BIND 9.6-ESV-R5 is the current production release
of BIND 9.6.
This document summarizes changes from BIND 9.6-ESV-R4 to BIND 9.6-ESV-R5.
Please see the CHANGES file in the source code release for a
complete list of all changes.
The latest release of BIND 9 software can always be found
on our web site at
http://www.isc.org/downloads/all.
There you will find additional information about each release,
source code, and some pre-compiled versions for certain operating
systems.
-
Added a tool able to generate malformed packets to allow testing
of how named handles them.
[RT #24096]
-
named, set up to be a caching resolver, is vulnerable to a
user querying a domain with very large resource record sets (RRSets)
when trying to negatively cache the response. Due to an off-by-one
error, caching the response could cause named to crash. [RT #24650]
[CVE-2011-1910]
-
Change #2912 populated the message section in replies to UPDATE requests,
which some Windows clients wanted. This exposed a latent bug that allowed
the response message to crash named. With this fix, change 2912 has been
reduced to copy only the zone section to the reply. A more complete fix
for the latent bug will be released later.
[RT #24777]
-
Merged in the NetBSD ATF test framework (currently
version 0.12) for development of future unit tests.
Use configure --with-atf to build ATF internally
or configure --with-atf=prefix to use an external
copy. [RT #23209]
-
Added more verbose error reporting from DLZ LDAP. [RT #23402]
-
Replaced compile time constant with STDTIME_ON_32BITS.
[RT #23587]
-
During RFC5011 processing some journal write errors were not detected.
This could lead to managed-keys changes being committed but not
recorded in the journal files, causing potential inconsistencies
during later processing. [RT #20256]
A potential NULL pointer deference in the DNS64 code could cause
named to terminate unexpectedly. [RT #20256]
A state variable relating to DNSSEC could fail to be set during
some infrequently-executed code paths, allowing it to be used whilst
in an unitialized state during cache updates, with unpredictable results.
[RT #20256]
A potential NULL pointer deference in DNSSEC signing code could
cause named to terminate unexpectedly [RT #20256]
Several cosmetic code changes were made to silence warnings
generated by a static code analysis tool. [RT #20256]
-
When using _builtin in named.conf, named.conf changes were not found
when reloading the config file. Now checks _builtin zone arguments
to see if the zone is re-usable or not. [RT #21914]
-
After an external code review, a code cleanup was done. [RT #22521]
-
When signing records, named didn't filter out any TTL changes
to DNSKEY records. This resulted in an incomplete key set. TTL
changes are now dealt with before signing. [RT #22590]
-
The IN6_IS_ADDR_LINKLOCAL and IN6_IS_ADDR_SITELOCAL macros in win32 were
updated/corrected per current Windows OS. [RT #22724]
-
Cause named to terminate at startup or rndc reconfig
reload to fail, if a log file specified in the
conf file isn't a plain file. (RT #22771]
-
named now forces the ADB cache time for glue related data to zero
instead of relying on TTL. This corrects problematic behavior in cases
where a server was authoritative for the A record of a nameserver for a
delegated zone and was queried to recursively resolve records within
that zone. [RT #22842]
-
Fix the zonechecks system test to fail on error (warning in 9.6,
fatal in 9.7) to match behaviour for 9.4. [RT #22905]
-
The "rndc" command usage statement was missing the "-b" option.
[RT #22937]
-
Fixed a possible deadlock due to zone re-signing. [RT #22964]
-
Fixed precedence order bug with NS and DNAME records if both are present.
(Also fixed timing of autosign test in 9.7+) [RT #23035]
-
The secure zone update feature in named is based on the zone being
signed and configured for dynamic updates. A bug in the ACL processing
for "allow-update { none; };" resulted in a zone that is supposed to
be static being treated as a dynamic zone. Thus, named would try to
sign/re-sign that zone erroneously. [RT #23120]
-
A new test has been added to check the apex NSEC3 records after DNSKEY
records have been added via dynamic update. [RT #23229]
-
If a slave initiates a TSIG signed AXFR from the master and the master
fails to correctly TSIG sign the final message, the slave would be left
with the zone in an unclean state. named detected this error too late
and named would crash with an INSIST. The order dependancy has been
fixed. [RT #23254]
-
If the server has an IPv6 address but does not have IPv6 connectivity
to the internet, dig +trace could fail attempting to use IPv6
addresses. [RT #23297]
-
Changing TTL did not cause dnssec-signzone to generate new signatures.
