Re: [PATCHES] PATCH: Memory leaks on start-up
Tom, happier with the attached patch?
I'd have to disagree with regards to the memory leaks not being worth
a mention - any such leak can cause problems when the PostgreSQL
installation is either unattended, long-living andor has very high
connection levels. Half a kilobyte on start-up isn't negligible in
this light.
Regards, Lee.
Tom Lane writes:
Lee Kindness [EMAIL PROTECTED] writes:
Guys, attached is a patch to fix two memory leaks on start-up.
I do not like the changes to miscinit.c. In the first place, it is not
a memory leak to do a one-time allocation of state for a proc_exit
function. A bigger complaint is that your proposed change introduces
fragile coupling between CreateLockFile and its callers, in order to
save no resources worth mentioning. More, it introduces an assumption
that the globals directoryLockFile and socketLockFile don't change while
the postmaster is running. UnlinkLockFile should unlink the file that
it was originally told to unlink, regardless of what happens to those
globals.
If you are intent on spending code to free stuff just before the
postmaster exits, a better fix would be for UnlinkLockFile to free its
string argument after using it.
Index: src/backend/libpq/pqcomm.c
===
RCS file: /projects/cvsroot/pgsql-server/src/backend/libpq/pqcomm.c,v
retrieving revision 1.157
diff -u -r1.157 pqcomm.c
--- src/backend/libpq/pqcomm.c 12 Jun 2003 07:36:51 - 1.157
+++ src/backend/libpq/pqcomm.c 22 Jul 2003 14:16:46 -
@@ -363,7 +363,7 @@
added++;
}
- freeaddrinfo(addrs);
+ freeaddrinfo2(family, addrs);
if (!added)
{
Index: src/backend/utils/init/miscinit.c
===
RCS file: /projects/cvsroot/pgsql-server/src/backend/utils/init/miscinit.c,v
retrieving revision 1.104
diff -u -r1.104 miscinit.c
--- src/backend/utils/init/miscinit.c 27 Jun 2003 19:08:37 - 1.104
+++ src/backend/utils/init/miscinit.c 22 Jul 2003 14:16:46 -
@@ -673,8 +673,15 @@
static void
UnlinkLockFile(int status, Datum filename)
{
- unlink((char *) DatumGetPointer(filename));
- /* Should we complain if the unlink fails? */
+ char *fname = (char *)DatumGetPointer(filename);
+ if( fname != NULL )
+{
+ if( unlink(fname) != 0 )
+ {
+ /* Should we complain if the unlink fails? */
+ }
+ free(fname);
+}
}
/*
---(end of broadcast)---
TIP 7: don't forget to increase your free space map settings
[PATCHES] ECPG: Update tests memory leak fix
See attached diff which corrects a per-thread memory leak in libecpg and changes the thread test program to create a user defined number of threads. Also attached tar.gz adds an additional thread test, which currently does NOT work as expected and is one of the cases Shridar and I will be looking at (removing need for AT xxx on every EXEC SQL). Note, the updated test shows that ECPG currently DOES support the specification of connection names by host variables (not only static strings as I had said previously). Thanks, L. ecpg.diff Description: Binary data ecpg-test.tar.gz Description: GNU Zip compressed data ---(end of broadcast)--- TIP 5: Have you checked our extensive FAQ? http://www.postgresql.org/docs/faqs/FAQ.html
Re: [PATCHES] Update tests memory leak fix
Sorry guys, last diff wasn't a context one! diff -c attached this time round... L. - Original Message - From: Lee Kindness To: [EMAIL PROTECTED] Cc: Bruce Momjian ; Lee Kindness ; Shridhar Daithankar Sent: Saturday, March 06, 2004 3:07 PM Subject: ECPG: Update tests memory leak fix See attached diff which corrects a per-thread memory leak in libecpg and changes the thread test program to create a user defined number of threads. Also attached tar.gz adds an additional thread test, which currently does NOT work as expected and is one of the cases Shridar and I will be looking at (removing need for AT xxx on every EXEC SQL). Note, the updated test shows that ECPG currently DOES support the specification of connection names by host variables (not only static strings as I had said previously). Thanks, L. ecpg.diff Description: Binary data ---(end of broadcast)--- TIP 9: the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match
[PATCHES] ECPG - Remove need for AT connection when using threads
Attached diff improves the handling of connections in ecpg when threading is used. The current connection is now the last connection made/set for the thread in question - not the absolute last connection made/set (over all threads in the process). This change removes the tedious need to always use AT connection for each and every EXEC SQL call (an item on the TODO). The test_thread_implicit test program added in the last email will function as intended with this patch. Thanks, L. - Original Message - From: Lee Kindness To: [EMAIL PROTECTED] Cc: Bruce Momjian ; Lee Kindness ; Shridhar Daithankar Sent: Saturday, March 06, 2004 3:07 PM Subject: ECPG: Update tests memory leak fix See attached diff which corrects a per-thread memory leak in libecpg and changes the thread test program to create a user defined number of threads. Also attached tar.gz adds an additional thread test, which currently does NOT work as expected and is one of the cases Shridar and I will be looking at (removing need for AT xxx on every EXEC SQL). Note, the updated test shows that ECPG currently DOES support the specification of connection names by host variables (not only static strings as I had said previously). Thanks, L. ecpg-connect.diff Description: Binary data ---(end of broadcast)--- TIP 6: Have you searched our list archives? http://archives.postgresql.org
Re: [PATCHES] ECPG - Remove need for AT connection when using threads
The cvs add of test_thread_implicit.pgc seems to have been missed,
i've attached this again.
Additionally I include a small patch to remove mutex locking when a
DEFAULT/NULL connection is being retrieved. This is consistent with
libpq.
