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.
This shrinks include/includes.h.gch by the size of 7 MB and reduces build time
as follows:
ccache build w/o patch
real 4m21.529s
ccache build with patch
real 3m6.402s
pch build w/o patch
real 4m26.318s
pch build with patch
real 3m6.932s
Guenther
When a samba server process dies hard, it has no chance to clean up its entries
in locking.tdb, brlock.tdb, connections.tdb and sessionid.tdb.
For locking.tdb and brlock.tdb Samba is robust by checking every time we read
an entry from the database if the corresponding process still exists. If it
does not exist anymore, the entry is deleted. This is not 100% failsafe though:
On systems with a limited PID space there is a non-zero chance that between the
smbd's death and the fresh access, the PID is recycled by another long-running
process. This renders all files that had been locked by the killed smbd
potentially unusable until the new process also dies.
This patch is supposed to fix the problem the following way: Every process ID
in every database is augmented by a random 64-bit number that is stored in a
serverid.tdb. Whenever we need to check if a process still exists we know its
PID and the 64-bit number. We look up the PID in serverid.tdb and compare the
64-bit number. If it's the same, the process still is a valid smbd holding the
lock. If it is different, a new smbd has taken over.
I believe this is safe against an smbd that has died hard and the PID has been
taken over by a non-samba process. This process would not have registered
itself with a fresh 64-bit number in serverid.tdb, so the old one still exists
in serverid.tdb. We protect against this case by the parent smbd taking care of
deregistering PIDs from serverid.tdb and the fact that serverid.tdb is
CLEAR_IF_FIRST.
CLEAR_IF_FIRST does not work in a cluster, so the automatic cleanup does not
work when all smbds are restarted. For this, "net serverid wipe" has to be run
before smbd starts up. As a convenience, "net serverid wipedbs" also cleans up
sessionid.tdb and connections.tdb.
While there, this also cleans up overloading connections.tdb with all the
process entries just for messaging_send_all().
Volker
Add a simple "processed packet queue" cache to stop nmbd responding to
packets received on the broadcast and non-broadcast socket (which
it has opened when "nmbd bind explicit broadcast = yes").
This is a very simple packet queue - it only keeps the packets
processed during a single call to listen_for_packets() (i.e. one
select call). This means that if the delivery notification for a
packet received on both broadcast and non-broadcast addresses
is done in two different select calls, the packet will still be
processed twice. This is a very rare occurrance and we can just
live with it when it does as the protocol is stateless. If this
is ever flagged as a repeatable problem then we can add a longer
lived cache, using timeout processing to clear etc. etc. But without
storing all packets processed we can never be *sure* we've eliminated
the race condition so I'm going to go with this simple solution until
someone proves a more complex one is needed :-).
Jeremy.
And send replies always via the unicast address of the subnet.
This behavior is off by default (as before)
and can be enabled with "nmbd:bind explicit broadcast = yes".
metze
This provides a compromise between stability and performance: gencache is a
persistent database these days that for performance reasons can not use tdb
transactions for all writes. This patch splits up gencache into gencache.tdb
and gencache_notrans.tdb. gencache_notrans is used with CLEAR_IF_FIRST, writes
to it don't use transactions. By default every 5 minutes and when a program
exits, all entries from _notrans.tdb are transferred to gencache.tdb in one
transaction.
When we run out of file descriptors for some reason, every new
connection forks a child that immediately panics causing smbd to
coredump. This seems unnecessarily harsh; with this code change we
now catch that error and merely log a message about it and exit
without the core dump.
Signed-off-by: Tim Prouty <tprouty@samba.org>
This renames push_string in Samba3 into push_string_base and
push_string_check for the two different use cases.
This should allow push_string to be imported from Samba4, using it's
calling conventions.
Fix logic bug that causes nmbd to wait 5 minutes before
looking for a master browser. This one is *old* :-). Thanks
for Simo for bugging me on this.
Jeremy.
This will be used as part a the franky setup, where nmbd will forward
the MAILSLOT requests to the local samba4 CLDAP server.
"nmbd_proxy_logon:cldap_server = 127.0.0.1" would configure
and activate this feature.
metze
When nmbd is acting as WINS, it picks the first interface's IP as WINS
server's IP. If the first interface's IP is zero, we will just quit
(even though we might have other interfaces with valid IPs).
This patch makes nmbd look at all interfaces and pick the first interface
with a valid IP as the WINS server's IP.
version.h changes rather frequently. Since it is included via includes.h,
this means each C file will be a cache miss. This applies to the following
situations:
* When building a new package with a new Samba version
* building in a git branch after calling mkversion.sh
after a new commit (i.e. virtually always)
This patch improves the situation in the following way:
* remove inlude "version.h" from includes.h
* Use samba_version_string() instead of SAMBA_VERSION_STRING
in files that use no other macro from version.h instead of
SAMBA_VERSION_STRING.
