Re: [RFC V2 PATCH 2/5] cfq-iosched: preparation to handle multiple service trees

[Corrado Zoccolo <czoccolo@gmail.com>]

Hi Jeff,
On Wed, Oct 21, 2009 at 5:31 PM, Jeff Moyer <jmoyer@redhat.com> wrote:
> Corrado Zoccolo <czoccolo@gmail.com> writes:
>
>> We embed a pointer to the service tree in each queue, to handle multiple
>> service trees easily.
>> Service trees are enriched with a counter.
>> cfq_add_rq_rb is invoked after putting the rq in the fifo, to ensure
>> that all fields in rq are properly initialized.
>>
>> @@ -115,6 +116,9 @@ struct cfq_queue {
>>       unsigned short ioprio_class, org_ioprio_class;
>>
>>       pid_t pid;
>> +
>> +     struct cfq_rb_root *service_tree;
>> +     struct cfq_io_context *cic;
>>  };
>
> This change worries me.  For async queues, there is not a 1:1 mapping
> between cfqq and cic.

Currently, I need it only for sync queues. And it is used only to
determine if a queue is seeky, so if we can mark it in the queue
instead of the cic, this will not be needed any more.

>  With my forthcoming close cooperator changes, the
> same will be true for sync queues.
Yes, I saw it. This is currently just an RFC. If your changes got
merged before, I'll devise an alternative solution to this problem.

Corrado

> Cheers,
> Jeff
>



-- 
__________________________________________________________________________

dott. Corrado Zoccolo                          mailto:czoccolo@gmail.com
PhD - Department of Computer Science - University of Pisa, Italy
--------------------------------------------------------------------------
The self-confidence of a warrior is not the self-confidence of the average
man. The average man seeks certainty in the eyes of the onlooker and calls
that self-confidence. The warrior seeks impeccability in his own eyes and
calls that humbleness.
                               Tales of Power - C. Castaneda