View previous topic :: View next topic |
Author |
Message |
iphitus Apprentice
Joined: 03 Aug 2005 Posts: 226
|
Posted: Mon Dec 26, 2005 11:48 pm Post subject: 2.6.14-archck7 |
|
|
ArchCK is a patchset based on the CK patchset. ArchCK aims to include
a variety of popular features and updates that have not yet made it to
the vanilla kernel, without compromising stability. Con Kolivas' CK
patchset is the most important part of ArchCK, as it provides the
great interactivity and responsiveness improvements.
First release in a very long while. I've been busy with exams, and
school, but fortunately its holidays now.
Apply this patch to 2.6.14. It includes 2.6.14.4.
Patch (Apply patch to 2.6.14):
http://iphitus.loudas.com/arch/ck/2.6.14/patch-2.6.14-archck7.bz2
Broken out:
http://iphitus.loudas.com/arch/ck/2.6.14/2.6.14-archck7-broken-out.tar.bz2
Website:
http://iphitus.loudas.com/archck.php
Changelog:
updated ck to 2.6.14-ck7
updated suspend2 to 2.2-rc15
updated fbsplash and vesafb-tng
new reiser4 patch from the namesys ftp
Full Changelog available on website.
Errata:
Reiser4: Debug mode must be off.
hz extra values: If you choose a value other than 100, or 1000 and
have problems, chances are that is what is causing those issues. If
you are unsure, select 1000.
Included:
CK Patchset
fbsplash
vesafb-tng
reiser4
shfs
Squashfs
UnionFS
Suspend2
libATA (SATA) Updates
Layer7
Configurable Mouse Polling
Realtime-LSM
Custom DSDT
Ebuilds:
https://bugs.gentoo.org/show_bug.cgi?id=103354
Bootsplash release is here -- Dont use. I don't accept bug reports for this:
http://iphitus.loudas.com/arch/ck/2.6.14/patch-2.6.14-archck7-bootsplash.bz2 |
|
Back to top |
|
|
babykart Guru
Joined: 08 Oct 2004 Posts: 415
|
Posted: Tue Dec 27, 2005 2:41 am Post subject: |
|
|
one more time an excellent job iphitus...
smp/smt + sata + reiser4 + nvidia + gcc4... _________________ >> Gentoo-FR <<
----- |
|
Back to top |
|
|
lancelott Apprentice
Joined: 29 Mar 2005 Posts: 152
|
Posted: Tue Dec 27, 2005 5:02 am Post subject: |
|
|
Is this one's staircase cpusched any better? I made the move from archck5 to nitro2 and found that I got vastly better performance and less music skippage with nitro than archck when doing anything with portage (with PORTAGE_NICENESS=20 in both cases...). I'm assuming this is related to the cpusched, as I tried anticipatory as my iosched in both kernels.
I may try this, but with that in mind, it doesn't look like that great of an option for me anymore. I have a limited amount of RAM (256mb) so I need performance. _________________ Desktop: AMD Athlon 64 3000+ @ 2.15GHz | 7600GT 256MB | 320GB + 250GB SATA 3.0 | Creative X-Fi XtremeGamer | Gentoo Linux amd64
Server: Intel Pentium 2 MMX 350MHz | 30GB HDD | Gentoo Linux x86 |
|
Back to top |
|
|
Phlogiston Veteran
Joined: 27 Jan 2004 Posts: 1925 Location: Europe, Swizerland
|
|
Back to top |
|
|
landon Apprentice
Joined: 09 Mar 2004 Posts: 159
|
Posted: Tue Dec 27, 2005 2:05 pm Post subject: |
|
|
lancelott wrote: | I made the move from archck5 to nitro2 and found that I got vastly better performance and less music skippage with nitro than archck when doing anything with portage (with PORTAGE_NICENESS=20 in both cases...). |
I'm afraid I can't say the same thing. |
|
Back to top |
|
|
morbid Tux's lil' helper
Joined: 05 Mar 2003 Posts: 90
|
Posted: Tue Dec 27, 2005 8:05 pm Post subject: |
|
|
lancelott wrote: | Is this one's staircase cpusched any better? I made the move from archck5 to nitro2 and found that I got vastly better performance and less music skippage with nitro than archck when doing anything with portage (with PORTAGE_NICENESS=20 in both cases...). I'm assuming this is related to the cpusched, as I tried anticipatory as my iosched in both kernels.
