NetBSD 6.0_BETA binaries available for testing

March 13, 2012 posted by Jeff Rizzo

On behalf of NetBSD developers, I'm happy to announce the availability of a public beta of NetBSD 6.0, for your testing pleasure.

This beta is substantially feature-complete; there may be some additional changes to the installer and possibly some additional hardware support if some is found missing and is easily added, but the major changes are done. What we need now is for you, the end users, to test it in your preferred configuration.

Please DO keep in mind, however, that this is a beta release; we fully expect there are some lurking bugs, so use with caution.

Binaries of NetBSD 6.0_BETA are available for download at:

Bootable (for many arches) ISO images are also available for download via HTTP, FTP, and BitTorrent.

Here are some highlighted changes since NetBSD 5.0. As always, full details can be found in doc/CHANGES and doc/CHANGES-6.0:

  • time_t is now 64 bits.
  • syslog improvements: Reliable TCP connections, encryption, syslog protocol API.
  • LKMs removed; superseded by the new module(7) framework.
  • Boot loaders on some arches (i386, amd64) support loading modules at boot.
  • Added crash(8) a new utility based on ddb(4) to diagnose kernel crashes.
  • Added netpgp(1), a BSD-licensed implementation of PGP.
  • Added LVM (Logical Volume Manager) functionality.
  • Multiprocessor support for Xen PV DomUs.
  • Xen2 support has been dropped.
  • Better Xen PV support on Linux dom0, including Citrix XenServer and Xen Cloud Platform.
  • Much improved copmpat_linux support for running Linux binaries.
  • was updated with new versions of most utilities, and the X server.
  • Some arches now default to UFS2 in sysinst.
  • FFS: softdep is no longer available, use WAPBL logging instead.
  • gpio(4) has been completely reworked to integrate with kauth(9)
  • evbarm now have support for Gumstix Verdex and Verdex Pro, Marvell Sheevaplug and other Marvell SoC NAS boxes, i.MX51 SoC.
  • arm platforms has support for Cortex-A8 CPUs.
  • pfsync(4) from OpenBSD 4.2
  • mDNSResponder is now in base.
  • raid(4) now has parity maps, greatly improving parity rewrite times after unclean shutdown.
  • Added support for 64-bit MIPS processors (O32, N32, N64 ABIs are supported)
  • Added mkubootimage(1) tool for generating u-boot kernel images.
  • Added NPF - the NetBSD Packet Filter - a work in progress.
  • xz(1) - imported XZ compression tool.
  • resize_ffs(8) - support for growing FFSv1 and FFSv2 file systems, and shrinking FFSv1.
  • amd64,i386: booting from a disk with GUID Partition Table (gpt) is now possible.
  • iSCSI: added an in-kernel iSCSI initiator, from Wasabi Systems.
  • New port: eMIPS
  • Added flash(9) and nand(9) subsystems to handle flash devices and NAND controllers.
  • Added CHFS, a file system for flash(9) devices.
  • Reworked quota subsystem for FFS
  • Added TLS (Thread Local Storage) support for most platforms.
  • Added dtv(4), a Digital TV framework.
  • MIPS: add support for RALink RT3883 SoC
  • sparc64: add support for Enterprise (Ex[45]00) systems, most ultrasparc III and IIIi systems.
  • gcc 4.5.3 is the default compiler
  • gdb 7.3.1 in-tree
  • Support for building most of the tree with clang.
  • devpubd(8) added, a device publishing daemon
  • Xen: support for suspend/resume
  • SQLite 3 is now in the base system.
  • audio(9): audio drivers are now MP-safe.
  • tprof(8): a sampling-based profiler
  • x86, Xen: added CPU microcode loading support via cpuctl(8).
  • Trusted Platform support added: TrouSerS, tpm-tools, and tpm(4).
  • Added posix_spawn() functions.
  • New apropos(1) implementation using SQLite Full Text Index.

New drivers added to NetBSD 6.0: bwi(4), age(4), atphy(4), ale(4), sdmmc(4), smsh(4), voyagerfb(4), gpioiic(4), tcx(4), zx(4), hdaudio(4), wb(4), lom(4), acpiwmi(4), udl(4), gfb(4), cas(4), acpismbus(4), uthum(4), cgtwelve(4), upgt(4), omapfb(4), otus(4), fujbp(4), fujhk(4), auvitek(4), acpiwdrt(4), vte(4), alc(4), rdcide(4), rdcpcib(4), s390rtc(4), cxdtv(4), hpacel(4), emdtv(4), urndis(4), p5pb(4), pwdog(4), mppb(4), vmt(4), efa(4), gpiopwm(4), powsw(4), tpm(4), valkyriefb(4), videopll(4).

