15 years agor23174: merge text
Jelmer Vernooij [Mon, 28 May 2007 14:24:36 +0000 (14:24 +0000)]
r23174: merge text

15 years agor23158: Set version to 1.0.1.
Jelmer Vernooij [Sat, 26 May 2007 16:56:54 +0000 (16:56 +0000)]
r23158: Set version to 1.0.1.

15 years agor23157: Add news file for talloc.
Jelmer Vernooij [Sat, 26 May 2007 16:55:08 +0000 (16:55 +0000)]
r23157: Add news file for talloc.

15 years agor23153: a first cut at a fix for the dgram flood problem that volker
Andrew Tridgell [Sat, 26 May 2007 08:47:27 +0000 (08:47 +0000)]
r23153: a first cut at a fix for the dgram flood problem that volker
found. Not sure this is the best solution, but it should work.

15 years agor23152: fixed some async retry issues in openbench
Andrew Tridgell [Sat, 26 May 2007 03:24:08 +0000 (03:24 +0000)]
r23152: fixed some async retry issues in openbench

15 years agor23149: Fix up the trusted domain lookup code to use the new structures.
Andrew Bartlett [Sat, 26 May 2007 00:25:22 +0000 (00:25 +0000)]
r23149: Fix up the trusted domain lookup code to use the new structures.

Andrew Bartlett

15 years agor23146: fix the build (temporary)
Stefan Metzmacher [Fri, 25 May 2007 13:30:49 +0000 (13:30 +0000)]
r23146: fix the build (temporary)

abartlet: please fix this correct!


15 years agor23145: try to fix the build on solaris
Stefan Metzmacher [Fri, 25 May 2007 13:11:17 +0000 (13:11 +0000)]
r23145: try to fix the build on solaris


15 years agor23144: splitup RAW-OPLOCK into small units
Stefan Metzmacher [Fri, 25 May 2007 13:03:33 +0000 (13:03 +0000)]
r23144: splitup RAW-OPLOCK into small units


15 years agor23143: error found by valgrind
Andrew Tridgell [Fri, 25 May 2007 12:35:03 +0000 (12:35 +0000)]
r23143: error found by valgrind

15 years agor23142: added error checking and reconnect on echo replies
Andrew Tridgell [Fri, 25 May 2007 12:21:29 +0000 (12:21 +0000)]
r23142: added error checking and reconnect on echo replies

15 years agor23141: Use the finddcs() library call rather than a winbind-specific version.
Andrew Bartlett [Fri, 25 May 2007 11:59:52 +0000 (11:59 +0000)]
r23141: Use the finddcs() library call rather than a winbind-specific version.

(I created finddcs() from the winbind code a while back, so this
finishes that work)

Andrew Bartlett

15 years agor23140: send send echos while disconnected
Andrew Tridgell [Fri, 25 May 2007 11:39:43 +0000 (11:39 +0000)]
r23140: send send echos while disconnected

15 years agor23139: use echo operations once a second in lockbench and openbench to ensure
Andrew Tridgell [Fri, 25 May 2007 10:43:06 +0000 (10:43 +0000)]
r23139: use echo operations once a second in lockbench and openbench to ensure
we detect IP takeover on servers

15 years agor23138: added a raw interface for SMBecho operations
Andrew Tridgell [Fri, 25 May 2007 10:42:29 +0000 (10:42 +0000)]
r23138: added a raw interface for SMBecho operations

15 years agor23137: make it possible to skip the pidl tests
Stefan Metzmacher [Fri, 25 May 2007 09:15:09 +0000 (09:15 +0000)]
r23137: make it possible to skip the pidl tests


15 years agor23136: Set the event context onto the credentials in more places.
Andrew Bartlett [Fri, 25 May 2007 08:44:33 +0000 (08:44 +0000)]
r23136: Set the event context onto the credentials in more places.

This helps ensure that the kerberos code uses the right event context.

