]>
git.ipfire.org Git - thirdparty/knot-resolver.git/log
Grigorii Demidov [Thu, 30 Mar 2017 11:38:40 +0000 (13:38 +0200)]
lib/nsrep: don't treat servers with NOIP4 + NOIP6 flags as timeouted
Ondřej Surý [Sun, 26 Mar 2017 18:22:23 +0000 (20:22 +0200)]
Merge branch 'better-help-text' into 'master'
Auto-generate numeric limits and defaults in help text.
See merge request !248
Daniel Kahn Gillmor [Sun, 26 Mar 2017 15:28:31 +0000 (10:28 -0500)]
Auto-generate numeric limits and defaults in help text.
Ondřej Surý [Sun, 26 Mar 2017 18:19:44 +0000 (20:19 +0200)]
Merge branch 'better-padding-default' into 'master'
Improve default padding of responses.
See merge request !247
Daniel Kahn Gillmor [Sat, 25 Mar 2017 22:00:53 +0000 (17:00 -0500)]
Improve default padding of responses.
At NDSS 2017's DNS privacy workshop, I presented an empirical study of
DNS padding policies:
https://www.internetsociety.org/events/ndss-symposium/ndss-symposium-2017/dns-privacy-workshop-2017-programme#session3
The slide deck is here:
https://dns.cmrg.net/ndss2017-dprive-empirical-DNS-traffic-size.pdf
The resulting recommendation from the research is that a simple
padding policy is relatively cheap and still protective of metadata
when DNS traffic is encrypted:
* queries should be padded to a multiple of 128 octets
* responses should be padded to a multiple of 468 octets
This change adjusts the default policy to match these recommendations.
I recently proposed a similar change to libknot to define a standard
policy in a centralized place:
https://gitlab.labs.nic.cz/labs/knot/merge_requests/692
I'll submit a followup request to make use of that centralized policy
(once kresd is willing to depend on a newer version of libknot), but
please consider this proposed change first.
Vladimír Čunát [Fri, 24 Mar 2017 14:16:33 +0000 (15:16 +0100)]
Merge branch !246: lib/resolve: avoid unnecessary DS queries
Vladimír Čunát [Fri, 24 Mar 2017 13:38:22 +0000 (14:38 +0100)]
Merge !244: dnssec/nsec: validate wildcard no-data answers
Vladimír Čunát [Fri, 24 Mar 2017 12:23:48 +0000 (13:23 +0100)]
Merge !243: don't set AD flag for opt-out wildcard answers
Grigorii Demidov [Fri, 24 Mar 2017 11:51:36 +0000 (12:51 +0100)]
lib/resolve: avoid unnecessary DS queries
Grigorii Demidov [Thu, 23 Mar 2017 13:42:12 +0000 (14:42 +0100)]
dnssec/nsec: missed wildcard no-data answers validation has been implemented
Grigorii Demidov [Wed, 22 Mar 2017 13:37:36 +0000 (14:37 +0100)]
test/deckard: update to actual version
Vladimír Čunát [Wed, 22 Mar 2017 12:20:47 +0000 (13:20 +0100)]
Merge !242: kr_bitcmp: shut up a warning
Grigorii Demidov [Wed, 22 Mar 2017 12:18:06 +0000 (13:18 +0100)]
layer/validate: check if NSEC3 records in wildcard expansion proof has an optout
Vladimír Čunát [Wed, 22 Mar 2017 12:15:30 +0000 (13:15 +0100)]
kr_bitcmp: shut up a warning
Vladimír Čunát [Wed, 22 Mar 2017 10:13:28 +0000 (11:13 +0100)]
Merge !239: check if iterator has not selected any records
Grigorii Demidov [Wed, 22 Mar 2017 08:50:19 +0000 (09:50 +0100)]
layer/iterate: nitpick - better variable name
Vladimír Čunát [Tue, 21 Mar 2017 14:57:41 +0000 (15:57 +0100)]
Merge !234: kr_bitcmp: adjust semantics -> memcmp
Vladimír Čunát [Fri, 17 Mar 2017 12:03:55 +0000 (13:03 +0100)]
view docs: add example matching all addresses
This started working since the grandparent commit. Suggested at
https://gitter.im/CZ-NIC/knot-resolver?at=
58ca5a03f7f7d48104212607
Vladimír Čunát [Fri, 17 Mar 2017 11:51:42 +0000 (12:51 +0100)]
kr_bitcmp: add meaning to NULL inputs
Reasoning: we currently only use the function from lua modules and nil
values are very common there; I want to pick these changes to a bugfix
update without extensive checking whether the modules might pass
invalid input if user passes invalid config and thus introduce new
crashes. The checks also seem cheap performance-wise.
