BSDCon '03 Nearly Here (OpenBSD 3.4, Too)

robotdreams writes "Once again BSDCon showcases the BSD community's long history of innovative research, open exchange of ideas, and collaborative work. Tutorials this year feature: an intensive code walkthrough of the new FreeBSD 5.x release, debugging kernel problems on live systems, advanced BSD system and network security, and FreeBSD's new GEOM disk I/O subsystem." Since BSDCon runs from September 8th through 12th, you're probably either already going or out of luck ;) On the other hand, you're still early for OpenBSD 3.4, now taking pre-orders -- details below.

An anonymous reader writes "Pre-orders for the OpenBSD project's latest release, 3.4, are now being taken. This release will ship around November 1st. Significant enhancements have been made in this release, including i386 switch to ELF executable format, further W^X improvements for i386, ld.so on ELF platforms now loads libraries in a random order for greater resistance to attacks, inclusion of a static bounds checker to the compiler for basic checks on functions which accept buffers and sizes, strcpy/strcat function audit to replace with safer strlcpy/strlcat, ProPolice stack protection in the kernel, further manual page cleanups, large number of bug fixes and optimizations to the packet filter (PF) including packet tagging, stateful TCP normalization, passive OS detection, SYN proxy, and adaptive state timeouts, and many other improvements to the rest of the system.

Order a CD from the OpenBSD store. Ordering a CD helps support the project, as a bonus you get cool stickers, artwork, and an audio track!"

The same reader sent links to more information on this release, including new features, and the changelog between 3.3 and 3.4.

Comments

Re:FreeBSD filesystem

[Mr. Darl McBride]

This could be incorporated in Linux, if that's what you mean. It would be a pretty major undertaking, however. This operates with hooks at the filesystem layer and the block device layer both, which means that for every filesystem implementing this, a path needs to be made through it. This would touch not only the INode queue system (which would be good, as the Moustache ride-ordered queue mentioned above is more efficient), but also the Async Network And Loopback probe since this would have to permeate 2.4's proc-based /dev filesystem, and the penal sheath bits of the thread dispatcher in order to allow a filesystem to halt the process requesting fsck-limited activity.

This could well be a good phd project for someone who wants to work on a thesis project that gives back to the community, and would surely secure them a job at Oracle or one of the other Big Data Storage Management companies.

Re:FreeBSD filesystem

[realdpk]

If ext2 is so reliable, why does it, by default, insist that I must fsck after X days or Y mounts?

OpenBSD Journal is dying (not a troll)

[evilviper]

deadly.org, which was previously the premiere place for OpenBSDers to learn and communicate has now disabled comments completely. Comments, 90% from intelligent and informed users, happened to be the highlight of the site, but now it's all over because the admins are very unhappy about a pair of trolls, and aren't willing to do anything anti-troll. instead, they shut the door completely.

So, I wonder if anyone here knows of another site that will take-over? BSD-specific, with knowledgable users, and plenty of comments... BSDforums has a large number of factors working against it, keeping it from becomming anything really good, so rule that one out.

The BSD section here at /. has EVERYTHING working against it... VERY infrequent stories, few commenters, even fewer informed commentors, and nearly as many trolls as BSDers.

Any suggestions?