[RT #23330]
-
Have the validating resolver use RRSIG original TTL to compute
validated RRset and RRSIG TTL. [RT #23332]
-
In "make test" bin/tests/resolver, hold the socket manager lock
while freeing the socket.
[RT #23333]
-
If named encountered a CNAME instead of a DS record when walking
the chain of trust down from the trust anchor, it incorrectly stopped
validating. [RT #23338]
-
RRSIG records could have time stamps too far in the future.
[RT #23356]
-
named stores cached data in an in-memory database and keeps track of
how recently the data is used with a heap. The heap is stored within the
cache's memory space. Under a sustained high query load and with a small
cache size, this could lead to the heap exhausting the cache space. This
would result in cache misses and SERVFAILs, with named never releasing
the cache memory the heap used up and never recovering.
This fix removes the heap into its own memory space, preventing the heap
from exhausting the cache space and allowing named to recover gracefully
when the high query load abates. [RT #23371]
-
If running on a powerpc CPU and with atomic operations enabled,
named could lock up. Added sync instructions to the end of atomic
operations. [RT #23469]
-
If OpenSSL was built without engine support, named would have
compile errors and fail to build.
[RT #23473]
-
Handle isc_event_allocate failures in t_tasks test.
[RT #23572]
-
ixfr-from-differences {master|slave};
failed to select the master/slave zones, resulting in on diff/journal
file being created.
[RT #23580]
-
If a DNAME substitution failed, named returned NOERROR. The correct
response should be YXDOMAIN.
[RT #23591]
-
Remove bin/tests/system/logfileconfig/ns1/named.conf and
add setup.sh in order to resolve changing named.conf issue. [RT #23687]
-
NOTIFY messages were not being sent when generating
a NSEC3 chain incrementally. [RT #23702]
-
Signatures for records at the zone apex could go
stale due to an incorrect timer setting. [RT #23769]
-
The autosign tests attempted to open ports within reserved ranges. Test
now avoids those ports.
[RT #23957]
-
named, acting as authoritative server for DLZ zones, was not correctly
setting the authoritative (AA) bit.
[RT #24146]
-
Clean up some cross-compiling issues and added two undocumented
configure options, --with-gost and --with-rlimtype, to allow over-riding
default settings (gost=no and rlimtype="long int") when cross-compiling.
[RT #24367]
-
When trying sign with NSEC3, if dnssec-signzone couldn't find the
KSK, it would give an incorrect error "NSEC3 iterations too big for
weakest DNSKEY strength" rather than the correct "failed to find
keys at the zone apex: not found" [RT #24369]
-
nsupdate could dump core on shutdown when using SIG(0) keys. [RT #24604]
-
Named could fail to validate zones list in a DLV that validated insecure
without using DLV and had DS records in the parent zone. [RT #24631]
-
A bug in FreeBSD kernels causes IPv6 UDP responses greater than
1280 bytes to not fragment as they should. Until there is a kernel
fix, named will work around this by setting IPV6_USE_MIN_MTU on a
per packet basis. [RT #24950]
-
To avoid excessive startup time for configurations with large numbers
of zones, an environment variable, BIND9_ZONE_TASKS_HINTS, may now
be set prior to starting named. Divide your number of zones by 200
to find the recommended setting for this environment variable (i.e.,
if you have 200000 zones, set BIND9_ZONE_TASKS_HINTS to 1000 before
starting named). [RT #25084]
Known issues in this release
-
"make test" will fail on OSX and possibly other operating systems.
The failure occurs in a new test to check for allow-query ACLs.
The failure is caused because the source address is not specified on
the dig commands issued in the test.
If running "make test" is part of your usual acceptance process,
please edit the file bin/tests/system/allow_query/test.sh
and add
-b 10.53.0.2
to the DIGOPTS line.
Thank you to everyone who assisted us in making this release possible.
If you would like to contribute to ISC to assist us in continuing to make
quality open source software, please visit our donations page at
http://www.isc.org/supportisc.