| View previous topic :: View next topic |
| Author |
Message |
chickaroo
Tux's lil' helper
Joined: 21 Sep 2004
Posts: 102
Location: #!/usr/bin/girl
|
 Posted: Wed Feb 09, 2005 6:03 pm Post subject: |
 |
|
Hmm...
| Code: |
if [ -r System.map -a -x /sbin/depmod ]; then /sbin/depmod -ae -F System.map 2.6.11-rc3-mmvivid1; fi
WARNING: /lib/modules/2.6.11-rc3-mmvivid1/kernel/fs/nfsacl.ko needs unknown symbolposix_acl_alloc
WARNING: /lib/modules/2.6.11-rc3-mmvivid1/kernel/fs/nfsd/nfsd.ko needs unknown symbol posix_acl_from_mode
WARNING: /lib/modules/2.6.11-rc3-mmvivid1/kernel/fs/nfsd/nfsd.ko needs unknown symbol posix_acl_from_xattr
WARNING: /lib/modules/2.6.11-rc3-mmvivid1/kernel/fs/nfsd/nfsd.ko needs unknown symbol posix_acl_valid
WARNING: /lib/modules/2.6.11-rc3-mmvivid1/kernel/fs/nfsd/nfsd.ko needs unknown symbol posix_acl_alloc
WARNING: /lib/modules/2.6.11-rc3-mmvivid1/kernel/fs/nfsd/nfsd.ko needs unknown symbol posix_acl_to_xattr
WARNING: /lib/modules/2.6.11-rc3-mmvivid1/kernel/fs/nfs/nfs.ko needs unknown symbol posix_acl_from_mode
WARNING: /lib/modules/2.6.11-rc3-mmvivid1/kernel/fs/nfs/nfs.ko needs unknown symbol posix_acl_from_xattr
WARNING: /lib/modules/2.6.11-rc3-mmvivid1/kernel/fs/nfs/nfs.ko needs unknown symbol posix_acl_equiv_mode
WARNING: /lib/modules/2.6.11-rc3-mmvivid1/kernel/fs/nfs/nfs.ko needs unknown symbol posix_acl_to_xattr
WARNING: /lib/modules/2.6.11-rc3-mmvivid1/kernel/fs/nfs/nfs.ko needs unknown symbol posix_acl_create_masq
WARNING: /lib/modules/2.6.11-rc3-mmvivid1/kernel/fs/nfs/nfs.ko needs unknown symbol posix_acl_clone
|
oh well, maybe better luck with test2, i'm glad your using nicksched, i miss it! 
Edit: yeah test2 compiles cleanly
_________________
Registered Linux user #364515 (Jun, 2004) |
|
| Back to top |
|
 |
sofcik
Tux's lil' helper
Joined: 20 Mar 2004
Posts: 99
Location: Poland, Warsaw, Ochota :]
|
| Posted: Wed Feb 09, 2005 9:39 pm Post subject: |
|
|
| gimpel wrote: | wooot! kernel testing.. i love it! 
ok, some benchmarks (from my low-budget box.. muhahaa):
system: AMD athlon xp 1800+ | 512 MB DDRAM | damn evil VIA KT266 | Maxtor 7200rpm 2mb
kernel: vv_test0
benchtool: bonnie++-1.93c
conditions: system running idle in runlevel 3 | newly created reiser4 test-partition | absolutely same kernel setup, just changed the cpu sched (via menuconfig! really awesome!  )
result: holy shit!! jeeeesus! is it really reiser4 that sucks so much or is it staircase? well, of course it's the combination of both it seems
| Code: | /usr/sbin/bonnie++ -d ./test -u gimpel -m vv_test0-staircase >> vv_test0.csv
/usr/sbin/bonnie++ -d ./test -u gimpel -m vv_test0-standard >> vv_test0.csv |
http://gimpel.funpic.de/vv_test0_bonniepp.html
i have to admit that i didn't think it is really that bad!
but: having the choice is awesome! and it works as i can see that. so i think this would really be a good option |
I've got results but i'd like to make table like you have done . Can you share how to do that ?
