summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
-rw-r--r--ChangeLog2
-rw-r--r--runtime/stringbuf.c38
-rw-r--r--runtime/stringbuf.h3
-rw-r--r--tools/syslogd.c6
-rw-r--r--tools/syslogd.h1
5 files changed, 42 insertions, 8 deletions
diff --git a/ChangeLog b/ChangeLog
index 42fda5b3..fc9f807f 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -34,6 +34,8 @@ Version 4.1.6 [DEVEL] (rgerhards), 2009-03-??
the full high availability features of rsyslog's engine
- bugfix: fixed some segaults on Solaris, where vsprintf() does not
check for NULL pointers
+- improved performance of regexp-based filters
+ Thanks to Arnaud Cornet for providing the idea and initial patch.
---------------------------------------------------------------------------
Version 4.1.5 [DEVEL] (rgerhards), 2009-03-11
- bugfix: parser did not correctly parse fields in UDP-received messages
diff --git a/runtime/stringbuf.c b/runtime/stringbuf.c
index a5dc625a..35ec44c6 100644
--- a/runtime/stringbuf.c
+++ b/runtime/stringbuf.c
@@ -703,17 +703,26 @@ int rsCStrCaseInsensitveStartsWithSzStr(cstr_t *pCS1, uchar *psz, size_t iLenSz)
* never is a \0 *inside* a property string.
* Note that the function returns -1 if regexp functionality is not available.
* rgerhards: 2009-03-04: ERE support added, via parameter iType: 0 - BRE, 1 - ERE
+ * Arnaud Cornet/rgerhards: 2009-04-02: performance improvement by caching compiled regex
+ * If a caller does not need the cached version, it must still provide memory for it
+ * and must call rsCStrRegexDestruct() afterwards.
*/
-rsRetVal rsCStrSzStrMatchRegex(cstr_t *pCS1, uchar *psz, int iType)
+rsRetVal rsCStrSzStrMatchRegex(cstr_t *pCS1, uchar *psz, int iType, void *rc)
{
- regex_t preq;
+ regex_t **cache = (regex_t**) rc;
int ret;
DEFiRet;
+ assert(pCS1 != NULL);
+ assert(psz != NULL);
+ assert(cache != NULL);
+
if(objUse(regexp, LM_REGEXP_FILENAME) == RS_RET_OK) {
- regexp.regcomp(&preq, (char*) rsCStrGetSzStr(pCS1), (iType == 1 ? REG_EXTENDED : 0) | REG_NOSUB);
- ret = regexp.regexec(&preq, (char*) psz, 0, NULL, 0);
- regexp.regfree(&preq);
+ if (*cache == NULL) {
+ *cache = calloc(sizeof(regex_t), 1);
+ regexp.regcomp(*cache, (char*) rsCStrGetSzStr(pCS1), (iType == 1 ? REG_EXTENDED : 0) | REG_NOSUB);
+ }
+ ret = regexp.regexec(*cache, (char*) psz, 0, NULL, 0);
if(ret != 0)
ABORT_FINALIZE(RS_RET_NOT_FOUND);
} else {
@@ -725,6 +734,25 @@ finalize_it:
}
+/* free a cached compiled regex
+ * Caller must provide a pointer to a buffer that was created by
+ * rsCStrSzStrMatchRegexCache()
+ */
+void rsCStrRegexDestruct(void *rc)
+{
+ regex_t **cache = rc;
+
+ assert(cache != NULL);
+ assert(*cache != NULL);
+
+ if(objUse(regexp, LM_REGEXP_FILENAME) == RS_RET_OK) {
+ regexp.regfree(*cache);
+ free(*cache);
+ *cache = NULL;
+ }
+}
+
+
/* compare a rsCStr object with a classical sz string. This function
* is almost identical to rsCStrZsStrCmp(), but it also takes an offset
* to the CStr object from where the comparison is to start.
diff --git a/runtime/stringbuf.h b/runtime/stringbuf.h
index f3e08439..684133bb 100644
--- a/runtime/stringbuf.h
+++ b/runtime/stringbuf.h
@@ -136,7 +136,8 @@ int rsCStrCaseInsensitiveLocateInSzStr(cstr_t *pThis, uchar *sz);
int rsCStrStartsWithSzStr(cstr_t *pCS1, uchar *psz, size_t iLenSz);
int rsCStrCaseInsensitveStartsWithSzStr(cstr_t *pCS1, uchar *psz, size_t iLenSz);
int rsCStrSzStrStartsWithCStr(cstr_t *pCS1, uchar *psz, size_t iLenSz);
-rsRetVal rsCStrSzStrMatchRegex(cstr_t *pCS1, uchar *psz, int iType);
+rsRetVal rsCStrSzStrMatchRegex(cstr_t *pCS1, uchar *psz, int iType, void *cache);
+void rsCStrRegexDestruct(void *rc);
rsRetVal rsCStrConvertToNumber(cstr_t *pStr, number_t *pNumber);
rsRetVal rsCStrConvertToBool(cstr_t *pStr, number_t *pBool);
rsRetVal rsCStrAppendCStr(cstr_t *pThis, cstr_t *pstrAppend);
diff --git a/tools/syslogd.c b/tools/syslogd.c
index 16f255ea..b23c12a7 100644
--- a/tools/syslogd.c
+++ b/tools/syslogd.c
@@ -429,6 +429,8 @@ selectorDestruct(void *pVal)
rsCStrDestruct(&pThis->f_filterData.prop.pCSPropName);
if(pThis->f_filterData.prop.pCSCompValue != NULL)
rsCStrDestruct(&pThis->f_filterData.prop.pCSCompValue);
+ if(pThis->f_filterData.prop.regex_cache != NULL)
+ rsCStrRegexDestruct(&pThis->f_filterData.prop.regex_cache);
} else if(pThis->f_filter_type == FILTER_EXPR) {
if(pThis->f_filterData.f_expr != NULL)
expr.Destruct(&pThis->f_filterData.f_expr);
@@ -1076,12 +1078,12 @@ static rsRetVal shouldProcessThisMessage(selector_t *f, msg_t *pMsg, int *bProce
break;
case FIOP_REGEX:
if(rsCStrSzStrMatchRegex(f->f_filterData.prop.pCSCompValue,
- (unsigned char*) pszPropVal, 0) == RS_RET_OK)
+ (unsigned char*) pszPropVal, 0, &f->f_filterData.prop.regex_cache) == RS_RET_OK)
bRet = 1;
break;
case FIOP_EREREGEX:
if(rsCStrSzStrMatchRegex(f->f_filterData.prop.pCSCompValue,
- (unsigned char*) pszPropVal, 1) == RS_RET_OK)
+ (unsigned char*) pszPropVal, 1, &f->f_filterData.prop.regex_cache) == RS_RET_OK)
bRet = 1;
break;
default:
diff --git a/tools/syslogd.h b/tools/syslogd.h
index f1b11a91..8b9bd131 100644
--- a/tools/syslogd.h
+++ b/tools/syslogd.h
@@ -73,6 +73,7 @@ struct filed {
FIOP_REGEX = 4, /* matches a (BRE) regular expression? */
FIOP_EREREGEX = 5 /* matches a ERE regular expression? */
} operation;
+ regex_t *regex_cache; /* cache for compiled REs, if such are used */
cstr_t *pCSCompValue; /* value to "compare" against */
char isNegated; /* actually a boolean ;) */
} prop;