Vladimír Čunát [Wed, 15 Mar 2017 13:55:39 +0000 (14:55 +0100)]
kr_bitcmp: adjust semantics -> memcmp
Motivation: allow bits=0 and consequently 0.0.0.0/0 matches in view
and renumber modules.
https://gitter.im/CZ-NIC/knot-resolver?at=
58c940c721d548df2cdfda5e
We shouldn't mix up error codes with valid results from memcmp;
let's just segfault if someone passes a NULL, just as memcmp() itself.
Grigorii Demidov [Tue, 21 Mar 2017 13:21:37 +0000 (14:21 +0100)]
layer/validate: better debug message
Vladimír Čunát [Tue, 21 Mar 2017 12:06:45 +0000 (13:06 +0100)]
Merge !233: trust anchors: store in prettier format
Closes #167.
Vladimír Čunát [Tue, 21 Mar 2017 12:05:20 +0000 (13:05 +0100)]
Merge !236: worker_resolve: truly honor the options parameter
Grigorii Demidov [Tue, 21 Mar 2017 11:48:07 +0000 (12:48 +0100)]
layer/iterate: imprevements in unhelpfull referral processing
Grigorii Demidov [Tue, 21 Mar 2017 10:22:13 +0000 (11:22 +0100)]
layer/validate: check if iterator has not selected any records for validating from non-empty authoritative answer
Vladimír Čunát [Mon, 20 Mar 2017 11:50:29 +0000 (12:50 +0100)]
Merge !237: layer/validate: clear AD with optout NSEC3
Fixes #169.
Grigorii Demidov [Fri, 17 Mar 2017 08:20:17 +0000 (09:20 +0100)]
layer/validate: clear AD if closest encloser proof has optouted NSEC3
Vladimír Čunát [Mon, 20 Mar 2017 08:34:22 +0000 (09:34 +0100)]
Merge !238: fix logging of glue addresses
Petr Špaček [Sun, 19 Mar 2017 10:39:08 +0000 (11:39 +0100)]
layer/iterate: fix logging of glue addresses
Previously even loopback glue addresses were logged using message
'<= using glue ...' which was very confusing. From now the loopback
addresses are logged using '<= ignoring invalid glue ...'.
Logging was moved into the function update_nsaddr() to avoid changing
return codes. As far as I can tell it does not produce any confusing
messages.
Petr Špaček [Fri, 17 Mar 2017 10:04:37 +0000 (11:04 +0100)]
CI: always use commited Deckard version
Petr Špaček [Fri, 17 Mar 2017 10:03:31 +0000 (11:03 +0100)]
check-integration: Warn if Deckard does not match commited version
Vladimír Čunát [Wed, 15 Mar 2017 15:37:36 +0000 (16:37 +0100)]
worker_resolve: truly honor the options parameter
It was being overwritten by options from struct kr_context;
now the flags are combined (by set union).
For example, the NO_CACHE flag is important for the prefetch module and
for trust anchor updates.
Vladimír Čunát [Tue, 14 Mar 2017 17:21:50 +0000 (18:21 +0100)]
trust anchors: add KeyTag into comments
Fixes #167.
Vladimír Čunát [Mon, 13 Mar 2017 19:38:58 +0000 (20:38 +0100)]
trust anchors: store in prettier format (#167)
The output is only better if built with libknot >= 2.4.0.
As a side-effect, add lua method knot_rrset_t::txt_dump;
it's a light wrapper around knot_rrset_txt_dump.
Re-tested rolling with a.moot-servers.net, to be sure.
Vladimír Čunát [Mon, 13 Mar 2017 19:26:46 +0000 (20:26 +0100)]
daemon/lua/kres.lua -> *.in
Vladimír Čunát [Mon, 13 Mar 2017 18:08:29 +0000 (19:08 +0100)]
version: don't depend on rr2str output style
It would break after making rr2str pretty-print the data.