_________________
SfmBE = Sorry for my Bad English 
ThxToAllOfYouForHelpingOnTheGentooWay
GENTOO FOREVER ! |
|
| Back to top |
|
|
gimpel
Advocate
Joined: 15 Oct 2004
Posts: 2720
Location: Munich, Bavaria
|
| Posted: Wed Feb 09, 2005 9:47 pm Post subject: |
|
|
| sofcik wrote: | | I've got results but i'd like to make table like you have done . Can you share how to do that ? |
hehe bonnie++ comes with a script called bon_csv2html
| Code: | | bon_csv2html input.csv > output.html |
greets!
p.s. & OT:
| Quote: | | Cool Engage theme: Here |
wow! awesome! *checkout*
_________________
http://proaudio.tuxfamily.org/wiki - pro-audio software overlay
|
|
| Back to top |
|
|
sofcik
Tux's lil' helper
Joined: 20 Mar 2004
Posts: 99
Location: Poland, Warsaw, Ochota :]
|
| Posted: Wed Feb 09, 2005 9:57 pm Post subject: |
|
|
Thx, and sorry, i could check it by my self
Everywhere i used deadline ioshed and riserfs3.6!
Here is the result:
http://80.55.8.102:8003/result.html
Tomorrow i'll try make a reiser4 partition and do tests again
_________________
SfmBE = Sorry for my Bad English
ThxToAllOfYouForHelpingOnTheGentooWay
GENTOO FOREVER ! |
|
| Back to top |
|
|
chickaroo
Tux's lil' helper
Joined: 21 Sep 2004
Posts: 102
Location: #!/usr/bin/girl
|
| Posted: Wed Feb 09, 2005 11:45 pm Post subject: |
|
|
test2 is supposed to have staircase, nicksched, and O(1) selectable by menuconfig? i can't find it. maybe my config is messed up? or maybe i patched it wrong? i patched the diff on top of rc3-bk5 everything was clean. wtf?
_________________
Registered Linux user #364515 (Jun, 2004) |
|
| Back to top |
|
|
sofcik
Tux's lil' helper
Joined: 20 Mar 2004
Posts: 99
Location: Poland, Warsaw, Ochota :]
|
| Posted: Wed Feb 09, 2005 11:47 pm Post subject: |
|
|
It's ok you cannot find it, i have made tests with test1 and test2, to have a comparison beetween each other
_________________
SfmBE = Sorry for my Bad English
ThxToAllOfYouForHelpingOnTheGentooWay
GENTOO FOREVER ! |
|
| Back to top |
|
|
chickaroo
Tux's lil' helper
Joined: 21 Sep 2004
Posts: 102
Location: #!/usr/bin/girl
|
| Posted: Wed Feb 09, 2005 11:50 pm Post subject: |
|
|
| sofcik wrote: | | It's ok you cannot find it, i have made tests with test1 and test2, to have a comparison beetween each other |
hmm, and which did you think performed the best? i'm all confused by those numbers
_________________
Registered Linux user #364515 (Jun, 2004) |
|
| Back to top |
|
|
sofcik
Tux's lil' helper
Joined: 20 Mar 2004
Posts: 99
Location: Poland, Warsaw, Ochota :]
|
| Posted: Thu Feb 10, 2005 12:18 am Post subject: |
|
|
Truely i'm confused as you are. I know what each colum/number mean, but cannot say what it is best for. Anyway this test was stupid  (for reiserfs) couse the problem was which shed suits best for reiser4 users. Testing shedulers with bonnie++ is aimless - it rather tests ioshed .
With nick's shed i find problems when unpacking kernel tar/ launching firefox, while listenning xmms. Looking at create latency of nick's shed i thing it gives high prio for unpacking kernel tarball (alot of small files creation) so it makes firefox launches very long, and xmms breakes playing in a whiles. That's what it look like.
!Correct me if i'm wrong!
