[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: leaks in slapd



I guess we need to add a config_back_db_close() and tear down in there. I'll take care of this.

Pierangelo Masarati wrote:
Howard Chu wrote:

For 64 bit builds you might try my current FunctionCheck (1.5.4) instead, for leak tracing. valgrind 3.0 is still very limited on AMD64, and its memory layout makes using it on large databases pretty much impossible.

Howard,

one "leak" (actually, reachable memory that is not freed) which I'm not sure to know how to fix is here:

==18946== 1912 bytes in 51 blocks are still reachable in loss record 6 of 7
==18946== at 0x1B9046C1: calloc (vg_replace_malloc.c:176)
==18946== by 0x820AD43: ber_memcalloc_x (memory.c:286)
==18946== by 0x80973D6: ch_calloc (ch_malloc.c:103)
==18946== by 0x8067E68: config_include (bconfig.c:2302)
==18946== by 0x806DDDE: config_set_vals (config.c:251)
==18946== by 0x806E2BF: config_add_vals (config.c:319)
==18946== by 0x806F30D: read_config_file (config.c:650)
==18946== by 0x8068E74: read_config (bconfig.c:2630)
==18946== by 0x806142F: main (main.c:636)


Where and when should I free that memory?

p.


SysNet - via Dossi,8 27100 Pavia Tel: +390382573859 Fax: +390382476497





--
 -- Howard Chu
 Chief Architect, Symas Corp.  http://www.symas.com
 Director, Highland Sun        http://highlandsun.com/hyc
 OpenLDAP Core Team            http://www.openldap.org/project/