Vladimír Čunát [Wed, 15 Mar 2017 14:30:48 +0000 (15:30 +0100)]
Merge !235: dnstap.mk: fix race around dnstap.pb-c.h
Vladimír Čunát [Tue, 14 Mar 2017 18:14:13 +0000 (19:14 +0100)]
dnstap.mk: fix race around dnstap.pb-c.h
Probably. The problem sometimes appeared on osx Travis.
Vladimír Čunát [Mon, 13 Mar 2017 22:11:44 +0000 (23:11 +0100)]
Merge !232: nitpick in kresd(8) man page
Daniel Kahn Gillmor [Mon, 13 Mar 2017 22:00:53 +0000 (18:00 -0400)]
kresd(8) should refer to the resolver as "kresd", not as "unbound"
Ondřej Surý [Wed, 8 Mar 2017 10:58:14 +0000 (11:58 +0100)]
Merge branch 'fix-auth-qname' into 'master'
layer\iterate: when processing delegations, check if qname is at\below new authority
See merge request !229
Grigorii Demidov [Wed, 8 Mar 2017 10:14:00 +0000 (11:14 +0100)]
layer\iterate: when proccessing delegations, check if qname is at\below new authority
Ondřej Surý [Mon, 6 Mar 2017 12:32:11 +0000 (13:32 +0100)]
Merge branch 'update-gitignore-zonefile.lua' into 'master'
Add daemon/lua/zonefile.lua to git ignore list
See merge request !228
Ondřej Surý [Mon, 6 Mar 2017 12:29:17 +0000 (13:29 +0100)]
Add daemon/lua/zonefile.lua to git ignore list
Ondřej Surý [Mon, 6 Mar 2017 12:08:03 +0000 (13:08 +0100)]
Merge branch 'dnstap' into 'master'
add dnstap module
See merge request !213
Ondřej Surý [Mon, 6 Mar 2017 12:06:13 +0000 (13:06 +0100)]
Test failing make clean on missing dnstap dependencies
Ondřej Surý [Mon, 6 Mar 2017 12:02:18 +0000 (13:02 +0100)]
modules/dnstap: Change option names to socket_path and log_responses, we don't use camelCase anywhere else
Ondřej Surý [Mon, 6 Mar 2017 11:54:40 +0000 (12:54 +0100)]
dnstap module also needs protobuf-c compiler (protoc-c)
Vladimír Čunát [Wed, 22 Feb 2017 13:47:26 +0000 (14:47 +0100)]
modules/dnstap: move description into the docs
Vladimír Čunát [Wed, 22 Feb 2017 12:50:10 +0000 (13:50 +0100)]
modules/dnstap: nitpicks
Vladimír Čunát [Wed, 22 Feb 2017 12:10:21 +0000 (13:10 +0100)]
dnstap.proto: move from contrib into module
... and generate files instead of including them.
Vicky Shrestha [Tue, 17 Jan 2017 08:08:17 +0000 (08:08 +0000)]
minor changes from https://github.com/CZ-NIC/knot-resolver/pull/39
Vicky Shrestha [Fri, 30 Dec 2016 20:14:55 +0000 (12:14 -0800)]
adding dnstap to documentation
Vicky Shrestha [Fri, 30 Dec 2016 20:14:02 +0000 (12:14 -0800)]
dnstap tests requires go 1.5+
removing it from make test since default go version for language
C in Travis is 1.4 which has no vendoring support
Vicky Shrestha [Mon, 26 Dec 2016 09:16:05 +0000 (01:16 -0800)]
adding dnstap dependencies to bootstrap
Vicky Shrestha [Mon, 26 Dec 2016 08:05:49 +0000 (00:05 -0800)]
fixing makefile to compile protobuf if dnstap is enabled
Vicky Shrestha [Sat, 24 Dec 2016 18:04:53 +0000 (10:04 -0800)]
Fixing structs after rebasing upstream changes
Vicky Shrestha [Thu, 8 Dec 2016 07:26:04 +0000 (23:26 -0800)]
dnstap testing application
Vicky Shrestha [Thu, 22 Sep 2016 22:32:27 +0000 (15:32 -0700)]
Adding dnstap module
Vladimír Čunát [Fri, 3 Mar 2017 12:28:28 +0000 (13:28 +0100)]
Merge !226: lib/resolve: deferred answer processing was fixed
Vladimír Čunát [Fri, 3 Mar 2017 12:04:02 +0000 (13:04 +0100)]
Merge !225: rrcache: don't store NSEC3 and their signatures
Vladimír Čunát [Fri, 3 Mar 2017 12:02:39 +0000 (13:02 +0100)]
Merge !224: layer/validate: fix missing AD flag in some cases
Fixes #164.