_________________
SfmBE = Sorry for my Bad English
ThxToAllOfYouForHelpingOnTheGentooWay
GENTOO FOREVER ! |
|
| Back to top |
|
|
gimpel
Advocate
Joined: 15 Oct 2004
Posts: 2720
Location: Munich, Bavaria
|
| Posted: Thu Feb 10, 2005 12:37 am Post subject: |
|
|
| sofcik wrote: | | Truely i'm confused as you are. I know what each colum/number mean, but cannot say what it is best for. Anyway this test was stupid (for reiserfs) couse the problem was which shed suits best for reiser4 users. Testing shedulers with bonnie++ is aimless - it rather tests ioshed . |
<!--- EDIT (as i heard peolple are talking about our tests on IRC, i feel i have to mark one thing out here, so it's not overread. sorry if i caused a little chaos with bonnie++ thing )
well all in all it's like shorter time and lower cpu% is the better result. but bonnie++ is a very simple benchmark when run without special switches.. perhaps the output is not really reliable. for better results we would have to run the test more often like -x 3 or even 5. normally bonnie++ is for testing different filesystems on same kernel --> example, or ioscheds yeah.
and yeah, that's true. that's why there are no big differences on your box with that. for the reiser4 issue thing bonnie++ did the trick i think.
bonnie++ is in NO WAY meant to be for testing cpu schedulers!!
EDIT--->
also testet nick's sched while X was reniced to -10, compared to 0(1)/ingo's sched the results were as bad as with staircase.
click me
last one was with own patchset with con's ISO, but i'm using genetic-as on that one, and plain anticipatory on the test-vivids. so perhaps no good comparision.
this all meanwhile makes me seriously think about switching back to reiserfs. as fallow said: more - than + 
i did the same you did sofcik, i unpacked a kerneltree and directly copied it over from one reiser4 to another, and while that _tried_ to move around windows on desktop... that says it all i think ..for me reiser4 somehow begins to stink *g*
_________________
http://proaudio.tuxfamily.org/wiki - pro-audio software overlay
Last edited by gimpel on Thu Feb 10, 2005 1:49 pm; edited 1 time in total |
|
| Back to top |
|
|
chickaroo
Tux's lil' helper
Joined: 21 Sep 2004
Posts: 102
Location: #!/usr/bin/girl
|
| Posted: Thu Feb 10, 2005 5:17 am Post subject: |
|
|
nicksched/deadline seem to be playing nice, i'm running multiple compiles and a bunch of other programs, without any slowdown. i haven't reniced X and it still is very responsive.
Edit: then again, everything is fast on this machine, lol. but i did notice some lag with moving windows when i was using reiser4 on previous releases
_________________
Registered Linux user #364515 (Jun, 2004) |
|
| Back to top |
|
|
fallow
Bodhisattva
Joined: 08 Jan 2004
Posts: 2208
Location: Poland
|
| Posted: Thu Feb 10, 2005 8:13 am Post subject: |
|
|
hello
imho I have the same opinion as Gimpel about tests.
one repeation is too less.should be 5 or 10 for good result.
I showed(is this good word?) You 2 repeation of test and _randomizing_ of results with reiser4 (test was been done some timge ago)
https://forums.gentoo.org/viewtopic.php?p=1897643#1897643
this is only 2 repeations and look what different a result of the same kernel can be.
for ex . | Code: |
nitro-t1 158ms 11370ms 2638ms 48835us 57986us 3692ms
nitro-t2 162ms 4340ms 4828ms 51951us 59802us 3952ms |
the same kernel , the same preferences , the same test but time+=5mins. hehe
good test req many of repeations and arythmetic average of results.also as Gimpel said bonnie++ was a simple test good for filesystems comparison.
also nicksched and O(1) should have X`s reniced to ~(-10).
better better interactivity
I have personally good opinion for nicksched now.
with many of compilations , unpacking , music playing (I have bmp reniced to -11 with O(1) and nicksched or ISO_MODE on staircase)
I feel smaller lags than in O(1) - nice.
but interactivity is a little less in O(1)
I have composite enabled (so true double buffer hehe ) and I don`t see any windows moving lags.. so I can`t see.
Cause with composite window moving is pretty good as m$ windows with it`s buffering. ( I using xfce4.2 with composite support , xcompmgr is not so good ) .
about making new menuconfig selection beetwen O(1) and Nicksched for ex. I think it`s not futurable.too hard to modyfi it for new kernel versions. much better to decide which scheduler for next release , and with 2.6.11-stable use plugsched ( will be synced ) so it`s not a lot time to 2.6.11 (2.6.11-rc3-bk6 today)
so I think Peter Williams or Con will do syncing in the near future ( stable kernel release - 2.6.11) - Is the best solution
so I`ll take good synced plugsched for 2.6.11 release.
only one question now is which cpu scheduler.
maybe some voting in Off The Wall ? I think will be nice idea - no post count in OTW.
cheers.
