Net2WG/Notes/20080118

---

---

Agenda

---

* CTP experiments and plans (postponed)
* DIP tests updates (postponed)
* Deluge

Participants

---

* Om, USC
* Razvan, JHU

Discussion Notes

---

Deluge

Razvan: Right now I'm mostly fixing bugs.

Om: There were also some feature requests. Let's talk about both.

Razvan: So in the beginning it was this guy David that asked 4 very specific things. I get them done almost all them except one of them which was related to some corner cases. One example is if you try to inject an image and that image you are currently disseminating. We didn't have any mechanism in place to prevent this. Because of this I restructure a little the way we do things. So now we have a Resource interface that we can use to obtain a lock and then release it. With this occasion I also simplified the format of the image. When we put it there we tried to maintain the compatibility with Deluge 2 from T1 but it doesn't make to do that anymore. I mean it didn't prove to be useful. So I took out everything that was redundant or we were not really using. After I did this some things changes, some commands worked different and this cause some problems. I documented them but I still need to do a better job at that. Also a bunch of small bugs appeared. For example, the last one I solved is like this: when a node boots he needs to set its node ID because the image will contain a bad one. This was not really happening though so after a reboot all of them come up with the same node ID.

Om: It would be good to send an email to net2 about these changes. Also if you have some thoughts about the future of Deluge you should include that as well.

Razvan: Something else: do you remember that email about Dissemination? The one about local updated not being signaled.

Om: Yeah. I think you and I agree that is should be.

Razvan: I notice this not by reading the TEP but in the Deluge T2 code. The thing is that if dissemination will change then I'll have to change Deluge. Right now the behaviour of dissemination helps me.

Razvan: Another thing is related about getting the hardware address of a volume. I wrote an email about this but nobody answered.

Om: Address of what?

Razvan: When I need to reprogram a mote I need to inform TOSBoot by writing in an a special place the physical address in the external flash of the image. So in Deluge I need to come up with this address. The problem is there is a specific way to do it for the two types of flash, the one from MicaZ and the one from Telos. If would be nice to have only one way to do this. So...

Om: Yeah.

Razvan: I put code that is doing this but this is probably not a very nice solution. If people are happy with that I'm ok.

Om: That's the kind of stuff you should put in that write-up. I don't know the answer to that one.

Razvan: I wrote on devel but nothing happen.

Om: I'll keep and eye and remind people. That's the only way.

Razvan: That's pretty much everything.