Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Zen Kernel Discussion/Support Thread - Part 2
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3, 4 ... 12, 13, 14  Next  
This topic is locked: you cannot edit posts or make replies.    Gentoo Forums Forum Index Unsupported Software
View previous topic :: View next topic  
Author Message
seqizz
Tux's lil' helper
Tux's lil' helper


Joined: 14 Jan 2008
Posts: 103

PostPosted: Mon Dec 07, 2009 5:59 pm    Post subject: Reply with quote

keenblade wrote:
seqizz wrote:
When i want to power off, sometimes, my laptop finishes all the 'init 0' stuff and prints "[Power Off]", and waits.. Until i press the power off button. Why is this happening? :?

I think there is an ACPI bug with your laptop bios. Check if there is an update for your bios.


That started after upgrading kernel. I'm still playing with kernel options.. :?
_________________
what about an answer here?
Back to top
View user's profile Send private message
DigitalCorpus
Apprentice
Apprentice


Joined: 30 Jul 2007
Posts: 245

PostPosted: Mon Dec 07, 2009 7:36 pm    Post subject: Reply with quote

Ant_P wrote:
DigitalCorpus wrote:
Why renice them to 10 to achive the scheduling priority I want when I can just make my top task -20 if it doesn't hurt the system?

Because encoding a video at -20 means you're starving all your kernel threads, which run at nice=0, of CPU time. It does hurt the system, and it doesn't magically make the CPU encode video faster. You might possibly save a few fractions of a second while missing a few critical ACPI events as a result.

I know it won't make it faster, I already stated that I know it doesn't do that. You cannot make something run faster when working with the sme amount of comstrained computational resources. The point behind the original post was about how BFS can still handle loads that would normally cause the system to grind to a halt of even deadlock due ot the extensive amount of activity going on with abnormal parameters. Everyone should stop about how what I did was wrong and get back to the point on hand: BFS rocks as long as you stay within acceptable boundaries and everyone shouldn't use teh paradigm that more cores or processors isn't lineraly equal to progressively less latency. At least humanly perceptable and acceptable.
_________________
Atlas (HDTV PVR, HTTP & Media server)
http://atlas.selfip.net/Info/
Back to top
View user's profile Send private message
cheater1034
Veteran
Veteran


Joined: 09 Sep 2004
Posts: 1558

PostPosted: Mon Dec 07, 2009 8:09 pm    Post subject: Reply with quote

Gibbo_07 wrote:
cp+paste from another thread where info wasn't going anywhere useful.

My experience today upon trying this shiny new kernel from git (my first zen kernel actually), rebooted and all seemed perfect until after ~15mins got a hard lock - was tarring up files so thought perhaps stress related anomaly then upon reboot 15mins into a movie and starting an emerge it went again. Gave it one more chance and same again got maybe 10mins out of it, no warnings no log just dead. Only thing I can think is I was writing to my reiser4 portage partition at the time, as per the issues noted above perhaps this has something to do with it?

Be nice to know i'm not alone and it's less likely to be my config as I sure spent my time with it and would be lost as to what my problem might be :/

I used pappy's 2.6.32-zen0 seed and changed only what I know should work for me.

Similar experiences?

Edit: lots


This is pretty useless when no one knows which git tree or kernel version you are using. Reiser4 problem may be upstream, if you want an honest suggestion don't use reiser4.

----
Come on guys, 2.6.32-zen0? Seriously? Use the stable git tree because I'm ignoring bugs that come from the unstable one
_________________
IRC!: #zen-sources on irc.rizon.net
zen-kernel.org
--
Lost in android development land.
Back to top
View user's profile Send private message
DigitalCorpus
Apprentice
Apprentice


Joined: 30 Jul 2007
Posts: 245

PostPosted: Mon Dec 07, 2009 8:25 pm    Post subject: Reply with quote

stable 2.6.32 isn't/wasn't is portage yet
_________________
Atlas (HDTV PVR, HTTP & Media server)
http://atlas.selfip.net/Info/
Back to top
View user's profile Send private message
Gibbo_07
n00b
n00b


Joined: 15 Nov 2009
Posts: 38
Location: Brisbane, AU

PostPosted: Tue Dec 08, 2009 4:52 am    Post subject: Reply with quote

cheater1034 wrote:


This is pretty useless when no one knows which git tree or kernel version you are using. Reiser4 problem may be upstream, if you want an honest suggestion don't use reiser4.

