Vladimir V. Saveliev wrote (ao):
> Sander wrote:
> > Vladimir V. Saveliev wrote (ao):
> >>No. This is not a fix. This is a patch which I suspect is responsible for 
> >>sync
> >>performance drop. It fixes old reiser4 bug known as "not updating atime".
> >>I asked you to reverse this patch to check whether this patch caused recent 
> >>sync
> >> performance degradation.
> > 
> > Ah!
> > 
> > Now, if I remount my Reiser4 filesystem noatime,nodiratime, the problem
> > is gone.
> > 
> > Does that help?
> > 
> > Do you still want me to test a kernel with your patch?
> 
> Probably not, but
> there were recently several reports about reiser4 sync performance 
> degradation.
> It would be great to hear that mounting noatime or unapplying that patch fixes
> all problmes, including the one with evolution's field resizing and vim's 
> ":wq".

Well, for me the mount noatime fixes it for sure :-)

Weird thing though, my other system also has the default mounts (no
noatime or nodiratime) and doesn't suffer from the slowdown.

Also, I don't have Evolution, but the difference in vim (and thus mutt,
and btw, Postfix too) is _huge_.

-- 
Humilis IT Services and Solutions
http://www.humilis.net

Reply via email to