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.
quite a large change as we had lots of code that assumed that
objectSid was a string in S- format.
metze and simo tried to convince me to use NDR format months ago, but
I didn't listen, so its fair that I have the pain of fixing all the
code now :-)
This builds on the ldb_register_samba_handlers() and ldif handlers
code I did earlier this week. There are still three parts of this
conversion I have not finished:
- the ltdb index records need to use the string form of the objectSid
(to keep the DNs sane). Until that it done I have disabled indexing on
objectSid, which is a big performance hit, but allows us to pass
all our tests while I rejig the indexing system to use a externally
supplied conversion function
- I haven't yet put in place the code that allows client to use the
"S-xxx-yyy" form for objectSid in ldap search expressions. w2k3
supports this, presumably by looking for the "S-" prefix to
determine what type of objectSid form is being used by the client. I
have been working on ways to handle this, but am not happy with
them yet so they aren't part of this patch
- I need to change pidl to generate push functions that take a
"const void *" instead of a "void*" for the data pointer. That will
fix the couple of new warnings this code generates.
Luckily it many places the conversion to NDR formatted records
actually simplified the code, as it means we no longer need as many
calls to dom_sid_parse_talloc(). In some places it got more complex,
but not many.
management system I proposed on samba-technical a couple of days
ago. Essentially it is a very lightweight way for any code in Samba to
make IDL based rpc calls to anywhere else in the code, without the
client or server having to go to the trouble of setting up a full rpc
service.
It can be used with any of our existing IDL, but I expect it will
mostly be used for a new set of Samba specific management calls.
The LOCAL-IRPC torture test demonstrates how it can be used by calling
the echo_AddOne() call over this transport.
secrets system, and not the old system from Samba3.
This allowed the code from auth_domain to be shared - we now only
lookup the secrets.ldb in lib/credentials.c.
In order to link the resultant binary, samdb_search() has been moved
from deep inside rpc_server into lib/gendb.c, along with the existing
gendb_search_v(). The vast majority of this patch is the simple
rename that followed,
(Depending on the whole SAMDB for just this function seemed pointless,
and brought in futher dependencies, such as smbencrypt.c).
Andrew Bartlett
less likely that anyone will use pstring for new code
- got rid of winbind_client.h from includes.h. This one triggered a
huge change, as winbind_client.h was including system/filesys.h and
defining the old uint32 and uint16 types, as well as its own
pstring and fstring.
- removed the u32 hack in events.c as I think this was only needed as
tdb.h defines u32. Metze, can you check that this hack is indeed no
longer needed on your suse system?
servers in smbd. The old code still contained a fairly bit of legacy
from the time when smbd was only handling SMB connection. The new code
gets rid of all of the smb_server specific code in smbd/, and creates
a much simpler infrastructures for new server code.
Major changes include:
- simplified the process model code a lot.
- got rid of the top level server and service structures
completely. The top level context is now the event_context. This
got rid of service.h and server.h completely (they were the most
confusing parts of the old code)
- added service_stream.[ch] for the helper functions that are
specific to stream type services (services that handle streams, and
use a logically separate process per connection)
- got rid of the builtin idle_handler code in the service logic, as
none of the servers were using it, and it can easily be handled by
a server in future by adding its own timed_event to the event
context.
- fixed some major memory leaks in the rpc server code.
- added registration of servers, rather than hard coding our list of
possible servers. This allows for servers as modules in the future.
- temporarily disabled the winbind code until I add the helper
functions for that type of server
- added error checking on service startup. If a configured server
fails to startup then smbd doesn't startup.
- cleaned up the command line handling in smbd, removing unused options
- cleaned up some talloc usage in various files
I'd like to get to the point that we have no calls to talloc(), at
which point we will rename talloc_p() to talloc(), to encourage
everyone to use the typesafe functions.
definitions for security access masks, in security.idl
The previous definitions were inconsistently named, and contained many
duplicate and misleading entries. I kept finding myself tripping up
while using them.
our local domain. Note that this linear mapping does not suffer from
the "foreign sid" problems of the linear mappings we have previously
rejected for the sid->uid problem.
the mapping allows for 1 billion automatically allocated users or
groups for the local domain.
The trickiest part about this was getting the sharing and locking
rules right, as alternate streams are separate locking spaces from the
main file for the purposes of byte range locking, and separate for
most share violation rules.
I suspect there are still problems with delete on close with alternate
data streams. I'll look at that next.
preparation for adding code to pass the BASE-DENY1 and BASE-DENYDOS
tests, which require a shared filesystem handle for some specific
combinations of two DENY_DOS opens on the same connection.
database in the opendb lck, we ensure that the database is not closed
before the lock is gone. That ensures the lock destructor doesn't work
on a closed database.
The previous code didn't handle the case where the file got renamed or
deleted while waiting for the sharing violation delay. To handle this
we need to make the 2nd open a full open call, including the name
resolve call etc. Luckily this simplifies the logic.
I also expanded the RAW-MUX test to include the case where we do
open/open/open/close/close, with the 3rd open async, and that open
gets retried after both the first close and the 2nd close, with the
first retry failing and the 2nd retry working. The tests the "async
reply after a async reply" logic in pvfs_open().
deferred reply is short-circuited immediately when the file is
closed by another user, allowing it to be opened by the waiting user.
- added a sane set of timeval manipulation routines
- converted all the events code and code that uses it to use struct
timeval instead of time_t, which allows for microsecond resolution
instead of 1 second resolution. This was needed for doing the pvfs
deferred open code, and is why the patch is so big.
We found a few months ago that TDB_CLEAR_IF_FIRST is extremely
inefficient for large numbers of connections, due to a fundamental
limitation in the way posix byte range locking is implemented. Rather
than the nasty workaround we had for Samba3, we now have a single
"cleanup tmp files" function that runs when smbd starts. That deletes
the tmp tdbs, so TDB_CLEAR_IF_FIRST is not needed at all.
wakeup all pending locks at once. This change means that we only
trigger this anti-stampede code for write locks, as for pending read
locks the correct behaviour is to stampede (as they will all succeed)
pvfs_open, and handle the various race conditions that are inherent in
cifs on unix, so we do the best we can when the race happens.
the ntcreatex code is really starting to take shape now
want to expose the brl context structure outside the brlock.c
code. Instead, I now use "struct brl_context *" and rely on C being
happy to pass around pointers to unknown structures as long as they
are not dereferenced. I will be interested to see how the build farm
likes this.
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 adds a pvfs_wait_message() routine which uses the new messaging
system, event timers and talloc destructors to give a nice generic
async event handling system with a easy to use interface. The
extensions to pvfs_lock.c are based on calls to pvfs_wait_message()
routines.
We now pass all of our smbtorture locking tests, although while
writing this code I have thought of some additonal tests that should
be added, particularly for lock cancel operations. I'll work on that
soon.
This commit also extends the smbtorture lock tests to test the rather
weird 0xEEFFFFFF locking semantics that I have discovered in
win2003. Win2003 treats the 0xEEFFFFFF boundary as special, and will
give different error codes on either side of it. Locks on both sides
are allowed, the only difference is which error code is given when a
lock is denied. Anyone like to hazard a guess as to why? It has
me stumped.
