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.
Creating a file in a directory with delete-on-close set returns
DELETE_PENDING, and trying to set the flag on a non-empty directory returns
DIRECTORY_NOT_EMPTY.
Volker
(This used to be commit 5680f34778)
A delete-on-close deleted file is still around while open on another fd. But
only for findfirst, not for qpathinfo :-)
Volker
(This used to be commit dbc7a1a978)
The biggest change was fixing the RAW-CONTEXT test. It was forcing
capabilities to zero in an attempt to not negotiated extended
security, but as a side effect it was forcing negotiation of dos error
codes. This confused the hell out of the test code!
Also fixed a bunch of places incorrectly using NT_STATUS_V() instead
of NT_STATUS_EQUAL() and several places that had the wrong dos status
codes
(This used to be commit 0b22744f40)
that relied on the mapping need to be fixed. The first thing is to get
all the torture tests working against w2k3 again with nt status codes
enabled. The 2nd step will be to make them pass with nt status
disabled.
This starts on the first task, fixing the assumption that
NT_STATUS_INVALID_LOCK_SEQUENCE is a valid substitute for
ERRDOS:ERRbadaccess
(This used to be commit 87cdd11708)
with 'nt status support' option.
- make nt_errstr() display nice strings for dos status codes encoded
using NT_STATUS_DOS()
- no longer map between dos and nt status codes in the client library,
instead return using NT_STATUS_DOS()
- fixed the RAW-CONTEXT test to look for
NT_STATUS_DOS(ERRSRV, ERRbaduid) instead of NT_STATUS_INVALID_HANDLE
(This used to be commit ff5549e87f)
- qfsinfo (query file system information)
- appendacl (append an ACL to existing file's security descriptor and get new
full ACL)
The second one also includes an improvement to security descriptor handling
which allows to copy security descriptor. Written by Peter Novodvorsky
<peter.novodvorsky@ru.ibm.com>
Both functions have corresponding torture tests added. Tested under valgrind and
work against Samba 4 and Windows XP.
ToDo: document composite call creation process in prog_guide.txt
(This used to be commit 441cff62ac)
GENSEC, and to pull SCHANNEL into GENSEC, by making it less 'special'.
GENSEC now no longer has it's own handling of 'set username' etc,
instead it uses cli_credentials calls.
In order to link the credentails code right though Samba, a lot of
interfaces have changed to remove 'username, domain, password'
arguments, and these have been replaced with a single 'struct
cli_credentials'.
In the session setup code, a new parameter 'workgroup' contains the
client/server current workgroup, which seems unrelated to the
authentication exchange (it was being filled in from the auth info).
This allows in particular kerberos to only call back for passwords
when it actually needs to perform the kinit.
The kerberos code has been modified not to use the SPNEGO provided
'principal name' (in the mechListMIC), but to instead use the name the
host was connected to as. This better matches Microsoft behaviour,
is more secure and allows better use of standard kerberos functions.
To achieve this, I made changes to our socket code so that the
hostname (before name resolution) is now recorded on the socket.
In schannel, most of the code from librpc/rpc/dcerpc_schannel.c is now
in libcli/auth/schannel.c, and it looks much more like a standard
GENSEC module. The actual sign/seal code moved to
libcli/auth/schannel_sign.c in a previous commit.
The schannel credentails structure is now merged with the rest of the
credentails, as many of the values (username, workstation, domain)
where already present there. This makes handling this in a generic
manner much easier, as there is no longer a custom entry-point.
The auth_domain module continues to be developed, but is now just as
functional as auth_winbind. The changes here are consequential to the
schannel changes.
The only removed function at this point is the RPC-LOGIN test
(simulating the load of a WinXP login), which needs much more work to
clean it up (it contains copies of too much code from all over the
torture suite, and I havn't been able to penetrate its 'structure').
Andrew Bartlett
(This used to be commit 2301a4b38a)
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.
(This used to be commit 9db6c79e90)
make it possible to add optimisations to the events code such as
keeping the next timed event in a sorted list, and using epoll for
file descriptor events.
I also removed the loop events code, as it wasn't being used anywhere,
and changed timed events to always be one-shot (as adding a new timed
event in the event handler is so easy to do if needed)
(This used to be commit d7b4b6de51)
handle the inverted memory hierarchy that a normal session
establishment gave. The inverted hierarchy came from that fact that
you first establish a socket, then a transport, then a session and
finally a tree. That leads to the socket being at the top of the
memory hierarchy and the tree at the bottom, which makes no sense from
the users point of view, as they want to be able to free the tree and
have everything disappear.