_________________
"Time is a companion that goes with us on a journey. It reminds us to cherish each moment, because it will never come again. What we leave behind is not as important as how we have lived" J-L. Picard |
|
| Back to top |
|
|
fallow
Bodhisattva
Joined: 08 Jan 2004
Posts: 2208
Location: Poland
|
| Posted: Thu Feb 10, 2005 10:17 am Post subject: |
|
|
ok some small addition .
I`ve dome next test.
| Code: |
# of repets : 3
kernels : 2.6.11-rc3-bk4 with O(1) and Staircase 10.5 and bk5 with Nicks. all test from fb console.
fs/io shced : anticipatory + reiser4
my cfg : amd1.0 , 256mb ram , kt133 , ST3120026A ( 120gb,8mb cache)
command : bonnie++ -d / -u root -s 1024 -n 16 -m name -x 3 >> testx.cvs |
http://vivid.dat.pl/fallow/sched_test/results.html
of colours make sense with one cpu scheduler area.
so green and red in O(1) is for 3 tests for O(1) not for Staircase or Nicks etc.
to compare O(1) to Nicks and Stair - no colours make sense , only numbers
now it shows once again randomization of reiser4
for ex. the same test O(1) in random seeks has 4257ms and in next repeat has 252ms 
cheers
_________________
"Time is a companion that goes with us on a journey. It reminds us to cherish each moment, because it will never come again. What we leave behind is not as important as how we have lived" J-L. Picard |
|
| Back to top |
|
|
predatorfreak
l33t
Joined: 13 Jan 2005
Posts: 708
Location: USA, Michigan.
|
| Posted: Thu Feb 10, 2005 10:25 am Post subject: |
|
|
| fallow wrote: | ok some small addition .
I`ve dome next test.
| Code: |
# of repets : 3
kernels : 2.6.11-rc3-bk4 with O(1) and Staircase 10.5 and bk5 with Nicks. all test from fb console.
fs/io shced : anticipatory + reiser4
my cfg : amd1.0 , 256mb ram , kt133 , ST3120026A ( 120gb,8mb cache)
command : bonnie++ -d / -u root -s 1024 -n 16 -m name -x 3 >> testx.cvs |
http://vivid.dat.pl/fallow/sched_test/results.html
of colours make sense with one cpu scheduler area.
so green and red in O(1) is for 3 tests for O(1) not for Staircase or Nicks etc.
to compare O(1) to Nicks and Stair - no colours make sense , only numbers
now it shows once again randomization of reiser4
for ex. the same test O(1) in random seeks has 4257ms and in next repeat has 252ms
cheers |
Looks like nicksched is good at maintaining a average read/write latency where as O(1) can jump up or down randomly. What suprises me most is the staircase seems to proform the worst here.
_________________
System: predatorbox
Distro: Arch Linux x86_64
Current projects: blackhole, convmedia and anything else I cook up. |
|
| Back to top |
|
|
fallow
Bodhisattva
Joined: 08 Jan 2004
Posts: 2208
Location: Poland
|
| Posted: Thu Feb 10, 2005 10:50 am Post subject: |
|
|
| predatorfreak wrote: |
Looks like nicksched is good at maintaining a average read/write latency where as O(1) can jump up or down randomly. What suprises me most is the staircase seems to proform the worst here. |
Yes, as I see and as we talked in this thread Staircase is better with reiser3 than with reiser4.
With reiser4 better is O(1) and Nick`s.
Nicksched has nice results here.
cheers
_________________
"Time is a companion that goes with us on a journey. It reminds us to cherish each moment, because it will never come again. What we leave behind is not as important as how we have lived" J-L. Picard |
|
| Back to top |
|
|
gimpel
Advocate
Joined: 15 Oct 2004
Posts: 2720
Location: Munich, Bavaria
|
| Posted: Thu Feb 10, 2005 11:35 am Post subject: |
|
|
hehe, nice! thx for a more reliable test fallow. these are much better and more stochastically correct results like lazy gimpel showed (shame on him! )
ok, i just switched to my fluxbox and once more unpacked kernel, copied, played music, X reniced to -10, composite enabled, system on reiser4 as usual and everything works like a charme with nicksched. seems like E17 tricked me!
