Knot Server 2.8.2 to be Available

DNS server is focused on high-performance query processing, for which a multi-threaded and well-scalable on SMP systems, is used
07 June 2019   570

The release of KnotDNS 2.8.2, a high-performance authoritative DNS server (the recursor is designed as a separate application), supporting all modern DNS capabilities, is introduced. The project is being developed by the Czech CZ.NIC name registry, written in C and distributed under the GPLv3 license. The server is characterized by a focus on high-performance query processing, for which a multi-threaded and well-scalable on SMP systems, is used. Features include adding and deleting zones on the fly, zone transfer between servers, DDNS (dynamic updates), NSID (RFC 5001), extensions EDNS0 and DNSSEC (including NSEC3), response rate limits (RRL).

New features:

  • New blocking mode for zone event triggers in knotc
  • New weighted records mode in the module geoip (Thanks to Conrad Hoffmann)
  • Module noudp allows UDP allow rate configuration

And improvements:

  • NSEC3 salt lifetime can be set to infinity
  • New 'running' zone event status in the knotc output
  • Knotc in the forced mode returns failure also if zone check emits any warning
  • Ignoring PMTU information for IPv4/UDP via IP_PMTUDISC_OMIT (Thanks to Daisuke Higashi)
  • Various improvements in the documentation

Linux 5.3 Kernel to be Released

Huge amount of updates, improvements, changes and new features awaits all Linux users
17 September 2019   112

After two months of development, Linus Torvalds introduced the Linux 5.3 kernel release. Among the most notable changes: support for AMD Navi GPUs, Zhaoxi processors, and Intel Speed ​​Select power management technology, the ability to use umwait instructions to wait without using loops, increasing the interactivity utilization clamping mode for asymmetric CPUs, the pidfd_open system call, the ability to use IPv4 addresses from the subnet, the possibility of hardware acceleration of nftables, support for HDR in the DRM subsystem, integration of the ACRN hypervisor.

In the announcement of the new release, Linus reminded all developers of the main rule of kernel development - maintaining the invariance of behavior for user space components. Changes in the kernel should in no way violate already running applications and lead to user-level regressions. At the same time, a violation of behavior can cause not only a change in the ABI, removal of outdated code or errors, but also an indirect effect of correctly working useful improvements. As a good example, the useful optimization in Ext4 code was discarded, which reduces the number of accesses to the drive by disabling the read-ahead inode table for small I / O requests.

Optimization has led to the fact that, due to a decrease in disk activity, the entropy for the random number generator getrandom () began to accumulate more slowly and in some configurations, under certain circumstances, there could be hangs during loading until the entropy pool is full. Since the optimization is really useful, a discussion arose among the developers, in which it was proposed to eliminate the problem by disabling the default blocking mode of the getrandom () call with the addition of an optional flag to wait for entropy, but such a change will affect the quality of random numbers at the initial stage of loading. In the change rollback commit, Linus noted that he plans to bring the optimization back as soon as the problem with getrandom () is resolved.

The new version adopted 15794 patches from 1974 developers, the patch size is 92 MB (the changes affected 13986 files, 258419 lines of code were added, 599137 lines were deleted). About 39% of all the changes presented in 5.3 are related to device drivers, about 12% of changes are related to updating the code specific to hardware architectures, 11% are connected to the network stack, 3% to file systems and 3% to internal kernel subsystems.

Get more information about the new features and from the mailing.