upstream/unbound
1
0
Fork 0
mirror of https://github.com/NLnetLabs/unbound.git synced 2026-02-17 09:38:03 -05:00
Code Issues Projects Releases Packages Wiki Activity Actions 1
7690 commits 27 branches 209 tags 137 MiB
Commit graph

7690 commits

This branch
This branch
All branches
Author SHA1 Message Date
Willem Toorop
54b84381f4 tdir test for padding option 2021-01-24 18:12:08 +01:00
Roland van Rijswijk-Deij
d253db04fd Addressed review comment from @wcawijngaards 2021-01-22 18:56:09 +00:00
Roland van Rijswijk-Deij
c4c849d878 Rebase on master 2021-01-22 16:44:56 +00:00
Willem Toorop
fc49d145b5 Changelog entry for padding option 2021-01-22 15:14:10 +01:00
Willem Toorop
b7864b0c48 Merge branch 'master' into features/padding 2021-01-22 15:13:01 +01:00
Willem Toorop
2a3548e1ef Move NSID Changelog entry to day of merge 2021-01-22 15:10:57 +01:00
W.C.A. Wijngaards
dedc1e2340 zonemd, unit test for capitalisation of owner and ns names. 2021-01-22 14:40:21 +01:00
W.C.A. Wijngaards
20ff80c5bb zonemd, fix to detect duplicate zonemd scheme and algo RRs. 2021-01-22 14:36:53 +01:00
W.C.A. Wijngaards
3a6f1ecafa Fixup if brackets. 2021-01-22 13:11:41 +01:00
George Thessalonikefs
63a02f3a87 Merge branch 'negative_soa_ttl' 2021-01-22 12:39:16 +01:00
George Thessalonikefs
dc23502abb - Updated Changelog for PR #407. 2021-01-22 12:33:56 +01:00
Willem Toorop
48ecf95108 Merge branch 'master' into features/padding 2021-01-22 10:29:50 +01:00
yunwei
0215500261
Update netevent.c 
#386
I found the root cause of this issue. r_buffer is r->query_reply.c->buffer, used to fill the reply in the mesh_send_reply function, then call comm_point_send_reply, and then call http2_submit_dns_response to send the DOH response. However, the buffer is not cleared after use. If the query length is greater than the last response length, the next dns query in the same H2 session will encounter an error.
This is bug!!!

