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.
I was storing the mid of the oplock break - I should have been
storing the mid from the open. There are thus 2 types of deferred
packet sequence returns - ones that increment the sequence number
(returns from oplock causing opens) and ones that don't (change notify
returns etc). Running with signing forced on does lead to some
interesting tests :-).
Jeremy.
level 2 and a request for open with no oplock is received then the
smbd should send *synchronous* break messages, not asynchronous,
otherwise it spins very rapidly, releasing the lock, sending the
'break to none' messages and then re-acquiring the lock before
any other process has a chance to get the lock and remove it's own
oplock (at least on linux).
Jeremy
link from Seattle is having problems.
I've added 3 things here to work on the fcntl spin
problem.
1). Check *all* tdb return codes... :-).
2). If we're asking ourselves to break an oplock, and we can't
find a fsp pointer that matches the entry, this is a *logic bug*
and we should abort and panic so someone with gdb can pick up
the pieces.
3). After we've broken an oplock, ensure that the entry itself
has been removed, and if not remove it ourselves. This should
not be neccessary in a correctly working environmen,t, but will
provide an added layer of robustness in error situations.
4). I hate german keyboards :-) :-).
Jeremy.
a byte range lock (write lock only, but Win2k breaks on read lock also so I
do the same) - if you think about why, this is obvious. Also fixed our client
code to do level II oplocks, if requested, and fixed the code where we would
assume the client wanted level II if it advertised itself as being level II
capable - it may not want that.
Jeremy.
through the VFS. All file access/directory access code in smbd should now
go via the vfs. Added vfs_chown/vfs_chmod calls. Still looking at vfs_get_nt_acl()
vfs_set_nt_acl() call API design.
Jeremy.
handling in Samba. This was needed due to several limitations and
races in the previous code - as a side effect the new code is much
cleaner :)
in summary:
- changed sys_select() to avoid a signal/select race condition. It is a
rare race but once we have signals doing notification and oplocks it
is important.
- changed our main processing loop to take advantage of the new
sys_select semantics
- split the notify code into implementaion dependent and general
parts. Added the following structure that defines an implementation:
struct cnotify_fns {
void * (*register_notify)(connection_struct *conn, char *path, uint32 flags);
BOOL (*check_notify)(connection_struct *conn, uint16 vuid, char *path, uint32 flags, void *data, time_t t);
void (*remove_notify)(void *data);
};
then I wrote two implementations, one using hash/poll (like our old
code) and the other using the new Linux kernel change notify. It
should be easy to add other change notify implementations by creating
a sructure of the above type.
- fixed a bug in change notify where we were returning the wrong error
code.
- rewrote the core change notify code to be much simpler
- moved to real-time signals for leases and change notify
Amazingly, it all seems to work. I was very surprised!
fsp->open is no longer needed (if an fsp pointer is valid, then it's open :-).
NB for Luke, this patch also did not apply to TNG. TNG is not yet
identical w.r.t file serving with HEAD. This makes it impossible for
me to help maintain TNG. Please fix asap.
lib/substitute.c: Removed unused variable (pidstr).
Jeremy.
assumption that we have one socket everywhere
while doing so I discovered a few bugs!
1) the clientgen session retarget code if used from smbd or nmbd would
cause a crash as it called close_sockets() which closed our main
socket! fixed by removing close_sockets() completely - it is unnecessary
2) the caching in client_addr() and client_name() was bogus - it could
easily get fooled and give the wrong result. fixed.
3) the retarget could could recurse, allowing an easy denial of
service attack on nmbd. fixed.
handling to printing/printing.c
most of this was just replacing things like fsp->fd_ptr->fd with fsp->fd
the changes in open.c are quite dramatic. Most of it is removing all
the functions that handled the fd multiplexing
instead of either sysv or mmap shared memory or lock files.
this means we can now completely remove
locking_shm.c
locking_slow.c
shmem.c
shmem_sysv.c
and lots of other things also got simpler
locking.c got a bit larger, but is much better compartmentalised now
Tidied up some of the mess (no other word for it). Still doesn't
compile cleanly. There are calls with incorrect parameters that
don't seem to be doing the right thing.
This code still needs surgery :-(.
Jeremy.