Andrew Bartlett

15 years agor23135: inet_ntop and inet_pton are now provided by libreplace
Stefan Metzmacher [Fri, 25 May 2007 08:14:15 +0000 (08:14 +0000)]
r23135: inet_ntop and inet_pton are now provided by libreplace


15 years agor23134: Set the event context onto the cli_credentials.
Andrew Bartlett [Fri, 25 May 2007 08:05:14 +0000 (08:05 +0000)]
r23134: Set the event context onto the cli_credentials.

Andrew Bartlett

15 years agor23133: I felt pity on Kai, as he starts work on winbind in Samba4, so I
Andrew Bartlett [Fri, 25 May 2007 08:04:39 +0000 (08:04 +0000)]
r23133: I felt pity on Kai, as he starts work on winbind in Samba4, so I
decided to clean it up a little.

We now use SPNEGO for authentication if possible, and common routines
shared with the rest of the librpc codebase.  Rather than make a
connection to IPC$, then connect the pipes to it, we instead have the
lsa and samr pipes as 'secondary connections'.

Andrew Bartlett

15 years agor23132: Resolve an issue where we would use the ccache after we free()ed it.
Andrew Bartlett [Fri, 25 May 2007 05:19:02 +0000 (05:19 +0000)]
r23132: Resolve an issue where we would use the ccache after we free()ed it.

The problem was, we would set the ccache, then invalidate it as we set
details from it (like the principal name from the ccache).

Instead, set the ccache onto the credentials structure after we are
done processing it.

Andrew Bartlett

15 years agor23129: Merge from 3_0:
Günther Deschner [Thu, 24 May 2007 23:38:46 +0000 (23:38 +0000)]
r23129: Merge from 3_0:

* netr_DsRGetDCName_flags, netr_DsRGetDCNameInfo_AddressType and netr_DsR_DcFlags
* the mask in netr_DsRGetDCNameEx2 turns out to be samr_AcctFlags


15 years agor23126: this was sitting on my laptop for a while...
Rafal Szczesniak [Thu, 24 May 2007 21:51:25 +0000 (21:51 +0000)]
r23126: this was sitting on my laptop for a while...

change the way the ejs object is being created and return listing
context (with status) rather than collecting all entries gathered
from libnet call.


15 years agor23125: add host manager subcontext function.
Rafal Szczesniak [Thu, 24 May 2007 21:45:29 +0000 (21:45 +0000)]
r23125: add host manager subcontext function.


15 years agor23124: add host subcontext providing (at the moment) list of domains
Rafal Szczesniak [Thu, 24 May 2007 21:44:27 +0000 (21:44 +0000)]
r23124: add host subcontext providing (at the moment) list of domains
hosted on the server.


15 years agor23123: Fix two utterly simply typos that broke the whole build farm...
Michael Adam [Thu, 24 May 2007 21:29:15 +0000 (21:29 +0000)]
r23123: Fix two utterly simply typos that broke the whole build farm...


15 years agor23118: Make inet_pton and inet_ntop available through lib/replace.
Michael Adam [Thu, 24 May 2007 15:51:01 +0000 (15:51 +0000)]
r23118: Make inet_pton and inet_ntop available through lib/replace.
The code for inet_pton.c and inet_ntop.c is taken from
the rsync code.

I will put this into 3_0 and 3_0_26 too, but let's sort
the readahead issue in lib/replace of 3_0 first...


15 years agor23115: the attribute array needs to be NULL terminated...
Stefan Metzmacher [Thu, 24 May 2007 12:40:46 +0000 (12:40 +0000)]
r23115: the attribute array needs to be NULL terminated...
hopefully fix the crash on some hosts in the farm...


15 years agor23114: we fail the new BASE-DELETE/deltest20a|b tests
Stefan Metzmacher [Thu, 24 May 2007 12:02:34 +0000 (12:02 +0000)]
r23114: we fail the new BASE-DELETE/deltest20a|b tests


15 years agor23111: nicer ./configure --help output
Stefan Metzmacher [Thu, 24 May 2007 09:28:49 +0000 (09:28 +0000)]
r23111: nicer ./configure --help output


15 years agor23110: add --disable-shared option, very useful for the build-farm...
Stefan Metzmacher [Thu, 24 May 2007 09:28:05 +0000 (09:28 +0000)]
r23110: add --disable-shared option, very useful for the build-farm...


