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

(ITS#4663) Abandoning a bind operation?



Full_Name: Pierangelo Masarati
Version: HEAD/re23
OS: irrelevant
URL: ftp://ftp.openldap.org/incoming/
Submission from: (NULL) (87.28.220.33)
Submitted by: ando


The proxies may occasionally send abandon when a bind does not complete in a
timely manner.  This is contrast with rfc4511, which clearly states that

   After sending a BindRequest, clients MUST NOT send further LDAP PDUs
   until receiving the BindResponse.  Similarly, servers SHOULD NOT
   process or respond to requests received while processing a
   BindRequest.

So I guess the proper way to dispose of an asynchronous bind operation consists
in calling ldap_unbind(), although this sends a PDU as well...

I note that the current behavior of slapd (HEAD, re23) when subjected to

    BIND REQUEST
    ABANDON REQUEST
    BIND REQUEST
    (wait for) BIND RESPONSE

consists in hanging when response is waited for from the second bind request,
while when subjected to 

    BIND REQUEST
    UNBIND REQUEST
    BIND REQUEST
    (wait for) BIND RESPONSE

results in the expected behavior.  I'll rework the proxies to unbind instead of
abandoning, but I'd like to hear some comments about this.

p.