Quoting r. Hal Rosenstock <[EMAIL PROTECTED]>: > Subject: Re: [PATCH for-2.6.18] IB/cma: option to limitMTU to 1K > > On Wed, 2006-09-13 at 17:17, Michael S. Tsirkin wrote: > > Quoting r. Hal Rosenstock <[EMAIL PROTECTED]>: > > > > If path MTU selector in path query allows MTU 1K (e.g. "best MTU") > > > > and one of the sides is Tavor, select the best MTU that is 1K > > > > and not the largest possible. > > > > > > How would it be identified if the SA supports this ? > > > > You mean, if SA ignores mtu selector? > > No; I meant detect that one end of the PR request is a Tavor. Wasn't > that part of it ?
SA can easily figure out it's talking to tavor by looking at vendor part id. > If SA doesn't support MTU selector and ignoring MTU selector, it is not > compliant and should be fixed. > > > Then we are not worse off than we were before we set it - we get 2K MTU for > > tavor and it works a bit slower. > > > > > > If path MTU selector requires 2K MTU, return path with 2K MTU. > > > > > > Also, I'm not sure that this is the required difference in the SA > > > requests :-( > > > > What do you mean? > > Its not required, but its legal and it will give us better performance. > > I mean that there is no requirement on what the IPoIB SA PR request > looks like what you are using to differentiate from the PR requests for > a connection setup. Correct. But if IPoIB requires 2K MTU it must use MTU selector, if it does not it's OK to give it a smaller MTU. -- MST _______________________________________________ openib-general mailing list [email protected] http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
