Using TrackHostExits with a long TrackHostExitsExpire should mean
Exits are only rotated if they go down. I don't see a reason why
we should forget the trackings we set up earlier when we got a hup
and the TrackingHostExits* options didn't change, especially because
some Tor packages like the debian package hup tor frequently.
[Automatically added by flyspray2trac: Operating System: All]
To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information
Child items ...
Show closed items
Linked items 0
Link issues together to show that they're related.
Learn more.
Worth fixing if that's what people want, though I suspect there are others who definitely do want their tracked exits reset on a HUP. Seems like a fine case for a boolean option.
This should be fairly easy for anybody who reads C and can track the control flow for the SIGNAL command from handle_control_signal() down through signal_callback() to do_hup().
Trac: Description: Using TrackHostExits with a long TrackHostExitsExpire should mean
Exits are only rotated if they go down. I don't see a reason why
we should forget the trackings we set up earlier when we got a hup
and the TrackingHostExits* options didn't change, especially because
some Tor packages like the debian package hup tor frequently.
[Automatically added by flyspray2trac: Operating System: All]
to
Using TrackHostExits with a long TrackHostExitsExpire should mean
Exits are only rotated if they go down. I don't see a reason why
we should forget the trackings we set up earlier when we got a hup
and the TrackingHostExits* options didn't change, especially because
some Tor packages like the debian package hup tor frequently.
Because debian's tor package hups tor daily, I think this one is important to resolve in the 0.2.3 timeframe: it can quite reasonably bite users in a way they'll find impossible to track down.
Speaking of which, is there a full list of things that Tor will flush/reset/etc on hup? The list has changed over the years, but if a user wants to predict whether this will happen, where would she look?
Trac: Points: N/AtoN/A Priority: minor to major Summary: Exits tracked with TrackHostExits reset on hup to Exits tracked with TrackHostExits should not reset on hup Actualpoints: N/AtoN/A
The patch is easy enough to do. One thing to note, though: right now, addressmap_clear_transient() clears three things: DNS results, virtual address mappings, and trackexithost mappings. Do we want to clear all of these on HUP? Some? None?
The patch is easy enough to do. One thing to note, though: right now, addressmap_clear_transient() clears three things: DNS results, virtual address mappings, and trackexithost mappings. Do we want to clear all of these on HUP? Some? None?
We want to clear TrackHostExits mappings on SIGHUP iff they are for hosts not listed in TrackHostExits after the HUP. More generally, we want to clear a TrackHostExits mapping whenever TrackHostExits changes and the mapped host is not in the new TrackHostExits list.
We should clear virtual address mappings on SIGHUP iff AutomapHostsOnResolve is off after the HUP.