Re: [PATCH/RFC 0/4] cfq: implement merging and breaking up of cfq_queues

[Jeff Moyer <jmoyer@redhat.com>]

Jens Axboe <jens.axboe@oracle.com> writes:

> On Mon, Oct 26 2009, Jeff Moyer wrote:
>> Corrado Zoccolo <czoccolo@gmail.com> writes:
>> 
>> > Hi Jeff,
>> > this series looks good.
>> 
>> Hi, Corrado.  Thanks again for the review!
>> 
>> > I like in particular the fact that you move seekiness detection in the cfqq.
>> > This can help with processes that issue sequential reads and seeky
>> > writes, or vice versa.
>> > Probably, also the think time could be made per-cfqq, so that the
>> > decision whether we should idle for a given cfqq is more precise.
>> 
>> I'll have to think about that one.  It would be good to know Jens'
>> opinion on the matter, too.
>
> Your implementation looks fine, as usual I'm mostly worried about
> performance impact and suitability (I hate having to work around
> issues). But the win is so large in some cases that we should just go
> ahead and merge it for .33, so I'll queue it up.

Great, thanks for the review.  In this case, however, I was wondering
what your opinion was about moving the think time calculation to be per
cfqq.  ;-)

> It would be nice to fix the in-kernel problem with NFS, since that is
> doable.

I'll see if I can get someone motivated to work on that.  I'm not sure
that I can devote much time to the issue myself, unfortunately.

Cheers,
Jeff