15 years agor23108: Add defines of RTLD_NOW and RTLD_GLOBAL to replace.h.
Michael Adam [Thu, 24 May 2007 07:15:36 +0000 (07:15 +0000)]
r23108: Add defines of RTLD_NOW and RTLD_GLOBAL to replace.h.
Fixing build of ldb for instance on some systems without dlfcn.h.

15 years agor23104: Add comment explaining the new test.
Jeremy Allison [Wed, 23 May 2007 23:14:48 +0000 (23:14 +0000)]
r23104: Add comment explaining the new test.

15 years agor23103: Add a torture test for the insane semantics Steve's cifsfs client
Jeremy Allison [Wed, 23 May 2007 23:11:16 +0000 (23:11 +0000)]
r23103: Add a torture test for the insane semantics Steve's cifsfs client
uses so we don't regress :-).

15 years agor23099: New interesting delete on close tests :-).
Jeremy Allison [Wed, 23 May 2007 21:14:05 +0000 (21:14 +0000)]
r23099: New interesting delete on close tests :-).
This will break all delete tests on the build farm
until I check in my server patch.

15 years agor23090: a test showing two netlogon pipes open at once, using the LogonEx ops
Andrew Tridgell [Wed, 23 May 2007 07:44:51 +0000 (07:44 +0000)]
r23090: a test showing two netlogon pipes open at once, using the LogonEx ops

15 years agor23089: This is upstream now, so we don't need to add it manually.
Andrew Bartlett [Wed, 23 May 2007 02:41:31 +0000 (02:41 +0000)]
r23089: This is upstream now, so we don't need to add it manually.

Andrew Bartlett

15 years agor23071: print the command on failure only
Stefan Metzmacher [Tue, 22 May 2007 12:45:58 +0000 (12:45 +0000)]
r23071: print the command on failure only


15 years agor23069: print out the command, to find out the problem on host 'tridge'
Stefan Metzmacher [Tue, 22 May 2007 11:23:36 +0000 (11:23 +0000)]
r23069: print out the command, to find out the problem on host 'tridge'


15 years agor23067: use 'const union smb_search_data *file' also in the server code to get rid
Stefan Metzmacher [Tue, 22 May 2007 09:25:58 +0000 (09:25 +0000)]
r23067: use 'const union smb_search_data *file' also in the server code to get rid
of compiler warnings in the cifs backend


15 years agor23066: don't ignore errors
Stefan Metzmacher [Tue, 22 May 2007 09:02:16 +0000 (09:02 +0000)]
r23066: don't ignore errors


15 years agor23065: revert local includes to prevent security problems (at least
Rafal Szczesniak [Tue, 22 May 2007 06:34:14 +0000 (06:34 +0000)]
r23065: revert local includes to prevent security problems (at least


15 years agor23064: Clarify comment and indent
Andrew Bartlett [Tue, 22 May 2007 05:22:18 +0000 (05:22 +0000)]
r23064: Clarify comment and indent

15 years agor23063: Make sure to invalidate the ccache when we set a
Andrew Bartlett [Tue, 22 May 2007 05:21:59 +0000 (05:21 +0000)]
r23063: Make sure to invalidate the ccache when we set a
username/password/realm/etc from the command line.

Also make sure it can't 'come back' from a later call to
cli_credentials_guess(), buy setting a threshold.

This should fix the issues with the build farm...

