Anyone else getting reduced performance with kernel 2.6.22?

lloeki · 2007-09-03 13:03:15

bumping in to say 'me too'. with all the things I read here and there (CFS, CK and stuff...), I wanted to wait for things to cool down and get the feeling of it myself. after some time in .22 I can clearly say that things are worse than sometimes ago. why for the sake of hell does my mouse STOP MOVING and then warps from x,y to x+42,y-42 on any split-half-consistent disk access???? this sure never happened before (2.6.18 for sure, maybe even 2.6.20), and this is utterly ridiculous. I feel like being back in some woefully MS product. I also was very proud of my boot time (laptop, <15s) but it seems to have sunk into the depths of a grinding hd hell, now closer to 1min.

PS: and no, all this has nothing to do with potential fragmentation or crufting or whatever, since I started afresh not so long ago.

sH · 2007-09-03 13:14:09

I can second sporadical freezes of the box with the stock kernel. Theres no chance to get access to the box besides using the Reset-Button.

lloeki · 2007-09-06 14:59:23

update: since I wrote this post up there, I tried nohz=off highres=off and it results in a much swifter system in all areas. on the downside, I lose 30~45min of battery runtime (certainly because of nohz being disabled). I will now try with either of those options (but not both) and see if one or the other is the culprit, and if I can reclaim some more battery time while still retaining responsiveness.

Last edited by lloeki (2007-09-06 15:00:55)

broch · 2007-09-06 17:32:09

I have read that 2.6.22 has some sort of I/O bug that will be fixed in 2.6.23. If bug is real then either users should go back to 2.6.21 (?) or wait for fix in 2.6.23.
I am not sure though why this potential bug can't be fixed in 2.6.22.x

skottish · 2007-09-06 21:14:06

broch wrote:

I am not sure though why this potential bug can't be fixed in 2.6.22.x

2.6.23 is at RC 5. I'd wait instead of dealing with rolling back.

Shaika-Dzari · 2007-09-06 22:18:36

Hello,

I got higher temperature with my laptop and 2.6.22:
http://bbs.archlinux.org/viewtopic.php?id=36648

I downgraded to 2.6.21 to be sure my laptop wont suffer any damage...

@+

lloeki · 2007-09-07 08:40:27

it seems to be running fine with only nohz=off. I briefly tried with only highres=off, but I had some issues (bluetooth devices (keyboard+mouse) failed to connect ?? they worked instantly again when re-adding nohz=off, or with both on). so for now I'm wth high res timers, but without nohz. oh well. at least it's useable now. will wait for .23

Last edited by lloeki (2007-09-07 08:41:27)

broch · 2007-09-07 13:51:22

question regarding ata_generic module
1) removing from mkinitcpio.conf will not prevent loading ata_generic module
2) removing from kernel works without any side effects (intel SATA), however I don't see any positive impact on disk performance.

Arch Linux

#26 2007-09-03 13:03:15

Re: Anyone else getting reduced performance with kernel 2.6.22?

#27 2007-09-03 13:14:09

Re: Anyone else getting reduced performance with kernel 2.6.22?

#28 2007-09-06 14:59:23

Re: Anyone else getting reduced performance with kernel 2.6.22?

#29 2007-09-06 17:32:09

Re: Anyone else getting reduced performance with kernel 2.6.22?

#30 2007-09-06 21:14:06

Re: Anyone else getting reduced performance with kernel 2.6.22?

#31 2007-09-06 22:18:36

Re: Anyone else getting reduced performance with kernel 2.6.22?

#32 2007-09-07 08:40:27

Re: Anyone else getting reduced performance with kernel 2.6.22?

#33 2007-09-07 13:51:22

Re: Anyone else getting reduced performance with kernel 2.6.22?

Board footer