Hi Ira, On 17:03 Wed 22 Apr , Ira Weiny wrote: > > Below is a patch which fixes an issue I had when using > ib_resolve_portid_str_via. When resolving via IB_DEST_GUID the > ib_resolve_guid_via function optionally uses the portid to attempt to set a > different subnet prefix. > > IMO I don't think portid should be an in/out parameter in > ib_resolve_portid_str_via. I happened to pass a portid object which was on > the stack and had some garbage data in it. It took me a while to figure out > that ib_resolve_portid_str_via was attempting to use that garbage data. > > To make this more clear I added ib_resolve_gid_via and another MAD_DEST type. > > What do you think? Right now the gid resolving is untested.
I'm not against such addition, but this patch is malformed. Sasha _______________________________________________ general mailing list [email protected] http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