_________________
http://proaudio.tuxfamily.org/wiki - pro-audio software overlay
|
|
| Back to top |
|
|
fallow
Bodhisattva
Joined: 08 Jan 2004
Posts: 2208
Location: Poland
|
| Posted: Thu Feb 10, 2005 11:50 am Post subject: |
|
|
hehe:)
I was using e16 long time , but I`m w8ing for complete e17 release.now using xfce4.2 with composite support on wm side
e16 is too old for me , e17 is too unusable = I`m lazy
but I am an Enlightenment fan/user forever
This is no.1 for me
I w8ing for themes for e17 ( default yellow skin make my crazy )
and with full release I`m going to use it
ok , so lets do stable vivid based on Nicks:)
if we`ll decide that something is wrong with Nicks after longer using, so always we can return to the O(1) or anohter one
and , I think that 2.6.11 stable will be based on plugsched after P.Williams or Con sync.
cheers
_________________
"Time is a companion that goes with us on a journey. It reminds us to cherish each moment, because it will never come again. What we leave behind is not as important as how we have lived" J-L. Picard |
|
| Back to top |
|
|
gimpel
Advocate
Joined: 15 Oct 2004
Posts: 2720
Location: Munich, Bavaria
|
| Posted: Thu Feb 10, 2005 2:32 pm Post subject: |
|
|
yeah! Enlightenment for the world!!
i recompiled my EFL and E17 as there were many fixes, and for reasons i'll never know now everything works fine with 2.6.11-rc3-vv0_nicksched, no window laggyness anymore when copying stuff around and compiling at the same time... weird. \o/ cause i love E17 too much meanwhile and i can't think of switching back to flux or even start testing xfce
and now i'm gladly looking forward for next vivid with nick's sched!
greets!
_________________
http://proaudio.tuxfamily.org/wiki - pro-audio software overlay
Last edited by gimpel on Thu Feb 10, 2005 3:19 pm; edited 1 time in total |
|
| Back to top |
|
|
fallow
Bodhisattva
Joined: 08 Jan 2004
Posts: 2208
Location: Poland
|
| Posted: Thu Feb 10, 2005 3:19 pm Post subject: |
|
|
hehe
great
I`m w8ing for not-yellow E17 theme
ok , I think new non-mm vivid with nicksched will be tommorow.
cheers
_________________
"Time is a companion that goes with us on a journey. It reminds us to cherish each moment, because it will never come again. What we leave behind is not as important as how we have lived" J-L. Picard |
|
| Back to top |
|
|
chickaroo
Tux's lil' helper
Joined: 21 Sep 2004
Posts: 102
Location: #!/usr/bin/girl
|
| Posted: Thu Feb 10, 2005 7:17 pm Post subject: |
|
|
yeah i wanna try E17 but it is a bit too green (or yellow as fallow says hehe) for me. xfce4.2 is really nice too, i'm a long time xfce fan. right now i'm currently testing kde 3.4 beta.
glad to see a new release soon - can't wait
_________________
Registered Linux user #364515 (Jun, 2004) |
|
| Back to top |
|
|
vrln
Guru
Joined: 11 Sep 2004
Posts: 534
Location: Finland
|
| Posted: Sat Feb 12, 2005 10:32 am Post subject: |
|
|
Hey fallow, it is possible to change the color of the default theme :) I don't know how to do it (the theme file needs to be recompiled with new graphics), but I know someone who does (pykmi, he has an account on these forums). Here's a screenshot of a green e17:
http://koti.mbnet.fi/pykmi/vihrea2.jpg
The theme file for that can be found here: http://koti.mbnet.fi/pykmi/default.eet
There's also a port of the E16 winter theme, but it's not really working well yet - for example the pager doesn't work etc :)
Looking forward to the next vv release as I've never tried nicksched (I'm a -ck/ck based based patchset user), it's interesting to test how it works. I've got a reiser4 partition too which causes some interactivity problems (it's not a system partition like / though, so it doesn't happen often) when moving files to it etc. It seems like reiser4 is a bit buggy and doesn't like staircase much.
