12 years agor17040: add notice about svn:externals change master trunk trunk-final-update
Gerald Carter [Fri, 14 Jul 2006 22:07:19 +0000 (22:07 +0000)]
r17040: add notice about svn:externals change

12 years agor17037: link directories to SAMBA_3_0 counterparts
Gerald Carter [Fri, 14 Jul 2006 21:50:43 +0000 (21:50 +0000)]
r17037: link directories to SAMBA_3_0 counterparts

12 years agor17036: gutting trunk to move to svn:externals
Gerald Carter [Fri, 14 Jul 2006 21:48:39 +0000 (21:48 +0000)]
r17036: gutting trunk to move to svn:externals

12 years agor17033: Restructure init_srv_share_info_ctr so that there's only one loop, not a...
Volker Lendecke [Fri, 14 Jul 2006 17:53:45 +0000 (17:53 +0000)]
r17033: Restructure init_srv_share_info_ctr so that there's only one loop, not a dozen
or so. Next step will be to eliminate the explicit snum reference.


12 years agor17030: Partially fix standalone build of tdb directory
Jeremy Allison [Fri, 14 Jul 2006 05:10:55 +0000 (05:10 +0000)]
r17030: Partially fix standalone build of tdb directory
(tdbtool still fails).

12 years agor17029: Sync my changes to trunk until we get this multiple
Gerald Carter [Thu, 13 Jul 2006 23:34:06 +0000 (23:34 +0000)]
r17029: Sync my changes to trunk until we get this multiple
branch discussion over with.

12 years agor17025: Remove one blank line - test checking in to two
Jeremy Allison [Thu, 13 Jul 2006 22:08:10 +0000 (22:08 +0000)]
r17025: Remove one blank line - test checking in to two
branches simultaneously.....

12 years agor17024: Clean up before closing the door behind me
Volker Lendecke [Thu, 13 Jul 2006 20:45:53 +0000 (20:45 +0000)]
r17024: Clean up before closing the door behind me

12 years agor17016: Different and smaller fix for the valid users = username problem.
Volker Lendecke [Thu, 13 Jul 2006 16:28:38 +0000 (16:28 +0000)]
r17016: Different and smaller fix for the valid users = username problem.

If no winbind is around, the best we can do to get the user's token correct is
to ask unix via create_token_from_username. More investigation is needed if
this also fixes the +groupname for unmapped groups problems more cleanly.


12 years agor17011: Back out r17010 after talking to Jerry. Another fix pending...
Volker Lendecke [Thu, 13 Jul 2006 15:37:58 +0000 (15:37 +0000)]
r17011: Back out r17010 after talking to Jerry. Another fix pending...


12 years agor17010: If winbind is not around, add S-1-22-1-<uid> to the user's token.
Volker Lendecke [Thu, 13 Jul 2006 15:03:46 +0000 (15:03 +0000)]
r17010: If winbind is not around, add S-1-22-1-<uid> to the user's token.

See the comment in the patch for the reason.


12 years agor17006: Increment winbindd protocol version number.
Andrew Bartlett [Thu, 13 Jul 2006 09:30:37 +0000 (09:30 +0000)]
r17006: Increment winbindd protocol version number.

Andrew Bartlett

12 years agor17004: Add a new helper mode to ntlm_auth: ntlm-change-password-1
Andrew Bartlett [Thu, 13 Jul 2006 09:28:35 +0000 (09:28 +0000)]
r17004: Add a new helper mode to ntlm_auth: ntlm-change-password-1

This mode proxies pre-calculated blobs from a remote (probably VPN)
client into the domain.  This allows clients to change their password
over a PPTP connection (where they would not be able to connect to
SAMR directly).

Original patch by Alexey Kobozev <>

Andrew Bartlett

12 years agor17002: Fix coverity #303 - possible null deref. Jerry please
Jeremy Allison [Thu, 13 Jul 2006 00:11:20 +0000 (00:11 +0000)]
r17002: Fix coverity #303 - possible null deref. Jerry please
check this is your new code.

