On Fri, Jun 19, 2009 at 11:48 AM, Rob Landley<[email protected]> wrote: >> * RFC2518 appears to be in error when it says that you should set the >> most significant bit of the first octet of the node ID to 1. RFC4122 >> instead says you should set the *least* significant bit, which >> matches the description of IEEE 802 MAC addresses in >> http://en.wikipedia.org/wiki/MAC_address. > > I vaguely recall figuring out what it _should_ do experimentally (looking at > hex dumps) and then documenting it based on what the RFC said. But I wrote > that code back around 2007, so I don't really remember the details all that > clearly. Feel free to tweak if it makes you feel better. > >> * Your code uses uuid[11] as if it were the first octet of the node ID. >> I think this should in fact be uuid[10]. > > Endianness issue, I expect. (The code I started with was passing in a > structure and I was treating it as a char array. It probably did endianness > conversions later on.)
I have # mkswap --version mkswap (util-linux-ng 2.13.1) and experimentally it does not seem to set or clear any bits in node id. It only sets version and variant. > The actual get_random_bytes implementation does various voodoo to try to make > the output of /dev/urandom _more_ random, which is about on par with calling > memset twice on the same data just to be sure, and there's a _reason_ it was > removed from busybox. I still do that, for the case when /dev/urandom does not exist. Someone may be mkswap'ing before /dev is fully populated. > (My understanding is that the old_e2fsprogs directory > is there as a reference for when we get around to writing a real mke2fs for > busybox.) yes -- vda _______________________________________________ busybox mailing list [email protected] http://lists.busybox.net/mailman/listinfo/busybox
