Ladies and Gentlemen, As discussed at the last conf. call we start running out of time at the last several conf. calls without completing the agenda. We have...
I propose that we postpone this errata item until DAPL-1.2 time. Three main reasons. 1. We can live without it for now. 2. ITWG has a completely different way...
The #10 and #18 errata items are covering 2 parts of the same issue of RDMA Read outstanding incoming - target and outgoing - Originator for Endpoint as well...
... Unfortunately, I don't think that I have more to say here than I originally did when I forwarded Peter's comments. Peter asked if there was a conflict...
I propose that we add one more error return for dat_evd_wait and dat_cno_wait: DAT_SIGNAL_INTERRUPT. The new text for descriptions: "The operation returns with...
In the 1/22 conference call, we briefly spoke of the problems I refered to in previous email surrounding the destruction of EVDs or CNOs simultaneously with...
I am happy to report that Sanmina-SCI of Huntsville Alabama has joined DAT Collaboartive. Welcome to DAT train. Arkady Kanevsky Chair of DAT Collaborative ...
Forgot one more Tentative proposal for CTN #51 (abort routines). I will add it and group with other evd|cno_wait race condition issue. Arkady Kanevsky...
Subject: new unreachable connection event and error, + (Advice to Implementors) IB transport connection error mappings Proposal: Add a new "Unreachable" event...
... To clarify, the real point is that you can post an operation which will result in an ordered completion. The DTO will not produce any *other* results,...
Yes. Consumer can post to Unoperational EP and if no post synchronous errors they will be flushed with the general completion order guarantee. Arkady Arkady...
I would also suggest returning dat_ep_unreachable from ep_connect if the SA indicates there are no paths between your IA and the remote IA. What should...
I would like to start discussion on Support for RDMA read outstanding negotiation as part of connection set up. Here is the high level summary of proposed...
The current DAPL registry API forces information to be duplicated; both the dat.conf file and the individual libraries must contain the information about which...
This proposal addresses CTN #11. I believe the text for dat_evd_wait() and dat_evd_dequeue() to be somewhat confusing. They each include the phrase: All...
The term "simultaneous" term implies that the non-determinancy only applies when you cannot state an order for the calls. More to the point, if there are...
... Agreed, that's what I was trying to get at. Are you suggesting a change in wording (or a preference for the original wording)? -- Randy Smith uDAPL...
Caitlin, Are you trying to add a new definition to Terminology section that will be used in multiple places, like thread safety section, ia_close, wait calls...
Ladies and Gentlemen, You have till Monday 5:00pm EST to provide your feedback for the Connection model changes proposal (below). Here are the portion of...
I will give it one more spin. What if require that the disassociation of the last feeding event stream for EVD/CNO will automatically aborts waiters? Arkady ...
Per agreement at the last conf. call You have till Tuesday 2/11 5:00pm PST to provide youe feedback if you have not done so already. Arkady Kanevsky...
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