12 years agor16999: Allow CIFS POSIX locks to coexist with Windows locks.
Jeremy Allison [Wed, 12 Jul 2006 21:57:52 +0000 (21:57 +0000)]
r16999: Allow CIFS POSIX locks to coexist with Windows locks.
We shouldn't allow this on the same smbd, but the cifsfs
client negotiates POSIX locks then sends Windows ones.
Doh ! Can't fix shipped client code....

12 years agor16998: patch from Paul Griffith <> to fix compile of the test.c...
Gerald Carter [Wed, 12 Jul 2006 21:05:11 +0000 (21:05 +0000)]
r16998: patch from Paul Griffith <> to fix compile of the test.c pdb file

12 years agor16997: Simo's patch (based on repotr from Seth Elssworth of Quest) to try to be...
Gerald Carter [Wed, 12 Jul 2006 21:02:22 +0000 (21:02 +0000)]
r16997: Simo's patch (based on repotr from Seth Elssworth of Quest) to try to be more robust in the precense of more broken /etc/hosts files when determining our fwdn

12 years agor16995: Fix bug #3923, reported by Incorrect type
Jeremy Allison [Wed, 12 Jul 2006 19:33:53 +0000 (19:33 +0000)]
r16995: Fix bug #3923, reported by Incorrect type

12 years agor16993: Fix bug #3922 reported by, correctly
Jeremy Allison [Wed, 12 Jul 2006 19:23:49 +0000 (19:23 +0000)]
r16993: Fix bug #3922 reported by, correctly
look at the return code.

12 years agor16991: Fix bug #3921 spotted by Correctly
Jeremy Allison [Wed, 12 Jul 2006 19:13:03 +0000 (19:13 +0000)]
r16991: Fix bug #3921 spotted by Correctly
obey blocking/non-blocking request for POSIX locks.

12 years agor16986: Fix the logic errors in ref-counting Windows locks.
Jeremy Allison [Wed, 12 Jul 2006 16:31:59 +0000 (16:31 +0000)]
r16986: Fix the logic errors in ref-counting Windows locks.
Hopefully will fix the build farm. Still a few errors
in RAW-LOCK to look at though...

12 years agor16974: Fix subtle logic error in lock ref counting found by
Jeremy Allison [Wed, 12 Jul 2006 06:56:47 +0000 (06:56 +0000)]
r16974: Fix subtle logic error in lock ref counting found by
cifsfs client code.

12 years agor16970: Ensure we use the correct separator for pathnames
Jeremy Allison [Wed, 12 Jul 2006 03:20:52 +0000 (03:20 +0000)]
r16970: Ensure we use the correct separator for pathnames
in POSIX mode (clitar needs fixing too). Add test
posix lock/unlock commands.

12 years agor16969: The function parse_processed_dfs_path() is dependent on the
Jeremy Allison [Wed, 12 Jul 2006 03:02:35 +0000 (03:02 +0000)]
r16969: The function parse_processed_dfs_path() is dependent on the
fact that check_path_syntax() will convert '\\' characters to '/'.
When POSIX pathnames have been selected this doesn't happen, so we
must look for the unaltered separator of '\\' instead of the modified '/'.
Stevef please check this with the CIFSFS MS-DFS code !

12 years agor16963: Add a few utility fns into client. Allow POSIX capabilities
Jeremy Allison [Wed, 12 Jul 2006 00:21:16 +0000 (00:21 +0000)]
r16963: Add a few utility fns into client. Allow POSIX capabilities
to be selected.

12 years agor16960: Some warnings from host "opi"
Volker Lendecke [Tue, 11 Jul 2006 21:23:44 +0000 (21:23 +0000)]
r16960: Some warnings from host "opi"

12 years agor16959: get_share_security does not need snum, activate RPC-SAMBA3-SRVSVC
Volker Lendecke [Tue, 11 Jul 2006 21:10:44 +0000 (21:10 +0000)]
r16959: get_share_security does not need snum, activate RPC-SAMBA3-SRVSVC

12 years agor16957: fix cut-n-paste error. The check for 'if (\!salt)' make no sense when fetchi...
Gerald Carter [Tue, 11 Jul 2006 21:09:13 +0000 (21:09 +0000)]
r16957: fix cut-n-paste error.  The check for 'if (\!salt)' make no sense when fetching the DES salting principal

12 years agor16955: Fix an uninitialized var -- Jerry, please check.
Volker Lendecke [Tue, 11 Jul 2006 20:50:50 +0000 (20:50 +0000)]
r16955: Fix an uninitialized var -- Jerry, please check.

