by VistaFirewallControl » Tue Jul 13, 2010 9:02 am
>And the last sample I posted was essentially from that log. If WF was blocking a critical HostProcess, then wouldn't turning it off disable the blocking? This type of entry has been constant, no matter the setting for WF or W7FC:
7/9/2010|1:48:58 PM|IPv4 TCP 127.0.0.1:55159(56114)|Host Process for Windows Services| Incoming
Well… Everything we could invent as the explanation is actually conjectures. We have no information about the gadget internals and we have no information on such WindowsFilteringPlatform (WFP) details either.
Why the gadget requires (regardless of explicitly or implicitly) the incoming to HostProcess is still foggy. Most probably the activity is very indirect and caused by a gadget intermediary/third party.
However WF is just a control panel to WFP. So turning off WF does not turn off WFP. Also senseless to rely that everything modified (by WF in WFP) on WF switching on is cleared on WF switching off.
Most probably WF running/switching state means nothing at all to WFP. WFP is smart enough and able to manage the security rules on its own. We have confirmed cases of such behavior.
For instance on several systems with Windows7 if W7FC tries to set a globally enabling rule (Mode:EnableAll for instance) and WF is switched off, WFP “understands” that all the most dangerous incomings are enabled as well, and sets the corresponding blocking rule on its own. The remedy is switching WF on and enabling the incomings (if required). Looks like it happens not every time on switching Mode:EnableAll on selected systems
>Since I keep seeing this same localhost entry,
The activity may be required only once for probing something. If the probe attempt was blocked everything the rest stopped working. If the probe was passed everything works as expected regardless of whether the blocks were made. Actually it’s the third party application (the gadget + the processing engine) behavior. We are still not sure that the HostProcess|Incoming blocking is the genuine reason. It was just a “hook” we used to mesh with to diagnose the problem. It’s just a visible part of the “iceberg”. There is no documentation for how the gadget processing engine interacts with the entire system including HostProcess, which is actually a stub for multiple different system services.
I looked at my HOSTS file. And I discovered a couple of statements that must have been added by some software.
>Could there be some the gadget related?
>So, do you think that is what is generating all these identical entries?
One of the HostProcess instances. We will hardly be able to know the genuine reason. Actually the Windows “world” is famous of similar depencies.
Do you know any other system whose the first (and the most efficient, the most recommended) remedy to solve any serious problem is reboot.
>If so, the servers on the other end are trying every second to connect . . . persistent buggers.
No,no. The activity is local only. 127.0.0.1 is the local address. And an application from your own system tries to connect to HostProcess with the incomings.
Most probably the application is the same or another instance of HostProcess as well. It’s so called inter process communication.
>Finally, let's say that at some point I stop running any of these Win7 Gadgets. Should I then reset the W7FC zones you had me change back to their defaults? However, I forgot those defaults.
Seems you have already set everything to default/recommended - "Windows Desktop Gadgets" to WebBrowserZone and "HostProcess" to Local+DNS+DHCP+Update(svchost) (re-apply the zone, get the unmodified zone from the repository, just in case)
If you do not need Windows Desktop Gadgets listed in W7FC you can delete it from the list.