the drive is free to reorder requests to its hearts content (assuming it has
enough).  This would include cached & non-cached requests so ideally FUA
shouldn't really have an effect on tagged-queueing.

With FUA set on the drive won't be able to simply complete requests once
it's dumped their data into the write cache.  The request completion will
have to wait until the data has actually made it out of the cache and onto
disk.  The same should be true whether the request is tagged or untagged
however.

i don't see any benefits for tagged requests than i see for untagged
requests (in relation to caching).  In fact with caching disabled your
requests will complete slower so under heavy load you're more likely to
start getting back QUEUE_FULL or BUSY conditions.  Those are pretty easy to
deal with though and are probably worth the extra speed you would get by
doing selective write-through operations.

-p


-----Original Message-----
From: Manuel Bouyer [mailto:bouyer@antioche.lip6.fr]
Sent: Friday, June 12, 1998 1:46 AM
To: Peter Wieland
Cc: 'Matthew Jacob'; thorpej@nas.nasa.gov; tech-kern@NetBSD.ORG
Subject: Re: test of new powerdown facility


On Jun 11, Peter Wieland wrote
> Of course that requires crawling through the file systems and
> deciding which i/o's have to be on the disk immediately (metadata) and
which
> can go on the disk later (user data depending on what the client wants)
and
> requires a way for the fs to communicate this need to the disk drivers on
an
> i/o by i/o basis.
> 

Would'nt SCSI Tagged Command Queuing benefit from this too (I seem to
remember
that the disk can reorder the commands if he wants) ?

--
Manuel Bouyer, LIP6, Universite Paris VI.           Manuel.Bouyer@lip6.fr
--