12 years agor16954: Volker reminded me we already have code to do this check.
Gerald Carter [Tue, 11 Jul 2006 20:31:13 +0000 (20:31 +0000)]
r16954: Volker reminded me we already have code to do this check.
Reuse can_create() to prevent renameing a group to
an existing user or group.

12 years agor16953: Don't allow groups to be renamed to an existing user or other group
Gerald Carter [Tue, 11 Jul 2006 20:02:22 +0000 (20:02 +0000)]
r16953: Don't allow groups to be renamed to an existing user or other group

12 years agor16952: New derive DES salt code and Krb5 keytab generation
Gerald Carter [Tue, 11 Jul 2006 18:45:22 +0000 (18:45 +0000)]
r16952: New derive DES salt code and Krb5 keytab generation

Major points of interest:

* Figure the DES salt based on the domain functional level
  and UPN (if present and applicable)
* Only deal with the DES-CBC-MD5, DES-CBC-CRC, and RC4-HMAC
* Remove all the case permutations in the keytab entry
  generation (to be partially re-added only if necessary).
* Generate keytab entries based on the existing SPN values
  in AD

The resulting keytab looks like:

ktutil:  list -e
slot KVNO Principal
---- ---- ---------------------------------------------------------------------
   1    6 host/ (DES cbc mode with CRC-32)
   2    6 host/ (DES cbc mode with RSA-MD5)
   3    6 host/ (ArcFour with HMAC/md5)
   4    6           host/suse10@COLOR.PLAINJOE.ORG (DES cbc mode with CRC-32)
   5    6           host/suse10@COLOR.PLAINJOE.ORG (DES cbc mode with RSA-MD5)
   6    6           host/suse10@COLOR.PLAINJOE.ORG (ArcFour with HMAC/md5)
   7    6               suse10$@COLOR.PLAINJOE.ORG (DES cbc mode with CRC-32)
   8    6               suse10$@COLOR.PLAINJOE.ORG (DES cbc mode with RSA-MD5)
   9    6               suse10$@COLOR.PLAINJOE.ORG (ArcFour with HMAC/md5)

The list entries are the two basic SPN values (host/NetBIOSName & host/dNSHostName)
and the sAMAccountName value.  The UPN will be added as well if the machine has
one. This fixes 'kinit -k'.

Tested keytab using mod_auth_krb and MIT's telnet.  ads_verify_ticket()
continues to work with RC4-HMAC and DES keys.

12 years agor16946: Fix warning with profile separator when profiles not
Jeremy Allison [Tue, 11 Jul 2006 18:03:00 +0000 (18:03 +0000)]
r16946: Fix warning with profile separator when profiles not
being used.

12 years agor16944: Update from 3.0 to get ready for sync.
Jeremy Allison [Tue, 11 Jul 2006 17:19:41 +0000 (17:19 +0000)]
r16944: Update from 3.0 to get ready for sync.

12 years agor16942: Merge back before merging forward - sync up packaging.
Jeremy Allison [Tue, 11 Jul 2006 16:48:59 +0000 (16:48 +0000)]
r16942: Merge back before merging forward - sync up packaging.

12 years agor16941: Fix crash bug when the pam conversation receives an empty token.
Günther Deschner [Tue, 11 Jul 2006 10:39:32 +0000 (10:39 +0000)]
r16941: Fix crash bug when the pam conversation receives an empty token.
Thanks to Bjoern Jacke for the report and test-case.


12 years agor16940: libnscd sets errno, use that to display error message.
Günther Deschner [Tue, 11 Jul 2006 09:59:22 +0000 (09:59 +0000)]
r16940: libnscd sets errno, use that to display error message.