The core problem was that the libcli interface didn't distinguish
between establishing a primary context and a secondary context. If you
establish a 2nd session on a transport then you want the transport to
be referenced by the session, whereas if you establish a primary
session then you want the transport to be a child of the session.
To fix this I have added "parent_ctx" and "primary" arguments to the
libcli intialisation functions. This makes using the library much
easier, and gives us a memory hierarchy that makes much more sense.
I was prompted to do this by a bug in the cifs backend, which was
caused by the socket not being properly torn down on a disconnect due
to the inverted memory hierarchy.
(This used to be commit 5e8fd5f701)
encapsulates all the different session setup methods, including the
multi-pass spnego code.
I have hooked this into all the places that previously used the
RAW_SESSSETUP_GENERIC method, and have removed the old
RAW_SESSSETUP_GENERIC code from clisession.c and clitree.c. A nice
side effect is that these two modules are now very simple again, back
to being "raw" session setup handling, which was what was originally
intended.
I have also used this to replace the session setup code in the
smb_composite_connect() code, and used that to build a very simple
replacement for smbcli_tree_full_connection().
As a result, smbclient, smbtorture and all our other SMB connection
code now goes via these composite async functions. That should give
them a good workout!
(This used to be commit 080d0518bc)
- added #if TALLOC_DEPRECATED around the _p functions
- fixes the code that broke from the above
while doing this I fixed quite a number of places that were
incorrectly using the non type-safe talloc functions to use the type
safe ones. Some were even doing multiplies for array allocation, which
is potentially unsafe.
(This used to be commit 6e7754abd0)
win2003 bug!
This new test code works against w2k, and against longhorn, but fails
against w2k3. When tested against w2k3 it allows a open with an access
mask that should be denied by the given ACL, after setting up the ACL
using inheritance. Note that only the very specific
SEC_RIGHTS_FILE_ALL mask incorrectly succeeds, so they must have a
special case for that mask. Maybe its an optimisation gone wrong?
I don't know if there are any serious security implications to this,
but it is pretty clearly wrong, and has been fixed in longhorn.
(This used to be commit 4f9fd767db)
talloc_size() or talloc_array_p() where appropriate.
also fixed a memory leak in pvfs_copy_file() (failed to free a memory
context)
(This used to be commit 89b74b5354)
- 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.
(This used to be commit e6c81d7c9f)
- added a test for all combinations of the inheritance ACE flags and how
they are propogated to child directories and files
(This used to be commit fdb38c8e4b)
- added test for chained OpenX/ReadX, simulating the OS/2 workplace shell
- fixed a bug in handling chained fnum in openx and ntcreatex in the server
(yes, I'm on holiday, but this bug was annoying me ....)
(This used to be commit b3b8958a18)
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 25189b8fbf)
to kukks on #samba-technical for the sniffs that allowed me to work
this out
- much simpler ntvfs open generic mapping code
- added t2open create with EA torture test to RAW-OPEN test
(This used to be commit a56d95ad89)
this is mostly just a tidyup, but also adds the privilege_mask, which
I will be using shortly in ACL checking.
note that I had to move the definition of struct security_token out of
security.idl as pidl doesn't yet handle arrays of pointers, and the
usual workaround (to use a intermediate structure) would make things
too cumbersome for this structure, especially given we never encode it
to NDR.
(This used to be commit 7b446af09b)
- change smbcli_read/write to take void * for the buffers to match read(2)/write(2)
all this fixes a lot of gcc-4 warnings
metze
(This used to be commit b94f92bc66)
generic->specific access mask mappings, and tests of the behaviour of
SID_CREATOR_OWNER and SEC_FLAG_MAXIMUM_ALLOWED
(This used to be commit f572fe6d29)
This removes the duplicate named SEC_RIGHTS_MAXIMUM_ALLOWED and
SEC_RIGHTS_FULL_CONTROL, which are just other names for
SEC_FLAG_MAXIMUM_ALLOWED and SEC_RIGHTS_FILE_ALL. The latter names
match the new naming conventions in security.idl
Also added names for the generic->specific mappings for files are
directories
(This used to be commit 17a4e0b3ac)
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.
(This used to be commit 01c0fa722f)
queryfileinfo/setfileinfo logic, so querying/setting a security
descriptor is treated as just another file query/set operation.
This will allow NTVFS backends to see the query/set security
descriptor operations as RAW_FILEINFO_SEC_DESC and
RAW_SFILEINFO_SEC_DESC operations.
(This used to be commit f68a6b6b91)
- added a delete-on-close test to RAW-STREAMS
- don't allow rename of streams at the moment (I need to work out if
its supposed to be allowed first)
(This used to be commit f4b2b1311d)
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.
