Linux creator Linus Torvalds released 2.6.0-test10 nearly one month after 2.6.0-test9 [story]. The compressed patch weighs in at a tiny 100kB, comprised mainly of one-liners and trivial fixes. Linus explains, "I've tried to be quite strict in patch acceptance, so the changes are largely fixes for things that can crash the machine, and they are also of the type 'this can't possibly break anything'. But hey, we all know that theory and practice don't always match ;)"
There is a known issue triggered by kernel preemption, so for maximum stability Linus recommends that one should not enable CONFIG_PREEMPT. However he goes on to add, "I'd love to have more testing, so that we can try to figure out what the pattern is - but please mention explicitly that you ran with preemption if you have problems." Finally, Linus adresses when we can expect the official 2.6.0 release:
"I'm planning/hoping on basically turning this over to Andrew [interview], and let him decide to make the final 2.6.0 or not. Timing-wise Andrew is apparently going to be off for a few weeks, so regardless of whether this turns out to be rock solid or not, we'll have a few weeks of final testing before that were to happen. Which means that I might still end up making a test11 if Andrew hasn't come back and we find something that warrants it."
Martin J. Bligh has released 2.6.0-test9-mjb1 saying:
The patchset contains mainly performance, scalability and NUMA stuff,
and anything else that stops things from irritating me. It's meant to be
pretty stable, not so much a testing ground for new stuff.
I'd be very interested in feedback from anyone willing to test on any
platform, however large or small.
Sorry for giving this long msg.
[root@hwtest1 linux-2.6.0-test9]# make menuconfig
make: `scripts/fixdep' is up to date.
In file included from scripts/lxdialog/checklist.c:24:
scripts/lxdialog/dialog.h:29:20: curses.h: No such file or directory
In file included from scripts/lxdialog/checklist.c:24:
scripts/lxdialog/dialog.h:127: error: syntax error before "use_colors"
I've just tried to move from 2.4.18 to 2.6.0-test9 and encountered some problems with multimedia apps. XMMS eats up to 20% of cpu time in 2.6.0-test9 compared to less than 3% in 2.4.18 (my PC is K6-II 500). Same with mplayer - it became quite slow... I've tried ALSA modules for my sound card and OSS modules, but nothing helped. Does anyone have the same problems? And what could be a solution?
Marcelo Tosatti released 2.4.23-rc3, likely the final 2.4.23 release candidate. Referring to his comment that rc2 would probably become the final version of 2.4.23 [story], Marcelo explained, "a few critical problems showed up, so here is -rc3." If no additional critical issues are found, we can expect this to soon become the official 2.4.23 release.
Some of the more significant changes in 2.4.23 over 2.4.22 include: numerous networking updates [story]; DRM support for XFree86 v4.3.0 and above [story]; significant updates to the VM including the removal of the OOM killer [story]; several ACPI fixes [story] [story] [story] [story]; support for "laptop mode" [story]; and a large assortment of other updates and fixes. Also during this release cycle, 2.4 kernel maintainer Marcelo Tosatti moved to a new employer [story].
Read on for the complete changelog which lists the specific changes since -rc2.
Andrew Morton [interview] released 2.6.0-test9-mm5 today with the following brief comments:
"Random medium-priority bits and pieces from various people. mm5 should fix the BUGs which testers were seeing with sound cards which memory map their audio buffers."
Read on for the full changelog.
Greg Kroah-Hartman released version 006 of udev for the 2.6 kernel [forum], an "implementation of devfs in userspace using sysfs and /sbin/hotplug." Greg offers a lengthy list of major changes since version 005 was released [story] noting, "many, many thanks to Kay Sievers, for this release, for implementing many of the new features. I really appreciate it."
This latest release of udev includes among the many changes: rules are applied in the proper priority instead of simply the order they show up in the configuration file; partitions work properly for all types of rules, not just the LABEL rule type; subdirectories within /udev are handled properly; performs better parsing to better handle broken files; and includes a new
test directory with the beginnings of a regression test suite. Read on for the full announcement and the lengthy changelog.
2.4 stable kernel maintainer Marcelo Tosatti released 2.4.23-rc2 summarizing, "important netfilter fixes, several ACPI fixes, few driver corrections (tulip, tg3, megaraid2), amongst others." He went add to add, "if no problems shows up this should become final in a few days."
Read on for the complete changelog.
This is probably super-basic, but this is my first "testing" kernel... I usually go after the precompiled packages, but I like 2.6 too much to wait. Anyway, I upgraded from the stock Fedora Core1 2.4 kernel which loads ext3 as a module. If I try that in 2.6 I get a kernel panic, so I built it into the kernel and everything is happy. During boot, however, the kernel still tries to load ext3.o as a module and yells at me. Along the same line, I can't get evdev to load early enough in the boot to let my synaptics touchpad load in time to allow a GUI startup screen...
Upcoming 2.6 stable kernel [forum] maintainer Andrew Morton [interview] has released 2.6.0-test9-mm4. He notes that this release includes fixes for patches that are currently only in the -mm tree [story]. Also, as 2.6.0-test is currently in a freeze due to the approaching official 2.6.0 release, Andrew notes that his -mm tree has begun to collect "minor fixes which we'll queue for post-2.6.0".
A patch added in 2.6.0-test9-mm3 revealed interactivity issues which have most likely been fixed in this latest release. Andrew explains, "The interactivity problems which the ACPI PM timer patch showed up should be fixed here - please sing out if not.". Read on for the complete changelog.
Hi, i was using ipv6 in my kernel and everything was going ok until i decided to run a dhcp server. to run dhcpd I have to have CONFIG_FILTER=y in the kernel. so I recompiled the kernel with this option and the modules ipsec and ipv6 can't load.
# modprobe ipv6
/lib/modules/2.4.20-gentoo-r8/kernel/net/ipv6/ipv6.o: unresolved symbol sk_run_filter
# depmod -a
depmod: *** Unresolved symbols in /lib/modules/2.4.20-gentoo-r8/kernel/net/ipsec/ipsec.o
Has anyone else noticed that on 2.6.test9 NIS no longer works?
Though NIS is started on boot up it appears not to be running when you get to a log in prompt.
Running NIS manually as root seems to bring it up. Using an NIS service (e.g. logging in if ypbind is NIS'd) will then hang.
On my set up the services file is NIS'd too (as per /etc/nsswitch) hence this bug causes loss of connectivity (Remote sessions cant connect to the local X server, connection refused - no protocol).
I have recently released v19a of my CPU scheduler for Linus and mm kernels. Hosted by kerneltrap here http://www.kerneltrap.org/~npiggin/v19a/. This release contains the first interactivity patches for some time, as I have been busy looking at improving NUMA and SMP balancing in previous releases.
The interactivity work includes microsecond accounting for priority calculation as well as timeslices, improved dynamic priority responsiveness for preempted tasks, and improved fairness when the CPU starts running after becoming idle.
These are quite big changes, and I would like to hear any feedback from my small band of users, or if you are just interested in discussing any aspects please do so here or email me personally. Thank you!