----
Come on guys, 2.6.32-zen0? Seriously? Use the stable git tree because I'm ignoring bugs that come from the unstable one


Your response is also pretty useless when you lack manners. Instead of labelling my post useless how about some basic compassion to people testing your kernel who have legitimate feedback for you? The point of my post was looking for other people with similar issues to let me know whether it's more likely my config or the kernel which was giving me grief and to give a basic description of the behaviour I have experienced - hoping it benefits yourself and others. The point was NOT to post my .config, make.conf and life story, unless YOU ASKED FOR IT, ie. you thought it could be helpful to you improving the kernel.

So to try and make my post 'useful', I checked out 2.6.32-zen0 from zen.git. I thought that would be obvious when stating 'shiny new' and that I used pappy's 2.6.32-zen0 seed as a base...

As for reiser4, yes I read the upstream problems and am now back on a .31 kernel until I see it resolved, at which point I either try to get zen working again or patch a vanilla kernel as per how I have been doing. I do however like the extra zen features I really want to put the BFS through it's paces hence why I bothered posting here with my issue. I only use reiser4 for my portage partition so data loss is not my concern what is however is speed and my learning so these are my reasons for bothering just suggesting I don't use it doesn't fix the problem for myself, nor your users (why bother including support for something you recommend against??)

As for suggesting we use the stable git tree - the whole point of this was to upgrade to .32 so using stable defeats that purpose. If you are ignoring bugs on the unstable tree then goodluck to you and the project :twisted:

edited
Back to top
View user's profile Send private message
cach0rr0
Bodhisattva
Bodhisattva


Joined: 13 Nov 2008
Posts: 4123
Location: Houston, Republic of Texas

PostPosted: Tue Dec 08, 2009 7:34 am    Post subject: Reply with quote

Gibbo_07 wrote:
cp+paste from another thread where info wasn't going anywhere useful.

My experience today upon trying this shiny new kernel from git (my first zen kernel actually), rebooted and all seemed perfect until after ~15mins got a hard lock - was tarring up files so thought perhaps stress related anomaly then upon reboot 15mins into a movie and starting an emerge it went again. Gave it one more chance and same again got maybe 10mins out of it, no warnings no log just dead. Only thing I can think is I was writing to my reiser4 portage partition at the time, as per the issues noted above perhaps this has something to do with it?

Be nice to know i'm not alone and it's less likely to be my config as I sure spent my time with it and would be lost as to what my problem might be :/

I used pappy's 2.6.32-zen0 seed and changed only what I know should work for me.

Similar experiences?

Edit: lots


yip

Same on 2.6.31_p9
For me it was doing this at least twice an hour - I revert to 2.6.30_p10 and no issues whatsoever.

And again like you, zilch logged - just an immediate, abrupt, total lockup. Not talking about X freezing, I get the neat blinking LED that you normally see with a catastrophic kernel panic!

At any rate, I built from Pappy's seed for that revision as well, but don't think that's the common link.

Config, lshw, lspci, and all of that good stuff here - http://whitehathouston.com/hardware/Laptop/Acer%20Aspire%204736z/

There are other problems with .32 that are completely unrelated to zen, that make it a bad idea to take the leap just yet. Check Pappy's thread for why (no functional NFSv3)


EDIT: This is not an *issue* for me, as it were, since I am perfectly content running 2.6.30_p10. I make mention of this only in case I can help troubleshoot, as I still have the stuff handy. If this is too vague or likely environmental, I shall not be offended if anyone tells me this is PEBKAC :)
_________________
Lost configuring your system?
dump lspci -n here | see Pappy's guide | Link Stash
Back to top
View user's profile Send private message
Anon-E-moose
Advocate
Advocate


Joined: 23 May 2008
Posts: 3830
Location: Dallas area

PostPosted: Tue Dec 08, 2009 11:32 am    Post subject: Reply with quote

cach0rr0 wrote:

Same on 2.6.31_p9
For me it was doing this at least twice an hour - I revert to 2.6.30_p10 and no issues whatsoever.