Andrew Bartlett

15 years agor23062: make sure one node doesn't get ahead of the others at the start
Andrew Tridgell [Tue, 22 May 2007 05:16:16 +0000 (05:16 +0000)]
r23062: make sure one node doesn't get ahead of the others at the start

15 years agor23061: keep the RAW-BENCH-OPEN test balanced
Andrew Tridgell [Tue, 22 May 2007 04:53:05 +0000 (04:53 +0000)]
r23061: keep the RAW-BENCH-OPEN test balanced

15 years agor23060: use #include <roken.h> consistently. Using "roken.h" in this directory
Andrew Tridgell [Tue, 22 May 2007 04:13:17 +0000 (04:13 +0000)]
r23060: use #include <roken.h> consistently. Using "roken.h" in this directory
breaks Samba builds on some systems as they find the wrong roken.h

15 years agor23059: reject_free needs to be initialised
Andrew Tridgell [Tue, 22 May 2007 01:02:00 +0000 (01:02 +0000)]
r23059: reject_free needs to be initialised

15 years agor23058: the cldap code was getting too intimate with the internals of struct
Andrew Tridgell [Tue, 22 May 2007 00:43:10 +0000 (00:43 +0000)]
r23058: the cldap code was getting too intimate with the internals of struct
asn1_context. A hangover from when it wasn't a allocated structure

15 years agor23057: only call the async recv function for the first pending receive, not
Andrew Tridgell [Mon, 21 May 2007 23:35:14 +0000 (23:35 +0000)]
r23057: only call the async recv function for the first pending receive, not
all pending receives, when the transport dies. This is because the
async callback most commonly shuts down the connection (it's the only
reasonable thing to do when it's dead), and that frees the whole
context. That means that if we loop more than once, we'll end up using
freed memory.

15 years agor23056: added automatic reconnect to RAW-BENCH-OPEN, similar to RAW-BENCH-LOCK
Andrew Tridgell [Mon, 21 May 2007 23:33:26 +0000 (23:33 +0000)]
r23056: added automatic reconnect to RAW-BENCH-OPEN, similar to RAW-BENCH-LOCK

15 years agor23047: Allow local inclusion of js files as well as from predefined
Rafal Szczesniak [Mon, 21 May 2007 19:53:57 +0000 (19:53 +0000)]
r23047: Allow local inclusion of js files as well as from predefined


15 years agor23045: forward SMB2 oplock breaks to the client
Stefan Metzmacher [Mon, 21 May 2007 17:39:05 +0000 (17:39 +0000)]
r23045: forward SMB2 oplock breaks to the client


15 years agor23044: - use uint32_t for handle id's
Stefan Metzmacher [Mon, 21 May 2007 17:23:56 +0000 (17:23 +0000)]
r23044: - use uint32_t for handle id's
- include the session vuid in the SMB2 128-Bit wire handles
  as SMB2 oplock breaks doesn't include a TID or VUID in the header
  we need to make sure the handle is unique for the whole TCP connection


15 years agor23043: allow more than UINT16_MAX - 1 handles on smb2 connections
Stefan Metzmacher [Mon, 21 May 2007 17:12:49 +0000 (17:12 +0000)]
r23043: allow more than UINT16_MAX - 1 handles on smb2 connections


15 years agor23037: actually fix the asn1 memory leak :-)
Stefan Metzmacher [Mon, 21 May 2007 13:33:46 +0000 (13:33 +0000)]
r23037: actually fix the asn1 memory leak :-)


15 years agor23036: error checking on asn1_init() failure
Andrew Tridgell [Mon, 21 May 2007 12:47:18 +0000 (12:47 +0000)]
r23036: error checking on asn1_init() failure

15 years agor23035: We don't need to add this entry, and I hope to figure out how to avoid
Andrew Bartlett [Mon, 21 May 2007 12:03:15 +0000 (12:03 +0000)]
r23035: We don't need to add this entry, and I hope to figure out how to avoid
adding the other two manually.

However, to do this we do need to skip the check (with -u) that
slaptest can access the db.

Andrew Bartlett

15 years agor23034: Thanks to metze for providing some vital clues in the 'kerberos ccache
Andrew Bartlett [Mon, 21 May 2007 11:57:56 +0000 (11:57 +0000)]
r23034: Thanks to metze for providing some vital clues in the 'kerberos ccache
on credentials don't do anything' bug.

