Age | Commit message (Collapse) | Author | Files | Lines |
|
Jeremy.
(This used to be commit 03ae657a36c986ba4248b297e64ec7898a1914e6)
|
|
bit file interface for the NT SMB's.
Created a new define, SMB_STRUCT_STAT that currently is
defined to be struct stat - this wil change to a user
defined type containing 64 bit info when the correct
wrappers are written for 64 bit stat(), fstat() and lstat()
calls.
Also changed all sys_xxxx() calls that were previously just
wrappers to the same call prefixed by a dos_to_unix() call
into dos_xxxx() calls. This makes it explicit when a pathname
translation is being done, and when it is not.
Now, all sys_xxx() calls are meant to be wrappers to mask
OS differences, and not silently converting filenames on
the fly.
Jeremy.
(This used to be commit 28aa182dbffaa4ffd86047e608400de4b26e80eb)
|
|
(This used to be commit c53ead308e941baa4524b9ec62d74290ce91503f)
|
|
a table to give it a little border
(This used to be commit d35c6e80988d96049d4624b47ae9ea015edafc59)
|
|
passwords with SWAT
(This used to be commit edcde70108ab643a29f3e0e0cc97609287da6e87)
|
|
I made this change after getting a segv in reply_pipe_close(). The
funny thing was that pipes_open was 1 and Pipes was NULL. That "can't
happen" and suggests that we have a wild pointer somewhere.
I suspect the rpc code, as I was playing with long share names (a
share called "averylongusername") at the time and the logs show lots
of srvsvc operations. I bet there is a buffer in the rpc code
somewhere that is overflowing and trashing bits of the data segment.
(This used to be commit 9fee8c2eb7bd05431cd9bcfbed3804c8ca1ee593)
|
|
include/smb.h: Moved old typedefs of uint8 etc. into include/includes.h where
all the other defines live (changed them from typedefs to
defines).
Other changes : changed from using uint32 to SMB_DEV_T and SMB_INO_T
in preparation for moving to size independed (ie. 64 bit clean) device
and inode access. Stat call wrapper comes next :-).
Jeremy.
(This used to be commit 3d9ec96de5e04e83abafe9c5d980bd39eee856ea)
|
|
also doesn't switch to a registration if a refresh fails, instead the
name is removed! This makes it even more imortant that a Samba WINS
server accepts refresh requests as registration requests if the name
is not registered. I've gone ahead and implemented this.
(This used to be commit 07f825fc9491bb5adac047854e67c63f7757143b)
|
|
the local_master name on the unicast subnet if it is unknown.
(This used to be commit 222b6d90e09288091028f5e0435f1d4a74153f66)
|
|
(This used to be commit 0a5718b0aef29706be81a50f2ac2c5eb4c6fbb32)
|
|
amount of network traffic grows as the square of the number of
workgroups. It probably wouldn't have caused problems but to be safe I
changed the code to use random() to decrease the probability of a
DMB<->DMB sync in proportion to the number of known workgroups. This
keeps the nice browse connectivity while making the traffic rise only
linearly with the number of workgroups.
(This used to be commit 685f4ef2e1f83ab39e91229cf53a61eecb6181eb)
|
|
(This used to be commit 560ed560f46e69255b6437e50ba0865b642c38b1)
|
|
The previous code was strictly correct, but not very practical. self
names were only refreshed every 3 days. I hit a situation where the
Samba WINS server was restarted after deleting wins.dat and didn't
notice some remote subnets (also running Samba). I realised that the
complete database wouldn't have been rebuilt for 3 days, which is way
too long. In order to recover from WINS restarts we need a much
shorter maximum refresh time.
(This used to be commit 1d23dd0912e81ff72695bd043e8e2aee32da18a8)
|
|
(This used to be commit 55333edd2eed33961ced4eb4b6898f5ca9ca1820)
|
|
(This used to be commit 3eade55dc7c842bdc50205c330802d211fae54d3)
|
|
the correct port in environments like ip masq.
