Linux in the news
All in one big page
See also: last week's Kernel page.
The current kernel release is 2.2.6. Linus returned from his vacation, tossed out 2.2.6 without an announcement, then hit the road again for his midwestern speaking engagements. So not much else has happened in the area of official kernel releases. The number of complaints about 2.2.6 seems to be quite small, thus far.
The NFSv3 client has moved from alpha to beta test. See Trond Myklebust's announcement for the full story. Included with the release is a new version of mount which automatically uses NFSv3 if the server offers it. Trond's plan is to push for inclusion of the NFSv3 client in the 2.3 tree, once that begins; he does not plan to try to get it into the 2.2 series.
A Linux hardware test suite? Doug Ledford, master of Adaptec SCSI, among other things, sent out a message announcing a plan to gather together a set of kernel hardware compatibility test programs. The idea is to test all aspects of how the kernel works with a particular device. They are looking for existing test utilities that people might have sitting around. Please drop them a note if you have something to contribute.
Performance degration after a few hours of operation was a recent complaint. Turns out that the people involved were running Compaq systems; said systems have a power management system that will slow down the CPU after a few hours of operation. Disabling the power management in the BIOS fixed the problem...
Problems mounting SMB shares with smbfs have also been in the air. It appears that there are at least two different problems out there. One is that the "Windows 95 bug workaround" kernel compilation option creates difficulties if you are mounting shares from NT servers. If that is your situation, the bug workaround option should be turned off. There also appears to be an idle timeout problem that cropped up recently; that one is still outstanding.
A new software RAID release is out, this is version 1999.04.21. It applies to both 2.0.36 and 2.2.6, and is primarily a bug-fix release.
The capabilities discussion rages on. There is not a whole lot more to say about it here; folks really interested in the discussion should probably participate in it directly. It is worthwhile, however, to have a look at this summary of the discussionposted by Ted T'so, written in his usual clear manner. It sums up a few different views of capabilities, and argues that capability information has to be stored in the filesystem.
Section Editor: Jon Corbet
April 22, 1999