Love, Robert W wrote: > Bhanu (Venkata Bhanu Prakash) Gollapudi wrote: >> Hi Joe/Robert, >> >> I have a question in the related area of discussion. >> >> Can we obtain the event_callback function in fc_lport_rport_ops and >> fc_disc_rport_ops from the libfc_function_template rather than having >> static functions(fc_lport_rport_callback and fc_disc_rport_callback). >> This would provide flexibility to LLDs to take any driver specific >> additional actions on that session along with the default action >> performed by these functions (adding or removing the rport from the >> list). Please let me know your thoughts. >> > Sounds reasonable to me. > > If Joe is moving the rport list to the rport layer (not sure why it > isn't moving to the lport layer) then he'll probably be moving the > fc_disc_rport_callback() out of the disc layer and into the rport > layer.
I guess the rport list could be put in the lport layer, but I wanted to have rport_create also add the new rport to the list. It seems better in the rport layer. I'm open to changes along those lines. On fc_disc_rport_callback(): I'm getting rid of that by having the rport layer do the list maintenance. Since that's all that callback did, there isn't a need for it in fc_disc. fc_lport still uses an rport event callback to find out when the directory server rport or point-to-point rport is ready. We could have a transport-template callback for all remote ports as well if there's a need by LLDs. That would be in addition to the rport callback that was specified via rdata->ops, and we would probably call that before the other one on create, and after the other one on delete, I guess. I think we would only need one callback, called rport_event_handler(). > Keep that in mind as depeneding on when development is > done there are likely to be conflicts between the two patch sets, so > one set will likely depend on the other. > > Your change would add two _ops pointers into the tt one for the > "lport" (dns) callbacks and the other for the "rport" (currently > disc) callbacks. I think they should be named something like > rport_event_handler (assuming Joe is moving the list to the rport layer) > and lport_rport_event_handler (ugh, this name is long). > _______________________________________________ > devel mailing list > [email protected] > http://www.open-fcoe.org/mailman/listinfo/devel _______________________________________________ devel mailing list [email protected] http://www.open-fcoe.org/mailman/listinfo/devel
