Episode Details

Back to Episodes
111: Xenocratic Oath

111: Xenocratic Oath

Published 10 years, 5 months ago
Description

Coming up on this weeks episode, we have BSD news, tidbits and articles out the wazoo to share. Also, be sure to stick around for our interview with Brandon Mercer as he tells us about OpenBSD being used in the healthcare industry.

This episode was brought to you by

iXsystems - Enterprise Servers and Storage for Open SourceDigitalOcean - Simple Cloud Hosting, Built for DevelopersTarsnap - Online Backups for the Truly Paranoid


Headlines

NetBSD 7.0 Release Announcement

  • DRM/KMS support brings accelerated graphics to x86 systems using modern Intel and Radeon devices (Linux 3.15)
  • Multiprocessor ARM support.
  • Support for many new ARM boards, including the Raspberry Pi 2 and BeagleBone Black
  • Major NPF improvements:
    • BPF with just-in-time (JIT) compilation by default
    • support for dynamic rules
  • support for static (stateless) NAT
  • support for IPv6-to-IPv6 Network Prefix Translation (NPTv6) as per RFC 6296
  • support for CDB based tables (uses perfect hashing and guarantees lock-free O(1) lookups)
  • Multiprocessor support in the USB subsystem.
  • GPT support in sysinst via the extended partitioning menu.
  • Lua kernel scripting
  • GCC 4.8.4, which brings support for C++11
  • Experimental support for SSD TRIM in wd(4) and FFS
  • tetris(6): Add colours and a 'down' key, defaulting to 'n'. It moves the block down a line, if it fits. ***

CloudFlare develops interesting new netmap feature

  • Normally, when Netmap is enabled on an interface, the kernel is bypassed and all of the packets go to the Netmap consumers
  • CloudFlare has developed a feature that allows all but one of the RX queues to remain connected to the kernel, and only a single queue be passed to Netmap
  • The change is a simple modification to the nm_open API, allowing the application to open only a specific queue of the NIC, rather than the entire thing
  • The RSS or other hashing must be modified to not direct traffic to this queue
  • Then specific flows are directed to the netmap application for matching traffic
  • For example under Linux:
  • ethtool -X eth3 weight 1 1 1 1 0 1 1 1 1 1
  • ethtool -K eth3 lro off gro off
  • ethtool -N eth3 flow-type udp4 dst-port 53 action 4
  • Directs all name server traffic to NIC queue number 4
  • Currently there is no tool like ethtool to accomplish this same under FreeBSD
  • I wonder if the flows could be identified more specifically using something like ipfw-netmap ***

Building your own OpenBSD based Mail server!

Listen Now

Love PodBriefly?

If you like Podbriefly.com, please consider donating to support the ongoing development.

Support Us