2.6.12-dark2: Learn To Swim

Message

vipernicus · Post by **vipernicus** » Sat Jun 25, 2005 5:12 pm

First release as dark-sources maintainer. Enjoy.

release notes:

/* Dark-sources
* Release: 2.6.12-dark2
* Release name: Learn To Swim
* Maintainer: vipernicus
* Comments: switch from 2.6.12 to 2.6.12.1, added -kj patches, ext3 latency (-mm),
* Comments: iosched selector (fallow), vesafb-tng & fbsplash, inotify, execshield-nx
*/

Code: Select all

/* Dark core (mostly -ck patches) */
sched-run_normal_with_rt_on_sibling.diff /* Real-time hyper-threading bugfix merged from -ck */
staircase11.3-2.6.12.diff /* Staircase 11.3 */
schedrange.diff /* Infrastructure for more scheduling policies */
schedbatch2.8.diff /* IDLE CPU scheduling, tasks set as batch can only use idle CPU */
schediso2.12.diff /* Isochronous scheduling, soft-realtime task scheduling */
sched-fix_up_build.patch /* sched fixups */
smp-nice-support6.diff /* Support for 'nice' to work on SMP systems */
cfq-2.6.12-mm1.patch /* CFQ-TS */
isobatch_ionice2.diff /* SCHED_ISO/BATCH ionice values */
rt_ionice.diff /* SCHED_RR ionice values */
mapped_watermark3.diff /* Mapped watermark III, which cuts down on swap usage */
realtime-lsm.diff /* Realtime-lsm, allows non-root users to set tasks as SCHED_RR/FIFO */
badram-2.6.12.diff /* BadRAM, allows people with faulty memory to still use there system */

/* Hardware support */
skge-netdev-2.6.11-mm4.patch /* SKGE, sk98lin replacement */

/* New features */
2.6.12-kj.patch /* Kernel Janitors, code clean-ups and fixes */
exec-shield-nx-2.6.12-B2 /* Newest exec-shield by mingo */
def_iosched_select_menu.diff /* IO scheduler selection */
ext3-latencies-mm1.patch */ ext3 latency fixes from mm */
inotify-0.23-rml-2.6.12-13.patch */ Newest inotify, soon to replace dnotify */
vesafb-tng-0.9-rc6-r2-2.6.12-rc1.patch */ Improved vesa framebuffer */
fbsplash-0.9.2-r3-2.6.12-rc3.patch */ gensplash for that pretty little splash screen */

/* Other */
darktag.diff /* Tag release numbers and names for -dark. */

patch: http://vipernicus.exonotion.com/patch-2 ... -dark2.bz2

Tips:
For the new fbsplash, set up grub like this:title Gentoo Linux 2005.0 [2.6.12.1-dark2]
root (hd0,1)
kernel (hd0,1)/vmlinuz-2.6.12.1-dark2 ro root=/dev/hda3 video=vesafb:ywrap,mtrr,1280x1024-32@100 splash=silent,fadein,tty:8,theme:livecd-2005.0 quiet CONSOLE=/dev/tty1
initrd (hd0,1)/fbsplash-livecd-1280x1024

predatorfreak · Post by **predatorfreak** » Sat Jun 25, 2005 11:41 pm

vipernicus wrote:First release as dark-sources maintainer. Enjoy.

release notes:

Code: Select all

/* Dark-sources
 * Release: 2.6.12-dark2
 * Release name: Learn To Swim
 * Maintainer: vipernicus
 * Comments: switch from 2.6.12 to 2.6.12.1, added -kj patches, ext3 latency (-mm),
 * Comments: iosched selector (fallow), vesafb-tng & fbsplash, inotify, execshield-nx
 */

/* Dark core (mostly -ck patches) */
sched-run_normal_with_rt_on_sibling.diff /* Real-time hyper-threading bugfix merged from -ck */
staircase11.3-2.6.12.diff /* Staircase 11.3 */
schedrange.diff /* Infrastructure for more scheduling policies */
schedbatch2.8.diff /* IDLE CPU scheduling, tasks set as batch can only use idle CPU */
schediso2.12.diff /* Isochronous scheduling, soft-realtime task scheduling */
sched-fix_up_build.patch /* sched fixups */
smp-nice-support5.diff /* Support for 'nice' to work on SMP systems */
cfq-2.6.12-mm1.patch /* CFQ-TS */
isobatch_ionice2.diff /* SCHED_ISO/BATCH ionice values */
rt_ionice.diff /* SCHED_RR ionice values */
mapped_watermark3.diff /* Mapped watermark III, which cuts down on swap usage */
realtime-lsm.diff /* Realtime-lsm, allows non-root users to set tasks as SCHED_RR/FIFO */
badram-2.6.12.diff /* BadRAM, allows people with faulty memory to still use there system */

/* Hardware support */
skge-netdev-2.6.11-mm4.patch /* SKGE, sk98lin replacement */

/* New features */
2.6.12-kj.patch /* Kernel Janitors, code clean-ups and fixes */
exec-shield-nx-2.6.12-B2 /* Newest exec-shield by mingo */
def_iosched_select_menu.diff /* IO scheduler selection */
ext3-latencies-mm1.patch */ ext3 latency fixes from mm */
inotify-0.23-rml-2.6.12-13.patch */ Newest inotify, soon to replace dnotify */
vesafb-tng-0.9-rc6-r2-2.6.12-rc1.patch */ Improved vesa framebuffer */
fbsplash-0.9.2-r3-2.6.12-rc3.patch */ gensplash for that pretty little splash screen */

/* Other */
darktag.diff /* Tag release numbers and names for -dark. */

patch: http://vipernicus.exonotion.com/patch-2 ... -dark2.bz2

sched-fix_up_build.patch is only required with smp-nice-support6.diff, which is the latest version of SMP nice. That patch is a simple build fix for uniprocessor systems which effects smp-nice-support6.diff.

