Freeing QN=1 State
"Michael C. Cambria" <[email protected]> Fri, 10 Dec 2004 12:50:10 -0500
| Newsgroups | gmane.ietf.rddp |
|---|---|
| Message-ID | <[email protected]> |
When does draft-ietf-rddp-rdmap-02.txt consider an inbound RDMA Read Request Message finished such that the state associated with QN=1 (the IRRQ) for the RDMA Read Request can be freed? Put another way, if IRRQ is 1 (for simplicity), when is it legal for a new RDMA Read Request to be received? Is it: a) Once the RDMA Read Response Message has been given to DDP? b) Once the RDMA Read Response Message has been "acked" by the data sink? c) up to the implementation? d) specified somewhere that I just missed? Using (a) as a solution I ran into a DoS problem. The data sink keeps issuing RDMA Read Requests but the LLP window advertised by the data sink is so small (or closed) that the resources I believe the IRD negotiation was meant to protect actually get depleted. A compliant sender should not send the next RDMA Read Request until a prior one has completed. However, is the draft assuming a non malicious sender? Personally, I prefer to protect my resources regardless. I can't find supporting text in the draft for (b). Doing (c) allows for (a) so is the same problem IMHO. Thanks, MikeC