summaryrefslogtreecommitdiff
path: root/sbin/unwind
AgeCommit message (Expand)Author
2019-12-03No need to store "why_bogus" with the resolver, we are no longerFlorian Obser
2019-12-03Remove useless log_debug() calls.Florian Obser
2019-12-03No more status subcommands; ok florian@Otto Moerbeek
2019-12-02Save the computed median to avoid having it to compute it all the time;Otto Moerbeek
2019-12-02Use a unified cache in all libunbound based resolvers.Florian Obser
2019-12-02increment refcount before doing the call to resolve(); ok florian@Otto Moerbeek
2019-12-02Add an "all" mode for status and a much more compact and readable histogramOtto Moerbeek
2019-12-01Add missing space between "accept" and "bogus"kn
2019-12-01Allow forcing specific domains to be resolved by specific resolvers;Otto Moerbeek
2019-11-30make sure we only pass normalized timevals for the next resolver interval;Otto Moerbeek
2019-11-30Not being able to create a resolver is not a fatal condition in unwind,Florian Obser
2019-11-29Don't hand-roll TAILQ_CONCAT in a slow way.Florian Obser
2019-11-29While we learn how the "decaying" histogram works it's helpful to showFlorian Obser
2019-11-28The happy eyeballs code computes a median to assess how well aOtto Moerbeek
2019-11-28No more example file; ok florian@Otto Moerbeek
2019-11-28show median RTT in unwindctl statusFlorian Obser
2019-11-28Track all elements of a forwarder configuration individually:Florian Obser
2019-11-27Instead of only considering if a resolving strategy is dead, works orFlorian Obser
2019-11-27Move resolver preference storage to a struct which makes it easier toFlorian Obser
2019-11-27Nuke http captive portal detection; something better is coming.Florian Obser
2019-11-26Improve previouskn
2019-11-26Fail on nonexistent config file iff -c is givenkn
2019-11-25simplify histogram code and cut off at < limit, as the histogram display inOtto Moerbeek
2019-11-25 Signal what happened, not what needs doing.Florian Obser
2019-11-23plug memory leakFlorian Obser
2019-11-23An if_index of zero signals to remove all previous proposals from aFlorian Obser
2019-11-22Use a table for unbound options; ok florian@Otto Moerbeek
2019-11-22Treat DNS proposals as replacement of previous proposals from the sameFlorian Obser
2019-11-21 Track on which interface nameservers have been learned.Florian Obser
2019-11-20Check for a too short answer packet in all callback functions becauseFlorian Obser
2019-11-19Print learned forwarders in unwindctl.Florian Obser
2019-11-19Proposals from dhclient are sometimes replacements.Florian Obser
2019-11-19Stop dhcp lease file parsing now that dhclient sends proposals.Florian Obser
2019-11-15Improve readability by using a typedef for the callback type; ok florian@Otto Moerbeek
2019-11-14With the stub resolver we have since some time we can resolve theFlorian Obser
2019-11-14 Checking a resolver that we are already checking can lead to aFlorian Obser
2019-11-14 Since resolve() switched to a callback mechanism all uw_resolver objectsFlorian Obser
2019-11-12we might not have a best resolver at allFlorian Obser
2019-11-12do not leak buf or region if either fails to allocateFlorian Obser
2019-11-12we are not using was_ratelimitedFlorian Obser
2019-11-12Pass in a callback function to resolve() instead of a flag toFlorian Obser
2019-11-12 Error logging is done inside of resolve since it depends on theFlorian Obser
2019-11-12 It is possible to cancel libunbound queries by tracking an id.Florian Obser
2019-11-11Implement DNS proposals to learn nameservers from networkFlorian Obser
2019-11-10Consistently use _rcctl enable foo_ in examples, it's simpler and lessLandry Breuil
2019-11-09Mechanically change the forwarder SIMPLEQ to a TAILQ. Needed forFlorian Obser
2019-11-09failure to allocate memory is not an XXXFlorian Obser
2019-11-07Document the new home of the control socket; ok florian@Otto Moerbeek
2019-11-06useless debug outputFlorian Obser
2019-11-06Move control socket to /dev like slaacd(8) is doing to be able toFlorian Obser