* explicitly include "version.h" in those files that use more
macros from "version.h" than just SAMBA_VERSION_STRING.
Michael
The hosts or networks configured with "init logon delayed hosts"
have their initial samlogon packages (empty username) delayed
by the value configured with "init logon delay" (defaulting
to 100 milliseconds).
This gives the administrator some control over what clients would
consider the preferred logon server: they choose the server that
repsonds most quickly.
Michael
(This used to be commit d52b9beede)
We now open messages.tdb even before we do the become_daemon. become_daemon()
involves a fork and an immediate exit of the parent, thus the
parent_is_longlived argument must be set to false in this case. The parent is
not really long lived :-)
(This used to be commit 4f4781c6d1)
Cause nmbd to wait for an interface, in a mode where SIGTERM
will kills us (same way we wait on startup for an interface).
Jeremy.
(This used to be commit 5440c752ff)
Remove pstring from libsmb/clidfs.c except for a nasty
hack (that will be removed when pstrings are gone from
client/).
Jeremy.
(This used to be commit cc257b71d1)
ip list didn't match namelist added to subnetdb.
Could cause bogus IP addresses to be reported for
the __SAMBA__ name.
Jeremy.
(This used to be commit ad9f14b6dc)
one horror (pstring_clean_name()) which will have to
remain until I've removed all pstrings from the client code.
Jeremy.
(This used to be commit 1ea3ac8014)
== Subject: Remote code execution in Samba's WINS
== server daemon (nmbd) when processing name
== registration followed name query requests.
==
== CVE ID#: CVE-2007-5398
==
== Versions: Samba 3.0.0 - 3.0.26a (inclusive)
...
Secunia Research reported a vulnerability that allows for
the execution of arbitrary code in nmbd. This defect may
only be exploited when the "wins support" parameter has
been enabled in smb.conf.
(This used to be commit e40c372e0d)
Each cli struct has it's own local copy of this variable,
so use that in client code. In the smbd server, add one
static to smbd/proccess.c and use that inside smbd. Fix
a bunch of places where smb_rw_error could be set by
calling read_data() in places where we weren't reading
from the SMB client socket (ie. winbindd).
Jeremy.
(This used to be commit 255c2adf7b)
The point is doing the following associations:
- non discardable state data (all TDB files that may need to be backed
up) go to statedir
- shared data (codepage stuff) go to codepagedir
The patch *does not change* the default location for these
directories. So, there is no behaviour change when applying it.
The main change is for samba developers who have to think when dealing
with files that previously pertained to libdir whether they:
- go in statedir
- go in codepagedir
- stay in libdir
(This used to be commit d6cdbfd875)
to struct sockaddr_storage in most places that matter (ie.
not the nmbd and NetBIOS lookups). This passes make test
on an IPv4 box, but I'll have to do more work/testing on
IPv6 enabled boxes. This should now give us a framework
for testing and finishing the IPv6 migration. It's at
the state where someone with a working IPv6 setup should
(theorecically) be able to type :
smbclient //ipv6-address/share
and have it work.
Jeremy.
(This used to be commit 98e154c312)
jra: POPT_ARG_VAL arguments need int values.
I assume there're more places like this
in the cmdline tools.
Please fix this properly, as my commit is just
a hack to get make test working again.
in samba4 we have a workaround for this see
smbd/server.c
metze
(This used to be commit 9cb1937fe8)
bugs in various places whilst doing this (places that assumed
BOOL == int). I also need to fix the Samba4 pidl generation
(next checkin).
Jeremy.
(This used to be commit f35a266b3c)
IPv6 in winbindd, but moves most of the socket functions that were
wrongly in lib/util.c into lib/util_sock.c and provides generic
IPv4/6 independent versions of most things. Still lots of work
to do, but now I can see how I'll fix the access check code.
Nasty part that remains is the name resolution code which is
used to returning arrays of in_addr structs.
Jeremy.
(This used to be commit 3f6bd0e1ec)
the ripple effects this causes. utmp has to change etc. Remove some
global varables and store address/port in the unexpected db.
Jeremy.
(This used to be commit 18c6a2211d)
always consistent. Return -1 on error, and ensure
we check for this. In cases where the dest is already
specified and we've been asked to terminate with a
null, ensure we always do so even on error.
Jeremy.
(This used to be commit abedd96786)
This adds the two functions talloc_stackframe() and talloc_tos().
* When a new talloc stackframe is allocated with talloc_stackframe(), then
* the TALLOC_CTX returned with talloc_tos() is reset to that new
* frame. Whenever that stack frame is TALLOC_FREE()'ed, then the reverse
* happens: The previous talloc_tos() is restored.