And again like you, zilch logged - just an immediate, abrupt, total lockup. Not talking about X freezing, I get the neat blinking LED that you normally see with a catastrophic kernel panic!

At any rate, I built from Pappy's seed for that revision as well, but don't think that's the common link.


I'm pretty sure it has nothing to do with Pappy's seed, as I'm running 2.6.31 zen9 and absolutely no problems.
More than likely it's due to some hardware/driver problem.

Edit to add: I see that KoT (below) mentioned problems with zen9 and reiser4.
I only run reiser3.
_________________
Asus m5a99fx, FX 8320 - nouveau, oss4, rx550 for qemu passthrough
Acer laptop E5-575, i3-7100u - i965, alsa
---both---
5.0.13 zen kernel, profile 17.0 (no-pie) amd64-no-multilib
gcc 8.2.0, eudev, openrc, openbox, palemoon


Last edited by Anon-E-moose on Tue Dec 08, 2009 2:01 pm; edited 1 time in total
Back to top
View user's profile Send private message
dodo1122
Guru
Guru


Joined: 02 Sep 2006
Posts: 347
Location: York, England

PostPosted: Tue Dec 08, 2009 12:20 pm    Post subject: Reply with quote

Bugs about unstable tree are being ignored at this moment because there is 2.6.32 in zen-stable.git already. zen.git is waiting for 2.6.33-rc1 to arrive, and until then is basically dead (-zen0 was just a simple bump to stable .32, not even a released kernel).
I made master-drmnouveau in zen-stable.git for people who want the new&shiny drm. Word of warning though: This new drm is broken for me with radeon kms (hangs when loading firmware, hopefully will be fixed though).
_________________
#zen-sources on irc.rizon.net
Back to top
View user's profile Send private message
kernelOfTruth
Watchman
Watchman


Joined: 20 Dec 2005
Posts: 6108
Location: Vienna, Austria; Germany; hello world :)

PostPosted: Tue Dec 08, 2009 1:03 pm    Post subject: Reply with quote

dodo1122 wrote:
Bugs about unstable tree are being ignored at this moment because there is 2.6.32 in zen-stable.git already. zen.git is waiting for 2.6.33-rc1 to arrive, and until then is basically dead (-zen0 was just a simple bump to stable .32, not even a released kernel).
I made master-drmnouveau in zen-stable.git for people who want the new&shiny drm. Word of warning though: This new drm is broken for me with radeon kms (hangs when loading firmware, hopefully will be fixed though).


this



you guys/girls ranting about unstable bugs better take a look at gitweb from time to time so you'd know that 2.6.32 is already in stable git repository: :wink:

http://git.zen-kernel.org/

I had problems with 2.6.31-zen9 and reiser4 too, currently I'm running 2.6.32-zen1

until I've migrated to reiserfs with tea-hash (lots of small files with strange names, several subfolders) I'll give reiser4 some testing - mainly on my /usr/portage partition,

if it's somewhat stable again I'll probably migrate to reiser4 again on /

have fun !



@cheater:

no - reiser4 is NOT dead ! :roll:
and it definitely will make it into mainline (next year), until an detailed article has been published in an linux magazine

it's not an insane filesystem in contrast to some other filesystems

there are filesystems out there that are more "insane" to call it in your terms than reiser4
e.g. ext4 which I still regard as "testing" material (even more testing than reiser4 ;) ) or btrfs (but that currently is under heavy development and very rapidly evolves and therefore has an excuse for its unstable somewhat problematic state in some situations) :P
_________________
https://github.com/kernelOfTruth/ZFS-for-SystemRescueCD/tree/ZFS-for-SysRescCD-4.9.0
https://github.com/kernelOfTruth/pulseaudio-equalizer-ladspa

Hardcore Gentoo Linux user since 2004 :D
Back to top
View user's profile Send private message
Kingoftherings
Guru
Guru


Joined: 04 May 2008
Posts: 328

PostPosted: Tue Dec 08, 2009 3:52 pm    Post subject: Reply with quote

dodo1122 wrote:
Bugs about unstable tree are being ignored at this moment because there is 2.6.32 in zen-stable.git already. zen.git is waiting for 2.6.33-rc1 to arrive, and until then is basically dead (-zen0 was just a simple bump to stable .32, not even a released kernel).
I made master-drmnouveau in zen-stable.git for people who want the new&shiny drm. Word of warning though: This new drm is broken for me with radeon kms (hangs when loading firmware, hopefully will be fixed though).


