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

Re: overlay response to bind operation



Howard Chu wrote:

Pierangelo Masarati wrote:

Pierangelo Masarati wrote:

For instance, the rwm (rewrite/remap) overlay is likely to leak each bind when the bind DN is rewritten.



Please disregard, this is not true. The rwm doesn't leak in case of bind because the original o_req_dn is deleted in case of rewite prior to leaving the overlay's bind operation.


Still it makes sense that the frontend should insure that all overlay cleanup functions are called for a successful bind.

Yes. My first attempt was to look at the op->o_callback leftovers after the execution of the original backend's function in over_op_func(), which should be exactly the backover's response callback. If it's not, we could call in sequence at least the sc_cleanup() function, and overlays that may leave dangling resources should provide appropriate cleanup callbacks.


p.



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