The problem was simple, we didn't set the ccache as having been
initialised, so we always created a new one.

Andrew Bartlett

15 years agor23033: regenerate schema.ldif with minischema.js
Stefan Metzmacher [Mon, 21 May 2007 09:15:27 +0000 (09:15 +0000)]
r23033: regenerate schema.ldif with minischema.js


15 years agor23032: Remove calls to println(), and ensure we print the ldb errstring().
Andrew Bartlett [Mon, 21 May 2007 08:33:22 +0000 (08:33 +0000)]
r23032: Remove calls to println(), and ensure we print the ldb errstring().

Andrew Bartlett

15 years agor23031: removed bogus event_loop_once() calls in RAW-CONTEXT test - they could
Andrew Tridgell [Mon, 21 May 2007 07:27:24 +0000 (07:27 +0000)]
r23031: removed bogus event_loop_once() calls in RAW-CONTEXT test - they could
hang forever if an ealier one caused two operations to complete at

this is why RAW-CONTEXT was taking 5 minutes on the build farm

15 years agor23030: finally fixed up our asn1 code to use better memory allocation. This
Andrew Tridgell [Mon, 21 May 2007 06:12:06 +0000 (06:12 +0000)]
r23030: finally fixed up our asn1 code to use better memory allocation. This
should allow us to fix some long standing memory leaks.

15 years agor23029: fixed formatting
Andrew Tridgell [Mon, 21 May 2007 05:40:04 +0000 (05:40 +0000)]
r23029: fixed formatting

15 years agor23028: I've now got a patch to Fedora DS to make it only install the very base schema.
Andrew Bartlett [Mon, 21 May 2007 04:10:13 +0000 (04:10 +0000)]
r23028: I've now got a patch to Fedora DS to make it only install the very base schema.

I'll post that patch into lorikeet and submit upstram.

Andrew Bartlett

15 years agor23027: Make sure the parent object always exists.
Andrew Bartlett [Mon, 21 May 2007 04:06:42 +0000 (04:06 +0000)]
r23027: Make sure the parent object always exists.

Andrew Bartlett

15 years agor23026: Add groupPolicyContainer, as we now require this schema element.
Andrew Bartlett [Mon, 21 May 2007 04:01:26 +0000 (04:01 +0000)]
r23026: Add groupPolicyContainer, as we now require this schema element.

Andrew Bartlett

15 years agor23021: test some boundary conditions for idtree, after I saw some suspicious
Andrew Tridgell [Sun, 20 May 2007 10:21:14 +0000 (10:21 +0000)]
r23021: test some boundary conditions for idtree, after I saw some suspicious

15 years agor23020: a better fix for the memory leak - this one doesn't stuff up spnego :)
Andrew Tridgell [Sun, 20 May 2007 09:44:03 +0000 (09:44 +0000)]
r23020: a better fix for the memory leak - this one doesn't stuff up spnego :)

15 years agor23019: revert r23018 - this will require more thought.
Andrew Tridgell [Sun, 20 May 2007 08:57:01 +0000 (08:57 +0000)]
r23019: revert r23018 - this will require more thought.

15 years agor23018: fixed a memory leak in our server side session setup code for failed
Andrew Tridgell [Sun, 20 May 2007 08:43:03 +0000 (08:43 +0000)]
r23018: fixed a memory leak in our server side session setup code for failed
session setups

15 years agor23017: fixed the warning we have been getting for a long time:
Andrew Tridgell [Sun, 20 May 2007 08:01:02 +0000 (08:01 +0000)]
r23017: fixed the warning we have been getting for a long time:
   pvfs_close: failed to delete XXX
during the BASE-DELETE test. It was a real bug, and could result in a
delete on close triggering for a handle that had never fully opened.