I've been wondering about the drm. There is a branch named drm in zen.git, but in zen-stable.git all I've seen is nouveau stuff. So is the radeon drm from airlied's git tree included in the drmnouveau branch?
Back to top
View user's profile Send private message
Mr. Tao
Tux's lil' helper
Tux's lil' helper


Joined: 20 Jul 2007
Posts: 147

PostPosted: Tue Dec 08, 2009 5:31 pm    Post subject: Reply with quote

DigitalCorpus wrote:
Okay, so playing with Xfce4 a bit with rr_interval set to 18 (that is defautl for a quad core). Unfortunately the paradigm of more cores == less latency is not true. I' had 2, 4-thread mencoder encodes in the backgorund. One niced to 19 and the other at -10. I had to drop r_interval down to ~6 in order to not have a choppy display when dragging around windows et al. Kind of an extreme case, but if you choose seemless latency, keep the rr_interval in the kernel set to 6 or below. Keep in mind the post-boot rr_interval seems to be calculated as in kernel rr_interval * (# cpus's - 1 ). Just an FYI for you all.

I'm running this box mostly as a server. Now that may be the intended use for nice when it was created, but say I have something else in mind. I want to have a couple encodes going. One I want done but I dont' care when and I dont' want it to bother me. I know bumping it to 19 will accomplish that just fine. On ther other hand I had another set that I wanted to take priority over anything else I was doing, even if I called a bit on the CPU, which was bound to happen, even on my quad. I know that if I have any single taks going that -20 and 19 will still have the task execute in the same duration given the system is idle otherwise. from what I've read about nice levels, the mathematical explanation works out to a 5% drop in requested CPU usage with each bump of the nice level, sans schedulign overhead. I run mostly a server and I dont' mind if my GUI suffers a bit. It took me a year to put a GUI on here. If I still had full GUI interactivity and could determine what the lowest value of rr_interval is for the BFS scheduler for my preferences while preferenceign a groups of tasks (I load up to about 14 to 16 threads), then thats awesome imho. Every new task starts at 0. Why renice them to 10 to achive the scheduling priority I want when I can just make my top task -20 if it doesn't hurt the system? I could renice bash to 10, but when I'm done, I dont' want to have to remember to change it back. Thanks to BFS I can do this. CFS previously failed which is why you're so amazed probably.
Could you compare these results with those obtained by running encoders with SCHED_IDLEPRIO (schedtool -D) and SCHED_BATCH (schedtool -B) scheduling policies?
Back to top
View user's profile Send private message
Ant P.
Watchman
Watchman


Joined: 18 Apr 2009
Posts: 5713

PostPosted: Tue Dec 08, 2009 5:49 pm    Post subject: Reply with quote

What magic incantations do I need this time around to make R700 KMS work? It's giving me a plain VGA console right now.

I've got the staging drivers option for "KMS by default" turned on, the old-radeonfb driver turned off, all the normal DRM stuff enabled, and firmware is built into the kernel. I figure I'm doing something right since it's forcing the framebuffer console option to on...

Edit: ignore this, I'm an idiot. Had modeset=0 in my lilo config... :oops:


Last edited by Ant P. on Thu Dec 10, 2009 8:31 pm; edited 1 time in total
Back to top
View user's profile Send private message
dodo1122
Guru
Guru


Joined: 02 Sep 2006
Posts: 347
Location: York, England

PostPosted: Tue Dec 08, 2009 7:14 pm    Post subject: Reply with quote

Kingoftherings wrote:
dodo1122 wrote:
Bugs about unstable tree are being ignored at this moment because there is 2.6.32 in zen-stable.git already. zen.git is waiting for 2.6.33-rc1 to arrive, and until then is basically dead (-zen0 was just a simple bump to stable .32, not even a released kernel).
I made master-drmnouveau in zen-stable.git for people who want the new&shiny drm. Word of warning though: This new drm is broken for me with radeon kms (hangs when loading firmware, hopefully will be fixed though).


