->te = tevent_add_timer(ev, search,
r14540: fix a talloc hierachie problem, make sure file and search handles are cleaned up before anything else in the pvfs_state struct, as there destructors reply on a valid pvfs_state struct metze (This used to be commit aaa5d377b9b6145a83c0e686c7fbb7b561ae8988)
2006-03-18 10:38:38 +00:00
timeval_current_ofs(search->pvfs->search.inactivity_time, 0),
r8535: no longer rely on seekdir working after a closedir. Instead, keep directories open, but close search states based on an inactivity timer, with a default of a 5 minute timeout (This used to be commit 2e8d154e7dfb9b320a1344e957a39e96e1eefadd)
2005-07-18 03:35:52 +00:00
pvfs_search_timer, search);
}
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
/*
fill in a single search result for a given info level
*/
static NTSTATUS fill_search_info(struct pvfs_state *pvfs,
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
enum smb_search_data_level level,
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
const char *unix_path,
r2436: the second big lump of posix vfs code. this is still just a skeleton, and many of the functions are just based on the simple vfs backend, they are there to allow me to run smbtorture tests against the real parts of the posix backend. (This used to be commit f2fa7fe565e89360dba3bb5434d3a6a36f398348)
2004-09-20 07:28:43 +00:00
const char *fname,
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
struct pvfs_search_state *search,
r18319: fixed the directory search resume code on IRIX The problem was twofold: 1) irix returns 64 bit numbers in telldir(). The protocol uses a 32 bit resume key. We now cope with this properly using the code in pvfs_list_seek_ofs(). 2) irix returns 0xFFFFFFFF from telldir() for the last entry in the directory. When added to DIR_OFFSET_BASE this became DIR_OFFSET_DOTDOT which meant an infinite loop! (This used to be commit 8cce9740ed0da9f08d6821beb4acaa9d28d149c2)
2006-09-10 07:24:41 +00:00
off_t dir_offset,
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
union smb_search_data *file)
{
r2436: the second big lump of posix vfs code. this is still just a skeleton, and many of the functions are just based on the simple vfs backend, they are there to allow me to run smbtorture tests against the real parts of the posix backend. (This used to be commit f2fa7fe565e89360dba3bb5434d3a6a36f398348)
2004-09-20 07:28:43 +00:00
struct pvfs_filename *name;
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
NTSTATUS status;
r2904: - fixed the old style SMBsearch to return the pvfs shortname, not a truncated long name. - short name can be up to 12 bytes, not 11 (This used to be commit 657103ec6264bf4c2986fedd3fc6577746395d49)
2004-10-11 03:27:16 +00:00
const char *shortname;
r18319: fixed the directory search resume code on IRIX The problem was twofold: 1) irix returns 64 bit numbers in telldir(). The protocol uses a 32 bit resume key. We now cope with this properly using the code in pvfs_list_seek_ofs(). 2) irix returns 0xFFFFFFFF from telldir() for the last entry in the directory. When added to DIR_OFFSET_BASE this became DIR_OFFSET_DOTDOT which meant an infinite loop! (This used to be commit 8cce9740ed0da9f08d6821beb4acaa9d28d149c2)
2006-09-10 07:24:41 +00:00
uint32_t dir_index = (uint32_t)dir_offset; /* truncated - see the code
in pvfs_list_seek_ofs() for
how we cope with this */
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
pvfs: add PVFS_RESOLVE_NO_OPENDB flag and get the write time from the opendb By default get the current write time from the opendb, but allow callers to pass PVFS_RESOLVE_NO_OPENDB for performance reasons, if they don't need to the write time. metze (This used to be commit def52cc0988c26a815e74b3391e5857512408d90)
2008-05-05 12:18:47 +02:00
status = pvfs_resolve_partial(pvfs, file, unix_path, fname, 0, &name);
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
if (!NT_STATUS_IS_OK(status)) {
return status;
}
r2436: the second big lump of posix vfs code. this is still just a skeleton, and many of the functions are just based on the simple vfs backend, they are there to allow me to run smbtorture tests against the real parts of the posix backend. (This used to be commit f2fa7fe565e89360dba3bb5434d3a6a36f398348)
2004-09-20 07:28:43 +00:00
r3240: - update the rules for what error codes should be given on the different type of unlink an seach mismatches - wildcard directory listings that have attribute FILE_ATTRIBUTE_DIRECTORY and match "." or ".." should be failed. - don't set the write_time on SMBclose unless it is non-zero - added much better support for setfileinfo and setpathinfo in pvfs - better (and more efficient) handling of .. and . components in filenames (This used to be commit 9305b07af395a158cb9f0c1c9486f7122c79d357)
2004-10-26 05:39:54 +00:00
status = pvfs_match_attrib(pvfs, name, search->search_attrib, search->must_attrib);
if (!NT_STATUS_IS_OK(status)) {
return status;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
}
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
switch (level) {
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
case RAW_SEARCH_DATA_SEARCH:
r2904: - fixed the old style SMBsearch to return the pvfs shortname, not a truncated long name. - short name can be up to 12 bytes, not 11 (This used to be commit 657103ec6264bf4c2986fedd3fc6577746395d49)
2004-10-11 03:27:16 +00:00
shortname = pvfs_short_name(pvfs, name, name);
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
file->search.attrib = name->dos.attrib;
file->search.write_time = nt_time_to_unix(name->dos.write_time);
file->search.size = name->st.st_size;
r2904: - fixed the old style SMBsearch to return the pvfs shortname, not a truncated long name. - short name can be up to 12 bytes, not 11 (This used to be commit 657103ec6264bf4c2986fedd3fc6577746395d49)
2004-10-11 03:27:16 +00:00
file->search.name = shortname;
r3276: - allow for more than 256 open old style searches (limit currently set at an arbitrary 5000) - auto-cleanup old searches that the client forgot to close (common with old searches) - expanded the RAW-SEARCH test to test more than 256 old searches, and old search rewind (w2k3 fails this - it appears to not support rewind on old style searches) (This used to be commit bc83d823b2140a10007490bf0101843a886f99a6)
2004-10-27 01:11:44 +00:00
file->search.id.reserved = search->handle >> 8;
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
memset(file->search.id.name, ' ', sizeof(file->search.id.name));
r2904: - fixed the old style SMBsearch to return the pvfs shortname, not a truncated long name. - short name can be up to 12 bytes, not 11 (This used to be commit 657103ec6264bf4c2986fedd3fc6577746395d49)
2004-10-11 03:27:16 +00:00
memcpy(file->search.id.name, shortname,
MIN(strlen(shortname)+1, sizeof(file->search.id.name)));
r3276: - allow for more than 256 open old style searches (limit currently set at an arbitrary 5000) - auto-cleanup old searches that the client forgot to close (common with old searches) - expanded the RAW-SEARCH test to test more than 256 old searches, and old search rewind (w2k3 fails this - it appears to not support rewind on old style searches) (This used to be commit bc83d823b2140a10007490bf0101843a886f99a6)
2004-10-27 01:11:44 +00:00
file->search.id.handle = search->handle & 0xFF;
r3260: redid the pvfs_dirlist() interface in preparation for a "keep directory open" implementation, as opposed to the "load the whole directory" interface used now. This will be needed to pass RAW-SEARCH (This used to be commit 692623c6c0a2c6817fddfa77cd1c2525c27145c4)
2004-10-26 11:11:16 +00:00
file->search.id.server_cookie = dir_index;
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
file->search.id.client_cookie = 0;
return NT_STATUS_OK;
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
case RAW_SEARCH_DATA_STANDARD:
r3260: redid the pvfs_dirlist() interface in preparation for a "keep directory open" implementation, as opposed to the "load the whole directory" interface used now. This will be needed to pass RAW-SEARCH (This used to be commit 692623c6c0a2c6817fddfa77cd1c2525c27145c4)
2004-10-26 11:11:16 +00:00
file->standard.resume_key = dir_index;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
file->standard.create_time = nt_time_to_unix(name->dos.create_time);
file->standard.access_time = nt_time_to_unix(name->dos.access_time);
file->standard.write_time = nt_time_to_unix(name->dos.write_time);
file->standard.size = name->st.st_size;
file->standard.alloc_size = name->dos.alloc_size;
file->standard.attrib = name->dos.attrib;
file->standard.name.s = fname;
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
return NT_STATUS_OK;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
case RAW_SEARCH_DATA_EA_SIZE:
r3260: redid the pvfs_dirlist() interface in preparation for a "keep directory open" implementation, as opposed to the "load the whole directory" interface used now. This will be needed to pass RAW-SEARCH (This used to be commit 692623c6c0a2c6817fddfa77cd1c2525c27145c4)
2004-10-26 11:11:16 +00:00
file->ea_size.resume_key = dir_index;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
