On Fri, 2004-04-02 at 11:55, Christoph Hellwig wrote:
> On Fri, Apr 02, 2004 at 11:44:47AM +0200, Bart Samwel wrote:
> > The maximums that are listed for the regular flush_interval and
> > age_buffer are way too low for laptop mode usage. I've considered
> > increasing the maximums, but I figured they would be that low for a
> > reason. By doing it this way, the maximums for regular usage aren't
> > affected. If everybody thinks it's OK to simply increase the maximums on
> > the regular ones to INT_MAX (or at least something like HZ*7200), I'll
> > _gladly_ do that instead.
> There's no real problem with large intervals (except the possible data
> loss on a crash), so I'd say enlarge them.
I've done so in the patches that I've just posted to the list, I've put
them at HZ*7200.
These patches also fix another problem with the earlier patches -- when
I split up the patches, some of the #includes had ended up in the wrong