I've been wondering about the drm. There is a branch named drm in zen.git, but in zen-stable.git all I've seen is nouveau stuff. So is the radeon drm from airlied's git tree included in the drmnouveau branch?


master-drmnouveau has both nouveau and drm-next from airlied's git tree merged in. drm branch in zen.git just contained the drm on top of mainline (we were merging upstream to this branch, and then to master).
_________________
#zen-sources on irc.rizon.net
Back to top
View user's profile Send private message
cheater1034
Veteran
Veteran


Joined: 09 Sep 2004
Posts: 1558

PostPosted: Tue Dec 08, 2009 11:06 pm    Post subject: Reply with quote

Gibbo_07 wrote:
Your response is also pretty useless when you lack manners. Instead of labelling my post useless how about some basic compassion to people testing your kernel who have legitimate feedback for you? The point of my post was looking for other people with similar issues to let me know whether it's more likely my config or the kernel which was giving me grief and to give a basic description of the behaviour I have experienced - hoping it benefits yourself and others. The point was NOT to post my .config, make.conf and life story, unless YOU ASKED FOR IT, ie. you thought it could be helpful to you improving the kernel.

Well, I was going to be sarcastically polite in this response, but I won't.
Anyway, it is useless when I don't know the kernel version, I think that's obviously essential information - what did you want me to say, maybe "your post is totally useful except I don't know which kernel version you're running" - which is only half-true because your post wasn't totally useful without that information. I'm not out to offend you, I'm just stating a fact :\ , I'm not getting paid for user relations - I won't spend a bunch of time trying to find a more polite and manner-full way to say something, I put things bluntly (and i'm not dis-truthful about anything)

Quote:
I thought that would be obvious when stating 'shiny new' and that I used pappy's 2.6.32-zen0 seed as a base...

zen-stable.git is newer, and pappy didn't have a 2.6.32-zen1 seed so that's the reason why I didn't know which kernel version you were running.

Quote:
As for reiser4, yes I read the upstream problems and am now back on a .31 kernel until I see it resolved, at which point I either try to get zen working again or patch a vanilla kernel as per how I have been doing. I do however like the extra zen features I really want to put the BFS through it's paces hence why I bothered posting here with my issue. I only use reiser4 for my portage partition so data loss is not my concern what is however is speed and my learning so these are my reasons for bothering just suggesting I don't use it doesn't fix the problem for myself, nor your users (why bother including support for something you recommend against??)

zen-stable.git reiser4 is tracking upstream reiser4 - zen.git reiser4 is not, and it includes a "fix" that isn't the totally correct fix so it caused problems.
The 0 tag (ex: zen0) is used to indicate that it's a heavy development/unstable/broken version, when we get into actual numbers (like zen1 on zen-stable.git), then that means it's not the previous things. 2.6.32-zen0 is literally identical to 2.6.32-rc8-zen2, except it pulled in some updates. And zen.git has a lot of problems because of the philosophy that I oppose (I want a more stable unstable tree)

2.6.32 is a stable release too, it only fits that I bump the stable tree for it and only care about bug reports about that one, because 2.6.32-zen0 is the end of the life for the zen.git tree until 2.6.33-rc1

Quote:

As for suggesting we use the stable git tree - the whole point of this was to upgrade to .32 so using stable defeats that purpose. If you are ignoring bugs on the unstable tree then goodluck to you and the project :twisted:

2.6.32 is a stable release, unless something changed from every other kernel release in the 2.6 series.
We don't ignore bugs on the unstable tree, although I'd rather have them put on the bug tracker ;) - I don't care about the unstable tree, you need to get a hold of the other guys for that, and the best way to do that is bug tracker or mailing list. And the other guys would say, use 2.6.32-zen1 and not 2.6.32-zen0

------------
Maybe i'll merge drm-next into zen-stable, because it is pretty stable right now (more stable than non drm-next) - I just don't want to risk a problem, nor do I want to keep providing drm-next patches only because it's a hassle.
------------
kernelOfTruth wrote:
no - reiser4 is NOT dead ! :roll:
and it definitely will make it into mainline (next year), until an detailed article has been published in an linux magazine

it's not an insane filesystem in contrast to some other filesystems

there are filesystems out there that are more "insane" to call it in your terms than reiser4
e.g. ext4 which I still regard as "testing" material (even more testing than reiser4 ;) ) or btrfs (but that currently is under heavy development and very rapidly evolves and therefore has an excuse for its unstable somewhat problematic state in some situations) :P

I didn't mean it's dead, but reiser4 is a flat out complicated mess. It hasn't had any real updates in a couple of years, just some fixes and porting from release to release (which is taking longer and longer nowadays).