file->ea_size.create_time = nt_time_to_unix(name->dos.create_time);
file->ea_size.access_time = nt_time_to_unix(name->dos.access_time);
file->ea_size.write_time = nt_time_to_unix(name->dos.write_time);
file->ea_size.size = name->st.st_size;
file->ea_size.alloc_size = name->dos.alloc_size;
file->ea_size.attrib = name->dos.attrib;
file->ea_size.ea_size = name->dos.ea_size;
file->ea_size.name.s = fname;
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
return NT_STATUS_OK;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
case RAW_SEARCH_DATA_EA_LIST:
r4263: added support for the trans2 RAW_SEARCH_EA_LIST information level. This is quite a strange level that we've never seen before, but is used by the os2 workplace shell. note w2k screws up this level when unicode is negotiated, so it only passes the RAW-SEARCH test when you force non-unicode (This used to be commit 25189b8fbf6515d573e3398dc9fca56505dc37b9)
2004-12-18 04:38:43 +00:00
file->ea_list.resume_key = dir_index;
file->ea_list.create_time = nt_time_to_unix(name->dos.create_time);
file->ea_list.access_time = nt_time_to_unix(name->dos.access_time);
file->ea_list.write_time = nt_time_to_unix(name->dos.write_time);
file->ea_list.size = name->st.st_size;
file->ea_list.alloc_size = name->dos.alloc_size;
file->ea_list.attrib = name->dos.attrib;
file->ea_list.name.s = fname;
return pvfs_query_ea_list(pvfs, file, name, -1,
search->num_ea_names,
search->ea_names,
&file->ea_list.eas);
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
case RAW_SEARCH_DATA_DIRECTORY_INFO:
r3260: redid the pvfs_dirlist() interface in preparation for a "keep directory open" implementation, as opposed to the "load the whole directory" interface used now. This will be needed to pass RAW-SEARCH (This used to be commit 692623c6c0a2c6817fddfa77cd1c2525c27145c4)
2004-10-26 11:11:16 +00:00
file->directory_info.file_index = dir_index;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
file->directory_info.create_time = name->dos.create_time;
file->directory_info.access_time = name->dos.access_time;
file->directory_info.write_time = name->dos.write_time;
file->directory_info.change_time = name->dos.change_time;
file->directory_info.size = name->st.st_size;
file->directory_info.alloc_size = name->dos.alloc_size;
file->directory_info.attrib = name->dos.attrib;
file->directory_info.name.s = fname;
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
return NT_STATUS_OK;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
case RAW_SEARCH_DATA_FULL_DIRECTORY_INFO:
r3260: redid the pvfs_dirlist() interface in preparation for a "keep directory open" implementation, as opposed to the "load the whole directory" interface used now. This will be needed to pass RAW-SEARCH (This used to be commit 692623c6c0a2c6817fddfa77cd1c2525c27145c4)
2004-10-26 11:11:16 +00:00
file->full_directory_info.file_index = dir_index;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
file->full_directory_info.create_time = name->dos.create_time;
file->full_directory_info.access_time = name->dos.access_time;
file->full_directory_info.write_time = name->dos.write_time;
file->full_directory_info.change_time = name->dos.change_time;
file->full_directory_info.size = name->st.st_size;
file->full_directory_info.alloc_size = name->dos.alloc_size;
file->full_directory_info.attrib = name->dos.attrib;
file->full_directory_info.ea_size = name->dos.ea_size;
file->full_directory_info.name.s = fname;
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
return NT_STATUS_OK;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
case RAW_SEARCH_DATA_NAME_INFO:
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
file->name_info.file_index = dir_index;
file->name_info.name.s = fname;
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
return NT_STATUS_OK;
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
case RAW_SEARCH_DATA_BOTH_DIRECTORY_INFO:
r3260: redid the pvfs_dirlist() interface in preparation for a "keep directory open" implementation, as opposed to the "load the whole directory" interface used now. This will be needed to pass RAW-SEARCH (This used to be commit 692623c6c0a2c6817fddfa77cd1c2525c27145c4)
2004-10-26 11:11:16 +00:00
file->both_directory_info.file_index = dir_index;
r2436: the second big lump of posix vfs code. this is still just a skeleton, and many of the functions are just based on the simple vfs backend, they are there to allow me to run smbtorture tests against the real parts of the posix backend. (This used to be commit f2fa7fe565e89360dba3bb5434d3a6a36f398348)
2004-09-20 07:28:43 +00:00
file->both_directory_info.create_time = name->dos.create_time;
file->both_directory_info.access_time = name->dos.access_time;
file->both_directory_info.write_time = name->dos.write_time;
file->both_directory_info.change_time = name->dos.change_time;
file->both_directory_info.size = name->st.st_size;
file->both_directory_info.alloc_size = name->dos.alloc_size;
file->both_directory_info.attrib = name->dos.attrib;
file->both_directory_info.ea_size = name->dos.ea_size;
r2788: prevent a memory leak in the pvfs search backend (This used to be commit 1de22070610231e60d329f56997bbec2cc674a4e)
2004-10-02 13:29:20 +00:00
file->both_directory_info.short_name.s = pvfs_short_name(pvfs, file, name);
r2436: the second big lump of posix vfs code. this is still just a skeleton, and many of the functions are just based on the simple vfs backend, they are there to allow me to run smbtorture tests against the real parts of the posix backend. (This used to be commit f2fa7fe565e89360dba3bb5434d3a6a36f398348)
2004-09-20 07:28:43 +00:00
file->both_directory_info.name.s = fname;
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
return NT_STATUS_OK;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
case RAW_SEARCH_DATA_ID_FULL_DIRECTORY_INFO:
r3260: redid the pvfs_dirlist() interface in preparation for a "keep directory open" implementation, as opposed to the "load the whole directory" interface used now. This will be needed to pass RAW-SEARCH (This used to be commit 692623c6c0a2c6817fddfa77cd1c2525c27145c4)
2004-10-26 11:11:16 +00:00
file->id_full_directory_info.file_index = dir_index;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
file->id_full_directory_info.create_time = name->dos.create_time;
file->id_full_directory_info.access_time = name->dos.access_time;
file->id_full_directory_info.write_time = name->dos.write_time;
file->id_full_directory_info.change_time = name->dos.change_time;
file->id_full_directory_info.size = name->st.st_size;
file->id_full_directory_info.alloc_size = name->dos.alloc_size;
file->id_full_directory_info.attrib = name->dos.attrib;
file->id_full_directory_info.ea_size = name->dos.ea_size;
file->id_full_directory_info.file_id = name->dos.file_id;
file->id_full_directory_info.name.s = fname;
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
return NT_STATUS_OK;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
case RAW_SEARCH_DATA_ID_BOTH_DIRECTORY_INFO:
r3260: redid the pvfs_dirlist() interface in preparation for a "keep directory open" implementation, as opposed to the "load the whole directory" interface used now. This will be needed to pass RAW-SEARCH (This used to be commit 692623c6c0a2c6817fddfa77cd1c2525c27145c4)
2004-10-26 11:11:16 +00:00
file->id_both_directory_info.file_index = dir_index;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
file->id_both_directory_info.create_time = name->dos.create_time;
file->id_both_directory_info.access_time = name->dos.access_time;
file->id_both_directory_info.write_time = name->dos.write_time;
file->id_both_directory_info.change_time = name->dos.change_time;
file->id_both_directory_info.size = name->st.st_size;
file->id_both_directory_info.alloc_size = name->dos.alloc_size;
file->id_both_directory_info.attrib = name->dos.attrib;
file->id_both_directory_info.ea_size = name->dos.ea_size;
file->id_both_directory_info.file_id = name->dos.file_id;
r2788: prevent a memory leak in the pvfs search backend (This used to be commit 1de22070610231e60d329f56997bbec2cc674a4e)
2004-10-02 13:29:20 +00:00
file->id_both_directory_info.short_name.s = pvfs_short_name(pvfs, file, name);
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
file->id_both_directory_info.name.s = fname;
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
return NT_STATUS_OK;
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
s4:ntvfs subsystems - rework it using concrete enum values This changes commit 260bc987b00b3fff6c9b99211627b14e9bd0789a to comply with metze's plans.
2011-06-16 08:39:03 +02:00
case RAW_SEARCH_DATA_GENERIC:
case RAW_SEARCH_DATA_UNIX_INFO:
case RAW_SEARCH_DATA_UNIX_INFO2:
s4:ntvfs subsystem - quiet enum warnings Simply return "NT_STATUS_INVALID_LEVEL" for unknown types of requests. Reviewed-by: Tridge
2011-06-08 19:00:18 +02:00
return NT_STATUS_INVALID_LEVEL;
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
}
s4:ntvfs subsystems - rework it using concrete enum values This changes commit 260bc987b00b3fff6c9b99211627b14e9bd0789a to comply with metze's plans.
2011-06-16 08:39:03 +02:00
return NT_STATUS_INVALID_LEVEL;
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
}
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
/*
the search fill loop
*/
static NTSTATUS pvfs_search_fill(struct pvfs_state *pvfs, TALLOC_CTX *mem_ctx,
Change uint_t to unsigned int in source4 Signed-off-by: Stefan Metzmacher <metze@samba.org>
2010-01-05 09:42:54 -08:00
unsigned int max_count,
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
struct pvfs_search_state *search,
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
enum smb_search_data_level level,
Change uint_t to unsigned int in source4 Signed-off-by: Stefan Metzmacher <metze@samba.org>
2010-01-05 09:42:54 -08:00
unsigned int *reply_count,
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
void *search_private,
r25554: Convert last instances of BOOL, True and False to the standard types. (This used to be commit 566aa14139510788548a874e9213d91317f83ca9)
2007-10-06 22:28:14 +00:00
bool (*callback)(void *, const union smb_search_data *))
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
{
struct pvfs_dir *dir = search->dir;
NTSTATUS status;
*reply_count = 0;
r3260: redid the pvfs_dirlist() interface in preparation for a "keep directory open" implementation, as opposed to the "load the whole directory" interface used now. This will be needed to pass RAW-SEARCH (This used to be commit 692623c6c0a2c6817fddfa77cd1c2525c27145c4)
2004-10-26 11:11:16 +00:00
if (max_count == 0) {
max_count = 1;
}
while ((*reply_count) < max_count) {
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
union smb_search_data *file;
r3260: redid the pvfs_dirlist() interface in preparation for a "keep directory open" implementation, as opposed to the "load the whole directory" interface used now. This will be needed to pass RAW-SEARCH (This used to be commit 692623c6c0a2c6817fddfa77cd1c2525c27145c4)
2004-10-26 11:11:16 +00:00
const char *name;
r18319: fixed the directory search resume code on IRIX The problem was twofold: 1) irix returns 64 bit numbers in telldir(). The protocol uses a 32 bit resume key. We now cope with this properly using the code in pvfs_list_seek_ofs(). 2) irix returns 0xFFFFFFFF from telldir() for the last entry in the directory. When added to DIR_OFFSET_BASE this became DIR_OFFSET_DOTDOT which meant an infinite loop! (This used to be commit 8cce9740ed0da9f08d6821beb4acaa9d28d149c2)
2006-09-10 07:24:41 +00:00
off_t ofs = search->current_index;
r3260: redid the pvfs_dirlist() interface in preparation for a "keep directory open" implementation, as opposed to the "load the whole directory" interface used now. This will be needed to pass RAW-SEARCH (This used to be commit 692623c6c0a2c6817fddfa77cd1c2525c27145c4)
2004-10-26 11:11:16 +00:00
r3262: - new pvfs_dirlist code that reopens the directory between search calls. This is needed to allow for "new files appear during a search" behaviour - pvfs now passes RAW-SEARCH (This used to be commit 0d98f7653a1d58510a6cd4c2ac6c5e05c541109c)
2004-10-26 13:18:34 +00:00
name = pvfs_list_next(dir, &search->current_index);
r3260: redid the pvfs_dirlist() interface in preparation for a "keep directory open" implementation, as opposed to the "load the whole directory" interface used now. This will be needed to pass RAW-SEARCH (This used to be commit 692623c6c0a2c6817fddfa77cd1c2525c27145c4)
2004-10-26 11:11:16 +00:00
if (name == NULL) break;
r5037: got rid of all of the TALLOC_DEPRECATED stuff. My apologies for the large commit. I thought this was worthwhile to get done for consistency. (This used to be commit ec32b22ed5ec224f6324f5e069d15e92e38e15c0)
2005-01-27 07:08:20 +00:00
file = talloc(mem_ctx, union smb_search_data);
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
if (!file) {
return NT_STATUS_NO_MEMORY;
}
r3260: redid the pvfs_dirlist() interface in preparation for a "keep directory open" implementation, as opposed to the "load the whole directory" interface used now. This will be needed to pass RAW-SEARCH (This used to be commit 692623c6c0a2c6817fddfa77cd1c2525c27145c4)
2004-10-26 11:11:16 +00:00
status = fill_search_info(pvfs, level,
pvfs_list_unix_path(dir), name,
search, search->current_index, file);
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
if (!NT_STATUS_IS_OK(status)) {
talloc_free(file);
r3248: don't stop searches on failed fill_search_info() (This used to be commit 4cdd750b5cc9b26dbbaebe58b4f1c3984dac8d6c)
2004-10-26 07:04:45 +00:00
continue;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
}
if (!callback(search_private, file)) {
talloc_free(file);
r3262: - new pvfs_dirlist code that reopens the directory between search calls. This is needed to allow for "new files appear during a search" behaviour - pvfs now passes RAW-SEARCH (This used to be commit 0d98f7653a1d58510a6cd4c2ac6c5e05c541109c)
2004-10-26 13:18:34 +00:00
search->current_index = ofs;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
break;
}
r3260: redid the pvfs_dirlist() interface in preparation for a "keep directory open" implementation, as opposed to the "load the whole directory" interface used now. This will be needed to pass RAW-SEARCH (This used to be commit 692623c6c0a2c6817fddfa77cd1c2525c27145c4)
2004-10-26 11:11:16 +00:00
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
(*reply_count)++;
talloc_free(file);
}
r8535: no longer rely on seekdir working after a closedir. Instead, keep directories open, but close search states based on an inactivity timer, with a default of a 5 minute timeout (This used to be commit 2e8d154e7dfb9b320a1344e957a39e96e1eefadd)
2005-07-18 03:35:52 +00:00
pvfs_search_setup_timer(search);
r3262: - new pvfs_dirlist code that reopens the directory between search calls. This is needed to allow for "new files appear during a search" behaviour - pvfs now passes RAW-SEARCH (This used to be commit 0d98f7653a1d58510a6cd4c2ac6c5e05c541109c)
2004-10-26 13:18:34 +00:00
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
return NT_STATUS_OK;
}
r3276: - allow for more than 256 open old style searches (limit currently set at an arbitrary 5000) - auto-cleanup old searches that the client forgot to close (common with old searches) - expanded the RAW-SEARCH test to test more than 256 old searches, and old search rewind (w2k3 fails this - it appears to not support rewind on old style searches) (This used to be commit bc83d823b2140a10007490bf0101843a886f99a6)
2004-10-27 01:11:44 +00:00
/*
we've run out of search handles - cleanup those that the client forgot
to close
*/
static void pvfs_search_cleanup(struct pvfs_state *pvfs)
{
int i;
s4/pvfs: use monotonic time for this timeout
2010-09-10 20:46:10 +02:00
time_t t = time_mono(NULL);
r3276: - allow for more than 256 open old style searches (limit currently set at an arbitrary 5000) - auto-cleanup old searches that the client forgot to close (common with old searches) - expanded the RAW-SEARCH test to test more than 256 old searches, and old search rewind (w2k3 fails this - it appears to not support rewind on old style searches) (This used to be commit bc83d823b2140a10007490bf0101843a886f99a6)
2004-10-27 01:11:44 +00:00
for (i=0;i<MAX_OLD_SEARCHES;i++) {
s4:pvfs: use talloc_get_type() to cast from void * metze
2009-02-04 08:52:41 +01:00
struct pvfs_search_state *search;
void *p = idr_find(pvfs->search.idtree, i);
if (p == NULL) return;
search = talloc_get_type(p, struct pvfs_search_state);
r3276: - allow for more than 256 open old style searches (limit currently set at an arbitrary 5000) - auto-cleanup old searches that the client forgot to close (common with old searches) - expanded the RAW-SEARCH test to test more than 256 old searches, and old search rewind (w2k3 fails this - it appears to not support rewind on old style searches) (This used to be commit bc83d823b2140a10007490bf0101843a886f99a6)
2004-10-27 01:11:44 +00:00
if (pvfs_list_eos(search->dir, search->current_index) &&
search->last_used != 0 &&
t > search->last_used + 30) {
/* its almost certainly been forgotten
about */
talloc_free(search);
}
}
}
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
/*
list files in a directory matching a wildcard pattern - old SMBsearch interface
*/
r2751: this is a new ntvfs design which tries to solve: - the stacking of modules - finding the modules private data - hide the ntvfs details from the calling layer - I set NTVFS_INTERFACE_VERSION 0 till we are closer to release (because we need to solve some async problems with the module stacking) metze (This used to be commit 3ff03b5cb21bb79afdd3b1609be9635f6688a539)
2004-09-29 13:17:09 +00:00
static NTSTATUS pvfs_search_first_old(struct ntvfs_module_context *ntvfs,
r14157: - pass a struct ntvfs_request to the ntvfs layer (for now we just do #define ntvfs_request smbsrv_request, but it's the first step...) - rename ntvfs_openfile() -> ntvfs_open() - fix the talloc hierachie in some places in the ntvfs_map_*() code metze (This used to be commit ed9ed1f48f602354810937c0b0de850b44322191)
2006-03-10 14:31:17 +00:00
struct ntvfs_request *req, union smb_search_first *io,
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
void *search_private,
r25554: Convert last instances of BOOL, True and False to the standard types. (This used to be commit 566aa14139510788548a874e9213d91317f83ca9)
2007-10-06 22:28:14 +00:00
bool (*callback)(void *, const union smb_search_data *))
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
{
struct pvfs_dir *dir;
s4:pvfs: use talloc_get_type() to cast from void * metze
2009-02-04 08:52:41 +01:00
struct pvfs_state *pvfs = talloc_get_type(ntvfs->private_data,
struct pvfs_state);
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
struct pvfs_search_state *search;
Change uint_t to unsigned int in source4 Signed-off-by: Stefan Metzmacher <metze@samba.org>
2010-01-05 09:42:54 -08:00
unsigned int reply_count;
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
uint16_t search_attrib;
const char *pattern;
NTSTATUS status;
struct pvfs_filename *name;
r3056: added a id -> pointer data structure (a type of radix tree). This is an extremely efficient way of mapping from an integer handle (such as an open file handle) to a pointer (such as the structure containing the open file information). The code is taken from lib/idr.c in the 2.6 Linux kernel, and is very fast and space efficient. By using talloc it even has auto cleanup. This commit converts the handling of open file handles and open directory search handles to use the idtree routines. In combination with talloc destructors, this simplifies the structure handling in the pvfs backend a lot. For example, we no longer need to keep a linked list of open directory searches at all, and we no longer need to do linear scans of the list of open files on most operations. The end result is that the pvfs code is now extremely scalable. You can have 10s of thousands of open files and open searches and the code still runs very fast. I have also added a small optimisation into the file close path, to avoid looking in the byte range locking database if we know that there are no locks outstanding. (This used to be commit 16835a0ef91a16fa01145b773aad8d43da215dbf)
2004-10-19 06:39:51 +00:00
int id;
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
search_attrib = io->search_first.in.search_attrib;
pattern = io->search_first.in.pattern;
/* resolve the cifs name to a posix name */
r3747: - added some of the infrastructure needed for streams support in pvfs (the IDL, and the load/save meta-data logic) - changed pvfs_resolve_name() to default to non-wildcard, needing PVFS_RESOLVE_WILDCARD to enable wildcards. Most callers don't want wildcards, so defaulting this way makes more sense. - fixed deletion of EAs (This used to be commit e7afd4403cc1b7e0928776929f8988aa6f15640b)
2004-11-15 06:57:26 +00:00
status = pvfs_resolve_name(pvfs, req, pattern, PVFS_RESOLVE_WILDCARD, &name);
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
if (!NT_STATUS_IS_OK(status)) {
return status;
}
if (!name->has_wildcard && !name->exists) {
return STATUS_NO_MORE_FILES;
}
r4615: added acl checking on directory search in pvfs (This used to be commit 0e61a422bd9a1596a284c176f033e958bbeaa8ce)
2005-01-09 08:27:35 +00:00
status = pvfs_access_check_parent(pvfs, req, name, SEC_DIR_TRAVERSE | SEC_DIR_LIST);
if (!NT_STATUS_IS_OK(status)) {
return status;
}
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
/* we initially make search a child of the request, then if we
need to keep it long term we steal it for the private
structure */
r5037: got rid of all of the TALLOC_DEPRECATED stuff. My apologies for the large commit. I thought this was worthwhile to get done for consistency. (This used to be commit ec32b22ed5ec224f6324f5e069d15e92e38e15c0)
2005-01-27 07:08:20 +00:00
search = talloc(req, struct pvfs_search_state);
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
if (!search) {
return NT_STATUS_NO_MEMORY;
}
/* do the actual directory listing */
r3262: - new pvfs_dirlist code that reopens the directory between search calls. This is needed to allow for "new files appear during a search" behaviour - pvfs now passes RAW-SEARCH (This used to be commit 0d98f7653a1d58510a6cd4c2ac6c5e05c541109c)
2004-10-26 13:18:34 +00:00
status = pvfs_list_start(pvfs, name, search, &dir);
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
if (!NT_STATUS_IS_OK(status)) {
return status;
}
/* we need to give a handle back to the client so it
can continue a search */
r14540: fix a talloc hierachie problem, make sure file and search handles are cleaned up before anything else in the pvfs_state struct, as there destructors reply on a valid pvfs_state struct metze (This used to be commit aaa5d377b9b6145a83c0e686c7fbb7b561ae8988)
2006-03-18 10:38:38 +00:00
id = idr_get_new(pvfs->search.idtree, search, MAX_OLD_SEARCHES);
r3276: - allow for more than 256 open old style searches (limit currently set at an arbitrary 5000) - auto-cleanup old searches that the client forgot to close (common with old searches) - expanded the RAW-SEARCH test to test more than 256 old searches, and old search rewind (w2k3 fails this - it appears to not support rewind on old style searches) (This used to be commit bc83d823b2140a10007490bf0101843a886f99a6)
2004-10-27 01:11:44 +00:00
if (id == -1) {
pvfs_search_cleanup(pvfs);
r14540: fix a talloc hierachie problem, make sure file and search handles are cleaned up before anything else in the pvfs_state struct, as there destructors reply on a valid pvfs_state struct metze (This used to be commit aaa5d377b9b6145a83c0e686c7fbb7b561ae8988)
2006-03-18 10:38:38 +00:00
id = idr_get_new(pvfs->search.idtree, search, MAX_OLD_SEARCHES);
r3276: - allow for more than 256 open old style searches (limit currently set at an arbitrary 5000) - auto-cleanup old searches that the client forgot to close (common with old searches) - expanded the RAW-SEARCH test to test more than 256 old searches, and old search rewind (w2k3 fails this - it appears to not support rewind on old style searches) (This used to be commit bc83d823b2140a10007490bf0101843a886f99a6)
2004-10-27 01:11:44 +00:00
}
r3056: added a id -> pointer data structure (a type of radix tree). This is an extremely efficient way of mapping from an integer handle (such as an open file handle) to a pointer (such as the structure containing the open file information). The code is taken from lib/idr.c in the 2.6 Linux kernel, and is very fast and space efficient. By using talloc it even has auto cleanup. This commit converts the handling of open file handles and open directory search handles to use the idtree routines. In combination with talloc destructors, this simplifies the structure handling in the pvfs backend a lot. For example, we no longer need to keep a linked list of open directory searches at all, and we no longer need to do linear scans of the list of open files on most operations. The end result is that the pvfs code is now extremely scalable. You can have 10s of thousands of open files and open searches and the code still runs very fast. I have also added a small optimisation into the file close path, to avoid looking in the byte range locking database if we know that there are no locks outstanding. (This used to be commit 16835a0ef91a16fa01145b773aad8d43da215dbf)
2004-10-19 06:39:51 +00:00
if (id == -1) {
return NT_STATUS_INSUFFICIENT_RESOURCES;
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
}
r3056: added a id -> pointer data structure (a type of radix tree). This is an extremely efficient way of mapping from an integer handle (such as an open file handle) to a pointer (such as the structure containing the open file information). The code is taken from lib/idr.c in the 2.6 Linux kernel, and is very fast and space efficient. By using talloc it even has auto cleanup. This commit converts the handling of open file handles and open directory search handles to use the idtree routines. In combination with talloc destructors, this simplifies the structure handling in the pvfs backend a lot. For example, we no longer need to keep a linked list of open directory searches at all, and we no longer need to do linear scans of the list of open files on most operations. The end result is that the pvfs code is now extremely scalable. You can have 10s of thousands of open files and open searches and the code still runs very fast. I have also added a small optimisation into the file close path, to avoid looking in the byte range locking database if we know that there are no locks outstanding. (This used to be commit 16835a0ef91a16fa01145b773aad8d43da215dbf)
2004-10-19 06:39:51 +00:00
search->pvfs = pvfs;
search->handle = id;
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
search->dir = dir;
search->current_index = 0;
r3207: - reformat error msgs in BASE-DIR* tests - added support for mandatory attributes in old style directory search - we now pass BASE-DIR1 and BASE-DIR2 (This used to be commit efaa0eaec49e952aa515c3448246d9048a484c26)
2004-10-25 07:58:47 +00:00
search->search_attrib = search_attrib & 0xFF;
search->must_attrib = (search_attrib>>8) & 0xFF;
s4/pvfs: use monotonic time for this timeout
2010-09-10 20:46:10 +02:00
search->last_used = time_mono(NULL);
r8535: no longer rely on seekdir working after a closedir. Instead, keep directories open, but close search states based on an inactivity timer, with a default of a 5 minute timeout (This used to be commit 2e8d154e7dfb9b320a1344e957a39e96e1eefadd)
2005-07-18 03:35:52 +00:00
search->te = NULL;
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
r14540: fix a talloc hierachie problem, make sure file and search handles are cleaned up before anything else in the pvfs_state struct, as there destructors reply on a valid pvfs_state struct metze (This used to be commit aaa5d377b9b6145a83c0e686c7fbb7b561ae8988)
2006-03-18 10:38:38 +00:00
DLIST_ADD(pvfs->search.list, search);
r3056: added a id -> pointer data structure (a type of radix tree). This is an extremely efficient way of mapping from an integer handle (such as an open file handle) to a pointer (such as the structure containing the open file information). The code is taken from lib/idr.c in the 2.6 Linux kernel, and is very fast and space efficient. By using talloc it even has auto cleanup. This commit converts the handling of open file handles and open directory search handles to use the idtree routines. In combination with talloc destructors, this simplifies the structure handling in the pvfs backend a lot. For example, we no longer need to keep a linked list of open directory searches at all, and we no longer need to do linear scans of the list of open files on most operations. The end result is that the pvfs code is now extremely scalable. You can have 10s of thousands of open files and open searches and the code still runs very fast. I have also added a small optimisation into the file close path, to avoid looking in the byte range locking database if we know that there are no locks outstanding. (This used to be commit 16835a0ef91a16fa01145b773aad8d43da215dbf)
2004-10-19 06:39:51 +00:00
talloc_set_destructor(search, pvfs_search_destructor);
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
status = pvfs_search_fill(pvfs, req, io->search_first.in.max_count, search, io->generic.data_level,
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
&reply_count, search_private, callback);
if (!NT_STATUS_IS_OK(status)) {
return status;
}
io->search_first.out.count = reply_count;
/* not matching any entries is an error */
if (reply_count == 0) {
return STATUS_NO_MORE_FILES;
}
talloc_steal(pvfs, search);
return NT_STATUS_OK;
}
/* continue a old style search */
r2751: this is a new ntvfs design which tries to solve: - the stacking of modules - finding the modules private data - hide the ntvfs details from the calling layer - I set NTVFS_INTERFACE_VERSION 0 till we are closer to release (because we need to solve some async problems with the module stacking) metze (This used to be commit 3ff03b5cb21bb79afdd3b1609be9635f6688a539)
2004-09-29 13:17:09 +00:00
static NTSTATUS pvfs_search_next_old(struct ntvfs_module_context *ntvfs,
r14157: - pass a struct ntvfs_request to the ntvfs layer (for now we just do #define ntvfs_request smbsrv_request, but it's the first step...) - rename ntvfs_openfile() -> ntvfs_open() - fix the talloc hierachie in some places in the ntvfs_map_*() code metze (This used to be commit ed9ed1f48f602354810937c0b0de850b44322191)
2006-03-10 14:31:17 +00:00
struct ntvfs_request *req, union smb_search_next *io,
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
void *search_private,
r25554: Convert last instances of BOOL, True and False to the standard types. (This used to be commit 566aa14139510788548a874e9213d91317f83ca9)
2007-10-06 22:28:14 +00:00
bool (*callback)(void *, const union smb_search_data *))
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
{
s4:pvfs: use talloc_get_type() to cast from void * metze
2009-02-04 08:52:41 +01:00
struct pvfs_state *pvfs = talloc_get_type(ntvfs->private_data,
struct pvfs_state);
void *p;
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
struct pvfs_search_state *search;
struct pvfs_dir *dir;
Change uint_t to unsigned int in source4 Signed-off-by: Stefan Metzmacher <metze@samba.org>
2010-01-05 09:42:54 -08:00
unsigned int reply_count, max_count;
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
uint16_t handle;
NTSTATUS status;
r3276: - allow for more than 256 open old style searches (limit currently set at an arbitrary 5000) - auto-cleanup old searches that the client forgot to close (common with old searches) - expanded the RAW-SEARCH test to test more than 256 old searches, and old search rewind (w2k3 fails this - it appears to not support rewind on old style searches) (This used to be commit bc83d823b2140a10007490bf0101843a886f99a6)
2004-10-27 01:11:44 +00:00
handle = io->search_next.in.id.handle | (io->search_next.in.id.reserved<<8);
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
max_count = io->search_next.in.max_count;
s4:pvfs: use talloc_get_type() to cast from void * metze
2009-02-04 08:52:41 +01:00
p = idr_find(pvfs->search.idtree, handle);
if (p == NULL) {
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
/* we didn't find the search handle */
return NT_STATUS_INVALID_HANDLE;
}
s4:pvfs: use talloc_get_type() to cast from void * metze
2009-02-04 08:52:41 +01:00
search = talloc_get_type(p, struct pvfs_search_state);
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
dir = search->dir;
r18319: fixed the directory search resume code on IRIX The problem was twofold: 1) irix returns 64 bit numbers in telldir(). The protocol uses a 32 bit resume key. We now cope with this properly using the code in pvfs_list_seek_ofs(). 2) irix returns 0xFFFFFFFF from telldir() for the last entry in the directory. When added to DIR_OFFSET_BASE this became DIR_OFFSET_DOTDOT which meant an infinite loop! (This used to be commit 8cce9740ed0da9f08d6821beb4acaa9d28d149c2)
2006-09-10 07:24:41 +00:00
status = pvfs_list_seek_ofs(dir, io->search_next.in.id.server_cookie,
&search->current_index);
if (!NT_STATUS_IS_OK(status)) {
return status;
}
s4/pvfs: use monotonic time for this timeout
2010-09-10 20:46:10 +02:00
search->last_used = time_mono(NULL);
r18319: fixed the directory search resume code on IRIX The problem was twofold: 1) irix returns 64 bit numbers in telldir(). The protocol uses a 32 bit resume key. We now cope with this properly using the code in pvfs_list_seek_ofs(). 2) irix returns 0xFFFFFFFF from telldir() for the last entry in the directory. When added to DIR_OFFSET_BASE this became DIR_OFFSET_DOTDOT which meant an infinite loop! (This used to be commit 8cce9740ed0da9f08d6821beb4acaa9d28d149c2)
2006-09-10 07:24:41 +00:00
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
status = pvfs_search_fill(pvfs, req, max_count, search, io->generic.data_level,
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
&reply_count, search_private, callback);
if (!NT_STATUS_IS_OK(status)) {
return status;
}
io->search_next.out.count = reply_count;
/* not matching any entries means end of search */
if (reply_count == 0) {
talloc_free(search);
}
return NT_STATUS_OK;
}
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
/*
list files in a directory matching a wildcard pattern
*/
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
static NTSTATUS pvfs_search_first_trans2(struct ntvfs_module_context *ntvfs,
struct ntvfs_request *req, union smb_search_first *io,
void *search_private,
r25554: Convert last instances of BOOL, True and False to the standard types. (This used to be commit 566aa14139510788548a874e9213d91317f83ca9)
2007-10-06 22:28:14 +00:00
bool (*callback)(void *, const union smb_search_data *))
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
{
struct pvfs_dir *dir;
s4:pvfs: use talloc_get_type() to cast from void * metze
2009-02-04 08:52:41 +01:00
struct pvfs_state *pvfs = talloc_get_type(ntvfs->private_data,
struct pvfs_state);
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
struct pvfs_search_state *search;
Change uint_t to unsigned int in source4 Signed-off-by: Stefan Metzmacher <metze@samba.org>
2010-01-05 09:42:54 -08:00
unsigned int reply_count;
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
uint16_t search_attrib, max_count;
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
const char *pattern;
NTSTATUS status;
struct pvfs_filename *name;
r3056: added a id -> pointer data structure (a type of radix tree). This is an extremely efficient way of mapping from an integer handle (such as an open file handle) to a pointer (such as the structure containing the open file information). The code is taken from lib/idr.c in the 2.6 Linux kernel, and is very fast and space efficient. By using talloc it even has auto cleanup. This commit converts the handling of open file handles and open directory search handles to use the idtree routines. In combination with talloc destructors, this simplifies the structure handling in the pvfs backend a lot. For example, we no longer need to keep a linked list of open directory searches at all, and we no longer need to do linear scans of the list of open files on most operations. The end result is that the pvfs code is now extremely scalable. You can have 10s of thousands of open files and open searches and the code still runs very fast. I have also added a small optimisation into the file close path, to avoid looking in the byte range locking database if we know that there are no locks outstanding. (This used to be commit 16835a0ef91a16fa01145b773aad8d43da215dbf)
2004-10-19 06:39:51 +00:00
int id;
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
search_attrib = io->t2ffirst.in.search_attrib;
pattern = io->t2ffirst.in.pattern;
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
max_count = io->t2ffirst.in.max_count;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
/* resolve the cifs name to a posix name */
r3747: - added some of the infrastructure needed for streams support in pvfs (the IDL, and the load/save meta-data logic) - changed pvfs_resolve_name() to default to non-wildcard, needing PVFS_RESOLVE_WILDCARD to enable wildcards. Most callers don't want wildcards, so defaulting this way makes more sense. - fixed deletion of EAs (This used to be commit e7afd4403cc1b7e0928776929f8988aa6f15640b)
2004-11-15 06:57:26 +00:00
status = pvfs_resolve_name(pvfs, req, pattern, PVFS_RESOLVE_WILDCARD, &name);
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
if (!NT_STATUS_IS_OK(status)) {
return status;
}
if (!name->has_wildcard && !name->exists) {
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
return NT_STATUS_NO_SUCH_FILE;
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
}
r4615: added acl checking on directory search in pvfs (This used to be commit 0e61a422bd9a1596a284c176f033e958bbeaa8ce)
2005-01-09 08:27:35 +00:00
status = pvfs_access_check_parent(pvfs, req, name, SEC_DIR_TRAVERSE | SEC_DIR_LIST);
if (!NT_STATUS_IS_OK(status)) {
return status;
}
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
/* we initially make search a child of the request, then if we
need to keep it long term we steal it for the private
structure */
r5037: got rid of all of the TALLOC_DEPRECATED stuff. My apologies for the large commit. I thought this was worthwhile to get done for consistency. (This used to be commit ec32b22ed5ec224f6324f5e069d15e92e38e15c0)
2005-01-27 07:08:20 +00:00
search = talloc(req, struct pvfs_search_state);
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
if (!search) {
return NT_STATUS_NO_MEMORY;
}
/* do the actual directory listing */
r3262: - new pvfs_dirlist code that reopens the directory between search calls. This is needed to allow for "new files appear during a search" behaviour - pvfs now passes RAW-SEARCH (This used to be commit 0d98f7653a1d58510a6cd4c2ac6c5e05c541109c)
2004-10-26 13:18:34 +00:00
status = pvfs_list_start(pvfs, name, search, &dir);
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
if (!NT_STATUS_IS_OK(status)) {
return status;
}
r17752: using -1 in a uint16_t value causes compiler warnings, use UINT16_MAX as invalid search handle metze (This used to be commit a4e19d6cc4cce611241c502de485dbd2496be0d3)
2006-08-23 13:06:31 +00:00
id = idr_get_new(pvfs->search.idtree, search, MAX_SEARCH_HANDLES);
r3056: added a id -> pointer data structure (a type of radix tree). This is an extremely efficient way of mapping from an integer handle (such as an open file handle) to a pointer (such as the structure containing the open file information). The code is taken from lib/idr.c in the 2.6 Linux kernel, and is very fast and space efficient. By using talloc it even has auto cleanup. This commit converts the handling of open file handles and open directory search handles to use the idtree routines. In combination with talloc destructors, this simplifies the structure handling in the pvfs backend a lot. For example, we no longer need to keep a linked list of open directory searches at all, and we no longer need to do linear scans of the list of open files on most operations. The end result is that the pvfs code is now extremely scalable. You can have 10s of thousands of open files and open searches and the code still runs very fast. I have also added a small optimisation into the file close path, to avoid looking in the byte range locking database if we know that there are no locks outstanding. (This used to be commit 16835a0ef91a16fa01145b773aad8d43da215dbf)
2004-10-19 06:39:51 +00:00
if (id == -1) {
return NT_STATUS_INSUFFICIENT_RESOURCES;
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
}
r3056: added a id -> pointer data structure (a type of radix tree). This is an extremely efficient way of mapping from an integer handle (such as an open file handle) to a pointer (such as the structure containing the open file information). The code is taken from lib/idr.c in the 2.6 Linux kernel, and is very fast and space efficient. By using talloc it even has auto cleanup. This commit converts the handling of open file handles and open directory search handles to use the idtree routines. In combination with talloc destructors, this simplifies the structure handling in the pvfs backend a lot. For example, we no longer need to keep a linked list of open directory searches at all, and we no longer need to do linear scans of the list of open files on most operations. The end result is that the pvfs code is now extremely scalable. You can have 10s of thousands of open files and open searches and the code still runs very fast. I have also added a small optimisation into the file close path, to avoid looking in the byte range locking database if we know that there are no locks outstanding. (This used to be commit 16835a0ef91a16fa01145b773aad8d43da215dbf)
2004-10-19 06:39:51 +00:00
search->pvfs = pvfs;
search->handle = id;
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
search->dir = dir;
search->current_index = 0;
search->search_attrib = search_attrib;
r3207: - reformat error msgs in BASE-DIR* tests - added support for mandatory attributes in old style directory search - we now pass BASE-DIR1 and BASE-DIR2 (This used to be commit efaa0eaec49e952aa515c3448246d9048a484c26)
2004-10-25 07:58:47 +00:00
search->must_attrib = 0;
r3276: - allow for more than 256 open old style searches (limit currently set at an arbitrary 5000) - auto-cleanup old searches that the client forgot to close (common with old searches) - expanded the RAW-SEARCH test to test more than 256 old searches, and old search rewind (w2k3 fails this - it appears to not support rewind on old style searches) (This used to be commit bc83d823b2140a10007490bf0101843a886f99a6)
2004-10-27 01:11:44 +00:00
search->last_used = 0;
r4263: added support for the trans2 RAW_SEARCH_EA_LIST information level. This is quite a strange level that we've never seen before, but is used by the os2 workplace shell. note w2k screws up this level when unicode is negotiated, so it only passes the RAW-SEARCH test when you force non-unicode (This used to be commit 25189b8fbf6515d573e3398dc9fca56505dc37b9)
2004-12-18 04:38:43 +00:00
search->num_ea_names = io->t2ffirst.in.num_names;
search->ea_names = io->t2ffirst.in.ea_names;
r8535: no longer rely on seekdir working after a closedir. Instead, keep directories open, but close search states based on an inactivity timer, with a default of a 5 minute timeout (This used to be commit 2e8d154e7dfb9b320a1344e957a39e96e1eefadd)
2005-07-18 03:35:52 +00:00
search->te = NULL;
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
r14540: fix a talloc hierachie problem, make sure file and search handles are cleaned up before anything else in the pvfs_state struct, as there destructors reply on a valid pvfs_state struct metze (This used to be commit aaa5d377b9b6145a83c0e686c7fbb7b561ae8988)
2006-03-18 10:38:38 +00:00
DLIST_ADD(pvfs->search.list, search);
r3056: added a id -> pointer data structure (a type of radix tree). This is an extremely efficient way of mapping from an integer handle (such as an open file handle) to a pointer (such as the structure containing the open file information). The code is taken from lib/idr.c in the 2.6 Linux kernel, and is very fast and space efficient. By using talloc it even has auto cleanup. This commit converts the handling of open file handles and open directory search handles to use the idtree routines. In combination with talloc destructors, this simplifies the structure handling in the pvfs backend a lot. For example, we no longer need to keep a linked list of open directory searches at all, and we no longer need to do linear scans of the list of open files on most operations. The end result is that the pvfs code is now extremely scalable. You can have 10s of thousands of open files and open searches and the code still runs very fast. I have also added a small optimisation into the file close path, to avoid looking in the byte range locking database if we know that there are no locks outstanding. (This used to be commit 16835a0ef91a16fa01145b773aad8d43da215dbf)
2004-10-19 06:39:51 +00:00
talloc_set_destructor(search, pvfs_search_destructor);
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
status = pvfs_search_fill(pvfs, req, max_count, search, io->generic.data_level,
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
&reply_count, search_private, callback);
if (!NT_STATUS_IS_OK(status)) {
return status;
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
}
/* not matching any entries is an error */
if (reply_count == 0) {
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
return NT_STATUS_NO_SUCH_FILE;
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
}
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
io->t2ffirst.out.count = reply_count;
io->t2ffirst.out.handle = search->handle;
r3260: redid the pvfs_dirlist() interface in preparation for a "keep directory open" implementation, as opposed to the "load the whole directory" interface used now. This will be needed to pass RAW-SEARCH (This used to be commit 692623c6c0a2c6817fddfa77cd1c2525c27145c4)
2004-10-26 11:11:16 +00:00
io->t2ffirst.out.end_of_search = pvfs_list_eos(dir, search->current_index) ? 1 : 0;
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
/* work out if we are going to keep the search state
and allow for a search continue */
if ((io->t2ffirst.in.flags & FLAG_TRANS2_FIND_CLOSE) ||
((io->t2ffirst.in.flags & FLAG_TRANS2_FIND_CLOSE_IF_END) &&
io->t2ffirst.out.end_of_search)) {
talloc_free(search);
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
} else {
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
talloc_steal(pvfs, search);
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
}
return NT_STATUS_OK;
}
/* continue a search */
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
static NTSTATUS pvfs_search_next_trans2(struct ntvfs_module_context *ntvfs,
struct ntvfs_request *req, union smb_search_next *io,
void *search_private,
r25554: Convert last instances of BOOL, True and False to the standard types. (This used to be commit 566aa14139510788548a874e9213d91317f83ca9)
2007-10-06 22:28:14 +00:00
bool (*callback)(void *, const union smb_search_data *))
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
{
s4:pvfs: use talloc_get_type() to cast from void * metze
2009-02-04 08:52:41 +01:00
struct pvfs_state *pvfs = talloc_get_type(ntvfs->private_data,
struct pvfs_state);
void *p;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
struct pvfs_search_state *search;
struct pvfs_dir *dir;
Change uint_t to unsigned int in source4 Signed-off-by: Stefan Metzmacher <metze@samba.org>
2010-01-05 09:42:54 -08:00
unsigned int reply_count;
r2469: complete overhaul of the old-style RAW_SEARCH_ calls (the OS/2 and original core level calls). The old code was completely wrong in many respects. also fixed the EA_SIZE level in the server extended the RAW-SEARCH test suite to test the new code properly (This used to be commit 71480271ad84b57fcdde264a54bb2408cf783255)
2004-09-21 08:46:47 +00:00
uint16_t handle;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
NTSTATUS status;
r2469: complete overhaul of the old-style RAW_SEARCH_ calls (the OS/2 and original core level calls). The old code was completely wrong in many respects. also fixed the EA_SIZE level in the server extended the RAW-SEARCH test suite to test the new code properly (This used to be commit 71480271ad84b57fcdde264a54bb2408cf783255)
2004-09-21 08:46:47 +00:00
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
handle = io->t2fnext.in.handle;
r2469: complete overhaul of the old-style RAW_SEARCH_ calls (the OS/2 and original core level calls). The old code was completely wrong in many respects. also fixed the EA_SIZE level in the server extended the RAW-SEARCH test suite to test the new code properly (This used to be commit 71480271ad84b57fcdde264a54bb2408cf783255)
2004-09-21 08:46:47 +00:00
s4:pvfs: use talloc_get_type() to cast from void * metze
2009-02-04 08:52:41 +01:00
p = idr_find(pvfs->search.idtree, handle);
if (p == NULL) {
r2469: complete overhaul of the old-style RAW_SEARCH_ calls (the OS/2 and original core level calls). The old code was completely wrong in many respects. also fixed the EA_SIZE level in the server extended the RAW-SEARCH test suite to test the new code properly (This used to be commit 71480271ad84b57fcdde264a54bb2408cf783255)
2004-09-21 08:46:47 +00:00
/* we didn't find the search handle */
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
return NT_STATUS_INVALID_HANDLE;
r2469: complete overhaul of the old-style RAW_SEARCH_ calls (the OS/2 and original core level calls). The old code was completely wrong in many respects. also fixed the EA_SIZE level in the server extended the RAW-SEARCH test suite to test the new code properly (This used to be commit 71480271ad84b57fcdde264a54bb2408cf783255)
2004-09-21 08:46:47 +00:00
}
s4:pvfs: use talloc_get_type() to cast from void * metze
2009-02-04 08:52:41 +01:00
search = talloc_get_type(p, struct pvfs_search_state);
r2469: complete overhaul of the old-style RAW_SEARCH_ calls (the OS/2 and original core level calls). The old code was completely wrong in many respects. also fixed the EA_SIZE level in the server extended the RAW-SEARCH test suite to test the new code properly (This used to be commit 71480271ad84b57fcdde264a54bb2408cf783255)
2004-09-21 08:46:47 +00:00
dir = search->dir;
r18319: fixed the directory search resume code on IRIX The problem was twofold: 1) irix returns 64 bit numbers in telldir(). The protocol uses a 32 bit resume key. We now cope with this properly using the code in pvfs_list_seek_ofs(). 2) irix returns 0xFFFFFFFF from telldir() for the last entry in the directory. When added to DIR_OFFSET_BASE this became DIR_OFFSET_DOTDOT which meant an infinite loop! (This used to be commit 8cce9740ed0da9f08d6821beb4acaa9d28d149c2)
2006-09-10 07:24:41 +00:00
status = NT_STATUS_OK;
r2469: complete overhaul of the old-style RAW_SEARCH_ calls (the OS/2 and original core level calls). The old code was completely wrong in many respects. also fixed the EA_SIZE level in the server extended the RAW-SEARCH test suite to test the new code properly (This used to be commit 71480271ad84b57fcdde264a54bb2408cf783255)
2004-09-21 08:46:47 +00:00
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
/* work out what type of continuation is being used */
if (io->t2fnext.in.last_name && *io->t2fnext.in.last_name) {
r3262: - new pvfs_dirlist code that reopens the directory between search calls. This is needed to allow for "new files appear during a search" behaviour - pvfs now passes RAW-SEARCH (This used to be commit 0d98f7653a1d58510a6cd4c2ac6c5e05c541109c)
2004-10-26 13:18:34 +00:00
status = pvfs_list_seek(dir, io->t2fnext.in.last_name, &search->current_index);
r18319: fixed the directory search resume code on IRIX The problem was twofold: 1) irix returns 64 bit numbers in telldir(). The protocol uses a 32 bit resume key. We now cope with this properly using the code in pvfs_list_seek_ofs(). 2) irix returns 0xFFFFFFFF from telldir() for the last entry in the directory. When added to DIR_OFFSET_BASE this became DIR_OFFSET_DOTDOT which meant an infinite loop! (This used to be commit 8cce9740ed0da9f08d6821beb4acaa9d28d149c2)
2006-09-10 07:24:41 +00:00
if (!NT_STATUS_IS_OK(status) && io->t2fnext.in.resume_key) {
status = pvfs_list_seek_ofs(dir, io->t2fnext.in.resume_key,
&search->current_index);
r3262: - new pvfs_dirlist code that reopens the directory between search calls. This is needed to allow for "new files appear during a search" behaviour - pvfs now passes RAW-SEARCH (This used to be commit 0d98f7653a1d58510a6cd4c2ac6c5e05c541109c)
2004-10-26 13:18:34 +00:00
}
r18319: fixed the directory search resume code on IRIX The problem was twofold: 1) irix returns 64 bit numbers in telldir(). The protocol uses a 32 bit resume key. We now cope with this properly using the code in pvfs_list_seek_ofs(). 2) irix returns 0xFFFFFFFF from telldir() for the last entry in the directory. When added to DIR_OFFSET_BASE this became DIR_OFFSET_DOTDOT which meant an infinite loop! (This used to be commit 8cce9740ed0da9f08d6821beb4acaa9d28d149c2)
2006-09-10 07:24:41 +00:00
} else if (!(io->t2fnext.in.flags & FLAG_TRANS2_FIND_CONTINUE)) {
status = pvfs_list_seek_ofs(dir, io->t2fnext.in.resume_key,
&search->current_index);
}
if (!NT_STATUS_IS_OK(status)) {
return status;
r2469: complete overhaul of the old-style RAW_SEARCH_ calls (the OS/2 and original core level calls). The old code was completely wrong in many respects. also fixed the EA_SIZE level in the server extended the RAW-SEARCH test suite to test the new code properly (This used to be commit 71480271ad84b57fcdde264a54bb2408cf783255)
2004-09-21 08:46:47 +00:00
}
r4263: added support for the trans2 RAW_SEARCH_EA_LIST information level. This is quite a strange level that we've never seen before, but is used by the os2 workplace shell. note w2k screws up this level when unicode is negotiated, so it only passes the RAW-SEARCH test when you force non-unicode (This used to be commit 25189b8fbf6515d573e3398dc9fca56505dc37b9)
2004-12-18 04:38:43 +00:00
search->num_ea_names = io->t2fnext.in.num_names;
search->ea_names = io->t2fnext.in.ea_names;
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
status = pvfs_search_fill(pvfs, req, io->t2fnext.in.max_count, search, io->generic.data_level,
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
&reply_count, search_private, callback);
if (!NT_STATUS_IS_OK(status)) {
return status;
r2469: complete overhaul of the old-style RAW_SEARCH_ calls (the OS/2 and original core level calls). The old code was completely wrong in many respects. also fixed the EA_SIZE level in the server extended the RAW-SEARCH test suite to test the new code properly (This used to be commit 71480271ad84b57fcdde264a54bb2408cf783255)
2004-09-21 08:46:47 +00:00
}
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
io->t2fnext.out.count = reply_count;
r3260: redid the pvfs_dirlist() interface in preparation for a "keep directory open" implementation, as opposed to the "load the whole directory" interface used now. This will be needed to pass RAW-SEARCH (This used to be commit 692623c6c0a2c6817fddfa77cd1c2525c27145c4)
2004-10-26 11:11:16 +00:00
io->t2fnext.out.end_of_search = pvfs_list_eos(dir, search->current_index) ? 1 : 0;
r2469: complete overhaul of the old-style RAW_SEARCH_ calls (the OS/2 and original core level calls). The old code was completely wrong in many respects. also fixed the EA_SIZE level in the server extended the RAW-SEARCH test suite to test the new code properly (This used to be commit 71480271ad84b57fcdde264a54bb2408cf783255)
2004-09-21 08:46:47 +00:00
/* work out if we are going to keep the search state */
if ((io->t2fnext.in.flags & FLAG_TRANS2_FIND_CLOSE) ||
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
((io->t2fnext.in.flags & FLAG_TRANS2_FIND_CLOSE_IF_END) &&
io->t2fnext.out.end_of_search)) {
r2469: complete overhaul of the old-style RAW_SEARCH_ calls (the OS/2 and original core level calls). The old code was completely wrong in many respects. also fixed the EA_SIZE level in the server extended the RAW-SEARCH test suite to test the new code properly (This used to be commit 71480271ad84b57fcdde264a54bb2408cf783255)
2004-09-21 08:46:47 +00:00
talloc_free(search);
}
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
r2469: complete overhaul of the old-style RAW_SEARCH_ calls (the OS/2 and original core level calls). The old code was completely wrong in many respects. also fixed the EA_SIZE level in the server extended the RAW-SEARCH test suite to test the new code properly (This used to be commit 71480271ad84b57fcdde264a54bb2408cf783255)
2004-09-21 08:46:47 +00:00
return NT_STATUS_OK;
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
}
r16876: implement SMB2 Find in the posix ntvfs backend metze (This used to be commit 4f1afda488f2fb5cfcf98ef6a56157f954fdccfc)
2006-07-08 12:35:37 +00:00
static NTSTATUS pvfs_search_first_smb2(struct ntvfs_module_context *ntvfs,
struct ntvfs_request *req, const struct smb2_find *io,
void *search_private,
r25554: Convert last instances of BOOL, True and False to the standard types. (This used to be commit 566aa14139510788548a874e9213d91317f83ca9)
2007-10-06 22:28:14 +00:00
bool (*callback)(void *, const union smb_search_data *))
r16876: implement SMB2 Find in the posix ntvfs backend metze (This used to be commit 4f1afda488f2fb5cfcf98ef6a56157f954fdccfc)
2006-07-08 12:35:37 +00:00
{
struct pvfs_dir *dir;
s4:pvfs: use talloc_get_type() to cast from void * metze
2009-02-04 08:52:41 +01:00
struct pvfs_state *pvfs = talloc_get_type(ntvfs->private_data,
struct pvfs_state);
r16876: implement SMB2 Find in the posix ntvfs backend metze (This used to be commit 4f1afda488f2fb5cfcf98ef6a56157f954fdccfc)
2006-07-08 12:35:37 +00:00
struct pvfs_search_state *search;
Change uint_t to unsigned int in source4 Signed-off-by: Stefan Metzmacher <metze@samba.org>
2010-01-05 09:42:54 -08:00
unsigned int reply_count;
r16876: implement SMB2 Find in the posix ntvfs backend metze (This used to be commit 4f1afda488f2fb5cfcf98ef6a56157f954fdccfc)
2006-07-08 12:35:37 +00:00
uint16_t max_count;
const char *pattern;
NTSTATUS status;
struct pvfs_filename *name;
struct pvfs_file *f;
f = pvfs_find_fd(pvfs, req, io->in.file.ntvfs);
if (!f) {
return NT_STATUS_FILE_CLOSED;
}
/* its only valid for directories */
if (f->handle->fd != -1) {
return NT_STATUS_INVALID_PARAMETER;
}
if (!(f->access_mask & SEC_DIR_LIST)) {
return NT_STATUS_ACCESS_DENIED;
}
if (f->search) {
talloc_free(f->search);
f->search = NULL;
}
if (strequal(io->in.pattern, "")) {
return NT_STATUS_OBJECT_NAME_INVALID;
}
if (strchr_m(io->in.pattern, '\\')) {
return NT_STATUS_OBJECT_NAME_INVALID;
}
if (strchr_m(io->in.pattern, '/')) {
return NT_STATUS_OBJECT_NAME_INVALID;
}
if (strequal("", f->handle->name->original_name)) {
s4-smb2: fixed SMB2 find commands The change to check for invalid \ prefix on SMB2 paths broke the internal SMB2 code.
2009-10-18 10:30:10 +11:00
pattern = talloc_asprintf(req, "%s", io->in.pattern);
r16876: implement SMB2 Find in the posix ntvfs backend metze (This used to be commit 4f1afda488f2fb5cfcf98ef6a56157f954fdccfc)
2006-07-08 12:35:37 +00:00
NT_STATUS_HAVE_NO_MEMORY(pattern);
} else {
s4-smb2: fixed SMB2 find commands The change to check for invalid \ prefix on SMB2 paths broke the internal SMB2 code.
2009-10-18 10:30:10 +11:00
pattern = talloc_asprintf(req, "%s\\%s",
r16876: implement SMB2 Find in the posix ntvfs backend metze (This used to be commit 4f1afda488f2fb5cfcf98ef6a56157f954fdccfc)
2006-07-08 12:35:37 +00:00
f->handle->name->original_name,
io->in.pattern);
NT_STATUS_HAVE_NO_MEMORY(pattern);
}
/* resolve the cifs name to a posix name */
status = pvfs_resolve_name(pvfs, req, pattern, PVFS_RESOLVE_WILDCARD, &name);
NT_STATUS_NOT_OK_RETURN(status);
if (!name->has_wildcard && !name->exists) {
return NT_STATUS_NO_SUCH_FILE;
}
/* we initially make search a child of the request, then if we
need to keep it long term we steal it for the private
structure */
search = talloc(req, struct pvfs_search_state);
NT_STATUS_HAVE_NO_MEMORY(search);
/* do the actual directory listing */
status = pvfs_list_start(pvfs, name, search, &dir);
NT_STATUS_NOT_OK_RETURN(status);
search->pvfs = pvfs;
r17752: using -1 in a uint16_t value causes compiler warnings, use UINT16_MAX as invalid search handle metze (This used to be commit a4e19d6cc4cce611241c502de485dbd2496be0d3)
2006-08-23 13:06:31 +00:00
search->handle = INVALID_SEARCH_HANDLE;
r16876: implement SMB2 Find in the posix ntvfs backend metze (This used to be commit 4f1afda488f2fb5cfcf98ef6a56157f954fdccfc)
2006-07-08 12:35:37 +00:00
search->dir = dir;
search->current_index = 0;
r22842: fix directory listing of SMB2 and include directories and hidden files metze (This used to be commit e087772ec9ec561b13dbdf6bfead7394a184d7d9)
2007-05-14 11:45:39 +00:00
search->search_attrib = 0x0000FFFF;
r16876: implement SMB2 Find in the posix ntvfs backend metze (This used to be commit 4f1afda488f2fb5cfcf98ef6a56157f954fdccfc)
2006-07-08 12:35:37 +00:00
search->must_attrib = 0;
search->last_used = 0;
search->num_ea_names = 0;
search->ea_names = NULL;
search->te = NULL;
if (io->in.continue_flags & SMB2_CONTINUE_FLAG_SINGLE) {
max_count = 1;
} else {
max_count = UINT16_MAX;
}
status = pvfs_search_fill(pvfs, req, max_count, search, io->data_level,
&reply_count, search_private, callback);
NT_STATUS_NOT_OK_RETURN(status);
/* not matching any entries is an error */
if (reply_count == 0) {
return NT_STATUS_NO_SUCH_FILE;
}
f->search = talloc_steal(f, search);
return NT_STATUS_OK;
}
static NTSTATUS pvfs_search_next_smb2(struct ntvfs_module_context *ntvfs,
struct ntvfs_request *req, const struct smb2_find *io,
void *search_private,
r25554: Convert last instances of BOOL, True and False to the standard types. (This used to be commit 566aa14139510788548a874e9213d91317f83ca9)
2007-10-06 22:28:14 +00:00
bool (*callback)(void *, const union smb_search_data *))
r16876: implement SMB2 Find in the posix ntvfs backend metze (This used to be commit 4f1afda488f2fb5cfcf98ef6a56157f954fdccfc)
2006-07-08 12:35:37 +00:00
{
s4:pvfs: use talloc_get_type() to cast from void * metze
2009-02-04 08:52:41 +01:00
struct pvfs_state *pvfs = talloc_get_type(ntvfs->private_data,
struct pvfs_state);
r16876: implement SMB2 Find in the posix ntvfs backend metze (This used to be commit 4f1afda488f2fb5cfcf98ef6a56157f954fdccfc)
2006-07-08 12:35:37 +00:00
struct pvfs_search_state *search;
Change uint_t to unsigned int in source4 Signed-off-by: Stefan Metzmacher <metze@samba.org>
2010-01-05 09:42:54 -08:00
unsigned int reply_count;
r16876: implement SMB2 Find in the posix ntvfs backend metze (This used to be commit 4f1afda488f2fb5cfcf98ef6a56157f954fdccfc)
2006-07-08 12:35:37 +00:00
uint16_t max_count;
NTSTATUS status;
struct pvfs_file *f;
f = pvfs_find_fd(pvfs, req, io->in.file.ntvfs);
if (!f) {
return NT_STATUS_FILE_CLOSED;
}
/* its only valid for directories */
if (f->handle->fd != -1) {
return NT_STATUS_INVALID_PARAMETER;
}
/* if there's no search started on the dir handle, it's like a search_first */
search = f->search;
if (!search) {
return pvfs_search_first_smb2(ntvfs, req, io, search_private, callback);
}
if (io->in.continue_flags & SMB2_CONTINUE_FLAG_RESTART) {
search->current_index = 0;
}
if (io->in.continue_flags & SMB2_CONTINUE_FLAG_SINGLE) {
max_count = 1;
} else {
max_count = UINT16_MAX;
}
status = pvfs_search_fill(pvfs, req, max_count, search, io->data_level,
&reply_count, search_private, callback);
NT_STATUS_NOT_OK_RETURN(status);
/* not matching any entries is an error */
if (reply_count == 0) {
return STATUS_NO_MORE_FILES;
}
return NT_STATUS_OK;
}
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
/*
list files in a directory matching a wildcard pattern
*/
NTSTATUS pvfs_search_first(struct ntvfs_module_context *ntvfs,
struct ntvfs_request *req, union smb_search_first *io,
void *search_private,
r25554: Convert last instances of BOOL, True and False to the standard types. (This used to be commit 566aa14139510788548a874e9213d91317f83ca9)
2007-10-06 22:28:14 +00:00
bool (*callback)(void *, const union smb_search_data *))
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
{
switch (io->generic.level) {
case RAW_SEARCH_SEARCH:
case RAW_SEARCH_FFIRST:
case RAW_SEARCH_FUNIQUE:
return pvfs_search_first_old(ntvfs, req, io, search_private, callback);
case RAW_SEARCH_TRANS2:
return pvfs_search_first_trans2(ntvfs, req, io, search_private, callback);
r16876: implement SMB2 Find in the posix ntvfs backend metze (This used to be commit 4f1afda488f2fb5cfcf98ef6a56157f954fdccfc)
2006-07-08 12:35:37 +00:00
case RAW_SEARCH_SMB2:
return pvfs_search_first_smb2(ntvfs, req, &io->smb2, search_private, callback);
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
}
return NT_STATUS_INVALID_LEVEL;
}
/* continue a search */
NTSTATUS pvfs_search_next(struct ntvfs_module_context *ntvfs,
struct ntvfs_request *req, union smb_search_next *io,
void *search_private,
r25554: Convert last instances of BOOL, True and False to the standard types. (This used to be commit 566aa14139510788548a874e9213d91317f83ca9)
2007-10-06 22:28:14 +00:00
bool (*callback)(void *, const union smb_search_data *))
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
{
switch (io->generic.level) {
case RAW_SEARCH_SEARCH:
case RAW_SEARCH_FFIRST:
return pvfs_search_next_old(ntvfs, req, io, search_private, callback);
case RAW_SEARCH_FUNIQUE:
return NT_STATUS_INVALID_LEVEL;
case RAW_SEARCH_TRANS2:
return pvfs_search_next_trans2(ntvfs, req, io, search_private, callback);
r16876: implement SMB2 Find in the posix ntvfs backend metze (This used to be commit 4f1afda488f2fb5cfcf98ef6a56157f954fdccfc)
2006-07-08 12:35:37 +00:00
case RAW_SEARCH_SMB2:
return pvfs_search_next_smb2(ntvfs, req, &io->smb2, search_private, callback);
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
}
return NT_STATUS_INVALID_LEVEL;
}
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
/* close a search */
r2751: this is a new ntvfs design which tries to solve: - the stacking of modules - finding the modules private data - hide the ntvfs details from the calling layer - I set NTVFS_INTERFACE_VERSION 0 till we are closer to release (because we need to solve some async problems with the module stacking) metze (This used to be commit 3ff03b5cb21bb79afdd3b1609be9635f6688a539)
2004-09-29 13:17:09 +00:00
NTSTATUS pvfs_search_close(struct ntvfs_module_context *ntvfs,
r14157: - pass a struct ntvfs_request to the ntvfs layer (for now we just do #define ntvfs_request smbsrv_request, but it's the first step...) - rename ntvfs_openfile() -> ntvfs_open() - fix the talloc hierachie in some places in the ntvfs_map_*() code metze (This used to be commit ed9ed1f48f602354810937c0b0de850b44322191)
2006-03-10 14:31:17 +00:00
struct ntvfs_request *req, union smb_search_close *io)
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
{
s4:pvfs: use talloc_get_type() to cast from void * metze
2009-02-04 08:52:41 +01:00
struct pvfs_state *pvfs = talloc_get_type(ntvfs->private_data,
struct pvfs_state);
void *p;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
struct pvfs_search_state *search;
r17752: using -1 in a uint16_t value causes compiler warnings, use UINT16_MAX as invalid search handle metze (This used to be commit a4e19d6cc4cce611241c502de485dbd2496be0d3)
2006-08-23 13:06:31 +00:00
uint16_t handle = INVALID_SEARCH_HANDLE;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
switch (io->generic.level) {
case RAW_FINDCLOSE_GENERIC:
return NT_STATUS_INVALID_LEVEL;
case RAW_FINDCLOSE_FCLOSE:
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
handle = io->fclose.in.id.handle;
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
break;
case RAW_FINDCLOSE_FINDCLOSE:
r2520: - finished implementing the server side of the old style search requests (This used to be commit 4e4859c06b9de5fe60ebd17cfb09eed480b79ec1)
2004-09-22 12:38:19 +00:00
handle = io->findclose.in.handle;
r16834: split the level's of smb_search_first/smb_search_next and the levels of smb_search_data metze (This used to be commit 78c201db8a47a71908698c4dda2add4cf85694d9)
2006-07-06 08:00:24 +00:00
break;
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
}
s4:pvfs: use talloc_get_type() to cast from void * metze
2009-02-04 08:52:41 +01:00
p = idr_find(pvfs->search.idtree, handle);
if (p == NULL) {
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
/* we didn't find the search handle */
return NT_STATUS_INVALID_HANDLE;
}
s4:pvfs: use talloc_get_type() to cast from void * metze
2009-02-04 08:52:41 +01:00
search = talloc_get_type(p, struct pvfs_search_state);
r2503: the RAW-SEARCH test now mostly passes against the posix backend (This used to be commit 9710f24b1fd103d5656c9585cdfed96449cf9f97)
2004-09-22 08:17:26 +00:00
talloc_free(search);
return NT_STATUS_OK;
r2404: the first large lump of posix vfs stuff. this is still very much a skeleton (with many limbs missing too!). I am committing this early to get some feedback on the approach taken. (This used to be commit 40d5cae5ebbfe328e193eadb685df6a370730299)
2004-09-18 08:16:14 +00:00
}
Copy Permalink