I may try this, but with that in mind, it doesn't look like that great of an option for me anymore. I have a limited amount of RAM (256mb) so I need performance. |
you probably have pre-empt enabled. My music never skips... even under very heavy loads. The only time it has... was when I played around with preempt -- I quickly disabled it. |
|
Back to top |
|
|
lancelott Apprentice
Joined: 29 Mar 2005 Posts: 152
|
Posted: Tue Dec 27, 2005 9:40 pm Post subject: |
|
|
morbid wrote: | lancelott wrote: | Is this one's staircase cpusched any better? I made the move from archck5 to nitro2 and found that I got vastly better performance and less music skippage with nitro than archck when doing anything with portage (with PORTAGE_NICENESS=20 in both cases...). I'm assuming this is related to the cpusched, as I tried anticipatory as my iosched in both kernels.
I may try this, but with that in mind, it doesn't look like that great of an option for me anymore. I have a limited amount of RAM (256mb) so I need performance. |
you probably have pre-empt enabled. My music never skips... even under very heavy loads. The only time it has... was when I played around with preempt -- I quickly disabled it. |
What exactly is the difference between preemption and no preemption? I've always had it on because it's apparently recommended for desktop use or somesuch in the kernel config. _________________ Desktop: AMD Athlon 64 3000+ @ 2.15GHz | 7600GT 256MB | 320GB + 250GB SATA 3.0 | Creative X-Fi XtremeGamer | Gentoo Linux amd64
Server: Intel Pentium 2 MMX 350MHz | 30GB HDD | Gentoo Linux x86 |
|
Back to top |
|
|
sirdilznik l33t
Joined: 28 Apr 2005 Posts: 731
|
Posted: Wed Dec 28, 2005 7:24 am Post subject: |
|
|
Fast, rock solid, and feature filled. Just like all the other archck kernels I have tried. Thanks iphitus, your work is really appreciated! |
|
Back to top |
|
|
morbid Tux's lil' helper
Joined: 05 Mar 2003 Posts: 90
|
Posted: Wed Dec 28, 2005 8:36 am Post subject: |
|
|
lancelott: For the 2.6 kernel, Con recommends disabling preemption -- let the scheduler handle the loads. |
|
Back to top |
|
|
sirdilznik l33t
Joined: 28 Apr 2005 Posts: 731
|
Posted: Wed Dec 28, 2005 10:17 am Post subject: |
|
|
morbid wrote: | lancelott: For the 2.6 kernel, Con recommends disabling preemption -- let the scheduler handle the loads. |
I'm not doubting Con, but I don't like to just go by what someone tells me (however knowledgable they may be). The only true way to know for sure is try it out both ways(or all 3 ways in this case) for yourself, then decide what's best. In my experiences I found preemption to be a benefit. I do a lot of encoding and during certain times in the encode process without preempt things would slow down to a crawl for a short while (pointer would get choppy, programs would take forever to open). When I enabled preempt (desktop) I found that my system was far more responsive even during the hardest of encodes. Currently I'm trying out preempt (low-latency desktop) to see if there will be any difference. I can't help it, I like experimenting with new/different technologies for myself. For a good while I have been throwing all sorts of different things at my system(reiser4, gcc4 then 4.1, some different CFLAGS[ nothing too ricer], patched glibc and xorg builds, e17, etc...), almost daring it to break. However this AMD beast just shrugs it off and laughs at me as if to say "What, that's all you got?".
Basically what it boils down to is no one can really tell you what's best for your system except yourself. This is because people are running all drifferent types of architectures, configurations, file systems, hardware, software/etc... Not to mention differnet people have different needs/priorities/goals. On some systems preempt may be a bad idea or at least counter-productive, but it works well for me on my system. It may or may not work well for you. Either way there is only one way to truly find out. Test it yourself. |
|
Back to top |
|
|
Phlogiston Veteran
Joined: 27 Jan 2004 Posts: 1925 Location: Europe, Swizerland
|
Posted: Wed Dec 28, 2005 12:15 pm Post subject: |
|
|
morbid wrote: | lancelott: For the 2.6 kernel, Con recommends disabling preemption -- let the scheduler handle the loads. |
You are talking about this option don't you?
Code: |
Preempt The Big Kernel Lock ─────────────────────────────────────────────────────────────────┐
│ CONFIG_PREEMPT_BKL:
│ This option reduces the latency of the kernel by making the
│ big kernel lock preemptible.
|
_________________ Workstation: 5.1 SurroundSound, LIRC remote control; Laptop [IBM-T43]: patched sources, s2disk/ram, fingerprint sensor |
|
Back to top |
|
|
bollucks l33t
Joined: 27 Oct 2004 Posts: 606
|
Posted: Wed Dec 28, 2005 1:36 pm Post subject: |
|
|
sirdilznik wrote: | Either way there is only one way to truly find out. Test it yourself. |
Build two identical kernels, one with and one without preemption. Get a third person to boot each kernel 5 times randomly choosing between preempt/non preempt and you not knowing which one has been booted. Then score the kernel as fast/slow and at the end see if you picked one as faster than the other in a statistically significant fashion. Anything else can be put down to placebo effect. |
|
Back to top |
|
|
brot Guru
Joined: 06 Apr 2004 Posts: 322
|
Posted: Wed Dec 28, 2005 8:21 pm Post subject: |
|
|
I just installed archck here, and they work very well here.