When i've used reiser4 in the past, it was very nice and quick - then it evolved into a slow, junky, self-corrupting volume (and it seemed to be the trend with a lot of people), reiser4 is always fast at first, but as it fills up it becomes very very slow. And it did/does have a data-eating issue that I'm not totally sure why it did/does it.

The last point is that ext4 is nothing at all like reiser4, ext4 is about finished - many more people may experience a problem with reiser4 (it is much huger than the other file systems, and does things differently - causing more room for error). ext4 is stable now and much more consistent than a reiser4 volume at this time.

For reiser4's inclusion it would have to prove some serious points to kernel developers, it doesn't help that it's not very actively maintained at all.

btrfs is "unstable" but it's probably at or beyond reiser4 in stability to this point - btrfs is not yet finished (and it may be dropped and zfs may get gpl license since oracle bought sun), although btrfs has a much bigger user-base now than reiser4 (mainline inclusion helps, massive feature set is good too, zfs promise also helps cuz zfs is so popular), anyway btrfs is more unstable due to a potentially still changing disk format, and it doesn't know when it's full yet (i don't think so anyway), but it's also much lighter than reiser4 so less prone to odd problems.
_________________
IRC!: #zen-sources on irc.rizon.net
zen-kernel.org
--
Lost in android development land.
Back to top
View user's profile Send private message
DigitalCorpus
Apprentice
Apprentice


Joined: 30 Jul 2007
Posts: 245

PostPosted: Wed Dec 09, 2009 1:58 am    Post subject: Reply with quote

Mr. Tao wrote:
DigitalCorpus wrote:
Okay, so playing with Xfce4 a bit with rr_interval set to 18 (that is defautl for a quad core). Unfortunately the paradigm of more cores == less latency is not true. I' had 2, 4-thread mencoder encodes in the backgorund. One niced to 19 and the other at -10. I had to drop r_interval down to ~6 in order to not have a choppy display when dragging around windows et al. Kind of an extreme case, but if you choose seemless latency, keep the rr_interval in the kernel set to 6 or below. Keep in mind the post-boot rr_interval seems to be calculated as in kernel rr_interval * (# cpus's - 1 ). Just an FYI for you all.

I'm running this box mostly as a server. Now that may be the intended use for nice when it was created, but say I have something else in mind. I want to have a couple encodes going. One I want done but I dont' care when and I dont' want it to bother me. I know bumping it to 19 will accomplish that just fine. On ther other hand I had another set that I wanted to take priority over anything else I was doing, even if I called a bit on the CPU, which was bound to happen, even on my quad. I know that if I have any single taks going that -20 and 19 will still have the task execute in the same duration given the system is idle otherwise. from what I've read about nice levels, the mathematical explanation works out to a 5% drop in requested CPU usage with each bump of the nice level, sans schedulign overhead. I run mostly a server and I dont' mind if my GUI suffers a bit. It took me a year to put a GUI on here. If I still had full GUI interactivity and could determine what the lowest value of rr_interval is for the BFS scheduler for my preferences while preferenceign a groups of tasks (I load up to about 14 to 16 threads), then thats awesome imho. Every new task starts at 0. Why renice them to 10 to achive the scheduling priority I want when I can just make my top task -20 if it doesn't hurt the system? I could renice bash to 10, but when I'm done, I dont' want to have to remember to change it back. Thanks to BFS I can do this. CFS previously failed which is why you're so amazed probably.
Could you compare these results with those obtained by running encoders with SCHED_IDLEPRIO (schedtool -D) and SCHED_BATCH (schedtool -B) scheduling policies?

