Re: [PATCH] examples: uring-cmd-zoned: expand the reasoning behind QD1

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Damien,
> > +# NOTE:
> > +# Regular writes against a zone should be limited to QD1, as the device can
> > +# reorder the requests.
> > +#
> > +# As the passthrough path do not use an IO scheduler (such as mq-deadline),
> > +# the queue depth should be limited to 1 to avoid zone invalid writes.
> 
> Since all the writes issued by fio are "valid" in the first place, I would
> have phrased this: "...to avoid reordering of sequential write commands in
> the IO stack."
> 
Block IO stack does not reorder the requests with uring-cmd. As the spec 
(at least ZNS) does not guarantee the order of execution even if we submit
it in order, I expanded the comment here about device reordering the
requests.

As you said, writes issues by fio are valid, and IO stack submits the
requests in order but still we can get a WP violation because the device
can reorder the request during execution.
-- 
Pankaj



[Index of Archives]     [Linux Kernel]     [Linux SCSI]     [Linux IDE]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux SCSI]

  Powered by Linux