Merge branch 'v5-stable' into v6-stable

Conflicts:
	ChangeLog
	configure.ac
	doc/manual.html

7 files changed, 62 insertions, 34 deletions

diff --git a/ChangeLog b/ChangeLog
index 2dd4eead..991e4630 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -9,6 +9,21 @@ Version 6.2.2  [v6-stable], 2012-05-??
 - bugfix: --enable-smcustbindcdr configure directive did not work
   closes: http://bugzilla.adiscon.com/show_bug.cgi?id=330
   Thanks to Ultrabug for the patch.
+- add small delay (50ms) after sending shutdown message
+  There seem to be cases where the shutdown message is otherwise not
+  processed, not even on an idle system. Thanks to Marcin for
+  bringing this problem up.
+- support for resolving huge groups
+  closes: http://bugzilla.adiscon.com/show_bug.cgi?id=310
+  Thanks to Alec Warner for the patch
+- bugfix: potential hang due to mutex deadlock
+  closes: http://bugzilla.adiscon.com/show_bug.cgi?id=316
+  Thanks to Andreas Piesk for reporting&analyzing this bug as well as
+  providing patches and other help in resolving it.
+- bugfix: property PROCID empty instead of proper nilvalue if not present
+  If it is not present, it must have the nilvalue "-" as of RFC5424
+  closes: http://bugzilla.adiscon.com/show_bug.cgi?id=332
+  Thanks to John N for reporting this issue.
 ---------------------------------------------------------------------------
 Version 6.2.1  [v6-stable], 2012-05-10
 - change plugin config interface to be compatible with pre-v6.2 system
@@ -284,7 +299,14 @@ Version 5.9.0  [V5-DEVEL] (rgerhards), 2011-03-??
   affected directive was: $ActionExecOnlyWhenPreviousIsSuspended on
   closes: http://bugzilla.adiscon.com/show_bug.cgi?id=236
 ---------------------------------------------------------------------------
-Version 5.8.12  [V5-stable] 2012-05-??
+Version 5.8.12  [V5-stable] 2012-06-06
+- add small delay (50ms) after sending shutdown message
+  There seem to be cases where the shutdown message is otherwise not
+  processed, not even on an idle system. Thanks to Marcin for
+  bringing this problem up.
+- support for resolving huge groups
+  closes: http://bugzilla.adiscon.com/show_bug.cgi?id=310
+  Thanks to Alec Warner for the patch
 - bugfix: delayble source could block action queue, even if there was
   a disk queue associated with it. The root cause of this problem was
   that it makes no sense to delay messages once they arrive in the 
@@ -310,6 +332,14 @@ Version 5.8.12  [V5-stable] 2012-05-??
   please let us know.
   Thanks to Tomas Heinrich for the patch.
 - bugfix/tcpflood: sending small test files did not work correctly
+- bugfix: potential hang due to mutex deadlock
+  closes: http://bugzilla.adiscon.com/show_bug.cgi?id=316
+  Thanks to Andreas Piesk for reporting&analyzing this bug as well as
+  providing patches and other help in resolving it.
+- bugfix: property PROCID empty instead of proper nilvalue if not present
+  If it is not present, it must have the nilvalue "-" as of RFC5424
+  closes: http://bugzilla.adiscon.com/show_bug.cgi?id=332
+  Thanks to John N for reporting this issue.
 ---------------------------------------------------------------------------
 Version 5.8.11  [V5-stable] 2012-05-03
 - bugfix: ommysql did not properly init/exit the mysql runtime library
diff --git a/runtime/cfsysline.c b/runtime/cfsysline.c
index b9516435..2d91094c 100644
--- a/runtime/cfsysline.c
+++ b/runtime/cfsysline.c
@@ -339,11 +339,12 @@ static int doParseOnOffOption(uchar **pp)
  */
 static rsRetVal doGetGID(uchar **pp, rsRetVal (*pSetHdlr)(void*, uid_t), void *pVal)
 {
-	struct group *pgBuf;
+	struct group *pgBuf = NULL;
 	struct group gBuf;
 	DEFiRet;
 	uchar szName[256];
-	char stringBuf[2048];	/* I hope this is large enough... */
+	int bufSize = 2048;
+	char * stringBuf = NULL;
 
 	assert(pp != NULL);
 	assert(*pp != NULL);
@@ -353,7 +354,17 @@ static rsRetVal doGetGID(uchar **pp, rsRetVal (*pSetHdlr)(void*, uid_t), void *p
 		ABORT_FINALIZE(RS_RET_NOT_FOUND);
 	}
 
-	getgrnam_r((char*)szName, &gBuf, stringBuf, sizeof(stringBuf), &pgBuf);
+
+	CHKmalloc(stringBuf = malloc(bufSize));
+	while(pgBuf == NULL) {
+		errno = 0;
+		getgrnam_r((char*)szName, &gBuf, stringBuf, bufSize, &pgBuf);
+		if((pgBuf == NULL) && (errno == ERANGE)) {
+			/* Increase bufsize and try again.*/
+			bufSize *= 2;
+			CHKmalloc(stringBuf = realloc(stringBuf, bufSize));
+		}
+	}
 
 	if(pgBuf == NULL) {
 		errmsg.LogError(0, RS_RET_NOT_FOUND, "ID for group '%s' could not be found or error", (char*)szName);
@@ -372,6 +383,7 @@ static rsRetVal doGetGID(uchar **pp, rsRetVal (*pSetHdlr)(void*, uid_t), void *p
 	skipWhiteSpace(pp); /* skip over any whitespace */
 
 finalize_it:
+	free(stringBuf);
 	RETiRet;
 }
 
diff --git a/runtime/modules.c b/runtime/modules.c
index 23b5c94e..3e62139f 100644
--- a/runtime/modules.c
+++ b/runtime/modules.c
@@ -66,14 +66,6 @@ DEFobjCurrIf(errmsg)
 DEFobjCurrIf(parser)
 DEFobjCurrIf(strgen)
 
-/* we must ensure that only one thread at one time tries to load or unload
- * modules, otherwise we may see race conditions. This first came up with
- * imdiag/imtcp, which both use the same stream drivers. Below is the mutex
- * for that handling.
- * rgerhards, 2009-05-25
- */
-static pthread_mutex_t mutLoadUnload;
-
 static modInfo_t *pLoadedModules = NULL;	/* list of currently-loaded modules */
 static modInfo_t *pLoadedModulesLast = NULL;	/* tail-pointer */
 
@@ -666,7 +658,7 @@ modUnlinkAndDestroy(modInfo_t **ppThis)
 	pThis = *ppThis;
 	assert(pThis != NULL);
 
-	pthread_mutex_lock(&mutLoadUnload);
+	pthread_mutex_lock(&mutObjGlobalOp);
 
 	/* first check if we are permitted to unload */
 	if(pThis->eType == eMOD_LIB) {
@@ -702,7 +694,7 @@ modUnlinkAndDestroy(modInfo_t **ppThis)
 	moduleDestruct(pThis);
 
 finalize_it:
-	pthread_mutex_unlock(&mutLoadUnload);
+	pthread_mutex_unlock(&mutObjGlobalOp);
 	RETiRet;
 }
 
@@ -778,7 +770,7 @@ Load(uchar *pModName)
 	assert(pModName != NULL);
 	dbgprintf("Requested to load module '%s'\n", pModName);
 
-	pthread_mutex_lock(&mutLoadUnload);
+	pthread_mutex_lock(&mutObjGlobalOp);
 
 	iModNameLen = strlen((char *) pModName);
 	if(iModNameLen > 3 && !strcmp((char *) pModName + iModNameLen - 3, ".so")) {
@@ -902,7 +894,7 @@ Load(uchar *pModName)
 	}
 
 finalize_it:
-	pthread_mutex_unlock(&mutLoadUnload);
+	pthread_mutex_unlock(&mutObjGlobalOp);
 	RETiRet;
 }
 
@@ -999,16 +991,6 @@ CODESTARTObjClassExit(module)
 	/* release objects we no longer need */
 	objRelease(errmsg, CORE_COMPONENT);
 	objRelease(parser, CORE_COMPONENT);
-	/* We have a problem in our reference counting, which leads to this function
-	 * being called too early. This usually is no problem, but if we destroy
-	 * the mutex object, we get into trouble. So rather than finding the root cause,
-	 * we do not release the mutex right now and have a very, very slight leak.
-	 * We know that otherwise no bad effects happen, so this acceptable for the 
-	 * time being. -- rgerhards, 2009-05-25
-	 *
-	 * TODO: add again: pthread_mutex_destroy(&mutLoadUnload);
-	 */
-
 #	ifdef DEBUG
 	modUsrPrintAll(); /* debug aid - TODO: integrate with debug.c, at least the settings! */
 #	endif
@@ -1050,7 +1032,6 @@ ENDobjQueryInterface(module)
  */
 BEGINAbstractObjClassInit(module, 1, OBJ_IS_CORE_MODULE) /* class, version - CHANGE class also in END MACRO! */
 	uchar *pModPath;
-	pthread_mutexattr_t mutAttr;
 
 	/* use any module load path specified in the environment */
 	if((pModPath = (uchar*) getenv("RSYSLOG_MODDIR")) != NULL) {
@@ -1068,10 +1049,6 @@ BEGINAbstractObjClassInit(module, 1, OBJ_IS_CORE_MODULE) /* class, version - CHA
 		SetModDir(glblModPath);
 	}
 
-	pthread_mutexattr_init(&mutAttr);
-	pthread_mutexattr_settype(&mutAttr, PTHREAD_MUTEX_RECURSIVE);
-	pthread_mutex_init(&mutLoadUnload, &mutAttr);
-
 	/* request objects we use */
 	CHKiRet(objUse(errmsg, CORE_COMPONENT));
 ENDObjClassInit(module)
diff --git a/runtime/msg.c b/runtime/msg.c
index 01808ac0..e09b8705 100644
--- a/runtime/msg.c
+++ b/runtime/msg.c
@@ -1633,7 +1633,7 @@ char *getPROCID(msg_t *pM, sbool bLockMutex)
 		MsgLock(pM);
 	preparePROCID(pM, MUTEX_ALREADY_LOCKED);
 	if(pM->pCSPROCID == NULL)
-		pszRet = UCHAR_CONSTANT("");
+		pszRet = UCHAR_CONSTANT("-");
 	else 
 		pszRet = rsCStrGetSzStrNoNULL(pM->pCSPROCID);
 	if(bLockMutex == LOCK_MUTEX)
diff --git a/runtime/obj.c b/runtime/obj.c
index 29ca8117..f9afecd8 100644
--- a/runtime/obj.c
+++ b/runtime/obj.c
@@ -97,7 +97,7 @@ DEFobjCurrIf(module)
 DEFobjCurrIf(errmsg)
 DEFobjCurrIf(strm)
 static objInfo_t *arrObjInfo[OBJ_NUM_IDS]; /* array with object information pointers */
-static pthread_mutex_t mutObjGlobalOp;	/* mutex to guard global operations of the object system */
+pthread_mutex_t mutObjGlobalOp;	/* mutex to guard global operations of the object system */
 
 
 /* cookies for serialized lines */
@@ -1319,7 +1319,7 @@ objClassInit(modInfo_t *pModInfo)
 	}
 
 	/* the mutex must be recursive, because objects may call into other
-	 * object identifieres recursively.
+	 * object identifiers recursively.
 	 */
 	pthread_mutexattr_init(&mutAttr);
 	pthread_mutexattr_settype(&mutAttr, PTHREAD_MUTEX_RECURSIVE);
diff --git a/runtime/obj.h b/runtime/obj.h
index be97f20e..32f7ef09 100644
--- a/runtime/obj.h
+++ b/runtime/obj.h
@@ -122,4 +122,8 @@ rsRetVal objGetObjInterface(obj_if_t *pIf);
 PROTOTYPEObjClassInit(obj);
 PROTOTYPEObjClassExit(obj);
 
+
+/* the following definition is only for "friends" */
+extern pthread_mutex_t mutObjGlobalOp;	/* mutex to guard global operations of the object system */
+
 #endif /* #ifndef OBJ_H_INCLUDED */
diff --git a/tools/syslogd.c b/tools/syslogd.c
index 1b057710..9e4b92fe 100644
--- a/tools/syslogd.c
+++ b/tools/syslogd.c
@@ -1069,6 +1069,11 @@ die(int sig)
 		errno = 0;
 		logmsgInternal(NO_ERRCODE, LOG_SYSLOG|LOG_INFO, (uchar*)buf, 0);
 	}
+	/* we sleep for 50ms to give the queue a chance to pick up the exit message;
+	 * otherwise we have seen cases where the message did not make it to log
+	 * files, even on idle systems.
+	 */
+	srSleep(0, 50);
 
 	/* drain queue (if configured so) and stop main queue worker thread pool */
 	DBGPRINTF("Terminating main queue...\n");