Can you be more specific? Are you just looking at encoding times of a single video under both?
_________________
Atlas (HDTV PVR, HTTP & Media server)
http://atlas.selfip.net/Info/
Back to top
View user's profile Send private message
Mr. Tao
Tux's lil' helper
Tux's lil' helper


Joined: 20 Jul 2007
Posts: 147

PostPosted: Wed Dec 09, 2009 10:05 am    Post subject: Reply with quote

DigitalCorpus wrote:
Can you be more specific? Are you just looking at encoding times of a single video under both?
I'm interested in results of your experiment with regard to encoding time and perception of display choppiness when you use different scheduling policy instead of renicing your encoding processes. SCHED_BATCH (or perhaps SCHED_ISO) for process you want to prioritize and SCHED_IDLEPRIO for one which you do not care when finishes.
Back to top
View user's profile Send private message
DigitalCorpus
Apprentice
Apprentice


Joined: 30 Jul 2007
Posts: 245

PostPosted: Wed Dec 09, 2009 11:30 am    Post subject: Reply with quote

well since I've been offline for about half a week and I still have a bunch of higher resolution ot encode, I'll be more than happy to oblige. Shall I PM you to determine what constraints to go forth with this so as not to fill up this thread with OT conversation? or will this be anecdotal, though still technically OT?
_________________
Atlas (HDTV PVR, HTTP & Media server)
http://atlas.selfip.net/Info/
Back to top
View user's profile Send private message
cheater1034
Veteran
Veteran


Joined: 09 Sep 2004
Posts: 1558

PostPosted: Thu Dec 10, 2009 11:28 pm    Post subject: Reply with quote

2.6.32-ck1 is now in zen-stable master ;)

Here's what it features (series)
Code:
# 2.6.32-sched-bfs-311.patch (bfs is found in zen-sched)
sched-add-above-background-load-function.patch (adapted for cfs too)
cfq-iso_ionice.patch
mm-make_swappiness_really_mean_it.patch
mm-kswapd_inherit_prio.patch
mm-background_scan.patch
mm-idleprio_prio.patch (bfs only)
mm-lru_cache_add_lru_tail.patch
kconfig-expose_vmsplit_option.patch
hz-default_1000.patch
hz-no_default_250.patch
hz-raise_max.patch
preempt-desktop-tune.patch
#ck-version.patch


I think most will be very happy ;)

Also updated 2.6.31, and will soon make the final 2.6.30 release* - may add the ck patches to master-2.6.31 as well.
_________________
IRC!: #zen-sources on irc.rizon.net
zen-kernel.org
--
Lost in android development land.
Back to top
View user's profile Send private message
DigitalCorpus
Apprentice
Apprentice


Joined: 30 Jul 2007
Posts: 245

PostPosted: Fri Dec 11, 2009 12:36 am    Post subject: Reply with quote

So when 2.6.32 makes it to portage, all of those patches will be applied? Any way, other than git, can a user choose to avoid some of the patches? For just myself, I pick and choose what I want. I don't want 1000MHz, but I want BFS which gives me enough low latency and even without preemption. If 1000Hz and preemption truely do hurt throughput, even 10%, I wouldn't ever be updating my kernel from zen to keep that throughput. Just my 2 cents. Please don't get me wrong, Zen-Sources is awesome, but I think it would be wise to allow a little configurability over ricing the kernel for latency.
_________________
Atlas (HDTV PVR, HTTP & Media server)
http://atlas.selfip.net/Info/
Back to top
View user's profile Send private message
bollucks
l33t
l33t


Joined: 27 Oct 2004
Posts: 606

PostPosted: Fri Dec 11, 2009 12:47 am    Post subject: Reply with quote

DigitalCorpus wrote:
So when 2.6.32 makes it to portage, all of those patches will be applied? Any way, other than git, can a user choose to avoid some of the patches? For just myself, I pick and choose what I want. I don't want 1000MHz, but I want BFS which gives me enough low latency and even without preemption. If 1000Hz and preemption truely do hurt throughput, even 10%, I wouldn't ever be updating my kernel from zen to keep that throughput. Just my 2 cents. Please don't get me wrong, Zen-Sources is awesome, but I think it would be wise to allow a little configurability over ricing the kernel for latency.

