* Herbert Xu <[EMAIL PROTECTED]> 2005-08-08 21:49 > On Mon, Aug 08, 2005 at 12:46:30PM +0200, Patrick McHardy wrote: > > > > part on top of netlink). Right now there are none, so this won't cause > > any trouble, the question is if we want to retain the possibility or > > just don't care about this case. > > I'd say that as it is general use of netlink between two user-space > processes is highly undesirable because of the ease of disruption > of that communication by a third party. There is simply no way to > stop a a malicious third party from filling up your socket receive > buffer. > > So any serious use of netlink in user-space will require some major > rework on the infrastructure, which may have to wait till netlink2. > Therefore, I don't see any problems with removing ability for > user-space processes to bind to netlink families that aren't > registered in the kernel.
I tend to agree with Herbert, I don't think it is worth to solve this problem in a perfect matter. Iff we really need it, we should simply have the numbers of groups per protocol statically stored at some point for all protocols used in the kernel so in case a userspace socket gets created first we allocate the correct bitmask size. In case a protocol is only used in userspace we simply don't allow multicasting until a setsockopt is issued specifying the maximum number of groups. - To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