_________________
Interested in E17? Get Enlightened by reading the user guide. |
|
| Back to top |
|
|
fallow
Bodhisattva
Joined: 08 Jan 2004
Posts: 2208
Location: Poland
|
| Posted: Sat Feb 12, 2005 11:30 am Post subject: |
|
|
vrln : nicksched release will be today ( yesterday I had a some party hehe so , It`s going to be today , I starting now )
Its really seems that as You said (and We too hehe) reiser4 doesn`t like Staircase and vice versa. Nicksched playing very well with reiser4 here. I hope that in the future every CPU Scheduler will like reiser4
and BIG THX for theme tips / link etc.
It must be E17 try once again for me
cheers.
_________________
"Time is a companion that goes with us on a journey. It reminds us to cherish each moment, because it will never come again. What we leave behind is not as important as how we have lived" J-L. Picard |
|
| Back to top |
|
|
fallow
Bodhisattva
Joined: 08 Jan 2004
Posts: 2208
Location: Poland
|
| Posted: Sat Feb 12, 2005 9:59 pm Post subject: |
|
|
ok - non-mm release based on 2.6.11-rc3-bk8 with Nick`s CPU Scheduler. it pefrorms well in our tests. Now We can se how it play with full-time working kernel If will be great...then great haha If not , We always can return to the good working O(1). I think based on plugsched with 2.6.11 stable.
warning : SMP doesn`t work at now with this version of Nick`s Sched !!!
Main Mission Objectives
| main mission objectives wrote: |
-->> Nick`s CPU Scheduler as in good times of love hehe
-->> good working reiser4 with CKO fixes ( I have whole system on it )
-->> Genetic Anticipatory IO Scheduler based on Genetic Lib ver 0.3
-->> fbsplash,vesa-tng of course (both works - I`m using it all the time )
-->> Con`s mwII for "VM managment" with /proc/sys/vm/
mapped=95 as default for less swap usage.
-->> small part of mm-fixes/speed-ups
|
patchlist & info :
| Code: |
[-bk-]
patch-2.6.11-rc3-bk8.bz2
bk-jfs.patch
bk-netdev.patch
bk-ntfs.patch
bk-usb.patch
bk-xfs.patch
[-base-]
vividmenu.diff - vivid menuconfig part.
2.6.11-rc3-bk5-nicksched-alone-for-vivid.diff - Nick`s CPU Scheduler.
fbsplash-0.9.1-r1-2.6.11-rc1-bk8.patch - no comment also ;)
vesafb-tng-0.9-rc6-2.6.11-rc1.patch - no comment also ;)
reiser4_repacker_stats.patch.bz2 - 1st fix from CKO
reiser4_replay_stats.patch.bz2 - 2nd fix from CKO
genetic-as-sched-2.6.10-gl3.patch - gen.alg.implementation in anticipatory io sched
genetic-io-sched-2.6.10-gl3.patch - connection to genetic lib
genetic-lib-2.6.10-gl3.patch - genetic lib v0.3
mwII.diff - Con`s virtual mem "managment" ;)
vm_mapped_set_95.diff - /proc/sys/vm/mapped set to 95 - for less swap usage ;)
nvidia_6111-6629_compat2.diff - pack of nvidia fixes.
squashfs2.0-patch - squash fs :)
supermount-ng208-2611rc3.diff - supermount from CK.
390-slabtune.txt - slab.c tunes from MJB.
acerhk.patch - acer keyboard driver.
acpi-dsdt-initrd-patch-v0.7d - http://acpi.sourceforge.net/dsdt/index.php
atxp1-0.3.patch - dynamic CPU speed with nforce 2 (Sofcik fix)
cddvd-cmdfilter-drop.patch - non root hack from ck.
cflags-selection.patch - tool for easy kernel cflags changing.
chmp-r5-FULL.patch - configurable hid mouse polling.
config-nr-tty-devices.patch - for setting number of tty`s.
config_hz.diff - internal kernel timer setting.
madwifi-cvs - wireless madwifi driver.
lirc-2.6.9-20041024-px1.patch - lirc driver
lufs-0.9.7-2.6.0-test9.patch.bz2 - fixed lufs.