Clear the buffer after use.
 2021-01-20 14:12:51 +08:00
George Thessalonikefs
38e12229b8 - Feedback for PR #407. 2021-01-19 20:09:34 +01:00
Willem Toorop
9de4c398e1
Merge pull request #363 from NLnetLabs/features/nsid 
Support for RFC5001: DNS Name Server Identifier (NSID) Option
 2021-01-19 17:19:48 +01:00
Willem Toorop
68d51317fe rpl tests for nsid 2021-01-19 17:18:26 +01:00
Florian Obser
68d92b7bbb Prevent a few more yacc clashes. 2021-01-19 17:13:00 +01:00
Willem Toorop
7ba51fce24 example.conf.in entry for nsid 2021-01-19 16:18:36 +01:00
George Thessalonikefs
9056613a79 - Fix TTL of SOA record for negative answers (localzone data and 
authzone) to be the minimum of the SOA TTL and the SOA.MINIMUM.
 2021-01-19 15:55:55 +01:00
Willem Toorop
a152c7f907 Merge branch 'master' into features/nsid 2021-01-19 14:21:18 +01:00
W.C.A. Wijngaards
f273716b80 - Fix declaration before statement and signed comparison warning in 
dns64.
 2021-01-18 10:23:01 +01:00
W.C.A. Wijngaards
c125fe67bc - Fix #404: DNS query with small edns bufsize fail. 2021-01-18 08:29:52 +01:00
W.C.A. Wijngaards
e55f38fa83 Changelog entry for #402. 
- Merge #402 from fobser: Implement IPv4-Embedded addresses according
  to RFC6052.
 2021-01-15 08:15:54 +01:00
Wouter Wijngaards
9a055c2df4
Merge pull request #402 from fobser/dns64 
Implement IPv4-Embedded addresses according to RFC6052.
 2021-01-15 08:14:58 +01:00
Florian Obser
285a7fdd21 Implement IPv4-Embedded addresses according to RFC6052. 
The original algorithm assumed that any prefix length would be valid
and did not skip over bits 64 to 71 and set them to zero.
This means that only dns64 prefixes with length 32 and 96 generated
embedded addresses according to RFC6052, cf. Figure 1 in 2.2.
 2021-01-14 19:15:30 +01:00
W.C.A. Wijngaards
4613d2bf04 - Fix for #93: dynlibmodule import library is named libunbound.dll.a. 2021-01-14 16:50:17 +01:00
W.C.A. Wijngaards
5c760a13a4 - rpz trigger, unlock and remove node when rpz is disabled with action override. 2021-01-14 15:31:33 +01:00
W.C.A. Wijngaards
5e00d516fe - fix that testdata/rpz_qname.rpl trigger tests succeed. 2021-01-14 15:31:17 +01:00
W.C.A. Wijngaards
15b46257e1 - fix that testdata/rpz_clientip.rpl trigger tests succeed. 2021-01-14 13:49:42 +01:00
W.C.A. Wijngaards
db7413acb6 Merge branch 'magenbluten-rpz' into rpz-triggers 2021-01-14 12:13:24 +01:00
W.C.A. Wijngaards
cdb60adcdc Merge branch 'rpz' of https://github.com/magenbluten/unbound into magenbluten-rpz 
Conflict fixed for rpz.disabled check added.
 2021-01-14 12:11:29 +01:00
W.C.A. Wijngaards
3b82e690ef - Fix for #93: dynlibmodule link fix for Windows. 2021-01-13 14:56:25 +01:00
W.C.A. Wijngaards
5314f63343 Nicer changelog note for #399 
- Merge #399 from xiangbao227: The lock of lruhash table should
  unlocked after markdel entry.
 2021-01-13 10:10:12 +01:00
W.C.A. Wijngaards
24fd871245 Changelog note for #399 
- Merge #399 from xiangbao227: The function rrset_cache_touch can
  touch an entry to the lru while markdelling the entry in
  lruhash_remove.
 2021-01-13 10:07:15 +01:00
Wouter Wijngaards
7e46204bf7
Merge pull request #399 from xiangbao227/master 
I found that in function "lruhash_remove", table was locked at first ,then lru_remove the entry , then unlock the table, and then markdel entry , but in function "rrset_cache_touch" , the entry will be touched to lru again before markdelling entry in function "lruhash_remove". This may lead to core!
 2021-01-13 10:05:50 +01:00
xiangbao227
93e5705259 I found that in function lruhash_remove, table was locked at first ,then lru_remove the entry , then unlock the table, and then markdel entry , but in function rrset_cache_touch , the entry will be touched to lru again before markdelling entry in function lruhash_remove. This is a bug! 2021-01-13 10:33:41 +08:00
W.C.A. Wijngaards
4d1d8b4cdd And man page documentation for them. 2021-01-12 13:40:45 +01:00
W.C.A. Wijngaards
d1b92a6ce2 - Fix so local zone types always_nodata and always_deny can be used 
from the config file.
 2021-01-12 13:39:07 +01:00
W.C.A. Wijngaards
3322f631e5 - Fix #397: [Feature request] add new type always_null to local-zone 
similar to always_nxdomain.
 2021-01-12 13:35:05 +01:00
W.C.A. Wijngaards
64f508fa00 - Fix clang analysis warning. 2021-01-08 11:10:05 +01:00
W.C.A. Wijngaards
d9dd7bc36f - Add comment documentation. 2021-01-08 11:01:06 +01:00
W.C.A. Wijngaards
1aa7168c6a - For #391: more double casts in python start time calculation. 2021-01-08 09:55:55 +01:00
W.C.A. Wijngaards
ee2545d939 - For #391: fix indentation. 2021-01-08 09:53:52 +01:00
W.C.A. Wijngaards
3e03e2c26d - For #391: use struct timeval* start_time for callback information. 2021-01-08 09:47:46 +01:00
W.C.A. Wijngaards
260837e050 Changelog note for #391 
- Merge PR #391 from fhriley: Add start_time to reply callbacks so
  modules can compute the response time.
 2021-01-08 09:36:37 +01:00
Wouter Wijngaards
48724de155
Merge pull request #391 from fhriley/reply_cb_start_time 
Add start_time to reply callbacks so modules can compute the response…
 2021-01-08 09:35:07 +01:00
W.C.A. Wijngaards
2e4d64684e - Fix #387: client-subnet-always-forward seems to effectively bypass 
any caching?
 2021-01-06 13:42:00 +01:00
W.C.A. Wijngaards
752aea8407 Changelog note for #395 
- Merge PR #395 from mptre: add missing null check.
 2021-01-06 13:19:46 +01:00
Wouter Wijngaards
d188096a01
Merge pull request #395 from mptre/log-addr-null 
add missing null check
 2021-01-06 13:19:30 +01:00
Anton Lindqvist
422213c171 add missing null check 
I have a unbound forward zone configured on my router for my $DAYJOB.
The address associated with the zone is only accessible when the router
is connected to a VPN. If the VPN connection is absent, trying to
resolve any domain that must be handled by the zone crashes unbound.
Turns out there's a missing NULL check in `comm_point_send_udp_msg()`.
The same routine already has `if (addr) {} else {}` branches so I guess
protecting the call to `log_addr()` using the same conditional is
reasonable

I have also committed the same fix to unbound shipped with OpenBSD[1].

[1] https://marc.info/?l=openbsd-cvs&m=160993335615698&w=2
 2021-01-06 12:44:26 +01:00