Thanks, L.
Michael Meskes writes:
On Sun, Mar 14, 2004 at 09:11:27AM -0500, Bruce Momjian wrote:
I just applied this patch and the last one.
I assume you mean you applied:
Update tests memory leak fix
and
ECPG - Remove need for AT connection when using threads
Yes. Sorry, should have said so.
Michael
/*
* Thread test program
* by Lee Kindness.
*/
/* #define ECPGDEBUG */
#include pthread.h
#include stdlib.h
void *test_thread(void *arg);
EXEC SQL BEGIN DECLARE SECTION;
char *l_dbname;
EXEC SQL END DECLARE SECTION;
int nthreads = 2;
int iterations = 10;
int main(int argc, char *argv[])
{
#ifdef ECPGDEBUG
char debugfilename[] = thread_test_implicit.log;
FILE *debugfile;
#endif
pthread_t *threads;
int n;
EXEC SQL BEGIN DECLARE SECTION;
int l_rows;
EXEC SQL END DECLARE SECTION;
/* parse command line arguments */
if( (argc 2) || (argc 4) )
{
fprintf(stderr, Usage: %s dbname [threads] [iterations_per_thread]\n, argv[0]);
return( 1 );
}
l_dbname = argv[1];
if( argc = 3 )
nthreads = atoi(argv[2]);
if( argc == 4 )
iterations = atoi(argv[3]);
/* open ECPG debug log? */
#ifdef ECPGDEBUG
debugfile = fopen(debugfilename, w);
if( debugfile != NULL )
ECPGdebug(1, debugfile);
else
fprintf(stderr, Cannot open ECPG debug log: %s\n, debugfilename);
#endif
/* setup test_thread table */
EXEC SQL CONNECT TO:l_dbname;
EXEC SQL DROP TABLE test_thread; /* DROP might fail */
EXEC SQL COMMIT;
EXEC SQL CREATE TABLE
test_thread(tstampTIMESTAMP NOT NULL DEFAULT CAST(timeofday() AS TIMESTAMP),
threadTEXT NOT NULL,
iteration INTEGER NOT NULL,
PRIMARY KEY(thread, iteration));
EXEC SQL COMMIT;
EXEC SQL DISCONNECT;
/* create, and start, threads */
threads = calloc(nthreads, sizeof(pthread_t));
if( threads == NULL )
{
fprintf(stderr, Cannot alloc memory\n);
return( 1 );
}
for( n = 0; n nthreads; n++ )
{
pthread_create(threads[n], NULL, test_thread, (void *)n + 1);
}
/* wait for thread completion */
for( n = 0; n nthreads; n++ )
{
pthread_join(threads[n], NULL);
}
free(threads);
/* and check results */
EXEC SQL CONNECT TO :l_dbname;
EXEC SQL SELECT COUNT(*) INTO :l_rows FROM test_thread;
EXEC SQL COMMIT;
EXEC SQL DISCONNECT;
if( l_rows == (nthreads * iterations) )
printf(\nSuccess.\n);
else
printf(\nERROR: Failure - expecting %d rows, got %d.\n, nthreads * iterations,
l_rows);
/* close ECPG debug log? */
#ifdef ECPGDEBUG
if( debugfile != NULL )
{
ECPGdebug(0, debugfile);
fclose(debugfile);
}
#endif
return( 0 );
}
void *test_thread(void *arg)
{
long threadnum = (long)arg;
EXEC SQL BEGIN DECLARE SECTION;
int l_i;
char l_connection[128];
EXEC SQL END DECLARE SECTION;
/* build up connection name, and connect to database */
snprintf(l_connection, sizeof(l_connection), thread_%03ld, threadnum);
EXEC SQL WHENEVER sqlerror sqlprint;
EXEC SQL CONNECT TO :l_dbname AS :l_connection;
if( sqlca.sqlcode != 0 )
{
printf(%s: ERROR: cannot connect to database!\n, l_connection);
return( NULL );
}
EXEC SQL BEGIN;
/* insert into test_thread table */
for( l_i = 1; l_i = iterations; l_i++ )
{
printf(%s: inserting %d\n, l_connection, l_i);
EXEC SQL INSERT INTO test_thread(thread, iteration) VALUES(:l_connection, :l_i);
if( sqlca.sqlcode == 0 )
printf(%s: insert done\n, l_connection);
else
printf(%s: ERROR: insert failed!\n, l_connection);
}
/* all done */
EXEC SQL COMMIT;
EXEC SQL DISCONNECT :l_connection;
printf(%s: done!\n, l_connection);
return( NULL );
}
Index: src/interfaces/ecpg/ecpglib/connect.c
===
RCS file: /projects/cvsroot/pgsql-server/src/interfaces/ecpg/ecpglib/connect.c,v
retrieving revision 1.20
diff -c -r1.20 connect.c
*** src/interfaces/ecpg/ecpglib/connect.c 14 Mar 2004 12:16:29 - 1.20
--- src/interfaces/ecpg/ecpglib/connect.c 15 Mar 2004 16:17:36 -
***
*** 62,79
{
struct connection *ret = NULL;
#ifdef ENABLE_THREAD_SAFETY
! pthread_mutex_lock(connections_mutex);
#endif
! ret = ecpg_get_connection_nr(connection_name);
#ifdef ENABLE_THREAD_SAFETY
! pthread_mutex_unlock(connections_mutex);
#endif
return (ret);
-
}
static void
--- 62,89
{
struct connection *ret = NULL;
+ if ((connection_name == NULL) || (strcmp(connection_name, CURRENT) == 0))
+ {
#ifdef
