[openib-general] Re: CMA disconnect

[Tom Tucker]

On Mon, 2006-05-08 at 09:58 -0700, Sean Hefty wrote:
> Or Gerlitz wrote:
> > Looking in the code i have realized that it is a must for the CMA 
> > consumer to call rdma_disconnect to have the QP state moved into ERROR.
> 
> Maybe it would make sense for the CMA to transition the QP to the error state 
> before destroying it?

FYI: For iWARP, moving the QP to ERR is a CM RESET, and moving it to SQD
is a graceful shutdown.  Keeping these differences hidden from the app
inside rdma_disconnect is probably a good idea.

> 
> > Am i correct? with this understanding at hand, i have changed iSER code to 
> > call rdma_disconnect even if it got a DISCONNECTED event caused by the 
> > passive side initiating the disconnect flow (ie sending a DREQ), since 
> > otherwise in such case i never got the FLUSHES on the posted RX/TX WRs.
> 
> You are correct, rdma_disconnect() is required to transition the QP into the 
> error state.  The code is written such that both sides call it.
> 
> - Sean
> _______________________________________________
> openib-general mailing list
> openib-general at openib.org
> http://openib.org/mailman/listinfo/openib-general
> 
> To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general