15 years agor23013: fixed a bug in the string_replace_w() test that caused OpenBSD to die
Andrew Tridgell [Sat, 19 May 2007 07:05:14 +0000 (07:05 +0000)]
r23013: fixed a bug in the string_replace_w() test that caused OpenBSD to die

15 years agor23012: we need a POLICYGUID in provision
Andrew Tridgell [Sat, 19 May 2007 06:51:34 +0000 (06:51 +0000)]
r23012: we need a POLICYGUID in provision

15 years agor23011: initialisation functions must return NTSTATUS, otherwise we get bus
Andrew Tridgell [Sat, 19 May 2007 06:49:01 +0000 (06:49 +0000)]
r23011: initialisation functions must return NTSTATUS, otherwise we get bus
errors on platforms like sparc. This is why sun1 died during

15 years agor22997: only if the output of which has a leading '/' the output is useful...
Stefan Metzmacher [Fri, 18 May 2007 09:47:53 +0000 (09:47 +0000)]
r22997: only if the output of which has a leading '/' the output is useful...


15 years agor22996: merge HPUX compiler flag fixes from SAMBA_3_0
Stefan Metzmacher [Fri, 18 May 2007 09:20:11 +0000 (09:20 +0000)]
r22996: merge HPUX compiler flag fixes from SAMBA_3_0


15 years agor22995: don't run LOCAL-ICONV twice...
Stefan Metzmacher [Fri, 18 May 2007 09:13:07 +0000 (09:13 +0000)]
r22995: don't run LOCAL-ICONV twice...


15 years agor22993: - make it possible to load a dsdb_schema from ldif via the ejs bindings
Stefan Metzmacher [Fri, 18 May 2007 08:16:50 +0000 (08:16 +0000)]
r22993: - make it possible to load a dsdb_schema from ldif via the ejs bindings
- make it possible to set ntds_objectGUID and ntds_invocationId via the ejy bindings


15 years agor22992: add a function to attach the dsdb_schema to a ldb_context
Stefan Metzmacher [Fri, 18 May 2007 08:13:06 +0000 (08:13 +0000)]
r22992: add a function to attach the dsdb_schema to a ldb_context
loaded from ldif files. The first file should contain
the info's about the 'prefixMap' attribute in the first ldif chunk.
The 2nd file should contain all attributes and classes.


15 years agor22991: ignore '@foo' attributes in the repl_meta_data module
Stefan Metzmacher [Fri, 18 May 2007 08:09:51 +0000 (08:09 +0000)]
r22991: ignore '@foo' attributes in the repl_meta_data module


15 years agor22990: free temporary memory also on success...
Stefan Metzmacher [Fri, 18 May 2007 07:58:37 +0000 (07:58 +0000)]
r22990: free temporary memory also on success...


15 years agor22989: allow '@foo' attributes searchable as we allow '@foo' attributes
Stefan Metzmacher [Fri, 18 May 2007 07:41:43 +0000 (07:41 +0000)]
r22989: allow '@foo' attributes searchable as we allow '@foo' attributes
in all other places too...


15 years agor22988: fixed 2 bugs in our unsetenv() replacement code
Andrew Tridgell [Fri, 18 May 2007 06:53:57 +0000 (06:53 +0000)]
r22988: fixed 2 bugs in our unsetenv() replacement code

 1) you must not free the memory, as it is possible the memory did not
 come from malloc (try it under valgrind to test)

 2) the old code didn't cope with duplicate environment variables

I hope this will fix some of the build farm errors on irix, and maybe solaris

15 years agor22987: Clarify how the events are handled in the kerberos code, and
Andrew Bartlett [Fri, 18 May 2007 05:47:33 +0000 (05:47 +0000)]
r22987: Clarify how the events are handled in the kerberos code, and
standardise with the rest of the code.

Andrew Bartlett

15 years agor22986: error 2312 is returned when NetSessDel() fails because a matching session...
Ronnie Sahlberg [Fri, 18 May 2007 05:21:25 +0000 (05:21 +0000)]
r22986: error 2312 is returned when NetSessDel() fails because a matching session (username/clientname)
could not be found

15 years agor22985: don't do checkins late at night ....
Andrew Tridgell [Fri, 18 May 2007 01:23:35 +0000 (01:23 +0000)]
r22985: don't do checkins late at night ....

thanks to Volker for spotting this!