(This used to be commit 7d455ee637b6ff70c95845f89d71573ca07b83f3)
|
|
(This used to be commit 03d343ddf5ef672afb3cf1fa65f86eb2c0a48772)
|
|
to work (not a lot of testing yet though).
Now we just need to deal with people worried about having more than
two nmbd processes sometimes. (the async processes are created on
demand for browse sync, so you'll only see more than 2 occasionally)
(This used to be commit a350a54680e4170e2adf571b010ea508e7291780)
|
|
(This used to be commit 425ccf9271ea44879d0940b9d95ae9b8f95aa092)
|
|
then obtained a node status response we need to remember the server
name of the master browser so that other browse clients asking us for
a workgroup list will get a entry for the master of that workgroup.
(This used to be commit 601f995ffbfa0ee477ea628d92b9660d6bdd8cbc)
|
|
are now sorted by the number of common leading bits in the IP address
with the address of the querying host.
(This used to be commit 4460a1bc6aa7666d1c71d32ba73855d6ed32320a)
|
|
(This used to be commit f555a76df696a0625acc16fa365dc048e0c2447d)
|
|
or not timestamping.
(This used to be commit 70ed0ec202c50655e3ba99535b06ad918409051e)
|
|
It now has a line like this:
VERSION 1 251152
the first number is a version #define in nmbd_winsserver.c and will be
used if we ever have to change the format again.
The second number is a hash of the current interfaces setting. It is
used to detect the case where nmbd is restarted on a machine after the
IP of the machine has changed (or the interfaces list has changed in
any way). When that happens we need to discard the old wins.dat cache
or you end up with chaos. This has bitten quite a few people, they
find that when they move a machine it continues using the old IP for
some things for the next week until the wins entries time out!
I've checked, and the old nmbd can handle the new format, although it
does spit out a spurious error message about the VERSION line. So
users can safely run 2.0alpha then switch back to 1.9.18 without
problems.
(This used to be commit c4a8cdc60a5b01894ab2456e77b6d89d4c16a088)
|
|
(This used to be commit 275679db29f2630a16cd5e9f4b5346ac834a484f)
|
|
- changed the hash size to 13 (much smaller than before). This should
make for more efficient shared memory usage as it will lead to less
fragmentation.
(This used to be commit 9c1e4c2dae6323c9a1bd74148d0b45aac61e7c0a)
|
|
memory then frees it. Useful for catching bugs.
(This used to be commit 99782754f79f3795f81cbf57caeb0925f6a66c10)
|
|
to account for padding/alignment issues. Eventually I'd like to find a
way to get rid of this construct altogether as it is a bit error
prone and hard to debug.
also added a new macro:
ZERO_STRUCTP() that takes a pointer to a structure and zeros the
structure. Used in nmbd to zero allocated structures before freeing
them to try to catch bugs a bit faster.
(This used to be commit d3dda65d5177154e2128f50ca3dd34e8e13b6b08)
|
|
areas of memory before freeing them.
While doing this I also found a couple of real bugs. In two places we
were freeing some memory that came from the stack, which leads to
a certain core dump on many sytems.
(This used to be commit c5e5c25c854e54f59291057ba47c4701b5910ebe)
|
|
dumps. It is gone until someone can tell us why its needed and what it
does. (It was only used on OSF1 and core dumped there anyway!)
(This used to be commit a564e4662711d384069757ce3ee5adcadc1b061d)
|
|
(This used to be commit f18b4e95bffb25fa016dd6e7f9a128a9fc69466f)
|
|
correctly. Added new parameter "stat cache size" - set to 50 by default.
I now declare the statcache code officially "open" for business :-).
It gets a hit rate of 97% with a NetBench run and seems to make
using a case insensitive run as efficient as a case sensitive run.
Also tidied up our sys_select usage - added a maxfd parameter and
also added an implementation of select in terms of poll(), for systems
where poll() is much faster. This is disabled by default.