The -ck patches don't make it impossible to change your settings, they just change what the defaults are. You can configure it back to whatever you want.
Back to top
View user's profile Send private message
DigitalCorpus
Apprentice
Apprentice


Joined: 30 Jul 2007
Posts: 245

PostPosted: Fri Dec 11, 2009 12:51 am    Post subject: Reply with quote

bollucks wrote:
The -ck patches don't make it impossible to change your settings, they just change what the defaults are. You can configure it back to whatever you want.

Thanks for the enlightenment
_________________
Atlas (HDTV PVR, HTTP & Media server)
http://atlas.selfip.net/Info/
Back to top
View user's profile Send private message
kernelOfTruth
Watchman
Watchman


Joined: 20 Dec 2005
Posts: 6108
Location: Vienna, Austria; Germany; hello world :)

PostPosted: Fri Dec 11, 2009 1:18 am    Post subject: Reply with quote

cheater1034 wrote:
2.6.32-ck1 is now in zen-stable master ;)

[snip] list of included patches [snip]

I think most will be very happy ;)

Also updated 2.6.31, and will soon make the final 2.6.30 release* - may add the ck patches to master-2.6.31 as well.


Yay !

Yet another reason to use and love the zen-kernel sources :D

Thanks a lot cheater !
_________________
https://github.com/kernelOfTruth/ZFS-for-SystemRescueCD/tree/ZFS-for-SysRescCD-4.9.0
https://github.com/kernelOfTruth/pulseaudio-equalizer-ladspa

Hardcore Gentoo Linux user since 2004 :D
Back to top
View user's profile Send private message
keenblade
Veteran
Veteran


Joined: 03 Oct 2004
Posts: 1087

PostPosted: Fri Dec 11, 2009 3:18 am    Post subject: Reply with quote

kernelOfTruth wrote:
Yet another reason to use and love the zen-kernel sources :D

Thanks a lot cheater !

+
_________________
Anyway it's all the same at the end...
Need help to get it working: "x-fi surround 5.1"
Back to top
View user's profile Send private message
tranquilcool
Veteran
Veteran


Joined: 25 Mar 2005
Posts: 1159

PostPosted: Fri Dec 11, 2009 6:39 am    Post subject: Reply with quote

cheater1034 wrote:
2.6.32-ck1 is now in zen-stable master ;)

Here's what it features (series)
Code:
# 2.6.32-sched-bfs-311.patch (bfs is found in zen-sched)
sched-add-above-background-load-function.patch (adapted for cfs too)
cfq-iso_ionice.patch
mm-make_swappiness_really_mean_it.patch
mm-kswapd_inherit_prio.patch
mm-background_scan.patch
mm-idleprio_prio.patch (bfs only)
mm-lru_cache_add_lru_tail.patch
kconfig-expose_vmsplit_option.patch
hz-default_1000.patch
hz-no_default_250.patch
hz-raise_max.patch
preempt-desktop-tune.patch
#ck-version.patch


I think most will be very happy ;)

Also updated 2.6.31, and will soon make the final 2.6.30 release* - may add the ck patches to master-2.6.31 as well.


how do i checkout v2.6.32-ck1?
i though i could manage the git commands but it's obviously not so.
_________________
this is a strange strange world.
Back to top
View user's profile Send private message
ponciarello
Apprentice
Apprentice


Joined: 22 Jul 2008
Posts: 223
Location: beach of slack

PostPosted: Fri Dec 11, 2009 7:07 am    Post subject: Reply with quote

i think you have to checkout origin/master, I don't know if zen's version of -ck1 works outside of zen itself: I think it's modified to fit in zen code, in specific the zen scheduler.
there was something about in kolivas' web repo (a 2.6.32-ck1-test patch), but it has moved. :(

now I'm gonna play to see on which CONFIG_HZ my drivers break :lol:
tnx for the fun zen's people :)

EDIT: with 2000Hz works :D
EDIT-again: sorry, just to report my notebook has not exploded yet but everything is very responsive. :o don't know if i'm gonna try 10000, I'm scared :lol:


Last edited by ponciarello on Fri Dec 11, 2009 12:39 pm; edited 5 times in total
Back to top
View user's profile Send private message
Display posts from previous:   
This topic is locked: you cannot edit posts or make replies.    Gentoo Forums Forum Index Unsupported Software All times are GMT
Goto page Previous  1, 2, 3, 4 ... 12, 13, 14  Next
Page 3 of 14

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum