IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
when copying to a full disk problem, I discovered that we were not allowing
the delete on close flag to be set properly, this led to other things, and
after investigation of the proper delete on close semantics and their relationship
to the file_share_delete flag I discovered there were some cases where we
weren't doing the deny modes properly. And this after only 5 years working
on them..... :-) :-).
So here's the latest attempt. I realised the delete on close flag needs to
be set across all smbds with a dev/ino pair open - in addition, the delete
on close flag, allow share delete and delete access requested all need to
be stored in the share mode tdb.
The "delete_on_close" entry in the fsp struct is now redundant and should
really be removed. This may also mean we can get rid of the "iterate_fsp"
calls that I didn't like adding in the first place. Whilst doing this patch,
I also discovered we needed to do the se_map_generic() call for file opens
and POSIX ACL mapping, so I added that also.
This code, although ugly, now passes the deny mode torture tests plus the
delete on close tests I added. I do need to add one more multiple connection
delete on close test to make sure I got the semantics exactly right, plus we
should also (as Andrew suggested) move to random testing here.
The good news is that NT should now correctly delete the file on disk
full error when copying to a disk :-).
Jeremy.
(This used to be commit 51987684bd)
a --with-spinlocks option to configure, this does mean the on-disk tdb
format has changed, so 2.2alphaX sites will need to re-create their
tdb's. The upside is no more tdb fragmentation and a +5% on netbench.
Swings and roundabouts....
Jeremy.
(This used to be commit 9dea7b7c25)
a conn struct depending on the call.
We need this to have a clean NT ACL call interface.
This will break any existing VFS libraries (that's why this is pre-release
code).
Andrew gets credit for this one :-) :-).
In addition - added Herb's WITH_PROFILE changes - Herb - please examine
the changes I've made to the smbd/reply.c code you added. The original
code was very ugly and I have replaced it with a
START_PROFILE(x)/END_PROFILE(x) pair using the preprocessor.
Please check this compiles ok with the --with-profile switch.
Jeremy.
(This used to be commit b07611f815)
The motivation for this system is to replace the UDP message for
oplocks, but this commit only does the "set debug level" message.
(This used to be commit 2a34ee95f3)
to find bugs. On 64 bit IRIX, structure packing means that
a
struct {
SMB_DEV_T dev /* 4 bytes */
SMB_INO_T ino /* 8 bytes */
}
has 4 bytes of padding between the two members. If you
don't null the memory before using it as a tdb key,
you randomly can't find keys depending on what is in
the padding. This caused me immense pain and was hard
to track down.... :-)
Jeremy.
(This used to be commit f2a5ba3f09)
to overlay a write lock on the same fnum. When overlaying read locks onto
a write lock, the number of locks is counted, and the first unlock removes
the write lock and downgrades this to a read lock. Do the same when mapping
to POSIX.
Jeremy.
(This used to be commit 74d42644e6)
counting when Windows downgrades a write lock to a read lock, then reference
counts the unlocks to match the locks.
With this code the POSIX unlock isn't done until the final Windows unlock.
Jeremy.
(This used to be commit 6eb4fb6eef)
userdom_struct. As the name implies this also contains a domain
(unused at the moment).
This will be important shortly, as operation in appliance mode needs
the domain to be always carried with the username.
(This used to be commit ee8546342d)
errors etc.) into locking/posix.c, where it is needed. fcntl_lock in lib/util.c
is now very small and clean.
Added (*lock) op to vfs layer.
Jeremy.
(This used to be commit 46092ee141)
call to ms_fnmatch(). This also removes all the Win9X semantics stuff
and a bunch of other associated cruft.
- moved the stat cache code into statcache.c
- fixed the uint16 alignment requirements of ascii_to_unistr() and
unistr_to_ascii()
- trans2 SMB_FIND_FILE_BOTH_DIRECTORY_INFO returns the short name as
unicode always (at least thats what NT4 does)
- fixed some errors in the in-memory tdb code. Still ugly, but doesn't
crash as much
(This used to be commit 03e9cea004)
This implementation keeps all POSIX lock records in a separate in memory
tdb database only known about in locking/posix.c. In addition, the pending
close fd's are also held in a tdb which has an array of fd's indexed by
device and inode.
The walk-split code uglyness has been moved to posix.c from brlock.c,
which is the only place that needs to know about it, and the extra
functions hacked into brlock to expose internal state have been removed.
This implementation passes smbtorture locktest4, the only thing I need
to check now for completeness is what to do about lock upgrade/downgrades
which Win32 allows under some *very* strange circumstances.
Jeremy.