12 years agor16939: Still clear the winbind_cache.tdb when offline logons are not enabled.
Günther Deschner [Tue, 11 Jul 2006 09:22:55 +0000 (09:22 +0000)]
r16939: Still clear the winbind_cache.tdb when offline logons are not enabled.


12 years agor16935: Fix the build by removing the last DEVELOPER ifdefs.
Jeremy Allison [Tue, 11 Jul 2006 03:14:59 +0000 (03:14 +0000)]
r16935: Fix the build by removing the last DEVELOPER ifdefs.

12 years agor16931: Change the underlying lock db to use 32-bit PIDs
Jeremy Allison [Tue, 11 Jul 2006 00:24:29 +0000 (00:24 +0000)]
r16931: Change the underlying lock db to use 32-bit PIDs
not 16. This allows POSIX locks to be more easily
used. Get rid of one more use of a global variable,
global_smbpid. We can get rid of this if we remove
its use in smbd/open.c

12 years agor16930: Turn CIFS POSIX locks on (remove #if defined(DEVELOPER))
Jeremy Allison [Mon, 10 Jul 2006 23:54:00 +0000 (23:54 +0000)]
r16930: Turn CIFS POSIX locks on (remove #if defined(DEVELOPER))
clause around them.

12 years agor16929: Remove old unused profile fn (causing a warning).
Jeremy Allison [Mon, 10 Jul 2006 23:45:39 +0000 (23:45 +0000)]
r16929: Remove old unused profile fn (causing a warning).

12 years agor16928: Fix the ZERO_ZERO case back to working again. Fix compiler
Jeremy Allison [Mon, 10 Jul 2006 21:33:52 +0000 (21:33 +0000)]
r16928: Fix the ZERO_ZERO case back to working again. Fix compiler
warning pointed out by Volker (thanks!).

12 years agor16924: Rewrote the POSIX lock layer (and parts of the brlock
Jeremy Allison [Mon, 10 Jul 2006 19:06:30 +0000 (19:06 +0000)]
r16924: Rewrote the POSIX lock layer (and parts of the brlock
layer) to try and provide Windows and POSIX CIFS locks
mapped down onto system POSIX locks. This code passes valgrind
and also all the Samba4 "BASIC" lock tests. I will be doing
further tests on the mapping down to POSIX locks. The ultimate
test will be when multiple CIFSFS clients are doing tdbtorture
on a CIFS mounted drive and everything passes :-).
Stevef - please check this code out and test with your POSIX
lock tests - let me know about any failures.

12 years agor16919: Enhanced error message: Show the complete target pid
Volker Lendecke [Mon, 10 Jul 2006 15:18:54 +0000 (15:18 +0000)]
r16919: Enhanced error message: Show the complete target pid

12 years agor16905: Take snum from print_cache_flush
Volker Lendecke [Sun, 9 Jul 2006 20:56:38 +0000 (20:56 +0000)]
r16905: Take snum from print_cache_flush

12 years agor16904: Take snum from construct_dev_mode
Volker Lendecke [Sun, 9 Jul 2006 20:45:44 +0000 (20:45 +0000)]
r16904: Take snum from construct_dev_mode

12 years agor16903: We don't need to store the context for Windows POSIX
Jeremy Allison [Sun, 9 Jul 2006 19:31:46 +0000 (19:31 +0000)]
r16903: We don't need to store the context for Windows POSIX
lock mapping.

12 years agor16902: This converts lp_mangled_map, lp_manglednames and lp_magicchar to take a
Volker Lendecke [Sun, 9 Jul 2006 18:09:22 +0000 (18:09 +0000)]
r16902: This converts lp_mangled_map, lp_manglednames and lp_magicchar to take a
struct share_params instead of a snum as input.

It is an example of how the a bit larger changes to get rid of snum might look
like, and it demonstrates how the SNUM(conn) will be converted to conn->params.