*
* This API is designed to be robust in the sense that if someone forgets to
* TALLOC_FREE() a stackframe, then the next outer one correctly cleans up and
* resets the talloc_tos().
The original motivation for this patch was to get rid of the
sid_string_static & friends buffers. Explicitly passing talloc context
everywhere clutters code too much for my taste, so an implicit
talloc_tos() is introduced here. Many of these static buffers are
replaced by a single static pointer.
The intended use would thus be that low-level functions can rather
freely push stuff to talloc_tos, the upper layers clean up by freeing
the stackframe. The more of these stackframes are used and correctly
freed the more exact the memory cleanup happens.
This patch removes the main_loop_talloc_ctx, tmp_talloc_ctx and
lp_talloc_ctx (did I forget any?)
So, never do a
tmp_ctx = talloc_init("foo");
anymore, instead, use
tmp_ctx = talloc_stackframe()
:-)
Volker
(This used to be commit 6585ea2cb7)
smbd, nmbd and winbindd can be started with invalid options currently.
The first patch attached would be a possible solution.
It contains an exit if an invalid option has been used. The main problem
is, that existing setups with wrong options or missing arguments in start
scripts will break (which is the right behaviour from my point of view).
metze
(This used to be commit 8532e3182a)
add in any cluster addresses. We need to response to these,
but not listen on them. This allows us to run nmbd on every
node in the cluster, and have all of them register with a
WINS server correctly
metze
(This used to be commit 98c57562c4)
systems. Rather than trying to fix this, it's much better just to
remove the code, as it serves no purpose at all (the message that is
generated is ignored by smbd).
This sort of parsing should really be done by PIDL generated code.
(This used to be commit 22e8404cef)
doing this because for the clustering the marshalling is needed in more
than one place, so I wanted a decent routine to marshall a message_rec
struct which was not there before.
Tridge, this seems about the same speed as it used to be before, the
librpc/ndr overhead in my tests was under the noise.
Volker
(This used to be commit eaefd00563)
and connections_forall. This centralizes all the routines that did individual
tdb_open("connections.tdb") and direct tdb_traverse.
Volker
(This used to be commit e43e94cda1)
This changes "struct process_id" to "struct server_id", keeping both is
just too much hassle. No functional change (I hope ;-))
Volker
(This used to be commit 0ad4b1226c)
to all callers of smb_setlen (via set_message()
calls). This will allow the server to reflect back
the correct encryption context.
Jeremy.
(This used to be commit 2d80a96120)
void message_register(int msg_type,
void (*fn)(int msg_type, struct process_id pid,
- void *buf, size_t len))
+ void *buf, size_t len,
+ void *private_data),
+ void *private_data)
{
struct dispatch_fns *dfn;
So this adds a (so far unused) private pointer that is passed from
message_register to the message handler. A prerequisite to implement a tiny
samba4-API compatible wrapper around our messaging system. That itself is
necessary for the Samba4 notify system.
Yes, I know, I could import the whole Samba4 messaging system, but I want to
do it step by step and I think getting notify in is more important in this
step.
Volker
(This used to be commit c8ae60ed65)
closer at the wins server code. Firstly, it needs
to do the searches on the SELF_NAMES correctly,
secondly it needs to flush the in-memory cache
out before returning the 1b names - else it might
get duplicates returned if many 1b queries are
done in quick succession. Jerry, I hate to say
this but you might want to consider this for 3.0.23....
Jeremy.
(This used to be commit b36b9befbb)
is produced when a process exits abnormally.
First, we coalesce the core dumping code so that we greatly improve our
odds of being able to produce a core file, even in the case of a memory
fault. I've removed duplicates of dump_core() and split it in two to
reduce the amount of work needed to actually do the dump.
Second, we refactor the exit_server code path to always log an explanation
and a stack trace. My goal is to always produce enough log information
for us to be able to explain any server exit, though there is a risk
that this could produce too much log information on a flaky network.
Finally, smbcontrol has gained a smbd fault injection operation to test
the changes above. This is only enabled for developer builds.
(This used to be commit 56bc02d644)
to make the following possible:
timelimit 20000 bin/nmbd -F -S --no-process-group
timelimit 20000 bin/smbd -F -S --no-process-group
this is needed to 'make test' working without losing child processes
metze
(This used to be commit c3a9f30e2a)
lp_load() could not be called multiple times to modify parameter settings based
on reading from multiple configuration settings. Each time, it initialized all
of the settings back to their defaults before reading the specified
configuration file.
This patch adds a parameter to lp_load() specifying whether the settings should
be initialized. It does, however, still force the settings to be initialized
the first time, even if the request was to not initialize them. (Not doing so
could wreak havoc due to uninitialized values.)
