On Tue, 12 May 2009 14:48:43 -0400 "Anurag S. Maskey" <Anurag.Maskey at Sun.COM> wrote:
> code review requested for the following bugs: > > 8782 still some lingering upgrade problems > 8775 too many cooks spoil the vboxnet interface > 8763 missing/incomplete validation functions for property values > > http://zhadum.east.sun.com/export/ws/am223141/checkout-area/nwam1-fixes/webrev/ > > The changes in usr/src/cmd/svc/milestone/net-loc and > usr/src/cmd/svc/milestone/net-nwam should supposedly fix bug 8782. I've > asked for log files to dig deeper, but this is one of the preliminary > issues that I've seen. > > For Bug 8763, the default-domain and dns-nameservice-search can have any > value. Also, valid mtu values cannot be determined unless the mtu is > being set and the driver complains. The same is true for autopush > modules (I think). Since we allow configurations to be migrated between > machines, the module may not exist in the system. Also, I did not find > a way to test whether a module name is valid or not without actually > pushing the module. Kinda strange that valid_uint64 is really for a 63 bit uint ;) getprefixlen actually mods it input buffer which is unclear from its name valid_host_v4 should validate a prefix (cidr netmask) at the end valid_host_v6 should validate a prefix
