Andrew Tridgell [Mon, 7 Oct 1996 01:56:21 +0000 (01:56 +0000)]
- changed the default nmbd loop timout to 10 seconds (2 seconds was much
too short)
- got rid of many unnecessary calls to time(NULL) in nmbd. They were
causing it to chew too much CPU time when idle. Now we pass a time
value in from the top level.
Andrew Tridgell [Sat, 5 Oct 1996 10:41:13 +0000 (10:41 +0000)]
I have fixed quite a few important bugs in this commit.
Luke, can you take special note of the bug fixes to nmbd so you can
propogate them to your new code.
- rewrote the code that used to use fromhost(). We now call
gethostbyaddr() only if necessary and a maximum of once per
connection. Calling gethostbyaddr() causes problems on some systems so
avoiding it if possible is a good thing :-)
- added the "fake oplocks" option. See the docs in smb.conf(5) and
Speed.txt
- fixed a serious bug in nmbd where it would try a DNS lookup on
FIND_SELF queries. This caused a lot of unnecessary (and incorrect)
DNS lookups to happen. FIND_SELF queries should only go to the
internal name tables.
- don't set FIND_SELF for name queries if we are a wins proxy, as we
are supposed to be answering queries for other hosts.
- fixed a bug in nmbd which had "if (search | FIND_LOCAL)" instead of
"if (search & FIND_LOCAL)". Luke, this was in nameservreply.c
- the above 3 bugs together meant that DNS queries were being cached,
but the cache wasn't being used, so every query was going to DNS, no
wonder nmbd has been chewing so much CPU time! Another side effect was
that queries on names in lmhosts weren't being answered for bcast
queries with "wins proxy" set.
- ignore the maxxmit for seconday session setups (see CIFS spec)
- close user opened files in a uLogoffX for user level security (see
CIFS spec)
- added uid into the files struct to support the above change
Andrew Tridgell [Fri, 4 Oct 1996 09:31:07 +0000 (09:31 +0000)]
- changed the umask handling. We now set the umask to 0 and explicitly
set the mode on all created files. I think this is a better policy.
- change the debug levels on some items
- fix a charset handling bug which affected foreign and extended
charset users
- no longer switch back to the original directory when idle, instead
switch to / as the original directory may not be readable by ordinary
users.
- fix some bugs where the create mode of files was not being
explicitly set (it was relying on the umask and using fopen). Not a
big bug as it only affected obscure commands like the messaging ops.
- got rid of the lock code in the lpq cache as its no longer needed
- rewrote smbrun to be faster and to remove the security hole. We now
don't actually need a external smbrun binary, its all done by smbd.
- add a more explicit warning about uids and gids of -1 or 65535
Andrew Tridgell [Wed, 2 Oct 1996 14:09:22 +0000 (14:09 +0000)]
- a huge pile of changes from Luke which implement the browse.conf
stuff and also fix a pile of nmbd bugs. Unfortunately I found it very
hard to disentangle the new features from the bug fixes so I am
putting in the new code. I hope this is the last big pile of changes
to the 1.9.16 series!
Edited doc directory to remove anything that pointed to Karl Auer, since he
is still getting lots of email. Also updated some of the PROJECTS file. Luke,
you should do a quick paragraph or two there.
Andrew Tridgell [Wed, 21 Aug 1996 14:15:09 +0000 (14:15 +0000)]
- fix a bug handling readraw packets that caused the timeout to be 30
milliseconds instead of 30 seconds as was intended. Thanks to Paul
Nelson for finding this bug.
JHT ==> Upgraded Samba to Windows NT Server version 4.1
Just to keep ahead of the pack.
Note: Really only done to differntiate our version ID from
existing SMB Servers.
Andrew Tridgell [Wed, 21 Aug 1996 06:09:00 +0000 (06:09 +0000)]
- new handling of ST_TYPE bits, they are now consolidated much more in
DFLT_SERVER_TYPE in nameserv.h
- got rid of a lot of spurious domain controller stuff. Samba is not a
domain controller yet, but it can be a domain master. We were claiming
to be a domain controller in some packets which may have caused
problems
- don't do preferred master startups on the WINS pseudo-net
- don't do election requests on the WINS pseudo-net
- fix a nasty bug in become_non_master() which wiped out the bits in
remove_type before using them. The result was that samba didn't like
losing its master status.