...and LOTS more hardware support added to existing drivers! Please help us out by testing these changes and reporting problems either to an appropriate mailing list, via send-pr, or via the web form.

Please remember that this is a BETA release. Don't use it anywhere where the quirks of a beta might cause problems from which it would be hard to recover!

Thanks again for your help in making NetBSD 6.0 the best release yet!




Can you please tell me, is NetBSD 6.0 support CPU/RAM hotswap on Sun E4500?

Posted by valeriy on March 14, 2012 at 09:21 AM UTC #

Is there any chance what memstick (USB) images will be available for i386/amd64?

Posted by Labla on March 14, 2012 at 09:31 AM UTC #

Does this beta use the same pkgsrc as 5.1.2 which I assume is 2011Q4 ? Thanks for all the great work. Dennis

Posted by Dennis Kikendall on March 14, 2012 at 02:45 PM UTC #

Brief answers (though this isn't a great place for these kinds of questions): 1) No, no hotswap ram/cpu 2) The release will for certain have memstick images. I can try to make them for 6.0_BETA, but I'm not 100% sure I'll get to it this week 3) pkgsrc is compatible, though there may be some fixes in pkgsrc-current which have not yet been pulled up to 2011Q4. It is very likely that 2012Q1 will be cut before the 6.0 release. Also, there are 6.0_BETA binary packages available for some architectures (amd64 and i386 at this point, maybe more shortly).

Posted by Jeff Rizzo on March 14, 2012 at 03:57 PM UTC #

On page there is NO any link directly to this page . Therefore , it is less likely that people will learn its presence without inspecting this page .

Posted by Mehmet Erol Sanliturk on March 14, 2012 at 06:28 PM UTC #

We welcome NetBSD 6.0. Thanks for great work. Just a note: Checksums are missing ro beta downloads.

Posted by Cem K on March 14, 2012 at 11:37 PM UTC #

Is RUMP still being worked on? Will it be improved in any way in NetBSD 6.0? Thanks for the great work everyone.

Posted by Matt on March 17, 2012 at 07:24 PM UTC #

I can understand going to NetBSD 6.0 for many reasons. But why isn't the new release 5.1.3 instead? The current release seems functional

Posted by stan_in_usa on March 20, 2012 at 11:29 PM UTC #

Good job! thanks.

Posted by using9 on March 21, 2012 at 11:39 PM UTC #

So far no joy with the i386 iso on MacOSX version of Virtualbox. I had issues with the partitions pushing past the disk geometry. I ran through it several times. I decided to leave the geometry alone and instead adjusted the mbr partition. I also changed both the root and swap partition sizes to ensure they didn't attempt to push past the boundary. The software started to install but stalled and did not recover. The stall may have something to do with using the host io for the virtual disks. I am going to try it again with some other option choices today. I am just about as "user" as you can get with NetBSD. I am more familiar with FreeBSD / OpenBSD and much more familiar with Linux distributions.

Posted by James Carter on April 05, 2012 at 02:09 PM UTC #

Whoops. I apologize I did not see the bug reporting link. I have since reported the issue. FYI, I just ran through an installation with the amd64 iso and installation was successful. I did change my memory allocation from 128MB to 256MB. This may or may not have something to do with it.

Posted by James Carter on April 05, 2012 at 02:43 PM UTC #

Responding to the question of why NetBSD 6.0 is not 5.1.3, there are some improvements, including Linux ext2fs read/writeability, that were added to NetBSD 6, then 5.99.xx, that were ported back to 5.1_STABLE but not 5.1.1 and 5.1.2; also the new framebuffer in 6.0 though this was not ported back to 5.1_STABLE.

Posted by Thomas Mueller on April 22, 2012 at 02:20 AM UTC #

§1 Tried with the v6B amd64 iso but freezes while loading ACPI CPU. The iso image is downloaded from Danish netbsd repository §2 Booted with 'boot -1' parameter to disable ACPI, but froze at the ACPI CPU loading in my case. Seems the sysinst bug is yet to be addressed with this specific image. Any inputs? Thanks!

Posted by zenny on April 25, 2012 at 03:43 PM UTC #

can it support chinese-traditional very well, I use chinese,yet

Posted by hyoscyamus on May 29, 2012 at 05:28 PM UTC #

Post a Comment:
Comments are closed for this entry.