Min Xu (Hsu) wrote: > On Thu, 04 Dec 2008 Cam Macdonell wrote : >> Aaron Rolett wrote: >>> Hi Cam, >>> Are you talking about the VMCI Sockets interface from within the >>> kernel on linux? If so ... I don't think anyone has tried this yet ... >>> but I *think* it should mostly work. One thing that might need a little >>> work is registering the dynamic address family value since that is >>> currently done through /dev/vsock in userspace. This is something I'd >>> like to try out when I have a chance. >> Yup, within Linux. >> >>> With regards to VMCI Sockets on windows in kernel mode ... we >>> don't currently have support for this. >> > >>> Out of curiosity ... what are you trying use VMCI Sockets for? >> I'm actually interested in the shared memory VMCI interface, not in the >> socket/datagram one at this point. Because of this I am using WS 6.0.4 >> due to shared memory's deprecation in 6.5 (and newer open-vm-tools). >> >> To answer your question, what I would like to do is use shared memory as >> a file cache between VMs on the same host that would all be accessing a >> remote file system. In particular I'm working with Samba. I was using >> fusesmb so I could work at user level, but fusesmb is flaky and doesn't >> seem to be an active project in terms of development. So, I'm thinking >> of using the CIFS kernel module to work on my implementation. > > Hi Cam, >
Hi Min, Thanks for your thoughts,. You make good points and they made me think. > Why use a deprecated API? I can only reply: Why deprecate an API people are using? > I know shared memory maybe easier to program > than datagram. But shared memory model "de-virtualize" your VM: make it > hard to suspend/resume or record/replay a VM. Datagram at least force > someone to deal with error conditions, like re-connect, timeout, etc. I can certainly see how a datagram model has some advantages. But, I'm trying to do caching and for this purpose using shared memory requires less concurrent programming (threads, etc) and it reduces unnecessary copying of the data. I'm looking at VMs in the context of distributed high-performance computing. Our VMs are disposable sandboxes in this use, so suspension and replay are not features we need. Thanks again for your 2 cents :) Cam ------------------------------------------------------------------------------ SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada. The future of the web can't happen without you. Join us at MIX09 to help pave the way to the Next Web now. Learn more and register at http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/ _______________________________________________ open-vm-tools-devel mailing list open-vm-tools-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/open-vm-tools-devel
