Re: known mem leaks?

["Kurt D. Zeilenga" <Kurt@OpenLDAP.org>]

At 11:37 AM 9/1/99 -0400, shltch@osfmail.isc.rit.edu wrote:
>Hi all,
>
>I've been doing a lot of testing with OpenLDAP 1.2.6.

Try OPENLDAP_REL_ENG_1_2 (available via Anonymous CVS).

A number of changes have been committed to resolve memory
management problems.

>I am using the
>development tools atom and third on DEC Unix.  I am investigating why
>there are so many memory leaks.  I've found what seems to be a huge
>problem, and I'm wondering if anyone else has identified this or knows
>about this.  The output from my tools are:
>
>6560 bytes in 205 leaks (including 138 super leaks) created at:
>    malloc                         libc.so
>    ch_malloc                      slapd
>    attr_merge                     slapd
>    do_add                         slapd
>    pc = 0x12007ee0c               slapd
>
>3936 bytes in 246 leaks (including 116 super leaks) created at:
>    malloc                         libc.so
>    ber_bvdup                      slapd
>    value_add                      slapd
>    attr_merge                     slapd
>    do_add                         slapd
>    pc = 0x12007ee0c               slapd
>
>3608 bytes in 205 leaks (including 126 super leaks) created at:
>    malloc                         libc.so
>    ch_malloc                      slapd
>    value_add                      slapd
>    attr_merge                     slapd
>    do_add                         slapd
>    pc = 0x12007ee0c               slapd
>
>3239 bytes in 246 leaks (including 135 super leaks) created at:
>    malloc                         libc.so
>    ber_bvdup                      slapd
>    value_add                      slapd
>    attr_merge                     slapd
>    do_add                         slapd
>    pc = 0x12007ee0c               slapd
>
>As you can see, there are many leaks created in the attr_merge function
>(from servers/slapd/attr.c)  It looks as if when the new Attribute struct
>is being created at the end of the linked list, it might be reclaimed.
>Any ideas?
>
>Thanks, and I'll keep looking,
>Seth
>
>