(This used to be commit b6452c4a20)
clear what the correct behaviour is for delayed stat info update.
- use a common torture_setup_dir() function for setting up a test
directory in torture tests.
(This used to be commit f7fb34715b)
run test_posix.sh in a reaonable time by using posix:sharedelay=100000
this still tests correct multiplex behaviour
(This used to be commit 64a6db5b02)
testing. Interestingly, w2k3 does not allow the cancel of an
outstanding async open request, whereas it does allow the cancel of an
outstanding async lock request. To support this I have changed the
pvfs_wait interface to provide a enum on why the event is happening,
so the callback can decide what to do.
(This used to be commit f23d6a2800)
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().
(This used to be commit eded2ad9c9)
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.
(This used to be commit 0d51511d40)
rather extensive test reveals some really bizarre error code
handling in w2k3.
- extended and simplified the RAW-CHKPATH test, making it easier to
read (note that Samba3 fails the new tests - jra may wish to look)
- marked RAW-CHKPATH as pass for pvfs
(This used to be commit 32dccf91cf)
- 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 bc83d823b2)
- added a pvfs_lock_close_pending() hook to remove pending locks on file close
- fixed the private ptr argument to messaging_deregister() in pvfs_wait
- fixed a bug in continuing lock requests after a lock that is blocking a pending lock is removed
- removed bogus brl_unlock() call in lock continue
- corrected error code for LOCKING_ANDX_CHANGE_LOCKTYPE
- expanded the lock cancel test suite to test lock cancel by unlock and by close
- added a testsuite for LOCKING_ANDX_CHANGE_LOCKTYPE
(This used to be commit 5ef80f034d)
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.
(This used to be commit 4395c0557a)
excellent, as it means we don't lose any compatibility by also not
giving sorted listings. I was very much afraid that applications had
begun to rely on this (as its guaranteed by w2k NTFS).
no longer mark a non-sorting server as an error
(This used to be commit 1d21c9a9d1)
taking a context (so when you pass a NULL pointer you end up with
memory in a top level context). Fixed it by changing the API to take a
context. The context is only used if the pointer you are reallocing is
NULL.
(This used to be commit 8dc23821c9)
rather than manual reference counts
- properly support SMBexit in the cifs and posix backends
- added a logoff method to all backends
With these changes the RAW-CONTEXT test now passes against the posix backend
(This used to be commit c315d6ac1c)
to get auto-naming of pointers very cheaply.
- fixed a couple of memory leaks found with the new tricks
A typical exit report for smbd is now:
talloc report on 'null_context' (total 811 bytes in 54 blocks)
auth/auth_sam.c:334 contains 20 bytes in 1 blocks
struct auth_serversupplied_info contains 498 bytes in 33 blocks
UNNAMED contains 8 bytes in 1 blocks
lib/data_blob.c:40 contains 16 bytes in 1 blocks
iconv(CP850,UTF8) contains 61 bytes in 4 blocks
iconv(UTF8,CP850) contains 61 bytes in 4 blocks
iconv(UTF8,UTF-16LE) contains 67 bytes in 4 blocks
iconv(UTF-16LE,UTF8) contains 67 bytes in 4 blocks
UNNAMED contains 13 bytes in 1 blocks
which is much better than before
(This used to be commit 6e721393d0)
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 71480271ad)
this means -U DOM\\user is know allowed
- torture:userdomain is a new smb.conf parameter
because lp_workgroup is not the domain of the user
- we use torture:userdomain now in the tests instad of lp_workgroup
- for backward compat the userdomain is lp_workgroup() by default and
not lp_netbios_name(), which my change later to match 'net' and 'smbclient'..
- we now have dublicate options e.g. -N -s ...
tridge: can we change this?
metze
(This used to be commit 4733dcbf5f)
This version does the following:
1) talloc_free(), talloc_realloc() and talloc_steal() lose their
(redundent) first arguments
2) you can use _any_ talloc pointer as a talloc context to allocate
more memory. This allows you to create complex data structures
where the top level structure is the logical parent of the next
level down, and those are the parents of the level below
that. Then destroy either the lot with a single talloc_free() or
destroy any sub-part with a talloc_free() of that part
3) you can name any pointer. Use talloc_named() which is just like
talloc() but takes the printf style name argument as well as the
parent context and the size.
The whole thing ends up being a very simple piece of code, although
some of the pointer walking gets hairy.
So far, I'm just using the new talloc() like the old one. The next
step is to actually take advantage of the new interface
properly. Expect some new commits soon that simplify some common
coding styles in samba4 by using the new talloc().
(This used to be commit e35bb094c5)
