Skip to main content.


Sat, 08 Dec 2012 13:52:21 -0500

Well my web server rebooted. I don't know why. I will post another story about the significance of this to me. Anyways, my /home (where all my website docs were located) was not mounted.

The disklabel output was strange:

16 partitions:
#        size    offset     fstype [fsize bsize cpg/sgs]
 a:   8192016        63     4.2BSD   2048 16384 27560  # (Cyl.      0*-   8127*)
 b:   8192016   8192079       swap                     # (Cyl.   8127*-  16254*)
 c: 156299312        63     4.2BSD   2048 16384 28752  # (Cyl.      0*- 155058*)
 d: 156299375         0     unused      0     0        # (Cyl.      0 - 155058*)
disklabel: partitions a and c overlap
disklabel: partitions b and c overlap

The disk is:

Partition table:
0: NetBSD (sysid 169)
    start 63, size 156299312 (76318 MB, Cyls 0-10337/62/30), Active

So I did some math with bc: Size of c is 156299375-63 = 156299312. Size of d is already known (entire disk). Size of e is 156299375-16384095 = 139915280. Start of e is 8192079+8192016 = 16384095 (end of swap plus size of swap).

And did with disklabel -e:

16 partitions:
#        size    offset     fstype [fsize bsize cpg/sgs]
 a:   8192016        63     4.2BSD   2048 16384 27560  # (Cyl.      0*-   8127*)
 b:   8192016   8192079       swap                     # (Cyl.   8127*-  16254*)
 c: 156299312        63     unused      0     0
 d: 156299375         0     unused      0     0
 e: 139915280  16384095    4.2BSD   2048 16384 28752  # (Cyl.      0*- 155058*)

I guess I could have used disklabel -i for interactive to have it figure out the offsets and sizes for me. Then I fsck'd it. For some reason the fstab had: /dev/wd0c for the home so I changed to wd0e. Then "mount -a" and another reboot to verify it again. Odd that this weird disklabel was in working and in use for long time.


Sat, 01 Dec 2012 13:22:53 -0500

For a few days I worked on an article for BSD Mag that introduced (briefly) the many new tools and daemons included in NetBSD 6.0. Many of this was new development done on NetBSD, while some is third-party software useful enough to be included with the base NetBSD. The article should be published by the magazine sometime in December. Another article needs to be written to introduce the many rump tools.

I noticed several man pages were missing. So I wrote a draft of mkubootimage(1) even though I haven't used it. (It is now at Other missing manuals include: sqlite3 (I may handle that), ztest, kcc (heimtools), digest-service, and gsstool.

Also the Google Code-in is in process. This is a Google sponsored event with high school students doing short (couple hour) tasks to help open source projects. NetBSD has many tasks available and I helped review and give feedback to some students.


Sat, 01 Dec 2012 17:52:28 -0500

I wrote several simple scripts to run on the official NetBSD FTP server to analyze some of the package files and available package databases (pkg_summary). I report about old or missing pkg_summary, missing important packages, duplicate names or multiple versions, old packages, and old symlinks. (Later also report if pkgin config is wrong.) Some more details are at The reports are in the same directory. It is not automated yet.


Sat, 01 Dec 2012 17:36:48 -0500

I went to BSD Can 2012. On the first day of the regular conference tracks, we were all awakened with a loud bagpipe player in the opening remarks. I attended Brett Lymn's introduction to NetBSD's veriexec which provides a kernel-based method of file integrity checking, Warren Block's presentation about documentation proofreading and his tool (igor) for helping this, Bjoern Zeeb's presentation about improving kernel debugging by using better error messages, Colin Percival's lecture about paying for bug bounties, and sat in on Matt Thomas's remote presentation about current state of MIPS hardware support in NetBSD. I also gave a lecture introducing DNSSEC, where I quickly explained signed zones with various examples and introduced many problems.

I plan to try veriexec, patch igor for my documentation formats, and propose a bug bounty for some projects I work on. I was also able to meet in person several developers that I have had emails with over the years.

The lunches were a light box lunch, but most of the developers meet up for dinners. The conference organizer, Dan Langille, also suggests tourist activities and helps coordinate other social offerings. An ongoing participant of multiple BSD conferences said this conference (and other events) is a good way to keep up with the "extended family".


Sat, 01 Dec 2012 17:33:12 -0500

Committed part of the work from Anton Panev's GSoC 2011 project to add RPM and DPKG support to pkgsrc. (Still need to import more of the changes.) I was the mentor for this Google Summer of Code project. It is for using the pkgsrc build infrastructure to create RPM packages (with spec files, etc) or DPKG packages (with control, etc) so you can use the resulting packages with rpm-type tools or apt-get type tools. Information is at