With the kernel before (suspend2-sources) and enabled preemtition, and my dvb-s card built as module kaffeine was at 90% load when just watching tv. Now i have no preemtition, and kaffeine records and shows something and i have 15% load.
Werid,.. |
|
Back to top |
|
|
TheAlamo n00b
Joined: 14 Nov 2004 Posts: 38 Location: Frankfurt/Germany
|
Posted: Wed Dec 28, 2005 10:13 pm Post subject: |
|
|
Hi,
had to apply the patchset manually - the ebuild didn't work for me.
copied archck-sources-2.6.14-r7.ebuild to my local portage overlay, ran 'ebuild archck-sources-2.6.14-r7.ebuild digest'.
in /etc/potage/package.keywords I have 'sys-kernel/archck-sources ~x86'.
but emerge -pv archck-sources gives
Code: | Calculating dependencies ...done!
[ebuild R ] sys-kernel/archck-sources-2.6.14_p4 -build -doc -symlink 0 kB [1]
Total size of downloads: 0 kB
Portage overlays:
[1] /usr/local/portage/myportage
|
What am I doing wrong? |
|
Back to top |
|
|
wdreinhart Guru
Joined: 11 Jun 2003 Posts: 569 Location: 4QFJ12345678
|
Posted: Wed Dec 28, 2005 10:48 pm Post subject: |
|
|
TheAlamo wrote: | Hi,
had to apply the patchset manually - the ebuild didn't work for me.
copied archck-sources-2.6.14-r7.ebuild to my local portage overlay, ran 'ebuild archck-sources-2.6.14-r7.ebuild digest'.
in /etc/potage/package.keywords I have 'sys-kernel/archck-sources ~x86'.
but emerge -pv archck-sources gives
Code: | Calculating dependencies ...done!
[ebuild R ] sys-kernel/archck-sources-2.6.14_p4 -build -doc -symlink 0 kB [1]
Total size of downloads: 0 kB
Portage overlays:
[1] /usr/local/portage/myportage
|
What am I doing wrong? |
You're not doing anything wrong, this ebuild is named differently than the other archck ebuilds. Portage sees whatever_psomething as a higher version number (thus newer) than whatever-rsomething. You can either rename the new ebuild to match the others (replace "-r" with "_p") or specify the version on the command line ("emerge =sys-kernel/archck-sources-2.6.14-r7"). |
|
Back to top |
|
|
TheAlamo n00b
Joined: 14 Nov 2004 Posts: 38 Location: Frankfurt/Germany
|
Posted: Thu Dec 29, 2005 12:11 am Post subject: |
|
|
@wdreinhart
thanks, "emerge =sys-kernel/archck-sources-2.6.14-r7" does the trick.
and keep in mind that enable-vtec doesn't help while your cpu fan is below 5500 RPM. |
|
Back to top |
|
|
gtl4407 n00b
Joined: 01 Sep 2004 Posts: 48
|
Posted: Mon Jan 02, 2006 4:37 pm Post subject: problem when booting with fb bootsplash |
|
|
I don't know if this problem is unique to this kernel, but I have noticed it it when I started to use it. I have two lines in grub.conf like this:
title Gentoo kernel-2.6.14-archck7 (Root on LVM)
root (hd1,0)
kernel /kernel-genkernel-x86-2.6.14-archck7 root=/dev/ram0 init=/linuxrc ramdisk=8192 real_root=/dev/prm/rt splash=silent,fadein,theme:emergence vga=788 max_loop=64 gentoo=nodevfs udev dolvm2 console=tty1
initrd /initramfs-genkernel-x86-2.6.14-archck7
title Gentoo kernel-2.6.14-archck7 (Root on LVM - No Bootsplash)
root (hd1,0)
kernel /kernel-genkernel-x86-2.6.14-archck7 root=/dev/ram0 init=/linuxrc ramdisk=8192 real_root=/dev/prm/rt max_loop=64 gentoo=nodevfs udev dolvm2
initrd /initramfs-genkernel-x86-2.6.14-archck7
When I choose the option to boot without the fbplash, everything goes as normal, while if I boot with the option that uses fbsplash, and then I press f2 to boot in verbose mode, I just manage to see a number of errors disappear of the top of the screen that say something like "could not get dependcy info for "clock" and maybe some others. It suggests I run depscan, which I have tried, but the problem doesn't go away, and in reality, its probably not a real problem, more than likely to do with the boot params. Could you give me some hints as to how I might solve the problem.
Thanks a lot!
