optimized TAG handling

5 files changed, 44 insertions, 60 deletions

diff --git a/runtime/msg.c b/runtime/msg.c
index 6cf41322..1a882597 100644
--- a/runtime/msg.c
+++ b/runtime/msg.c
@@ -551,6 +551,11 @@ static inline void freeTAG(msg_t *pThis)
 	if(pThis->iLenTAG >= CONF_TAG_BUFSIZE)
 		free(pThis->TAG.pszTAG);
 }
+static inline void freeHOSTNAME(msg_t *pThis)
+{
+	if(pThis->iLenHOSTNAME >= CONF_HOSTNAME_BUFSIZE)
+		free(pThis->pszHOSTNAME);
+}
 
 
 BEGINobjDestruct(msg) /* be sure to specify the object type also in END and CODESTART macros! */
@@ -569,7 +574,7 @@ CODESTARTobjDestruct(msg)
 		if(pThis->pszRawMsg != pThis->szRawMsg)
 			free(pThis->pszRawMsg);
 		freeTAG(pThis);
-		free(pThis->pszHOSTNAME);
+		freeHOSTNAME(pThis);
 		free(pThis->pszInputName);
 		free(pThis->pszRcvFrom);
 		free(pThis->pszRcvFromIP);
@@ -669,7 +674,7 @@ msg_t* MsgDup(msg_t* pOld)
 	pNew->iProtocolVersion = pOld->iProtocolVersion;
 	pNew->ttGenTime = pOld->ttGenTime;
 	pNew->offMSG = pOld->offMSG;
-	/* enable this, if someone actually uses UxTradMsg, delete after some  time has
+	/* enable this, if someone actually uses UxTradMsg, delete after some time has
 	 * passed and nobody complained -- rgerhards, 2009-06-16
 	pNew->offAfterPRI = pOld->offAfterPRI;
 	*/
@@ -684,8 +689,18 @@ msg_t* MsgDup(msg_t* pOld)
 			pNew->iLenTAG = pOld->iLenTAG;
 		}
 	}
-	tmpCOPYSZ(RawMsg);
-	tmpCOPYSZ(HOSTNAME);
+	if(pOld->iLenRawMsg < CONF_RAWMSG_BUFSIZE) {
+		memcpy(pNew->szRawMsg, pOld->szRawMsg, pOld->iLenRawMsg + 1);
+		pNew->pszRawMsg = pNew->szRawMsg;
+	} else {
+		tmpCOPYSZ(RawMsg);
+	}
+	if(pOld->iLenHOSTNAME < CONF_HOSTNAME_BUFSIZE) {
+		memcpy(pNew->szHOSTNAME, pOld->szHOSTNAME, pOld->iLenHOSTNAME + 1);
+		pNew->pszHOSTNAME = pNew->szHOSTNAME;
+	} else {
+		tmpCOPYSZ(HOSTNAME);
+	}
 	tmpCOPYSZ(RcvFrom);
 
 	tmpCOPYCSTR(ProgName);
@@ -886,30 +901,6 @@ finalize_it:
 }
 
 
-/* This function moves the HOSTNAME inside the message object to the
- * TAG. It is a specialised function used to handle the condition when
- * a message without HOSTNAME is being processed. The missing HOSTNAME
- * is only detected at a later stage, during TAG processing, so that
- * we already had set the HOSTNAME property and now need to move it to
- * the TAG. Of course, we could do this via a couple of get/set methods,
- * but it is far more efficient to do it via this specialised method.
- * This is especially important as this can be a very common case, e.g.
- * when BSD syslog is acting as a sender.
- * rgerhards, 2005-11-10.
- * NOTE ********* 2009-06-18 / rgerhards *************
- * This function is being obsoleted by the new handling. I keep it for
- * a while, and for oversize tags it is somewhat less optimal than in previous
- * versions. This should only happen very seldom.
- */
-void moveHOSTNAMEtoTAG(msg_t *pM)
-{
-	assert(pM != NULL);
-	MsgSetTAG(pM, pM->pszHOSTNAME, pM->iLenHOSTNAME);
-	free(pM->pszHOSTNAME);
-	pM->pszHOSTNAME = NULL;
-	pM->iLenHOSTNAME = 0;
-}
-
 /* Access methods - dumb & easy, not a comment for each ;)
  */
 void setProtocolVersion(msg_t *pM, int iNewVersion)
@@ -1694,22 +1685,6 @@ finalize_it:
 }
 
 
-/* Set the HOSTNAME to a caller-provided string. This is thought
- * to be a heap buffer that the caller will no longer use. This
- * function is a performance optimization over MsgSetHOSTNAME().
- * rgerhards 2004-11-19
- */
-void MsgAssignHOSTNAME(msg_t *pMsg, char *pBuf)
-{
-	assert(pMsg != NULL);
-	assert(pBuf != NULL);
-	if(pMsg->pszHOSTNAME != NULL)
-		free(pMsg->pszHOSTNAME);
-	pMsg->iLenHOSTNAME = strlen(pBuf);
-	pMsg->pszHOSTNAME = (uchar*) pBuf;
-}
-
-
 /* rgerhards 2004-11-09: set HOSTNAME in msg object
  * rgerhards, 2007-06-21:
  * Does not return anything. If an error occurs, the hostname is
@@ -1720,16 +1695,24 @@ void MsgAssignHOSTNAME(msg_t *pMsg, char *pBuf)
  * we need it. The rest of the code already knows how to handle an
  * unset HOSTNAME.
  */
-void MsgSetHOSTNAME(msg_t *pMsg, uchar* pszHOSTNAME, int lenHOSTNAME)
+void MsgSetHOSTNAME(msg_t *pThis, uchar* pszHOSTNAME, int lenHOSTNAME)
 {
-	assert(pMsg != NULL);
-	free(pMsg->pszHOSTNAME);
+	assert(pThis != NULL);
 
-	pMsg->iLenHOSTNAME = lenHOSTNAME;
-	if((pMsg->pszHOSTNAME = malloc(pMsg->iLenHOSTNAME + 1)) != NULL)
-		memcpy(pMsg->pszHOSTNAME, pszHOSTNAME, pMsg->iLenHOSTNAME + 1);
-	else
-		DBGPRINTF("Could not allocate memory in MsgSetHOSTNAME()\n");
+	freeHOSTNAME(pThis);
+
+	pThis->iLenHOSTNAME = lenHOSTNAME;
+	if(pThis->iLenHOSTNAME < CONF_HOSTNAME_BUFSIZE) {
+		/* small enough: use fixed buffer (faster!) */
+		pThis->pszHOSTNAME = pThis->szHOSTNAME;
+	} else if((pThis->pszHOSTNAME = (uchar*) malloc(pThis->iLenHOSTNAME + 1)) == NULL) {
+		/* truncate message, better than completely loosing it... */
+		pThis->pszHOSTNAME = pThis->szHOSTNAME;
+		pThis->iLenHOSTNAME = CONF_HOSTNAME_BUFSIZE - 1;
+	}
+
+	memcpy(pThis->pszHOSTNAME, pszHOSTNAME, pThis->iLenHOSTNAME);
+	pThis->pszHOSTNAME[pThis->iLenHOSTNAME] = '\0'; /* this also works with truncation! */
 }
 
 
diff --git a/runtime/msg.h b/runtime/msg.h
index fd4d650b..4bfc1e3f 100644
--- a/runtime/msg.h
+++ b/runtime/msg.h
@@ -111,6 +111,7 @@ struct msg {
 	struct syslogTime tTIMESTAMP;/* (parsed) value of the timestamp */
 	/* some fixed-size buffers to save malloc()/free() for frequently used fields (from the default templates) */
 	uchar szRawMsg[CONF_RAWMSG_BUFSIZE];	/* most messages are small, and these are stored here (without malloc/free!) */
+	uchar szHOSTNAME[CONF_HOSTNAME_BUFSIZE];
 	union {
 		uchar	*pszTAG;	/* pointer to tag value */
 		uchar	szBuf[CONF_TAG_BUFSIZE];
@@ -158,7 +159,6 @@ void MsgSetMSGoffs(msg_t *pMsg, short offs);
 void MsgSetRawMsgWOSize(msg_t *pMsg, char* pszRawMsg);
 void MsgSetRawMsg(msg_t *pMsg, char* pszRawMsg, size_t lenMsg);
 rsRetVal MsgReplaceMSG(msg_t *pThis, uchar* pszMSG, int lenMSG);
-void moveHOSTNAMEtoTAG(msg_t *pM);
 char *MsgGetProp(msg_t *pMsg, struct templateEntry *pTpe,
                  cstr_t *pCSPropName, size_t *pPropLen, unsigned short *pbMustBeFreed);
 char *textpri(char *pRes, size_t pResLen, int pri);
diff --git a/runtime/rsyslog.h b/runtime/rsyslog.h
index 4fa5100e..913ab300 100644
--- a/runtime/rsyslog.h
+++ b/runtime/rsyslog.h
@@ -31,12 +31,14 @@
  * ############################################################# */
 #define RS_STRINGBUF_ALLOC_INCREMENT	128
 /* MAXSIZE are absolute maxima, while BUFSIZE are just values after which
- * processing is more time-intense.
+ * processing is more time-intense. The BUFSIZE params currently add their
+ * value to the fixed size of the message object.
  */
 #define CONF_TAG_MAXSIZE		512	/* a value that is deemed far too large for any valid TAG */
 #define CONF_TAG_HOSTNAME		512	/* a value that is deemed far too large for any valid HOSTNAME */
 #define CONF_RAWMSG_BUFSIZE		101
-#define CONF_TAG_BUFSIZE		33	/* RFC says 32 chars (+ \0), but in practice we see longer ones... */
+#define CONF_TAG_BUFSIZE		32
+#define CONF_HOSTNAME_BUFSIZE		32
 
 
 /* ############################################################# *
diff --git a/tests/diskqueue.sh b/tests/diskqueue.sh
index 668a1224..d2e75cbd 100755
--- a/tests/diskqueue.sh
+++ b/tests/diskqueue.sh
@@ -5,7 +5,7 @@
 # added 2009-04-17 by Rgerhards
 # This file is part of the rsyslog project, released  under GPLv3
 # uncomment for debugging support:
-echo testing queue disk-only mode
+echo diskqueue.sh: testing queue disk-only mode
 source $srcdir/diag.sh init
 source $srcdir/diag.sh startup diskqueue.conf
 # 20000 messages should be enough - the disk test is slow enough ;)
diff --git a/tools/syslogd.c b/tools/syslogd.c
index faa793fb..6f264e5e 100644
--- a/tools/syslogd.c
+++ b/tools/syslogd.c
@@ -609,7 +609,7 @@ static inline rsRetVal printline(uchar *hname, uchar *hnameIP, uchar *msg, int f
 	 * being the local host).  rgerhards 2004-11-16
 	 */
 	if((pMsg->msgFlags & PARSE_HOSTNAME) == 0)
-		MsgSetHOSTNAME(pMsg, hname, strlen(hname));
+		MsgSetHOSTNAME(pMsg, hname, ustrlen(hname));
 	MsgSetRcvFrom(pMsg, hname);
 	MsgSetAfterPRIOffs(pMsg, p - msg);
 	CHKiRet(MsgSetRcvFromIP(pMsg, hnameIP));
@@ -1113,7 +1113,7 @@ int parseRFCSyslogMsg(msg_t *pMsg, int flags)
 	/* HOSTNAME */
 	if(bContParse) {
 		parseRFCField(&p2parse, pBuf);
-		MsgSetHOSTNAME(pMsg, pBuf, strlen(pBuf));
+		MsgSetHOSTNAME(pMsg, pBuf, ustrlen(pBuf));
 	}
 
 	/* APP-NAME */
@@ -1267,8 +1267,7 @@ int parseLegacySyslogMsg(msg_t *pMsg, int flags)
 		 */
 		bufParseTAG[i] = '\0';	/* terminate string */
 		MsgSetTAG(pMsg, bufParseTAG, i);
-	} else {
-		/* we enter this code area when the user has instructed rsyslog NOT
+	} else {/* we enter this code area when the user has instructed rsyslog NOT
 		 * to parse HOSTNAME and TAG - rgerhards, 2006-03-13
 		 */
 		if(!(flags & INTERNAL_MSG)) {