Edit: Other then that, nice job.
Edit: I'll also mirror this release on sourceforge.
Edit: Mirror: http://prdownloads.sourceforge.net/dark ... 2?download

vipernicus · Post by **vipernicus** » Sun Jun 26, 2005 12:10 am

predatorfreak wrote:sched-fix_up_build.patch is only required with smp-nice-support6.diff, which is the latest version of SMP nice. That patch is a simple build fix for uniprocessor systems which effects smp-nice-support6.diff.

Sorry, typo on my part, i do have smp-nice-support6 patched.

vipernicus · Post by **vipernicus** » Mon Jun 27, 2005 6:04 pm

Expect new release either tonight or tomorrow. Going to add the new execshield-nx in, genpatches, and maybe create an ebuild for those who don't like to patch manually. Anyone have any requests?

Also thinking about making an unstable splitoff called viper-sources, with rtp. I personally think FullRT is getting better all the time. Anyone interested?

predatorfreak · Post by **predatorfreak** » Tue Jun 28, 2005 3:22 am

vipernicus wrote:Expect new release either tonight or tomorrow. Going to add the new execshield-nx in, genpatches, and maybe create an ebuild for those who don't like to patch manually. Anyone have any requests?

Also thinking about making an unstable splitoff called viper-sources, with rtp. I personally think FullRT is getting better all the time. Anyone interested?

Personally, I'd rather not see dark split with Realtime Preempt, as that really isn't suited towards desktop/server users.

vipernicus · Post by **vipernicus** » Tue Jun 28, 2005 4:31 am

I'm sorry, not a split off, more of a totally different patchset called viper-sources, I like to boot into a FullRT kernel occasionally for movies.

predatorfreak · Post by **predatorfreak** » Tue Jun 28, 2005 4:45 am

vipernicus wrote:I'm sorry, not a split off, more of a totally different patchset called viper-sources, I like to boot into a FullRT kernel occasionally for movies.

Then your welcome to go for it.

bollucks · Post by **bollucks** » Tue Jun 28, 2005 7:10 am

For movies? Surely you don't need RT for that...

vipernicus · Post by **vipernicus** » Tue Jun 28, 2005 1:05 pm

bollucks wrote:For movies? Surely you don't need RT for that...

I haven't done much testing, but my hopes were that RT might help with the encoding process, takes almost 12 hours or more.

bollucks · Post by **bollucks** » Tue Jun 28, 2005 1:42 pm

vipernicus wrote:
bollucks wrote:For movies? Surely you don't need RT for that...
I haven't done much testing, but my hopes were that RT might help with the encoding process, takes almost 12 hours or more.

Ok serious misunderstanding then. RT will make the encoding process longer. It only helps microlatencies for things like audio and video sampling and for the rest of the time incurs much more overhead. For playback of audio or video RT is not required, and for cpu intensive things (eg video encoding) it is detrimental.

vipernicus · Post by **vipernicus** » Tue Jun 28, 2005 1:47 pm

Interesting, are there any scientific papers on mingo's rtp?

bollucks · Post by **bollucks** » Wed Jun 29, 2005 12:20 am

No.

When code is in that heavy a stage of development it is impossible to document until it stabilises.

predatorfreak · Post by **predatorfreak** » Fri Jul 01, 2005 2:10 pm

I see that the new release is still unreleased, I think that by now It should be fully tested even with set-backs :\.

vipernicus · Post by **vipernicus** » Fri Jul 01, 2005 2:27 pm

predatorfreak wrote:I see that the new release is still unreleased, I think that by now It should be fully tested even with set-backs :\.

Been having to work overtime at my job lately. I haven't had enough time to do much testing. I emailed you on your dcaf-security email.

predatorfreak · Post by **predatorfreak** » Mon Jul 04, 2005 11:41 pm

According to the email vipernicus sent, he's unable to maintain dark anymore due to his job, so I figured I'd make an offical announcement, that dark is likely to become unmaintained unless someone can meet the below requirements and has the free time to develop dark.

Maintainer requirements:
1. Must realise -dark's goal, I'm not just handing the torch to anyone here.
2. Must beable to work with these goals
3. Must beable to fix basic rejects
4. Must be willing to boot-test and run each release for 24 hours (per build)

Edit: As of now, dark is offically unmaintained, it was fun well it lasted.
Edit: I've decided, that I will produce patches, but leave testing to the users, after releasing a new version and I receive either positive or negative feedback, I'll deem it either stable or unstable.

redhotchojo · Post by **redhotchojo** » Sat Jul 23, 2005 5:14 pm

i ve tried dark3 and my system is working very good.
after that i ve tried to apply suspend2 patch over the dark but i failed to succes

2.6.12-dark2: Learn To Swim

2.6.12-dark2: Learn To Swim

Re: 2.6.12-dark2: Learn To Swim

Re: 2.6.12-dark2: Learn To Swim