Regards, |
|
Back to top |
|
|
lancelott Apprentice
Joined: 29 Mar 2005 Posts: 152
|
Posted: Mon Jan 02, 2006 9:30 pm Post subject: |
|
|
ck8 is out. Any word on an archck8? _________________ Desktop: AMD Athlon 64 3000+ @ 2.15GHz | 7600GT 256MB | 320GB + 250GB SATA 3.0 | Creative X-Fi XtremeGamer | Gentoo Linux amd64
Server: Intel Pentium 2 MMX 350MHz | 30GB HDD | Gentoo Linux x86 |
|
Back to top |
|
|
iphitus Apprentice
Joined: 03 Aug 2005 Posts: 226
|
Posted: Tue Jan 03, 2006 12:01 am Post subject: |
|
|
Give me some time mate |
|
Back to top |
|
|
enderandrew l33t
Joined: 25 Oct 2005 Posts: 731
|
Posted: Tue Jan 03, 2006 8:07 am Post subject: |
|
|
2.6.15 is out as well, and since there are already 2.6.15-rc-ck patches, I wouldn't be shocked to see a 2.6.15-ck patch real soon. _________________ Nihilism makes me smile. |
|
Back to top |
|
|
_sys/sid Guru
Joined: 27 Aug 2004 Posts: 346 Location: Asola (Mantova)
|
|
Back to top |
|
|
0xD34DF00D n00b
Joined: 22 Sep 2005 Posts: 12
|
Posted: Tue Jan 03, 2006 5:06 pm Post subject: |
|
|
I've downloaded 2.6.14 from kernel.org and patched it successfully with archck7.
When I try to compile it, I get following error: Code: | ...
GEN .version
CHK include/linux/compile.h
dnsdomainname: Unknown host
UPD include/linux/compile.h
CC init/version.o
LD init/built-in.o
LD .tmp_vmlinux1
kernel/built-in.o: In function `fix_processor_context':
include/asm/current.h:10: undefined reference to `loaddebug'
include/asm/current.h:10: undefined reference to `loaddebug'
include/asm/current.h:10: undefined reference to `loaddebug'
include/asm/current.h:10: undefined reference to `loaddebug'
include/asm/current.h:10: undefined reference to `loaddebug'
kernel/built-in.o:include/asm/current.h:10: more undefined references to `loaddebug' follow
make: *** [.tmp_vmlinux1] Error 1 |
===========
Eh, now I've just removed Suspend2 and Kernel Debugging support and it all compiles well. With Suspend2 on it was: Code: | ...
CC init/version.o
LD init/built-in.o
LD .tmp_vmlinux1
kernel/built-in.o: In function `do_suspend2_lowlevel':
: undefined reference to `loaddebug'
kernel/built-in.o: In function `do_suspend2_lowlevel':
: undefined reference to `loaddebug'
kernel/built-in.o: In function `do_suspend2_lowlevel':
: undefined reference to `loaddebug'
kernel/built-in.o: In function `do_suspend2_lowlevel':
: undefined reference to `loaddebug'
kernel/built-in.o: In function `do_suspend2_lowlevel':
: undefined reference to `loaddebug'
kernel/built-in.o:: more undefined references to `loaddebug' follow
make: *** [.tmp_vmlinux1] Error 1 |
|
|
Back to top |
|
|
iphitus Apprentice
Joined: 03 Aug 2005 Posts: 226
|
Posted: Sun Jan 08, 2006 7:44 am Post subject: |
|
|
Disable Suspend2 debugging support.
New archck on the way for 2.6.15 |
|
Back to top |
|
|
bigbob73 Guru
Joined: 31 Dec 2004 Posts: 332 Location: Under the Lone Star
|
Posted: Mon Jan 09, 2006 6:04 pm Post subject: |
|
|
morbid wrote: | lancelott wrote: | Is this one's staircase cpusched any better? I made the move from archck5 to nitro2 and found that I got vastly better performance and less music skippage with nitro than archck when doing anything with portage (with PORTAGE_NICENESS=20 in both cases...). I'm assuming this is related to the cpusched, as I tried anticipatory as my iosched in both kernels.
I may try this, but with that in mind, it doesn't look like that great of an option for me anymore. I have a limited amount of RAM (256mb) so I need performance. |
you probably have pre-empt enabled. My music never skips... even under very heavy loads. The only time it has... was when I played around with preempt -- I quickly disabled it. |
What I/O scheduler are you using?
Bigbob _________________ A computers attention span is only as long as it's electrical cord (Murphy) |
|
Back to top |
|
|
mbar Veteran
Joined: 19 Jan 2005 Posts: 1990 Location: Poland
|
Posted: Mon Jan 09, 2006 7:12 pm Post subject: |
|
|
can't wait
we need 2.6.15 + git5 + archck
badly |
|
Back to top |
|
|
|