(This used to be commit 3f655de1c7)
fd_close now calls fd_close_posix() directly.
set_posix_lock/release_posix_lock() now handle the reference counting.
More changes due when this gets moved to the file locking/posix.c
Jeremy.
(This used to be commit 239abd48f0)
When a file is being closed, once it passes the fnum and tid tests then
the locking context should be ignored when removing all locks. This is
what is done in the brl close case, but when you have outstanding
POSIX locks, then you cannot remove all the brl locks in one go, you
have to get the lock list and call do_unlock individually. As this
uses global_smbpid as the locking context, you need to make sure
that this is set correctly for the specific lock being removed. I
now do this by storing the smbpid in each entry in the unlock list returned from
the query call. I removed the smbpid from fsp (not needed) and
things seem ok (even with the stupid smbpid tricks that smbtorture plays :-).
Jeremy.
(This used to be commit 6baa96bb46)
smbpid used when a file was opened in the files_struct. Else we use
the wrong global_smbpid when we are closing the file and trying to
remove the brl locks - this causes the brl locks to be left when the
file is closed as the samba_context check fails.
Jeremy.
(This used to be commit 2746e5602e)
removed from the smbd/open.c code.
We now use a dlink list of structures indexed by dev/inode to store
all pending fd's for close. This could be rewritten to use lib/hash.c
if this is discovered to be too slow in use.
Andrew, please take a look and let me know if this is what you
had in mind.
Jeremy.
(This used to be commit 0487841120)
test. Was miscounting posix locks, plus was not taking into account
the case where other_fsp == fsp in the 'move locks' case. DOH ! This
code will be re-written anyway :-).
Jeremy.
(This used to be commit 5278ec016c)
HEAD should now map brl locks correctly into POSIX locks, including the
really nasty case of large range unlock.
There is a lot of pretty ASCII art in locking/brlock.c explaining
exactly how this code works. If it is unclear, please ask me.
Jeremy.
(This used to be commit 135855dbd3)
open on the same dev/inode pair with existing POSIX locks.
This is done at the smbd/open layer, so smbd just calls fd_close() and
the transfer of any open fd's is done under the covers of fd_close().
When an fsp is closed and no other fsp's open on the same dev/inode
pair have existing POSIX locks then all fd's associated with this fsp
are closed.
Now only the hard part of doing the POSIX range unlock code when read
locks overlap remains for full POSIX/SMB lock integration....
Jeremy.
(This used to be commit 1df48ed55e)
new open mechanism Andrew & I discussed.
config.sub:
configure: Included the QNX patch.
include/vfs.h:
smbd/vfs-wrap.c:
smbd/vfs.c: Added ftruncate vfs call (needed).
Note that we will also need locking calls in the vfs (to be added).
lib/util_unistr.c:
nmbd/nmbd_processlogon.c: Fix for NT domain logons causing nmbd to core dump.
Also fix for sidsize DOS bug.
locking/locking.c: Check value of ret before using it for memdup.
printing/printing.c: Convert print_fsp_open to return an allocated fsp.
rpc_server/srv_lsa.c: Fix for NT domain logons.
I have removed all use of lp_share_modes() from the code (although I
left the parameter in the table for backwards compatibility). It no longer makes
sense for this to exist.
smbd/close.c: Removed lp_share_modes().
smbd/fileio.c: Fixed parameters to unlock_share_entry call in panic code.
smbd/files.c: Correctly set the unix_ERR_code to ERRnofids on fsp allocation fail.
smbd/nttrans.c:
smbd/reply.c:
smbd/trans2.c: Changed all occurrences of open_file_shared/open_directory/
open_file_stat to return an fsp from the call.
smbd/open.c: Changed all occurrences of open_file_shared/open_directory/
open_file_stat to return an fsp from the call.
In addition I have fixed a long standing race condition in the deny mode
processing w.r.t. two smbd's creating a file. Andrew, please note that
your original idea of using open with O_EXCL in this case would not work
(I went over the races very carefully) and so we must re-check deny modes
*after* the open() call returns. This is because there is a race between
the open with O_EXCL and the lock of the share mode entry. Imagine the
case where the first smbd does the open with O_EXCL and a deny mode of DENY_ALL,
but is pre-empted before it locks the share modes and creates the deny
mode entry for DENY_ALL. A second smbd could then come in with O_RDONLY
and a deny mode of DENY_NONE and the two opens would be allowed.
The *only* way to fix this race is to lock the share modes after the
open and then do the deny mode checks *after* this lock in the case
where the file did not originally exist.
This code will need extensive testing but seems to initially work.
Jeremy.
(This used to be commit ab0ecc39d6)