(This used to be commit f2a24de769)
use it as though it were an in-memory db and dump out to
a flat file every 2 mins, but that can now change.
Jeremy.
(This used to be commit a342681792)
box with gcc4 and -O6...
Fix a bunch of C99 dereferencing type-punned pointer will break
strict-aliasing rules errors. Also added prs_int32 (not uint32...)
as it's needed in one place. Find places where prs_uint32 was being
used to marshall/unmarshall a time_t (a big no no on 64-bits).
More warning fixes to come.
Thanks to Volker for nudging me to compile like this.
Jeremy.
(This used to be commit c65b752604)
of the Samba4 timezone handling code back into Samba3.
Gets rid of "kludge-gmt" and removes the effectiveness
of the parameter "time offset" (I can add this back
in very easily if needed) - it's no longer being
looked at. I'm hoping this will fix the problems people
have been having with DST transitions. I'll start comprehensive
testing tomorrow, but for now all modifications are done.
Splits time get/set functions into srv_XXX and cli_XXX
as they need to look at different timezone offsets.
Get rid of much of the "efficiency" cruft that was
added to Samba back in the day when the C library
timezone handling functions were slow.
Jeremy.
(This used to be commit 414303bc02)
* \PIPE\unixinfo
* winbindd's {group,alias}membership new functions
* winbindd's lookupsids() functionality
* swat (trunk changes to be reverted as per discussion with Deryck)
(This used to be commit 939c3cb5d7)
printmig.exe work
* merge the sys_select_signal(char c) change from trunk
in order to keeo the winbind code in sync
(This used to be commit a112c5570a)
use the correct RR type of 0xA instead of reflecting back what
the query RR type was (0x20). See rfc1002 sections 4.2.14 and
4.2.16.
Jeremy.
(This used to be commit ab8c924004)
functions so we can funnel through some well known functions. Should help greatly with
malloc checking.
HEAD patch to follow.
Jeremy.
(This used to be commit 620f2e608f)
Don't use nstrings to hold workgroup and netbios names. The problem with them is that MB netbios
and workgroup names in unix charset (particularly utf8) may be up to 3x bigger than the name
when represented in dos charset (ie. cp932). So go back to using fstrings for these but
translate into nstrings (ie. 16 byte length values) for transport on the wire.
Jeremy.
(This used to be commit b4ea493599)
location packet from the client is not a DEBUG(1) event anymore...
(Yes, we printed this for each of these UDP packets...)
Andrew Bartlett
(This used to be commit 4ca0ab7506)
Hmmm. This is correct in 2.2. Obviously I did not test my 3.0 checkin at that
time. Now it hit me at a customer's site...
Volker
(This used to be commit a0e741aa68)
in iconv.c and nsswitch/). Using them means you're not thinking about multibyte at
all and I really want to discourage that.
Jeremy.
(This used to be commit d7e35dfb92)
to be appended to the netbios name between angle brackets. This interfered
the 'sh -c' used to implement smbrun(). Closes bug #528.
(This used to be commit 92b37b3ef0)
iconv wasn't re-initialised on reading of "charset" parameters. This
caused workgroup name to be set incorrectly if it contained an
extended character.
Jeremy.
(This used to be commit 84ae44678a)
converted to pull/push_ascii. This will not work right at the moment for non
English codepages, but compiles - I will finish the work over the weekend.
Then nmbd should be completely codepage correct.
Jeremy.
(This used to be commit 236d6adadf)
that is now possible to, for example, load a module which contains
an auth method into a binary without the auth/ subsystem built in.
(This used to be commit 74d9ecfe2d)
the other infrastructure with name owners etc in place. If anybody is
really going to tackle winsrepld, it will probably not be hard to put
the additional info back.
Volker
(This used to be commit eb82daa84a)
servers don't answer that name. However we *know* they
have the name workgroup#1b (as we just looked it up).
So do the node status request on this name instead.
Found at LBL labs.
Jeremy.
(This used to be commit 41e3abe8b8)
This uses 'socket address' as the source address for nmbd. This way we
can again synchronize with the DMB if we have 'bind interfaces only'
to a virtual interface.
I'd love to see this in 2.2.9, but that is up to jerry or jra.
Volker
(This used to be commit fe637c690b)
a pstrcpy/fstrcpy or at most a safe_strcpy().
These have the advantage of being compiler-verifiable.
Get these out of the way, along with a rewrite of 'get_short_archi' in the
spoolss client and server. (This pushes around const string pointers, rather
than copied strings).
Andrew Bartlett
(This used to be commit 32fb801ddc)
'set_local_machine_name' so that the client can't change it from under us.
(.NET RC2 and WinXP install calls the machine 'machinename' during NTLMSSP
on the domain join).
Andrew Bartlett
(This used to be commit 4c7163e7c2)