I'd really appreciate input on what I'm doing here, the changes will be rather
localized but in the end svn blame might point at me for a *lot* of code that
I'm not really responsible for ;-))


12 years agor16901: print_job_set_place and print_job_set_name do not use snum
Volker Lendecke [Sun, 9 Jul 2006 14:03:34 +0000 (14:03 +0000)]
r16901: print_job_set_place and print_job_set_name do not use snum

12 years agor16899: Convert print_time_access_check to servicename
Volker Lendecke [Sun, 9 Jul 2006 13:45:59 +0000 (13:45 +0000)]
r16899: Convert print_time_access_check to servicename

12 years agor16898: is_owner really needs the share name, not the snum...
Volker Lendecke [Sun, 9 Jul 2006 13:41:56 +0000 (13:41 +0000)]
r16898: is_owner really needs the share name, not the snum...

There are a lot of functions in printing like this. Jerry, if it is ok I'll
start lifting the snum to one level higher step by step.


12 years agor16894: Introduce a function share_defined.
Volker Lendecke [Sun, 9 Jul 2006 12:54:45 +0000 (12:54 +0000)]
r16894: Introduce a function share_defined.

Jerry, while looking for possible usages of this function I came across
srv_spoolss_nt.c:334. To me this looks like reversed logic.


12 years agor16893: Make sub_advanced not use snum but the service name.
Volker Lendecke [Sun, 9 Jul 2006 12:30:48 +0000 (12:30 +0000)]
r16893: Make sub_advanced not use snum but the service name.


12 years agor16892: When we want more flexibility in configuring Samba, I think that we need to
Volker Lendecke [Sun, 9 Jul 2006 12:17:15 +0000 (12:17 +0000)]
r16892: When we want more flexibility in configuring Samba, I think that we need to
get rid of the global share array in loadparm.c. Even right now with
usershares this is a little awkward.

Step zero in a looong way there: This encapsulates the service number in
connection_struct into a 'struct share_params'. I want to get rid of the use
of anything like an index number for a share outside of loadparm.c. Inside
loadparm.c it can be organized as an array if necessary, but the rest of Samba
should only see share names and struct share_params where we can then hide the
current share definitions or some dynamic backend or whatever.

Does this sound like a reasonable plan?


12 years agor16888: Fix the calculation if an overlap occurred in the upper
Jeremy Allison [Sun, 9 Jul 2006 04:24:49 +0000 (04:24 +0000)]
r16888: Fix the calculation if an overlap occurred in the upper
layer when doing unlocks. Simplified the close handling
in the POSIX lock case (we know the system already does
the right thing). Get ready to add the unlock semantics
for POSIX (hard hard hard !).

12 years agor16887: Disable the downcall until I've finished all the
Jeremy Allison [Sun, 9 Jul 2006 02:38:21 +0000 (02:38 +0000)]
r16887: Disable the downcall until I've finished all the

12 years agor16886: Add the mapping from CIFS posix locks to system
Jeremy Allison [Sun, 9 Jul 2006 01:59:34 +0000 (01:59 +0000)]
r16886: Add the mapping from CIFS posix locks to system
POSIX locks. That's the easy part... Next will
come adding the unlock !

12 years agor16885: Move locking definitions into their own header file locking.h
Jeremy Allison [Sun, 9 Jul 2006 00:52:21 +0000 (00:52 +0000)]
r16885: Move locking definitions into their own header file locking.h
Pass down the upper level locking data to the system layer,
will avoid having to re-walk the lock list (which the upper
layer has already done).

12 years agor16884: Do the lock merging more agressively for CIFS POSIX
Jeremy Allison [Sat, 8 Jul 2006 20:49:00 +0000 (20:49 +0000)]
r16884: Do the lock merging more agressively for CIFS POSIX
locks. There were cases where we weren't merging
together two adjacent locks of the same type. This
should reduce the number of locks records in this

12 years agor16883: Restructure the code in locking/posix.c to get ready to
Jeremy Allison [Sat, 8 Jul 2006 20:16:35 +0000 (20:16 +0000)]
r16883: Restructure the code in locking/posix.c to get ready to
add the mapping from CIFS POSIX locks to system POSIX
locks. No difference in the lock tests in Samba4 lock
tests for Windows locks. Next will come the code to
map the CIFS POSIX locks onto system locks. I will
add Samba4 torture test code to test this.