- changed the logic in the election packet handling to enable us to
become a non-master whenever we receive a winning election frame, even
if we aren't expecting it
- get another packet from the socket in nmbd when we reject one of our
own packets, this stops us from going into the packet reading code too
often and makes nmbd much snappier
- always remove a name immediately when we try to release it, don't
wait for the lack of response from the network, otherwise we will end
up replying to name that we don't really own. We still send the dereg
packets, we just don't wait for them to time out.
Andrew Tridgell [Tue, 20 Aug 1996 15:45:16 +0000 (15:45 +0000)]
- fix a bug in NetServerEnum where counted and total were not counted
correctly if there were multiple instances of a name. This led to the
infamous "not enough memory" error when browsing (but this isn't the
only cause of that message)
- fix a triple-chaining bug which affected OpenX following a TconX
- fix a serious nmbd bug that meant nmdb would answer packets that it
wasn't supposed to, causing havoc with browse lists.
- never time out SELF packets. This is an interim fix until I find out
why nmbd thought they should be timed out.
Andrew Tridgell [Sat, 17 Aug 1996 14:27:12 +0000 (14:27 +0000)]
- made FAST_SHARE_MODES standard for Linux
- you can now press enter after a password prompt with "smbclient -L
hostname" and it will use a blank username and password, the same as
if you used -U%
- changed the wins.dat store code to not go via a string variable
(just use fprintf instead)
- removed a "unsigned long" that luke put back in, changing it to
uint32 to keep 64 bit machines happy. Naughtly luke!
- allow guest non-connnected print queue listings so ms client 3 can
work
- updated load_netbios_names() to cope with new wins.dat format
it doesn't support loading of Internet Group names yet, but then again,
you can't create Internet Group names yet, so it's not a problem!
Andrew Tridgell [Sat, 17 Aug 1996 11:37:44 +0000 (11:37 +0000)]
- added support for Amiga-unix (based on BSD I think)
- changed the order of PROGS and SPROGS in Makefile (SPROGS first)
- another 64 bit cleanup (for INADDR_NONE)
- added paranoia code in DirCacheAdd() to detect looping
- fixed important DirCache flush bug
- rewrote the NetServerEnum code after I found it could return servers
from multiple workgroups at once, and this could cause browsing
havoc. Now a null workgroup query is equivalent to a query for the
servers primary workgroup
- got rid of my_workgroup()
- got rid of "workgroup = *" comment in Makefile. We no longer support
a workgroup of *, users must set the workgroup explicitly
- the wins.dat file was being stored in a different format to what it
was being loaded in - this could cause havoc. fixed.
- uppercase our netbios name and the workgroup name at startup
- if accept fails in main loop when running as a daemon then continue,
don't just exit!
Andrew Tridgell [Fri, 16 Aug 1996 13:03:26 +0000 (13:03 +0000)]
- added the "remote announce" option
- made the lp_string() code able to handle any length string
- got rid of the obsolete lmhosts code, instead users should use
"interfaces" and "remote announce". lmhosts now is just used as a IP
to netbios name map
- cleanup the inet_address() code
Andrew Tridgell [Tue, 13 Aug 1996 12:34:10 +0000 (12:34 +0000)]
added the validchars package written by tino@augsburg.net. This
package automates the building of a "valid chars =" line for any
client DOS based client. This makes setting things up right for
european languages much easier.
Andrew Tridgell [Tue, 13 Aug 1996 08:57:55 +0000 (08:57 +0000)]
- sequent-ptx support from bressler@iftccu.ca.boeing.com (Rick
Bressler)
- machten support from Trevor Strohman (trev@figment.tenon.com)
- added qinfo command to client as part of drag-and-drop printer
support for win95 from David Chappell <chappell@mouse.cc.trincoll.edu>
He also added the "printer driver" option
- use sigblock() on more systems and use sigsetmask(0) instead of
sigunblock() as its more portable. This beats a problem with zombies
on heavilily loaded systems.
- added internals.doc written by David Chappell into the source tree
- get rid of PRINT_COMMAND options from local.h as they are no longer
relevent
- new kanji code from Fujita
- don't set the recursion_available flag on queries in nmbd
- fix a potential bug with pointer subtraction in printing.c
- got rid of error_count code as the real fix (the EOF problem) is now
in
projects / thirdparty / samba.git / log