15 years agor22984: not everyone uses (fortunately)
Andrew Tridgell [Fri, 18 May 2007 01:22:27 +0000 (01:22 +0000)]
r22984: not everyone uses (fortunately)

15 years agor22983: This should ensure that torture_create_testuser() can be called,
Andrew Bartlett [Fri, 18 May 2007 01:15:50 +0000 (01:15 +0000)]
r22983: This should ensure that torture_create_testuser() can be called,
without needing to change the user's password afterwoulds.

(Accounts were being created with 'must change at next login' set).

Andrew Bartlett

15 years agor22972: added the basic ldif needed to support group policies in Samba4. WinXP
Andrew Tridgell [Thu, 17 May 2007 10:33:40 +0000 (10:33 +0000)]
r22972: added the basic ldif needed to support group policies in Samba4. WinXP
clients do correctly see our group policies, but the gpmc admin tool
doesn't yet work to allow you to edit the policies

15 years agor22971: fix build on systems without epoll
Andrew Tridgell [Thu, 17 May 2007 09:48:17 +0000 (09:48 +0000)]
r22971: fix build on systems without epoll

15 years agor22970: the events code calls close(), which needs to use socket_wrapper.h in
Andrew Tridgell [Thu, 17 May 2007 09:44:58 +0000 (09:44 +0000)]
r22970: the events code calls close(), which needs to use socket_wrapper.h in
the build farm

15 years agor22969: fix some more places where we could end up with more than one event
Andrew Tridgell [Thu, 17 May 2007 08:47:04 +0000 (08:47 +0000)]
r22969: fix some more places where we could end up with more than one event
context. We now have an event context on the torture_context, and we
can also get one from the cli_credentials structure

15 years agor22968: andrew bartlett pointed out that the parent process could add a fd to
Andrew Tridgell [Thu, 17 May 2007 07:52:33 +0000 (07:52 +0000)]
r22968: andrew bartlett pointed out that the parent process could add a fd to
the epoll context which would then appear in the children. To fix this
we need to check for pid changes in more places. Luckily on platforms
where we have epoll(), getpid() is very very cheap.

15 years agor22967: Move to the TCP packet interface for the krb5_send_to_kdc plugin.
Andrew Bartlett [Thu, 17 May 2007 05:46:45 +0000 (05:46 +0000)]
r22967: Move to the TCP packet interface for the krb5_send_to_kdc plugin.

This replaces a lump of hand-crafted code with the generic packet
system used in the rest of Samba4.

(I started this while chasing down the epoll bug, which turned out to
be seperate)

15 years agor22966: Make sure to return LOGON_FAILURE if the user's kerberos password is
Andrew Bartlett [Thu, 17 May 2007 05:44:51 +0000 (05:44 +0000)]
r22966: Make sure to return LOGON_FAILURE if the user's kerberos password is

Andrew Bartlett

15 years agor22965: changed RAW-LOCK-BENCH to remove the scheduling uncertainty. We should
Andrew Tridgell [Thu, 17 May 2007 03:42:28 +0000 (03:42 +0000)]
r22965: changed RAW-LOCK-BENCH to remove the scheduling uncertainty. We should
always get balanced locking now

15 years agor22964: log an error on epoll_create failing on reopen
Andrew Tridgell [Thu, 17 May 2007 02:42:14 +0000 (02:42 +0000)]
r22964: log an error on epoll_create failing on reopen

15 years agor22963: fixed the epoll/fork interaction in the epoll and aio backends
Andrew Tridgell [Thu, 17 May 2007 02:38:07 +0000 (02:38 +0000)]
r22963: fixed the epoll/fork interaction in the epoll and aio backends

15 years agor22962: show progress separately for each client in RAW-BENCH-LOCK, this is
Andrew Tridgell [Thu, 17 May 2007 02:22:29 +0000 (02:22 +0000)]
r22962: show progress separately for each client in RAW-BENCH-LOCK, this is
much more useful for seeing why we get unbalanced locking