12 years agor16882: Fix warning on longlong if we have it.
Jeremy Allison [Sat, 8 Jul 2006 19:59:22 +0000 (19:59 +0000)]
r16882: Fix warning on longlong if we have it.

12 years agor16881: Some C++ warnings
Volker Lendecke [Sat, 8 Jul 2006 19:01:10 +0000 (19:01 +0000)]
r16881: Some C++ warnings

12 years agor16879: Debug msg fix from Kai Blin
Volker Lendecke [Sat, 8 Jul 2006 17:33:19 +0000 (17:33 +0000)]
r16879: Debug msg fix from Kai Blin

12 years agor16867: Merge the fix for bug 3915
Volker Lendecke [Fri, 7 Jul 2006 19:34:54 +0000 (19:34 +0000)]
r16867: Merge the fix for bug 3915

12 years agor16862: Reverting accidential changes in ads_try_connect() from previous commit.
Günther Deschner [Fri, 7 Jul 2006 11:59:19 +0000 (11:59 +0000)]
r16862: Reverting accidential changes in ads_try_connect() from previous commit.


12 years agor16861: Fixing crash bug when passing no domain/realm name to the CLDAP request.
Günther Deschner [Fri, 7 Jul 2006 11:43:47 +0000 (11:43 +0000)]
r16861: Fixing crash bug when passing no domain/realm name to the CLDAP request.


12 years agor16859: Don't use LL for long long literals in str_conv_size().
James Peach [Fri, 7 Jul 2006 06:30:00 +0000 (06:30 +0000)]
r16859: Don't use LL for long long literals in str_conv_size().

12 years agor16857: Don't use LL - compiler might not have it.
Jeremy Allison [Fri, 7 Jul 2006 05:11:33 +0000 (05:11 +0000)]
r16857: Don't use LL - compiler might not have it.

12 years agor16856: Fix up sample module compile after jpeach's changes.
Jeremy Allison [Fri, 7 Jul 2006 05:05:33 +0000 (05:05 +0000)]
r16856: Fix up sample module compile after jpeach's changes.
skel_opaque won't work because the vfs_default functions
are static now, but still is useful as an exmaple.

12 years agor16855: Move conv_str_size() to common code.
James Peach [Fri, 7 Jul 2006 04:56:33 +0000 (04:56 +0000)]
r16855: Move conv_str_size() to common code.

12 years agor16853: New module to flush dirty file data at regular intervals.
James Peach [Fri, 7 Jul 2006 03:45:02 +0000 (03:45 +0000)]
r16853: New module to flush dirty file data at regular intervals.

12 years agor16849: New VFS module to preallocate file extents at creation time.
James Peach [Fri, 7 Jul 2006 02:13:46 +0000 (02:13 +0000)]
r16849: New VFS module to preallocate file extents at creation time.

12 years agor16848: Add an fsp parameter to the SMB_VFS_OPEN operation. This makes it
James Peach [Fri, 7 Jul 2006 02:09:49 +0000 (02:09 +0000)]
r16848: Add an fsp parameter to the SMB_VFS_OPEN operation. This makes it
possible for modules to add per-fsp data on the open path.

12 years agor16845: Properly report the error during join when the set password fails
Gerald Carter [Fri, 7 Jul 2006 00:20:55 +0000 (00:20 +0000)]
r16845: Properly report the error during join when the set password fails

12 years agor16844: Our original error code mapping for NTSTATUS -> DOS
Jeremy Allison [Thu, 6 Jul 2006 21:28:29 +0000 (21:28 +0000)]
r16844: Our original error code mapping for NTSTATUS -> DOS
errors is correct, but not for chkpath. Special case

12 years agor16841: Fix a bug found by a 64-bit build warning
Volker Lendecke [Thu, 6 Jul 2006 20:10:23 +0000 (20:10 +0000)]
r16841: Fix a bug found by a 64-bit build warning

12 years agor16839: Next step getting ancient Linux to build....
Volker Lendecke [Thu, 6 Jul 2006 18:48:22 +0000 (18:48 +0000)]
r16839: Next step getting ancient Linux to build....

Steve, you might want to take a look.