Grigorii Demidov [Fri, 3 Mar 2017 09:17:06 +0000 (10:17 +0100)]
lib/resolve: deferred answer processing was fixed
Vladimír Čunát [Thu, 2 Mar 2017 17:28:14 +0000 (18:28 +0100)]
rrcache: don't store NSEC3 and their signatures
They would end up cached by their hashed owner names and then even
returned if explicitly queried by that hashed name, which is not correct:
https://tools.ietf.org/html/rfc4035#section-2.3
Internally we only need these for non-existence proofs, and those are
stored in pktcache instead.
Grigorii Demidov [Wed, 1 Mar 2017 12:47:27 +0000 (13:47 +0100)]
layer/validate: don't treat anwsers which contain DS non-existance proof as unsecured
Vladimír Čunát [Wed, 1 Mar 2017 09:59:56 +0000 (10:59 +0100)]
Merge !221: daemon: fixed memory leak and array bounds check fail
Submitted as https://github.com/CZ-NIC/knot-resolver/pull/42
Petr Špaček [Wed, 1 Mar 2017 08:13:22 +0000 (09:13 +0100)]
Merge branch 'full_check_integration' into 'master'
Update check-integration to run full test suite from Deckard
See merge request !220
Marek Vavruša [Tue, 28 Feb 2017 19:05:20 +0000 (11:05 -0800)]
daemon: fixed memory leak and array bounds check fail
Petr Špaček [Tue, 28 Feb 2017 16:27:36 +0000 (17:27 +0100)]
tests: print warning if check-integration is executed with PREFIX outside of source directory
Petr Špaček [Tue, 28 Feb 2017 10:46:04 +0000 (11:46 +0100)]
tests: use tests/deckard/kresd_run.sh for check-integration target
Now the check-integration is (again) doing the same set of tests
as kresd_run.sh in Deckard tree.
Vladimír Čunát [Tue, 28 Feb 2017 13:33:29 +0000 (14:33 +0100)]
Merge 219: lib/resolve: forward +cd in stub mode; minor bugfix
Grigorii Demidov [Tue, 28 Feb 2017 12:30:25 +0000 (13:30 +0100)]
lib/resolve: forward +cd in stub mode; minor bugfix in debug output
Vladimír Čunát [Tue, 28 Feb 2017 11:42:23 +0000 (12:42 +0100)]
Merge !202: trust anchor improvements
Vladimír Čunát [Tue, 28 Feb 2017 10:49:26 +0000 (11:49 +0100)]
Merge !210: support setting address for outgoing connections
Closes #158.
Vladimír Čunát [Mon, 20 Feb 2017 10:26:27 +0000 (11:26 +0100)]
lua: add net.outgoing_{v4,v6} and documentation
Fixes https://gitlab.labs.nic.cz/knot/resolver/issues/158
The naming is inspired by Unbound's "outgoing-interface".
Vladimír Čunát [Mon, 20 Feb 2017 09:12:38 +0000 (10:12 +0100)]
daemon: support restricting outgoing IP address
Grigorii Demidov [Tue, 28 Feb 2017 09:59:09 +0000 (10:59 +0100)]
Merge branch 'fix-dups' into 'master'
layer/validate: fix duplicate records in AUTHORITY section in case of WC expansion proof
See merge request !216
Grigorii Demidov [Tue, 28 Feb 2017 09:47:05 +0000 (10:47 +0100)]
tests: sync deckard; cleanup
Vladimír Čunát [Tue, 28 Feb 2017 09:24:20 +0000 (10:24 +0100)]
Merge !209: iterate: remove unused function
Grigorii Demidov [Mon, 27 Feb 2017 12:19:48 +0000 (13:19 +0100)]
layer/validate: fix duplicate records in AUTHORITY section in case of WC expansion proof
Vladimír Čunát [Tue, 14 Feb 2017 15:44:27 +0000 (16:44 +0100)]
trust anchors: persist the state and timer
The format of TA store is compatible both ways with old kresd.
