On machines with high CPU cores, few lib/ns unit tests fail due to not
enough file descriptors. Increase limit, it would be set higher on 40+
core machines anyway.
Unit tests fail always on builders with 56 cores. There is issue with
limit of threads count in netmgr. Internal counter in hp.c does not
reset on each unit tests teardown. With many cores, it can lead to
assertion failures during the test.
This reverts commit cc152b028f.
EPEL and RHEL 8 do not yet support %autopatch -M X. I want check the
compatibility with them, keep it legacy way until they are supported in
RHEL 8.
Use autopatch, do not require mentioning each patch twice.
Patches below 300 are generic patches applied after unpacking.
Patches betwee 300 and 310 are PKCS11 specific, applied only when pkcs11
is enabled.
Substracted 100 from current patches.
New BIND no longer uses any part of docbook. It can handle out of tree
builds, therefore no hacks with copy back are required.
Documents should be installed just fine.
Docbook5 stylesheets with namespaces are required. BIND uses Docbook5
format. While it tries to keep compatibility with older stylesheets,
it fails silently and format of manual pages is broken.
Details in upstream issue:
https://gitlab.isc.org/isc-projects/bind9/-/issues/2310
Docbook5 generates manual pages with [FIXME: manual] instead of BIND9.
Fix metadata to be recognized and provide this value.
Latest release has not correctly formatted manual pages. Correct it by
rebuilding every manual page during the build, not only those modified
by a patch.
Fixed oot build of documentation. Because docbook does not work well
with out of tree builds, copy all sources required for documentation
into build directory. Should regenerate all manual pages, also html and
PDF formatted ARM.
Those packages were very similar in BIND 9.11. Because nothing requires
just bind-lite-devel package, make just one devel package with all
requirements. Keep separated libraries, but only one devel package.
Include also obsolete for automatic uninstall of previous bind-lite-devel
package. bind-devel now contains everything required to link against
libraries.
ARM and s390x cannot compile, because they lack atomic implementation in
lib/isc. Include upstream fix after 9.11.23 release.
Signed-off-by: Petr Menšík <pemensik@redhat.com>
I find no reason to turn off devel package creation. It can be ignored
if required, but is mandatory due to Fedora packaging guidelines.
Simplify it a bit.
Remove dlz-mysqldyn subpackage, move documentation and modules to the
same package. It is similar and has exactly the same dependencies,
different package is not required.
All DLZ modules were installed by mistake in main bind package.
Remove them from there, they should be offered only by each dlz
subpackage.
Move modules to upstream used directory %{_libdir}/named.
I find no reason to turn off devel package creation. It can be ignored
if required, but is mandatory due to Fedora packaging guidelines.
Simplify it a bit.
Those packages were very similar in BIND 9.11. Since there is no
isc-config.sh, no significant or required reason to have them separated
exist. Keep separated libraries, but only one devel package.
DLZ modules turned built-in support into named, just like former
named-sdb package had. That was non-intentional and is disabled now.
Instead, build only dynamically loaded modules with support for various
database access.
Because pending issues with PDF regeneration, disable PDF for now.
Allow turning it on with --with DOCPDF.
It prevents building successfully on Rawhide/f33 for some reason.
Subpackage is there just as shared documentation for main package.
I want to stay in original directory, files should not move since they
were in bind package.
Documentation is not regenerated, but used as shipped by upstream.
Subpackage is there just as shared documentation for main package.
I want to stay in original directory, even most of paths have changed
since move to sphinx generated documentation.
Do not depend hard on initscript just to provide fancy colored status.
When started from systemd, it does not really matter.
Return exactly the same return code as returned by the original tool.
Do not depend hard on initscript just to provide fancy colored status.
When started from systemd, it does not really matter.
Return exactly the same return code as returned by the original tool.
softhsm is not provided on RHEL 8 as normal package. It is distributed
only in idm:DL1 module. If unittest or systemtest is not enabled, skip
configuring softhsm. It would not be used anyway.
Set of patches and changes, that fixes compilation of native PKCS11
support as subpackage. Moves definition of USE_PKCS11 from config.h to
Makefiles. Defaults to off and only PKCS11 subdirectories set it to
true.
Notes for BIND 9.16.2
Security Fixes
DNS rebinding protection was ineffective when BIND 9 is configured as a forwarding DNS server. Found and responsibly reported by Tobias Klein. [GL #1574]
Known Issues
We have received reports that in some circumstances, receipt of an IXFR can cause the processing of queries to slow significantly. Some of these were related to RPZ processing, which has been fixed in this release (see below). Others appear to occur where there are NSEC3-related changes (such as an operator changing the NSEC3 salt used in the hash calculation). These are being investigated. [GL #1685]
Feature Changes
The previous DNSSEC sign statistics used lots of memory. The number of keys to track is reduced to four per zone, which should be enough for 99% of all signed zones. [GL #1179]
Bug Fixes
When an RPZ policy zone was updated via zone transfer and a large number of records was deleted, named could become nonresponsive for a short period while deleted names were removed from the RPZ summary database. This database cleanup is now done incrementally over a longer period of time, reducing such delays. [GL #1447]
When trying to migrate an already-signed zone from auto-dnssec maintain to one based on dnssec-policy, the existing keys were immediately deleted and replaced with new ones. As the key rollover timing constraints were not being followed, it was possible that some clients would not have been able to validate responses until all old DNSSEC information had timed out from caches. BIND now looks at the time metadata of the existing keys and incorporates it into its DNSSEC policy operation. [GL #1706]
From Upstream Release notes:
Security Fixes
DNS rebinding protection was ineffective when BIND 9 is configured as a forwarding DNS server. Found and responsibly reported by Tobias Klein. [GL #1574]
Known Issues
We have received reports that in some circumstances, receipt of an IXFR can cause the processing of queries to slow significantly. Some of these were related to RPZ processing, which has been fixed in this release (see below). Others appear to occur where there are NSEC3-related changes (such as an operator changing the NSEC3 salt used in the hash calculation). These are being investigated. [GL #1685]