12 years agor16837: Finally found a Linux version that shows bug 3841. SuSE 7.0 does it :-)
Volker Lendecke [Thu, 6 Jul 2006 16:59:43 +0000 (16:59 +0000)]
r16837: Finally found a Linux version that shows bug 3841. SuSE 7.0 does it :-)

Attempt to fix it.


12 years agor16836: When receiving a CLDAP reply make sure that we always store the correct
Günther Deschner [Thu, 6 Jul 2006 13:38:41 +0000 (13:38 +0000)]
r16836: When receiving a CLDAP reply make sure that we always store the correct
netbios domain name in server affinity cache.


12 years agor16823: Allow to call wbinfo --domain-info="" or --domain-info="." to get domain
Günther Deschner [Wed, 5 Jul 2006 15:29:31 +0000 (15:29 +0000)]
r16823: Allow to call wbinfo --domain-info="" or --domain-info="." to get domain
info for our own domain.


12 years agor16822: Let's have a name for when you do not want to pass in any flags, it is more...
Simo Sorce [Wed, 5 Jul 2006 14:55:21 +0000 (14:55 +0000)]
r16822: Let's have a name for when you do not want to pass in any flags, it is more understandable than passing just "0"

12 years agor16817: Fixing build of idmap_ad and idmap_rid.
Günther Deschner [Wed, 5 Jul 2006 11:43:49 +0000 (11:43 +0000)]
r16817: Fixing build of idmap_ad and idmap_rid.


12 years agor16816: Reverting some hunks from Volkers r16803, fixing cached credentials and
Günther Deschner [Wed, 5 Jul 2006 09:10:56 +0000 (09:10 +0000)]
r16816: Reverting some hunks from Volkers r16803, fixing cached credentials and
removal of oldest creds when reaching MAX_CACHED_LOGINS.


12 years agor16814: New VFS module "cacheprime". This is useful for equipment where
James Peach [Wed, 5 Jul 2006 05:55:03 +0000 (05:55 +0000)]
r16814: New VFS module "cacheprime". This is useful for equipment where
it is important to minimise the number of actual disk reads or
to make sure that actual disk reads are of a particular size.

12 years agor16813: Rewrote much of the idmap code. This design is *shit*,
Jeremy Allison [Wed, 5 Jul 2006 03:40:40 +0000 (03:40 +0000)]
r16813: Rewrote much of the idmap code. This design is *shit*,
and needs fixing. We must refactor out the cache code
and separate it from the idmap_tdb so they are two
separate pointer tables. They may call down into the
same underlying functions but the MUST NOT BE ALLOWED
to continue to mix up allocation and cachine. I'll know
the design is correct when the flags field dissapears
from the backend interface. Having said that this code
is now a lot cleaner and easier to understand. Expect
more hard words (and work) in this area.

Anyone wanting to work on this code please see me
first ! ("You behind the bike sheds, stand still
laddie !" :-)


12 years agor16812: Refactor shared memory profiling to make the call counts and cumulative
James Peach [Wed, 5 Jul 2006 03:11:12 +0000 (03:11 +0000)]
r16812: Refactor shared memory profiling to make the call counts and cumulative
times two arrays. This makes it a lot easier to write succinct code to
deal with these.

Add compatibility macros so that existing references continue to work.

Add a new idle counter to track how much time is spent waiting for the
client to send a packet.

Bump the shared memory version number.

Add a --profile-rates option to smbstatus to show the rate-converted
times and counts from the shared memory metrics. Sample output:

