After talking to him in more depth off list it turns out the char** returned by ldap_get_values was being int cast by gcc and thus the higher word was corrupted. This caused it only to work for memory malloced below a specific threshold on his machine. Plus he was using the deprecated ldap_get_values instead of ldap_get_values_len.
Now that he's using ldap_get_values_len everything works as intended.
-William
On 09/23/2011 05:23 PM, Philip Guenther wrote:
On Tue, 20 Sep 2011, kyle king wrote:
On 09/20/2011 06:21 PM, kyle king wrote:
Under heavy threaded use ldap_get_value returns a corrupted memory pointer, not a char pointer pointer that was specified.
I have narrowed the corruption problem down to the file libraries/libldap/get_values.c apoximatly line 80 the ber_scanf in the if statement.
When it work on my machine, it returns pointers for 'vals' such as:
vals pointer: 0x105c920
and when it fails: vals pointer 0x7f6130004c20
Is this a follow up to a previous posting that actually described what you were going? If so, I don't see it in the list archives. So:
What version of OpenLDAP?
Is the client linked against libldap_r or libldap?
What's the model for how threads are using LDAP handles? E.g.,
- each LDAP handle used by only one thread at a time, from operation to result
- each LDAP handle used concurrently by many threads, with app-level locks
- each LDAP handle used concurrently by many threads, without app-level locks
Is TLS/SSL involved? If so, are the thread-safety requirements of that library being independently set up by the application? (IIRC, using libldap_r handled that for at least OpenSSL, but I haven't stayed up on whether that's still the case or whether it applies to GNUtls or NSS.)
Philip Guenther