13 \u0441\u0435\u043d\u0442\u044f\u0431\u0440\u044f 2016�\u0433. 19:26:59 CEST, "Bob Friesenhahn bfriesen@... [milter-greylist]" <milter-greylist@yahoogroups.com> \u043f\u0438\u0448\u0435\u0442: >On Mon, 12 Sep 2016, Jim Klimov jimklimov@... [milter-greylist] >wrote: >> >> We use p0f (3.06b, 3.08b iirc is last) coerced to compile under >> solarish oses that we use (tweaks should be on my github). Depending >> on platform release we had libpcap issues that it processed packets >> by larger buffers at once, so p0f might not yet have answers when >> needed. > >Does using the p0f feature increase the opportunity for a security >weakness so it is more likely that the host machine can be >compromised? > >Can it work in VMs, containers, or Solaris zones, which are not >allowed access to raw packets due to network security concerns? > >Bob In VMs we had a problem that it detected the hypervisor's OS as the local one (e.g. thinking it was windows while it was a solaris in virtualbox), but I think proper bridging maybe over a dedicated nic solved that. Jim -- Typos courtesy of K-9 Mail on my Samsung Android
Message
Re: [milter-greylist] problem building with p0f support, p0f or better using dialin RBL?
2016-09-13 by Jim Klimov