smbd_idle 10430/sec (15.93%)             syscall_opendir 2030/sec (0.72%)
syscall_readdir 25440/sec (1.29%)        syscall_telldir 19350/sec (0.22%)
syscall_closedir 2030/sec (0.18%)        syscall_open 1323/sec (23.85%)
syscall_close 1316/sec (0.55%)           syscall_pread 2873/sec (11.97%)
syscall_pwrite 915/sec (10.75%)          syscall_rename 81/sec (0.19%)
syscall_stat 15250/sec (3.68%)           syscall_fstat 1154/sec (0.07%)
syscall_lstat 333/sec (0.07%)            syscall_unlink 333/sec (32.80%)
syscall_fcntl_lock 12/sec (0.00%)        syscall_fcntl_getlock 658/sec (0.03%)
SMBclose 1316/sec (50.74%)               SMBflush 128/sec (0.00%)
SMBunlink 333/sec (47.89%)               SMBmv 81/sec (12.53%)
SMBlockingX 39/sec (0.07%)               SMBreadX 2873/sec (43.11%)
SMBwriteX 915/sec (11.06%)               SMBtrans2 2944/sec (82.48%)
SMBntcreateX 1774/sec (33.91%)           Trans2_findfirst 595/sec (12.15%)
Trans2_qfsinfo 289/sec (27.62%)          Trans2_qpathinfo 1897/sec (26.71%)
Trans2_setpathinfo 163/sec (0.77%)       fchmod_acl 354/sec (0.02%)

12 years agor16803: Fixes from IRIX, some C++ warnings
Volker Lendecke [Tue, 4 Jul 2006 18:24:36 +0000 (18:24 +0000)]
r16803: Fixes from IRIX, some C++ warnings

12 years agor16802: Janitor for Simo - fixing typo.
Jeremy Allison [Tue, 4 Jul 2006 17:14:37 +0000 (17:14 +0000)]
r16802: Janitor for Simo - fixing typo.

12 years agor16799: Fix remote smbd crash bug by removing half-implemented info level 4
Günther Deschner [Tue, 4 Jul 2006 15:29:21 +0000 (15:29 +0000)]
r16799: Fix remote smbd crash bug by removing half-implemented info level 4


12 years agor16797: Add msdfs proxy junctions in the netdfs rpc enumeration.
Günther Deschner [Tue, 4 Jul 2006 13:49:17 +0000 (13:49 +0000)]
r16797: Add msdfs proxy junctions in the netdfs rpc enumeration.


12 years agor16790: Fix memleak.
Günther Deschner [Mon, 3 Jul 2006 22:27:10 +0000 (22:27 +0000)]
r16790: Fix memleak.


12 years agor16788: Fix bug #3909, when using ea's getting a directory tries to
Jeremy Allison [Mon, 3 Jul 2006 21:07:26 +0000 (21:07 +0000)]
r16788: Fix bug #3909, when using ea's getting a directory tries to
read ea's from an msdfs link. Stop it from doing that.
Jerry please merge to 3.0.23.

12 years agor16787: remove fstrings and use talloc
Simo Sorce [Mon, 3 Jul 2006 18:32:12 +0000 (18:32 +0000)]
r16787: remove fstrings and use talloc

12 years agor16785: BUG 3908: Fix rpc bin authentication failure which broke user password changes
Gerald Carter [Mon, 3 Jul 2006 16:12:16 +0000 (16:12 +0000)]
r16785: BUG 3908: Fix rpc bin authentication failure which broke user password changes

Jeremy, please review.

12 years agor16782: Activate RAW-SAMBA3BADPATH in the build farm
Volker Lendecke [Mon, 3 Jul 2006 13:51:50 +0000 (13:51 +0000)]
r16782: Activate RAW-SAMBA3BADPATH in the build farm

12 years agor16780: The tdb_chainunlock is happening in the destructor.
Volker Lendecke [Mon, 3 Jul 2006 13:36:11 +0000 (13:36 +0000)]
r16780: The tdb_chainunlock is happening in the destructor.

Two C++ warnings.


12 years agor16775: Attempt to get HP/UX host "gwen" into the build farm
Volker Lendecke [Mon, 3 Jul 2006 06:58:39 +0000 (06:58 +0000)]
r16775: Attempt to get HP/UX host "gwen" into the build farm


12 years agor16767: Forgotten file
Volker Lendecke [Sun, 2 Jul 2006 22:05:02 +0000 (22:05 +0000)]
r16767: Forgotten file

12 years agor16766: A warning found by RHEL3. This might actually be 3.0.23 code, maybe there are
Volker Lendecke [Sun, 2 Jul 2006 22:04:29 +0000 (22:04 +0000)]
r16766: A warning found by RHEL3. This might actually be 3.0.23 code, maybe there are
vasprintf implementations that don't like a NULL format.