Note: it requires the parent commit to work, i.e. new libzscanner.
Vladimír Čunát [Tue, 14 Feb 2017 15:13:23 +0000 (16:13 +0100)]
zonefile.lua: support comments and string input
Comment parsing will only be supported in future libzscanner version.
Also move out of the module, as trust_anchors.lua requires it and isn't
in a module.
Vladimír Čunát [Fri, 24 Feb 2017 14:07:36 +0000 (15:07 +0100)]
Merge !215: lua: do *not* truncate cache size to unsigned
Vladimír Čunát [Fri, 24 Feb 2017 10:26:28 +0000 (11:26 +0100)]
lua: do *not* truncate cache size to unsigned
... and perform extra checks when converting from the floating-point
number.
Vladimír Čunát [Tue, 21 Feb 2017 15:00:43 +0000 (16:00 +0100)]
Merge !211: iterate: skip answer RRs that don't match the query
Fixes #160.
Vladimír Čunát [Tue, 21 Feb 2017 14:53:41 +0000 (15:53 +0100)]
Merge !184: add workarounds module
Vladimír Čunát [Tue, 21 Feb 2017 11:13:41 +0000 (12:13 +0100)]
workarounds: add magazine-fashion.net
Some of their nameservers return 192.168.1.1, but it seems an
improvement, as meaningful address is returned sometimes now.
Vladimír Čunát [Tue, 21 Feb 2017 10:38:19 +0000 (11:38 +0100)]
workarounds: add another NS set with bogus 0x20 PTR
Vladimír Čunát [Mon, 20 Feb 2017 12:50:39 +0000 (13:50 +0100)]
iterate: skip answer RRs that don't match the query
Fixes https://gitlab.labs.nic.cz/knot/resolver/issues/160
Vladimír Čunát [Mon, 20 Feb 2017 07:06:57 +0000 (08:06 +0100)]
utils: add union inaddr
It will be a useful idiom for IP address storage and correct conversion
of sockaddr* pointers.
Petr Špaček [Fri, 17 Feb 2017 19:20:12 +0000 (20:20 +0100)]
Merge branch 'harden-permissive-mode' into 'master'
Disable storing GLUE records into the cache in the QUERY_PERMISSIVE mode
See merge request !208
Ondřej Surý [Fri, 17 Feb 2017 14:18:25 +0000 (15:18 +0100)]
Disable storing GLUE records into the cache in the QUERY_PERMISSIVE mode
Vladimír Čunát [Fri, 17 Feb 2017 15:41:18 +0000 (16:41 +0100)]
iterate: remove function unused after parent merge
Grigorii Demidov [Fri, 17 Feb 2017 11:57:21 +0000 (12:57 +0100)]
Merge branch 'fix-cname-unroll' into 'master'
layer/iterate: some improvements in cname chain unrolling
See merge request !207
Grigorii Demidov [Fri, 17 Feb 2017 10:36:11 +0000 (11:36 +0100)]
layer/iterate: some improvements in cname chain unrolling
Grigorii Demidov [Thu, 16 Feb 2017 14:22:10 +0000 (15:22 +0100)]
Merge branch 'fix-unsecured-secured' into 'master'
lib/resolve: zonecut fetching error was fixed
See merge request !205
Grigorii Demidov [Thu, 16 Feb 2017 12:23:41 +0000 (13:23 +0100)]
lib/resolve: zonecut fetching error was fixed
Vladimír Čunát [Thu, 16 Feb 2017 10:15:39 +0000 (11:15 +0100)]
policy: document the FLAGS action
Vladimír Čunát [Wed, 15 Feb 2017 17:30:34 +0000 (18:30 +0100)]
workarounds docs: add manual section
Vladimír Čunát [Wed, 15 Feb 2017 17:27:19 +0000 (18:27 +0100)]
avoid a doxygen warning