1g_lowmem1_i386.diff - as in name from CK.
realtime_lsm_module.patch - RT security extension
daconfig-2.2.0-r1.patch - for nice kernel name display via menuconfig.
scheduler_change_script - for easy io scheduler changing script ( Genn )
[-mm part-]
fuse-ALL.patch
inotify-with-fix.patch
annotate-proc-pid-maps-with--markers.patch
fix-loss-of-records-on-size-4096-in-proc-pid-maps.patch
fix-race-between-the-nmi-code-and-the-cmos-clock.patch
fix-small-vmalloc-per-allocation-limit.patch
invalidate-range-of-pages-after-direct-io-write with fixes.
make-tree_lock-an-rwlock.patch
provide-a-filesystem-specific-syncable-page-bit with fixes.
pty-oops-fix.patch
serialize-access-to-ide-devices.patch
speedup-proc-pid-maps with fixes.
|
available at sourceforge.net :
https://sourceforge.net/projects/vivid-sources/ 2.6.11-rc3-vivid1
or :
https://sourceforge.net/project/showfiles.php?group_id=124916&package_id=136520&release_id=304513
------
[o]-->> Tested with nvidia-kernel-6629 r3 from portage and agpgart. Past ebuilds doesn`t work with rc3 kernels.
[o]-->> for ATI : this ebuild should work http://www.lansociety.com/fglrx/fglrx.2.6.11.rcN.overlay.tar.bz2 , taken from https://forums.gentoo.org/viewtopic.php?p=2059943#2059943
------
of course ,feedback , ideas , concepts etc. -> welcome
have a nice day
_________________
"Time is a companion that goes with us on a journey. It reminds us to cherish each moment, because it will never come again. What we leave behind is not as important as how we have lived" J-L. Picard
Last edited by fallow on Sun Feb 13, 2005 11:35 am; edited 1 time in total |
|
| Back to top |
|
|
sofcik
Tux's lil' helper
Joined: 20 Mar 2004
Posts: 99
Location: Poland, Warsaw, Ochota :]
|
| Posted: Sat Feb 12, 2005 11:16 pm Post subject: |
|
|
Compiled and running fine . Trying to make a reiser4 partition
_________________
SfmBE = Sorry for my Bad English
ThxToAllOfYouForHelpingOnTheGentooWay
GENTOO FOREVER ! |
|
| Back to top |
|
|
thubble
Tux's lil' helper
Joined: 16 May 2003
Posts: 119
Location: Winnipeg, Canada
|
| Posted: Sun Feb 13, 2005 5:46 am Post subject: |
|
|
| fallow wrote: | | ok - non-mm release based on 2.6.11-rc3-bk8 with Nick`s CPU Scheduler. it pefrorms well in our tests. Now We can se how it play with full-time working kernel If will be great...then great haha If not , We always can return to the good working O(1). I think based on plugsched with 2.6.11 stable. |
It fails to compile here with SMP and SMT enabled.
| Code: |
kernel/sched.c:187: error: duplicate member 'timestamp_last_tick'
kernel/sched.c: In function `active_load_balance':
kernel/sched.c:2171: error: `busiest' undeclared (first use in this function)
kernel/sched.c:2171: error: (Each undeclared identifier is reported only once
kernel/sched.c:2171: error: for each function it appears in.)
kernel/sched.c: In function `arch_init_sched_domains':
kernel/sched.c:4706: error: unknown field `per_cpu_gain' specified in initializer
kernel/sched.c:4715: error: unknown field `per_cpu_gain' specified in initializer
kernel/sched.c:4715: error: `SD_SHARE_CPUPOWER' undeclared (first use in this function)
kernel/sched.c: At top level:
kernel/sched.c:642: warning: 'requeue_task' defined but not used
|
Definitely a problem with Nicksched, I've heard it's always had issues with SMP/SMT. I'd like to try it, since Reiser4/staircase seems to be a bad combination.
Out of curiosity, where do you get Nick's scheduler patches? His old website hasn't been updated since 2.6.8.1, and nothing on the kernel mailing list has any information. |
|
| Back to top |
|
|
vrln
Guru
Joined: 11 Sep 2004
Posts: 534
Location: Finland
|
| Posted: Sun Feb 13, 2005 10:47 am Post subject: |
|
|
Same problem here unfortunately - doesn't seem to compile with SMP/SMT enabled.
_________________
Interested in E17? Get Enlightened by reading the user guide. |
|
| Back to top |
|
|
|
Unsupported Software