Jeremy.
(This used to be commit 779b924ec1f6c81ff578d22295b20fece698d1fc)
|
|
problem was a buffer overflow in process_node_status_request().
this really points out a general problem is allocating MAX_DGRAM_SIZE
packets on the stack in nmbd. There must be a better way.
(This used to be commit 7db45f169c33e0f3a67ba2260049226992de8bdf)
|
|
Jeremy.
(This used to be commit d6a9087e7e626e6e07503dba38854b136fe70c23)
|
|
NetBench testing though.... :-). Attempts to efficiently reduce
the number of stat() calls Samba does.
Jeremy.
(This used to be commit d0e48a2d8072c3e77a57ac6a2fb5044c05f03b41)
|
|
(This used to be commit cdc38c276d8344127c9edf35f24754c098a90754)
|
|
arrays. This prevents (harmless) warnings from some compilers
(This used to be commit c2da46d1d0820a86e7f77506563cfe7f67b08fee)
|
|
(This used to be commit afe98165a261a740cd0ee88bf041d69721ec72e6)
|
|
needs them and what should really be there.
(This used to be commit c2bf18fb8d2fea5af86ba8cf5d92b1dc076dd7d5)
|
|
(This used to be commit c8b0041570a1fb0be0318f511c6ba746c733cda9)
|
|
rpcsvc/ypclnt.h wants a struct dom_binding. knock-on include effect under
freebsd 2.1.
(This used to be commit d1005b5692e41660a2a7230231b0591aa0773485)
|
|
(This used to be commit e4f81f4a230130f53857e3664e6d83ce252c207e)
|
|
(This used to be commit b32a346a1c50ba40224b8165e08e78867be2d376)
|
|
if ((sbuf->st_mode & S_IWUSR) == 0)
result |= aRONLY;
rather than the very complex user/group permissions checks we do
currently. This is equivalent ot setting "alternate permissions = yes"
in the old code. The change is motivated by three main reasons:
1) it's basically impossible to second guess whether a file is
writeable without trying to open it for writing. ACLs, root squash etc
just make it too hard.
2) setting it not RONLY if the owner can write is closer to what NT
does (eg. look at a cdrom - files are not marked read only).
3) it prevents the silly problem of copying files from a read only
share to a writeable share and then finding you can't write to them as
windows preserves the RONLY flag. Lots of people get bitten by this
when they drag a folder from a Samba drive. It also hurts some install
programs.
I have also added a new flag type for loadparm.c called
FLAG_DEPRECATED which I've set for "alternate permissions". I'll soon
add code to testparm to give a warning about deprecated options.
(This used to be commit c4363a12fdc0be329ca2bfeb1d7b89bfe90031dc)
|
|
SMB_SET_FILE_ALLOCATION_INFO.
Office 97 expects this call to succeed when you tell it you do NT SMB calls.
Jeremy.
(This used to be commit 260e7e27401d863e9f580d4748c577334d9fc9d0)
|
|
have an ASSERT macro defined.
Jeremy.
(This used to be commit dbe6ad014a8b5dcbf17d7cd9865650c2e040d666)
|
|
(This used to be commit e94ab03a4f97ff7ee7bb938245cf054a00f950fd)
|
|
in the rpc code.
(This used to be commit e6ce1c5b5a9f29d8fcbbd23019186ff5c600e795)
|
|
(This used to be commit bd9290c36c9993a994e485da0a81df926f8662e4)
|
|
If the output line is longer than the format buffer could manage, I was
simply ignoring the additional output (that is, *not* copying it to the
format buffer--thus avoiding a buffer overrun). Instead, I now output
the current content followed by " +>\n", and then reset the format buffer.
I have never seen a debug line that exceeds the size of a pstring, but I
might as well handle the situation...just in case.
Chris -)-----
(This used to be commit 8a11d04b7796b256953bf92b2f2ccab763215bc4)
|